Proceedings of the Thirteenth Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst

Material Information

Proceedings of the Thirteenth Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst
Series Title:
NCKRI Symposia
Added title page title:
Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst
Added title page title:
NCKRI Symposia 2
Land, Lewis
Doctor, Daniel H.
Stephenson, J. Brad
National Cave and Karst Research Institute
University of South Florida
Publication Date:


Subjects / Keywords:
Environment ( local )
Conference Proceeding
serial ( sobekcm )


These proceedings represent the talks, posters, and symposia presented at the 13th Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst, which took place in Carlsbad, New Mexico, at the NCKRI headquarters, May 6-10, 2013. This international conference series creates a better understanding of environmental issues and geohazards associated with karst environments. This 480 page volume contains 52 peer-reviewed papers organized under the following headings: a) Engineering and Geotechnical Aspects of Karst, b) Evaporite Karst, c) Geophysical Investigations in Karst Terrain, d) Formation Processes of Karst and Sinkholes, e) Karst Hydrology, and f) Mapping and Management of Karst Regions.
Open Access - Permission by Publisher
General Note:
See Extended description for more information.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
K26-03297 ( USFLDC DOI )
k26.3297 ( USFLDC Handle )
11735 ( karstportal - original NodeID )

USFLDC Membership

Karst Information Portal

Postcard Information



This item has the following downloads:

Full Text




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`$@9$M1(52#1)L$;$D--5.+1'# -H$>0.AH$1-$.5-/50-52H$10,$'"#$T2:10+]+0:$4 5==+''##$ 2#=1+0$10$+0'#:21*$/12'$56$'"#$.506#2#0.#A $ $ !"#$/25.##,+0:-$56$'"#-#$.506#2#0.#-$"1<#$E##0$<1*)1E*#$1,,+'+50-$'5$812-'$*+E212+#-$125)0,$'"#$C52*,A$ F#* 5C$ +$ 1$*+-'$56$'"#$/25.##,+0:-$625=$ '"#$E#:+00+0:$.506#2#0.#$'5$B_%% $ C"+."$,#'1+*-$'"#$'5/+.-$.5<#2#,$10,$'"#$ -/50-52-$56$#1."a$ $ $ T1%W4'P,& )'*(%1%*.% ,&"*J& T1'.%%J4*G, & ]D ,+ I&DBX6I&F1$"*J'I&<0^& & ?4*=@'$%,_&!@%41&V%'$'G5I&Z*G4*%%14*G&"*J&Z*W41'*/%*+"$&7/A".+ & @25.##,+0:-$56$'"#$S+2-'$()*'+,+-.+/*+0123$4506#2#0.#$50$7+08"5*#-$10,$'"#$90:+0##2+0:$10,$90<+250=#0'1*$>=/1.'-$56$?12-'H$T2 *10,5H$ S*52+,1H$T.'5E#2$ %U V %WH$%XYRH$9,+'#,$E3$F1223$SA$F#.8`$7/50-52#,$E3$S*52+,1$7+08"5*#$P#-#12."$>0-'+')'#H$45**#:#$56$90:+0##2+0:H$ Z0+<#2-+'3$56$4#0'21*$S*52+,1A $ $ ]S *J I&DBX>I&F1$"*J'I&<0^ & N"1,+&35J1'G%'$'G5_&Z*G4*%%14*G&"*J&Z*W41'*/%*+"$&YAA$4."+4'*, & @25.##,+0:-$56$'"#$7#.50 ,$()*'+,+-.+/*+0123$4506#2#0.#$50$7+08"5*#-$10,$'"#$90:+0##2+0:$10,$90<+250=#0'1*$>= /1.'-$56$?12-'H$ T2*10,5H$S*52+,1 H$S#EA $ X V %%H$%XYW`$9,+'#,$E3$F1223$SA$F#.8$;$b+**+1=$MA$b+*-50H$D--+-'#,$E3$M1)21$S#*,=10H$7"10050$c53.#H$?1'"3$ (.K501*,$;$7"12250$(+8#-#**` $ 7/50-52#,$E3$S*52+,1$7+08"5*#$P#-#12."$>0-'+' )'#H$45**#:#$56$90:+0##2+0:H$Z0+=/1.'-$56$?12-'H$7'A$ @#'#2-E)2:$F#1."H$S*52+,1$T.'5E#2$B V RH$%XYX$9,+'#,$E3$F1223$SA$F#.8H$ D--+-'#,$E3$D,2+100#$G1:#0H$7.5''$41<+0H $ F2+10$F126)-$;$d+2:+0+1$ (#28*#`$ 7/50-52#,$E3$ $ S*52+,1$7+08"5*#$P#-#12."$> 0-'+')'#H$K+<+-+50$56$7/50-52#,$P#-#12."H$ 10,$ Z0+<#2-+'3$56$4#0'21*$S*52+,1 A $


! $ !"#$% & '" $ ()*'+,+-.+/*+0123$4506#2#0.#$50$7+08"5*#-$10,$'"#$ $ 90:+0##2+0:$;$90<+250=#0'1*$>=/1.'-$56$?12-' @1:#$ U $ $ ]6 +@ I&DBB2I&T"*"/"&)4+5I&<0^ $ YAA$4%J&N"1,+&V%'$'G5 & @25.##,+0:-$56$'"#$S5)2'"$()*'+,+-.+ /*+0123$4506#2#0.#$50$7+08"5*#-$ 10,$'"#$90:+0##2+0:$10,$90<+250=#0'1*$>=/1.'-$56$?12-'H$ @101 =1$4+' 3H$S*52+,1H$c10)123$BU V BWH$%XX&`$ 9,+'#,$E3$F1223$SA$F#.8 `$ 7/50-52#,$E3 $ Z0+<#2-+'3$56$4#0'21*$S*52+,1H $ K+<+-+50$56$7/50-52#,$ P#-#12."H$10,$ @A9A$M1(52#1)L$;$D--5.+1'#-H$>0.A $ $ ]Q +@ I&DBBQI&V"+$4*#P1GI&!K^ & N"1,+&V%'3"H"1J,_&Z*G4*%%14*G&"*J&Z*W41'*/%*+"$&T 1'#$%/,&4*&N"1,+&!%11"*%, & @25.##,+0:-$56$'"#$S+6'"$()*'+,+-.+/*+0123$4506#2#0.#$50$7+08"5*#-$10,$'"#$90:+0##2+0:$10,$90<+250=#0'1*$>=/1.'-$56$?12-'H$ ^1'*+0E)2:H$!#00#--##H$D/2+*$B V UH$%XXU `$9,+'#,$E3$F1223$SA$F#.8H$ D--+-'#,$E3$S#*+.+'3$(A$@#12-50 `$7/50-52#, $ E3$ @A9A$M1(52#1)L$;$ D--5.+1'#-H$>0.A H$ O1 '+501*$^25)0,$b1'#2$D--5.+1'+50H$ D=#2+.10$75.+#'3$56$4+<+ *$90:+0##2-$N!#00#--##$7#.'+50Q`$ Z0+<#2-+'3$56$ !#00#-##$>0-'+')'#$652$^#5'#."05*5:3H$10,$ ?12-'$b1'#2-$>0-'+')'# $ $ ]R +@ I&DBB>I&?A14*G(4%$JI&8F^ & !@%&Z*G4*%%14*G&V%'$'G5&"*J&35J1'G%'$'G5&'(&N"1,+&!%11"*%, & @25.##,+0:-$56$'"#$7+L'"$()*'+,+-.+/*+0123$4506#2#0.#$50$7+08"5*#-$10,$'"#$90:+0##2+0:$10,$90<+250=#0'1*$>=/1.'-$56$?12-'H$ 7/2+0:6 +#*,H$(+--5)2+H$D/2+*$e V XH$%XXW`$ 9,+'#,$E3$F12 23$SA$F#.8$;$cA$F21,$ 7'#/"#0-50H$D--+-'#,$ E3$cA$^13*#$G#22+0: `$ 7/50-52#,$E3 $ @A9A$ M1(52#1)L$;$D--5.+1'#-H$>0.A H$ 75)'"C#-'$(+--5)2+$7'1'#$Z0+<#2-+'3H$K#/12'=#0'$5 6$^#5:21/"3H$^#5*5:3$;$@*100+0:H$ D=#2+.10$ 75.+#'3$56$4+<+*$90:+0##2-H$(+, V (+--5)2+$10,$7'A$M5)+-$7#.'+50-H$ O1'+501*$ ^25)0,$b1'#2$D--5.+1'+50H$D--5.+1'+50$56$^25)0, $ b1'#2$ 7.+#0'+-'-$10,$90:+0##2-H$ D--5.+ 1'+50$56$90:+0##2+0:$^#5*5:+-'-H$ 10,$ ?12-'$b1'#2-$>0-'+')'# A $ $ ]> +@ I&DBBBI&3"114,#P1G`3%1,@%5I& TY^ & 35J1'G%'$'G5&"*J&Z*G4*%%14*G&V%'$'G5&'(&?4*=@'$%,&"*J&N"1,+ DBBB & @25.##, +0:-$56$'"#$7#<#0'"$()*'+,+-.+ /*+0123$4506#2#0.#$50$7+08"5*#-$ 10,$'"#$90:+0##2+0:$10,$90<+250=#0'1*$>=/1.'-$56$?12-'H$ G122+-E)2:IG#2-"#3H$ @#00-3*<10+1H$D/2+*$%_ V %RH$%XXX`$ 9,+'#,$E3$F1223$SA$F#.8H$D2'")2$ cA$@#''+'H$10,$cA$^13*#$G#22+0:`$ 7/50-52#,$E3 $ @A9A$ M1 (52#1)L$;$D--5.+1'#-H$>0.A H$ ZA7 A$K#/12'=#0'$56$!210-/52'1'+50HH$ S#,#21*$G+:"C13$D,=+0+-'21'+50 H$ ZA7A$90<+250=#0'1*$@25'#.'+50$ D:#0.3 H$@#00-3*<10+1$^#5*5:+.1*$7)2<#3H$ ^#5 V >0-'+')'#$56$'"#$D=#2+.10$75.+# '3$56$4+<+*$90:+0##2-H$ D--5.+ 1'+50$56$90:+0##2+0:$ ^#5*5 :+-'-H$ O1 '+501*$^25)0,$b1'#2$D--5.+1'+50H$ d+2:+0+1$b1' #2$P#-5)2.#-$P#-#12."$>0-'+')'#H$ (+, V D'*10'+.$?12-'$450-52'+)= H$10,$ ?12-'$b1'#2-$>0-'+')'# A $ $ ]X +@ I&SEEDI&0'P4,W4$$%I&Na^ & V%'+%.@*4."$&"*J&Z*W41'*/%*+"$&YAA$4."+4'*,&'(&N"1,+&V%'$'G5&"*J&35J1'$'G5 SEED & @25.##,+0:-$56$'"#$9+:"'"$()*'+,+-.+ /*+0123$4506#2#0.#$50$7+08"5*#-$ 10,$'"#$90:+0##2+0:$10,$90<+250=#0'1*$>=/1.'-$56$?12-'H$ M5)+-< +**#H$?#0').83H$D/2+*$% V RH$B__%`$ 9,+'#,$E3$F12 23$SA$F#.8$10,$cA$^13*#$G#22+0:`$ 7/50-52#,$E3 $ @A9A$M1(52#1)L$;$D--5.+1'#-H$>0.A H $ ^#5 V >0-'+')'#$56$'"#$D=#2 + .10$75.+#'3$56$4+<+*$90:+0##2-H$ D--5.+1'+50$56$^25)0,$b1'#2$7.+#0'+-'-$10,$90:+0##2-$56$'"#$O1'+501*$ ^25)0,$b1'#2$D--5.+1'+50 $ $ ]B +@ I&SEE2I&3P*+,W4$$%I&Y0^ & V%'+%.@*4."$&?A%.4"$&TP#$4."+4'*&K':&DSS & @25.##,+0:-$56$'"#$O+0'"$()*'+,+.+/*+0123$4506#2#0.#$50$7+08"5*#-$10,$'"#$90:+0##2+0:$10,$90<+250=#0'1*$>=/1.'-$56$?12-'H$ G)0'-<+* *#H$D*1E1=1 H$7#/'$e V %_H$B__&`$ 9,+'#,$E3$F1223$SA$F#.8 `$ 7/50-52#,$E3 H$^#5 V >0-'+')'#$56$'"#$ D=#2 +.10$75.+#'3$56$4+<+*$90:+0##2-H$ !"#$O1 '+501*$^25)0,$b1'#2$D--5. +1'+50H$10,$ @A9A$M1(52#1)L$;$D--5.+1'#-H$>0.A $ $ ]DE +@ I&SEEQI&?"*&Y*+'*4'I&!b^ & V%'+%.@*4."$&?A%.4"$&TP#$4."+4'*&K':&D66 & @25.##,+0:-$56$'"#$!#0'"$()*'+,+-.+/*+0123$4506#2#0.#$50$7+08"5*#-$10,$'"#$90:+0##2+0:$10,$90<+250=#0'1*$>=/1.'-$56$?12-'H$71 0$ D0'50+5H$!# L1H$7#/'$BR V BYH$B__U`$ 9,+'#,$E3$F1223$SA$F#.8 `$ 7/50-52#,$E3 H$ ^#5 V >0-'+')'#$56$'"#$D=#2 +.10$75.+#'3$56$4+<+*$90:+0##2-H$ !"#$9,C12,-$DJ)+6#2$D)'"52+'3 H$ @A9A$M1(52#1)L$;$D--5.+1'#-H$>0.A H$45 V 7/50-52#,$E3$ !"#$75)'"C#-'$P#-#12."$>0-'+')'# $


! $ !"#$% & '" $ ()*'+,+-.+/*+0123$4506#2#0.#$50$7+08"5*#-$10,$'"#$ $ 90:+0##2+0:$;$90<+250=#0'1*$>=/1.'-$56$?12-' @1:#$ e $ $ ]DD +@ I&SEEXI& !"$$"@",,%%I&<0^ & V%'+%.@*4."$&?A%.4"$&TP#$4."+4'*&K':&DX2 & @25.##,+0:-$56$'"#$9*#<#0'"$()*'+,+-.+/*+0123$4506#2#0.#$50$7+08"5*#-$10,$'"#$90:+0##2+0:$10,$90<+250=#0'1*$>=/1.'-$56$?12-'H $ !1 **1"1--##H$S*52+,1 H$7#/'$BB V BeH$B__Y`$ 9,+'#,$E3$M300$FA$f)"2H$9A$ 41*<+ 0$D*#L10,#2H$F1223$SA$F#.8`$ 7/50-52#,$E3 $ ^#5 V >0-'+')'#$56$'"#$ D=#2+.10$75.+#'3$56$4+<+*$90:+0##2 -A $ $ ]DS +@ I&SEDDI&?+:&0'P4,I&8F^ $ 3$*4,.$2"%&$.-&5)$6,*(2"% & W'$P/%&S>I&*',:&S U 2 $ @25.##,+0:-$56$'"#$!C#*6'"$()*'+,+-.+/*+0123$4506#2#0.#$50$7+08"5*#-$10,$'"#$90:+0 ##2+0:$10,$90<+250=#0'1* $ >=/1.'-$56$?12-'H$7'A$ M5)+-H$(+--5)2+ H$c10)123$%_ V %RH$B_%%`$9,+'#,$E3$M300$f)"2$10,$c1=#-$?1)6=10`$7/50-52#,$E3$@A9A$M1(52#1)L$10,$D--5.+1'#-A $ $ $ $$ $$$$$ $$$ $ %%%%%%%%% K-+&#%B#-?%,A&%4#-'&&2)*@0%-B%,A&% M *2 % '-*B&#&*'&8 % % % % %%% %%% K-+&#%B#-?%,A&%4#-'&&2)*@0 % 4"#,%9 % -B%,A&%9 M ,A % '-*B&#&*'&8


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d+.# V @2#-+,#0'$652$?12-'$10,$4"+#6$56$T/#21'+50-$652$'"#$T18$P+,:#H$ !#00#--##H$566+.#$56$@A9A$M1(52#1)L$;$D--5.+1'#-H$>0.A$N@9MDQA $ $ >0$B__XH$1$-'258#$+0'#22)/'#,$"+-$/*100+0:$652$'"#$%B'"$.506#2#0.#$+0$ c10)123$B_%%H$10,$=#,+.1*$.5=/*+.1'+50-$625=$'"1'$-'258#$)*'+=1'#*3$ .*1+=#,$"+-$*+6#$+0$O5<#=E#2$B_%%A $$ D-$/12'$56$'"#$%& '" $ .506#2#0.#H$1$ -/#.+1*$=#=52+1*$-#--+50$C+**$E#$"#*,$50$b# ,0#-,13H$(13$Y$1'$Ua__$/=$ 80#C$10,$C528#,$C+'"$F1223A $ $ 8$+0$812-'A$>'$1*-5$ +0'25,).#-$"+=$'5$'"5-#$C"5$0#<#2$"1,$'"#$/*#1-)2#$'5$805C$"+=A$!"#$F1223$SA$F#.8$7+08"5*#$4506#2#0.#$ 7'),#0'$7."5*12-"+/$"1-$E##0$#-'1E*+-"#,$+0$"+-$=#=523A & & The following obituary appeared in the August 2012 issue of NSS News and is r eprinted here with permission. Dr. Barry F. Beck passed away at the age of 67, following a series of severe blood disorders and a brain stem stroke. The str oke in 2009 left him debilitated with "Locked in Syndrome." Born in Hershey, Pennsylvania, Barry was raised by his parents, Daniel and Verna Beck, in an economically challenged neighborhood in Rochester, New York. He earned a Bachelor's degree in geology at Rensselaer Polytechnic Institute, followed by Master's an d PhD degrees in geology from Rice Uni versity, Texas. His thesis and dissertation focused on the caves and karst of central Texas. y in the late 1960s. He was an adviso r to a Boy Scout Explorer Post from the Houston area and led them on caving trips in Texas, Mexico, and other areas. In July 1969, Barry joined the National Speleological Society (NSS) and the Texas Speleological Association, whe re he served as chairman in 1971. He was instrumental in founding the short lived Rice University Grotto and was a member of the Texas caving organization known as Carta Valley S.U.C.K.S. (Society of Underground Cavers, Karstologists and Speleologists). Foreshadowing events to come, caving trips seemed to invite heavy rain and associated challenges. During one expedition, several vehicles were washed from a campsite and nearly floated down the Colorado River, after which he was presented with a high water warning device. After receiving his PhD in 1972, Barry's interest in caves led him and his family (including first wife Pat and son Erik) to Puerto Rico, where Hans and Jenny were born. In Puerto Rico, Barry worked for the Department of Natural Resources, conducting investigations in tropical karst and assessing the commercial potential of the Aguas Buenas Caves. He was active in the study and exploration o f the Rio Camuy Cave system and st udied the sea margin caves on Isla Mona. Barry also hosted numerous expeditions for cavers exploring Puerto Rico and provided lodging (sofas and floors at his house) and transportation. Barry played a significant role in the a ttempt to rescue Francis McKin ney, who fell 75 feet to his death while climbing from a pit in the Rio Camuy Cave System in 1975.


! $ !"#$% & '" $ ()*'+,+-.+/*+0123$4506#2#0.#$50$7+08"5*#-$10,$'"#$ $ 90:+0##2+0:$;$90<+250=#0'1*$>=/1.'-$56$?12-' @1:#$ Y $ $ In 1976, Barry joined the Georgia Department of Natural Resources in Atlanta, where he applied electrical resistivity geophys ical techniques to the detection of caves and other karst features. Subsequently, he moved to Americus, Georgia where he taught geology, hydrology, and geomorphology at Georgia Southwestern College (now known as Georgia Southwestern State University). He also me t Dr. Penny Lukin, who wou ld become his second wife and the love of his life, and had two more children, Daniel and Sonja. During the late 1970s and early 1980s, Barry served as a review/associate editor for the NSS News and exploration editor for the NSS Bulletin. He was selected as a Fellow of the NSS in August 1977. Barry authored An Introduction to Caves and Cave Exploring in Georgia in 1980 and edited the proceedings of the Eighth International Congress of Speleology in 1981. In 1981, Barry became the first and only director of the Florida Sinkhole Research Institute at the University of Central Florida in Orlando. In that capacity, Barry applied his research to the geotechnical and environmental issues associated with sinkhole c ollapse, land subsidence, and karst groundwater flow. In 1984, he organized the first in a series of Multidisciplinary Conferences on Sinkholes In 1992, Barry joined P.E. LaMoreaux & Ass ociates, Inc. (PELA), where he served as Chief of Operations for the Oak Ridge office and Vice President for Karst. He managed a variety of sinkhole and karst consulting/research projects throughout the U.S. and around the world. He authored or coauthored numerous technical papers and articles on karst topics. With significant support from PELA, he continued to direct the sinkhole conferences, edit the proceedings, and host the meetings in karst regions around the U.S. Ba rry was the primary force behind the cy, hosting the 13th confe rence in Carlsbad, New Mexico during May 2013. Barry was a member of the East Tennessee Grotto and explored caves in east Tennessee, as well as those associated with his pr oject work around the world. In 2004, the NSS presented Barry with an Honorary Memb ership for his lifetime of contributions to the study of speleology and karst. In addition to his professional endeavors, Barry was a loving father to Erik, Jenny, Hans, Sonja, and Daniel. He was an activ e, energetic, and passionate individual, biking or running almost daily. He was also an avid hiker and tennis player. Barry and Pen ny shared a passion for dancing, specifically clogging and various forms of folk dancing. At one time, Barry served as President of the Muckalee Mudstompers, a clogging group in Americus, Georgia. Barry loved to explore the world through travel, visiting k arst regions in China, Guam, Italy, eastern Europe, Russia, Kuwait, and Argentina. nickname during a March 1979 trip to Anderson Sp ring Cave in Walker County, Georgia. Barry and six of his stu dents from Georgia Southwestern spent 28 hours trapped underground when heavy precipitation caused flash flooding that completely submerged the his own body for the final two years of his life. Barry leaves behind family, friends, and associates that will miss him and remember him fondly. His contributions to caves an d karst will continue. Memorial contributions may be made to the Nature Conservancy and earmarke d for the Barry Beck Memorial Fund (, which is designated for preservation of karst areas and research on caves and sinkholes. Memo rial contributions to the Barry F. Beck Sinkhole Conference Student Scholarship (https://sites.g will support the participation and professional development of at least one student at each conference. Art Pettit, NSS #18565RL J. Brad Stephenson, NSS #22755RL Dr. Penny Lukin Jim McLane, NSS #14628RG Carl Kunath, NSS #6230RE (PH FE) Dr. Brian Smith, NSS #15208RE Dr. Wanfang Zhou


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a $ $ %A T0#$62##$7+08"5*#$.506#2#0.#$2#:+-'21'+50A $ BA T0#$62##$2#:+-'21'+50$'5$1$6+#*,$'2+/$10,$-"52'$. 5)2-#$N/#0,+0:$-/1.#$1<1+*1E+*+'3QA $ &A D0$1C12,$.#2'+6+.1'#A $ RA P#.5:0+'+50$'"25):"$01=#$E1,:#$2+EE50H$=#0'+50$+0$'"#$7+08"5*#$4506#2#0.#$/25:21=$10,$C#E-+'#H$ 10,$1005)0.#=#0'-$1'$'"#$5/#0+0:$.#2#=503$10,$E10J)#'A $ UA P#+=E)2-#=#0'$652$)/$'5$g%___$56$/#2-501*$+0,+<+ ,)1*$'21<#*H$655,H$10,$*5,:+0:$#L/#0-#-$1--5.+1'#,$ C+'"$1''#0,+0:$'"#$7+08"5*#$4506#2#0.#A $ $ S52$=52#$+0652=1'+50$1E5)'$'"#$F#.8$7."5*12-"+/$10,$'5$6)0,$6)')2#$-."5*12-H$C1'."$6)')2#$7+08"5*#$ 4506#2#0.#$C#E-+'#-$52$.50'1.'$'"#$O1'+501*$41<#$10,$?12-'$P#-#12."$>0-'+')'#$1' $ +065h0.882+A52: $ 52$E3$.1**+0:$ UWU V YYW V U U%YA $ & & & $ &


! $ !"#$% & '" $ ()*'+,+-.+/*+0123$4506#2#0.#$50$7+08"5*#-$10,$'"#$ $ 90:+0##2+0:$;$90<+250=#0'1*$>=/1.'-$56$?12-' @1:#$ %_ $ $ ;"115&<:& ;%.= & ?4*=@'$%&)'*(%1%*.% & ?+PJ%*+& ?.@'$"1,@4A&C%.4A4%*+ & & 91:&8"1.'&c"++"*'I&T@9&]8?.&?+PJ%*+^ $ K#/12'=#0'$56$912'"$10,$7#1$7.+#0.#-H $ Z0+<#2-+'3$56$@1*#2=5 H$>'1*3 $ $ (12.5$d1''105 $ C1-$E520$50$B&$K#.#=E#2$%XW&$+0$41'10+1$N7+.+*3H$>'1*3QH$C"#2#$ "#$:2#C$)/$10,$:21,)1'#,$625=$"+:"$-."55*A$!"#0$"#$=5<#,$'5$@1*#2=5H$ C"#2#$"#$E#:10$"+-$:21,)1'#$-'),+#-A$>0$@1*#2=5$"#$E#:10$"+-$.1<+0:$1.'+<+'+#-$ '"108-$'5$1$.5)2-#$52:10+]#,$E3$'"#$>'1*+10$ D*/+0#$4*)E$+0$%XXeA$(12.5$d1''105$ :21,)1'#,$C+'"$6+2-'$.*1--$"5052-$+0$B__U$+0$'"#$^#5*5:3$10,$^#5,#-3$ K#/12'=#0'H$S1.)*'3$56$7.+#0.#H$Z0+<#2-+'3$56$@1*#2=5H$>'1*3A$G#$'"#0$#120#,$ "+-$@256#--+501*$^#5*5:+-'$2#:+-'21'+50A$>0$B__eH$"#$-'12'#,$"+-$@"K$+0$^#5 *5:3$1'$'"#$Z0+<#2-+'3$56$@1*#2=5$10,$ V 4#0'21*$7+.+*3$E3$2#*1'+50-"+/$101*3-+-$E#'C##0$"3/5:#10$812-'$*10,652=-$10,$-)261.#$*10,652 B__e$1 0,$B__WH$(12. 5$C1-$-)//52'#,$E3$1$6#**5C-"+/$-/50-52#,$E3$'"#$>-'+')'5$O1]+501*#$,+$^#56+-+.1$#$ d)*.105*5:+1$N>O^d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


! $ !"#$% & '" $ ()*'+,+-.+/*+0123$4506#2#0.#$50$7+08"5*#-$10,$'"#$ $ 90:+0##2+0:$;$90<+250=#0'1*$>=/1.'-$56$?12-' @1:#$ %% $ $ F1G"*4H4*G&)'//4++%% & $ V%*%1"$&)'*(%1%*.%&)' U )@"41, & ^#52:#$d#0+H$@"AKAH$@A^AH$O1'+501*$41<#$10,$?12-'$P#-#12."$>0-'+')'#$NO4?P>QH$412*-E1,H$O( $ $ c+=$M1(52#1)LH$@"AKAH$@A9A$M1(52#1)L$;$D--5.+1'#-H$>0.AH$!)-.1*55-1H$DM $ $ T1'G1"/& )' U )@"41, & M300$FA$f)"2H$@A^AH$!#."05-H$>0.AH$(+1=+H$SM $ M#C+-$M10,H$@"AKAH$O#C$(#L+.5$F)2#1)$56$^#5*5:3$;$(+0#21*$P#-5)2.#-$10,$O1'+501*$41<#$10,$?12-'$ P#-#12."$>0-'+')'#H$412*-E1,H$O#C$(#L+.5 $ $ T1'.%%J4*G,&8"*"G4*G&"*J&Y,,4,+"*+&ZJ4+'1, & M#C+-$M10,H$@"AKAH$O #C$(#L+.5$F)2#1)$56$^#5*5:3$;$(+0#21*$P#-5)2.#-$10,$O1'+501*$41<#$10,$?12-'$ P#-#12."$>0-'+')'#H$412*-E1,H$O( $ K10+#*$GA$K5.'52H & @"AKAH$ZA7A$^#5*5:+.1*$7)2<#3H$91-'#20$^#5*5:3$;$@1*#5.*+=1'#$7.+#0.#$4#0'#2H$ P#-'50H$dD$ $ cA$F21,$7'#/"#0-50H$ @A^AH$MAPA7AH$ 4F;>H $ ?05L<+**#H$!O $ $ $ <4%$J&!14A, & M#C+-$M10,H$@"AKAH$O#C$(#L+.5$F)2#1)$56$^#5*5:3$;$(+0#21*$P#-5)2.#-$10,$O1'+501*$41<#$10,$?12-'$P#-#12."$ >0-'+')'#H$412*-E1,H$O( $ $ ?@'1+&)'P1,%, & 9A$41*<+0$D*#L10,#2H$c2AH$@"AKAH$K#/12'=#0'$56$912'"$7.+#0.#-H$Z0+<#2-+'3$56$(+00#5'1H$(+00#1/5*+-H$(O $ $ 7*W4+%J&?A%"=%1, & f50:*+$^15H$@"AKAH$Z0+<#2-+'3$56$!#L1V 710$D0'50+5H$710$D0'50+5H$!i $ $ ;%.=&?.@'$"1,@4A & 9A$41*<+0$D*#L10,#2H$c2A$@"AKAH$K#/12'=#0'$56$912'"$7.+#0.#-H$Z0+<#2-+'3$56$(+00#-5'1H$(+00#1/5*+-H$(O $ K+100#$c55/H$O1'+501*$41<#$ 10,$?12-'$P#-#12."$>0-'+')'#H$412*-E1,H$O( $ >21$KA$71-5C-83H$ @"AKAH$@A^AH$^#5-.+#0.#H$Z0+<#2-+'3$56$D8250H$D8250H$TG$ $ $ ;%.=&8%/'14"$ & c+=$M1(52#1)LH$@"AKAH$@A9A$M1(52#1)L$;$D--5.+1'#-H$>0.AH$!)-.1*55-1H$DM $ K+100#$c55/H$O1'+501*$41<#$10,$?12-'$P#-#12."$>0-'+ ')'#H$412*-E1,H$O( $ F2+10$7=+'"H$ @"AKAH$F12'50$7/2+0:-I9,C12,-$DJ)+6#2$450-#2<1'+50$K+-'2+.'H$D)-'+0H$!i$ $ cA$F21,$7'#/"#0-50H$ @A^AH$MAPA7AH$ 4F;>H$?05L<+**#H$!O $ $ $ & &


! $ !"#$% & '" $ ()*'+,+-.+/*+0123$4506#2#0.#$50$7+08"5*#-$10,$'"#$ $ 90:+0##2+0:$;$90<+250=#0'1*$>=/1.'-$56$?12-' @1:#$ %B $ $ & )'*(%1%*.%&8"*"G%/%*+ & K+100#$c55/H$O1'+501*$41<#$10,$?12-'$P#-#12."$>0-'+')'#H$412*-E1,H$O( $ K#EE+#$G#22H$O1'+501*$41<#$10,$?12-'$P#-#12."$>0-'+')'#H$412*-E1,H$O( $ 7)]1001$M10:5C-8+H$O1'+501*$41<#$10,$?12-'$P#-#12."$>0-'+')'#H$412*-E1,H$O( $ $ )41.P$"1,&"*J&TP#$4.4+5 & 71=)#*$dA$@1005H$4^b@H$>**+05+-$7'1'#$^#5*5:+.1*$7)2<#3H$@21+2+#$P#-#12."$>0-'+')'#H$ Z0+<#2-+'3$56$ >**+05+-H$4"1=/1+:0H$>M $ G1223$MA$(552#H$@A^AH$^5*,#2$D--5.+1'#-H$D'*10'1H$^D $ $ T1'G1"/&L4+@&Y#,+1".+, & F2+10$7=+'"H$ @"AKAH$F12'50$7/2+0:-I9,C12,-$DJ)+6#2$450-#2<1'+50$K+-'2+.'H$D)-'+0H$!i$ $ F2+10$G)0'H$ F12'50$7/2+0:-I9,C12,-$DJ)+6#2$450-#2<1'+50$K+-'2+.'H$D)-'+0H$!i$ $ $ -%#,4+% & K+100#$c55/H$O1'+501*$41<#$10,$?12-'$P#-#12."$>0-'+')'#H$412*-E1,H$O( $ ^"#52:"#$@50'1H$@A^AH$@A9A$M1(52#1)L$10,$D--5.+1'#-H$!)-.1*55-1H$DM$ $ $ ?%,,4'*&)@"41, & 90:+0##2+0:$10,$^#5'#."0+. 1* $ D-/#.'-$56$?12-' $ V $ c5#$S+-."#2H$@"AKAH$@A9AH$^#5-.+#0.#$7#2<+.#-H$4*+0'50H$ Oc $ 9<1/52+'#$?12-'$ V $ ?#0$c5"0-50H$T8*1"5=1$^#5*5:+.1*$7)2<#3H$O52=10H$T? $ ^#5/"3-+.1*$9L/*521'+50$+0$?12-' $ !#2210# $ V $ ()-'161$712+E),18H$@"AKAH$@A^AH$90<+250=#0'1*$^#5/"3-+.-$ D--5. +1'#-H$D)-'+0H$!i $ S52=1'+50 $ @25.#--#$ 56$?12-'$10,$7+08"5*#-$ V $ G1223$MA$(552#H$^5*,#2$D--5.+1'#-H$D'*10'1H$^D$ $ G3,25*5:+.1*$D-/#.'-$56$?12-' $ V $ ^#123$7."+0,#*H$@A^AH$9,C12,-$DJ)+6#2$D)'"52+'3H$710$D0'50+5H$!i $ (1//+0:$10,$ (101:#=#0' $ 56$?12-'$P#:+50-$ V $ 71=)#* $ dA$@1005H$4^b@H$>**+05+-$7'1'#$^#5*5:+.1*$7)2<#3H$ @21+2+#$P#-#12."$>0-'+')'#H$Z0+<#2-+'3$56$>**+05+-H$4"1=/1+:0H$>M $ $ ?%,,4'*&?P#.'//4++%%, & !503$455*#3H$ @A9AH$@A^AH$ 994 V K9@$K+<+-+50$56$b1-'#$(101:#=#0'H$75*+,$b1-'#$F210."H$4*5-)2#$7#.'+50H$ S2108652'H$?f $ c# 66$7."1)=E)2:H$@A9AH$K301=+.$912'"H$4"#-'#2H$Oc$ $ K1<#$b#123H$O1'+501*$455/#21'+<#$^#5*5:+.$(1//+0:$@25:21=H$ZA7A$^#5*5:+.1*$7)2<#3H$P#-'50H$dD $ @"+*$412/#0'#2H$@"AKAH$K#/'A$56$^#5*5:3$10,$90<+250=#0'1*$^#5-.+#0.#-H$O52'"#20$>**+05+-$Z0+<#2-+'3H$ K#?1*EH$>M$ $ !"5=1-$MA$K5E#.8+H$ @"AKAH$@A^AH$ 7K>>$^*5E1*$452/521'+50H$!1=/1H$SM $ F2).#$7=+'"H$ZA7A$^#5*5:+.1*$7)2<#3H$K#0<#2$S#,#21*$4#0'#2H$K#0<#2H$4T$ $ & $ &


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b1-'#$ >-5*1'+50$@+*5'$@*10'$Nb>@@QH$1$,+-/5-1*$ 61.+*+'3$652$,#6#0-# V 2#*1'#,$ '210-)210+.$21,+51.'+<#$C1-'#A $ K)2+0:$'"#$'+=#$56$'"#$.506#2#0.#H$,13'+=#$"+:"$'#=/#21')2#-$12#$*+8#*3$'5$210:#$E#'C##0$Y_ V XU$jS$NBW V &U$j4Q$ C+'"$5<#20+:"'$*5C-$56$U_ V W _$jS$N%_ V B%$j4QA$P1+0$+-$)0*+8#*3$E)'$#L/#.'$/5--+E*#$-'250:$C+0,-A $ *5.1'#,$50*3$&_$=+0)'#-$E3$.12$'5$'"#$-5)'"C#-'$1*50:$G+:"C13$eBI%Y_A$D05'"#2$&_ $ =+0)'#-$612'"#2$10,$+0'5$ !#L1-$+-$^)1,1*)/#$(5)0'1+0-$O1'+501*$@128A$!"#-#$/128-$566#2$C52*, V .*1--$<+#C-$56$=+,,*#$@#2=+10$:#5*5:3$ 10,$"3/5:#0+.$812-'A$G+:"C13$eBI%Y_$'21<#2-#-$1$-'2+8+0:$10,$)0+J)#$:3/-)=$812-'$ '#2210# $ '"1'$+0.*),#-$@128-$ P10."$41<#H$'"# $ -#.50, V *50:#-'$:3/-)=$.1<#$+0$'"#$Z0+'#,$7'1'#-A$ $ >0$,5C0'5C0$412*-E1,$+-$'"#$412*-E1,$()-#)=$10,$D2'$4#0'#2H$C"+."$"1-$10$+0'#2#-'+0:$#L"+E+'$50$'"#$ M+<+0:$ K#-#2'$7'1'#$@128$10,$k55H$1$E#1)'+6)**3$,#<#*5/#,$,+-/*13$56$6*521$10,$61)01$56$'"#$,#-#2'$-5)'"C#-'$2#:+50$56$ ZST$()-#)=$10,$P#-#1 2."$4#0'#2$+-$1*C13-$#0'#2'1+0+0:A $


! $ !"#$% & '" $ ()*'+,+-.+/*+0123$4506#2#0.#$50$7+08"5*#-$10,$'"#$ $ 90:+0##2+0:$;$90<+250=#0'1*$>=/1.'-$56$?12-' @1:#$ %R $ $ $ $ $ $ $ $ ^#0#21*+]#,$=1/$56$412*-E1,$10,$*5.1'+50-$56$7+08"5*#$4506#2#0.#$<#0)#-A $


! $ !"#$% & '" $ ()*'+,+-.+/*+0123$4506#2#0.#$50$7+08"5*#-$10,$'"#$ $ 90:+0##2+0:$;$90<+250=#0'1*$>=/1.'-$56$?12-' @1:#$ %U $ $ )'*(%1%*.%&0'."+4'* _& & K"+4'*"$&)"W%&"*J&N"1,+&C%,%"1.@&7*,+4+P+%&]K)NC7^ & R__ V %$41-.1,#-$D<#0)# $ 412*-E1,H$O#C$(#L+.5H$YYBB_ V eB%U$Z7D $ "''/aIICCCA0.82+A52: $ !#*a$__A%A UWUAYYWAUU%Y $ $ $ O4?P>$G#1,J)12'#2-$+-$*5.1'#,$0#12$'"#$.#0'#2$56$ 412*-E1,$1*50:$'"#$@#.5-$P+<#2A$>'$+-$/12'$56$1$0#C$ -5.+1*$10,$2#'1+*$.#0'#2$652$412*-E1,$'"1'$+-$.)22#0'*3$ )0,#2$.50-'2).'+50$10,$+-$#L/#.'#,$'5$E#$.5=/*#'#,$ +0 $ B V R$3#12-A$450-'2).'+50$56$O4?P>$+-$1*-5$50 V :5+0:H$ E)'$05'$,)2+0:$'"#$.506#2#0.#H$C+'"$/*10-$)0,#2C13$'5$6+**$'"#$=1+0$"1**$10,$*5EE3$C+'"$+0'#21.'+<#$=)-#)=$ #L"+E+'-A $ $ 4506#2#0.#$-#--+50-$10,$= 5-' $ #<#0'-H$+0.*),+0:$'"2##$56$'"#$-"52'$.5)2-#-H$C+**$E#$*5.1'#,$50$'"#$:25)0,$6*552A$ T0#$C528-"5/$C+**$E#$"#*,$)/-'1+2-A$^#0#21*$.506#2#0.#$+0652=1'+50$10,$/+.8 V )/$56$+0652=1'+50$+-$*5.1'#,$[)-'$ +0-+,#$'"#$*5EE3$1'$'"#$#0'210.#$'5$'"#$ 6)')2#$ E558-'52#$12#1A $ !"#$/5-'#2$-#--+50-$C+**$E#$"#*,$+0$'"#$6)')2#$ E558-'52#A$S52$50 V -+'#$2#:+-'21'+50$10,$+0652=1'+50$+6$'"#$:25)0,$6*552$,#-8$+-$05'$5..)/+#,H$/*#1-#$:5$)/-'1+2-$ '5$'"#$566+.#$2#.#/'+50$,#-8A $ & D=#0+'+#-$1'$O4?P>$G#1,J)12'#2-$+0.*),#$62##$C+2#*#--$>0'#20#'$N P55=$+0$'"#$566+.#$E#"+0,$'"#$:25)0,$6*552$2#:+-'21'+50$,#-8$+6$35)$0##,$1$J)+#'$-/1.#$'5$2#<+#C$35)2$ /2#-#0'1'+50H$E5''*#,$C1'#2$10,$-56'$,2+08-$+0$'"#$2#62+:#21'52$+0$'"#$.506#2#0.#$255=H$10,$"5'$,2+08-$10,$ -01.8-$+ 0$'"#$=1+0$"1**C13A$O4?P>$+-$1*-5$/25),$'5$"5-'$'"#$.506#2#0.#$+0$+'-$#0<+250=#0'1**3 V 62+#0,*3$61.+*+'3A$ S52$,#'1+*-H$1-8$103$O4?P>$-'166$=#=E#2$1E5)'$'"#$C1'#2$10,$#0#2:3 V -1<+0:$6#1')2#-H$1-$C#**$1-$'"#$2#.3.*#,H$ *5C V dT4H$10,$-)-'1+01E+*+'3 V /25,).#,$=1'#2 +1*-A $ %% %


! $ !"#$% & '" $ ()*'+,+-.+/*+0123$4506#2#0.#$50$7+08"5*#-$10,$'"#$ $ 90:+0##2+0:$;$90<+250=#0'1*$>=/1.'-$56$?12-' @1:#$ %e $ $ <4%$J&!14A & & & -",+%&7,'$"+4'*&T4$'+&T$"*+ & (50,13H$(13$ e H$Wa_ _$1=$ $ %a&_$/= $ 10, $ 2#/#1'#,$50$ !")2-,13H$(13$ X H$Wa _$1=$ $ %a&_$/= $ $ !2+/$M#1,#2-a$ b>@@$7'166 $ $ D--#=E*#$1'a$7'#<#0-$>00$Wa__$1=`$S1+26+#*,$>00$Wa%U$1=A$ F )$ C+**$/25.##,$'5$78##0 V b"+'*5.8$NK#/12'=#0'$56$ 90#2:3Q$F)+*,+0:$652$52+#0'1'+50$10,$'210-6#2$'5$KT9$<10A $ $ T0$(50,13$=520+0: H $ C#$12#$566#2+0:$10$)0,#2:25)0,$'5)2$56$'"#$b1-'#$>-5*1'+50 $ @+*5'$@*10'$Nb>@@Q$#1-'$56$ 412*-E1, $ 652 $ % _$/#5/*#$=1L+=)= $ 652$'"#$(50,13$'2+/$10,$B_$/#5/*#$=1L+=)=$652$'"#$!")2-,13$'2+/ A $ !"#$ $ ,+-/5-1*$56$,#6#0-# V 2#*1'#,$ '210-)210+.$21,+51.'+<#$C1-'#A$D$-")''*#$C+** $ /25<+,#$'210-/52'$'5$'"#$b>@ @$-+'#$+0$#1-'#20$9,,3$45AH$C"#2#$ 1''#0,##-$C+** $ "1<#$'"#$)0+J)#$5//52')0+'3$'5$'5)2$'"#$-+'#$E5'"$50$'"#$-)261.#$10,$eU_$=#'#2$ E#*5C$:25)0,$ *#<#*$N:#0#21**3$/)E*+.$'5)2-$56$b>@@$12#$05'$1<1+*1E*#QA$!"+$ "1*6 V ,13$1.'+<+'3$C+**$+0.*),#$1$:)+,#,$'5)2$56$ -)2 61.#$5/#21'+50-H$65**5C#,$E3$10 $ )0,#2:25)0,$'5)2$56$'"#$2#/5-+'523$255=-H$C"+."$"1<#$E##0$#L.1<1'#,$+0$ E#,,#, $ -1*'$56$'"#$)//#2$@#2=+10$71*1,5$S52=1'+50A$D''#0,##-$=13$.5**#.'$"10, $ -1=/*#-$56$"1*+'# $ 625=$'"#$ ,2+6'-$1'$05$."12:#A$M)0."$+-$1<1+*1E*# $ 652$/)2. "1-#$+0$'"#$b>@@$.16#'#2+1$16'#2$'"#$)0,#2:25)0,$'5)2A $ $ $ !"#$71*1,5$C1-$."5-#0$1-$'"#$"5-'$652=1'+50$652$b>@@ $ E#.1)-#$-1*'$,#652=-$+0$1$,).'+*#$=100#2$)0,#2$"+:"$ .506+0+0: $ /2#--)2#A$D6'#2$'"#$C1-'#$+-$,#/5-+'#,$+0$'"#$2#/5-+'523 $ 255=-H$'"#$-1*'$C+**$#<#0') 1**3$N16'#2$1$6#C$ 3#12-$'5$,#.1,#-Q $ 6*5C$10,$#0<#*5/$'"#$.1 0+-'#2-H$.2)-"+0:$'"#=$10, $ #0'5=E+0:$'"#$-'52#,$C1-'#A$!"#$ +=/#2=#1E*#$01')2#$56 $ 25.8$-1*'$/2#<#0'-$C1-'#$625=$*#18+0:$5)'$56$'"#$2#/5-+'523 $ 10,$.50'1=+01'+0:$-)261.#$ 52$:25)0,C1'#2$2#-5)2.#-A $ DE)0 ,10'$-)261.#$812-'$6#1')2#-$12#$,#<#*5/#,$+0$'"#$5<#2*3+0: $ P)-'*#2$S52=1'+50$1$ 6#C$8+*5=#'#2-$C#-'$56$'"#$b>@@ $ -+'#A$9L'#0-+<#$2#-#12."$E3$710,+1$O1'+501*$M1E$10,$'"# $ K#/12'=#0'$56$ 90#2:3$+0,+.1'#-$'"1'$1$.500#.'+50$E#'C##0 $ '"#-#$-)261.#$6#1')2#-$10,$'"#$b >@@$2#/5-+'523$+-$)0*+8#*3A $ $ E-,&%,A", % B-# % B-#&)@*%*",)-*"10%,-%",,&*2%,A&$%?50,%A"+&%'-?C1&,&2%"*2%05>?),,&2%"*%V*'1"00)B)&2%L-#&)@*% U)0),-#%B-#?%*-%1",&#%,A"*%9W%."#'A%MN9O3%"*2%C#-+)2&%"%'-C$%-B%,A&)#%C"00C-#,%-#%+)0"%"*2%"%CA-,-%/X%PA&*% ,A&$%"##)+&8 % $ $ &


! $ !"#$% & '" $ ()*'+,+-.+/*+0123$4506#2#0.#$50$7+08"5*#-$10,$'"#$ $ 90:+0##2+0:$;$90<+250=#0'1*$>=/1.'-$56$?12-' @1:#$ %W $ $ & Y&-1-@)0,0%"*2%&*@)*&�%2)0'500%C&#?"*&*,%0,-#"@&%-B%,#"*05#"*)'%P"0,&%7)*%,A&%'-*,")*�%)*%,A&%>"'(@#-5*2=% )*%2&&C1$%>5#)&2%0"1,%",%Z/448 % 7/*0&,%CA-,-=% U)0),-#0%",%Z/44%"#&%P&1'-?&%,-%'-11&',%,A&)#%-P*%0"?C1&0%-B% 4&#?)"*%"@&%0"1,%B#-?%,A&%P"110%-B%, A&%0,-#"@&%B"')1),$%7CA-,0 [%E",)-*"1%K"+&%"*2%F"#0,%6&0&"#'A%/*0,),5,&=8 % & ZW"A'14+%&N"1,+&'(&+@%&0'L%1&T%.',&c "$$%5 $ !)#-,13H$(13$ W H$Wa _$1=$ $ Ua&_$/= $ !2+/$M#1,#2 a$ M#C+-$M10, $ $ D--#=E*#$1'$7'#<#0-$>00H$eaRU$ $ Wa%U $ 1=`$E)-$,#/12'-$/25=/'*3$1 '$Wa%UA$D--#=E*#$ 10,$E512,$E)$ 1'$S1+26+#*,$ >00H$WaB_$ V $ Wa&_ $ 1=`$E)-$,#/12'-$/25=/'*3$1'$Wa&_ $ 1= A $ $ T0$!)#-,13$C#$566#2$1$6)** V ,13$-)261.#$'5)2$56$+08"5*#-$10,$5'"#2$:3/-)=$812-'$ 6#1')2#-$56$'"#$*5C#2$@#.5-$ <1**#3$E#'C##0$412* -E1,$10,$P5-C#**H$O(A$9<1/52+'#$ 812-'$/25.#--#$ "1 <#$/*13#,$1$6)0,1=#0'1*$25*#$+0$ -"1/+0:$'"#$:#5=52/"5*5:3$56 $ '"#$*5C#2$@#.5-$2#:+50$10,$.50'25**+0:$:25)0,C1'#2$6*5C$/1''#20-A$!"+-$'2+/$C+** $ 65.)-$50$#0:+0##2+0:$10,$#0<+250=#0'1*$"1]12, -$1--5.+1'#,$C+'"$:3/-)=$812-'H$ +0.*),+0:$'"#$652=1 '+50$56$


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a&_$ $ Ua&_$/= $ $ !2+/$ M#1,#2 a$^#52:#$d#0+ $ $ F)-#-$,#/12'$625= $ O4?P>$1'$%Ba&_ $ /= A $ $ T0$S2+,13$16'#20550$C#$12#$566#2+0:$1$C1*8+0:$'5)2$56$'"#$F+:$P55=$+0$412*-E1, $ 41<#20H$1$.*1--+.$#L1=/*#$56$ "3/5:#0+. V -)*6)2+ .$1.+,$-/#*#5:#0#-+-A$D$-")''*#$ C+**$,#/12'$625=$'"#$.506#2#0.#$.#0'#2$1 '$%Ba&_$ /=A$M)0."$ C+**$E#$1<1+*1E*#$652$ /)2."1-#$1'$'"#$412*-E1,$41<#20-$<+-+'52-$.#0'#2 H$10,$'"#$'5)2$C+**$E#:+0$16'#2$ *)0."$1'$Ba__$


! $ !"#$% & '" $ ()*'+,+-.+/*+0123$4506#2#0.#$50$7+08"5*#-$10,$'"#$ $ 90:+0##2+0:$;$90<+250=#0'1*$>=/1.'-$56$?12-' @1:#$ %X $ $ /=A $ P#')20$'5$O4?P>$C+**$E#$1'$Ua&_$/= A $ D..#--$C+**$E#$<+1$#*#<1'52$+0'5$'"# $ F+:$P55=$B&_$=$E#*5C$ :25)0,$ *#<#*A$412*-E1,$4 1<#20$+-$:#0#21**3$. 50-+,#2#,$'5$E#$50#$56$'"#$=5-'$ -/#.'1.)*12*3 V ,#.521'#,$*12:#$.1<#-$50$ #1 2'"H$10,$+-$1$ZO974T V ,#-+:01'#,$ b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


! $ !"#$% & '" $ ()*'+,+-.+/*+0123$4506#2#0.#$50$7+08"5*#-$10,$'"#$ $ 90:+0##2+0:$;$90<+250=#0'1*$>=/1.'-$56$?12-' @1:#$ B_ $ $ ?@'1+&)'P1,%, & (50,13H$(13$eH$B_%& $ $ V1'P+4*G&4*&N"1,+&!%11"*% & $ >0-'2).'52a $ c5-#/"$DA$S+-."#2H$@A9AH$^#5-.+#0.# $ 7#2<+.#$ (1''$d10 P#0-*#2 H$@A9AH$45=/1.'+50$^25)'+0:$7#2<+.#$ $ M5.1'+50a$O4?P>$4506#2#0.#$P55=$D $ $ 45)2-#$*#0:'" a$R$"5)2-$NYa&_ 1=$'5$%Ba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d1-'*3$>=/25<#$^25)0,C1'#2$l)1*+'3$ 10,$7)//*3$>0<#-'+:1'+50-A $ $ >0-'2).'52 a $ b+**+1=$@#,*#2H$@2#-+,#0'$56$PD7H$>0'#:21'#,$ 7)E-)261.#$9<1*)1'+50 $ $ M5.1'+50a$O4?P>$4506#2#0.#$P55=$D $ $ 45)2-#$*#0:'" a$R$"5)2-$N %a__/=$ $ Ua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


! $ !"#$% & '" $ ()*'+,+-.+/*+0123$4506#2#0.#$50$7+08"5*#-$10,$'"#$ $ 90:+0##2+0:$;$90<+250=#0'1*$>=/1.'-$56$?12-' @1:#$ B% $ $ ,#'1+*#,$)0,#2-'10,+0:$56$'"#$:#5*5:3H$-'2).')2#H$<#2'+.1*$"3,21)*+.$:21,+#0'$10,$<#2'+.1*$,+-'2+E)'+50$56$C1'#2$ J)1*+'3$1--5.+1'#,$C+'" $ '"#$1.'+<#$C1'#2$E#12+0:$)0+'-A$ $ $ >=/52'10'$1//*+.1'+50-$652$+0'#:21'+50$56$'"#-#$=#'"5,-$+0.*),#$."121.'#2+]1'+50$-'),+#-$1--5.+1'#,$C+'"$ .50'1=+01'#,$-+'#-$NP>S7Q`$C#**"#1,$/25'#.'+50`$C1'#2$-)//*3$+=/1.'-$2#*1'#,$'5$2#-5)2.#$,#<#*5/=#0'`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a$ $ ^#123$(A$7."+0,#*H$@A^AH$4"+#6$!#."0+.1*$T 66+.#2H$9,C12,-$DJ)+6#2 $ D)'"52+'3H$710$D0'50+5H$!#L1$ 9A$41*<+0$D*#L10,#2H$c2AH$@"AKAH$@256#--52$56$912'"$7.+#0.#-H $ Z0+<#2-+'3$56$(+ 00#-5'1H$(+00#1/5*+-H$(+00#-5'1 $ $ M5.1'+50a$O4?P>$4506#2#0.#$P55=$F $ $ 45)2-#$*#0:'" a$ W $ "5)2-$N Ya&_1=$ $ Ra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a$ $ M#C+-$M10,H $ @"A$KAH $ ?12-'$G3,25*5:+ -'H$O#C$(#L+.5$F)2#1)$56$^#5*5:3 $ $ M5.1'+50a$O4?P>$F512,$P55=$N-#.50,$6*552Q$10,$+0$'"#$6+#*, $ $ 45)2-#$*#0:'" a$ W $ "5)2-$N Ya&_ $ 1=$ $ Ra&_ $ /= Q $ $


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b"#0$ '"#$-) 2<#3$+-$.5=/*#'#$'"#$.*1--$C+**$2#')20$'5$O 4?P>$566+.#-$10,$/12'+.+/1'#$+0$ ,5C0*51,+0:$10,$/25.#--+0:$'"#$,1'1$652$6+01*$101*3-+-A $ $ $ $ I&P)0%I"*2%'-*25',)*@%&1& ',#)'"1%#&0)0,)+),$%05#+&$%-B%,A&% /]Z%>#)*&%P&11%'"+),$%)*%K"#10>"2^%D,-C%9%-B%,A&% J+"C-#),&% F"#0,%B)&12%,#)C 8 % 4A-,-[%E",)-*"1%K"+&%"*2%F"#0,%6&0&"#'A%/*0,),5,&8 % % $ $ $ $


! !"#$% & '" $ ()*'+,+-.+/*+0123$4506#2#0.#$50$7+08"5*#-$10,$'"#$ $ 90:+0##2+0:$;$90<+250=#0'1*$>=/1.'-$56$?12-' @1:#$ B& $ T1'G1"/&"+ U U V$"*.% & & 8'*J"5 & & & !P%,J"5 & & & -%J*%,J"5 & & & !@P1,J"5 & & & <14J"5 & & 8"5&RI&SED2 & & 8"5&>I&SED2 & & 8"5&XI&SED2 & & 8"5&BI&SED2 & & 8"5&DEI&SED2 & Wa&_ V Ya__ $ $ $ P#:+-'21'+ 50$T/#0-$ NWa&_1=Q $ $ $ $ $ $ $ P #:+-'21'+50$T/#0-$NWa&_1=Q $ $ $ P#:+-'21'+50$T/#0-$ NWa&_1=Q $ $ $ $ $ P#:+-'21'+50$T/#0-$ NWa&_1=Q $ Ya__ V Ya&_ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ Ya&_ V Xa__ $ $ $ $ $ $ $ $ $ $ $ Z*G4*%%14*G&N%5*'+%_& & V"1",4. & $ $ N"1,+&<'1/"+4'* & N%5*'+%_& & T"14,% & $ $ $ $ $ $ Xa__ V Xa&_ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ Xa&_ V %_a__ $ -7TT&!'P1 & ?@'1+&)'P1,%, & $ $ $ $ $ $ T$%*"15&?%,,4'* & $ $ T$%*"15&?%,,4'* & -7TT&!'P1 & $ $ T$%*"15&?%,,4'* & %_a__ V %_a&_ $ ]>_2E"/ U S_EEA/^ & ]X_2E"/ U DS_2EA/^ & $ $ $ $ $ $ ]X_DQ"/& U & DS_DEA/^ & $ $ ]X_DQ"/& U & DS_EEA/^ & ]>_2E"/ U S_EEA/^ & $ $ ]X_DQ& U & DD_SE"/^ & %_a&_ V %%a__ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ & & $ $ $ $ $ $ %%a__ V %%a&_ $ $ $ $ $ $ $ ZW"A'14+% & $ $ $ $ $ $ $ $ $ $ $ $ $ $ %%a&_ V %Ba__ $ $ $ $ $ $ $ N"1,+ & $ $ $ $ $ $ $ $ $ $ $ $ $ $ %Ba__ V %Ba&_ $ $ $ $ $ $ $ <4%$J&!14A & $ $ 41'#2#,$M)0." $ $ 41'#2#,$M)0." $ $ $ $ $ $ $ %Ba&_ V %a__ $ $ $ M)0."$T0$f5)2$TC0 $ $ $ ]>_DQ"/& U R_EEA/^ & $ $ N%Ba%_$ V $ %a&_/=Q $ $ N%Ba__$ V $ %a&_/=Q $ $ $ $ $ $ %a__ V %a&_ $ $ $ N%Ba&_$ V $ %a&_/=Q $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ %a&_ V Ba__ $ $ $ $ $ $ $ $ $ $ $ & & $ $ $ $ $ $ $ $ Ba__ V Ba&_ $ $ $ $ $ $ $ $ $ $ $ C%G4'*"$&V%'$'G5 & N%5*'+%_ & 9P)@%*% & $ $ $ $ $ )"1$,#"J&)"W%1* & Ba&_ V &a__ $ $ $ $ $ $ $ $ $ $ $ & & $ $ $ $ $ <4%$J&!14A & &a__ V &a&_ $ $ $ ?@'1+&)'P1,%, & $ $ $ $ $ $ T$%*"15&?%,,4'* & $ T$%*"15&?%,,4'* & $ $ ]DS_2E& U & Q_2E^ & &a&_ V Ra__ $ $ $ ]D_EE U Q_E EA/^ & $ $ $ $ $ $ ]D_2E& U & 6_QEA/^ & $ ]D_2E& U & 6_2EA/^ & $ $ $ $ Ra__ V Ra&_ $ $ $ $ $ $ $ $ $ $ $ & & $ $ $ $ $ $ $ $ $ Ra&_ V Ua__ $ $ $ $ $ $ $ $ $ $ $ & & $ $ $ $ $ $ $ $ Ua__ V Ua&_ $ $ $ $ $ $ $ $ $ $ $ ;%.=&8%/'14"$`-4*%&"*J & $ $ $ $ $ $ $ $ Ua&_ V ea__ $ $ $ $ $ $ $ $ $ $ $ )@%%,%&1%.%A+4'*`T',+%1, U & $ $ $ $ $ $ $ $ ea__ V ea&_ $ $ $ $ $ $ $ $ $ $ 8"**%J&]Q_EE U >_EEA/^ & $ $ $ $ $ $ $ $ ea&_ V Wa__ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ Wa__ V Wa&_ $ $ $ $ $ $ $ -%$.'/%& C%.%A+4'* & $ $ T$"**4*G&8%%+4*G&('1 & $ ;"*OP%+&"*J&VP%,+&?A%"=%1 & $ $ $ $ Wa& V Ya__ $ N%*&d'@*,'* & $ $ ]R_2E& U & X_2EA/^ & $ $ D6+@&?4*=@'$%&)'*(& ]>_EEA/^ & $ $ ;4$$&N'.@"*'W&]R_2E U B_EEA/^ & $ $ $ $ Ya__ V Ya&_ $ K)NC7& 94,+4*GP4,@%J&0%.+P1% & $ $ "+&?+%W%*,&7** & $ $ $ $ $ $ "+&)'P*+15&)$P# & $ $ $ $ Ya&_ V Xa__ $ ]>_2E U B_EEA/^ & $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ F#<#21:#$E2#18-$b#,0#-,13$1'$%_a%U $ 1= $ 10,$&a%_ $ /= H$!")2-,13$1'$%_a__ $ 1= $ 10,$Ba&_ $ /= H$S2+,13$1'$XaR_ $ 1= A $


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d'@*,'*_& K)NC7& 94,+4*GP4,@%J&0%.+P1%e& ZW"A'14+%&N"1,+&]'A%*&+'&AP#$4.^ & ?"$+&94,,'$P+4'*&"*J&?4*=@'$%,&4*&-%,+&!%M",&"*J&? 'P+@%",+&K%L& 8%M4.' & $ $ $ $ $ !P%,J"5I&8"5&> & $ $ >_2E"/ & <4%$J&+14A_&ZW"A'14+%&N "1,+&'(& +@%&0'L%1&T%.',&c"$$%5 & R_2EA/ & -%$.'/%&1%.%A+4'*I&?+%W%*,&7** & $ $ $ -%J*%,J"5I&8"5&X & $ $ X_DQ"/ & 7*+1' JP.+'15 & 1%/"1=,& U & V%'1G%&c%*4&"*J&d4/ & 0"8'1%"PM&]DQ&/4*P+%,^ & X_2E"/ & ?%*"+'1&\J"$$ & 1%A 1%,%*+"+4W% & L%$.'/4*G&.'//%*+,&]Q&/4*P+%,^ & X_2Q"/ & N%5*'+%&?A%"=%1_&Z*G4*%%14*G& & & ?A%$%'$'G4."$I&35J1'G%'$'G4."$&"*J&Z*G4*%%14*G&V%'$'G4."$& )@"$$%*G%,&'(&!P**%$4*G&4*&N"1,+&Y1%", & $ $ $ $ $ T$%*"15&?%,,4'*_&Z*G4*%%14*G&"*J& V%'+%.@*4."$ & Y,A%.+,&'(&N"1,+ I&d'%&<4,.@%1I&.@"41 & Xa%U1= $ P5E#2'$K#0'50$c2A $ !5C12,-$1$812-'$1--#--=#0'$-'10,12,$/21.'+.# $ Xa&U1= $ c5"0$@)-#3$c2A $ ^#5'#."0+.1*$.1-#$"+-'523$652$-+08"5*#$+0<#-'+:1'+50$10,$-'1E+*+]1'+50$ =#'"5,-$1*50:$1$"+:"$/2#--)2#$/#'25*#)=$/+/#*+0# $ XaUU1= $ 9A$KA$k+-=10 $ @25E*#=-$1--5.+1'#,$C+'"$'"#$ )-#$56$.5=/1.'+50$:25)'$652$-+08"5*#$ 2#=#,+1'+50$+0$C#-' V .#0'21*$S*52+,1 $ DE_DQ"/ & ; 1%"=&4*&%M@4#4+&@"$$&]DQ&/4*P+%,^ & %_a&_1= $ b+**+1=$F10:-)0, $ 9<1*)1'+0:$812-'$2+-8$1'$/25/5-#,$C+0,/5C#2$/25[#.'$ %_aU_1= $ !+=5'"3$7+#:#* $ D//*+.1'+50$56$-'1E+*+'3$."12'-$10,$ 2#*+1E+*+'3$.50.#/'-$652$-+=/*+6+#,$ 101*3-+-$56$1$<5+,$+0$-5+*$5<#2*3+0:$812-'$E#,25.8 $ %%a%_1= $ 9A$KA$k+-=10 $ >6$+'\-$C#+:"'$56$"1==#2$.50,+'+50-H$+'$=)-'$E#$1$-+08"5*#m $ %%a&_1= $ c5-#/" $ S+-."#2 $ 9L/*521'523$:25)'+0:$56$1$-)E-)261.#$,#'#0'+50I+06+*'21'+50$ -3-'#= $ %%aU_1= $ P5E#2'$F1.")$ D0$+005<1'+<#$1//251."$'5$."121.'#2+]+0:H$/#2=+''+0:H$10,$ .50-'2).'+0:$*10,6+**-$+0$812-'$:#5*5:+.$-#''+0:$ DS_DEA/ & 0\K)3&]D&@'P1I&SE&/4*P+%,^ & D_2EA/ & N%5*'+%&?A%"=%1_&C%G4'*"$&V %'$'G5& & 3"1W%5&C:&&9P)@%*%&]6E&/4*P+%,^ & V%'$'G4.&F W%1W4%L&'(&+@%& T%1/4"*&#",4*&'(&?'P+@%",+%1*&K%L& 8%M4.'&"*J&%,+&!%M", & $ $ $ $ $


! $ !"#$% & '" $ ()*'+,+-.+/*+0123$4506#2#0.#$50$7+08"5*#-$10,$'"#$ $ 90:+0##2+0:$;$90<+250=#0'1*$>=/1.'-$56$?12-' @1:#$ BU $ $ T$%*"15&?%,,4'*_&ZW"A'14+%&N "1,+I&N%*&d'@*,'*I&.@"41 & Ba%_/= $ ?#00#'" $ c5"0-50$ $ 71*'$?12-'$10,$45**1/-#$7'2).')2#-$+0$'"#$D01,1285$F1-+0$56$T8*1"5=1$ 10,$!#L1$ Ba&_/= $ M#C+-$M10,$ $ 9<1/52+'#$?12-'$+0$'"#$@#2=+10$F1-+0$P#:+50$56$b#-'$!#L1-$10,$ 75)'"#1-'#20$O#C$(#L+.5a$!"#$G)=10$>=/1.'$ $ BaU_/= $ ?#<+0 $ 7'16652, $ 9<1/52+'#$?12-'$10,$G3,25:#5*5:3$56$'"#$41-'+*#$S52=1'+50a$4)*E#2-50$ 45)0'3H$!#L1-$10,$9,,3$45)0'3H$O#C$(#L+.5 $ 2_DE A/ & ; 1%"=&4*&%M@4#4+&@"$$&]SE&/4*P+%,^ & &a&_/= $ ?#00#'" $ c5"0-50 $ ^3/-)=$?12-'$41)-#-$P#*5.1'+50$56$@25/5-#,$4#,12$P+,:#$K1=H$ !"25.8=52'50$45)0'3H$!#L1-$$ $ &aU_/= $ D0'"503$455/#2 $ !"#$P5*#$56$7)*61'# V P+."$7/2+0:-$10,$^25)0,C1'#2$+0$'"#$S52=1'+50$56$ 7+08"5*#-$ 5<#2$^3/-)=$+0$91-'#20$90:*10, $ Ra%_/= $ ?#00#'" $ c5"0-50 $ ^3/-)=$?12-'$10,$@5'#0'+1*$@25E*#=-$+0$7+'+0:$b+0,$!)2E+0#-$+0$ F*1+0#$45)0'3H$T8*1"5=1$ $ Ra&_/= $ K10+#* $ K5.'52 $ 9<1/52+'#$?12-'$+0$'"#$F*1.8$G+**-H$75)'"$K185'1$10,$b35=+0:H$10,$ '"#$T+*$@*13$+0$'"#$b+**+ -'50$F1-+0H$O52'"$K185'1$10,$(50'101 $ Q_EE&T8 & ;"115&;%.=&8%/'14"$& ?%,,4'* & $ $ & & ]-4*%&"*J&)@%%,%&1%.%A+4'*^ & Ua__/= $ F2+10$7=+'" $ $ @)#2'5$P+.5 $ UaB_/= $ G1223$(552# $ @2#-#0'1'+50$56$Z7KT!$812-'$/25[#.'$'"1'$F1223$C1-$+0<5*<#,$ C+'" $ UaR_/= $ T/#0$S52)= $ $ $ R_2EA/ & 8"**%J&T',+%1&?%,,4'*&& & $ $ & & F2+10 $ 450C13$ $ (50+'52+0:$9<1/52+'#$?12-'$D.'+<+'3$10,$M10,$7)E-+,#0.#$+0$'"#$ G5*E2558$F1-+0H$D2+]501$)-+0:$>0'#26#25=#'2+.$730'"#'+.$ D/#2')2#$P1,12$N>07DPQ$ $ & & F2+10$G)0' $ 45<#2 V 45**1/-#$ 7+08"5*#$K#<#*5/=#0'$+0$'"#$42#'1.#5)-$ 9,C12,-$M+=#-'50#H$4#0'21*$!#L1$ & & Fuping Gan >0'#:21'#,$:#5/"3-+.1*$=#'"5,-$652$:25)0,C1'#2$#L/*521'+50$ +0$1$812-'$12#1-$C+'"$5)2$C+'"5)'$'"+0$.5<#2$ V $ 1$.1-#$-'),3$625=$ !1+\10$.+'3H$7"10,50:$/25<+0.#H$4"+01 $ & & Henok Kiflu 7'1'+-'+.1*$D01*3-+-$56$^@P$10,$7@!$(#'"5,-$652$7+08"5*#$ >0<#-'+:1'+50-$+0$45<#2#,$?12-'$ !#2210# H$b#-' V 4#0'21*$S*52+,1H$ Z7D $ & & Mingtang Lei !3/+.1*$=#'"5,-$652$652#.1-'+0:$812-'$.5**1/-#$+0$4"+01 $ & & c1=#$ O#1*$ $ d12+1'+50-$+0$9<1/52+'#$?12-'$ +0$'"#$G5*E2558$F1-+0H$D2+]501 $ & & ?#<+0 $ 7'16652, $ 4*1-'+.$-+08"5*#$10,$/-#),5812-'$,#<#*5/=#0'$+0$#1-'$'#L1$ & & (#*+0,1$S1)*80#2 $ K#*+0#1'+50$10,$.*1--+6+.1'+50$56$812-'$,#/2#--+50-$)-+0:$*+,12a$ S52'$G55,$(+*+'123$>0-'1**1'+50H$!#L1$ & & (+."1#* $ (.^##$ $ 4"1**#0:#-$10,$*#--50-$*#120#,$625=$=101:+0:$1$0#C*3$ ,#<#*5/#,$.1'1-'25/"+.$-+08"5*#$M18#$*5.1'#,$50$=)*'+/*#$)-#$ =101:#,$/)E*+.$*10,$+0$'"#$*5C#2$/#.5-$<1**#3H$O( $ & & (+01$P1"+=+ $ M5.1'+0:$7+08"5*#-$+0$M+KDP$45<#21:#$56$1$^*1.+5 V 6*)<+1*$?12-'H$ b+0501$ 45)0'3H$(O $


! $ !"#$% & '" $ ()*'+,+-.+/*+0123$4506#2#0.#$50$7+08"5*#-$10,$'"#$ $ 90:+0##2+0:$;$90<+250=#0'1*$>=/1.'-$56$?12-' @1:#$ Be $ $ & & !#,$7=+'" $ >0,).#,$7+08"5*#$S52=1'+50$D--5.+1'#,$C+'"$>0-'1**1'+50$56$1$ G+:" V @2#--)2#$O1')21*$^1-$@+/#*+0#H$b#-' V 4#0'21*$S*52+,1 $ & & i+15]"#0$$c+10:$ $ 4"121.'#2+]1'+50$56$?12-'$45**1/-#$G1]12,$E1-#,$50$ :25)0,C1'#2 $ 6*).')1'+50-$+0$l+0:3)0$ H$ ^)+:10:H$^)10:L+H$4"+01 $ & & f50:*+$^15 $ >0<#-'+:1'+50-$56$M12:#$7.1*#$7+08"5*#$45**1/-#-H$M1+E+0H$ ^)10:L+H$4"+01 $ & & i+15]"#0$c+10: $ D$.1*+E21'+50$'#-'$56$812-'$.5**1/-#$=50+'52+0:$,#<+.#$E3$T/'+.1*$ !+=#$K5=1+0$P#6*#.'5=#'23$NFT!KPQ$'#."0+J)# $ >_E EA/ & D6+@& N"1,+&T$"**4*G&8%%+4*G & ].'*.P11%*+&L4+@&A',+%1&,%,,4'*^ & $ $ $ $ $ $ $ $ $ $ !@P1,J"5I&8"5&B & $ $ >_2E"/ & \*J%1G1'P*J&+'P1&'(&-",+%&7,'$"+4'*&T4$'+&T$"*+&]-7TT^ U & R&@'P1, & X_DQ"/ & 7*+1'JP.+'15&1%/"1=,_&V%'1G%&c%*4&"*J&d4/ & 0"8'1%"PM&]Q&/4*P+%,^ & X_SE"/ & N%5*'+%&?A%"=%1_& <'1/"+4'*&'(&? 4*=@'$%, & K"+P1"$&"*J& Y *+@1'A'G%*4.& ? 4*=@'$%,_&(1'/& 7 J%*+4(4."+4'*I&+'& ? P1W%54*GI& ? +PJ54*G&"*J& 8 'J%$4*G&"& ? P#+$%& 3 "H"1J & 8"14'&T"14,%&]6E&/4*P+%,^ & $ $ $ $ $ T$%*"15&?%,,4'*_&V%'A@5,4."$&ZMA$'1"+4'*&4*&N"1,+I & 8P,+"("&?"14#PJ"=I&.@"41 & Xa__1= $ (12.)-$^123 $ ^#5/"3-+.1*$+0<#-'+:1'+50-$56$'"#$9,C12,V !2+0+'3$1J) + 6#2$-3-'#=$1'$ =)*'+/*#$-.1*#-a$>0'#2/2#'+0:$1+2E520#$10,$,+2#.' V .)22#0'$2#-+-'+<+'3$+0$ 812-' $ XaB_1= $ !"5=1-$K5E#.8+ $ 7)EE5''5=$/256+*+0:$+0<#-'+:1'+50$56$-+08"5*#$*18#$-'2).')2#$+0$F13$ 10,$b1 -"+0:'50$45)0'+#-H$S*52+,1 $ XaR_1= $ K1<+,$G1225 $ >=/25<#,$+=1:+0:$56$.5<#2#,$812-'$C+'"$'"#$=)*'+ V #*#.'25,#$2#-+-'+<+'3$ +=/*10'$'#."0+J)# $ DE_EE"/ & ; 1%"=&4*&%M@4#4+&@"$$&]SE&/4*P+%,^ & %_aB_1= $ (+."1#*$P).8#2 $ P#.5001+--10.#$#<1*)1'+50$56$1$/5'#0'+1*$6)')2#$ -+08"5*#$)-+0:$ +0'#:21'#,$-+=/*#$-)261.#$:#5/"3-+.-$10,$-)261.#$=50+'52+0:$/5+0'$ %_aR_1= $ @"+*+/$412/#0'#2 $ ^25)0, V /#0#'21'+0:$21,12H$2#-+-'+<+'3$10,$-/50'10#5)-$/5'#0'+1*$ +0<#-'+: 1 '+50-$56$1$.50'1=+01'#,$1J)+6#2$0#12$410.)0H$(#L+.5 $ T$%*"15&?%,,4'*_& <'1/"+4'* & T1'.%,,%, & '(&N"1,+&"*J&?4*=@'$%,&A"1+&D I & 3"115&8''1%I&.@"41 & %%a__1= $ (12.5$d1''105 $ 9L1=/*#-$56$10'"25/5:#0+.$-+08"5*#-$+0$7+.+*3$10,$.5=/12+-50$C+'"$ -+=+*12$/"#05=#01$+0$-5)'"#20$>'1*3 $ %%aB_1= $ 71=$Z/.")2." $ K#<#*5/=#0'$56$-+08"5*#-$+0$1$'"+.8*3$.5<#2#,$812-'$'#2210# $ %%aR_1= $ M+0$(5) $ @1*#5812-'$.2)-'$56$T2,5<+.+10$*+=#-'50#$10,$+'-$.1/1E+*+'3$+0$2#-+-'+0:$ C1'#2$+02)-"#-$+0$.51*$=+0#-$56$052'"$4"+01 $ DS_EE & 0P*.@&]D&@'P1&2E&/4*P+%,^ & $ $ T$%*"15&?%,,4'*_&& <'1/"+4'* & T1'.%,,%, & '(&N"1,+&"*J&?4*=@'$%,&A"1+&SI&3"115&8''1%I&.@"41 & %a&_/= $ 41*<+0$D*#L10,#2 $ K##/$'+=#$52+:+0-$56$-+08"5*#$.5**1/-#$61+*)2#-$+0$-#C1:#$*1:550-$+0$ -5)'"#1-'$(+00#-5'1 $ %aU_/= $ (+0:'10:$M#+ $ 9=#2:#0.3$+0<#-'+:1'+50$56$#L'2#=#*3$*12:#$-+08"5*#-H$(15"#H$ ^)10:L+H$4"+01 $ Ba%_/= $ ^"#52:"#$@50'1 $ ?12-'$*10,652=-$+0$'"#$7121E)2+$^25)/$*+=#-'50#-H$!"1+*10, $


! $ !"#$% & '" $ ()*'+,+-.+/*+0123$4506#2#0.#$50$7+08"5*#-$10,$'"#$ $ 90:+0##2+0:$;$90<+250=#0'1*$>=/1.'-$56$?12-' @1:#$ BW $ $ S_2EA/ & ; 1%"=&4*&%M@4#4+&@"$$&]SE&/4*P+%,^ & T$%*"15&?%,,4'*_&35J1'$'G4."$&Y,A%.+,&'(&N"1,+ I&V%"15& ?.@4*J%$I&.@"41 & BaU_/= $ D201)*,$(1*12, $ (1//+0:$6*55,$2#*1'#,$"1]12,-$+0$812-'$)-+0:$?DP7f7$1//251."A$ D//*+.1'+50$'5$'"#$F#)."+2# V 42#):#01'$812-'$-3-'#=$NcZH$7C+']#2*10,Q $ &a%_/= $ c#00+6#2$D,8+0$ 450.#/')1*+]1'+50$56$^25)0,C1'#2$S*5C$+0$'"#$9,C12,-$DJ)+6#2$ !"25):"$'"#$?0+//1$^1/$G3,25:#5*5:+.$450-'2+.'+50H$Z<1*,#$45)0'3H$ !#L1$ &a&_/= $ O+.5$G1)C#2' $ K#*+0#1'+50$-5)2.#$12#1-$'5$.1<#$,2+/-$10,$.1<#$-'2#1=-$+0$D)-'+0$ !#L1-H$Z7D $ &aU_/= $ F2+10$45C10 $ Z-#$5 6$/"3-+.1*$10,$."#=+.1*$2#-/50-#$+0$.1<#$,2+/-$'5$."121.'#2+]#$ )/*10,$2#."12:#$+0$'"#$F12'50$7/2+0:-$-#:=#0'$56$'"#$9,C12,-$ DJ)+6#2H$4#0'21*$!#L1-H$Z7D $ Ra%_/= $ 7"+*5"$F##=10 $ !"#$O##,$652$@2#-)=/'+<#$G1E+'1'$450-+,#21'+50-$+0$b528+0:$C+'"$ 7)E'#2210#10$ DJ)1'+.$7/#.+#-$56$450.#20a$!"2##$T]128$P#:+50$41-#$ G+-'52+#-H$ZA7ADA $ R_2EA/ & ;"*OP%+_&;4$$&N'.@"*'WI&,A%"=%1 & -@%*&+@%&)"1#'*"+%&T$P/#4*G&V'%,&;"J_ & ?4*=@'$%,I&+@%&35J1"I&"*J&+@%&V%*%1"$&TP#$4. & $ $ $ $ $ $ $ $ $ $ <14J"5I&8"5&DE & $ $ X_DQ"/ & 7*+1' JP.+'15&1%/"1=,& U & V%'1G%&c%*4&"*J&d4/ & 0"8'1%"PM&]Q&/4*P+%,^ & T$%*"15&?%,,4'*_&8"AA4*G&"*J&8"*"G%/%*+&'(&N"1,+&C%G4'*,I & ?"/&T"**'I&.@"41 & YaB_1= $ (12+5$@12+-# $ @5/)*1'+0:$1$."2505*5:+.1*$.1'1*5:)#$56$-+08"5*#-$+0$>'1*3H$1-$'"#$6+2-'$ -'#/$652$1--#--+0:$'"#$-+08"5*#$"1]12, $ YaR_1= $ (12+5$@12+-# $ M#--50-$*#120#,$625=$5..)22#0.#$56$-+08"5*#$2#*1'#,$'5$=10 V =1,#$ .1<+'+#-$+0$1$'5C0$56$75)'"#20$>'1*3 $ Xa__1= $ ?#<+0$!"#)-#0 $ P#-'52+0:$M10,$10,$(101:+0:$?12-'$'5$@25'#.'$b1'#2$l)1*+'3$10,$ l)10'+'3$1'$F12'50$7/2+0:-H$D)-'+0H$!#L1-A $ XaB_1= $ 71=)#* $ @1005 $ !"#$)-#$56$,25):"' V 56$812-'$ '#2210# $ B_6E"/ & ; 1%"=&4*&%M@4#4+&@"$$&]SE&/4*P+%,^ & %_a__1= $ G1223$(552# $ (1//+0:$-)261.#$10,$-)E-)261.#$812-'$:#5"1]12,$ 652$"+:"C13$ /25[#.'-a$7P$W%$75)'"$?05L<+**#$F*<,A$9L'#0-+50H$?05L$45)0'3H$! O $ %_aB_1= $ ^#52:#$d#0+ $ ^5<#20=#0'$410350$7'1'#$O1')21*$D2#1a$D0$9=#2:+0:$(5,#*$652$?12-'$ (101:#=#0' $ %_aR_1= $ 7.5' $ D*#L10,#2 $ 45=E+0+0:$M+KDPH$D#2+1*$@"5'5:21/"3H$10,$@+.'5=#'2+. $ !55*-$652$?12-'$ S#1')2#-$K1'1E1-#$(101:#=#0' $ %%a__1= $ NK10+#* $ K5.'52$10,$c5"0$f5)0:Q $ D0$#<1*)1'+50$56$1)'5=1'#,$^>7$'55*-$652$,#*+0#1'+0:$812-'$-+08"5*#-$ 10, $ .*5-#,$,#/2#--+50-$625=$% V = $ M+KDP V ,#2+<#,$,+:+'1*$#*#<1'+50$,1'1 $ DS_2EA/ & ;P,%,&J%A"1+&('1& )"1$,#"J&)"W%1*,&(4%$J&+14A & $ $ M )0."$1'$412*-E1,$41<#20-$<+-+'52-$.#0'#2 $ $ $ D+2/52'$-")''*#$,#/12'-$652$9*$@1-5$N,#/12')2#$'+=#$!FKQ $ $


! $ !"#$% & '" $ ()*'+,+-.+/*+0123$4506#2#0.#$50$7+08"5*#-$10,$'"#$ $ 90:+0##2+0:$;$90<+250=#0'1*$>=/1.'-$56$?12-' @1:#$ BY $ $ 7*W4+%J&"*J&N%5*'+% & ?A%"=%1, & & 94,+4*GP4,@%J&0%.+P1% _&Y&K)NC7&$%.+P1%&,%14%,&'((%1%J&+'&+@%&G%*%1"$&AP#$4.&"*J&?4*=@'$%& )'*(%1%*.%&"++%*J%%, & 8'*J"5I&8"5&RI&SED2I&>_2EA/ & ?"$+&94,,'$P+4'*&"*J&?4*=@'$%,&4*&-%,+&!%M",&"*J&? 'P+@%",+&K%L&8%M4.' $ ?#00#'"$7A $ c5"0-50 $ ^#5*5:+-'H$O52=10H$T8*1"5=1 $ $ !"+.8$-1*'$,#/5-+'-$"1<#$E##0$,+--5*<#,$+0'#0'+501**3$'5$/25,).#$-1*' V C1'#2$E2+0#-$+0$/12'-$56$b#-'$!#L1-$10,$ -5)'"#1-'$O#C$(#L+.5H$10,$*5.1**3$'"+-$"1-$.2#1'#,$*12:#$)0,#2:25)0,$E2+0# V 6+**#,$.1<+'+#-A$$9*-#C"#2#H$'"#-#$ -1*'-$"1<#$E##0$,+--5*<#,$1..+,#0'1* *3$,)2+0:$/#'25*#)=$5/#21'+50-H$'")-$*#1<+0:$E2+0# V 6+**#,$.1<+'+#-$-+=+*12$'5$ '"5-#$.2#1'#,$+0$E2+0+0:$5/#21'+50-A$$7+0.#$%XY_H$-5=#$56$'"#-#$.1<+'+#-$"1<#$.5**1/-#,$10,$.2#1'#,$*12:#$ -+08"5*#-$0#12$b+08H$!iH$10,$'"#$cb7$10,$M5.5$G+**-$-+08"5*#-$+0$O(A$$ $ 4 5**1/-#$56$'"#-#$#12*+#2$.1<+'+#-$"1-$21+-#,$.50.#20-$1E5)'$1$-+=+*12$E2+0#$.1<+'3$+0$412*-E1,H$O(A$$!"#$>;b$ -+=+*12+'3$'5$'"#$cb7$10,$M5.5$G+** -$E2+0+0:$5/#21'+50-A$$D$-#2+#-$56$:#5*5:+.$10,$:#5/"3-+.1*$-'),+#-$"1<#$ ."121.'#2+]#,$'"#$/25E1E*#$-+]#H$-"1/#H$10,$,#/'"$56$'"#$)0,#2:25)0,$ .1<+'3 $ 10,$'"#$-+'#$+-$05C$E#+0:$ =50+'52#,$.50-'10'*3$'5$/25<+,#$#12*3$C120+0:-$+0$'"#$#<#0'$56$.5**1/-#$56$'"#$.1 <+'3A$$ $ ;4'G1"A@5_ & K2A$?#00#'"$7A$c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


! $ !"#$% & '" $ ()*'+,+-.+/*+0123$4506#2#0.#$50$7+08"5*#-$10,$'"#$ $ 90:+0##2+0:$;$90<+250=#0'1*$>=/1.'-$56$?12-' @1:#$ BX $ $ -%J*%,J"5I&8"5&XI&SED 2 I&X_2Q"/ & ?A%$%'$'G4."$I&35J1'G%'$'G4."$&"*J&Z*G4*%%14*G&V%'$'G4."$& )@"$$%*G%,&'(&!P**%$4*G&4*& N"1,+& Y1%", & & Z0+<#2-+'3$56$k1:2#EH$4251'+1 & >0$'"#$4*1--+.1*$K+012+.$?12-'$56$4251'+1H$5<#2$%%HU__$.1<#-$"1<#$E##0$#L/*52#,$ -5$612H$=52#$'"10$%H___$56$C"+."$C#2#$,+-.5<#2#,$,)2+0:$.50-'2).'+50$C528-A$$ 41<#-$,+-.5<#2#,$50$'"#$.50-'2).'+50$-+'#-$56$"+:"C13-$*1.8#,$01')21*$ #0'210.#-$50$'"#$-)261.#A$T<# 2$'"#$/1-'$B_$3#12-$'"#3$"1<#$E##0$ -3-'#=1'+.1**3$+0<#-'+:1'#,$10,$2#=#,+1'#,$'5$1**5C$.5=/*#'+50$56$'"#$251,-A$ 75=#$-/#.+1*$#L1=/*#-$C+**$E#$/2#-#0'#,$,)2+0:$'"#$*#.')2#H$-)."$1-$'"#$*12:#$ "1**$+0$'"#$d21'1$!)00#*$56$'"#$k1:2#E$ $ P+[#81$G+:"C13H$10,$.1<#$ *50:#-'$')00#*-A$K)#$'5$'"#$-+]#H$-"1/#H$/5-+'+50H$10,$"3,25:#5*5:+.1*$ /121=#'#2-$56$'"#$.1<#$C+'"+0$'"#$812-'$-3-'#=H$+'$C1-$0#.#--123$'5$,#-+:0$10,$ "1,$'5$E#$2#+0652.#,$ 10,$-'1E+*+]#,A$!"+-$/2#-#0'1'+50$C+**$+0.*),#$<+,#5$10,$/"5'5-$56$'"#$=5-'$+0'#2#-'+0:$812-'$10,$.1<#$*5.1'+50-$ +0$K+012+.$?12-'A $ ;4'G1"A@5_ & +1H$+0$%XU%H$K2A$ :#5-.+#0.#-$10,$:#5*5:+.1*$#0:+0##2+0:$+0$%XYeA$G#$+-$1$-.+#0'+-'H$10,$1$/256#--52$56$:#5*5:3H$812-'$ "3,25:#5*5:3H$1//*+#,$:#5*5:3H$#0:+0 ##2+0:$:#5*5:3$10,$-/#*#5*5:3$1'$'"#$Z0+<#2-+'3$56$k1:2#EH$10,$"1-$ 1)'"52#,$=52#$'"10$&&_$-.+#0'+6+.$10,$/256#--+501*$/1/#2-A$G#$-#2<#-$1-$1$.5==+''##$=#=E#2$652$'"#$ 4251'+10$D.1,#=3$56$7.+#0.#$10,$D2'-H$ZO974T$b52*,$G#2+'1:#$!#1=$652$'"#$K+012+.$?12-'H$>0 '#201'+501*$ D--5.+1'+50$56$G3,25:#5*5:+-'-H$10,$>0'#201'+501*$D--5.+1'+50$652$90:+0##2+0:$^#5*5:3$10,$'"#$90<+250=#0'A & =#=E#2$56$'"#$4251'+10$(5)0'1+ 0##2+0:$D--5.+1'+50$-+0.#$%XUU$10,$C1-$1C12,#,$E3$'"#$D--5.+1'+50$+0$%XeX$ 10,$%XY%A$G#$-'12'#,$.1<+0:$+0$%Xe&$10,$+-$'"#$65)0,#2$10,$/2#-+,#0'$56$-#<#21*$.1<+0:$.*)E-$+0$4251'+1A$G#$ -#2<#,$1-$6+2-'$/2#-+,#0'$56$'"#$4251'+10$7/#*#5*5:+.1*$S#,#21'+50$625=$%X X_$'5$B_%_$10,$+-$1$*+6#$=#=E#2$56$ Z0+50$56$7/#*#5*5:3$10,$'5$'"#$9)25/#10$7/#*#5*5:+.1*$S#,#21'+50$E#:+00+0:$+0$B__XA$ $ 2#-#12."$+0H$10,$#L/*52#,$10,$<+-+'#,$0#12*3$UH___$.1<#-$+0$eR$.5)0'2+#-A$G#$"1-$*#,$ =103$-/#*#5*5:+.1*$#L/#,+'+50-$+0$'"#$*50:#-'$10,$,##/#-'$.1<#-$+0$4251'+1H$9)25/#H$10,$'"#$C52*,A$G#$"1-$1*-5$ -'),+#-$1E5)'$%H___$.1<#-$C+'"5)'$01')21*$#0'210.#-H$,+-.5<# 2#,$E3$')00#*-$10,$J)122+#-H$10,$#<1*)1'#,$'"#+2$ "3,25:#5*5:3$10,$#0:+0##2+0:$:#5*5:3A $


! $ !"#$% & '" $ ()*'+,+-.+/*+0123$4506#2#0.#$50$7+08"5*#-$10,$'"#$ $ 90:+0##2+0:$;$90<+250=#0'1*$>=/1.'-$56$?12-' @1:#$ &_ $ $ -%J*%,J"5I&8"5&XI&SED 2 & D_2EA/ & V%'$'G4.&FW%1W4%L&'(&+@%&T%1/4"*&;",4*&'(&?'P+@%",+%1*&K%L& 8%M4.'&"*J&%,+&!%M", & & 3"1W%5&C:&9P)@%*% & d#.'1$T+*$10,$^1-H$M@ $ & 412*-E1,H$O#C$(#L+.5$+-$*5.1'#,$+0$'"#$052'" V .#0'21*$/12'$56$'"#$K#*1C12#$ E1-+0H$C"+."$+-$/12'$56$'"#$:2#1'#2$@#2=+10$E1-+0H$50#$56$'"#$2+."#-'$5+*$10,$:1-$ /25<+0.#-$+0$'"#$C52*,A$>0$M1'#$(+--+--+//+10$'5$912*3$@#00-3*<10+10$'+=#H$'"#$ .5**+-+50$56$75)'"$D= #2+.1 V D62+.1$C+'"$-5)'"#20$O52'"$D=#2+.1$.1)-#,$'"#$ T)1."+'1 V (121'"50$525:#03$10,$652=#,$'"#$,##/$-'2).')21*$E1-+0-$10,$)/*+6'-$ 56$'"#$10.#-'21*$P5.83$(5)0'1+0-A$(1'#2+1*$#25,#,$625=$)/*+6'-$/25:2#--+<#*3$ 6+**#,$E1-+0-$,)2+0:$@#00-3*<10+10$10,$@#2=+10$'+=# A$(1[52$-'2).')21*$.5=/50#0'-$56$'"#$10.#-'21*$P5.8+#-$+0$ '"#$@#2=+10$E1-+0$2#:+50$12#$'"#$(+,*10,$10,$K#*1C12#$E1-+0-H$-#/121'#,$E3$'"#$4#0'21*$F1-+0$/*1'652=A$$ $ $ @1*#5]5+.$-'21'+:21/"3$56$'"#$2#:+50$.10$E#$,+<+,#,$+0'5$'C5$E251,$/1.81:#-A$T*,#2$@1*#5]5+. $ -'21'1$210:#$625=$ 41=E2+10$'5$(+--+--+//+10$1:#$10,$C#2#$,#/5-+'#,$+0$10$10.#-'21*$,#/2#--+50$805C0$1-$'"#$!5E5-1$E1-+0A$ !"#-#$-'21'1$12#$=5-'*3$.12E501'#-$10,$12#$'2)0.1'#,$E3$1$=1[52$M1'#$(+--+--+//+10$'5$912*3$@#00-3*<10+10$ 2#:+501*$)0.50652=+'3A$$7'21 '1$56$@#00-3*<10+10$10,$@#2=+10$1:#$C#2#$,#/5-+'#,$+0$E1-+0-$652=#,$,)2+0:$'"#$ (121'"50 V T)1."+'1$525:#03A$!"#$-#,+=#0'-$.50-+-'#,$56$+0'#2E#,,#,H$52:10+. V 2+."$-"1*#H$-10,-'50#$10,$ .12E501'#$,#/5-+'#,$+0$E1-+0-$2+==#,$E3$.12E501'#$-"#*6$=12:+0-A$!3/+.1**3H$ '+=# V #J)+<1*#0'$-'21'1$56$ @#00-3*<10+10$10,$@#2=+10$1:#$:21,#$625=$=+L#,$*+'"5*5:+#-$+0$E1-+0$.#0'#2-$'5$.12E501'#$50$-"#*6$=12:+0-$'5$ =+L#,$#<1/52+'#$10,$.*1-'+.$-#,+=#0'-$50$'"#$E1-+0$-"#*6A$!5C12,$'"#$#0,$56$@#2=+10$'+=#H$'"+.8$E#,-$56$ #<1/52+'#-$C#2#$, #/5-+'#,$1'5/$5*,#2$.12E501'#$10,$.*1-'+.$-'21'1$1.25--$'"#$2#:+50A $ $ 7'21'1$+0$'"#$@#2=+10$E1-+0$C#2#$2#/#1'#,*3$+=/1.'#,$E3$."10:#-$+0$.*+=1'#$10,$-#1$*#<#*A$450,+'+50-$-)+'#,$ '5$.12E501'#$,#/5-+'+50$,5=+01'#,$,)2+0:$'+=#-$56$"+:"$-#1$*#<#*$C+'"$#25-+50$5 ..)22+0:$,)2+0:$'+=#-$C"#0$ -#1$*#<#*$C1-$*5CA$M5C$-'10,-$C#2#$.50,).+<#$'5$,+--5*)'+50$56$.12E501'#$10,$'"#$,#<#*5/=#0'$56$812-'A$(103$ #0"10.#,$/525-+'3$1 0,$/#2=#1E+*+'3A$(52#$.5==50*3H$/1*#5812-'$0#:1'+<#*3$+=/1.'#,$"3,25.12E50$2#-#2<5+2-$ E3$.1)-+0:$#12*3$#0.251."=#0'$56$C1'#2$1*50:$621.')2#-H$52$E3$5..*),+0:$=1'2+L$/525-+'3$10,$6+**+0:$621.')2#-$ C+'"$-#.50,123$.#=#0'$2#-)*'+0:$+0$.5=/12'=#0'1*+]1'+50A$ $ $ K #<#*5/=#0'$56$'"#$=5,#20$*10,-.1/#$E#:10$C+'"$'"#$50-#'$56$M121=+,#$'#.'50+-=$+0$M1'#$42#'1.#5)-$10,$ 912*3$!#2'+123$'+=#A$P#:+501*$.5=/2#--+50$.1)-#,$'"#$:21,)1*$2+-#$56$1$052'" V '2#0,+0:$12."$'"1'$#L'#0,#,$625=$ b35=+0:$'5$052'"#20$(#L+.5$805C0$1-$'"#$D*<12 1,5$P+,:#A$Z/*+6'$.)*=+01'#,$1//25L+=1'#*3$&U$=1$C"#0$ '"#$1L+-$56$'"#$D*<121,5$P+,:#A $ $


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eYQ$10,$(A7A$N%XW&Q$,#:2##-$+0$ :#5*5:3A$G#$"1-$&X$3#12-$56$#L/#2+#0.#$1-$1$/#'2 5*#)=$:#5*5:+-'H$C528+0:$652$D=5.5$@25,).'+50$45=/103H$ K1<+-$T+*H$DL#=$P#-5)2.#-$10,$5'"#2-A$G#$.)22#0'*3$+-$1$*+=+'#,$/12'0#2$+0$d#.'1$T+*$10,$^1-H$M@H$10$5+*$10,$:1-$ #L/*521'+50$10,$/25,).'+50$.5=/103$"#1,J)12'#2#,$+0$K1**1-H$!#L1-A$G+-$/2+=123$12#1$56$# L/#2'+-#$+-$ /#'25*#)=$#L/*521'+50$+0$E1-+0-$56$'"#$P5.83$(5)0'1+0$/25<+0.#$10,$C#-'$!#L1-H$C+'"$1,,+'+501*$#L/#2+#0.#$+0$ '"#$=+,.50'+0#0'H$D//1*1."+10$E1-+0$10,$566-"52#$b#-'$D62+.1A $ $ $ G12<#3$"1-$1*-5$=52#$'"10$&_$3#12-$-'),3+0:$'"#$-/#*#5:#0#-+-$56$"3/5:# 0+.$.1<#-H$/12'+.)*12*3$'"5-#$652=#,$ E3$-)*6)2+.$1.+,A$$G#$+-$+0'#2#-'#,$+0$'"#$.500#.'+50$E#'C##0$'"#$#<5*)'+50$56$"3/5:#0+.$.1<#$-3-'#=-$10,$'"#$ '#.'50+.$10,$:#5*5:+.$"+-'523$56$2#:+50-A $ $ G12<#3$+-$1$=#=E#2$56$'"#$^#5*5:+.1*$75.+#'3$56$D=#2+.1H$D=#2+.10$ D--5.+1'+50$56$@#'25*#)=$^#5*5:+-'-H$ D=#2+.10$^#5*5:+.1*$>0-'+')'#H$P5.83$(5)0'1+0$D--5.+1'+50$56$^#5*5:+-'-H$O#C$(#L+.5$^#5*5:+.1*$75.+#'3H$ b#-'$!#L1-$^#5*5:+.1*$D--5.+1'+50$10,$?12-'$b1'#2-$>0-'+')'#H$10,$"#$+-$1$S#**5C$56$'"#$O1'+501*$7/#*#5*5:+.1*$ 75.+ #'3A $ & & &


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`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a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`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


! $ !"#$% & '" $ ()*'+,+-.+/*+0123$4506#2#0.#$50$7+08"5*#-$10,$'"#$ $ 90:+0##2+0:$;$90<+250=#0'1*$>=/1.'-$56$?12-' @1:#$ &R $ $ ;4'G1"A@5_ & b+**+1=$NF+**Q$?5."105<$N/2505)0.#,$?T V V TSSQ$+-$1$7#0+52$^#5*5:+-'$C+'"$'"#$@#00-3*<10+1 $ K#/12'=#0'$ 56$450-#2<1'+50$10,$O1')21*$P#-5)2.#-H$F)2#1)$56$!5/5:21/"+.$10,$^#5*5:+.$7)2<#3H$^#5*5:+.$(1//+0:$ K+<+-+50A$7+0.#$%XYUH$"#$"1-$E##0$1.'+<#*3$=1//+0:$:#5*5:+.1**3$"1]12,5)-$12#1-$C+'"+0$'"#$*+=#-'50#$2#:+50-$ 56$@#00-3*<10+1$10,$=1+0'1+0-$'"#$F) (5)0'1+0-$P#:+50$'5$'"#$4"#-'#2$d1**#3$56$-5)'"#1-'#20$@#00-3*<10+1A$F+**$+-$=5-'$05'#,$6 52$1)'"52+0:$'"#$ -#2+#-$56$.5)0'3$2#/52'-H$-/#.+6+.1**3$,#-+:0#,$'5$."121.'#2+]#$812-'$-)261.#$6#1')2#-H$'"#+2$,+-'2+E)'+50H$10,$ '"#$:#5*5:3$5 G122+-E)2:$C+'"$"+-$C+6#$c10#$10,$."+*,2#0H$O1'1*+#$10,$D*#LH$.*5-#$'5$'"#$652#-'-$56$7'503$42##8$C"#2#$"#$ -/#0,-$=103$"5)2-$'21.8+0:$,5C0$'"#$#*)-+<#$#,+E*#-$ 56$'"#$=)-"255=$C52*,A & $

Sinkholes and the Engineering and Environmental
Impacts of Karst
Proceedings of the Thirteenth Multidisciplinary
Carlsbad, New Mexico, May 6-10, 2013 / Lewis Land,
Daniel H. Doctor, J. Brad Stephenson, Editors.
Keynote: Speleological, Hydrogeological and
Engineering Geological Challenges Of Tunneling in Karst Areas
/ Mladen Garaic --
Natural and Anthropogenic Sinkholes: From
Identification, to Surveying, Studying and Modeling a Subtle
Hazard / Mario Parise --
Tectonic Influences on Petroleum Migration and
Speleogenesis in the Guadalupe Mountains, New Mexico and
Texas / Harvey R. Duchene --
When the Carbonate Plumbing Goes Bad: Sinkholes, the
Hydra, and the General Public / William Kochanov.
Session I: Engineering and Geotechnical Aspects of
Towards a
karst assessment standard practice / Robert K. Denton,
case history for sinkhole investigation and stabilization
methods along a high pressure petroleum pipeline / John T.
Pusey, Jr. and John M. Caccese--
associated with the use of compaction grout for sinkhole
remediation in west-central Florida / Edward D. Zisman and
Daniel J. Clarey--
karst risk at proposed windpower projects / William J.
Bangsund and Kenneth S. Johnson--
of stability charts and reliability concepts for simplified
analysis of a void in soil overlying karst bedrock / Timothy
C. Siegel, Danner F. Drake, and Eric C. Drumm--
If it's
weight of hammer conditions, it must be a sinkhole? / Edward
D. Zisman and Daniel J. Clarey--
grouting of a subsurface detention/infiltration system /
Joseph A. Fischer, Todd K. Miller, Michael J. Miluski, and
Joseph J. Fischer--
Need for a
standardized approach to characterizing, permitting, and
constructing landfills in karst geologic settings / Robert C.
Bachus and Richard B. Tedder--
A calibration
test of karst collapse monitoring device by optical time
domain reflectometry (BOTDR) technique [Poster] / Guan
Zhende, Jiang Xiaozhen, and Gao Ming--
sinkhole formation associated with installation of a
high-pressure natural gas pipeline, west-central Florida
[Poster] / Ted J. Smith and George C. Sinn--
Cover-collapse sinkhole development in the Cretaceous Edwards
limestone, central Texas [Poster] / Brian B. Hunt, Brian A.
Smith, Mark T. Adams, Scott E. Hiers, and Nick Brown.
Session II: Evaporite Karst
Salt karst
and collapse structures in the Anadarko Basin of Oklahoma and
Texas / Kenneth S. Johnson--
karst in the Permian Basin region of west Texas and
southeastern New Mexico: The human impact / Lewis Land--
karst and hydrogeology of the Castile Formation: Culberson
County, Texas and Eddy County, New Mexico / Kevin W.
Gypsum karst
causes relocation of proposed Cedar Ridge Dam, Throckmorton
County, Texas / Kenneth S. Johnson and J. Mark
The role of
sulfate-rich springs and groundwater in the formation of
sinkholes over gypsum in eastern England / Anthony H. Cooper,
Noelle E. Odling, Phillip J. Murphy, Claire Miller,
Christopher J. Greenwood, and David S. Brown--
Gypsum karst
and potential risk in siting wind turbines in Blaine County,
Oklahoma / Kenneth S. Johnson, William J. Bangsund, and Neal
A. Hines--
karst in the Black Hills, South Dakota and Wyoming, and the
oil play in the Williston Basin, North Dakota and Montana /
Jack B. Epstein and Daniel H. Doctor--
Variations in
evaporite karst in the Holbrook Basin, Arizona [Poster] /
James T. Neal, Kenneth S. Johnson, and Paul Lindberg--
evaporite karst activity and land subsidence in the Holbrook
Basin, Arizona using interferometric synthetic aperture radar
(InSAR) [Poster] / Brian D. Conway and Joseph P.
Session III: Geophysical Investigations in Karst
investigations of the Edwards-Trinity aquifer system at
multiple scales: Interpreting airborne and direct-current
resistivity in karst / Marcus O. Gary, Dale F. Rucker, Bruce
D. Smith, David V. Smith, and Kevin Befus--
profiling investigation of sinkhole lake structure in Bay and
Washington Counties, Florida / Thomas L. Dobecki, Sam B.
Upchurch, Thomas M. Scott, Beth Fratesi, and Michael C.
imaging of covered karst with the multi-electrode resistivity
implant technique / David Harro and Sarah Kruse--
Reconnaissance evaluation of a potential future sinkhole
using integrated simple surface geophysics and surface
monitoring points / Michael L. Rucker, Sean Hulburt, and Mark
D. Edwards--
Ground-penetrating radar, resistivity and spontaneous
potential investigations of a contaminated aquifer near
Cancn, Mexico / Philip J. Carpenter, Ryan F. Adams, Melissa
Lenczewski, and Rosa M. Leal-Bautista--
methods for forecasting karst collapse in China [Poster] /
Yan Meng, Jianling Dai, Long Jia, Mingtang Lei, and Feng
analysis of GPR and SPT methods for sinkhole investigation in
covered karst terrain, west-central Florida, USA [Poster] /
Henok Kiflu, Michael Wightman, and Sarah Kruse--
geophysical methods for groundwater exploration in a karst
area with or without thin cover a case study from Tai'an
City, Shandong Province, China [Poster] / Fuping Gan, Yixiang
Chen, Wei Zhao, Yuling Chen, and Wei Liu.
Session IV: Formation Processes of Karst and
Examples of
anthropogenic sinkholes in Sicily and comparison with similar
phenomena in southern Italy / Marco Vattano, Mario Parise,
Piernicola Lollino, Marco Bonamini, Di Maggio, and Giuliana
of sinkholes in a thickly covered karst terrane / Sam B.
Upchurch, Thomas L. Dobecki, Thomas M. Scott, Steven H.
Meiggs, Sarah E. Fratesi, and Michael C. Alfieri--
crust of Ordovician limestone and its capability in resisting
water inrushes in coal mines of north China / Lin Mou and
Gongyu Li--
Deep time
origins of sinkhole collapse failures in sewage lagoons in
southeast Minnesota / E. Calvin Alexander, Jr., Anthony C.
Runkel, Robert G. Tipping, and Jeffrey A. Green--
investigation of extremely large sinkholes, Maohe, Guangxi,
China / Mingtang Lei, Yongli Gao, Xiaozhen Jiang, and Zhende
landforms in the Saraburi Group limestones, Thailand /
Gheorghe Ponta, Bashir Memon, James LaMoreaux, Jade Julawong,
and Somchai Wongsawat--
sinkhole and pseudokarst development in east Texas [Poster] /
Kevin W. Stafford, Melinda G. Shaw-Faulkner, and Wesley A.
Characterization of karst collapse hazard based on
groundwater fluctuations in Qingyun Village, Guigang,
Guangxi, China [Poster] / Xiaozhen Jiang, Mingtang Lei ,
Yongli Gao, and Zhende Guan--
Investigations of large scale sinkhole collapses, Laibin,
Guangxi, China [Poster] / Yongli Gao, Weiquan Luo, Xiaozhen
Jiang, Mingtang Lei, and Jianling Dai.
Sessions V: Karst Hydrology
flood-related hazards in karst using the KARSYS approach:
Application to the Beuchire-Creugenat karst system (Ju,
Switzerland) / Jonathan Vouillamoz, Arnauld Malard, Gabrielle
Schwab Rouge, Eric Weber, and Pierre-Yves Jeannin--
Conceptualization of groundwater flow in the Edwards Aquifer
through the Knippa Gap hydrogeologic constriction, Uvalde
County, Texas / Jennifer Adkins--
source areas to cave drips and cave streams in Austin Texas,
USA / Nico Hauwert and Brian Cowan--
Use of
physical and chemical response in cave drips to characterize
upland recharge in the Barton Springs segment of the Edwards
Aquifer, central Texas, USA / Brian Cowan and Nico
The need for
presumptive habitat considerations in working with
subterranean aquatic species of concern: Three Ozark region
case histories, USA / Shiloh L. Beeman, Thomas J. Aley, and
Michael Slay.
Part VI: Mapping and Management of Karst
chronological catalogue of sinkholes in Italy: The first step
toward a real evaluation of the sinkhole hazard / Mario
Parise and Carmela Vennari--
learned from occurrence of sinkholes related to man-made
cavities in a town of southern Italy / Pietro Pepe, Nunzia
Pentimone, Giuditta Garziano, Vincenzo Martimucci, and Mario
land and managing karst to protect water quality and quantity
at Barton Springs, Austin, Texas / Kevin Thuesen--
The use of
drought-induced "crop lines" as a tool for characterization
of karst terrain / Samuel V. Panno, Donald E. Luman, Walton
R. Kelly, and Matthew B. Alschuler--
surface and subsurface karst geohazards for highway projects:
SR 71 South Knoxville Boulevard Extension, Knox County,
Tennessee / Harry L. Moore--
Canyon State Natural Area: An emerging model for karst
management / George Veni--
LiDAR, aerial photography, and Pictometry tools for karst
features database management / Scott C. Alexander, Mina
Rahimi, Erik Larson, Cody Bomberger, Brittany Greenwaldt, and
E. Calvin Alexander, Jr.--
An evaluation
of automated GIS tools for delineating karst sinkholes and
closed depressions from 1-meter LiDAR-derived digital
elevation data / Daniel H. Doctor and John A. Young--
and classification of karst depressions using LiDAR: Fort
Hood military installation, Texas [Poster] / Melinda G. Shaw
Faulkner, Kevin W. Stafford, and Aaron W. Bryant--
sinkholes in LiDAR coverage of a glacio-fluvial karst, Winona
County, MN [Poster] / Mina Rahimi and E. Calvin Alexander,
Field Trip Guide
Evaporite Karst of the Lower Pecos Valley, New Mexico /
L. Land.


NATIONAL CAVE AND KARST RESEARCH INSTITUTESYMPOSIUM 2SINKHOLES AND THE ENGINEERING AND ENVIRONMENTAL IMPACTS OF KARST PROCEEDINGS OF THE THIRTEENTH MULTIDISCIPLINARY CONFERENCE May 6 through 10, 2013 Carlsbad, New Mexico EDITORS:Lewis LandNew Mexico Bureau of Geology and Mineral Resources and National Cave and Karst Research InstituteDaniel H. DoctorU.S. Geological SurveyJ. Brad Stephenson CB&I


Published and distributed by National Cave and Karst Research InstituteDr. George Veni, Executive Director Peer-review: Organizing Committee of the Thirteenth Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst. The citation information: Land L, Doctor DH, Stephenson JB, editors. 2013. Sinkholes and the Engineering and Environmental Impacts of Karst: Proceedings of the Thirteenth Multidisciplinary Conference, May 6-10, Carlsbad, New Mexico: NCKRI Symposium 2. Carlsbad (NM): National Cave and Karst Research Institute. ISBN 978-0-9795422-7-5TECHNICAL PROGRAM CHAIRS AND EDITORS Lynn B. Yuhr Technos Inc. Lewis Land New Mexico Bureau of Geology and Mineral Resources and the National Cave and Karst Research Institute Daniel H. Doctor United States Geological Survey J. Brad Stephenson CB&ICover Photo: Aerial view of the JWS sinkhole, Eddy Co., New Mexico, about six weeks after initial collapse. Photo compliments of the National Cave and Karst Research Institute.


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2III CONTENTSOrganizing Committee .............................................................................................. IX Foreword ......................................................................................................................... XI Keynote Speakers .............................................................................................. XII-XVI Engineering and Geotechnical Aspects of Karst Towards a karst assessment standard practice Robert K. Denton, Jr. ........................................................................................................1-12 Geotechnical case history for sinkhole investigation and stabilization methods along a high pressure petroleum pipeline John T. Pusey, Jr. and John M. Caccese .........................................................................13-21 Problems associated with the use of compaction grout for sinkhole remediation in west-central Florida Edward D. Zisman and Daniel J. Clarey ........................................................................23-26 Evaluating karst risk at proposed windpower projects William J. Bangsund and Kenneth S. Johnson ................................................................27-36 Application of stability charts and reliability concepts for simplified analysis of a void in soil overlying karst bedrock Timothy C. Siegel, Danner F. Drake, and Eric C. Drumm .............................................37-44 Edward D. Zisman and Daniel J. Clarey ........................................................................45-52 Exploratory grouting of a subsurface detention/infiltration system Joseph A. Fischer, Todd K. Miller, Michael J. Miluski, and Joseph J. Fischer ..............53-59 Need for a standardized approach to characterizing, permitting, and constructing landfills in karst geologic settings Robert C. Bachus and Richard B. Tedder .......................................................................61-69 A calibration test of karst collapse monitoring device by optical time domain reflectometry (BOTDR) technique [Poster] Guan Zhende, Jiang Xiaozhen, and Gao Ming ...............................................................71-77


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEIV Induced sinkhole formation associated with installation of a high-pressure natural gas pipeline, west-central Florida [Poster] Ted J. Smith and George C. Sinn ....................................................................................79-88 Cover-collapse sinkhole development in the Cretaceous Edwards limestone, central Texas [Poster] Brian B. Hunt, Brian A. Smith, Mark T. Adams, Scott E. Hiers, and Nick Brown ........89-102 Evaporite Karst Salt karst and collapse structures in the Anadarko Basin of Oklahoma and Texas Kenneth S. Johnson ..................................................................................................... 103-112 Evaporite karst in the Permian Basin region of west Texas and southeastern New Mexico: The human impact Lewis Land .................................................................................................................. 113-121 Evaporite karst and hydrogeology of the Castile Formation: Culberson County, Texas and Eddy County, New Mexico Kevin W. Stafford ........................................................................................................123-131 Gypsum karst causes relocation of proposed Cedar Ridge Dam, Throckmorton County, Texas Kenneth S. Johnson and J. Mark Wilkerson ...............................................................133-139 The role of sulfate-rich springs and groundwater in the formation of sinkholes over gypsum in eastern England Anthony H. Cooper, Noelle E. Odling, Phillip J. Murphy, Claire Miller, Christopher J. Greenwood, and David S. Brown ........................................................141-150 Gypsum karst and potential risk in siting wind turbines in Blaine County, Oklahoma Kenneth S. Johnson, William J. Bangsund, and Neal A. Hines ..................................151-159 Evaporite karst in the Black Hills, South Dakota and Wyoming, and the oil play in the Williston Basin, North Dakota and Montana Jack B. Epstein and Daniel H. Doctor ........................................................................161-175 Variations in evaporite karst in the Holbrook Basin, Arizona [Poster] James T. Neal, Kenneth S. Johnson, and Paul Lindberg ............................................177-186


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2V Monitoring evaporite karst activity and land subsidence in the Holbrook Basin, Arizona using interferometric synthetic aperture radar (InSAR) [Poster] Brian D. Conway and Joseph P. Cook ........................................................................187-194 Geophysical Investigations in Karst Terrain Geophysical investigations of the Edwards-Trinity aquifer system at multiple scales: Interpreting airborne and direct-current resistivity in karst Marcus O. Gary, Dale F. Rucker, Bruce D. Smith, David V. Smith, and Kevin Befus ..........................................................................................................195-206 Subbottom profiling investigation of sinkhole lake structure in Bay and Washington Counties, Florida Thomas L. Dobecki, Sam B. Upchurch, Thomas M. Scott, Beth Fratesi, ................................................................................................. 207-211 Improved imaging of covered karst with the multi-electrode resistivity implant technique David Harro and Sarah Kruse ....................................................................................213-219 Reconnaissance evaluation of a potential future sinkhole using integrated simple surface geophysics and surface monitoring points Michael L. Rucker, Sean Hulburt, and Mark D. Edwards ..........................................221-229 Ground-penetrating radar, resistivity and spontaneous potential investigations of a contaminated aquifer near Cancn, Mexico Philip J. Carpenter, Ryan F. Adams, Melissa Lenczewski, and Rosa M. Leal-Bautista .........................................................................................231-237 Typical methods for forecasting karst collapse in China [Poster] Yan Meng, Jianling Dai, Long Jia, Mingtang Lei, and Feng Ji ..................................239-245 Statistical analysis of GPR and SPT methods for sinkhole investigation in covered karst terrain, west-central Florida, USA [Poster] ....................................................247-253 Integrated geophysical methods for groundwater exploration in a Shandong Province, China [Poster] Fuping Gan, Yixiang Chen, Wei Zhao, Yuling Chen, and Wei Liu ..............................255-261


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEVI Formation Processes of Karst and Sinkholes Examples of anthropogenic sinkholes in Sicily and comparison with similar phenomena in southern Italy Marco Vattano, Mario Parise, Piernicola Lollino, Marco Bonamini, Di Maggio, and Giuliana Madonia ................................................................................................263-271 Development of sinkholes in a thickly covered karst terrane Sam B. Upchurch, Thomas L. Dobecki, Thomas M. Scott, Steven H. Meiggs, ....................................................................273-277 Paleokarst crust of Ordovician limestone and its capability in resisting water inrushes in coal mines of north China Lin Mou and Gongyu Li ..............................................................................................279-284 Deep time origins of sinkhole collapse failures in sewage lagoons in southeast Minnesota E. Calvin Alexander, Jr., Anthony C. Runkel, Robert G. Tipping, and Jeffrey A. Green ....................................................................................................285-292 Emergency investigation of extremely large sinkholes, Maohe, Guangxi, China Mingtang Lei, Yongli Gao, Xiaozhen Jiang, and Zhende Guan .................................293-297 Karst landforms in the Saraburi Group limestones, Thailand Gheorghe Ponta, Bashir Memon, James LaMoreaux, Jade Julawong, and Somchai Wongsawat ............................................................................................299-309 Clastic sinkhole and pseudokarst development in east Texas [Poster] Kevin W. Stafford, Melinda G. Shaw-Faulkner, and Wesley A. Brown ....................... 311-319 Characterization of karst collapse hazard based on groundwater fluctuations in Qingyun Village, Guigang, Guangxi, China [Poster] Xiaozhen Jiang, Mingtang Lei Yongli Gao, and Zhende Guan ................................321-326 Investigations of large scale sinkhole collapses, Laibin, Guangxi, China [Poster] Yongli Gao, Weiquan Luo, Xiaozhen Jiang, Mingtang Lei, and Jianling Dai ............327-331


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2VII Karst Hydrology Mapping flood-related hazards in karst using the KARSYS approach: Application to the Beuchire-Creugenat karst system (Ju, Switzerland) Jonathan Vouillamoz, Arnauld Malard, Gabrielle Schwab Rouge, Eric Weber, and Pierre-Yves Jeannin ..............................................................................................333-342 Conceptualization of groundwater flow in the Edwards Aquifer through the Knippa Gap hydrogeologic constriction, Uvalde County, Texas Jennifer Adkins ............................................................................................................343-352 Delineating source areas to cave drips and cave streams in Austin Texas, USA Nico Hauwert and Brian Cowan .................................................................................353-365 Use of physical and chemical response in cave drips to characterize upland recharge in the Barton Springs segment of the Edwards Aquifer, central Texas, USA Brian Cowan and Nico Hauwert .................................................................................367-375 The need for presumptive habitat considerations in working with subterranean aquatic species of concern: Three Ozark region case histories, USA Shiloh L. Beeman, Thomas J. Aley, and Michael Slay ................................................377-381 Mapping and Management of Karst Regions A chronological catalogue of sinkholes in Italy: The first step toward a real evaluation of the sinkhole hazard Mario Parise and Carmela Vennari ............................................................................383-392 Lessons learned from occurrence of sinkholes related to man-made cavities in a town of southern Italy Pietro Pepe, Nunzia Pentimone, Giuditta Garziano, Vincenzo Martimucci, and Mario Parise ........................................................................................................393-401 Restoring land and managing karst to protect water quality and quantity at Barton Springs, Austin, Texas Kevin Thuesen .............................................................................................................403-409


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEVIII characterization of karst terrain Samuel V. Panno, Donald E. Luman, Walton R. Kelly, and Matthew B. Alschuler ........................................................................................... 411-419 Mapping surface and subsurface karst geohazards for highway projects: SR 71 South Knoxville Boulevard Extension, Knox County, Tennessee Harry L. Moore ...........................................................................................................421-431 Government Canyon State Natural Area: An emerging model for karst management George Veni .................................................................................................................433-440 Combining LiDAR, aerial photography, and Pictometry tools for karst features database management Scott C. Alexander, Mina Rahimi, Erik Larson, Cody Bomberger, Brittany Greenwaldt, and E. Calvin Alexander, Jr. ........................................................................................441-448 An evaluation of automated GIS tools for delineating karst sinkholes and closed depressions from 1-meter LiDAR-derived digital elevation data Daniel H. Doctor and John A. Young .........................................................................449-458 Delineation and classification of karst depressions using LiDAR: Fort Hood military installation, Texas [Poster] Melinda G. Shaw Faulkner, Kevin W. Stafford, and Aaron W. Bryant ........................459-467 Locating sinkholes in LiDAR coverage of a glacio-fluvial karst, Winona County, MN [Poster] Mina Rahimi and E. Calvin Alexander, Jr. ..................................................................469-480


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2IXORGANIZING COMMITTEE Conference Co-Chairs George Veni, Ph.D., P.G., National Cave and Karst Research Institute (NCKRI), Carlsbad, NM Jim LaMoreaux, Ph.D., P.E. LaMoreaux & Associates, Inc., Tuscaloosa, AL Program Co-Chairs Lynn B. Yuhr, P.G., Technos, Inc., Miami, FL Lewis Land, Ph.D., New Mexico Bureau of Geology & Mineral Resources and National Cave and Karst Research Institute, Carlsbad, New Mexico Proceedings Managing and Assistant Editors Lewis Land, Ph.D., New Mexico Bureau of Geology & Mineral Resources and National Cave and Karst Research Institute, Carlsbad, NM Daniel H. Doctor, Ph.D., U.S. Geological Survey, Eastern Geology & Paleoclimate Science Center, Reston, VA J. Brad Stephenson, P.G., L.R.S., CB&I, Knoxville, TN Field Trips Lewis Land, Ph.D., New Mexico Bureau of Geology & Mineral Resources and National Cave and Karst Research Institute, Carlsbad, NM Short Courses E. Calvin Alexander, Jr., Ph.D., Department of Earth Sciences, University of Minnesota, Minneapolis, MN Invited Speakers Yongli Gao, Ph.D., University of TexasSan Antonio, San Antonio, TX Beck Scholarship E. Calvin Alexander, Jr. Ph.D., Department of Earth Sciences, University of Minnesota, Minneapolis, MN Dianne Joop, National Cave and Karst Research Institute, Carlsbad, NM Ira D. Sasowsky, Ph.D., P.G., Geosciences, University of Akron, Akron, OH Beck Memorial Jim LaMoreaux, Ph.D., P.E. LaMoreaux & Associates, Inc., Tuscaloosa, AL Dianne Joop, National Cave and Karst Research Institute, Carlsbad, NM Brian Smith, Ph.D., Barton Springs/ Edwards Aquifer Conservation District, Austin, TX J. Brad Stephenson, P.G., L.R.S., CB&I, Knoxville, TN Conference Management Dianne Joop, National Cave and Karst Research Institute, Carlsbad, NM Debbie Herr, National Cave and Karst Research Institute, Carlsbad, NM Suzanna Langowski, National Cave and Karst Research Institute, Carlsbad, NM Circulars and Publicity Samuel V. Panno, CGWP, Illinois State Geological Survey, Prairie Research Institute, University of Illinois, Champaign, IL Harry L. Moore, P.G., Golder Associates, Atlanta, GA Program with Abstracts Brian Smith, Ph.D., Barton Springs/ Edwards Aquifer Conservation District, Austin, TX Brian Hunt, Barton Springs/Edwards Aquifer Conservation District, Austin, TX


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEXWebsite Dianne Joop, National Cave and Karst Research Institute, Carlsbad, NM Gheorghe Ponta, P.G., P.E. LaMoreaux and Associates, Tuscaloosa, AL Session Chairs Engineering and Geotechnical Aspects of Karst Joe Fischer, Ph.D., P.E., Geoscience Services, Clinton, NJ Evaporite Karst Ken Johnson, Oklahoma Geological Survey, Norman, OK Geophysical Investigations in Karst Terrain Mustafa Saribudak, Ph.D., P.G., Environmental Geophysics Associates, Austin, TX Formation Processes of Karst and Sinkholes Harry L. Moore, Golder Associates, Atlanta, GA Karst Hydrology Geary Schindel, P.G., Edwards Aquifer Authority, San Antonio, TX Mapping and Management of Karst Regions Samuel V. Panno, CGWP, Illinois State Geological Survey, Prairie Research Institute, University of Illinois, Champaign, IL Session Subcommittees Tony Cooley, P.E., P.G., EEC-DEP Division of Waste Management, Solid Waste Branch, Closure Section, Frankfort, KY Jeff Schaumburg, P.E., Dynamic Earth, Chester, NJ Dave Weary, National Cooperative Geologic Mapping Program, U.S. Geological Survey, Reston, VA Phil Carpenter, Ph.D., Dept. of Geology and Environmental Geosciences, Northern Illinois University, DeKalb, IL Thomas L. Dobecki, Ph.D., P.G., SDII Global Corporation, Tampa, FL Bruce Smith, U.S. Geological Survey, Denver Federal Center, Denver, CO AEG and SAGEEP Liaison Mustafa Saribudak, Ph.D., P.G., Environmental Geophysics Associates, Austin, TX ASCE/Geo-Institute Liaison Rob Schweinfurth, Geo-Institute of ASCE, Reston, VA EWRI Liaison Brian Parsons, P.E., Environmental and Water Resources Institute, Reston, VA Highway Geology Symposium Liaison Harry L. Moore, P.G., Golder Associates, Atlanta, GA Members at Large Ralph Ewers, Ph.D., EWC Ewers Water Consultants Inc., Richmond, KY Geary Schindel, P.G., Edwards Aquifer Authority, San Antonio, TX Wanfang Zhou, ERT, Inc., Knoxville, TN Gheorghe Ponta, P.G., P.E. LaMoreaux and Associates, Tuscaloosa, AL


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2XI FOREWORDWelcome to the Thirteenth Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst in sunny Carlsbad, New Mexico. This will be the farthest west the Sinkhole Conference, as it is informally known, has met since its inception in 1984. The setting will provide conference participants with a unique opportunity to view karst phenomena such as gypsum cenotes that are uncommon outside the southwestern United of the Permian Basin region. In 2011 the National Cave and Karst Research Institute (NCKRI) assumed responsibility for hosting the Sinkhole and karst hydrology is well-positioned to assume a leadership role in organizing and hosting the conference. Several of the staff of NCKRI have a long history of participation in past Sinkhole Conferences, and we look forward to supporting and hosting future meetings in other areas of the United States and abroad. The fourteenth conference will be held in Minneapolis, Minnesota in 2015, and discussion has begun on the possibility of an international setting for a future conference. We wish to dedicate this years proceedings volume to the memory of Barry Beck, who died in 2011. Barry initiated the Sinkhole Conference series in 1984 and was instrumental in maintaining the series of meetings over the years through several sponsors. Although his energy and enthusiasm will be greatly missed by future conference organizers, we are honored to carry Barrys legacy into the future. Edited by:


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEXIIKEYNOTE SPEAKERSPELEOLOGICAL, HYDROGEOLOGICAL AND ENGINEERING GEOLOGICAL CHALLENGES OF TUNNELING IN KARST AREASDr. Mladen GaraicUniversity of Zagreb, Croatia In the Classical Dinaric Karst of Croatia, over 11,500 caves have been explored so far, more than 1,000 of which were discovered during construction works. Caves discovered on the construction sites of highways lacked natural entrances on the surface. Over the past 20 years they have been systematically investigated and remediated to allow completion of the roads. Some special examples will be presented during the lecture, such as the large hall in the Vrata Tunnel of the Zagreb Rijeka Highway, and caves in Croatias longest tunnels. Due to the size, shape, position, and hydrogeological parameters of the cave within the karst system, it was necessary to design and construct a special bridge through the cave in the Vrata Tunnel. The caves vaulted ceiling had to be reinforced and stabilized. This presentation will include video and photos of the most interesting karst and cave locations in Dinaric Karst.BiographyMladen Garaic PhD. Geology, Hydrogeology, and Geological Engineering. Born in Zagreb, Croatia, in 1951, Dr. Garaic graduated in geology and karst hydrogeology in 1977, master of science 1981, and doctorate in geosciences and geological engineering in 1986. He is a scientist, and a professor of geology, karst hydrogeology, applied geology, engineering geology and speleology at the University of Zagreb, and has authored serves as a committee member for the Croatian Academy of Science and Arts, UNESCO World Heritage Team for the Dinaric Karst, International Association of Hydrogeologists, and International Association for Engineering Geology and the Environment. Dr. Garaic started skiing in 1955 and won the Junior Skiing competition of Croatia in 1963. He has been a member of the Croatian Mountaineering Association since 1955 and was awarded by the Association in 1969 and 1981. He started caving in 1963 and is the founder and president of several caving clubs in Croatia. He Federation from 1990 to 2010 and is a life member of the U.S. National Speleological Society. Since 1993, he has served as Croatias delegate to the International Union of Speleology and to the European Speleological Federation beginning in 2009. Dr. Garaic has conducted research in, and explored and visited nearly 5,000 caves in 64 countries. He has led many speleological expeditions in the longest and deepest caves in Croatia, Europe, and the world. He has also studies about 1,000 caves without natural entrances, discovered by tunnels and quarries, and evaluated their hydrogeology and engineering geology.


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2XIIIKEYNOTE SPEAKERNATURAL AND ANTHROPOGENIC SINKHOLES: FROM IDENTIFICATION, TO SURVEYING, STUDYING AND MODELING A SUBTLE HAZARDMario PariseNational Research Council of Italy, Institute of Research for the Hydrogeological Protection Sinkholes are the most common hazard in karst, being related to the presence of natural caves, and to their interaction with the ground surface. In the last decades, however, the study of sinkholes widened well beyond the boundaries of karst, including situations where cavities produced by man in different epochs and for different purposes interact in some way with the built-up environment, and represent a likely threat to the society. As a matter of fact, several urban areas in many countries worldwide have been recently affected by sinkhole occurrence which caused severe damage; sinkholes in Guatemala City, and other events in Italy, Germany and Turkey are only some of the many that characterize the last several years. In terms of civil protection issues, the topic has become of high interest in Italy, and much work has been devoted activities carried out, as they concern both natural and anthropogenic sinkholes, and to share the experiences so far developed. These latter cover all the phases of prone areas, to surveying the underground environment (by combining speleological techniques and modern technologies in order to get reliable and precise surveys), to recognizing the type of rock failures and characterizing the rock mass in terms of mechanical properties, to eventually modeling the case studies through numerical codes in order to forecast the likely evolution of underground failures, their upward propagation, and evaluating the possibility of sinkhole occurrence at the ground surface. A particular focus will be given to historical research, and its use in identifying ancient and/ sinkhole susceptibility and hazard. All of this will be illustrated through a number of case studies in southern Italy, dealing with natural karst caves and anthropogenic also cover some issues related to land-use problems in sinkhole-prone areas, and the utilization of the outcomes from sinkhole studies in civil protection programs at the local and national level, aimed at safeguarding and protecting private and public properties and the local populations.BiographyAfter graduating with honors in Geology in 1988 at the Faculty of Sciences of the University of Naples, Mr. Parise received grants from the National Research Council of Italy and spent several periods working in cooperation with the U.S. Geological Survey at Golden, Colorado, and the University of South Florida at Tampa, Florida. Since 1994 he has worked as a Research Geologist at the National Research Council, Institute of Research for Hydrogeological Protection (CNR-IRPI) in Bari, Italy. He has organized and convened several international workshops and conferences on the topics of karst, karst hazards, and slope movements (European Geosciences Union Assemblies, Geological Society of America Meetings, Italian Forums of Earth Sciences), and is the scientist responsible for several projects between CNR-IRPI and different public administrations and private companies. Since 1990, Mr. Parise has developed research mainly into the geological and geomorphological analysis of slope movements. Much of his research deals with the deformations, mass wasting processes, etc.) by means of stereoscopic interpretations of aerial photographs and analyses, aimed at understanding the likely evolution of slopes, even in relationship with anthropogenic events (rainstorm, earthquakes, etc.). For several sites


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEXIVof natural and anthropogenic hazards that occur in karst territories, with particular regard to sinkholes related to both natural caves and man-made cavities. He is the author of over 100 papers published in international journals and proceedings of international conferences. He has given several presentations in international symposia and workshops. Mr. Parise has guest edited 10 special issues for ISI international journals, published two books with the Geological Society of London, and reviews papers for several international journals. in southern Italy, he has created a framework of the movements. He has also contributed to the analysis of rapid landslides (debris avalanches, rock avalanches) in different geological settings in Italy and abroad, and to He began caving in 1998 and since 2002 he also works


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2XVdissolution eastward and downward. Cave levels formed by sulfuric acid record the position of the water table at a given time, and the elevation difference between levels may correlate with episodes of Rio Grande Rift tectonism since 12 Ma.BiographyHarvey DuChene is a graduate of the University of New Mexico, earning B.A. (1968) and M.S. (1973) degrees in geology. He has 39 years of experience as a petroleum geologist, working for Amoco Production Company, Davis Oil, Axem Resources and others. He currently is a limited partner in Vecta Oil and Gas, LP, an oil and gas exploration and production company headquartered in Dallas, Texas. His primary area of expertise is petroleum exploration in basins of the Rocky Mountain province and west Texas, with additional experience in the midcontinent, Appalachian basin and offshore West Africa. Harvey has also more than 30 years studying the speleogenesis of hypogenic caves, particularly those formed by sulfuric acid. He is interested in the connection between the evolution of hypogenic cave systems and the tectonic and geologic history of regions. Harvey is a member of the Geological Society of America, American Association of Petroleum Geologists, American Geological Institute, Rocky Mountain Association of Geologists, New Mexico Geological Society, West Texas Geological Association and Karst Waters Institute, and he is a Fellow of the National Speleological Society.KEYNOTE SPEAKERTECTONIC INFLUENCES ON PETROLEUM MIGRATION AND SPELEOGENESIS IN THE GUADALUPE MOUNTAINS, NEW MEXICO AND TEXASHarvey R. DuCheneVecta Oil and Gas, LP Sulfuric acid speleogenesis in the Guadalupe Mountains is a consequence of the rise of the Alvarado ridge and subsequent opening of the Rio Grande Rift during Cenozoic time. Uplands of the late Laramide (~38 35 Ma) Alvarado Ridge provided an immense recharge area that supplied water to aquifers draining eastward to the Permian basin. Evidence for east-directed hydrodynamic subsequent recharging of hydrocarbons in large structural and stratigraphic traps in Artesia Group (Permian, Guadalupian) reservoirs in southeast New Mexico and adjacent west Texas. Prior to, or during the early stages of the development of the Rio Grande Rift, hydrostatic through Artesia Group strata toward the Permian basin. Some of this water moved upward along fractures to artesian springs in the area of the Guadalupe Mountains. This resulted in solutional enlargement of fractures and development of early stage caves. Extensional faulting progressively reducing the size of the upland recharge introduced microbes into Artesia Group (Permian, Guadalupian) reservoirs causing biodegradation of petroleum and generating copious H2S.The water table within the Guadalupe Mountains began to fall 14-12 Ma in response to erosion and tectonism. During this time, oxygen-rich meteoric water mixed with H2S water to form sulfuric acid, which enlarged passages and galleries at the water table. Tectonic spasms related to the opening of the Rio Grande Rift caused abrupt drops in the water table, shifting the locus of sulfuric acid


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEXVIBiographyWilliam (Bill) Kochanov (pronounced KO-CHANOFF) is a Senior Geologist with the Pennsylvania Department of Conservation and Natural Resources, Bureau of Topographic and Geologic Survey, Geologic Mapping Division. Since 1985, he has been actively mapping geologically hazardous areas within the limestone regions of Pennsylvania and maintains the Bureaus sinkhole database. He has also conducted bedrock mapping projects spanning much of the Paleozoic from Pennsylvanias northern anthracite coal Valley of southeastern Pennsylvania. Bill is most noted designed to characterize karst surface features, their distribution, and their relation to physiographic setting. He is strongly involved with the Surveys outreach programs; translating the geology of Pennsylvania for, as Joe Fischer puts it, the greater unwashed. Bill lives in the suburbs of Harrisburg with his wife Jane and children, Natalie and Alex, close to the forests of Stony Creek where he spends many hours tracking down the elusive edibles of the mushroom world. KEYNOTE SPEAKERWHEN THE CARBONATE PLUMBING GOES BAD: SINKHOLES, THE HYDRA, AND THE GENERAL PUBLICWilliam KochanovPennsylvania Department of Conservation and Natural Resources In 1985, a program was initiated by the Pennsylvania Geological Survey to inventory (catalog) existing sinkholes and to map areas of potential sinkhole development. The program was developed to provide general background information for the initial stages of site investigations, aid in sinkhole remediation efforts, and serve as a tool for developing regional land-use planning strategies. Although the methods of data collection and distribution have evolved over the past 25 years, it has been interesting to note that the practicing professional continually has had to refine the means of sorting and sifting data much like that of a forensic specialist; each investigator having their own special challenges as the clues for remediation often lie hidden beneath the veneer of urbanization, are squirreled away in files of the local Historical Society or are muted for fear of liability. Bill will take you on a savage journey through the karstlands of Pennsylvania to marvel at some of its many wonders, examine yawning portals to the underworld, grapple with the paradox of the cultural hydra, and the ultimate in trepidation, entering the lair of the general public.


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 21AbstractThe assessment of karst conditions and putative karst geohazards prior to residential and commercial aspect. Borrowing from the medical lexicon, most karst features at proposed building sites are dealt with using an approach wherein the symptoms and conditions are treated (e.g. sinkhole remediation), often only after site development activities have commenced. If karst hazards are suspected, roadways, foundations various geophysical methods; however the results of these investigations require specialized knowledge to be interpreted and understood. Thus stakeholders results indecipherable, often leading to unnecessary and expensive supplemental studies, the need for which is to attach cursory karst assessments to due diligence study reports, particularly Phase I Environmental Site Assessments. These combined assessments are often performed by individuals who are inexperienced in geology. Not unexpectedly, this can lead to numerous mistakes, errors, and oversights. More troubling, these studies often report a lack of karst risks at the site under study, a result that the stakeholders may initially embrace, but which later can result in substantial and the environment. To address these concerns, we propose a proactive, preventative standard practice for karst assessments. Ideally, this proactive approach will help to delineate potential karst hazards so that they can be avoided, managed, or corrected by remediation. Requirements for investigators, a proposed scope of services, fieldwork and data review checklist, and a presented. It is our hope that if carried out and reported accurately, technical stakeholder to make informed decisions regarding the relative risk of karst geohazards, the need for further studies, and potential corrective actions that site development may entail. Introduction and BackgroundThe study of karst features, in particular karst springs and groundwater stretches back into earliest written caves and their hydrography was written in 221 B.C.E. in China, and the solution process of carbonate rocks was described accurately by the Roman Philosopher Seneca (4 B.C.E. 65 C.E.). Commentary by naturalists and philosophers on karst features and hydrology continued in both Europe and Asia through the subsequent centuries and entered into the era of systematic geomorphological investigation in the 19th century (LaMoreaux and LaMoreaux, 1998). Not surprisingly, in regions where much of the land surface was underlain by soluble bedrock and prone to the development of karst terrain, karst studies were advanced by the interests of regional politics (Ztl, 1974). One such Empire had acquired extensive tracts of karst lands. The need to ensure that water supplies were adequately developed and infrastructure was protected drove these studies forward, and arguably the Austrian studies Das Karstphnomen. was little emphasis on assessing the environmental and economic impacts of human development in karst terrains until the latter half of the 20th Century (LaMoreaux, et al, ; Rauch and Werner, 1974). An increased sense of environmental awareness, coupled with increasing residential and commercial development in karst terrains during the 1970s and 1980s led to increased interest in the characterization and mitigation TOWARDS A KARST ASSESSMENT STANDARD PRACTICERobert K. Denton Jr., CPG, CPSS GeoConcepts Engineering Inc, 19955 Highland Vista Dr., Ste. 170, Ashburn, Virginia 20147 USA,


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE2of karst hazards and environmental impacts. The Center for Cave and Karst Studies at Western Kentucky level, the importance of karst studies was heralded by the creation of the National Cave and Karst Research Institute (NCKRI). speleogenesis, karst hydrology and karst biology, yet ironically there was little attempt to advance the development of a karst site assessment as a standard practice. The putative process languished at the same stage of evolution as environmental site assessments Materials (ASTM) E1527 practice. Karst assessments ranged in nature from cursory sinkhole inventory and rudimentary geophysical subsurface investigation (often without any interpretation), to geologically detailed and of which would assist municipal planners, regulators Frequently the lack of any obvious surface karst features investigator(s) that there were no karst issues at a site. In contrast, investigators might recommend lengthy and Errors and misstatements of these sorts made karst studies misleading and essentially useless for responsible development and land planning.Towards a Standard of PracticeIn response to the polyglot of assessment schemes a movement towards a karst assessment standard of st century. Notable examples were the Virginia Sinkhole et al, 2001), Kentucky Model Karst Ordinance (Currens, 2009), the Clarke County Virginia Sinkhole Ordinance (Code of Clarke County, 1997) and Karst Plan Requirements (Teetor, 2004), and Chapter 6 of the Loudoun County Virginia Facility Standards Manual Limestone Overlay District (2010). Nevertheless, a single karst assessment standard of practice similar to the ASTM standard practice for Environmental Site Assessments (ASTM, 2005) was lacking. Thus, what we present in this article is a proposed model standard of practice that embodies a set of basic elements that should be included in any karst site characterization. It must be emphasized that this approach is not to be considered the exclusive requisite elements in a karst assessment, but the essential starting points for a basic (preliminary) evaluation. Karst assessments will vary according to the needs of the user(s), the requirements embodied in local ordinances and the scope and nature karst geohazard issues on their own, driven by various incidents that brought caves and karst to the forefront of public interest. (Figure 1). As a result of this increased public interest and concern, a series of karst model ordinances were proposed at both state and county levels across the United States (Karst Portal, 2012). Typically, these model ordinances dealt with the what and where of karst, but not the who and how. Karst studies were increasingly being required by planning boards and zoning commissions as part of the studies for approval and permitting of residential and commercial development in potential karstlands, but the manner in which the studies were During the last decades of the 20th century there was a veritable renaissance in academic studies regarding Figure 1. A Virginia newspaper story detailing the 1992 collapse of a house in the Shenandoah Valley into a sinkhole.


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 23in accord with this scheme, and reported accurately, the stakeholder to make informed decisions regarding the relative risk, the need for further studies, and potential corrective actions that site development may entail.Requirements for Karst Investigators for the karst professional investigator is as follows: A Professional Engineer (PE) with a geotechnical (or) and engineering geology; supporting documentation (e.g. resume, curriculum vitae, etc.) should be part of the assessment report, including a statement specifying that the investigator It is important to understand that a P.E. license does not necessarily qualify an individual to be a karst investigator, or make recommendations regarding engineering solutions for karst geohazards. By the same token, many licensed geologists have never had any formal training or experience with engineering geology or geotechnical with karst issues is the most critical factor in designating an individual as a karst professional investigator. karst investigator can be found in the Clarke County Va. Karst Plan Requirements: professional engineer (PE) engaged in the engineering geology. somewhat of an exaggeration in the above statement of to be called such, and a licensed geologist is not an engineer although in the Clarke County regulation they the Berryville, Clarke County Va. statute falls short is could theoretically sign and seal an investigation, within which recommendations have been made that could be poorly informed at best, or lead to disastrous consequences at worst. Finally, it cannot be emphasized more that karst is not a uniform geomorphological process, and varies considerably from region to region. A geologist or engineer Tertiary carbonates of Florida may not be familiar with issues affecting the stronger and more competent Paleozoic carbonates of the Appalachian region, or the Mesozoic carbonates of the Texas plateaus. Thus, it is important that karst where the assessment is being conducted.Definitions and TerminologyThe lexicon of karst literature is among the most varied and complex of the earth sciences, due to much of the countries. Thus, myriad terms are often used for the same structure (e.g. swallet, insurgence, sinking stream, ponor, swallow hole, perte de riviere, all of which refer to the same feature). As much of karst description is describe a feature must be consistent and understandable Thus, each assessment should include at least a brief glossary wherein the specialized terms being used are this glossary should be the publication A Lexicon of Cave and Karst Terminology with Special Reference to Environmental Protection Agency (Field, 2002).Recommended Scope of Services undertake an inspection of the site area and prepare an investigation report which shall include (but not be limited to) the following elements: a. Site description and terrain analysis;


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE4b. Description of published soils and underlying bedrock and comparison to onsite observations; c. features; d. and drainages including, but not limited to: sinkholes (both active and incipient), caves, insurgences (swallow sinkholes), resurgences (springs), losing streams, and potential for covered karst (i.e. sinkholes lying beneath soils cover); e. Inferred locations of shallow bedrock (based on evidence from rock outcrops) with any recommendations made by the investigator for additional studies which may include electrical resistivity studies, seismic studies, subsurface borings, or any other appropriate method to determine if the proposed development may have negative impact on human health, safety, property or the environment. No evidence of karst features If the investigator bedrock, or there is no evidence of karst features (including covered karst or pinnacled bedrock), they shall so indicate. Evidence of karst features In cases where the would be impacted by development, detailed subsurface investigations shall be required within foot radius shall be measured from their discernable edge. At the completion of the investigation the investigator should prepare a Karst Management Plan and the developer directed to follow the Presence of karst features on the site which will not be impacted If no karst features are to be affected by the planned development, there will be no need to submit a stand alone karst plan. A statement should be included in the Karst Site Assessment certifying that no features will be impacted. Description of the Scope ElementsSite Description and Terrain Analysis The investigator should describe the site, based on examination of the closest topographic mapping available (Light Detection and Ranging) be utilized if available photograph pairs and aerial photo fracture trace analysis may be utilized. Any karst features visible on the topographic map and remote sensing resources (i.e. caves entrances, sinkholes, closed depressions, etc.) should phase of the assessment. The site description should also include a careful current use and condition (i.e. vacant land, agricultural land, developed land etc.). Any proposed changes to the site, especially development plans, should be noted and explained in the assessment report. Description of Soils and Bedrock Geology The investigator should access the National Resource Conservation Service soil maps properties, with particular emphasis on the parent materials Figure 2. Two-foot contour map of a project site, showing a series of closed depressions (sinkhole) in lineaments.


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 25(i.e. whether the soils are residual or transported), their hydrologic characteristics, and textural analysis. Certain soils are noted in NRCS survey data as being prone to sinkhole formation. These soils should be noted and should be carefully examined even if no closed depressions or sinkholes are noted in the terrain analysis. Understanding the soils is critical to predicting whether sinkholes will form after a site has been stripped and grubbed (i.e. cleared), as highly cohesive soils can often create a covered or mantled karst condition hidden beneath the seemingly homogeneous soils cover. Upon removal of the vegetation, the soil will begin to ravel, and previously undetected sinkholes will begin to form. Bedrock geology should be determined by referencing the highest resolution geological mapping available, ideally at a Map Database ( or the websites of the local state geological survey. Dip and strike of the bedrock, and or synclines, etc.) should be noted. Field inspection should attempt to verify the mapped soils and bedrock by comparison to the available investigator should note whether or not the soils and bedrock conform to the published description(s). If they compare favorably, then no further explanation is required. If they do not, then a detailed description of the differences should be provided. Description of Surface Drainages and Water Features The investigator should determine the drainage patterns at the site by examination of the topography. The investigator should also check to see if any publicly available hydrological assessments have been performed for the region of interest by state or federal entities. The analysis of drainage patterns should determine if the site has outlets (i.e. if drainage is directed offsite) or if it is internally drained as these factors can profoundly affect site planning, especially in regards to stormwater management. Drainages to sinkholes should be clearly delineated (Figure 4). Figure 3. Topographic Position Index (TPI) showing local topographic concavity and convexity derived from a 1m LIDAR elevation model and overlain on aerial imagery. Figure 4. Example drainage map showing sinkhole drainage areas. Note that the drainage area for sinkhole K1 is primarily outside of the site boundary (red line).


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE61A sinkhole which acts as a spring during groundwater highstand conditions, and an insurgence during low stand conditions.Closed Depressions/Sinkholes The locations of any closed depression (CD) or area of closed descending contours should be located and examined. The investigator should describe the feature, noting the following parameters: 1. What is the general shape of the CD? 2. Is the CD actively forming (i.e. are there soil tension cracks around the perimeter of the structure?) or has most of the soil already raveled into the subsurface? (See Figure 5A,5B) (Figure 5C,5D) 4. Are there mature trees in the structure? What are the estimated ages of the trees? (Figure 5C) 5. Does the CD have a throat or opening(s) leading into the subsurface? (Figure 5D, 5E) 6. estavelle1, such as watermarks, saturated soils, or 7. Is the CD in a topographic position such that it receives drainage from the surrounding area? 8. If the answer to question 7 is yes, does the CD have an obvious drainage channel leading into it, The CD should then be measured and delineated. This can depth and circumference should be determined as closely as possible and noted, as well as any nesting of smaller depressions within the larger ones. The investigator should be aware of any area where there as this may be an indicator of a subsurface conduit that this regard, distinct changes in vegetation can be a clue if topographic is slight or absent. These areas should be carefully noted and investigated if they are to be impacted by proposed site development, as they can be the site of sudden and catastrophic subsidence if not managed properly. closed depressions and sinkholes, as cave entrances The locations of perennial springs, streams and water bodies (lakes, pond, etc.) should be noted. The locations of losing streams (i.e. streams that lose water to the subsurface through their bed), gaining streams, and sinking streams should be carefully noted. Location and Delineation of Karst Features the investigator should access available karst and cave survey databases to determine if any features have been areas. The National Speleological Society (NSS) has survey number of caves, and although the databases of these surveys are technically proprietary, the surveys will share these data with legitimate investigators to assist in conservation and protection efforts. In addition, many karst features have been publications. Various state surveys have also published compendiums of cave locations and descriptions in book form, but these publications are seldom complete and need to be supplemented by data that has been collected from the regional NSS surveys. The NSS also has made available through their publication bookstore numerous county level cave surveys which should be accessed if pertinent to the area of interest. Finally, it is extremely helpful to interview the land owner features known to them that may exist on or near the survey area. Residents may also know of sinkholes that have physically closed, or other features not readily observable during the site inspection. They may also have useful information regarding locations of wet weather springs, seeps, or ephemeral karst lakes and ponds (turloughs) resurgences that are not present during dry weather periods. Alternately, residents may know of locations advantage to examine and verify these observations. Once the potential locations of karst features have been accessed and noted, the investigator can begin the


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 27Figure 5A. Actively forming cover collapse sinkhole in granular sediments. Figure 5B. Actively forming cover collapse sinkhole in cohesive, fine-grained sediments.Figure 5C. Mature, stable sinkholes in cohesive soils.Figure 5D. Mature, rock-walled sinkhole with open throat (i.e. cave entrance). Figure 5E. Soil-bottomed sinkhole with open throat. A 40 deep vertical cave lies below the opening. This type of structure is sometimes called a natural trap. the subsurface large enough to allow the passage of a human being. As caves are frequently the home for rare, threatened and endangered species (RTES), often contain important cultural and historic resources, and are environmentally sensitive, it is imperative that they be managed, conserved and protected. The investigator should attempt to locate and examine any mapped or reported caves on the site. Locations of from the regional speleological survey of the NSS. A plan view of the cave showing its route beneath the site is


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE8called grottoes, generally are glad to help with an assessment by exploring, photographing and mapping a new or unexplored cave. either entering the subsurface through a solution feature, or exiting the subsurface through a resurgence (spring) should be located and examined. The locations of perennial topographic maps. In addition, the landowner or neighbors may have knowledge of springs that have not been mapped using accepted hydrological methods and reported. Insurgences, sinking streams or valley drains (open throat sinkholes that receive surface drainage through a should be noted that if a site is internally drained, and a established by the US EPA, and should be registered own regulatory requirements for stormwater disposal into sinkholes as well, and these should be checked and referenced if applicable. The determination of subsurface drainage patterns in karst is a technically demanding and specialized activity, and is typically beyond the scope of a preliminary karst important, as caves that are located close to the surface can present a risk to planned development. In contrast, shallow caves can be more readily impacted by releases of contaminants, redirection of surface drainage, and by the investgator unless they are an experienced spelunker and familiar with the methods and techniques of cave exploration. Caving is an inherently dangerous the proper equipment. The local chapters of the NSS, Figure 6B. The same structure as shown in Figure 6A during groundwater high stand conditions. When this pho tograph was taken the estavelle was an active, ephemeral spring with an outflow measured at 60 gpm.Figure 7. An area of snowmelt marking a closed depres sion where water was actively infiltrating into the subsurface. This depression had a relief of less than 2-feet below the surrounding terrain and was not indicated on the site civil engineers 2-foot contour map. Subsequent Electrical Resis tivity Survey (ERS) showed the presence of a soil-filled throat in the bedrock below the structure that was actively channel ing surface drainage into the subsurface. Figure 6A. An estavelle in groundwater low-stand conditions. Note the tell-tale water mark along the rock wall of the structure.


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 29geology, soils, and prior experience with sites in similar geological settings. site is located in an area that the investigator suspects where there may be covered karst conditions present, this should be clearly indicated in the assessment construction, and increase the costs of site development strongly recommended that the investigator include a section as follows: As indicated in this report, the bedrock and overly ing soil below the site are susceptible to sinkhole development, and karst features are likely hidden beneath the existing soil stratum. Risk associated with sinkhole formation can be minimized during development with proper foundation design and construction, and the control of site hydrology. The must evaluate the risks and attendant costs of devel opment, and must be willing to accept them. Location of Shallow Bedrock The karst terrain is notorious for the presence of shallow bedrock, often with large areas of exposed ledges and shelves. This is particularly problematic due to the fact that much of the carbonate rocks can be resistant to scaling or scarping, and must be either rammed or blasted during the grading process. Areas of shallow or surface exposed bedrock need to be clearly delineated and described in the assessment report. In areas where the bedrock is steeply inclined, differential solution activity can produce a pinnacled bedrock surface, often with exposed bedrock ledges and deep intervening cutters in between containing residual soil (Figure 9). The ledge and cutter terrain is often not considered a sensitive environmental feature by site developers or impact to the subsurface environment if not managed properly. Surface water can migrate rapidly along interstice. During periods of extended drought, the soil using dye tracing techniques, and the literature should be searched by the investigator to see if any previous studies have been conducted in or near the area where the assessment is being performed. If ground water monitoring is to be included in the scope of work, then the investigator should employ the techniques embodied in the US EPA guidelines for groundwater monitoring in karst (Quinlan, 1989). Finally, it should be noted that although they are not natural features, abandoned quarries, drilled wells subsurface, and often have direct connection to the phreatic aquifer. As such, these features should also be included in any comprehensive karst assessment. Covered or Mantled Karst In many karst settings there is often a relatively thick stratum of cohesive soils lying no obvious karst features to be seen in this type of natural setting, however upon removal of the vegetation and topsoil (i.e. stripping and grubbing) during the preliminary stages of grading a site, cover collapse sinkholes will rapidly form where there seemingly were none before (Figure 8). Figure 8. A pair of cover collapse sinkholes that opened at a site under development after the vegetation and topsoil was stripped. Open throat, air-filled conduits in the bedrock were located at the bottoms of both of these features.


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE10or retention ponds or impoundments, must be carefully examined for the presence of pinnacled bedrock. Exposed pinnacles (Figure 11) can lead both to subsurface from the base of the pond, or in the worst case scenario, catastrophic development of sinkholes into which the entire contents of a pond (i.e. water, collected sediment and entrained contaminants) can be disgorged. If pinnacled bedrock is present in these areas the users of the assessment should be made aware of the condition and the risks associated with it. Exposed pinnacles (Figure 11) can lead both to subsurface from the base of the pond, or in the worst case scenario, catastrophic development of sinkholes into which the entire contents of a pond (i.e. water, collected sediment and entrained contaminants) can be disgorged. If pinnacled bedrock is present in these areas the users of the assessment should be made aware of the condition and the risks associated with it.Follow-Up StudiesIf the planned site development will impact karst be necessary to thoroughly characterize the impact and help the developer and regional planners understand the risks involved. These studies may include detailed subsurface investigations such as geophysical exploration (e.g. electrical resistivity survey, seismic survey, microgravimetric survey, etc.), borings, track drill exploration, or any combination of the methods. It should be noted that geophysical studies, in particular will form, allowing surface water to plunge into the subsurface, often with direct connection to the phreatic can also begin the process of soil raveling, sometimes resulting in the sudden formation of sinkholes. In swell prone clays, there is often a condition informally referred to as sinkhole weather which is characterized by extended dry weather or drought punctuated by periods of heavy rain. Sinkholes will often form when these conditions are present. Finally, areas of a site designated for storm water Figure 9. Excavated site cross-section showing pinnacled bedrock with intervening soil-filled cutters. Figure 10. The epikarst exposed in an abandoned lime stone quarry wall, showing steeply-angled open solutionmodified fractures extending down to the quarry lake. The lake is representative of the local phreatic base-level, and demonstrates how contaminants and surface water can readily migrate to the underlying water table.Figure 11. Exposed bedrock pinnacles located in the base of a stormwater detention structure in West Virginia.


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 211 ClosureIt is our hope that this article may serve as a template to assist investigators in conducting comprehensive regulators, engineers and legislators in determining the minimum elements that should be expected in a site evaluation. It should be emphasized that the scheme presented herein is not intended to serve as a substitute for detailed subsurface investigations, or to supersede any existing References for Environmental Site Assessments: Phase I Environmental Site Assessment Process, ASTM International, West Conshohocken, PA, 2005, Available from: Code of Clarke County, Virginia. 1997. 1st Revision Adopted October 21, 1997. Currens, J. C., 2009. Model Ordinance for Development University of Kentucky, Lexington, KY. Field, M. 2002. A Lexicon of Cave and Karst Terminology with Special Reference to Environmental Karst Protection Agency, Washington, DC. LaMoreaux, P. E., LaMoreaux, R., 1998. A history of karst studies from: From Stone Age to the present. Focus, Summer98, 45(2), 6 pp. Loudoun County Facility Standards Manual, Model ordinance [Internet]. 2012. [Place of publication unknown]:; Available from: ordinances.html Colbert, and Franklin Counties, Alabama. U.S., Orndorff, W., Fagan, J., 2001. Towards a Comprehensive Use Planning. 2001 National Cave and Karst Management Symposium (abstract only). electrical resistivity survey (ERS), require experienced addition, the use of ERS or other geophysical methods without attendant rock probes (coring, track drill, etc.) can often be misinterpreted; however coring or air track investigations carried out without any supporting geophysical evidence of subsurface structures can be wasteful and expensive with little to show for the effort. The two methods should always be used in concert with one another.The Karst Management PlanA karst management plan should be prepared for any sites where there is evidence of karst features (i.e. sites upon which karst features are fully or partially located, The Karst Management Plan shall include (but not be limited to) the following elements: a. A karst feature inventory showing the areal extent of each structure, and a (minimum) 100 foot radius buffer area around the feature; b. A topographic map prepared at a maximum sufficient to determine low points or discernible edges; c. ensure structural stability of principal structures proposed within 100 feet of a sinkhole or other tests that will be completed to determine subsurface conditions. d. Mitigation recommendations for each karst feature requiring this action. All sinkholes mitigated or separated from construction. Mitigation should be carried out under the careful observation of the karst professional predicated in the karst assessment study, and measures in the event actual site conditions differ from the estimated conditions presented in the study. e. The management plan should be reviewed or planning staff prior to approval of site development or issuance of plats.


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE12 Karst Terranes: Recommended Protocols and Environmental Protection Agency, Environmental Monitoring Systems Laboratory, Las Vegas, Nevada. Teetor, A., Clarke County Karst Plan Requirements. Ztl, J. 1974. Kartshydrologie. Wien and Noew York:


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 213GEOTECHNICAL CASE HISTORY FOR SINKHOLE INVESTIGATION AND STABILIZATION METHODS ALONG A HIGH PRESSURE PETROLEUM PIPELINEJohn T. Pusey, Jr., and John M. Caccese, P.E.Earth Engineering Inc., 115 W. Germantown Pk., Ste. 200, East Norriton, Pennsylvania 19401 USA,, johnc@earthengineering.comAbstractInstallation of underground pipelines through unpopulated land situated over pinnacled carbonate bedrock can lead to the development of sinkholes. The formation of sinkholes beneath buried pipelines has the potential of damaging the pipeline and more importantly causing hazardous environmental incidents. This paper presents a case history at a site where a 400 foot (112 meters) long section of high pressure local creek in Plymouth Meeting, Pennsylvania. Various geophysical investigation techniques consisting (MASW), and two dimensional electrical resistivity borings to effectively evaluate the subsurface conditions at the site. Three options were considered as a solution property. Following the investigation process, relocating improvements via subsurface grouting represented the IntroductionIn January 2009 a sinkhole developed below an active petroleum pipeline that crossed under a local creek in Plymouth Meeting, Pennsylvania. Upon initial discovery, it was reported that the sinkhole measured approximately depth causing the pipeline to be unsupported across a portion of the open void. Representatives of the pipeline the owner recognized the severity of the problem and the Initially, a feasibility study was conducted to determine the site. The options considered include: 1) subsurface supporting the pipeline on a deep foundation system or reconnaissance and a stereographic aerial photograph review. Due to the site being primarily wooded, inconclusive results were found from the aerial photograph review. During the site reconnaissance, the streambed was dry on each side of the pipeline crossing. The stream bed remained dry for approximately 500 to 600 yards (457 to 549 meters) upstream of the sinkhole at the pipeline crossing. Further inspection revealed a large sinkhole had created a disappearing stream condition upstream of the pipeline crossing. Photograph 1 shows the large sinkhole upstream of the pipeline crossing. slopes gently to moderately downwards toward the creek and sinkhole locations. Photograph 2 shows the area of slopes gently to moderately downwards toward the creek and sinkhole locations. Photograph 2 shows the area of Project Description and Background into the upstream sinkhole leaving the downstream side dry. During periods of steady rainfall, storm water


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE14 down past the pipeline crossing. Numerous additional sinkholes are present on the western bank of the stream between the disappearing stream location and the sinkhole at the pipeline crossing. At the conclusion of immediate region is highly active and warranted further means of investigation. Two separate geophysical investigation methods were initially performed within the referenced section of microgravity, provides a broad interpretation of the channel analysis of surface waves (MASW), provides The combination of the geophysical methods provides a relatively accurate depiction of the subsurface conditions within the area of study. The microgravity investigation provides spatial coverage of the investigation area. Broad areas of higher gravity indicate relatively shallow rock (potential pinnacles) and broad areas of lower gravity indicate relatively deeper rock (voids). In microgravity surveying, fractures and faults are typically observed as linear low gravity anomalies because the fractured rock tends to be less email communication). The study conducted at the site consisted of recording sharp contrast to dense soil or bedrock. In addition, potential faults and fracture traces were also generated from the microgravity investigation. The results of the microgravity readings at each grid station are plotted in color and a microgravity contour map is generated to provide a clear interpretation of the subsurface conditions to the viewer. Figure 1 shows the results of the microgravity investigation. The results of the microgravity investigation clearly depict that subsurface conditions in the vicinity of the 2009 sinkhole location are highly variable with dense, shallow rock on the eastern and southern side of the creek and less dense, deeper overburden soils on the north and west side of the creek. Interpretation of the survey also revealed the presence of a potential fault that fault extends between the pipeline and the northern edge is on the north side of the fault. Two parallel fractures also exist in line with the creek and perpendicular to have an increased risk of development in proximity to the intersection of fractures and faults in the underlying bedrock. Accordingly, at this site, sinkhole locations coincide with the location of intersecting fractures and faults. Further exacerbating sinkhole activity is that the topography generally slopes downwards in all directions Photograph 1. Sinkhole within creek bed. Photograph 2. Pipeline right-of way crossing local creek.


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 215 was generated from the interpretation of shear wave velocities generated by striking a plate attached to the of the array record shear wave velocities as function of distance from strike point. From this data, material properties and depth to bedrock were estimated (Lee, 2010). The results are presented in Figure 2. After the borings were performed in proximity to open sinkholes, geophysical surveys, and where dense shallow bedrock was interpreted to exist. The intent of the test borings was to verify the conditions found in the geophysical investigations. Standard Penetration Tests (SPT) were performed at regular intervals throughout the borings until auger refusal was achieved. Following refusal toward sinkhole areas. The sinkholes are also at an apparent transition location between the shallow dense rock on the south and east side of the pipeline and deeper and less dense rock on the north and west side. Competent off the shallow, pinnacled rock surface and carries away loose soil and rock material to accelerate the sinkhole activity (Lee, 2010). Following the microgravity investigation, the secondary geophysical method consisting of MASW was of subsurface conditions directly below the pipeline. The MASW could not be completed within the stream Figure 1. Microgravity results.


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE16 A permeation grouting method was chosen based bedrock as well as minimizing the potential for heaving the active pipeline. The permeation grout consists of a clayey nature of the overburden soils, grouting was only sized void was accepting the grout. A total of 40 cubic subsurface without recorded backpressure. The results of the geotechnical investigation revealed that active sinkhole conditions were present in the work, a budgetary value of $600,000 was estimated for the large amount of grout required during the stopgap grouting operations, and the potential for extensive active sinkhole conditions near the pipeline, concerns were raised that the grouting costs could easily exceed the budget estimate. Therefore, a subsurface grouting program within the existing right of way was considered subsurface conditions and depth to competent bedrock to structurally support the pipeline was not considered viable due to the high costs associated with this option. on the bedrock surface, rock coring was performed to evaluate the condition of the underlying bedrock. As expected from the geophysical testing, the results of the test borings revealed highly variable conditions. The depth to bedrock ranges from seven feet beneath the existing ground surface to in excess of 70 feet (21 meters). The large variation in depth to bedrock exists meters) apart. Interpretation of a boring drilled near the 2009 sinkhole location and near the pipeline revealed This void was encountered during the rock coring operation. In areas where subsurface anomalies were found in the geophysical investigation, the test borings the relationship between the results of the microgravity investigation to the conditions found in the borings. The situated above a plan view of the microgravity results using microgravity to determine subsurface conditions. As a result of the conditions found in proximity to the by the test borings, an emergency stopgap grouting operation was performed utilizing a permeation grout. This stopgap grouting program was developed in an attempt to minimize the potential of failure below the Figure 2. MASW results on western side of creek.


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 217After evaluation of the microgravity results, a proposed shallow rock and minimal anomalies. Figure 4 displays the results of the microgravity results within the available proposed pipeline relocation route. After the preferred relocation alignment was chosen, (2D ERI) surveys were conducted to provide a linear alignment. The 2D ERI was used in areas of steep slopes or undulating ground surface. Following the geophysical surveys, test borings were drilled at anomaly locations 400 foot (122 meters) long section of pipeline to a less sinkhole prone location were evaluated. A triangular shaped, undeveloped tract of land is situated directly south of the 400 foot (122 meters) long section of existing pipeline. Once permission was granted to investigate the land to the south, a second phase of work at the site commenced. Since the purpose of the second phase was to evaluate an optimum route to relocate the pipeline, a proactive approach was taken. The new process consisted of performing an initial microgravity survey in a grid pattern to spatially identify subsurface conditions. Figure 3. Test boring profiles results and a plan view of the corresponding microgravity results along the creek perpendicular to pipeline.


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE18 Figure 4. Microgravity results included with MASW & 2D ERI locations over the proposed realignment route.Figure 5. Test boring locations conducted in realignment route.


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 219recommended in between the targeted locations, every the proposed pipeline. Figure 6 displays the proposed grouting location plan. The grouting program is recommended to be performed in phases. As shown in Figure 6, the black circles display the phase 1 grouting locations and the red triangles display the phase 2 grouting locations. The phase 1 grouting locations consist of installing casing into the bedrock where voids, soil seams, or phase 1 operation extends from the voided areas within the bedrock to a depth of 2 feet (0.6 meters) below the proposed bottom of trench elevation. During the bedrock interface and extends upward to the same depth criteria referenced for phase 1. Within the initial zones for each application, a higher slump material casing is raised into the overburden soils, the slump compaction grout. The pumping rate is maintained at per 2 foot (0.6 meters) stage is recommended based on the backpressure recorded at the given depths. Table 1 provides the recommended pressure versus grout volumes per 2 foot (0.6 meters) stage. bedrock suggests that the upper crust close to the ground displays the test boring locations chosen based on the geophysical testing. Analysis of microgravity data from the new alignment revealed that relatively shallow and dense rock was present with isolated anomalies in most and MASW data in the relocated alignment revealed low shown in Figure 5. Additionally, isolated anomalies are located along the remainder of the proposed relocation route. As before, strong correlation was found between the geophysical data and the new boring data. and generally poor quality rock were found in the bedrock. Along the remainder of relocation route, some isolated areas of weak soil, voids in the bedrock and poor quality rock exist. Further complicating the new alignment is that poor quality carbonate rock is more susceptible to dissolution and weathering than higher quality rock. investigation for the new alignment indicated that recommended program consists of a subsurface grouting program along portions of the proposed relocation route prior to installation of the new pipeline. The grouting program is required within the initial 70 as in areas where the isolated anomalies exist. A grout way where a fracture trace exists. Since sinkholes have a tendency to develop over fractures in the bedrock, the grout curtain is expected to reduce the potential path(s) in the underlying bedrock. The recommended subsurface grouting program is based on the level of risk for potential sinkhole formation test boring operation performed. In areas that possess the greatest risk for sinkhole activity, targeted grouting is recommended to be performed in a grid pattern around for sinkhole activity, additional compaction grouting is Recorded Backpressure Volume of Grout to be Injected 1.0 yd (0.764 m 0.5 yard (2068 2758 KPa) 0.25 yard (0.191 m > 400 psi (2758 KPa) raise to next stage TABLE 1. Grout volume cut-off criteria.


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE20matrix and grouting continues, a decreased volume of higher backpressures. Figure 7 displays the expected results of the recommended grouting operation.ConclusionsThis paper shows that geophysical testing using microgravity, MASW, and 2D ERI can predict the occurrence of active sinkholes in pinnacled carbonate bedrock. If subsurface grouting is being considered as a method for sinkhole stabilization or treating sinkhole prone site, a comprehensive geophysical in developing an effective scope of work for the ground improvements. The information gathered is also surface is typically characterized by stiffer clayey soils. As the depth increases, a decrease in stiffness or more coarse soils is typically found above the fractured carbonate bedrock surface. When steel casing from the grouting program is installed to the bedrock surface, intent is to seal the open throat in the rock surface. As the steel casing is raised into the lower zones of the overburden soils, a low mobility displacement grout is used to densify surrounding weak zones of soils and interface. Within the zones of the weak soils, it should be expected that the higher volumes of grout will be experienced. As the casing is raised into a denser soil Figure 6. Proposed grouting location plan within realignment route.


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 221 Shultz C, Kochanov W, Wilshusen J. 1999. The geology foundation soil. In: Practical foundation instrumental in developing a cost estimate for the ground improvement work. Furthermore, by basing the volume reduce the risks of sinkhole development and reduce costs for potentially problematic sinkhole recurrence.References Figure 7. Conceptual sketch of grouting. Base sketch used to show grout from J.P. Wilshusen & W.E. Kochanov, The Geology of Pennsylvania, 1999.




13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2(Warner, 2004). In theory, the volume of grout placed in the ground will cause an increase in density in loose sandy soils as the expanding grout displaces soil and thus compacts and increases the strength of soil between will be obtained from the compressive strength of the grout columns placed typically six to ten feet apart but ground surface; however, the primary use of this method function of the grout is to seal any seepage paths that may exist at the rock soil interface. Grout Application Problems rate causing hydraulic fracturing of the soil. In this instance, high pore pressures develop that cause the soil to fail in an undrained state, remolding the soil into a pressures generated by the rapidly expanding grout front (see Figure1). The hydraulic fracturing interferes with the orderly compaction process and can cause damage in the buildings. Damage to overlying structures can be caused by the increase in overburden weight from the soil that has been intruded by lenses of grout as shown in Figure 1. The increase in soil weight can sometimes result in settlement of the building being remediated. Nearby buildings can also be damaged from the intrusion of grout into utilities and into the building. Some assert that contactors monitor heave while pumping and pumping can be stopped when movement is seen. This sounds reasonable in theory but in practice there are a number of problems. First, there is a time lag from the time the inspector happens to notice movement till the time he communicates that to the pump operator. AbstractCompaction grouting is a widely used method for sinkhole remediation. It is generally less costly than other methods of remediation and provides a less intrusive method of repairing adverse subsurface practice have improperly applied compaction grouting as a method of remediation. In some cases, improper use of compaction grout has resulted in the deterioration of the cost of repair and inconvenience to the homeowner. Another consideration in the selection of compaction grout is the occurrence of subsurface conditions in which deep foundation support should be used rather than the seemingly less expensive compaction grout method. This manuscript discusses techniques in the proper use of compaction grouting and the precautions that should be taken before, during and after compaction grouting. It also discusses potential conditions when compaction grouting should be supplemented or replaced with deep foundations. Included in the manuscript are compaction grouting case studies and recommendations for the proper application of compaction grout. IntroductionCompaction grouting is a common method used to remediate homes affected by sinkhole activity. It generally provides a relatively fast, effective and economical method of soil improvement. The compaction grouting and compact the surrounding soil. The common practice is to apply compaction grout from the rock surface upward (upstage grouting) by building successive segments of grout such that one segment rests on the segment below until the grout reaches the desired depth. Fundamental to the success of the grouting procedure is deposition of the grout in a globular mass (typically PROBLEMS ASSOCIATED WITH THE USE OF COMPACTION GROUT FOR SINKHOLE REMEDIATION IN WEST -CENTRAL FLORIDAEdward D. Zisman, P.E., P.G., M ASCE, and Daniel J. ClareyCardno ATC, 5602 Thompson Center Ct., Ste. 405, Tampa, Florida 33634 USA,, daniel.clarey@atcassociates.com23


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE24wrong such as damage to the home, grout deposited in a neighboring property or settlement sometime after completion of remediation is the grouting procedure questioned. By that time it is too late to correct the conclude that this is one of the shortfalls of compaction grouting. The delay in determining if the grouting was successful is a concern for all and is minimized by the procedures discussed. Second, once movement starts it may continue for a period because of pressure in the formation. Third, when movement occurs, even if it stops when pumping is stopped, it may be too late, the building can be immediately damaged. These problems, in many cases, pale relative to the greatest impetus to increase grout flow rates, to the highest possible rate. This is the increased cost for pumping grout at low flow rates. The lower flow rate increases the time required to complete the grouting hence labor and equipment costs increase for the grouting contractor and for inspection. Costs for supply of grout also increase because of the increased time to use the grout. Typically, most contracts adhere to ASTM C94 requirements for discharge of the concrete within a 1 hour period from batch to placement. If this time is exceeded the concrete cannot be used. This means that instead of the grout supplier providing 10 cubic yard trucks they must deliver grout in 5 cubic yard trucks. This therefore increases cost. one actually sees the completed productit is unseen below the ground surface. Only when something goes Figure 1. Adverse effects of high flow rates. Grout Flow < 2 5 cf/min Grout Flow > 5 cf/min Downward Movement of Unsupported Grout and Soil Downward Movement of Unsupported Grout and Soil Recommended Methods of Resolving Grout Flow ProblemsA solution to the dichotomy of cost verses compaction grout quality that considers both technical and economic hydraulic fracturing occurs in soft soil areas. This is done pressure occurs (presumed to be the onset of hydraulic fracturing of the soil). The procedure is performed in known areas of soft soil found in existing borings or at the location of soft soil conditions found in the newly installed results in a decrease in pressure. In other areas, with minute (0.142 to 0.198 cubic meters per minute) is used.


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 225Variable Soil and Rock ConditionsIt is important to use all subsurface information that is available to analyze the diverse conditions that occur in karst terrains. To determine potential areas where soft soil is recommended that consideration be given to the depths to sound rock found in the grout drill holes. Figures 2 through 5 show two sites where grout hole information is known. The point in illustrating this data is to show the stark difference in the interpretation that occurs when additional information is available. Compare the differences in the depth to rock found from grout holes as opposed to information obtained from SPT borings where distances between data points are very great. The grouting data points show the extreme variability in the rock surface that was not found in the SPT data. Therefore, the advantage in using grout hole data is that one can anticipate where soft soil conditions may occurin karst areas this is common in locations of abrupt changes in in areas of abrupt changes in depth to limestone. When Not to Use Compaction Grout If more than several inches of settlement have occurred in a structure, lifting a building component should be accomplished through means other than compaction can be accomplished by the use of chemical grout (polyurethane foam in low viscosity liquid form pumped at low pressure into cohesionless soils) where loads and be accomplished with chemical grout on some slabs with moderate loads depending on geometry and loading. piles. The limestone rock surface tends to be highly solutioned and weathered resulting in a surface of questionable integrity to support a load. Unfortunately, the quality of the limestone investigated to determine its competency. Figure 6 provides an illustration of a typical limestone surface that may be encountered for support of pile loads. When these conditions are anticipated, an additional subsurface investigation should be performed to determine the integrity of the rock. Figure 2. Depth (in meters) to sound rock from borings in Site 1 (no horizontal scale).Figure 3. Depth (in meters) to sound rock from grout holes in Site 1 (no horizontal scale).Figure 4. Depth (in meters) to sound rock from borings in site 2 (no horizontal scale).Figure 5. Depth (in meters) to sound rock from grout holes in Site 2 (no horizontal scale).


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE26 The net increase in soil weight, due to a high underlying soil and the building foundation supported by the soil (Warner). 4. Compaction grout is not a process where the weight of the building is supported on a column of grout; it is a process where compaction of the soil occurs from the inclusion of a volume of grout between successive grout columns compacting the soil and increasing soil strength. The strength of compaction grout is only required to meet or exceed the in situ soil.ReferencesASCE. 2010. Compaction grouting consensus guide. Warner J. 2004. Practical handbook of grouting. John Wiley & Sons, Inc. handbook, Vol. 2: Procedures, Ernst & Sohn. Standards, Vol. 04.02: Concrete and Aggregates. the order of greater than 20 feet (6.1 meters), a cost comparison should be performed to determine which method is economical. The analysis should include the for additional investigation of the quality of the rock surface to support a pile load. Other ConsiderationsThe use of compaction grouting is directed to remediating soft soil conditions; however, in doing so, areas of dense The net result is that the grouting process may loosen these areas. When large areas of dense soil are known to be present on a site, the extent of the grouting program to determine the grouting effort to be used in the various grout holes.ConclusionsIt has been discussed that: 1. unacceptable lateral displacement of the grout distances beyond the point of placement. This causes remolding of the soil greatly adding to (Figure 1). 2. causes a decrease in pressure in soft soil areas or in any area where a decrease in pressure is found. minute (0.142 to 0.198 cubic meters per minute) is used.Figure 6. Typical limestone surface.


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 (settlement) risks, and mitigation options Explain the uncertainty Negotiate liability Costs of investigation and mitigation reduction they will receive The discussion of karst risk should be ongoing and to investigate all sites underlain by a potentially karstic unit or try to rank the sites based on risk before focusing the turbine karst investigation costs can easily reach $20,000 possible, start karst evaluation early, manage available cash with a stepwise approach, and communicate.Introduction seismicity risk, karst risk is not addressed by the Federal may or may not be addressed by local building codes. are often chosen for wind farm sites. In these areas, there is a limited frame of reference for observing subsidence, fewer eyes observing the ground, and, normally, no reason for anyone to care about sinkholes. A sinkhole in downtown Miami gets more attention than a sinkhole in rural Texas. Karst can lead to a wind turbine tilting and even toppling. AbstractKarst can cause a litany of problems for a windpower footprints that can cost $2 million to $5 million each. can be split into the two categories outlined below. Desktop studies: Search for relevant literature lineament Search for existing well and boring logs Survey local expertslandowners, U.S. cavers, etc. Field studies: Perform site reconnaissance Conduct pit tests if bedrock is shallow DrillA normal geotechnical investigation includes one boring per turbine, while karst investigations may include multiple borings per turbine Use a downhole cameraMay be useful in evaluating extent of voids and convincing clients of risk. Conduct geophysical studies Effectively communicating with developers is critical. They want to know the location of the problem sites and may ask, If there is a cave, what is the chance that a following effectively:EVALUATING KARST RISK AT PROPOSED WINDPOWER PROJECTSWilliam J. BangsundBarr Engineering Co., 4700 W. 77 St., Minneapolis, Minnesota 55435 USA, BBangsund@barr.comKenneth S. JohnsonOklahoma Geological Survey, 1321 Greenbriar Dr., Norman, Oklahoma 73072 USA, ksjohnson@ou.edu27


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEstepping forward toward viabilitymore funding becomes available. The additional funding affects the stepwise so the early karst evaluation phases can be completed inexpensively, and the more expensive phases are done later when more funding is available. If possible, the karst professional needs to educate the geotechnical investigation and foundation design are packaged with the construction. In these cases, the issues and implications of karst may come as a surprise, at a point when there is no turning backthe turbines have typically already been purchased. Once in construction, a client has reaction, until faced with what karst evaluation can cost.28 the turbine to be out of tolerance and lead to expensive widely spaced for optimum performance (see Figure 1), so each proposed turbine location may need to be evaluated independently for karst risk. An installed turbine can cost $2 million to $5 million, so the liability is high.Figure 1. Typical wind farm. Note widely-spaced wind turbines in a remote setting.Figure 2. Typical turbine section and major forces. Spread footings are most common. Hub heights 80100 m (but can go up to 120 m); foundation width 15-22 m; foundation embedment 2-3 m; overturning moment: 35,000 kN*m 110,000 kN*m; dead load: 1,850 kN 5,100 kN forces: the wind load, dead load, lateral load, and overturning moment. Turbines have relatively low dead loads but relatively high overturning moments. While there are several types of foundations that can be used, the most common by far is the spread footing shown on Figure 2. The discussion in this paper generally assumes and relates to the use of spread footings. Note that the ground strength rarely affects the foundation diameter. years are spent completing the development phase. relatively speculative and available funds are limited. NCKRI SYMPOSIUM 2 Figure 1. Typical wind farm. Note widely spaced wind turbines in a remote setting. forces: the wind load, dead load, lateral load, and overturning moment. Turbines have relatively low dead loads but relatively high overturning moments. While there are several types of foundations that can be used, the most common by far is the spread footing shown on Figure 2. The discussion in this paper generally assumes and relates to the use of spread footings. Note that the o verturning moment is such a significant factor that ground strength rarely affects the foundation diameter. Commercial scale typically include 10 100 turbines. Employing a common foundation or foundations to address site specific conditions, the become untenable. Figure shows the basic timeline for building a typical wind farm. Once a promising site is identified, several years are spent completing the development phase. hase it is still relatively speculative and available funds are limited. along the development process stepping forward t oward viability more funding becomes available. The additional funding affects the karst evaluation process K arst evaluation should be stepwise so the early ka rst evaluation phases can be completed inexpensively and the more expensive phases are done later when more funding is available. If possible, the karst professional needs to educate the developer and work with to use funds efficiently Note that many developers structure so that the geotechnical investigation and foundation design are packaged with the construction In these cases the issue s and implications of karst may come as a surprise at a point when there is no turning back t he t urbines have typically already been purchased. Once in construction, a client has little patience is the common reaction until faced with what karst evaluation can cost. This paper will address : the typical karst investigation methods the ways karst risk can be mitigated the issues that must be addressed in communicating with the client Figure 2. Typical turbine section and major forces. Spread footings are most common. Hub heights 80 100 m (but can go up to 120 m); foundation width 1522 m; foundation embedment 2 3 m; overturning moment: 35,000 kN*m 110,000 kN*m; dead load: 1,850 kN 5,100 kN


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2Field studies: Site reconnaissance Pit tests Drilling (may include downhole camera and downhole mapping methods) These methods are listed, approximately, in the order of increasing cost. Because of their cost, drilling and geophysics are usually not undertaken until late in especially from geological surveys, is often extremely Literature searches are commonly used on all manner of geologic studies, and there is no need to discuss them further here. One example of something that may be karst occurrence. For example, much of southeastern Minnesota is underlain by carbonate bedrock, but in This paper will address: the typical karst investigation methods the ways karst risk can be mitigated the issues that must be addressed in communicating with the client Investigation Methods to evaluate karst risk while keeping costs under control. We have followed a commonly used program (Fischer et al. 1987; Roux, 1987; Tonkin & Taylor LTD, 2011). Not every tool is necessary or appropriate for every site: Desktop studies: Literature search Existing well and boring logs search Survey of local experts29Figure 3. Timeline for developing and operating a typical wind farm.


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE and epikarst development associated with the deeper karst is commonly why lineaments are expressed on the should be conducted, where appropriate, to identify terrains have relatively thick soil covers unrelated to the bedrock that can obscure bedrock lineaments. Lineament analyses have limited or no application in these areas. There is more than one type of karst, and investigations and mitigation must be appropriate to the local whose experience in Oklahoma with evaporite karst was invaluable in evaluating evaporite karst risk at the addition to geological surveys, other geologic experts can include landowners and speleological societies. Well logs are a valuable source of information. More and online. Some examples include: method with a long history. Maps often show the locations of karstic features, especially springs and sinkholes (Figure 4). USDA Natural Resources Conservation Service soil mapping also includes sinkholes and other karst features for many areas and is available nearly staff can review aerial photographs and topographic maps for apparent karstic features. Today, much of this information is available online, but it is still important to look for historic aerial photographs so the site can be viewed from different perspectives relative to the season and time of day. Modern methods such as interferometric synthetic aperture radar and digital elevation models may be particularly valuable.30Figure 4. Map of proposed wind farm development area showing mapped karst features and one example of lineation of features.Figure 5. Map of a Scurry County, Texas Wind farm project area showing mapped lineaments. Labeled dots are proposed turbine locations.


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 anomalies must then be determined through drilling. Risk characterization has a number of questions: formation? Are there any known karst features in the region? Are there karst features at the proposed turbine sites? The results at each stage of evaluation are used to determine if more investigation is required and, if so, the scope of the next phase. shallow carbonate or evaporite bedrock and no evidence of karst from the desktop phase or reconnaissance. The lack of evidence does not mean there is no risk. The question then is, how much investigation is required? Lineament analysis has been used to identify areas with higher potential risk. Then, intense investigation can be completed in these areas. If no subsurface voids are found, it may be acceptable to forego further karst investigation in other areas.Risk mitigation mitigation must be applied. More than one method of are several ways of mitigating karst risk: Move the turbines at risk. It may be possible learned to include alternative locations early in the process for this type of outcome. Depending on the number of sites that are eliminated and the number of alternate sites, the cost may range from practically nothing to the loss of the investment and revenue related to the net lost sites. Conduct detailed investigation. Some developments may have very limited constraints risk alternative locations may not be available. A developer can then decide to do more intensive uncooperative because landowners may be concerned about the effect of karst on their land value, and cavers are often reluctant to share private mapping with outsiders Site reconnaissance is important for the general characterization of the area. It may also identify karst features near or at individual turbine sites. Classic are important because so much cost and risk can be to quarries is especially valuable even if outside the Where bedrock is shallow, test pits can be useful in evaluating the bedrock surface and investigating the nature of depressions to determine whether or not they are related to karst formation. equal to the width of the turbine foundation, and the depth is chosen based on the vertical stress induced by the foundation (Das, 2010). Karst investigations may include multiple borings per turbine. The question is, how many are required to assess karst risk? Advanced geotechnical modeling can provide an indication of the size of void verses depth that may be problematic. vary across the proposed wind farm, requiring multiple models. The cost of drilling multiple borings per turbine quickly increases the cost of investigation. This can be especially useful in convincing the client that there is a risk. Although not used by these authors, laser The use of geophysics in karst evaluations is well studied and reported, and it is regularly addressed at karst conferences (Beck and Wilson, 1987; Beck and tools, but it rarely attains a useful depth of penetration; below grade. In fact, most geophysical methods lack the 31


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE Dont build the project. Developers typically best approach may be to move on to the next one. This means losing the investment to that point, so this is not done lightly. There is often great pressure to move forward despite the evidence of karst. As noted previously, the earlier that karst risk can Risk communicationThe cost of failure of a single turbine can range from hundreds of thousands of dollars (slight but unacceptable differential settlement) to millions of dollars (extreme tilt to catastrophic collapse). It is therefore important to communicate the cost implications to the client as risk of karst is the apportionment of risk amongst the Karst risk and risk apportionment is a very important conversation. follow through with the level of investigation needed to completely characterize the risk or carry all the liability per turbine, which does not offset the potential for a lost $5 million turbineespecially when that risk is multiplied by tens or hundreds of turbines. Therefore, it is important to educate the client about karst and karst risk to the extent that the client can carry the bulk of the risk and can make informed decisions regarding the degree of risk and how extensive the risk characterization will be. Effectively communicating with developers is critical. They want to know the exact location of the problem sites and may ask, if there is a cave, what is the chance that a turbine will fail? The developers cases, the issue is cover collapse or soil piping, not cave collapse. It is also important to communicate the inherent uncertainty of karst risk and the cost of reducing the uncertainty. investigation of a proposed turbine location to see if moving the turbine a short distance can reduce risk. This method of mitigation can add tens of thousands of dollars and may not be successful. Provide thick soil cover to mitigate the risk of subsidence. In some areas, thick soil unrelated may provide an effective bridge over bedrock karst features, and soil thickness may be preliminarily determined based on existing mapping and drilling logs. Eventually, each proposed turbine site should be drilled to determine actual soil thickness. needs to be answered. There may be precedents. expressions in Minnesota occur where there is less than 15 m of glacial cover. The Minnesota of karst features was not found for areas with similar for terrace and dune deposits (Johnson developer need to come to their own conclusion. Since a typical geotechnical investigation for foundation design includes borings at each proposed turbine site, this mitigation method is Use construction methods. Most turbine spread foundations are relatively shallow (~2 the foundation can be placed on piles that are supported on rock below the karst zone. This may require additional investigation of the bedrock for the design of a pile foundation. Another option is to grout the underlying voids full to eliminate the potential for collapse. One advantage with grouting is that you can complete the detailed investigation to identify voids at the same time as the mitigation is being completed. Another possible construction method not encountered by these authors is to construct a foundation that bridges the risk zone. While a typical spread foundation is likely capable of bridging a small gap, the normal design process does not evaluate that possibility. Such a design consideration basis. Constructed mitigation adds hundreds of thousands of dollars to the cost of each turbine. Note that implementing constructed mitigation often means that detailed karst characterization is no longer required.32


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2Project Examples where karst risk was evaluated mainly by the senior author. Following are some brief descriptions of a few of these sites.North Central Iowa indicated that the bedrock is dolomitie (as opposed to limestone), with which karst development is linked in this exists over most of the site to mitigate risk (Figure 6) and did liability: Ignore the issue. This is clearly unacceptable. Add a disclaimer. The disclaimer will state that it is impossible to completely know what is underground. This is a typical practice. Keep the investigation and evaluation of karst out of scope. In other words, pass the buck. Educate the client. workload and risk.33 Table 1. Project Summaries. NA = Project did not advance Site Location No. of turbines Built? Lit Search Remote Sensing/ Lineament Experts Recon Drill Geophysics Comment Arizona 1 62 No Yes Yes Yes NA NA NA progressed past desk top phase Arizona 2 62 Yes Yes Yes Yes Yes Yes Yes Developed area was reduced Iowa 79 Yes Yes NA Yes Yes Yes Yes Kansas 100 Yes Yes Yes Yes Yes Yes Yes Minnesota ~140 No Yes Yes Yes Yes NA Yes Unbuilt as of spring 2012 New York ~90 No Yes Yes Yes Yes NA Yes early. Develop er kept looking for a different answer Ohio 175 Yes Yes No Yes Yes Yes Yes Oklahoma 1 129 Yes Yes No No No Yes Yes Due to constraints and schedule, investigation Oklahoma 2 ~90 No Yes No Yes Yes NA NA Dune cover Watonga Pennsylvania 24 Yes Yes NA Yes Yes Yes Yes Expensive mitigation Texas 1 160 Yes Yes Yes Yes Yes Yes Yes Field investiga tion was limited based on linea ment analysis Texas 2 242 Yes Yes Yes No Yes No Yes 260 Yes Yes Yes Yes No Yes Texas 4 28` Yes Yes Yes Yes Yes Yes Yes cal modeling


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE34Southwest PennsylvaniaLiterature review indicated, and site reconnaissance confirmed, that karst features were present in the area. Karst was associated with particular stratigraphic units, so areas of relative risk could restrictions on where development could take place, and those limitations took precedence over karst risk. before any subsurface investigation was completed. Once drilling began, numerous subsurface voids were found beneath most of the proposed turbine enough detail, so multiple drill holes were completed at turbine locations that were at risk. Although not budgeted for, the developer ended up installing deep pile foundations at some sites and grouting voids in others, at great expense.South Central MinnesotaThe client was a contractor bidding on constructing Figure 6. Cross section of wind project in North Central Iowa showing depth to bedrock. Thicker soil=less risk. Figure 7. Map showing relative risk for a wind farm in southwest Pennsylvania.


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 235areas of Minnesota (Figure 4). The contractor was continued to try to bring it to fruition for several years.Northwest Oklahoma Investigations in Blaine County, in northwestern Oklahoma, evaluated potential problems that gypsum karst may pose for the proposed Watonga Windpower depths ranging from 10 to 45 m below ground level. The Blaine is overlain by the Permian Dog Creek Shale and by unconsolidated Quaternary sands, clays, and gravels that may obscure karst features. Field showed that there is no direct evidence of gypsum gypsum beds was appropriate risk mitigation: where gypsum is 25 m below ground level or deeper, the risk related to gypsum karst is low, and where gypsum beds are less than 25 m deep, risk was medium to high. A map (Figure 9) was prepared showing areas of low, medium, and high risk related to gypsum karst.Figure 8. Cross section of shear wave velocity showing a sinkhole underlying a proposed wind turbine site in southwest Pennsylvania. Boring blow count decreased with depth. Figure 9. Risk categories at Watonga Windpower Project, based upon depth to the Shimer Gypsum at top of the Blaine Formation (Johnson et al., 2013)


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE36ConclusionsKarst can lead to dramatic tilting and even toppling of a wind turbine. Subtle differential settlement of turbine foundation can cause the turbine to be out of tolerance, requiring remedial action. There are many tools available for evaluating karst risk desktop methods and field methods with widely ranging costs from reconnaissance to intensive drilling. The right tools at any given phase of a windpower development will be based on the management discussions with the client.References Special Publication No. 122. Proceedings of the Alabama. ASCE. Beck BF, Wilson WL, editors. 1987. Karst hydrogeology: Engineering and Environmental Applications. Proceedings of the 2nd Multidisciplinary Conference on Sinkholes and the Environmental Impacts of Karst. Orlando, Florida. Rotterdam (Netherlands): A.A. Balkema. Beck BF, Stephenson JB, editors. 1997. The Engineering Proceedings of the 6th Multidisciplinary Conference on Sinkholes and the Engineering Missouri. Rotterdam (Netherlands): A.A. Balkema. Das, BM. 2010. Principles of geotechnical engineering. Planning & design consideration in karst terrain. In: Engineering and Environmental Applications. Proceedings of the 2nd Multidisciplinary Conference on Sinkholes and the Environmental Impacts of Karst. Orlando, Florida. Rotterdam Department of Natural Resources. County Atlas and potential problems in siting wind turbines in Blaine County, Oklahoma. NCKRI Conference Proceedings. fracture trace and occurrence of groundwater in of sites proposed for development in the dolomite karst regions of southern Africa. In: Beck BF, Engineering and Environmental Applications. Proceedings of the 2nd Multidisciplinary Conference on Sinkholes and the Environmental Impacts of Karst. Orlando, Florida. Rotterdam Tonkin & Taylor, Ltd. 2011. Puketoi Wind Farm Prepared for Mighty River Power Ltd. Monroe Counties, West Virginia. Map ECS 17.


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 237 stability charts may be combined with reliability concepts to characterize the risk of collapse of a void in the soil overlying the rock surface.Simplified Charts for Soil StabilityStability charts are widely used for the evaluation of 1960) where the charts were developed in terms of the slope height and inclination, and the soil shear strength is expressed in terms of the soil cohesion intercept, c, These stability charts are typically presented in terms of a dimensionless stability number, (1) where N is a dimensionless stability number, is the c is the cohesion component of the soil shear strength. Typically, the charts allow the potential for failure to be of the available soil strength to the strength required to maintain stability. (2) where the parameters cdd are the corresponding values of the cohesion intercept and friction angle required to maintain equilibrium. Using some of the concepts originally applied to soil slopes, Drumm et al. stability of a void in the soil overlying the rock surface.AbstractThe karst belt stretching from Alabama to New England clay soil above a highly weathered rock surface. As part of the natural weathering process, subterranean voids frequently develop in the overburden soil, which can lead to surface subsidence or collapse (sinkholes). Furthermore, construction activities can promote instability, especially where a portion of the soil overburden is removed. A rational method for addressing the potential for void perform probabilistic analysis for likely ranges of void and soil conditions. This paper demonstrates the application evaluating the collapse potential of voids within the soil overlying the rock surface.IntroductionSubterranean voids in the bedrock and in the overburden soil develop as part of the natural weathering process in the karst belt stretching from Alabama to New England, clay soil above a highly weathered rock surface. A methodology for evaluating the static stability of discrete voids (i.e., caves) within shallow rock is presented by Siegel et al. (2001). Drumm and Yang, (2005) and evaluating the static stability of a void within the soil determination of representative void sizes and geometry, APPLICATION OF STABILITY CHARTS AND RELIABILITY CONCEPTS FOR SIMPLIFIED ANALYSIS OF A VOID IN SOIL OVERLYING KARST BEDROCKTimothy C. Siegel, PE, GE, DGEDan Brown and Associates, PC, 1808 Northshore Hills Blvd., Knoxville, Tennessee 37922 USA, tsiegel@danbrownandassociates.comDanner F. Drake, PECDG Engineers & Associates, Inc., 1830 Hartford Hwy., Dothan, Alabama 36301 USA, ddrake@cdge.comEric C. Drumm, PE, PhDDept. of Biosystems Engineering and Soil Science, The University of Tennessee, Knoxville, Tennessee 37996-4531 USA, = = = tan tan


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE38Stability Chart for Void in SoilA subterranean void will be stable where the overlying material below. The ability of the soil to redistribute the stresses will depend on the void geometry, the soil thickness, the soil strength and the magnitude of the surcharge load, if present.Characteristic Subsurface Profile From the ground surface, there is a blanket of soil that is composed of the insoluble portion of the karst consolidation as a result of exposure to multiple cycles of wetting and drying. With depth, the residual soil generally increases in water content and decreases in stiffness and strength. Competent karst bedrock (e.g., limestone) typically exhibits high strength but contains slots, caves, and other openings created by the solutioning process. Voids in the soil or domes are created as the soil ravels openings in the underlying rock (Figure 1).Finite Element ModelThe dimensionless chart developed by Drumm et al. (2009) to evaluate the stability of a void in soil overlying karst The analyses were conducted for a range of hypothetical soil properties and void geometries expressed in terms of the ratio of an assumed hemispherical void diameter (D) to soil overburden thickness above the void (h). The analyses are shown in Figure 2.Figure 1. Conceptual subsurface profile in karst with an enlarging void in the residual soil (Sowers,1996).Figure 2. Axisymmetric idealization of void in soil over rock (Drumm et al., 2009). summarized in the following: 1. The geometric conditions around the void were model, implying a hemispherical void of diameter D. The soil was assumed to be homogeneous except for analyses that assume a weaker soil 2. The stiffness of the rock was much greater (typically 104 times) than that of the soil and, as a result, the rock was considered to provide a rigid support at the base of the soil. Therefore, the rock the model; lateral extent (L) for the largest diameter was extended until it had negligible effect on stability. The results indicated that there was no boundary 4. m was incorporated into the dimensionless terms; 5. the soil to act as an elastic solid at stress levels plastically at stress levels equal to the strength. inherently assumes that the intermediate principle 212 condition (Chen and Liu, 1990) and the failure is = +


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 239The dimensionless stability number (Nc) was determined by applying the shear strength reduction (SSR) method proposed by Zheng et al. (2006). In the SSR method, which is widely used in both soil Swan and Seo, 1999), the strength parameters of the model are reduced by a strength reduction factor (SRF), such that (6) incrementally increasing values of SRF until the analysis does not converge to equilibrium. This determines the The critical SRF can be used to calculate the critical strength and Nc.Soil Friction AngleAnalysis using only undrained shear strength may be considered representative of short term conditions. To extend the analysis to long term (or effective stress) conditions, the stability was also evaluated using the corresponding to a convergent solution for values of and the initial stress ratio following Eq. (6). The stability Inverted Strength Profile increases with depth (as is the case in most geologic settings), karst often exhibits a soft zone above the rock surface. This is often referred to as an inverted shear strength (c*). (7) where c* is the reduced undrained shear strength undrained shear strength of the soil; and is the stability numbers for undrained conditions with inverted strength factors of 0.25, 0.5, and 1.0. where strength parameters c and represent the cohesion intercept and angle of internal friction, respectively, and dilation angle ( which results in the soil experiencing zero volume change during yield. The tensile strength was assumed to be 20% of the undrained shear strength values (cu). This assumption, while somewhat arbitrary, allows for a variation in tensile strength in proportion to cu while maintaining the dimensionless stability factors; 6. The elastic modulus of the soil (E) was assumed to be 22 MPa (4.6 x 105 psf). Although the stability is not sensitive to the elastic modulus provided it is a constant, this value is consistent with published correlations with the undrained shear strength (Das, 1999). (4) where cu is the initial value of undrained shear and the surface subsidence were not considered; 7. is consistent with published values for a variety of soil types (Bowles, 1988). In general, the results of the evaluation are somewhat sensitive 8. by restraining the soil around the void while applying the gravitational force with a stress ratio Ko according to Equation 5 (5) after which the soil around the void was released allowing deformation, and; 9. The water table is assumed to remain constant at a position below the top of the rock surface. This assumption results in the greatest effective stress for any of the conditions considered. Enlargement of the void due to soil loss is neglected and seepage effects on stability are not considered.Determination of Collapse Load geometry where h is the minimum soil thickness over the void (h = H-D/2) and D is the void diameter (Figure 2). = 440 = 1 = c + tan SRF c* = c


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE40Case History: Landfill in Karst Terrain presented herein were used to evaluate the collapse potential of voids within the soil during the permitting following paragraphs.Geologic and Subsurface ConditionsPublished maps show that site is located within the light gray and light brown, locally sandy dolostone, dolomitic The geotechnical exploration consisted of soil test resistivity. On the basis of the exploration results, the subsurface conditions are characterized by a thick layer of residual soil consisting of very stiff (average SPT N of 28), sandy clays and silts with interbedded seams of clayey gravel (chert) and sand. The soil thickness ranged There was a slight decrease in SPT N from 20 to 50 ft below the ground surface. The soil strength was characterized based on the results triaxial tests that were performed on soil samples obtained in similar geologic and geotechnical conditions. The strength test results are summarized in Figures 4 (total stress or undrained strength) and 5 (effective stress or drained strength). Figure 3. Stability chart for estimating Nc for a void in soil overlying rock (Drumm et al., 2009). Table 1. Constants and r2 values for curves in Figure 3 (Drumm et al., 2009).Functional Form of Stability ChartTo allow direct use of the stability chart shown in Figure following form. Nc2 where a, b, c and d are constants determined by regression analyses. The values of constants a, b, c, and d for a range of values of and are presented in Table 1.Reliability ConceptsReliability concepts provide a useful framework for analysis 1987; Whitman, 1996). For application of the stability chart presented herein, it is proposed to incorporate the approach proposed by Duncan (2000) which allows an assessment of probability of collapse using the following steps. 1. Estimate the standard deviations of the parameters of all values of a normally distributed parameter fall within three standard deviations of the average. The standard deviation is computed using the Equation 9. (9) LCV is the lowest conceivable value. 2. Use the Taylor series technique (Wolff, 1994; U.S. Army Corps of Engineers, 1997 and 1998) to Determine the probability of failure and the lognormal distribution of values. Duncan (2000) presents a table that summarizes the mathematical results necessary to apply a lognormal distribution. Constants a, b, c and d along with r2 a b c D r2 =


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 241kPa (1550 psf). An inverted strength factor ( ) of 0.6 was applied for undrained conditions. The effective friction angle ranged from 20.4 to 20.9 degrees and averaged 20.6 degrees. The effective cohesion ranged Probability of Void CollapseFollowing the Duncan approach (2000), the Taylor Series was used to compute the probability of void determined where each parameter is individually increased and decreased one standard deviation (s.d.) for the most likely values (MLV) are 2.74 and 2.79 for total stress conditions and effective stress conditions, respectively. The standard deviations of the calculated safety may be determined using Equation 10. (10) conditions) and 56.2% (effective stress conditions). The lognormal reliability index (LN) values are calculated using Equation 11. (11) by a void within the soil. The method involves passing direct current through the earth between two electrodes and measuring the resulting voltage drop between an were considered potential subterranean voids.Void and Soil ParametersNo voids were encountered within the test borings, including those that were drilled at anomalies (extremely high resistivity values or extremely low resistivity values) testing. Considering the results of the geotechnical exploration and published data of doline diameter (Newton and Tanner, 1986; Martin, 1995; Qubain et al., 1995, Abdulla and Mollah, 1997; Mishu et al., 1997; Smith, 1997; and Thomas and Roth, 1997), a void diameter of 6 feet was considered to be a realistic, conservative assumption. It was anticipated that voids having a diameter greater than 6 feet, if present, would be detected during the resistivity testing and borings that target resistivity anomalies. This would allow application of corrective actions (e.g., cap diameter (or any other variable) could have been explicitly considered in the reliability analysis. (114.5 (120.5 pcf). The soil thickness (i.e., the overburden height (h) ranged from 7.8 to The undrained shear strength ranged from 40.2 to Figure 4. Total stress strength data (1 ksf = 47.88 kPa). = = ( (


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE42construction activities can promote instability, especially where a portion of the soil overburden is removed. A rational method for addressing the potential for and the probability of void collapse (Pf) can be calculated using Equation 12. (12) (total stress conditions) and 4.5% (effective stress conditions). According to Vick (2002), these values correspond to conditions where void collapse is between almost impossible to very improbable.ConclusionsSubterranean voids in the overburden soil develop as part of the natural weathering process in karst terrain. Even in cases where the soil strength is well characterized, there is often uncertainty with respect to the size and geometry of the potential subterranean voids. Furthermore, Figure 5. Effective stress strength data (1 ksf = 47.88 kPa).Figure 6. Typical resistivity profile (1 ft = 0.305 m) (Examples of anomalies noted by red circles). = 1 Table 2. Summary of factors-of-safety. Total Stress Conditions Variable c h g FOS (+s.d.) 4.09 2.21 2.61 1.49 2.89 2.60 0.28 Effective Stress Conditions Variable h FOS (+s.d.) 2.81 2.74 2.66 2.78 2.87 2.94 0.28


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 43 remedies for residential construction over karst limestone in Nashville, Tennessee. In Beck BF, Engineering and Environmental Problems in Karst Terrane, Proceedings of the 6th Multidisciplinary Conference on Sinkholes and the Engineering karst activity in the Valley and Ridge Province, Eastern Tennessee, Phase 1. Oak Ridge National to investigate and remedy sinkholes. In Beck Engineering and Environmental Problems in Karst Terrane, Proceedings of the 6th Multidisciplinary Conference on Sinkholes and the Engineering electrode earth resistivity testing in karst. Siegel TC, Belgeri JJ, McCrackin DM. 2001. Case history of east Tennessee karst: Static stability of shallow Smith JT. 1997. Sinkhole damage investigations for the Engineering and Environmental Problems in Karst Terrane, Proceedings of the 6th Multidisciplinary Conference on Sinkholes and the Engineering construction of foundations in karst terrain, ASCE, 202 p. Swan CC, Seo YK. 1999. Limit state analysis of earthen International Journal of Numerical and Analytical Reprinted in: Contributions to Soil Mechanics 1925 Drumm et al. (2009) to perform probabilistic analysis for likely ranges of void and soil conditions. In such a way, the potential for void collapse may be described in both numerically (i.e., probability of collapse) and verbally (e.g., very improbable, almost improbable, very unlikely). The example presented herein represents a snapshot of a hypothetical void under static condition. It is important to note that multiple analyses may be required to fully characterize the risk of void collapse.ReferencesAbdulla WA, Mollah MA. 1997. Sinkhole investigation: Engineering and Environmental Problems in Karst Terrane, Proceedings of the 6th Multidisciplinary Conference on Sinkholes and the Engineering Bowles JE. 1988. Foundation design and analysis. New York (NY): Elsevier Science. Das BM. 1999. Principles of foundation engineering. 4th Stability charts for the collapse of residual Drumm EC, Yang MZ. 2005. Preliminary screening of residual soil stability in karst terrain. Journal of Environmental & Duncan JM. 2000. Factor of safety and reliability in and Environmental Problems in Karst Terrane: Proceedings of the 5th Multidisciplinary Conference on Sinkholes and the Engineering and


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE44Thomas B., Roth JS. 1997. Site characterization for sinkholes in Pennsylvania and New Jersey. In Beck Engineering and Environmental Problems in Karst Terrane, Proceedings of the 6th Multidisciplinary Conference on Sinkholes and the Engineering U.S. Army Corps of Engineers. 1997. Engineering and design introduction to probability and reliability methods for use in geotechnical engineering. Department of the Army Engineering Technical Whitman RV. 1996. Organizing and evaluating uncertainty in geotechnical engineering. Wolff TF. 1994. Evaluating the reliability of levees. Rep. for the U.S. Army Engineer Waterways Experiment element slope stability analysis. Computers and


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 245exist. This hasty conclusion does not consider the high stresses imparted to the soil by the drill string and the inability of loose soil to support a void at relatively 2005). This paper will discuss the formation and testing of these conditions, their meaning in the context of sinkhole formation and suggested steps for determining sinkhole presence. An example of this condition as it occurred in an actual sinkhole investigation will also be discussed. relatively weak Cenozoic carbonates of Florida. These sediments consist predominantly of residual soils known to decrease in strength with increasing depth as opposed to transported soils which increase in strength with increasing depth (Sowers, 1996). This phenomenon is discussed in more detail in Section 6. The important the result of soil arching but the result of soft zones normally found in residual soils. Determination of whether soil arching has affected the subsurface is found from the characteristics of the underlying soil or rock are present in the underlying rock then one cannot rule out the possibility of soil arching. This is discussed in Also discussed are the requirements in the Florida statute that aid in the determination of sinkhole activity. showing conditions that are not indicative of sinkhole formation and the reasons for these conclusions.Stress Associated with SPT SamplingThe Standard Penetration Test (SPT) adopted by ASTM in Test Method D1586 is widely used in sinkhole investigations to determine the consistency and type of AbstractIn a Florida sinkhole investigation, many people (engineers, geologists, lawyers, insurance agents, public and weight of rod (WR) as a void, and by association, a sinkhole (author is a Florida Neutral Evaluator). This causes some to allege the site contains a sinkhole maintenance, construction or design issues. The concept sinkhole investigations becoming a gamble with the homeowner or their representative wagering against conditions found and therefore a sinkhole present under the building likely giving the homeowner a payoff for a sinkhole. The rules for the game are mandated in Chapter 726.706 of the Florida Statute that ultimately results in given set of conditions is or is not a sinkhole. Since the WR conditions and its meaning in terms of stress that develops during soil sampling. It will further consider the distribution of stress and the potential for these Conversely, it will also discuss the factors necessary for these conditions to impact a structure and other conditions that can give false indications of sinkhole activity. Also provided are examples of case studies where critical subsurface conditions were resolved using considerations discussed in this manuscript.Introduction soils, it is not uncommon to see reports written by professional engineers and geologists with the assertion (WR) conditions are present it implies a void is present below the ground surface and hence sinkhole conditions IT MUST BE A SINKHOLE? Edward D. Zisman, P.E., P.G., M ASCE, and Daniel J. Clarey Cardno ATC, 5602 Thompson Center Ct., Ste. 405, Tampa, Florida 33634 USA,,


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE46pressure with the drill string weight by a factor of over 2, which accentuates loose or soft soil zones that cannot support the increasing weight of the drill string resulting Another consideration is the stresses at the tip of the sample spoon are very large, for example, at 20 feet (6.1 meters) the stress exerted by the sampler on the soil psi (27.9 bar) and at 80 feet (24.4 meters) it is 800 psi (55.2 bar). Compare these stresses to the stress a women, wearing high heel shoes, places on asphalt that has been warmed by the sun. If the heel is one square inch in area, and a woman places 100 pounds (45.4 kg) on each leg they will apply a pressure of 100 psi (6.9 bars) enough bars) to 800 psi (55.2 bars) (see Figure 1) some consider a void present if the soil at that depth will not support the drill string. For the WR conditions, many consultants only report the condition is present without providing information on the rate of rod fall. Depending on the type of soils, the rate of rod fall can be useful in determining the type and material occurring at depth below the ground surface. and the drill string drops under its own weight (WR) or one considers the stresses that exist at the end of the drill string in relation to insitu stress. First, consider the stresses that are present at the tip of the drill string during SPT sampling as shown in Figure 1. These stresses are based on the following assumptions: 1) buoyant conditions are present with a buoyant soil m) length of drill rod, the difference in weight between drill string weight with depth together with a plot of the buoyant soil weight of the column of soil replaced by the drill string with depth. It is apparent that the drill string weight exceeds the soil weight at all depth intervals and that the rate of increase in the drill string weight is greater than the rate of increase of soil weight with depth. So as we drill deeper, we exceed the overburden Figure 1. Comparison of Soil Weight with Weight and Stress of Drill String.


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 247or raveling of soils, sediments, or rock materials into subterranean voids created by the effect of water on a limestone or similar rock formation. (Florida Statute 627.706) Figure 2 provides a further explanation of the statue. From Figure 2, it is seen that two conditions must be present: dissolution of the limestone and the overburden (supporting material) must be affected for sinkhole activity to exist (see Steps 1 & 2 in of the words: earth supporting the covered building implies that the building must be damaged in the area where the soil has been weakened. Therefore, it is structure related to systematic weakening of the soil, separate from damage related to poor construction and maintenance to declare a sinkhole is present. The determination of the cause of building damage requires a thorough forensic investigation of soil conditions and, in particular, structural conditions to distinguish between damage from sinkhole activity verses damage from design, construction and maintenance consistency of the material. Consider if there is a gradual fall of the rods, one may conclude that a zone of soft clay or sand is present, depending on the material in the cuttings found in the wash water. The point is that many accurately provide a picture of what occurred during the drilling of the boring and consultants simply conclude the worst in the absence of this information. We must also consider the conditions that can occur in some of the soft soils that are commonly susceptible to remolding from the removal and insertion of the drill string. Rapid movement of the drill string can cause extreme changes in the state of stress at the sampling depth resulting in further disturbance and consequent loss in soil strength. Figure 2. Steps in Determining if Sinkhole Activity has Occurred According to .706.Florida Statute Requirements for a SinkholeThe Florida statute in .706 has established that sinkhole activity is present when: settlement or systematic weakening of the earth supporting the covered building only if the settlement or systematic weakening results from contemporaneous movement


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE48Geologic Conditions central Florida, we must not lose sight that, for the most part, we are analyzing Coastal Plain sediments deposited in diverse shallow marine environments. The geology of Florida is composed of strata formed during three During this time sands containing varying amounts of silt and clay were deposited on the bottom of shallow seas that existed during interglacial time when sea levels were higher than present (Kuhns, et al, 1987). During this time the great expanses of limestone that underlie most of the State of Florida were formed in these shallow seas. Most of the limestones contain impurities that resulted from depositional conditions during the formation of the limestone in the shallow marine environment. For to erosion from streams and offshore currents that resulted in inclusions of sediments that now serve as pervious conduits that facilitate weathering. Moreover, the clastic components of the limestone mass vary, thus creating areas within the indurated mass that are more permeable, and therefore more prone to dissolution. An important factor in the discussion of sinkhole development is to consider the time required for the dissolution of limestone. The rate of limestone dissolution is from 5 to 200 mm per 1,000 years. For the climate in eastern U.S. and Western Europe, the rate is between 25 and 40 mm per 1,000 years (Sowers, 1996). Geotechnical Conditions Related to the OverburdenThe overburden covering the limestone may consist of transported or residual soils. In transported soils and has had the longest time to consolidate under the weight of the overlying soil. In residual soils overlying limestone, the opposite is generally true with the youngest soil occurring at the bottom of the section. In decreasing with increasing depth until at a short distance above the limestone surface the soil may become softer (Sowers). The lower SPT value may result from erosion of soil raveling into solution slots or discontinuities in the limestone, which results from depositional features. The progression of these zones is generally very slow measure in the statute as to what constitutes sinkhole activity; there is much left to interpretation. Therefore, becomes a very critical aspect of a sinkhole investigation.Boring LogsThe information contained in the boring logs for a site investigation is generally the most useful data developed at the site. Overall, when we consider that the boring logs cover less than 1% of the site area (the area sampled by four borings compared to the area under the structure methods is limited in depth of coverage, we then must place great emphasis on information from borings. In meters) below the ground surface. Although good radar penetration is achieved in dry sandy soils, the penetration conductivity is sometimes only a few centimeters. but its depth of penetration is limited to about 25% of the length of the traverse, which presents a problem with depth of penetration at many residential and commercial sites with limited property. The marginal amount of data that may be obtained by geophysical methods places additional emphasis on developing complete information in the boring logs. Because of the complexity of subsurface conditions in karst terrains, we must carefully analyze subsurface conditions and not oversimplify them by only using contain a complete description of the circumstances under which these conditions occurred. The boring log should provide a record of not only the soil material found but also a detailed discussion of what occurred while sampling the soil and rock material. This information is typically absent from many effort must be placed into analyzing the origin of all building damage and relating this damage to potential subsurface conditions by considering the building as a giant test cell and analyzing building damage to explain its source relative to sinkhole causes or


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 249Another consideration in the evaluation of subsurface are present, is the investigator should perform an analysis of settlement at each boring location and determine the amount of settlement that will occur at each location. The magnitude of settlement determined at each boring location conditions on overall building performance during the past and future life of the structure. If the analysis of settlement at each boring location results in essentially the same magnitude of settlement, this becomes a compelling factor are not at play such as building damage that results from Case Studies Case Study No. 1 consultant found sinkhole activity present while another concluded no sinkhole activity was present (the dashed From analysis of subsurface conditions shown in this concluded that sinkhole conditions are not present. The following summarizes the reasons for this conclusion: 1) no evidence of loss of circulation was found in the correlation can be made to locations of exterior distress there is no evidence of movement of soil or raveling of soil into voids created by effects of water on limestone therefore there is no effect on the overburden (see maintenance, 4) all borings generally show similar lithologic conditions, 5) loose material found in the to be expected in residual overburden soils as opposed occurs in transported soils, and 7) the site is located near the east coast of Florida in an area not known for sinkhole activity. Case Study No. 2 central Florida. No sinkhole activity was found at the site. This conclusion was based upon several factors as follows: and their presence is normally investigated by the use of geophysical methods or may be detected by trends in the SPT borings drilled for the investigation. Because of when clayey soils are present, it is recommended that used in the sinkhole investigation. is used inside the building to determine if settlement has occurred under the building slab. ERI traverses can however, ERI is limited to a depth of approximately 25% of the traverse length. In many residential and commercial sites where property is limited this presents Sinkhole DeterminationIn analyzing subsurface conditions, there are a number drill string excursions, loss of circulation during limestone, relief of the limestone surface, associate conditions coupled with an understanding of the site geology need to be present before sinkhole activity can be established. From what has been discussed, it is seen how easily the investigation can be influenced by the results of may result from a boring intercepting a raveling zone at a depth substantially below the ground surface where this zone will not affect the structure in the foreseeable future. Depending on site condition it is common for some to ignore features occurring at meters) depending on the damage in the building presence of distress in the home and determine if there is a correlation between distress in the building and subsurface conditions or if distress is related to structural deficiencies. For this reason it is important where distress is found in the structure.


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE50 Figure 3. Case study No. 1a Site Near the East Coast of Florida (red indicates N-values less than or equal to 4, depth in feet).Figure 4. Case study No. 2a Site in West-Central Florida (red indicates N-values less than or equal to 4, depth in feet).


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 251mechanisms causing damage and determining if this damage can be caused by subsurface conditions. 5. Explain the origin of all distress found in the building. This may require an evaluation of the structural integrity of roof trusses, structural connections and modeling all distress to determine the overall building movement. 6. The Florida Sinkhole Statute requires that overburden material supporting the structure should be weakened or settled as a result of movement of the soil into pervious conduits in the limestone. 7. An analysis of the potential settlement that may occur at each boring location should be performed to determine if differential settlement can occur from the conditions determined in the investigation. Since borings may not be located in the exact areas of building damage, all assumptions are reasonable.References Subsurface indicators of potential sinkhole activity at In: Beck BF, Wilson, WL, editors. Proceedings of the Second Multidisciplinary Conference on Sinkholes and construction of foundations in karst terrain. ASCE Press. The 2012 Florida Statutes. .706 Sinkhole insurance; Proceedings of the Ninth Multidisciplinary Conference on Sinkholes and Engineering and In: Beck BF, editor. Proceedings of the Ninth Multidisciplinary Conference on Sinkholes and Engineering and Environmental Impact of Karst; 75 feet (22.9 meters) bls (below land surface), is associated with localized weathering or depositional conditions often found in this region and has had no effect on overburden conditions, 2) the soft zone in common occurrence in karst areas, and is considered to be related to localized increases in permeability associated with fractures and depositional features at or near the limestone surface, 5) the 55 foot (16.8 meter) difference in the depth to rock across the property is not uncommon in karst terrains and is not necessarily associated with sinkhole activity, 6) there is no focus to the damage found in the home and all damage appears to result from construction issues, and 7) there is no evidence of movement of soil or raveling of soil into voids created by the effects of water on limestone therefore there is no effect on the overburden (see Figure 2).Conclusions and RecommendationsThe following is a summary of some of the conclusions made in this paper: 1. It is misleading to consider that the occurrence of stresses imposed at the tip of the sample spoon are higher than insitu conditions, one must conclude that soil material at the bottom of the drill string has at least enough strength to support the weight does not represent a void. 2. WR conditions may or may not represent a void depending on the speed with which the rods fall. If the drill undergoes a slow gradual drop, one may be compelled to consider that there is some material at the bottom of the hole that can partially fall of the rods is found than one can conclude that void may be present. More information should be placed on the boring logs, in particular, a record of the rate of fall of present. 4. Determine if a correlation is present between the location of building damage and location of subsurface conditions. A very important part of a sinkhole investigation is determining the


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE52 a sinkhole? In: Beck BF, editor. Proceedings of the Tenth Multidisciplinary Conference on Sinkholes and Engineering and Environmental Impact of Karst; Zisman ED. 2010. The use of forensic engineering 2010: Advances in Analysis, Modeling & Design. ASCE Press.


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 253 remediation and to help in choosing a suitable contractor. water, mason sand and bentonite, in varying proportions, delivered under varying pressures, and using two different grout mixing methods was deemed the most appropriate cooperation between the grouting engineers, the grout crew, and the college administration and maintenance personnel provided useful insight and support. The various procedures used and the bases for their use are discussed in this paper. A total of 41 probe holes were was placed. Voids as large as 5 m (18 feet) in vertical probe hole. Subsurface connection between probe holes (10 feet) laterally.Introduction performing arts center, student cafeteria, radio station water from the structures and additional parking. The college hired a development company that had previously managed construction at the school to spearhead the available data, the geotechnical engineers employed for the planned construction did not seem to understand the atop karst terrane.EXPLORATORY GROUTING OF A SUBSURFACE DETENTION/INFILTRATION SYSTEMJoseph A. Fischer, Joseph J. FischerGeoscience Services, 1741 Route 31, Clinton, New Jersey 08809, USA, geoserv@hotmail.comTodd K. Miller400 Jefferson St., Hackettstown, New Jersey 07840, USA, millert02@centenarycollege.eduMichael J. MiluskiCompaction Grouting Services, P.O. Box 55, Wallingford, Pennsylvania 19086, USA, mmiluski@cgsinc.netAbstractIn 2007, a geotechnical investigation was performed for a student center at a New Jersey college. Even after reviewing the results of that study, the Municipality student center. Dolomite. The Allentown likely dips shallowly below the Beekmantown. This local suite of carbonate bedrock Ordovician rocks. parking area and remediation was attempted by others. placed in subsurface voids. Shortly thereafter, more sinkholes opened, some within the area remediated. Technical problems at the site included a lack of reliable subsurface information; the basin functioning need to vary the grout and delivery procedures based upon encountered conditions and probe hole locations in relation to the basin; the need to remediate solution features trending beyond the original area of interest; and the possibility of unrecognized solution features outside the area of interest and below the student center. These potential problems were brought to the attention of the current college administration. They quickly recognized the concerns and requested a different


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE54stringers at base of sequence. Lower contact placed at top with its many variations in structure, material types and tectonic history is of great value to geotechnical consultants. In this instance, comparing good test boring Beekmantown Formation and Allentown Dolomite would have allowed a better understanding of the site subsurface. The basin site is mapped as being very close to a formational contact with the Allentown Dolomite, which likely dips below the site at a relatively shallow depth. In our experience, the Lower Beekmantown and encountered in New Jersey and Pennsylvania. dissolution in this region is generally related to stress conditions and resultant fracturing. The southeasterly as solutioning varies with differences in the bedrock constituents affecting cavity formation along fracture trends as well as bedding. The Conclusions and Recommendation section of the 2007 geotechnical report starts by stating Neither the borings Additionally, the municipal engineer exacerbated by 12.2 m (120 by 40 feet) in plan dimension and the existing subgrade and the system; typical construction for such systems in the northeastern U.S. The parking lot delivery trucks. After one year of use, sinkholes formed within the continued to grow in size and number despite repairs, the construction contractors removed approximately a structures (installed to prevent debris from compromising to remediate the sinkholes affecting the basin area and the system was reinstalled and the parking lot replaced.Geology (Figure 1). The Lower Beekmantown Formation is The limestone beds grade laterally and down section into Figure 1. Bedrock geology map.


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 255 lean concrete. In the authors experience, conventional throats with a graded rock porous plug. This alternate which requires prior approval from the NJDEP, which has not been granted in any such proposal to our knowledge. consultant and general contractor recommended that the areas of concern be excavated for exploration under their technical supervision, resulting in a hole some 15.2 m Before the next phase of the remediation was initiated, two more sinkholes opened. A combination of graded mobility (compaction) grouting by a specialty contractor was attempted to complete the remediation of four areas of concern (Figure 2). A total of six grout holes were planned and ten were actually drilled. The total amount nor our observations revealed any evidence of solutioning, subsidence, sinkhole or other karst topographic features of the drilling logs indicate that of the 20 borings drilled deeper than 2.4 m (8 feet) below grade, 18 showed some soils atop the bedrock surface, variations in rock depth over short horizontal distances, open fractures and seams, and the redirection of the drill string from pinnacles.Stormwater Detention/Infiltration of sediment removal is generally considered mandatory (with some exceptions) in New Jersey. Originally, the design proposed a surface detention basin, presumably groundwater regime with an equivalent amount of precipitation that would be lost to impermeable cover (a New Jersey Department of Environmental Protection [NJDEP] requirement for new construction). During the municipal review process, the Planning Board advised recommendations for a subsurface stormwater disposal system without noting any concerns for the carbonate bedrock below the basin area. basins had been installed throughout the campus during earlier construction; their age evident by their brick and mortar construction. So essentially, the college went ahead with the various engineering recommendations without any warning from their professionals as to the problems that could exist as a result of the karstic subsurface.Sinkhole Occurrence and RemediationsDepressions and two sinkholes began to form in and near the parking lot surface in the fall of 2010 (Figure to punctured and cracked. The next step was to excavate the northwestern corner of the system. After inspecting the excavated area, one of the solutions offered Figure 2. Sinkholes in the subsurface stormwater system.


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE56ability to mix on site were important in selecting the Grouting ConceptsIn consideration of the potential problems extant at investigation and remediation program (e.g., Fischer and grouted. This work was completed in December of 2010. The authors were contacted in the fall of 2011 when additional sinkholes started to form in the parking lot 6). After discussions with the client and reviewing the available data for the stormwater system (which included a report from college maintenance personnel that the subsurface stormwater system had never detained water, even subsequent to large precipitation events), a Request for Proposal was prepared and sent to three prospective bidders, including the grouting contractor (who declined to bid). The other two contractors contacted provided closely competitive proposals, but previous history and the Figure 3. Exposed rock and sinkhole throat at bottom of stormwater system.Figure 4. Reinstallation of stormwater system (note graded rock in sinkhole at bottom right of photo).Figure 5. Parking lot sinkhole adjacent to stormwater system.Figure 6. Parking lot sinkhole adjacent to stormwater system.


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 257 cement pump (Figures 8 and 9). angled exploratory probes (so as to reach areas below the system without compromising the existing system) that would be logged by experienced geotechnical personnel. Alternative drilling methods were invited in economy and expediency, the grout holes were advanced in karst grouting. Mixers and pumps had to be able to handle a range of expected grout blends and viscosities, including the provision for a setting agent, which could change from grouting technician in charge. Potential ground heave was closely monitored during the grouting operations. variety of conditions expected during drilling within and drilling and grouting program. The lack of useful and reliable subsurface information increased the original from the system and was initially drilled to isolate the stormwater system for remediation.Grouting OperationsThe stabilization program began near the subsidence the aforementioned system. These holes were either tremie 20 pounds per square inch [psi]). The grouting began using 7) in an attempt to seal off small passages leading to the system. This system used a colloidal mixer, agitation tank was used except at select locations where drilling air losses Figure 7. ChemGrout CG 600 3X8DH.Figure 8. 10 cubic yard mobile site mixer with Putzmeister TK 15 HP pump at rear. Figure 9. Installing casing into the grout hole.


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE58Site Subsurface ConditionsAs should be expected in any grouting operation, drilling and grouting operations. The exploratory drilling likely controlled by fracturing roughly perpendicular to the general geologic strike of the region. Another solution feature seemed to parallel one edge (long axis) of the system. Although no linear pattern of sinkhole formation was evident by reconnaissance, these features became evident through exploratory drilling by cavities at varying depths and a generally deeper bedrock surface, The most problematic of these solution features trended through a corner of the system and into the area of the student center, in line with one of the borings performed prior to construction where concerns were noted during the data review. This feature was followed well outside the stormwater system in an effort to preclude further collapse in the parking lot and loading ramp areas. One other aforementioned feature appeared to be below the system, parallel to its long axis. This feature was grouted using angle holes drilled from outside the system at about a 10 degree angle so as to penetrate below the mobility grout was then placed only to the depth of the bottom of the system. During the operations, two solution features indicated by drilling and grouting intersected near the northeasterly corner of the system where the largest grout takes were experienced. This area evidenced extensive grout hole connection, mostly through drilling air exiting another evidenced drilling air connection through two probe holes bypassing another almost directly in the middle. was noted, but was far less prevalent than the drilling air connection. Another feature appeared to be related to a stormwater inlet and pipe some 12.2 m (40 feet) from the system and 22.9 m (75 feet) from the closest area of concern. area was grouted through two holes bracketing the basin rock was to be penetrated prior to terminating drilling. encountered installing the grout pipe to the bottom of the drilled hole due to ledges of rock and poor quality was to seal the causal throats of the new sinkholes isolate it from potential areas of concern. As the work progressed (in heavy rains), a lengthy crack appeared to open in the central portion of the previously repaired system and parking lot requiring a revision to the planned program. One unexpected problem with the grouting operations did arise as a result of the unusual subsurface pressures (measured at the grout hole head) were the removal of the grout from the system by a bucket helped stabilize a small section of the system with connection to a bedrock cavity. At the location of the aforementioned crack that appeared during initial grouting operations, the centerline between the two closest, linear chamber runs of the system was marked out. An attempt was then with a bit that can be extracted through the installed casing. The idea was to grout below the stormwater the system; then removing the remaining casing while surface and that drilling effort was abandoned before the bottom of the basin was reached.


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 259have revealed themselves in the vicinity of the subsurface More than 245 m (800 feet) away, however, a sinkhole appeared at a combination catch basin and dry well located in an older portion of the campus underlain by the Allentown Dolomite. As important infrastructure ReferencesFischer JA, Fischer JJ. 1995. Karst site remediation Engineering and Environmental Problems in Karst Terrane. Proceedings of the Fifth Multidisciplinary Conference on Sinkholes and the Engineering and and topographic base maps of New Jersey, in one grout hole, which then appeared within the other.Summary and ConclusionsNew Jersey regulations and space concerns are making (with some form of preliminary treatment) more ground stormwater detention, karst concerns have been accepted by some Municipal and State regulatory groups portion of the system. Thus, impermeable liners and personnel have been more commonly recommended at sites underlain by carbonate bedrock. As a result of the sinkhole problems in the stormwater well basins installed during previous construction were being eliminated by sealing the bottoms with concrete. The exploration and remediation work for this effort by a number of groups and individuals that has karst site. College administrative and maintenance personnel provided information and assistance that grouting program directed by experienced geotechnical personnel was implemented through a competent and cooperative grouting contractor and experienced crew. The various combinations of vertical and angled ability to vary the grout mix upon short notice was invaluable considering the highly variable conditions below the site. Additionally, the system appears to detain water after precipitation events as a result of the remediations described herein, yet still effectively deemed complete without full knowledge of the extent of solutioning in the area of concern and the ability of any grouting concept to eliminate all future problems.




13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 261issues regarding the investigation, characterization, design, and construction of engineered facilities in karst geologic settings. The authors recognize that these technical issues impact all engineered facilities, not other agencies, owners, and consultants who face similar challenges. The participants at this conference likely have In this presentation, the authors will actively engage the participants and will request input based of their experience and expertise.IntroductionIt is often said that we can only be certain of two things professionals can safely add three more relative certainties: (i) as a society we continue to generate large amounts of garbage (i.e., MSW) that require safe and (iii) geohazards that restrict the location of these performance, there is a reticence of the populace to view this as a societal need and prefer that the problem be shifted to others at other locations. Regarding geohazards Florida, which is one of the most populated states in the country. Across Florida, and particularly in Central Florida where the karst is prevalent and the population societal need runs headlong into geologic constraints. NEED FOR A STANDARDIZED APPROACH TO LANDFILLS IN KARST GEOLOGIC SETTINGSRobert C. Bachus, Ph.D., P.E., D.GE.Geosyntec Consultants, 1255 Roberts Blvd. NW, Ste. 200, Kennesaw, Georgia 30144 USA, rbachus@geosyntec.comRichard B. Tedder, P.E.Florida Dept. of Environmental Protection, 2600 Blair Stone Rd., MS 4565, Tallahassee, Florida 32399 USA, AbstractThe challenges presented by geohazards play a facilities, particularly those situated in karst geologic must make decisions regarding the siting and permitting of these facilities. While these decisions are based on their information, their decisions are often scrutinized by often (and usually) have contrasting interpretations of the same site characterization information. The Florida Department of Environmental Protection (FDEP) has initiated an innovative approach to help the agency in in the investigation, characterization, permitting, and construction of engineered facilities in karst settings. Through a process involving the compilation and permitees, consultants, and the agency in developing investigation, characterization, design, construction, operations, and monitoring strategies for facilities overlying karst geologic conditions. The activities of who have also been requested by FDEP to participate in provide information to and then solicit information from the conference participants (and readers). The approach


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE62In anticipation of the collision course, the Florida Department of Environmental Protection (FDEP) has taken a proactive course of action to develop technically rigorous recommendations regarding the siting, permitting, design, construction, operations, and monitoring of MSW facilities in the State that need to be located over karst terrain. This paper will identify geotechnical, water resources, and geoenvironmental professionals who must deal with the often competing demands placed by society in dealing with the disposal of MSW and the locations of the disposal facilities. proactive strategy for addressing the waste disposal problems caused by the challenging geologic conditions, in Florida. Finally, the authors will solicit opinions and experiences from the participants of the conference regarding improvements to this initiative, recognizing that do nothing or take the waste elsewhere is not a sustainable alternative.Before a strategy can be developed, a sense for the magnitude of the problem needs to be recognized. In Florida (as well as in many parts of the country), the problem prevalence of karst in the underlying geologic formations. A brief summary of these problems follows.Regarding solid waste practices and experiences, Florida follows many of the trends evident across the country. Figure 1 shows the reality of solid waste generation in Florida over the past 20 years. The downward trend since 2005 is a combination of and on the recent economic conditions in the U.S. If these trends are compared to national trends and coupled with the population, results indicate that in Florida, the waste generation can be represented as approximately national average of 2.0 kg (4.4 pounds) per person per day. Consistent with national trends, prosperity leads to an increase in MSW generation per person. When these trends are coupled with the future estimated population growth in Florida (Figure 2), the impact of population growth on solid waste disposal needs is staggering. Interestingly, the Florida population growth trend of about 250,000 people per year (ppy) is approximately of 2,500,000 ppy (FAIR, 2006). Clearly, the popularity of the 4th to increase over the next several generations. As can be seen in Figure 1, it would require an extreme paradigm shift in public policy, public response, and waste disposal disposal needs. To further demonstrate the MSW disposal issues facing Florida, consider the locations in Florida where people density across the State. People clearly like to live in Central Florida. Finally, over the past several years, most states have seen an overall reduction in the number of solid waste disposal facilities. This is demonstrated in Figure 4, which reports the number of active MSW disposal facilities across the country. The national trend over the past 20 years clearly shows that the number of facilities Figure 1. Solid Waste Disposal Trends in Florida (FDEP, 2012, written communication).Figure 2. Florida Population Projections (FL EDR, 2011).


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 263 MSW disposal facilities per state. Currently Florida has is Where do Floridians place waste in the future and how much capacity is needed?Karst Geohazards in FloridaKarst and the underlying problems associated with the geologic conditions are well known to most Floridians, University Of South Florida in Tampa. Perhaps the most famous (infamous) is the May 1981 Winter Park Sinkhole Although detailed formal historical records may be infrequent, and published records, primarily to assess the impacts of subsidence and sinkholes on groundwater resources. Figure present locations of reported subsidence. Florida (i.e., Southwest District and Central District) of subsidence. When the Northeast District is added to this list, the locations of nearly 95 percent of the reported episodes are included. Independent records maintained verbally provided to the authors provide nearly identical results. Clearly, the problems of subsidence and sinkholes around the State to develop Florida Aquifer Vulnerability Floridan Aquifer is presented in Figure 6.Figure 3. Florida Population Density (FL EDR, 2011).Figure 4. MSW Landfills in the U.S. (USEPA, 2009).Figure 5. Map Showing Reported Subsidence Areas (FDEP, 2010, written communication). Figure 6. Florida Aquifer Vulnerability Assessment (FAVA) Map for the Floridan Aquifer (FDEP, 2010, written communication).


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE64 to the groundwater table; (ii) hydraulic head difference distance to known karst features; (v) overburden soil permeability; and (vi) aquifer system overburden. Comparing Figures 5 and 6 provides the compelling observation that the most valuable groundwater resource in the State is most vulnerable in the areas where virtually 95 percent of the reported subsidence is located. and hydrogeologic conditions with the previous section regarding solid waste needs, a foreboding observation develops. It is anticipated that the areas where the in the future. Further, this area is where the potential for subsidence and sinkholes is highest (Figure 5) and where the Floridan Aquifer is most vulnerable (Figure 6). Furthermore, it is noted that the areas of subsidence and aquifer vulnerability, hereinafter referenced as sensitive areas, comprise nearly 60 percent of the total land area in the State. Clearly, a hasty reaction to simply would place a hardship on other areas of the State where residence of Central Florida due to high transportation challenges across the State. Technical differences of opinions are inevitable between environmental groups, consistent, defensible, and fair solid waste policies and guidelines are developed and enforced. Guidance DocumentsThe FDEP has developed and currently maintains and enforces solid waste regulations in the State that exceed the U.S. Environmental Protection Agency (USEPA) Subtitle D requirements regarding the siting, design, construction, operations, and performance of MSW disposal facilities. The FDEP has followed USEPA guidelines and like other states that experience karst geologic conditions (including Alabama, Arkansas, Kentucky, Minnesota, Pennsylvania, and Tennessee), has taken aggressive regulatory positions regarding the protection of groundwater resources by establishing: (i) guidelines. With regards to groundwater monitoring requirements, these State regulations acknowledge that the groundwater regime in karst geologic settings comprised of granular media. The FDEP has taken a strong position that its policies are directed to protecting groundwater and minimizing potential adverse risks to its aquifer systems. Therefore, the FDEP regulations explicitly recognize the importance for the applicant to regime and develop a groundwater monitoring system FDEP Rules and Regulations regime, karst can also impact the structural stability the regulations in most other states) address issues of of the Florida Administrative Code (FAC) are cited to provide examples of how regulations (and regulators) by authors for emphasis): prohibition for siting requirements for all solid waste disposal facilities states unless authorized by a another Department rule or a Department license geological, design, or operational features, no person shall store or dispose of solid waste. in an area where geological formations or other subsurface features will not provide support for the solid waste; location requirements site shall provide structural support for the facility including total wastes to be disposed of and structures to be built on the site; design requirements soil and geosynthetic liners shall be installed upon a base and in a geologic setting capable


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 265 test for reasonable assurance. The permit application and and scrutiny. In many cases the interpretations of the geotechnical investigation and geologic characterization an independent assessment by the public reviewers regarding regulatory compliance and reasonable assurance. In addition, particularly for permits involving professionals. This often leaves the FDEP in the professionals and the reality that regardless of its decision and potential litigation from either the applicant or the public. The FDEP has successfully faced the realities of this regulatory environment since the promulgation topics where controversy or technical challenges are anticipated, FDEP (and regulators in other states) have expectations regarding the permitting process. applicant and public response regarding recent MSW In addition, FDEP recognizes future challenges facing this region as summarized in the previous section of this paper. To address these issues, the next section describes a proactive approach that FDEP has taken regarding the siting, permitting, design, construction, operation, and monitoring of MSW disposal facilities located in karst geologic settings.Development of a Technical Advisory Group (TAG)To assist the agency in this initiative, the FDEP has comprised of a number of engineers, geologists, and scientists from both the public and private sectors with expertise in karst assessment to help the agency in the development of additional technical guidance. This guidance will assist: (i) the applicant in its preparation of MSW permit applications; (ii) the FDEP personnel responsible for technical review of the permit application of providing structural support to prevent overstressing of the liner due to settlements and applied stresses; and geotechnical site investigation requirements and construction and demolition (C&D) debris disposal facilities states the prior to any that are necessary for the design, construction, the facility and shallexplore and address the ground, lineaments, and sinkholes. applicant to make the requisite demonstrations and example, with regards to geotechnical site investigation requirements above, one engineer may believe that the site can be adequately characterized using 1 boring per hectare (2.4 borings per acre), while another may believe that variability at the site warrants a density of greater than 4 borings per hectare (10 borings per acre). In many cases, the regulations leave decisions to the discretion of the professionals tasked with preparing the permit application. The FDEP, however, recognizes that even and geologic characterization studies require the educated interpretation of data and facts. This interpretation of these study results must be provided in an application prepared on behalf of the applicant that demonstrates compliance with the FDEP regulations. To issue an FDEP permit for an MSW facility, the applicant must provide reasonable importantly states that reasonable assurance means the existence of a substantial likelihood, although not an absolute guarantee, that the proposed activity and applicant will comply with agency rules, laws, orders and permit conditions. It does not mean proof that a facility will not fail. It is noted that this section of the regulations recognizes that the permitting test is for reasonable assurance not for absolute assurance. other states) requires that professional engineers and geologists prepare technical applications that provide the previously stated reasonable assurance. The


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE66determining that stabilization was achieved; and (v) monitoring a disposal facility for sinkhole formation. A brief discussion of the approach used to address each of these tasks and preliminary recommendations by the Characterizing Site for Sinkhole Potential characterize the potential site. At a minimum, this task includes: (i) review of geologic information regarding mile) radius of the site; (ii) review of historical aerial spanning several years (or decades when possible) followed by physical inspection of the site with photos in hand; (ii) geophysical investigation along several transects, including orthogonal transects that intersect soundings; and (iv) physical invasive investigation, sampling, and in situ testing. This strategy recognizes that the potential for sinkhole development starts at should be cited and details of the features should be included in the permit application. With regards to the geophysical testing, electrical resistivity and ground that have been used successfully in Florida. Other techniques will be considered. It is critically important sounding locations. Invasive testing can include hollow stem auger or mud rotary drilling, with the latter being preferred due to the ability to note rod drop and slurry loss. Soil samples and rock cores should be collected. In situ testing can include the Standard Penetration Test is currently considering the recommended minimum number of geophysical transects, the depth and extent of as well as the recommended laboratory tests. The from the geological and aerial photograph review. Assessing Sinkhole Potential Risks assessment of the risk of a sinkhole developing at the proposed MSW disposal site. The FDEP would like to verify compliance and reasonable assurance; and (iii) the public in its review and critique of the permit FDEP in the development of technical guidance for the siting, permitting, design, construction, operation, and monitoring of MSW disposal facilities sited in that will help: (i) the FDEP decide how to evaluate these permit applications and then issue the solid waste disposal permits; and (ii) the applicant know what information should be submitted in these permit acknowledged that in developing the guidance, there needs to be a balance between cost of assessment and investigation and the risk of failure. Furthermore, the guidance needs to apply both good science and group of professionals, FDEP required that members set aside personal interests, if any exist, and focus on what is really good for Florida.Specific Objectives of the TAG regulations was protection of groundwater resources. validated the intention of the Subtitle D regulations. (2002) demonstrated that the composite liner system design and the leachate management system design and operations requirements promulgated by the Subtitle D protective of groundwater. As mentioned previously, the challenge in the geologic setting in Central Florida is to assure the structural integrity of the liner system. physical and geophysical techniques for characterizing sinkhole potential of a site; (ii) determining if potential sinkhole risks for a site are low, moderate, or high; (iii) deciding when a site cannot be used or can be used if properly stabilized; (iv) stabilizing a site and


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 267 and replacement. The stabilization efforts will require that the applicant demonstrate the effectiveness of the selected stabilization remedy. With reference to Figure overburden soil followed by treatment of the foundation bedrock. One aspect of stabilization that concerns the in which the plugging of one hole in the dike simply caused a new hole to form. Stabilization alternatives will need to consider site wide stabilization efforts or at least the impacts of localized stabilization efforts on overall site stability. Monitoring for Sinkhole Formation The FDEP acknowledges that the construction of a the design of surface water management ponds, (i.e., unloading) the foundation soils can increase the the assessment to report a high, medium, or low information regarding the potential size of the sinkhole, location and size of the reported subsidence features so that regional lessons can be reported based on past low risk. Evaluating Site Suitability assessing risk should a sinkhole develop) is to develop MSW disposal facility. Although in its preliminary state, the previous step and the assessment of site suitability. that all sites are potentially suitable for development engineering control. The FDEP does not necessarily want to condemn a site a priori, but clearly wants to make the applicant aware that certain geologic conditions will render a site essentially unsuitable due to the likelihood of sinkhole development and the system. Figure 7 provides an example of a potentially unsuitable site. This aerial map, when combined with historical photos from the previous 20 years, showed a gradual and steady development of large sinkholes that extend to the ground surface and grow over time. For most sites (and in particular this site), it is important to mechanism to assess whether it is economical to arrest future sinkhole development or better to simply abandon the site. Defining Site Stabilization Measures that may be required to improve the suitability of the site to a level that provides reasonable assurance to the FDEP that the site can be developed in compliance site conditions, techniques may include (but are not Figure 7. Example of a Potentially Unsuitable Site.


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE68 ConclusionThe FDEP has developed and currently maintains and enforces solid waste regulations in the State that exceed the national standards but desires to improve the MSW currently the 4th most populated State and Floridians generate solid waste at a rate that exceeds the national Central and Northeast Florida comprise nearly 60 percent of the total land area in the State and is founded subsidence due to sinkholes. These same areas are within zones where the valuable groundwater resources are considered most vulnerable and include areas of the highest population density. The FDEP has developed a the permit applicants provide the FDEP a reasonable assurance that the siting, design, construction, operations, and monitoring of the proposed facility is in compliance with FDEP regulations. The authors solicit feedback from conference participants (and proceedings readers) regarding techniques to improve the strategies ReferencesAlabama Department of Environmental Management: Land division Solid waste program, Division Environmental Management [ADEM]; [cited 2012 Nov 24]. 179 p. Available from: Arkansas Pollution Control and Ecology Commission: Regulation 22, Solid waste management rules [Internet]. Arkansas Department of Environmental Quality [ADEQ]; [cited 2012 Nov 24]. 202 p. Available fromt: Bonaparte R, Daniel DE, Koerner RM. 2002. Assessment and recommendations for improving the performance of waste containment systems. [Internet]. Florida Department of Environmental Protection [FDEP]; [cited 2012 Nov 24]. 114 p. Available from: that there will be recommendations for monitoring the site, as well as the surrounding parcels of land, for early indications of new sinkhole formation. Unfortunately, the solid waste itself decomposes over time resulting These recommendations will include provisions by the applicant for modifying operations and addressing these features should they occur. recommendations and minimum expectations regarding exploration and investigation programs that are based applicants. Feedback and Recommendations from Karst ExpertsThe purpose of this paper was to describe a strategy currently being implemented by the FDEP to karst geologic settings. Several of the charges to geologic phenomenon. The authors recognize that the participants at this conference (and readers of authors explicitly solicit feedback and suggestions recommendations regarding: (i) other similar efforts Florida that need to be considered; (iii) geophysical should be considered; (iv) stabilization options that have (or have not) worked effectively; and (v) considered. The authors recognize that the experience the experience of the participants will be useful in


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 269 Florida: demographics report 2011 [Internet]. Demographic Research [EDR]; [cited 2012 Nov 24]. 25 p. Available from: Kentucky legislature: Kentucky administrative regulations: Title 401: Energy and environment cabinet department for environmental protection; Kentucky Department for Environmental Protection [KDEP]. [cited 2012 Nov 24]. Available from: 2050: four immigration scenarios [Internet]. 2006. [Place of publication unknown]: Federation for American Immigration Reform [FAIR]; [cited 2012 Nov 24]. Available from: waste: [Internet]. Minnesota Pollution Control Agency [MPCA]; [cited 2012 Nov 24]. Available from: Tennessee Department of Environment and Conservation: Rules of Tennessee Department of Environment and Conservation: Solid Solid waste processing and disposal [Internet]. Tennessee Department of Environment and Conservation [TDEC]; [cited 2102 Nov 24]. Available from: solid waste [Internet]. Pennsylvania Department of Environmental Protection [PDEP]; cited 2012 Nov 24]. Available from: USEPA. 2009. Municipal solid waste in the United




13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 271 The most effective means to avoid geological disasters is prevention. Therefore, monitoring and early warning of karst collapse are particularly important. Current water or air pressure changes in underground karst collapse abnormalities, but due to its strict working environment, limited detection depth, professional monitoring TDR technique has many advantages, such TDR cannot be used to monitor the formation process of karst collapse because it receives only the signal from force, tension or both combined. Monitoring water and air pressure changes in underground karst system can only forecast the collapse risk of the karst fracture around the monitoring points. But it cannot point out Therefore, traditional monitoring methods cannot meet the demand for sinkhole collapses monitoring or system, which can detect temporal and spatial changes continuous basis (Tang et al., 2006). Nevertheless, there are still many problems in the application (Jiang et al., 2006; Li et al., 2005; Meng et al., 2011). According to the deformation characteristics obtained from sinkhole collapse modeling and calibration testing, we analyzed the mechanic relation between the soil and sensing Abstract on Brillouin scattering. This technique may potentially become a useful tool to monitor and predict karst collapse, especially for linear infrastructure such as roads, highways, and railways. This paper introduces a calibration device which is used to establish the characteristics of the sinkhole collapse, the mechanical design. The experimental tests are carried out through the designed equipment to investigate the effect of the sinkhole's size and the overburden stratum's thickness on process was stimulated with the orderly changes in load analyzed. It shows from the experiment results that the strain change in the sinkhole formation process sensing technology to sinkhole forming and evolving, it can be adopted in the monitoring for potential sinkhole collapse.PrefaceKarst is widely distributed in Southwest China Along of human activities, geological disasters related to karst have become prominent, especially karst collapse A CALIBRATION TEST OF KARST COLLAPSE MONITORING DEVICE BY OPTICAL TIME DOMAIN REFLECTOMETRY(BOTDR) TECHNIQUEGuan Zhende, Jiang Xiaozhen, Gao MingInstitute of Karst Geology, CAGS, Qixing 50 Rd., Guilin, GuangXi 130026 P.R. China,


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE72 et al., 2006) when the incident pulse wavelength is temperature changes. The experimental temperature variation is less than 5C, so the temperature effects were not considered. (OTDR) technique. According to the OTDR principle, the scattering position can be determined by measuring the scattered laser echo time. The distance between the (2 where: a pulse laser. According to Equation (1), the axial strain distribution Shi et al., 2005). According to Equation (2), the position where strain occurred can be calculated.The karst collapse monitoring modelThe working principle of BOTDR for collapse monitoring is based on the development of a karst soil void that manifests as deformation of the under the load coming from overlying stratum due to the development of a soil void. The location, scale and development of soil void can be well understood based on the analysis of temporal and spatial variation sinkhole collapse.Monitoring principle of optical fiber sensing technology on three spectroscopic analysis methods including Rayleigh scattering, Brillouin scattering and Raman scattering. Rayleigh scattering is an elastic scattering Brillouin and Raman scattering are nonelastic scattering 2006). Brillouin scattering arises from the interaction between optical and acoustic waves propagating in Brillouin scattering frequency and the temperature or temperature or axial strain can be calculated according to the amount of the frequency drift in the optical a temperature sensor is adopted to offset the drift by temperature change. The relationship between the center frequency drift and (1) where: Figure1. The principle of BOTDR. ( )


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 273 overburden soil; According to Equations (4) and (5): (6)Deformation compatibility of fiber and soilThe formation of soil void is the result of varied superimposing collapse factors, which causes overlying soil deformation or potential collapse. The key to the deformation. Reasonable distribution of the sensing is important during soil void monitoring. The placement features of karst collapse in the monitoring region. overburden stratum. During the development of collapse, deformation of the soil mass occurs gradually, and also deformation called compatible deformation. this interaction will be explained by mechanical analysis in the following discuss. (dT) can be demonstrated as below: where: Thus (Li et al.): (4 smaller than maximum static friction, so sliding friction is taken in the analysis. (5 where: = = N= .G = h Figure2. The mechanic relationship between soil and optical fiber. As explained in the theory mentioned above, when thickness and bulk density, stress is directly proportional happens while soil is deformed.Simplification of collapse monitoring modelAs soil void develops, incumbent soil load and void scale are critical to the magnitude and distribution of the stress around the developing void. According to the key monitoring factors and the deformation simulation experiment system was designed. During the formation of soil voids, the friction imposed on optical fiber at the edge of void and its influence area are changeable. Thus, optical fiber fixation should be considered in the model design (Liu et al., 2010; Yao et al., 2005). Intertwist, one of the fixation methods, is adopted which can express the way how friction varies with loads effectively


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE74Experiment process respectively. The positions were recorded by the labels at 1068.75m. The loading point deformation and strain loading and unloading. Test data processing and analysis According to its principle, the strain measured by the strain instrument is the integrated strain within 1m starting from the monitoring point (Wu et al., 2005; Yue et al., 2007). Taking the value got from connectivity test difference between the loading test value and the initial value. According to the strain change distribution as shown in be increased to obtain the enough friction The strain Experiment system with the load. This can be simulated by an experiment supporter, dial indicator and vertical loading system. And the formation of soil voids with 0.5~2.5m span under different loads can be simulated as well.Test and analysisTwo types of experiments, variable load in certain distance and variable distance under certain load out, respectively. produced by the ANDO corporation of Japan was used to measure the strain distribution in the optical fiber. The main index of the instrument is shown in Table 1.Test under stepwise variable load in certain distanceThe experiment simulates the load changes of the soil to analyze the changes of the axial strain and the optical Figure 3. Simplified sinkhole model. Technical Index Optional parameter Measure distance 1, 2, 5, 10, 20, 40, 80km Pulse width 10ns 20ns 50ns 100ns 200ns Dynamic range .004%(2s) 2dB 6dB 10dB 15dB 8dB 11dB Length resolution 1m 2m 5m 11m 22m Strain test accuracy .004%(2s) (.01%) (.005%) Strain test repeatability <0.04% <0.02%Table 1. AQ8603(BOTDR)Technical Index. In the experiment, .004%(2s) strain, 10cm sampling distance and 1m length resolution were adopted. Fiber connectivity was tested before the experiment.


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 275The experiment process experiment simulates the sinkhole span starting from 1 meter to 2.5 meter with 0.5 m interval. Test data processing and analysis strain, there is a mechanical analysis about the certain load experiment (Figure 8). (7) where: point; is suitable to be used for the soil bearing low pressure or unloaded. Following the soil void overburden collapse, the mass this process can be simulated by unloading experiment (Figures 6 and 7). Unloading experiment demonstrates that sinkhole collapse, and the position of the coverboard loading overburden stratum thickness of an incipient sinkhole, the to avoid the elastic modulus value exceeding the test range.Test under variable distance in certain load axial strain changes in different spans of the sinkhole by applying certain load.Figure 4. The loading point strain change (GFRP).Figure 5. The loading point strain change (Ordinary optical fiber).Figure 6. The strain change after unloading (GFRP).Figure 7. The strain change after unloading (Ordinary optical fiber).


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE76 and the load.ConclusionIn the process of soil void formation and subsequent sinkhole collapse, axial strain and deformation of the technology to monitor the location, size and collapsing process of the void in soil. effects on the strain value, which must be understand and choose appropriately. the deformation coordination between the soil and the response sensitivity. According to Figure 9, the maximal vertical displacement of the loading point is 55 mm and the minimum distance small deformation, the hypotenuse is approximately enough. (8) Not considering the material factor, the relationship the included angle is inverse proportion. For ordinary 1m to 2.5 m (Figure 10). Figure 8. Loading section stress analysisFigure 9. Vertical displacement in different sinkhole span.Figure 10. Strain change in different sinkhole span (Ordinary optical fiber).Figure 11. Strain change in different sinkhole span (GFRP).


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 277 to structure health monitoring. China Civil model, it still demonstrates that the strain characteristics useful tool for predicting sinkhole collapse.References of monitoring sinkhole collapse by using BOTDR Li K, Shi B, Tang C, et al. 2010. Feasibility research on soil deformation monitoring with distributed Li Y, Zhu P, Lei M, et al. 2005. Monitoring technique and methods of the karst collapses. Carsologica Liu J, Shi B, Zhang D, et al. 2006. Experimental study fixation technique of fiber based on BOTDR. Chinese Journal of Sensors and Actuators problem in monitoring and predicting sinkhole measurement for tunnel health diagnosis. China Journal of Rock Mechanics and Engineering Tang T, Zhu Y, CAI D, et al. 2006. Experimental sensing. Chinese Journal of Rock Mechanics and study on the measuring characteristics of BOTDR for structure health monitoring. China Civil




13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 279those triggered by water level declines, typically from ground water withdrawals (pumping), or those related to Florida, sinkhole formation and ground subsidence accompanying heavy ground water pumping are common occurrences and are typically associated with rapid declines in the potentiometric surface of the Floridan aquifer. This causes an increase in effective stress over formed by dissolution. Increased pumping can alter the downward raveling of overburden soils. Pumping can incipient raveled soil zones in the subsurface. Sinkhole formation can also be triggered by construction activities such as water impoundment in reservoirs and retention basins, ground loading, ground vibrations from heavy equipment, changes to natural drainage patterns by diversion of stormwater, and drilling of borings and water wells. The mechanisms activating sinkhole formation would include increased ground rocks associated with incipient sinkhole conditions. This paper presents a case history of a geophysical and geotechnical investigation conducted at a site in northern investigation was to evaluate impacts from installation karst subsidence feature(s), potential for induced ground subsidence, and impacts on building structures AbstractInduced sinkholes are a known geologic hazard and may be associated with construction activities that soils and rocks. This study describes the geophysical and geotechnical investigation of a site in northern installed utilizing horizontal directional drilling to evaluate the impacts of: 1) pipeline installation on bodies and building structures. The site was investigated utilizing geophysical testing (electrical resistivity), standard penetration test (SPT) borings, and ground vibration monitoring during pipeline construction. In the investigation, subsurface conditions indicative of possible preexisting weakened soil and rock materials associated with incipient raveled zones in overburden found in proximity to the pipeline corridor. During the vibrations occurred, along with formation of several two mechanisms of induced sinkhole formation: erosion of weak zones in overburden soils by the high pressure in limestone bedrock. In addition to current settlement impacts to the property, the investigation found a potential for future ground subsidence associated with undetected eroded and raveled zones that may in the future propagate to the land surface. IntroductionInduced sinkholes are caused or accelerated by human activities and are associated with two broad conditions: INDUCED SINKHOLE FORMATION ASSOCIATED WITH INSTALLATION OF A HIGH-PRESSURE NATURAL GAS Ted J. Smith, P.G., C.P.G.Earth & Water Resources, LLC, 2000 S. Florida Ave., Lakeland, Florida, 33803 USA, tjsmith@earth-waterllc.comGeorge C. Sinn, Jr., P.E.Central Florida Testing Laboratories, Inc., 12625 40th St., Clearwater, Florida, 33762 USA,


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE80clayey sediments and carbonate rocks of the Miocene landforms including sinkhole lakes, cypress domes, and broad wetland basins. Numerous lakes and swamps are were created by karst processes resulting in broad and water bodies near the property. The study methods included an electrical resistivity survey of the pipeline corridor, subsurface testing by standard penetration test (SPT) borings, and ground vibration monitoring during pipeline installation. Location and Geologic Setting approximately 20 acres in size and is bordered by and vegetated wetland areas on the northern side. The with ground elevations ranging from approximately 22 Lake to the south and east, and the fringing wetland areas on the north side. Two large residential structures and associated outbuildings are currently located on the property (Figure 2). Geology Karst Plain (Scott, 2005), which encompasses much of The geomorphic province is formed by a series of Pleistocene age marine terraces developed on sandy and Figure 1. Location Map, Hillsborough County, Florida.Figure 2. Geophysical Testing and Boring Location Plan showing pipeline easement, building structures, and testing locations.


13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 281are common sites of paleo or relic sinkhole activity and and Floridan aquifers.Karst and Sinkhole Development sediments. They also constitute a geologic hazard since greatly depending on the degree of consolidation of and conduits in the limestone formation, and the hydraulic connection with the deeper aquifers. Lakes and wetland and may serve as recharge areas to the Floridan Aquifer. Pipeline Construction horizontal directional borehole located within an property. The easement is approximately 15 meters wide to east alignment extending from the northwest corner Island Lake. The easement passes in close proximity to the main residential building and outbuildings (Figure 2). The borehole for the pipeline was completed using wetland basins and lakes formed by multiple coalescing sinkhole and subsidence features. Peninsular Florida is underlain by a thick sequence of Paleogene carbonate rocks that form the Florida platform, which is capped by a thin series of Miocene sediments. Important geologic and hydrogeologic units Table 1. Descriptions of each unit are taken from Arthur et al. (2008), Campbell (1984), and Scott, (1988).HydrogeologyTwo principal hydrogeologic units are present in the is hosted primarily by permeable sandy soils within Upper Floridan aquifer, which is hosted primarily in the Suwannee Limestone and deeper limestone formations (Ocala Limestone and Avon Park Formation). potentiometric surface corresponding with the water table, which occurs at depths of less than 1.5 meters (5 the Upper Floridan Aquifer system in the area occurs at permeable sandy sediments. Lakes and wetland basins System Series Lithostratigraphic Unit Lithology Hydrostratigraphic Unit Quaternary Undifferentiated Sands and Clays minor silty, clayey and organic soils, local deeper clayey soils and shell beds Pleistocene Tertiary Miocene Tampa Member: Arcadia Formation White to tan, quartz sandy, locally clayey, fossiliferous limestone and dolostone, per portion Upper Floridan Aquifer Oligocene Suwannee Limestone grainstoneTable 1. Regional Geologic and Hydrogeologic Units, Northern Hillsborough County.


NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE82 The modeled data were then analyzed with respect to identifying anomalous electrical resistivity signatures associated with potential subsurface karst features based on the following criteria: Columnar patterns of higher resistivity materials extending to the depth of apparent limestone bedrock. zones of soil raveling. Patterns indicative of localized increased depth of features. soils overlying moderate to low resistivity materials correlated with limestone occurring at approximate portion of the survey area located north and northwest of the main residential building. The anomalies were characterized by areas of higher resistivity materials at depth representing possible zones of downward raveling sands extending to limestone bedrock. The anomalies were subsequently tested by SPT borings Geological ConditionsStandard penetration test (SPT) borings were used to investigate subsurface conditions and geophysical anomalies within the pipeline corridor. The borings were conducted using conventional mud rotary drilling methods, in general accordance with ASTM performed along the alignment of the proposed gas pipeline and advanced to depths ranging from ground surface. The boring locations are shown on Figure 2. The subsurface geological and geotechnical data were used to construct a geologic profile of the pipeline corridor, depicted in Figure 4. Three generalized subsurface units or soil strata were encountered in the borings: the easement for the new pipeline. Although the construction details and depth of the older pipeline are not known, it was presumed that it was installed using concern at the site is an existing ground collapse feature located near the existing smaller gas pipeline (Figure 2). The collapse feature is oval in shape with dimensions of opinion that it is an induced sinkhole collapse associated with the installation of the existing gas pipeline. Minimal geotechnical data were collected by the pipeline contractor, consisting of two SPT borings drilled at each end of the pipeline alignment and advanced to depths of Electrical Resistivity SurveyAn electrical resistivity imaging (ERI) survey was conducted over accessible portions of the pipeline corridor and surrounding area to assist in characterizing the general subsurface geology and identify potential anomalous subsurface geological features for further investigation. The ERI survey consisted of four transect lines using up transects oriented approximately west to east were located within the pipeline corridor, with lengths of approximately 270 and 250 meters (885 ft. and 820 ft.). Additionally, two shorter transects, approximately 128 and 41 meters the pipeline corridor, were completed in the eastern and western portions of the property. The western transect was located near the western side of the masonry outbuilding and the existing ground collapse feature. The locations of the ERI transects are shown on Figure 2. The resistivity data were collected using a combined

PAGE 100

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 283 was encountered at variable depths, ranging from 10 to the site. The limestone bedrock deepens to the west to depths in excess of 24 meters (80 ft.). The subsurface soil and rock materials exhibited a variable density and consistency over the depth of the SPT borings. Isolated zones of very loose and very soft soil and rock materials over intervals of 0.5 to 2 meters (1.5 to 6.5 ft.) were encountered in several borings in the western and eastern portions of the site. The weak zones within the limestone bedrock. circulation losses typically occurred in association with very loose zones within the clayey soil unit (Stratum 2) within the limestone bedrock. Pleistocene deposits. The limestone bedrock correlates Stratum 1 Sand, Sand with Clay, Peat Stratum 2 Sand with Clay, Clayey Sand, Sandy Clay sand, with localized near surface deposits of peat (1 to sand (Stratum 2). The soil unit has a variable thickness, extending to depths ranging from approximately 7 to 14 meters (22 to 47 ft.) in eastern and central portions the entry point of the gas pipeline at the northwestern corner of the property. Sandy soils similar in appearance portion the property. Stratum 2 consists of variably clayey sand and minor sandy clay and was encountered as local shallow lenses zones within the limestone. Figure 3. Electrical Resistivity Profiles.

PAGE 101

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE84 Figure 4. Geologic section of pipeline alignment.

PAGE 102

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 285employed to maintain neutral buoyancy in the pipeline. was apparently utilized to facilitate the installation of the pipeline through the borehole, resulting in noticeable ground vibrations in the area. Monitoring of site conditions were conducted over the The monitoring consisted of a review of available daily as well as periodic inspections of the property. Several boring and installation of the gas pipeline. These included features and associated discharges of drilling mud near the pipeline easement in the northwestern portion of the site (Figure 5). Noticeable ground vibrations also occurred largely during installation of the pipeline within the borehole.Ground Vibration Monitoring concerns regarding potential damage to building structures. Excessive ground vibrations and ground The ground vibration monitoring utilized a remote seismograph system, which measures peak particle velocity, frequency, and air overpressures produced by vibration sources. The system records on a continuous basis in a histogram recording at a rate of 1000 samples per second with a maximum peak particle velocity with Suwannee Limestone (Oligocene). A thin section of the Tampa Member of the Arcadia Formation (Miocene) also appears to occur in the upper portion of the limestone Suwannee Limestone (Arthur et al., 2008).Monitoring of Pipeline ConstructionAs described, the gas pipeline was installed through a process involving the following phases: drilled. The borehole orientation is controlled by varying the angle of the drill bit. During drilling, the location of the borehole is surveyed using electromagnetic methods. Reaming Process: Upon completion of the pilot hole, the borehole is enlarged. Reaming tools with increasing diameters are alternatively pushed and pulled in multiple diameter. a mud pass or packer reamer is passed through the directional borehole to clean and remove soil and rock back of the steel pipeline. through the reamed borehole. A weld cap and swivel are welded to the end of the pipeline, which is then attached to the drill string to prevent rotation of the pipeline as it is pulled through the borehole. Depending on the Figure 5. Photographs of blowout and ground collapse features.

PAGE 103

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE86protection of strawberry crops. These sinkholes caused substantial damages to private and public buildings and infrastructure. Sinkhole formation in karst areas can also be triggered by construction activities such as water impoundment in reservoirs and retention basins, loading and ground vibrations from heavy equipment, changes to natural drainage patterns by diversion of stormwater, and drilling of borings and water wells. Examples of sinkholes have frequently formed in stormwater retention basins, typically following heavy rain known to occur during drilling of water wells and even geotechnical borings. In addition, the authors are familiar with an occurrence of induced sinkhole activity in 2007 during construction of the Land O' Lakes water reuse reservoir. At the site, numerous bentonite seepage cutoff wall within a dike associated with the reservoir. The sinkholes likely developed in response to the increased hydraulic loadings imposed voids near the dissolutioned limestone bedrock surface. Induced Sinkhole Formation collapse features or induced sinkholes formed in to greater than 10 ft. in diameter and were located within or near the pipeline easement in the western portion of the site (Figures 2 and 5). Drilling mud was discharged indicating a hydraulic connection with the underlying These induced sinkholes are theorized to have been triggered by increased stresses on preexisting weakened soil and rock materials associated with incipient sinkhole conditions in the subsurface. Based on the SPT borings, such conditions appear to exist as weakened or partially raveled zones within the overburden soils above the dissolution conduits within the limestone bedrock. These induced sinkholes which formed in association with the previous and current pipeline construction, and (PPV) recorded every minute. Data are transmitted via a cellular telemetry system. The geophone records vibrations on the longitudinal, transverse, and vertical pipeline easement near one of the masonry outbuildings (Figure 2). to evaluate ground vibrations, vibration data are often compared to U.S. Bureau of Mines (USBM) criteria for evaluating damage to building structures from blasting in mines and quarries (Siskind et al., 1980). While ground vibrations can be perceived by humans at very low levels cause damages to building structures are much higher, vibration levels were measured during the monitoring period. Maximum daily peak particle velocities (PPVs) correlate, in part, with installation of the gas pipeline use of the hammering device and perceived ground ground vibrations were well below USBM criteria for building damage, ranging from 1% to 2% of the limits not met, neighboring homes felt the vibrations associated with advancement of the pipe and complained of the noise created by the hammering efforts associated with installation of the pipe through the borehole.Geologic Hazards Analysis Geologic Hazards Based on local geologic conditions at the site, induced numerous examples of induced sinkholes related to ground water pumping in the region, including the vicinity of the Section 21 and South Pasco well recently, large numbers of sinkholes developed in 1998 development of a large capacity water production well. In 2009 and 2010, multiple sinkholes formed in association with heavy irrigation pumping for freeze

PAGE 104

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 287any overburden soils below the boring and into the limestone bedrock. These weakened soil zones would remain in the subsurface and may trigger future ground settlement as the drill mud dissipates with time. There is also a probability that additional zones of eroded and raveled soil zones are present in the subsurface that have not yet propagated to the land surface. The subsurface data from the geotechnical investigation are favorable for incipient sinkhole conditions over much of the pipeline corridor. An abrupt increase in depth to limestone bedrock along with zones of weak soils was found in SPT borings located in the northwestern portion of the site. The large wetland basin in the area appears to correspond with the underlying basin feature in the limestone, suggesting the presence of a large strength overburden soils and very loose and soft soil borings. Collectively, these data suggest a probability limestone. Depending on the degree of erosion, these weakened zones may have developed to the point to serve as potential locations of future ground settlement. would encompass the pipeline easement and potentially affecting building structures on the site.AcknowledgementsWe gratefully acknowledge support of the property owner who funded the investigation. The geophysical survey and ground vibration monitoring were performed & Scientists, Inc., who employed the senior author during ReferencesArthur JD, Fischler C, Kromhout C, Clayton JM, Kelley the Southwest Florida Water Management District. Florida Department of Natural Resources Bureau indicate that induced sinkholes should be considered a and pipeline construction in the region. Two mechanisms of induced sinkhole formation appear to be occurring at the site, depending on whether the limestone. In the western portion of the site, where the soils, the large diameter borehole entailed the removal lead to further weakening of any incipient raveled zones volumes of drilling mud utilized in the drilling would the weak soil zones, and triggering further raveling. The larger of these weak zones eventually propagated upward to the land surface resulting in ground settlement and collapse. boring reached its target depth and was advanced through features within the Suwannee Limestone, occurring as that often connect with deeper cavern systems (Yon the abundance of these features in the region and their presence in the SPT borings at the site, it is probable these structures were encountered during advancement of the the drilling mud would cause erosion of weaker zones in near the bedrock surface and triggering raveling of the overburden soils. These conditions could also result in ground settlement or collapse if the raveling propagates close to the land surface. events at the site, there is an elevated risk of further and pipeline installation. It is important to note that the induced sinkhole features that occurred in association connection between the limestone, the pipeline bore, and the land surface. These connections allowed the drilling mud from the borehole to discharge to the surface during drilling operations, but may have also extended through

PAGE 105

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE88 Publication No. 122. ASCE, Reston, VA. Survey Bulletin No. 59. Scott TM. 2005. Revisions to the geomorphology of Florida focusing on the Eastern Panhandle and RL. 1985. Types, features, and occurrence of sinkholes Siskind DE, Stagg MS, Kopp JW, and Dowding produced by ground vibration from surface mine blasting. US Bureau of Mines Report of In Land subsidence in the United States. US Wilson WL, Shock EJ. 1996. A description of quantitative methods for modeling new sinkhole frequency, size distribution, probability, and risk, based on actuarial statistical analysis of the new sinkhole data spreadsheet. Unpublished report by Subsurface Evaluations, Inc., Winter Park, Florida.

PAGE 106

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 289of friable, highly altered, clayey limestone consistent of the feature. Dipping beds, and fractured bedrock and surface observations suggested a lateral pathway liner for the retention pond. The collapse appears to consolidated sediments from beneath the SWRP and into boulders to gravel, a geomembrane cover, and reinforced concrete cap. Additional improvements to the SWRP included a new compacted clay liner overlain by a geomembrane liner on the side slopes of the retention pond.IntroductionKarst is a terrain with distinctive hydrology resulting developed solution channel porosity underground (Ford, 2004). Karst terrains and aquifers are characterized by sinking streams, sinkholes, caves, springs, and an transport groundwater from recharge features to springs (White, 1988; Todd and Mays, 2005). Sinkholes (also known as dolines) have long been characteristic of many karstic terrains in many areas of the world (White, age limestones of Texas in the Edwards Plateau and Balcones Fault Zone (Kastning, 1987). The purpose of Abstract study of a sinkhole that formed within a stormwater retention pond (SWRP) in southwest Austin. Results presented include hydrogeologic characterizations, fate of stormwater, and mitigation of the sinkhole. 26.5 million liters (7 million gallons) of stormwater drained into the aquifer through this opening. To determine the path, velocity, and destination of stormwater entering the sinkhole a dye trace was and arrived at Barton Springs in less than 4 days for a and geologic maps reveals that the SWRP was built within a broad (5,200 m2; 6 acre), shallow depression Photographs taken during SWRP construction showed wall. Following collapse of the sinkhole, additional hydrogeologic characterization included excavation to a depth of 6.4 m (21 ft), surface geophysics (resistivity), COVER-COLLAPSE SINKHOLE DEVELOPMENT IN THE Brian B. Hunt, P.G., Brian A. Smith, Ph.D., P.G.Barton Springs/Edwards Aquifer Conservation District, 1124 Regal Row, Austin, Texas 78748 USA, brianh@bseacd.orgMark T. Adams, P.G.ACI Consulting, 1001 MoPac Cir., Ste. 100, Austin, Texas 78746 USA, madams@aci-group.netScott E. Hiers, P.G.City of Austin-Watershed Protection Dept., 505 Barton Springs Rd., Austin, Texas 78704 USA, Scott.Hiers@austintexas.govNick Brown, P.E.Bury+Partners, 221 W. 6th St. #600, Austin, Texas 78701 USA,

PAGE 107

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE90The collapse of sinkholes is clearly a natural phenomenon. humans can accelerate the process and activate or induce a collapse sinkhole. This occurs by increasing unconsolidated materials, creating a large void and caves in the soil or regolith, resulting in collapse. Sinkhole development in the karstic areas of Texas is a common occurrence and is documented in Kastning Many studies of the eastern United States document environmental problems (Newton and Tanner, 1987). damage in the karstic Edwards, although examples may exist in areas with thick soils. Instead, the Edwards has many relatively stable sinkholes that do not cause collapse sinkholes are more accurately described as the land surface due to erosion of the overlying strata. Other stable sinkholes are formed by more recent vadose dissolution (often with a combination of collapse) and are directly linked to the current surface hydrology. cover throughout central Texas as the karst bedrock is often exposed directly at the surface. Other factors include Setting and consists of an area of about 10,800 km2 (4,200 mi2). source of water for about two million people, plus numerous industrial, commercial, and irrigation users. The Edwards Aquifer system also supports 11 threatened or endangered species, aquatic habitats in rivers of three segments. North of the Colorado River is the Northern segment, and south of the southern hydrologic divide near the City of Kyle is the San Antonio segment (Figure 1). The Barton Springs segment is located between this paper is to document the development and mitigation limestones. This sinkhole occurred in the Arbor Trails retail development stormwater pond and is referred to as the Arbor Trails Sinkhole (ATS). This case study will lead to insights into how to avoid activating or inducing sinkhole collapse in the future.SinkholesA broad discussion of sinkholes is beyond the scope as a natural enclosed depression found in karst landscapes (Williams, 2004). The mechanisms of washing), and regional subsidence. These mechanisms produce sinkholes described broadly as either a solution sinkhole, or a collapse sinkhole. A typical limestone sinkhole develops as a depression formed by the slow process of dissolution forming a broad bowl with a gentle slope. Solution sinkholes collapse rapidly due to gravitational forces following upward stoping of the cavern (void) ceiling. Sudden collapse, due to mechanically weakened unconsolidated sediments have a different morphology and can form and not collapse of mappable geologic units (Veni, 2012, also called dropout dolines, or simply collapse dolines (Williams, 2004; White, 1988). Development of sinkholes is related to the ability of (Williams, 2004). Recharge water dissolves the rock behind a highly corroded and permeable zone termed epikarst (Williams, 2004).

PAGE 108

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 291 into recharge features in the upland areas of the recharge zone (Slade et al., 1986). More recent water balance estimates of the Barton Springs segment suggest that more water could be recharged in the upland or intervening The Edwards Aquifer is inherently heterogeneous and et al., 2005). The Edwards Aquifer can be described as a triple porosity and permeability system consisting structural history, and hydrologic evolution (Lindgren et al., 2004). In the Barton Springs segment groundwater north toward Barton Springs. Numerous tracer tests have been performed on portions of the Edwards Aquifer demonstrating that rapid groundwater Johnson et al., 2012). In the Barton Springs segment these (secondary) fault and fracture trends presented on geologic al., 2002).Arbor Trails Pre-Development Site Characterization and Planning2 Edwards Rules). These requirements include geologic and environmental assessments, and reduction of pollution in stormwater leaving the site. The City of Austin has the most stringent requirements (so called SOS Ordinance) that limit impervious cover and set nondegradation standards for the treatment of stormwater these two larger segments. The Shops at Arbor Trails is the is located within the recharge zone of the Barton Springs segment of the Edwards Aquifer (Figure 1). Development of the aquifer is also thought to have speleogenesis (Klimchouk, 2007; Schindel et al., 2008). zone, located upgradient and primarily west of the sinks into numerous caves, sinkholes, and fractures along numerous (ephemeral to intermittent) losing streams. For the Barton Springs segment, Slade et al. (1986) estimated that as much as 85% of recharge to the aquifer is from Figure 1. Location map of the study area. Indicated are the Brush Country well (BC well) and a USGS stream gage station on Williamson Creek.

PAGE 109

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE92karst recharge features, springs, and wetlands. From 1994 to 2006, several development permit applications were submitted for the study property resulting in numerous environmental and geologic assessments. Beside the assessments provided in 1994 and 2004, respectively, at least two phase one environmental assessments were prepared to address hazardous material and general environmental concerns (Kleinfelder, 2005). In 2004 a karst survey and geologic assessment was property (Figure 2). The fault zone and the geologic units are consistent with the geologic map of Small et al., 1996. The on the Edwards Aquifer recharge zone. To achieve this standard, a variety of water quality measures, including construction of Storm Water Retention Ponds (SWRP) are required for development sites. Within the Edwards Aquifer recharge zone SWRPs are a type of permanent runoff and sediment so that sediments and other contaminants are not carried further downstream or into the Edwards Aquifer. The failure of a SWRP permits sediment and contaminated stormwater to leave a site and likely enter the aquifer. Both the State and the City permitting processes stipulate that a karst survey be completed to identify and evaluate all karst recharge features. In addition to the State permitting, the City requires an environmental assessment Figure 2. Predevelopment topographic map. Basemap is USGS Oak Hill Quadrangle (10-ft contours in brown). Geologic information from HBC/Terracon (2004). Geologic units and faults are consistent with Small et al., 1996. Black lines are City of Austin 2-ft topographic contours dated 1981, prior to major highway (MoPac). Contours create a depression centered around the SWRP (shown as dashed lines).

PAGE 110

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 293SWRP is to capture storm runoff from impervious areas (buildings and parking lots) and then irrigate vegetative areas with the stormwater throughout the property. The SWRP consists of two water quality controls; a and main permanent pool area that are separated by a berm. The wet pond was constructed for aesthetics within the retention basin. The retention pond has its capture volume above permanent pool elevation for the wet pond. The capture volume for the retention pond extends up 1.8 m (6 ft) onto the slope areas of the quality control structure for the surrounding shopping center. During a rain event stormwater captured by the retention basin is held and then irrigated on vegetated Hydrologic Conditions and Sinkhole CollapsePrior to collapse of the ATS, central Texas had been experiencing a severe drought. Beginning in late January, rainfall and subsequent recharge brought the aquifer out of drought conditions. On January 24, 2012, an 11 cm (4.5 in) rainfall event occurred in the area of the Arbor Trails development 4). On January 25, 2012, maintenance crews noticed the pond was draining, and that a sinkhole had developed ft) in diameter and 4 m (12 ft) deep. About 26.5 million liters (7 million gallons) of storm water drained into the aquifer through this opening. associated with the late January (and March) rainfall. These types of increases are relatively common in Conservation District (District) staff observed the runoff and recharge into swallets (Brodie Cave) within nearby tributaries of Slaughter Creek from the same rainfall event that created the ATS. It was noted that the stormwater entering those features was very turbid. to the failure of the SWRP.Sinkhole Characterization StudiesFollowing the collapse, the sinkhole was further characterized by excavation, surface geophysics, and three features were evaluated and scored as sensitive (i.e., they could be pathways for contamination) in the report, they had a small surface catchment area. The fault zone had no surface expression observed and was located based upon published maps (Small et al., 1996). The fault was not from the map alone. The geologic assessment concluded Edwards Aquifer beneath the site is considered very low As part of the site permitting processes, City staff evaluated described in the reports. This resulted in an additional karst recharge features on the study site, or a large depression in subsurface voids encountered during construction. Review of topographic contours from the City of Austin 1) reveals a very shallow and large (5,200 m2; 6 acre) depression centered on the SWRP (Figure 2). The contours agree with an even more subtle depression appears well drained from the aerial as no ponded features the subdued nature of the feature and the subsequent disturbance from the highway that bisected the eastern portion of the depression would make detection of the As part of the site engineering studies, geotechnical cores and borings were conducted throughout the site. quality designation (RQD) of very poor to incompetent The location of the SWRP for the Arbor Trails

PAGE 111

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE94 Figure 3. Detailed site map with key elements of the stormwater retention pond (SWRP), sinkhole location, and 2012 geophysical lines and boreholes. Figure 4. Photograph of sinkhole, all photos facing north. A) photo taken the day the sinkhole was observed (credit Heather Beatty, TCEQ). B) Photo taken two days after collapse and prior to excavation. Note the limestone beds are dipping to the west.

PAGE 112

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 295borehole (core) drilling by ACI Consulting (Austin, Texas). Prior to those studies the District and City of Austin (CoA) conducted the dye tracing studies. The ATS was excavated to a total depth of 6.4 m (21 ft) (Figure 6). Most of the excavated geologic material in the sinkhole consisted of friable, highly altered (weathered), clayey limestone fragments consistent Figure 5. Photograph locations indicated in Figure 4. A) Photo during construction of SWRP showing west-dipping beds in the northern wall of the forebay (photo credit Andrew Backus, 4/2/2006); B) Photo of the northern wall of the sinkhole taken two days after collapse and prior to excavation.Figure 6. Sketch of sinkhole after excavation (by Mike Warton of ACI Consulting). Very little competent bedrock was encountered in the excavations. Solution fractures striking to the north, the northern retaining wall, were observed (Figure 5). steeply dipping bedrock suggesting the ATS developed proximal to a fault zone.GeophysicsThe nature of collapse suggested the possible existence unstable material to further collapse into a void of unknown dimensions. To assess the void and assure structural stability for equipment and workers safety, a mechanism for subsurface evaluation was needed. Based on an initial review of the collapse, ACI proposed a geophysical approach. ACI uses geophysics on by geotechnical borings and subsequent construction authorities, a geophysical electrical resistivity array was Inc. (Round Rock, Texas) to evaluate the shallow surface for anomalies and take a deeper look at the subsurface.

PAGE 113

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE96spacing on lines 1 and 2 was 1.5 m (5 ft), which allowed for moderate penetration depth (18 m, 60 Other survey lines had spacing on the order of 2.1 m (7 ft), reducing resolution, but increasing the depth to over 24 m (80 ft). Each probe is connected to an Each probe alternated acting as an electrical source and receiver. The electrical pulses were recorded and the electrical energy loss recorded and the results are illustrated in Figure 7. conditions near the void and assess the surrounding area. The second bay (permanent pool) of the pond was not accessible as it was being used as a backup water quality control for development. For the array, metal spikes were driven into the ground to a depth determined based on desired resolution and survey depth. As this investigation was designed to evaluate the subsurface for the collapse geometry and to assure worker safety, a moderate spacing was chosen. Probe Figure 7. Resistivity profile from lines 1 and 2 (shown on Figure 3). The sinkhole was located between these two lines. Note the interpretation of water infiltration. This is based upon the resistivity data and the voids observed in the compacted clay material of the retention pond.

PAGE 114

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 297so many karstic features in the area, the results indicate system of the aquifer.Sinkhole Mitigation and SWRP ImprovementsAn engineered closure design by Bury + Partners (Austin, Texas) was reviewed and approved by the City and State to mitigate the sinkhole. The plan consisted vapor barrier lined the top of the gravel and a reinforced concrete slab was poured on the top and anchored into the splitter box. A compacted clay liner was installed over the concrete followed by a geomembrane liner, both of which covered the entire SWRP (Figure 11). In addition to the closure of the sink, the owners of the Since resistivity is a relative measure, two geotechnical to physically evaluate the subsurface and calibrate the geophysical model. Based on the borings, warmer (red) colors representing higher resistivity were determined to be relatively competent (crystalline) limestone. Cooler blue colors representing lower resistivity (high conductivity) were determined from Boring 1 to be wet recovery also suggesting highly fractured rock.Activation of CollapseSmall voids observed in the compacted clay liner of the side of the SWRP, suggest the most likely pathway for observations along with the geophysics and other data suggest that water from the SWRP was bypassing the liner. Other interpretations of pathways beneath the liner stormwater, resulted in a collapse of the relatively weak cover material and development of the sinkhole.Sinkhole Recharge and Groundwater Flow path, velocity, and destination of groundwater in a karst setting. A dye trace was performed to better understand therefore springshed, the ATS was developed within. The results will help scientists understand the fate of the stormwater in the ATS, and also how future contaminant study site, will move. 2012 (Figure 8). The dye was detected at one well and within the Sunset Valley groundwater basin as previously Figure 8. Phloxine B dye injection at Arbor Trails sinkhole. Dye was injected on February 3, 2012. A mass of about 7 kg (16 lbs) was mixed in a trash can and then gravity injected via a hose and polyvinyl chloride (PVC) pipe using water from an adjacent wet pond.

PAGE 115

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE98 Figure 9. Map of results from the Arbor Trails dye trace. Pink circles indicate positive detections (very high confidence, both labs) of Phloxine B. White circles are wells with tentative detections (single detections from EAA lab), and solid black circles are locations with non-detects (both labs). Dashed pink line represents estimated flow route and is coincident with the Sunset Valley Flow Route defined by Hauwert et al., 2004. Small gray circles are existing water-supply wells. Light gray potentiometric lines are from February 2002 high flow conditions (10-ft contour intervals). Groundwater basins are defined in Hauwert et al., 2004. to prevent future leakage and sinkhole development (Figure 11). Existing geomembrane liner was replaced level of the retention pond (previously the liner only existed for the wet pond). The subgrade underneath the geomembrane liner within the retention pond was replaced with new high quality compacted clay liner and geomembrane line. All masonry walls in the SWRP were grouted and sealed to prevent leakage.DiscussionFigure 12 illustrates a conceptual hydrogeologic model of depression is indicative of a solution sinkhole (Figures 2 and 12A). Evidence of a fault zone include fractures and borings revealed highly fractured and altered epikarst rock within the SWRP. The SWRP removed about 6 likely acted as a mantle of poorly consolidated material

PAGE 116

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 299covered sinkhole. In addition, geotechnical studies occur without the input from geologists surveying for karst features. Finally, geologists are not required to inspect the SWRP excavation during its construction. Despite these problems inherent in the development process, the studies and site remediation were a model of communication, transparency, and cooperation among the various regulators, scientists, engineers, and owners. All of these parties have a goal to understand the problem and provide the best solution.Conclusions can develop in the central Texas Cretaceous karst system. In this case the cover is a thick horizon of terra over a fractured and dissolved karstic fault zone (Figure the geomembrane liner and through the epikarst zone pipes (Figure 5), and upwards stoping of the void ceiling at depth. Sudden failure occurred as mechanically weak the bedrock (Figures 5 and 12C). Dye tracing established system (Figure 9). The sinkhole was mitigated with Improvements to the SWRP included extending the liner Under the current development process it is unlikely that the regulators or developers of the area in which the sinkhole occurred would have recognized the risk associated with the location of the SWRP, or predicted the failure. Only after compiling all the information does it become clear that human activities (placement of the SWRP on the sinkhole) activated the sinkhole collapse. Part of the challenge is that the land development process in the karstic Edwards Aquifer has inherent problems of communication between geologists, engineers, consultants, and owners over the life of engineered, and then the geologic assessment occurs, SWRP was located precisely in the lowest portion of the property, which makes sense from a engineering standpoint. But in this case the low elevation was a Figure 10. Photographs during sinkhole mitigation. A) Boulders and coarse fill and filter fabric; 5/2/12, B) graded cobble to gravel fill; 5/7/12, C) Gravelfilled sinkhole and filter fabric; 5/9/12, D) Reinforced concrete cap and blue vapor barrier; 5/10/12. Figure 11. A) Looking east from the splitter box showing new compacted clay liner overlain by new geomembrane. B) Looking south at the stone splitter box and the finished SWRP after significant rainfall event. New soil and vegetation cover in place over geomembrane in SWRP; 7/11/12. Note the sinkhole was located in front of the splitter box.

PAGE 117

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE100AcknowledgmentsThis study would not have been possible without the full support of the property owner of the Arbor Trails Christopher Commercial Inc. (CCI). We would also like to thank the CCI executive team of Vice President Maintenance Manager Dan Manchiela, and SWRP We would like to extend our thanks to the Edwards and Steve Johnson for their support of the dye tracing portion of this study. The EAA provided Phloxine B the dye tracing portion of this study. Justin Camp (CoA) helped sample the springs as part of the dye trace study. study confirms how human activities, superimposed upon natural karst features, can activate a connected these features can be with the aquifer be avoided if geologists and engineers are aware of the potential risks associated with SWRPs initiating sinkhole collapse. To reduce the risk of future SWRP failures, studies should be performed beyond current standards for areas impounding water, such as an SWRP. Additional studies could include detailed mapping, topographic surveys, traditional karst surveys, geophysics, and additional geotechnical borings (extending below the final grade) focused around a potential location of an SWRP. Excavations should be inspected periodically by geoscientists and engineers during construction looking for features that could contribute to sinkhole initiation. Figure 12. Conceptual hydrogeologic model of sinkhole in four stages: A) Pre-SWRP development, B) SWRP and sinkhole activation, C) covercollapse, and D) mitigation.

PAGE 118

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2101 The Shops at Arbor Trails, MoPac Expressway at William Cannon Drive, Austin, Texas. Report by at Arbor Trails, MoPac at William Cannon Drive, distribution of permeability in the Edwards Aquifer. controls on porosity development in platform carbonates, South Texas. The University of Texas structure of the Edwards Aquifer, South Texas Implications for aquifer management. The University of Texas at Austin, Bureau of Economic evidenced by groundwater dye tracing in the Barton Springs segment of the Edwards Aquifer, Central Texas: Implications for Modeling San Antonio, Texas. April 1. central Texas: Ordovician to Quaternary. In: Beck Engineering and Environmental Applications, Proceedings 2nd Multidisciplinary Conference on Sinkholes and the Environmental Impacts of Karst. Kleinfelder. 2005. Phase I environmental assessment: National Cave and Karst Research Institute Special Paper 1. Carlsbad, New Mexico. Painter S. 2004. Conceptualization and simulation of the Edwards Aquifer, San Antonio region, Texas. We would like to further thank the careful reviews and suggestions that improved this paper by Robert K. Denton, Jr., John M. Caccese, and Tony L. Cooley.References Beck B, Sinclair WC. 1986. Sinkholes in Florida: An introduction. Florida Sinkhole Research Institute studies (1996), Barton Springs segment of the Aquifer Conservation District unnumbered publication, Austin, Texas. caves and karst science. New York (NY): Fitzroy Dearborn, p. 902. New York (NY): Fitzroy Dearborn. Antonio segment of the Edwards Aquifer: matrix, fractures, or conduits? In: Wicks CM, Sasowsky transport in carbonate aquifers. Rotterdam Edwards Aquifer, southern Travis and northern and the City of Austin Watershed Protection and Development Review Department. within the Barton Springs segment of the Edwards to the Barton Springs segment of the Edwards Aquifer. World Lake Conference, Austin, Texas. channel and river contributions. City of Austin short report in preparation.

PAGE 119

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE102Maclay RW, Small TA. 1986. Carbonate geology and hydrogeology of the Edwards Aquifer in the San Antonio Area, Texas. Texas Water Development Board Report 296. sinkholes in the eastern United States. In: Beck Engineering and Environmental Applications, Proceedings 2nd Multidisciplinary Conference on Sinkholes and the Environmental Impacts of Karst. new conceptual model to explain aquifer dynamics. Sharp JM Jr. 1990. Stratigraphic, geomorphic and structural controls of the Edwards Aquifer, Texas: U.S.A.. In: Simpson ES, Sharp, JM Jr, editors. framework and hydrogeologic characteristics of the Edwards Aquifer outcrop (Barton Springs water quality of the Edwards Aquifer associated with Barton Springs in the Austin area, Texas. US Slade R Jr, Ruiz L, Slagle D. 1985. Simulation of the Edwards Aquifer in the Austin Area, Texas. US ed. New York (NY): John Wiley & Sons, Inc. terrains. Oxford (England): Oxford University Press. Encyclopedia of caves and karst science. New York

PAGE 120

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2103SALT KARST AND COLLAPSE STRUCTURES IN THE ANADARKO BASIN OF OKLAHOMA AND TEXASKenneth S. JohnsonOklahoma Geological Survey, 1321 Greenbriar Dr., Norman, OK 73072, USA, structure and distribution of overlying rocks.IntroductionThe current study summarizes years of investigations of salt karst and resultant collapse features in and around the Anadarko Basin of western Oklahoma and the Texas Panhandle. These investigations have involved integrated studies of: 1) the subsurface distribution and thickness of where outcropping strata are disrupted and disturbed. Subsurface studies have been carried out mainly by recognition of salt and associated strata on electric logs (also known as geophysical logs) of oil and gas tests, as well as examination of several cores of salt units in rock types on electric logs is well established (Alger and Crain, 1966), and I have carried out many studies using various types of well logs to identify, correlate, and map salt and gypsum beds in the subsurfacesome of the studies are in public documents (Johnson, 1967, 1981, Also of special value for this study are the cross sections of Permian evaporites in the Anadarko Basin by Jordan In the Anadarko Basin, Permian strata typically are gently dipping or horizontal, with dips normally being axis of the basin. Therefore, where dips are several degrees, or more, and particularly where strata dip chaotically in different directions within short distances, it is most likely a result of dissolution of underlying Permian evaporites (salt and gypsum) and collapse of younger strata into the solution cavities. In most areas the gypsum beds are 1 m thick, so their dissolution would not normally disrupt overlying rocks, except where the gypsum bed(s) are less than about 20 m below the land surface. On the other hand, salt units typically e AbstractPermian bedded salt is widespread in the Anadarko Basin of western Oklahoma and the Texas Panhandle, where partial or total dissolution of the shallowest salt in salts at depths of 10 m. The distribution (presence of many petroleum tests and a few scattered cores. Salt dissolution by ground water is referred to as salt karst. Chaotic structures, collapse features, breccia pipes, and other evidence of disturbed bedding are present in Permian, Cretaceous, and Tertiary strata that overly strata in the region normally is less than one degree, mainly towards the axis of the Anadarko Basin. Where strata locally dip in various directions at angles of 5 degrees or more, and underlying salt units show clear evidence of dissolution, these chaotic dips must result (mostly, if not totally) from subsidence and collapse into have been proposed in many parts of the Anadarko m thick and more than 100 m deep, and cannot contribute to disruption of outcropping strataexcept where they are within 10 m of the surface. Typical areas of disturbed bedding comprise several hectares, or more, with outcrops of moderately dipping strataas though large blocks of rock have foundered and subsided into large underground cavities. Other cia pipes, or chimneys, that extend vertically up from overlying strata. The best evidence of these chimneys are collapsed blocks of Cretaceous strata, chaotically of the Anadarko Basin. Any study of surface or shal

PAGE 121

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE104both of which are outside the region under investigation. The northern shelf of the basin extends across northern Oklahoma and much of western Kansas. The basin is Following a Late Cambrian through Mississippian epeirogenic episode, the basin underwent a period of orogenic activity when the WichitaAmarillo uplift was carbonates. This was followed by a late epeirogenic episode that began in Permian time and has persisted till today. Most Permian sediments in the Anadarko Basin are redbeds, evaporites, and marine carbonates. They were deposited mainly in fairly shallow water, although some of the redbeds are of deltaic, aeolian, or alluvial origin. Evaporites include anhydrite (gypsum at and near the outcrop), salt (halite), and variable mixtures of salt and shale; potash salts and other evaporites have not been found in the basin. The principal evaporite units involved in this study are, in ascending order, the Flowerpot salt, the Blaine salt (Figure 2). These units have been grouped into the the region underlain by salt in these units is shown in evaporites is as much as 150 m along the axis of the Anadarko Basin in Beckham and Washita Counties, Oklahoma, where all three of the evaporite units contain Salt KarstSalt (halite, NaCl) is extremely soluble in fresh at 25C, and it increases at higher temperatures. The four basic requirements for dissolution of salt and development of salt karst, are: 1) a deposit of salt against which, or through which, water can flow; 2) whereby the resulting brine can escape; and 4) energy (such as a hydrostatic head or density gradient) to cause the flow of water through the system (Johnson, 1981). When all four of these requirements are met, dissolution of salt can be quite rapid, in terms of geologic time. quite large and can result in collapse of several hundred meters of overlying strata. Four examples of salt dissolution and collapse are described in this report to show the geographic range of salt karst and resultant collapse in the Anadarko Basin. These examples are in Beckham, Washita, and Beaver Counties, Oklahoma, Geologic Setting of Anadarko BasinThe Anadarko Basin is the deepest sedimentary and structural basin in the interior of the United States (Johnson and others, 1988; Johnson, 1989a). Paleozoic sedimentary rocks are as much as 12,000 m thick along the axis of the basin, and Permian strata (which contain the salt units) are as much as 2,100 m thick. Outcropping strata in the basin are predominantly Permian, Tertiary, and Quaternary in age, with some scattered small outliers of mostly collapsed Cretaceous rocks. The basin is bounded on the south by the Wichita and Amarillo uplifts (Figure 1), and on the east and west by the Nemaha uplift and the Cimarron arch, respectively, Figure 1. Map of Anadarko Basin in western Oklahoma and Texas Panhandle, and dissolution limit of areas underlain by the Flowerpot, Blaine, and/or Yelton salts.

PAGE 122

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2105depth to the top of these salts typically ranges from 10 10 m.Collapse StructuresThe normal dip of strata in the Anadarko Basin is less Permian salt deposits in the Anadarko Basin have been dissolved locally by ground water, and this is still going on, as attested by several natural springs that emit brines formed by dissolution of the salt on the north and south 1981). Most salt karst in the Anadarko Basin occurs in the Yelton, Blaine, and Flowerpot salts (Beckham evaporites), because these are the shallowest salts; the Figure 2. Cross sections showing dissolution of Permian salt units in eastern Beckham and western Washita Counties, Oklahoma. Upper stratigraphic cross section (Johnson, 1967, Plate II) and lower structural cross section (Johnson, 1963) are along the same alignment.

PAGE 123

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE106outcropping strata in the overlying Cloud Chief, Doxey, and Elk City are undisturbed and dip at angles of but 1 or 2 degrees. At several places where the salt is known to be absent the surface beds are highly disturbed and dip in all directions at angles up to 25 degrees. I later served on the thesis committees of several mapping in Beckham and Washita Counties, and asked them to make maps showing locations within their study areas where surface rocks were undisturbed or disturbed. The work by these students (Smith, 1964; Richardson, the spatial relationship between the absence of Yelton salt and the presence of collapse structures in outcropping rocks. Suneson (2012) also recognized the relationship between Yelton salt karst and the distribution of collapsed blocks of Cretaceous strata in southwest Custer County, (discussed below). The original extent of the Yelton salt in the Anadarko of the synclinal basin, all the salt appears to have been dissolved (Figure 2). Along the axis of the basin, where the salt is thicker and more deeply buried, either all of the salt is still present or only some of the upper salt beds have been dissolved. Most of the collapse structures are characterized by hectares in size) where Permian strata dip at angles of another direction, or may even appear to be undisturbed. The various blocks have subsided irregularly into the underlying cavities as salt was being dissolved. In addition, it is likely that some of the collapse structures are breccia pipes, which are cylindrical or irregular outcrops of breccia pipes wherein Cretaceous strata were Permian strata.Study AreasFour areas of salt dissolution and collapse show the geographic range of salt karst and resultant collapse in the Anadarko Basin (Figure 1). Oklahoma examples are: number of collapse structures, and in places underlain by salt karst the overlying strata have subsided, settled, or collapsed into the underground cavities (Figure 2). This has resulted in disturbed bedding in outcropping rocks. Disturbed bedding and collapse structures caused by salt karst in the Oklahoma portion of the Anadarko Basin can be seen not only in outcrops of overlying Permian units (the Cloud Chief, Doxey, and Elk City Formations), but also in the large number of chaotic blocks of Cretaceous strata that are scattered from northern Washita County northward across much of northwestern Oklahoma Suneson, 2012). these erratic structures to dissolution in underlying gypsum beds (probably those in the Blaine Formation). and also attributed the chaotic structures to gypsum dissolution and collapse. It is now clear that gypsum dissolution is not the cause of these chaotic structures, except around the fringes of the Anadarko Basin where gypsum beds are within 10 m of the land surface. parts of the Anadarko Basin, but they are too thin and too deep in most areas for their dissolution to cause disrupted anhydrite beds in the Blaine Formation are 1 m thick, but typically they are at depths of 50 m. The only in the basin is the Cloud Chief Formation: the basal and southeastern Custer Counties, thins sharply to the west and is only 1 m thick and is 50 m deep in bearing units are too thin and too deep to contribute to the extensive areas of disrupted outcropping strata, although they may cause subsidence and collapse where they are only about 10 m deep. Recognition that the collapse structures were due to salt dissolution and collapse (not gypsum dissolution) the Yelton salt is known to be present in subsurface,

PAGE 124

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2107the surface geology by Richardson (1970), and later by collapsed blocks of Cretaceous strata, and this can be As shown above (Figure 2), the Yelton salt is dissolved subsidence and collapse of overlying strata. The Yelton is about 50 m thick in the southwest quarter of the CanuteBurns Flat study area, and it thins, by dissolution, completely missing in the northern and far eastern parts of the area. Suneson (2012) describes an excellent exposure several kilometers north of the CanuteBurns Flat area. By examining outcrops in the area, and comparing them 1) BeckhamWashita County line; 2) CanuteBurns Flat County, Texas.The Yelton salt is dissolved on the north and south sides of the Anadarko Basin in Beckham and Washita Counties, Oklahoma (Figure 2). The Yelton is the youngest salt unit in the basin, and it has been encountered at depths ranging from about 200 to nearly 400 m along the axis or totally dissolved, and overlying strata have collapsed Yelton well (the type well for the Yelton salt). It is as much as 87 m thick at the salt depocenter, about 8 km to and it thins gradually to the north and south until it thins abruptly at its dissolution front (Figure 2). Cores and electric logs of the Yelton salt in the type well indicate that the unit is about 75% halite, and the remainder or as irregular masses within impure salt. Collapse structures and disturbed bedding in outcropping strata of Beckham and Washita Counties have been are outside of the area now underlain by Yelton salt, and correctly have been interpreted to result from salt dissolution and subsequent collapse. Breccia pipes may well be present in this area (Figure 2), as there are many small chaotic collapse blocks of Cretaceous rocks exposed in the Canute area at the north end of the cross section (Richardson, 1970; Johnson and area, and farther north, are commonly less than one hectare in size, but locally are up to 5 hectares. Oklahoma is an excellent site to study dissolution of the Yelton salt and resultant collapse of overlying strata. Mapping of Figure 3. Thickness of Yelton salt showing saltdissolution area and collapse of overlying strata in the CanuteBurns Flat area of northwest Washita County, Oklahoma. Cretaceous collapse structures are mostly less than 1 hectare (2.5 acres).

PAGE 125

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE108m thick and at depths of 100 m in the CanuteBurns Flat areaclearly too thin and too deep to contribute to this collapse. Northwest Beaver County, OklahomaThe northwest corner of Beaver County, Oklahoma, contains a dramatic example of salt dissolution and collapse. In most of Beaver County the Flowerpot salt typically is 100 m thick (Figure 4) (Jordan of the county the salt unit has been dissolved along a SWNE trend, and overlying strata have collapsed and subsided about 100 m (Figure 5) (Jordan and The structure at the base of the Blaine Formation (Figure Formation in the area are 1 m thick and are equally thick on both sides of the dissolution zone, so there is no evidence that dissolution of these gypsums plays a part in the subsidence and collapse. to determine where the salt is present and where it is mapping of areas where outcropping strata are disrupted or disturbed (that is, where they dip an angles of several degrees, or more), and where they appear to not be disrupted and disturbed, and assuming that the disruption is due to salt dissolution and collapse of all overlying strata. Unfortunately, much of the south half of this study area is mantled by several meters, or more, of Quaternary silt and sands (Richardson geologic map, 1970), and this conceals evidence of whether the bedrock is disrupted or not. Therefore, the precise limits of salt dissolution here CanuteBurns Flat area, but they are too thin and too deep for their dissolution to cause disruption of overlying anhydrite beds in the Blaine Formation are 1 m thick, Figure 4. Thickness of Flowerpot salt in northwest Beaver County, Oklahoma (from Jordan and Vosburg, 1963). Figure 5. Structure map on base of Blaine Formation in northwest Beaver County, Oklahoma (from Jordan and Vosburg, 1963).

PAGE 126

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 109little or no contribution to subsidence in Meade County or in Beaver County. to the thick accumulation of the MiocenePliocene aged Ogallala Formation on the west side of the dissolution front, most of the dissolution apparently occurred prior to or during deposition of the Ogallala Formation (Irwin and Morton, 1969). Marine and Schoff (1962, Plate 1) also show great thicknesses of Ogallala and younger sediments on the west side of the dissolution front. Dissolution and collapse appear to be continuing: Irwin and Morton (1969) report that recent sudden movement County, Kans., where the land surface subsided several feet over a period of several months. Merriam and Mann (1957) report a number of recent and past sinks and sinkholes in southwest Kansas. Strata beneath the Flowerpot salt show no evidence of disturbance in this collapse zone. Structure mapping on the base of the much deeper Wellington evaporites (at a depth of about 1,000 m) shows that these strata dip gently, and uninterrupted, to the southeast at a rate of elevation of the Blaine Formation in northwest Beaver County is due to the dissolution front, and not to any This dissolutioncollapse feature is a southern continuation of the Crooked Creek fault that has been well documented and described in Meade County, Kansas (Frye, 1942, 1950). Frye attributed the Crooked Creek fault in Kansas to dissolution of Permian salt and gypsum in the shallow subsurface, and it is now clear that it is principally dissolution of the Flowerpot salt in that area; gypsum (Blaine Formation) dissolution makes Figure 6. Location maps showing area of salt dissolution and collapse in Hansford County, Texas. Cross section AB shown in Figure 8. Figure 7. Resistivity and Spontaneous Potential log showing principal lithology in Palo Duro Lake area, Hansford County, Texas (modified from Johnson, 1989). Well is R. R. Fulton #3 Lasater; H&TC Survey, Block 45, sec. 72.

PAGE 127

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE110of which is shown in Figure 8). Inasmuch as the caprock is disturbed, salt dissolution and subsidence continued, at least locally, until after development of the caprock County is a southwestward continuation of the northwest Beaver County feature (described above) and the Crooked Creek fault in Meade County, Kansas. There are, however, some patches or areas of Flowerpot salt that are partially or totally unaffected by dissolution farther west in both the Oklahoma and Texas Panhandles (Figure 1). Therefore, the structure of strata above the Flowerpot (the Blaine, Dog Creek, Whitehorse, and Ogallala strata) can be quite variable: in places they will be horizontal and undisturbed, and elsewhere they will be at least somewhat chaotic and SummaryWidespread Permian salts in the Anadarko Basin are collapse of overlying strata in parts of western Oklahoma Palo Duro Lake Area, Hansford County, Texas karst has resulted in subsidence and collapse of overlying strata in the vicinity of Palo Duro Lake (Figures 1, 6) (Johnson, 1989b). The Flowerpot salt is 0 m thick in the area (Figure 7), and at depths ranging from 180 beneath most of the study area (Figures 6, 8). Strata undisturbed (note the Cimarron anhydrite in Figure 8), zones are chaotic and are structurally low. Elevation of the overlying Blaine Formation collapses and undulates, following closely the areas where dissolution occurs in the Flowerpot salt and even in the deeper Upper Cimarron salt. Principal outcrops in the Palo Duro Lake area are sands and gravels of the MiocenePliocene age Ogallala of the Ogallala base is highly irregular, and the caprock, consisting of calcrete and silcrete, is disturbed and collapsed in two large subsidence basins in the area (one Figure 8. Cross section showing salt dissolution and collapse of overlying strata at Palo Duro Lake in Hansford County, Texas (modified from Johnson, 1989). See Figure 6 for location.

PAGE 128

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2111 of the Permian Basin in the southwestern U.S.A. Johnson KS, Amsden TW, Denison RE, Dutton SP, Midcontinent, in Sloss, L.L., ed., Sedimentary Anadarko basin, in K. S. Johnson, ed., Anadarko Johnson KS. 1989b. Salt dissolution, interstratal karst, and ground subsidence in the northern part of the Texas Panhandle. In: Beck BF, editor. Engineering and environmental impacts of sinkholes and Conference on Sinkholes and the Engineering and Environmental Impacts of Karst, St. Petersburg, Florida. October 2, 1989. p. 115. during Salado time in the Wink area, Winkler County, Texas. In: Love DW, and others, editors. Carlsbad region, New Mexico and west Texas. Annual Field Conference. p. 211. Mills, and Washita Counties, Oklahoma. Oklahoma evaporites in the Anadarko Basin of the western OklahomaTexas Panhandle region. Oklahoma Survey Bulletin 97. 74 p. Merriam DF, Mann CM. 1957. Sinkholes and related geologic features in Kansas. Kansas Academy of central Washita County, Oklahoma [MS thesis]. Norman (OK); University of Oklahoma 76 p. Richardson JL. 1970. Areal geology of western Washita County, Oklahoma [MS thesis]. Norman (OK): University of Oklahoma. 67 p. and the Texas Panhandle. Chaotic structures, collapse features, and breccia pipes are present in outcropping Permian, Cretaceous, and Tertiary strata. These strata normally dip at an angle of less than one degree towards the axis of the Anadarko Basin, but may dip at angles of 5 degrees in various directions where an underlying salt unit is partly or totally dissolved. It is clear that most (if not all) of these collapse features and chaotic dips result from dissolution of salt and salt karst. Whereas gypsum karst and resultant collapse had been proposed in many parts of the Anadarko Basin before, the gypsum beds are only 1 m thick and are more than 100 m deep in these study areas and cannot contribute to the collapse of overlying strata.References with electrical well logs. In: Rau JL, editor. Second Fay RO. 2010. Preliminary geologic map of the Foss Dewey, Ellis, and Roger Mills Counties, Oklahoma. Bulletin 45. 152 p. Formation in the Oklahoma Panhandle and Elk City area, Beckham and Washita Counties, Oklahoma. University of Oklahoma, Biennial Johnson KS. 1967. Stratigraphy of the Permian Blaine Formation and associated strata in southwestern Oklahoma [PhD dissertation]. Champaign (IL): University of Illinois. 247 p. Johnson KS. 1976. Evaluation of Permian salt deposits in the Texas Panhandle and western Oklahoma for underground storage of radioactive wastes. Union Carbide Corporation, Oak Ridge National

PAGE 129

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE112 Beckham and Roger Mills Counties, Oklahoma [MS thesis]. Norman (OK): University of Oklahoma. 64 p. and Woodward Counties, Oklahoma. Oklahoma 422. Zabawa PJ. 1976. Investigation of surficial structural geology of portions of Beckham, Custer, Roger Mills, and Washita Counties, Oklahoma [MS thesis]. Norman (OK): University of Oklahoma. 98 p.

PAGE 130

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2113 in Winkler Co., Texas (Figure 1). Wink Sink no. 1 formed in 1980 outside the small community of Wink, sinkhole ultimately expanded to a diameter of 110 m, with a total estimated volume of 159,000 cubic meters. Wink Sink no. 1 appears to have been largely inactive formed less than 2 km south of Wink Sink no. 1. Wink assumed to have formed by dissolution of salt beds in reports that a sinkhole that formed near Jal, New Mexico (Figure 1), was probably the result of Salado dissolution sinkholes all overlie the Middle Permian Capitan Reef aquifer. In the case of the Wink sinks, Johnson et al. Capitan Reef is locally above the elevation of the Salado Formation. Undersaturated water rising along a borehole by artesian pressure may have contributed to subsurface dissolution and collapse of the Wink sinkholes. Sinkholes in the greater Permian Basin region have also resulted from solution mining of Permian salt beds in the shallow subsurface. The Borger sinkholes, mining operations conducted to extract brine from the Upper Permian Flowerpot salt beds. Surface subsidence of the surface. Within the next 14 years two sinkholes ~15 m deep and 50 m in diameter had formed above the Abstract Permian Basin region of west Texas and southeastern New Mexico are of human origin. These anthropogenic activity, or with solution mining of Permian salt beds in Winkler Co., Texas formed in 1980 and 2002 within the the result of subsurface dissolution of salt caused by fresh water leakage in improperly cased abandoned oil wells. In 2008 two catastrophic sinkhole events occurred a few months apart in northern Eddy Co., New Mexico, and a third formed a few months later in 2009 near Denver City, Texas. All three sinkholes were the result of solution mining operations for brine production from Upper Permian salt beds. The Eddy Co. sinkholes kilometers from populated areas. In the aftermath of within the city limits of Carlsbad, New Mexico as having a similar geologic setting and pumping history. That well has been abandoned and geotechnical monitoring of the site has been continuous since 2008. Although there is no indication of imminent collapse, geophysical beds beneath the brine well extending north and south a railroad, and an irrigation canal.Introduction are common features of the lower Pecos region of west Texas and southeastern New Mexico. New sinkholes form almost annually, often associated with upward of regional extent that underlie evaporitic rocks at the EVAPORITE KARST IN THE PERMIAN BASIN REGION OF WEST TEXAS AND SOUTHEASTERN NEW MEXICO: THE HUMAN IMPACTLewis LandNational Cave and Karst Research Institute, and New Mexico Bureau of Geology and Mineral Resources, New Mexico Tech 400-1 Cascades Ave., Carlsbad, NM, 88220, USA,

PAGE 131

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE114Figure 1. Regional map of the lower Pecos region of southeastern New Mexico and adjoining areas of west Texas, showing locations of sinkholes discussed in text, and their position with respect to the Capitan Reef. WIPP = Waste Isolation Pilot Plant.

PAGE 132

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2115 gypsum, and greenish and reddish clay in chaotic outcrops. In most areas the Salado outcrop is covered by a few meters to tens of meters of pediment gravels and windblown sand (Kelley, 1971; McCraw and Land, 2008).Geologic settingThe lower Pecos region includes the city of Carlsbad in Eddy Co., New Mexico (Figures 1 and 4). Evaporitic rocks, primarily gypsum, are widely distributed in the Carlsbad area both at the surface and in the subsurface Northwest Shelf of the Delaware Basin (Figures 1 and The uppermost part of the Delaware Basin section is comprised of ~1700 m of redbeds and evaporites of Upper Permian (Lopingian) age (Lucas, 2006a; 2006b). This section includes the Salado Formation (Figures 2 consists of ~710 m of bedded halite and argillaceous halite, with lesser amounts of anhydrite and polyhalite. Rare amounts of potassium salts (sylvite and langbeinite) occur in the the McNutt Potash Zone near the center of the formation (Cheeseman, 1978). Clastic material makes up less than 4% of the Salado (Kelley, 1971). Potash ore is mined from the McNutt Potash Zone in underground mines a few kilometers east of Carlsbad. The formation is also the host rock for the Waste Isolation Pilot Plant (WIPP), a repository for transuranic radioactive waste in The Salado Formation thins to the north and west by erosion, halite dissolution, and onlap onto the Northwest Shelf of the basin. Because of the soluble nature of Salado rocks, the unit is very poorly exposed in an outcrop belt ~5 km east of the Pecos River valley (Figure 5). In that Figure 4. Map of study area in Eddy Co., New Mexico, showing locations of the Eddy County sinkholes. Southernmost filled circle shows the location of an abandoned brine well within city limits of Carlsbad. Figure 2. Diagramatic north-south stratigraphic section showing shelf-to-basin facies relationships in the Delaware Basin region. Line of section shown in Figure 1. Figure 3. Upper Permian (Lopingian) stratigraphy in the northern Delaware Basin in the vicinity of the WIPP Site, southeastern New Mexico.

PAGE 133

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE116air photo interpretation, with an estimated depth of 45 m (Land and Aster, 2009). Less than four months after collapse of the JWS sink, another brine well collapse occurred in northern Eddy Co. result of a brine well operation in the Salado Formation on state trust land. The well had been shut in three months earlier after it failed a mechanical integrity test as part of a statewide review of brine wells conducted in the aftermath of the JWS collapse. Then in July, 2009 another sinkhole abruptly formed near Denver City, Texas, ~115 km east of the product of a solution mining operation.Eddy Co. and Denver City sinkholesAround 8:15 on the morning of July 16th, 2008, a driver for a local water service company was inspecting a brine Carlsbad. While on location the driver noticed a rumbling noise and quickly vacated the site. Minutes later, a large associated structures (Figure 6). The well operator had fresh water and circulating it through the 86 m thick section of halite until the water reached saturation. The The brine well was being operated under permit from the New Mexico Oil Conservation Division (NMOCD). This sinkhole, referred to as the JWS sinkhole from the initials of the well operator, was originally several tens ~12 m below land surface. Large concentric fractures developed around the perimeter of the sink, threatening the integrity of County Road 217, 100 m to the south. By July 28, the walls of the sink had developed an level, above which the sides of the sink were vertical, and the water originally present had subsided into the of groundwater at shallow depths in the immediate vicinity of the JWS sinkhole. Thus we assume the water debris chimney in the initial stages of collapse, and was now stored in pore space in the resulting collapse breccia in the subsurface cavern. By this time the sinkhole had attained a diameter of ~111 m, based on Figure 5. West-east cross-section showing stratigraphic section penetrated by JWS sinkhole. Qpe = Quaternary pediment gravels and windblown sand. Figure 6. JWS sinkhole on 7/19/2008, three days after initial catastrophic collapse. Water in sink is ~12 m below ground level. View to south, with County Road 217 in background.

PAGE 134

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2117 top of the denser brine. A cushion of crude oil or diesel fuel roof and ensure that cavern excavation occurs outward rather than upward. This procedure was not applied in the sinkholes. To prevent surface subsidence and collapse, brine well operators in New Mexico are required to conduct annual pressure tests and downhole sonar surveys to assess the size and proportions of the cavern being excavated. conducting these surveys. Apparently, the mechanical strength of the mudstone and gypsum in the overlying to prevent upward stoping of the cavern roof, causing an unanticipated catastrophic surface collapse (Figure 9).I&W brine wellFormation of the Eddy Co. sinkholes in 2008 prompted NMOCD to review its regulations regarding brine well Seismic recordOn March 15, 2008, an EarthScope Transportable Array 4). This transportable seismograph is a component of continental seismic investigation program. About 6 hours before surface disruption at the site of the brine seismic signals, with vertical ground motion velocity spalling during upward stoping of the cavern roof, with seismic energy resulting from the fall of material into the solution cavity (Land and Aster, 2009). Solution miningDuring solution mining operations a subsurface cavern is excavated. Most cavern excavation occurs at the top Figure 7. JWS sinkhole on 7/28/2008, showing post-collapse draining and broadening. Note concentric fractures around margins of sink. Boulders in the sinkhole are approximately car-size.

PAGE 135

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE118 damage to individual property and civic infrastructure, and possibly cause fatalities. Following the collapse of the JWS Sinkhole, NMOCD ordered closure of the I&W brine well. The City of Carlsbad and Eddy County developed a monitoring, alarm, and emergency response system to prevent loss of life in the event that a catastrophic collapse should occur. related devices that measure shifts, subsidence, and cracks in the immediate vicinity of the brine well. within the city limits of Carlsbad (Figure 4) as having a located in relatively remote areas in northern Eddy County, the I&W operation is sited in a more densely populated area within the city of Carlsbad near the highways (Figure 4). The Carlsbad Irrigation District (CID) South Canal is about 50 m south of the wellhead, and the immediate area also includes a feed store, truck Figure 8. Transportable array seismograph TA126-A 3-day high-pass (filtered above 5 Hz) record of vertical ground velocity (top), located 13.9 km southeast of the JWS sinkhole, showing more than 6 hours of apparent precursor ground motion associated with sinkhole formation. Estimated time of surface breaching (8:15am) indicated by vertical red line. Seismograph TA125 (lower plot), located 50.3 km from the site, showed no obvious candidate precursor signals. Figure 9. Sequence of events associated with solution mining that led to development of the JWS sinkhole. Unnamed uppermost section consists of ~20 m of Quaternary sand, gravel, and calcrete.

PAGE 136

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2119 Figure 10. South-north electrical resistivity profile across I&W brine well site. This line passes within 2 m of the I&W wellhead, thus crossing directly over the subsurface cavern excavated during solution mining operations. Low resistivity zones, shown in blue and purple, probably indicate brine-filled cavities or brine-saturated breccia zones. The low resistivity zones labelled A, B, C, and D correspond to potentially hazardous areas indicated in Figure 11. Figure 11. Geophysical surveys conducted at I&W brine well facility from 2009 to 2011. Low resistivity zones A, B, C, D, and E defined by electrical resistivity survey are indicated by solid or dashed yellow lines; red filled area shows probable extent of the cavity excavated by the I&W brine well, as defined by resistivity surveys. Purple shading shows area where magnetotelluric surveys identified subsurface void thickness greater than 3 meters. White outline indicates area where a cavern signature was identified on 2D seismic reflection surveys. Inner white oval shows the area of greatest seismic disruption. Note that none of the seismic lines extended south of the CID South Canal.

PAGE 137

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE120 resources of Delaware Basin, Eddy and Lea Mexico Bureau of Mines and Mineral Resources seismic evaluation of the I&W brine cavern, Carlsbad, New Mexico. RESPEC Topical Report Johnson KS. 2002. Karst in evaporite rocks of the United land subsidence owing to salt dissolution near Wink, Texas, and other sites in western Texas and New Mexico. In: Johnson KS, Neal JT, editors. problems in the United States. Oklahoma southeastern New Mexico. New Mexico Bureau of Mines and Mineral Resources Memoir 24. water circulation in the lower Pecos Valley. In: Johnson KS. Neal JT, editors. Evaporite Karst Water Resources of the Lower Pecos Region, New Mexico: Science, Policy, and a Look to the Future. Park. In: Land L, Lueth V, Raatz B, Boston P, Love D, editors. Caves and Karst of Southeastern Land L, Aster R. 2009. Seismic recordings of an anthropogenic sinkhole collapse. Proceedings of to Engineering and Environmental Problems. Society 2009 Annual Meeting, Fort Worth, NMOCD also initiated geophysical investigations, including electrical resistivity, magnetotelluric, and shape, and lateral extent of the cavity excavated by the I&W solution mining operation. A Technical Advisory cavity, but only in general terms, since a reliable selection of the best methods and materials to prevent a collapse is not possible until site characterization is complete. Electrical resistivity surveys of the I&W brine well site, conducted by the National Cave and Karst Research Institute (NCKRI), indicate that the area is underlain by extensive low resistivity zones that represent either open cavities in the Rustler and Salado Formations that may have been caused by sagging and collapse into underlying cavities (Land and Veni, 2011; 2012) These low resistivity zones extend to the north beneath beneath residential areas south of the CID South Canal. The data suggest that solution mining of the Salado overlying Rustler strata (Figure 10). This interpretation (Woods, 2011) of the I&W site (Figure 11).ConclusionsSinkholes formed in evaporitic rocks are common features in the Permian Basin region of southeastern New of these features are of human origin, the product of solution mining of salt beds in the shallow subsurface. oped. Proper, modern design has virtually eliminated this problem in new facilities. It would appear that developing engineering safeguards for solution mining is still an evolving science.References evaporites, northern part of Delaware Basin. New Mexico Bureau of Mines and Mineral Resources Circular 184.

PAGE 138

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2121 I&W brine well, Eddy Co., NM: Unpublished report submitted to New Mexico Oil Conservation Division, NM Dept. of Energy, Minerals and Natural Resources; [cited 2012 Nov 14]. Available from: NCKRIResistivityFinalReport.pdf anthropogenic karst phenomena, southeastern New Lueth V, Raatz B, Boston P, Love D, editors. Caves and Karst of Southeastern New Mexico. New upper Permian of west Texas and southeastern New Mexico. In: Land L, Lueth V, Raatz B, Boston P, Love D, editors. Caves and Karst of Southeastern Martinez JD, Johnson KS, Neal JT. 1998. Sinkholes in Map, Eddy Co., New Mexico. New Mexico Bureau Mexico: Evaporite dissolution, drill holes, and the potential for sinkhole development. In: Johnson KS. Environmental Problems in the United States. 226. Carlsbad, New Mexico: DMT Technologies Division; [cited 2012 Nov 14]. Available from: IWMagnetotelluricReport.pdf.

PAGE 139


PAGE 140

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2123caves and rapidly evolving landscapes, all of which are coupled to a complex and poorly understood eastern New Mexico and far west Texas, with Castile strata outcrops limited to Eddy County, New Mexico, and Culberson County, Texas, along the western edge of the Delaware Basin (Figure 1). The region occurs within the northern portion of the Chihuahuan Desert with annual Annual precipitation averages 267 mm, with greatest summer monsoonal storms. Cave and karst development in far west Texas and southeastern New Mexico is extensive and widespread; however, most people envision the famous carbonate Lechuguilla Cave) when they think of this region, which overlooked and generally less studied, evaporite karst in the region is more extensive and widespread. In interbedded gypsum, carbonate, and clastic strata host extensive karst development in the Castile Formation, while strata overlying (Salado and Rustler formations) similar karst development (Stafford and Nance, 2009). Most research on evaporite karst associated with Permian deposition in the region has received little study, including geology (e.g. Forbes and Nance, 1997; Nance, AbstractKarst development in Permian Castile evaporites has resulted in complex speleogenetic evolution with multiple phases of diagenetic overprinting. More than throughout Culberson County, Texas, and Eddy County, Castile sulfates crop out. Cave development is largely from the underlying Bell Canyon Formation migrate near vertically through the Castile Formation, creating caves up to 100 meters deep and over 500 meters long, which have been breached through a combination of collapse and surface denudation. Numerous small and laterally limited epigene features occur throughout the region, as well as the anomalously large Parks Ranch Cave System with more than 6.5 kilometers of cave development and caves exhibit varying degrees of epigenic overprinting Water resources in the Castile Formation are directly related to karst development with extremely colonies. Isolated stream passages in northern Culberson Local water tables vary greatly over the region and few of hydrocarbons complicates regional hydrology and diagenesis, resulting in extensive evaporite calcitization, permeability structures.IntroductionThe arid southwestern United States hosts unique EVAPORITE KARST AND HYDROGEOLOGY OF THE Kevin W. StaffordDepartment of Geology, Stephen F. Austin State University, P.O. Box 13011, SFA Station, Nacogdoches, Texas, 75962-3011,

PAGE 141

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE124 a complex and evolving hydrogeologic system.Study AreaThe Castile Formation crops out over approximately 1800 km2 in Eddy County, New Mexico, and Culberson in the north to the Apache Mountains in the south (Figure 1). The western portion of the Castile Formation truncation associated with the uplifted Delaware Mountains. To the east, the Castile Formation dips into the subsurface beneath younger strata where it reaches a maximum thickness of 480 meters (Kelley, 1971). Intrastratal speleogenesis and diagenetic alterations occur throughout the eastern portion of the Delaware Basin where Castile evaporites are exposed at the land surface (Stafford et al., 2008a). Permeable siliciclastics of the Bell Canyon and Cherry Canyon formations underlie the Castile Formation throughout the Delaware Basin; the carbonate reef facies of the Capitan Formation (Figure 2) (Lee and Williams, 2000). The Castile Formation was deposited in the Ochoan, Reef, which is now exposed at the surface in the at the end of the Permian (Kirkland and Anderson, 1970; halite. Sulfates are hydrated to gypsum near the surface in the deeper subsurface. Similarly, halite interbeds are with limited reporting associated with karst inventories (e.g. Eaton, 1987; Belski, 1992; Lee, 1996). In the last in an attempt to characterize its speleogenetic and Stafford et al., 2008a, 2008b). Castile evaporites host a complex history that includes phases of hypogene and epigene cave development, high rates of landscape denudation, extensive evaporite calcitization, and a Figure 1. Castile Formation outcrop area showing relationship to major depositional features of the Delaware Basin and geomorphic features. Small inset shows relationship of Delaware Basin to the Orogrande Basin (OB), Val Verde Basin (VB) and Midland Basin (MB) (from Stafford et al., 2008c).Figure 2. Simplified stratigraphic section of Permian strata in the Delaware Basin region (adapted from Scholle et al., 2004).

PAGE 142

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2125 bedrock, with typical varieties associated with the slope of the rock surface they are developed upon (Stafford et al., 2008b). Deeply incised rillenkarren form long and length of rillenkarren increasing proportionally from moderately dipping surfaces to near vertical surfaces. apexes of dipping rock surfaces and converge through dendritic channels into rillenkarren. Kamenitzas form on algal crusts that hydrate during monsoonal storms. In addition to the normal karren forms, tumuli commonly limited to more deeply buried portions of the Castile Formation where they have not been removed by intrastratal dissolution (Kelley, 1971). Calcite laminae fabrics have been diagenetically altered since deposition, creating irregularly laminated, massive, nodular, and Throughout the Castile Formation, evaporite calcitization is common and widespread. Stafford et al. (2008c) documented more than one thousand regions of calcitized evaporites exposed at the surface across the into calcitization associated with vertical breccia pipes and calcitization associated with intrastratal dissolution Kirkland and Evans (1976) attributed these calcitized evaporites to bacterial sulfate reduction; however, isotope analyses by Stafford et al. (2008c) showed that differentiate the calcitization as being the result of bacterial sulfate reduction or thermal sulfate reduction. These extensive calcitized evaporites have formed along gases delivered from the underlying Bell Canyon and Cherry Canyon formations have migrated through Castile strata (Stafford et al, 2008c). Associated with most of the calcitized evaporites, intrastratal dissolution and recrystallization of microcrystalline sulfate into macrocrystalline sulfate (selenite) is common, as well as native sulfur, which is limited due to high rates of oxidation in near surface calcitized zones. Surficial-Karst Development crops out is widespread and diverse, ranging from large sinkholes to small karren developed on exposed bedrock (Stafford et al., 2008b). The Castile Formation has been 2) of the surface area is composed of exposed bedrock and less than one percent is calcitized evaporites. Alluvium associated with Quaternary gravels commonly occurs along incised drainages, while residual outcrops of the Rustler Formation occur as patchy remnants across the Figure 3. Castile Formation outcrop area geomorphic map, delineating zones of exposed bedrock, gypsite, calcitized evaporites, and karst features.

PAGE 143

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE126Subsurface Karst DevelopmentCave development in the Castile Formation is diverse and widespread, but not uniform. The spatial distribution of caves mimics that of sink development delineated cave development scattered amongst regions of poor cave development. Caves exhibit varying degrees of structural control, with many features being purely developed along fracture planes, while others show no distinct correlation to structural deformation (Stafford et al., 2008b). Cavernous porosity includes hypogene caves and intrastratal breccias, epigene caves, and hybrids that have resulted from epigenic overprinting of hypogene systems (Figure 6). resulting in a buckling of the local land surface over regions up to several meters. In regions where karren bedrock as a result of dissolution and reprecipitation of gypsum at the land surface (Stafford et al., 2008b). et al. (2008a) predict that more than 10,000 individual closed depressions exist across the study area based on on length to width ratios, and it was determined that at least 55% of the closed depressions were collapse structures while the remaining 45% were a mix of purely where surface denudation has enabled collapse features of subsurface voids that may not have been originally The widespread occurrence of collapse sinks and year investigation of surface denudation across the region utilizing standard gypsum tablets. Their data fall monsoonal season, with rates generally increasing northward away from the Apache Mountains to the southern and northern margins of the outcrop area, respectively (Shaw et al., 2011). When compared with the distribution of karst development assessed through sinkhole delineation (Figure 4), a similar trend is observed; however, sinkhole distribution is highly localized and occurs more frequently towards the western portion of the Castile outcrop area where total Castile thickness is reduced on the updip side of strata. Figure 4. Distribution of closed depressions across the Castile outcrop area derived from GIS analyses (modified from Stafford et al., 2008b).

PAGE 144

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2127humanly impassable within a few tens of meters as a result of the rapid solution kinetics associated with calcium sulfate (Klimchouk, 2000). Zombie Cave (Figure 6A) and Dead East Cave (Figure 6B) are the largest purely epigene caves that have been documented within the 41 meters, respectively. Although exceptionally large for epigene caves in the Castile Formation, they exhibit typical characteristics, including: 1) narrow apertures developed along dominant fractures; 2) laterally of secondary fractures proximal to the main conduit development. Most epigene features are developed in laminated, massive, and nodular facies; however, they occasionally develop in tabular (selenite) gypsum along crystal planes and in gypsite soils along zones of permeability contrast. extensive and deep cave systems within the Castile systems and isolated solutional chambers (Stafford et al., features, including risers, cupolas, wallchannels, and ceiling channels, often connected in series that show the complete suite of morphological features created (Klimchouk, 2007). Unlike epigene caves, hypogene features do not show rapid decreases in aperture width but instead exhibit zones of increasing pore volume beneath permeability boundary horizons. While hypogene caves do still exhibit passage correlation with fracture planes, they do not show speleogenetic dominance along preferential planes, but instead anticlinal structures and fracture planes with complex interconnected passages at multiple levels, while Bee Line Cave (Figure 6E) is largely a single solutional dome room in an ascending series of cupolas that have been overprinted by meteoric processes. Crystal series with interconnected ceiling and wall channels, in which the lowest portions of the cave descend below the current water table and are developed in selenite. Other hypogene caves in the study area (e.g. Black Dog Cave) exhibit inverse dendritic patterns where multiple Most caves within the Castile outcrop area are small and laterally limited epigene features (Figures 6A & 6B), having formed as either local groundwater recharge features or shallow bypass features proximal to incised arroyos. Epigene caves exhibit scalloped walls, are developed along fracture planes, and can be associated with solutional sinkholes that cover several hectares and routinely pass through these features, likely associated with extensive runoff during monsoonal storm events (Stafford et al., 2008b). Epigene cave apertures rapidly decrease away from entrance areas, generally becoming Figure 5. Average surface denudation for the Castile outcrop area based on standard tablet studies (modified from Shaw et al., 2011).

PAGE 145

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE128 Figure 6. Simplified plan and profile maps of Castile Formation caves, including epigene caves (A,B), hypogene caves (C,E,G), stream caves (D), collapse structures that breach the water table (F) and the deepest cave in the study are which intersects the water table (G).

PAGE 146

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2129 the incised valley, indicating that this spring location is the result of downward cutting of the incised valley and conduit breaching. Parks Ranch Cave appears to contain minor hypogene components that have been connected and heavily overprinted by epigene processes, forming a complex, hydrologically active system. Water ResourcesWater resources in the Castile outcrop area are scarce and limited to occasional springs, seeps, and caves that breech conduits. Most springs and seeps in the study area exhibit high total dissolved solids, primarily sulfate, as a result of the rapid saturation of waters passing through gypsum facies (Stafford et al., 2008b). solids to support diverse ecological systems, including several species of arthropods and healthy, riparian zones rich and relatively fresh water resources, are utilized by ranchers for livestock. Many of the springs provide perennial discharge and degas elevated hydrocarbons that provide source material for current evaporite calcitization in the subsurface and represent upward leakage of active hypogene processes to the land surface (Stafford et al., 2008b). China Mine, subsurface; however, these patterns are not common in the study area. While hypogene caves are the largest features in the study area, they are not the dominant cave type documented (Stafford et al., 2008b). This is likely a sampling bias because of the nature by which these caves form from hypogene features exhibit strong correlation with regions of evaporite calcitization and diagenetic recrystallization of original gypsum fabrics into tabular fabrics (Stafford et al., 2008c). Therefore, it is logical to assume that the development and evaporite calcitization were utilized by both ascending waters and hydrocarbons from the underlying Bell Canyon and Cherry Canyon Formations (Figure 7).The largest individual cave currently documented in the Castile Formation is Parks Ranch Cave with over 6.5 kilometers of surveyed passage (Figure either hypogene or epigene karst and likely represents a incised solutional sinkholes contribute meteoric runoff to the system which primarily discharges to Chosa Draw, a locally incised valley. Most passages exhibit exist that are effectively scallop free and contain minor (Stafford et al, 2008b). The current discharge point in Chosa Draw is a bisected cave passage and continues as Figure 7. Simplified paleohydrology diagram associated with evaporite calcitization and upward migration of hypogene fluids (adapted from Lee and Williams, 2000). White arrows show migration paths of meteoric waters; blue areas show migration paths of upward migrating, hydrocarbon-rich fluids.

PAGE 147

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE130 fresh. These features along with other isolated caves across the study area provide windows into the current hydrologic system, which are distinctively different microbial communities. glimpse into the extremely heterogeneous hydrogeologic system. Fluid chemistry varies rapidly over distances of several kilometers, with water tables that appear hydrogeologic system couples components of shallow that often exhibit a minor thermal component, high total dissolved solids, and host complex bacterial colonies. with aquifer compartmentalization.ConclusionsThe Castile Formation crops out in the western Delaware Basin in Eddy County, New Mexico, and Culberson County, Texas, and hosts extensive karst development. Surface denudation is rapidly modifying the landscape at breaching of numerous caves throughout the region. More than ten thousand individual sinkholes have been three thousand documented. Sinkholes provide windows into the complex speleogenetic evolution of the region, where hypogene caves form deep and complex systems that are being actively overprinted by epigene processes. Associated diagenetic alteration of the Castile sulfates has produced widespread regions of evaporite calcitization in association with ascending hydrocarbons from underlying clastic strata. The Castile Formation records the evolving hydrogeologic history of the Delaware Basin as multiple strata, creating a system that is continuing to be overprinted by current and active processes. Initial Castile Formation studies indicate a complex hydrogeologic system that couples ascending transverse chemistries have resulted in an extremely heterogeneous system in which perched water tables and a high an early 20th century sulfur mine in northern Culberson secondary sulfur deposition but does not exhibit spring discharge (Stafford et al., 2008c); however, it is likely that it is a nonaqueous version of the smaller hydrogen structure approximately 15 meters in diameter that drops approximately 10 meters to water which forms a lake 15 to 20 meters deep (Figure 6F). The feature degasses occur throughout the lake. While this feature does not discharge to the surface as a spring, it does provide a window into the local water table, while a nearby water well descends over 50 meters to the local water table. Several caves in the area are vadose conduits, where subsurface. Brantley Stream Cave (Figure 6D) contains over 100 meters of cave passage with an active stream that sumps in regions and discharges to the Delaware River, a perennial stream in northern Culberson County. Similarly, a series of collapse sinkholes intersect a separate conduit system in Sinkhole Flat approximately 20 kilometers southeast of Brantley Stream Cave on the opposite side of the Delaware River. These collapse sinks provide windows into a relatively freshwater conduit system approximately 10 meters beneath the land surface with no obvious surface connections; however, the stream exhibits low total dissolved solids. In the southern portion of the study area, Crystal Cave Figure 8. Simplified plan view map of Parks Ranch Cave with entrances labeled (e). Note that the easternmost entrance was formed by conduit truncation as Chosa Draw entrenched, while the western entrances are associated with sinkholes (from Stafford et al., 2008a).

PAGE 148

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2131 Park. In: Land L, Lueth V, Raatz B, Boston P, Love D, editors. Caves and karst of southeastern Alamogordo (NM): Southwestern Region of the National Speleological Society. Lee MK, Williams DD. 2000. Paleohydrology of the Delaware Basin, western Texas: overpressure development, hydrocarbon migration, and ore genesis. Bulletin of the American Association of a study of denudation in gypsum karst, Chosa Draw, University of Northern Colorado. 82 p. development of a low relief evaporite karst drainage basin, southeastern New Mexico [MS Thesis]. Albuquerque (NM): University of New Classic upper Paleozoic reefs and bioherms of west Texas and New Mexico. Socorro (NM): New Mexico Institute of Mining and Technology. Delaware Basin, southeastern New Mexico and far west Texas. In: White WB, editor. Proceedings of the 15th International Congress of Speleology. 2008a. Epigene and hypogene karst manifestations of the Castile Formation: Eddy County, New Mexico and Culberson County, Texas, USA. 2008b. Castile evaporite karst potential map methodological comparison. Journal of Cave diagenesis in the western Delaware Basin. degree of aquifer partitioning are prevalent in aquifer system. Future studies should focus on delineating this compartmentalization and unraveling geochemical creates a continuously evolving system that can pose continues to breach the system.References Albuquerque (NM): Southwestern Region of the National Speleological Society. Alamogordo (NM): Southwestern Region of the National Speleological Society. Forbes J, Nance R. 1997. Stratigraphy, sedimentology, and structural geology of gypsum caves in east central New Mexico. Carbonates and Evaporites Basin Section of the SEPM. Studies 62 (2). southeastern New Mexico. Socorro (NM): New Mexico Bureau of Mines and Mineral Resources. outcrop examination of the Capitan shelf margin, Kirkland DW, Anderson RY, 1970. Microfolding in the Castile and Todilto evaporites, Texas and New Kirkland DW, Evans R. 1976. Origin of limestone Klimchouk A. 2000. Speleogenesis in gypsum. In: Klimchouk A, Ford DC, Palmer AN, Dreybrodt W, editors. hydrogeological and morphometric perspective. Carlsbad (NM): National Cave and Karst Research Institute.

PAGE 149


PAGE 150

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2133IntroductionThis study examines aspects of the subsurface geology of an area along the Clear Fork of Brazos River (Clear Fork) in parts of Throckmorton, (Figure 1). The study area extends from the town of Lueders in the southwest to Paint Creek in the northeast. It includes the originally proposed Cedar Ridge Reservoir damsite (CR) and the newly proposed damsite (site A), located about 8 km farther upstream (to the southwest). The study focuses on the distribution, thickness, and structure of a series AbstractCedar Ridge Dam and Reservoir will be built to supply water for the city of Abilene, Texas. The original damsite (CR) was to be located on Clear Fork of Brazos River in Throckmorton County, but initial coring of the damsite soluble rock that typically contains karst features, and its presence in a dam foundation or impoundment area could allow water to escape from the reservoir. A decision was made to look at potential sites farther upstream (to the be deeper underground and karst problems would be minimized or eliminated. The first phase of the relocation was a comprehensive field study of Clear Fork Valley, upstream of the original damsite, to identify gypsum outcrops; above damsite CR. The second phase of the study geophysical logs to identify, correlate, and map the subsurface gypsum and associated rock layers upstream of the original damsite. The gypsiferous damsite (A), about 8 km to the southwest. Subsequent coring and other studies of the new damsite proposed dam. There is no evidence of solution channels or other karst features beneath this site, and thus there is little likelihood of water loss from the reservoir at the new site due to gypsum karst. GYPSUM KARST CAUSES RELOCATION OF PROPOSED Kenneth S. JohnsonOklahoma Geological Survey, 1321 Greenbriar Dr., Norman, OK 73072, USA, J. Mark WilkersonFugro Consultants, Inc., 2880 Virgo Lane, Dallas, Texas 75229, USA, Figure 1. Location map showing originally proposed Cedar Ridge Reservoir damsite (CR), and site of the newly proposed dam (A) and reservoir.

PAGE 151

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE134because of extensive gypsum karst in the abutments gypsum karst that underlie the impoundment area (Jarvis, the 1980s and 1990s (Pearson, 2002).Methods of StudyDetermining the subsurface distribution, thickness, and structure of gypsum beds in the study area required examining the electric logs (also known as geophysical Clear Fork. Recognition of gypsum beds and associated rock types on electric logs is well established (Alger and Crain, 1966), and the senior author has conducted many studies using various types of well logs to identify, correlate, and map gypsum beds in the subsurface some of these studies are available in public documents many others are contained within consulting reports. On each well log examined in the study area, individual gypsum beds (and interbedded shale units) that are at with continuous cores that were drilled near several of 21, 2008, at the original Cedar Ridge damsite. The core correlated with gypsum beds interpreted to be present on electric logs for two wells (#69 and #66) drilled 100 Well 66 contains several thin gypsum beds at the top of Farther to the southwest, in the vicinity of newly proposed 2009) are readily correlated with those in the electric The current study was prompted by the unexpected the originally proposed damsite during a preliminary investigation in the summer of 2008. Because gypsum was not known to crop out in the area (it is eroded, and studies of the area made no mention of gypsum not only was it a surprise to discover karstic gypsum in preliminary cores at the original damsite, but also a blowout of natural gas was encountered at a depth of 20 m beneath the proposed dam alignment at CR. susceptible to partial or total dissolution by ground water, and may develop karst features such as caves, sinkholes, crop out at one small site about 1 km upstream of CR, and should also be present along the river for several km farther upstream in the Clear Fork Valley (however, they do not crop out). Due to the potential for gypsum karst along this portion of the river, the distribution and depth of the various gypsum beds are important factors to consider location and construction because it has had an adverse impact on holding water behind a dam at several sites in the United States. Dams built upon gypsum karst result in collapse and failure of the dam (Johnson, 2008a, 2008b). If gypsum karst is located within the proposed impoundment area of a reservoir, water can penetrate the karst features and may escape from the reservoir. Several articles have been published on properties of dam foundations built upon gypsum deposits (James and in the United States are: Quail Creek Dike (Utah), Upper Mangum Dam (Oklahoma), Anchor Dam (Wyoming), (Johnson, 2008b). Quail Creek Dike failed in 1989 Mangum Dam was abandoned before construction,

PAGE 152

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2135 showing up very clearly is another rock unit that is herein referred to informally as the Upper Shale: this shale is are generally 1 m thick.Results of StudyWith recognition of gypsum and shale beds on examination of nearby cores, it is then possible to electric logs throughout the study area (Figure 4). Figure 4 is a structural cross section showing that the gypsum beds dip to the west, and therefore are deeper below the land surface and below Clear Fork to the west. It also shows that some of the gypsum beds present in the west are thinner to the east, and some of them disappear and even grade laterally into shale to the east. The entire gypsum sequence is about 45 m thick Figure 2. Gypsum beds in Core B, drilled at the original Cedar Ridge damsite (CR), are correlated with electric logs of nearby oil wells. Figure 3. Gypsum beds in Core B, drilled near newly proposed damsite A, are correlated with electric log of a nearby oil well.

PAGE 153

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE136Local irregularities do exist, where the dip is slightly higher or lower, and the direction of dip varies slightly. the gypsum sequence near damsite A, and about 25% of the total thickness near CR. Upon establishing the recognition of gypsum and shale units on electric logs, all 100 of the well logs within a larger study area were examined and the gypsum and shale units were identified and correlated. The depth to the top of the uppermost was identified and plotted on a map (Figure 5). In some areas, mainly in the western part of the study area, additional gypsum beds are present above gypsum sequence in those areas. Similarly, towards the east, the upper and lower gypsum beds disappear Valera gypsum sequence becomes thinner. Valera Formation. It shows that the gypsum units dip Figure 5. Structure-contour map on top of youngest gypsum bed in Jagger Bend/Valera Formation. Elevations at each well are in feet, but contour interval is 6 m ( feet). Figure 4. Structural cross section showing gypsum beds dipping down to the west. Gypsum beds thin and grade laterally into shale to the east.

PAGE 154

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2137The elevation of the top of the gypsum sequence is about 17 m above stream level of Clear Fork at the original Cedar Ridge damsite, and thus the upper part of the gypsum sequence is, or should be, exposed in the valley walls (Figure 6). The top of the gypsum is then at successively lower heights above stream level in the valley upstream from CR because of: a) westward dip of the gypsum sequence (Figures 4, 5, 6); and b) the 6). The uppermost gypsum dips beneath stream level in the vicinity of borehole SB. Therefore, gypsum interpretation), in all parts of Clear Fork Valley from CR up to the vicinity of borehole SB (Figures 5, 6). believed to be deep enough below the proposed reservoir (below) and the impounded reservoir water (above). Another result of this subsurface study is recognition that a large number of oil and gas wells have been drilled along and near Clear Fork in the study area. These provide many electric logs that can be used to evaluate of proposed damsite A, near the common corner of (above sea level) of the top of the highest gypsum bed throughout the area. By comparing this map (the elevation of the highest gypsum) with topographic maps, it is possible to determine how deep the gypsum is below the land surface, and also whether gypsum beds should be exposed in the valley walls of Clear Fork. The uppermost gypsum beds are exposed, or should be exposed, in the valley of Clear Fork at and near the originally proposed out at one location near CR, but at other places where it should crop out the gypsum is either eroded, dissolved, or is covered by alluvium, colluvium, or soil. If a dam is constructed upon gypsum, or if lake water is impounded too closely above gypsum in Clear Fork Valley, it could be detrimental to dam integrity. Potential karst development in the gypsum could provide pathways for impounded water to escape from the reservoir and be discharged downstream of the dam. Also, if such a pathway is established, the gypsum would undoubtedly be further dissolved, and the pathway would be enlarged. Therefore, it is important to know where gypsum does, or should, crop out in Clear Fork Valley. Figure 6. Schematic cross section showing west dip of gypsum beds beneath Clear Fork of Brazos River and damsites CR and A. Top of gypsum sequence is above stream level at CR, and is about 23 m below stream level at damsite A.

PAGE 155

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE138 and other strata are present to separate reservoir water from the gypsum sequence. At the newly proposed damsite (A), the Clear Fork above the gypsum sequence, is 6 m thick. The latest core drilling at this site does not indicate the presence of any karst features in any of the gypsum beds. Therefore, this site appears to be favorable and warrants further investigation.References with electrical well logs. In: Rau JL, editor. Second properties of a dam foundation containing gypsum seams. Rock Mechanics and Rock in foundations of hydraulic structures. The Institution of Civil Engineers (London), 1989. Investigation of the cause of Quail Creek Dike failure. Unpublished report of the Independent of Utah. 155 p. Dam. In: Johnson KS, Neal JT, editors. Evaporite Johnson KS. 1967. Stratigraphy of the Permian Blaine Formation and associated strata in southwestern Oklahoma [PhD dissertation]. Champaign (IL): University of Illinois. 247 p. of the Permian Basin in the southwestern U.S.A. Johnson KS. 1985. Structure contour map and Scale 1:125,000. Johnson KS. 1989a. Development of the Wink Sink in West Texas, U.S.A., due to salt dissolution and Sciences 14 (2): 81. the thickness, depth, and distribution of gypsum beds. of wells in the impoundment area of the proposed reservoir that could impact the reservoir and its water quality. These boreholes are potential pathways for surface and mix with reservoir water. They also are into the gypsum beds. Producing oil and gas wells in or and sealed; and even dry or abandoned wells within the impoundment area must be found, to ensure that they have been properly plugged and sealed. Summary contains cavities, sinkholes, and caves (karst features), and its presence in a dam foundation or in an impoundment area could allow water to escape from the reservoir. The presence of gypsum at the original Cedar Ridge damsite (CR) on Clear Fork of Brazos River was confirmed in core holes, and a decision was made to look at potential sites farther minimized or eliminated. The current study focused on examination of nearly and map the gypsum and associated rock layers of the km in the vicinity of prospective damsite A (Figures they grade laterally into shale and pinch out in that direction. Clear Fork Valley upstream from the original Cedar Ridge damsite, all the way to the vicinity of borehole SB (Figures 5, 6). The presence of gypsum beds in this portion of the valley means that there may be karst pathways whereby impounded water could escape a reservoir built downstream of SB. Therefore, the best location for a dam on Clear Fork would be at a site located some distance upstream from SB, at a

PAGE 156

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2139Johnson KS. 1989b. Salt dissolution, interstratal karst, and ground subsidence in the northern part of the Texas Panhandle. In: Beck BF, editor. Engineering and environmental impacts of sinkholes and karst. on Sinkholes and the Engineering and Environmental Impacts of Karst, St. Petersburg, Florida: A. A. during Salado time in the Wink area, Winkler County, Texas. In: Love DW, and others, editors. Carlsbad region, New Mexico and west Texas. Annual Field Conference. p. 211. United States, In: Johnson KS, Neal JT, editors. problems in the United States. Oklahoma of the Upper Mangum Damsite in southwestern Oklahoma, In: Johnson KS, Neal JT, editors. problems in the United States. Oklahoma constructing dams in the USA: Environmental dams, reservoirs, grouting, groundwater protection, water tapping, tunneling. Belgrade, Yugoslavia: cause of the Quail Creek dike failure. Association southwestern Utah: Failure and reconstruction of Quail Creek Dike, In: Johnson KS, Neal JT, editors. problems in the United States. Oklahoma and effects for Colorado Front Range water

PAGE 157


PAGE 158

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2141These gypsiferous sequences are both underlain by dolomite or dolomitic limestone aquifers (Figure 2). Abstract active gypsum dissolution causes subsidence and abundant Darlington, through Ripon to Doncaster. The topography and easterly dip of the strata feed artesian water through the dolomite up into the overlying gypsum sequences. The oC, many emanating from sinkholes that steam and do not freeze in the winter the east through the overlying Triassic sandstone aquifer. attest to the passage and escape of this groundwater. The sizes of the sinkholes, their depth and that of the associated breccia pipes are controlled by the thickness of gypsum that can dissolve and by the bulking factors associated with the aggravates the subsidence problems, both by active dissolution and drawdown. Furthermore, the gypsum and dolomite where active subsidence is expected; their presence along with records of subsidence can inform planning and development of areas requiring mitigation measures. Introduction and Roxby Formations in northeast England (Figure 1).THE ROLE OF SULFATE-RICH SPRINGS AND GROUNDWATER IN THE FORMATION OF SINKHOLES OVER GYPSUM IN EASTERN ENGLANDAnthony H. Cooper British Geological Survey, Keyworth, Nottingham, NG12 5GG, United Kingdom, E. Odling, Phillip J. Murphy School of Earth and Environment, The University of Leeds, LS2 9JT, United Kingdom,, Claire Miller, Christopher J. Greenwood, David S. BrownFormerly School of Earth and Environment, The University of Leeds, LS2 9JT,, Figure 1. Map of study area showing locations of main groups of sulfate-rich springs.

PAGE 159

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE142Spa (Thorpe Arch), Knaresborough, Ripon, and east of 1986, 1998) (Figure 1). The area is prone to subsidence and sinkhole formation caused by the dissolution of the gypsum and the evolution of gypsum cave systems beneath the area. Some places are more susceptible or buried valleys cut through the sequence allowing escapes where the Tees Valley cuts the Permian sequence sinkholes, one of which collapsed in the 12th century. Farther south, sulfurous water is recorded at Snape from here south to Ripon where the Permian sequence 1986) and similar water emanates from the petrifying spring of the Dropping Well near Knaresborough. In Brotherton and subsidence ponds at Askern. The ponds prevent them freezing easily in the winter. leads to the formation of sinkholes that can be up to 20 m or more across and up to 20 m deep (Cooper, 1986, 1989, The Edlington Formation containing up to 40 m of gypsum is underlain by the Cadeby Formation aquifer and overlain by the Brotherton Formation aquifer. This in turn is overlain by the Roxby Formation with another The Permian sequence dips gently eastwards with a dip of a degree or so. It presents a wide dip slope of Cadeby area that collects water and channels it down dip to the east where it rises up in the lower ground through the gypsiferous sequence (Figure 2 and Cooper, 1986, 1988, 1998). The local rainfall averages around 594mm year averages over 6 sites; Leeming, Topcliffe, Church Fenton, Linton on Ouse and Dishforth); the average oC with a minimum average of 4.7oC and a maximum average o 1oC to 20ooC or above 20oC slope and across the outcrop of the overlying strata (Figure 2). Many springs emerge under artesian pressure, some from within sinkholes that do not freeze in the winter due to the groundwater being at a temperature of 912oC. The outcrop of these strata, where gypsum is present in near Doncaster in the south through Askern, Boston Figure 2. Generalised cross-section through the easterly dipping Permian sequence showing water flow through the Permian dolomites and limestones into the Permian gypsum and then to the surface. The geological sequence is shown in Figure 3.

PAGE 160

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2143From the 18th century onwards, the mineral springs of the country were more extensively documented for their medicinal properties and the ailments they were thought to cure. Consequently, many medicinal scripts were published on the sulphur waters of Croft, near Darlington quoting analyses done by Peacock in 1805 showed the Croft water to be high in calcium sulfate. The waters of Askern north of Doncaster were Brewerton (1818) and Lankester (1842) both with analyses presented in grains per gallon showing the waters to be high in dissolved calcium sulfate. Edwin Lee (1854) wrote an extensive treatise on The Watering Places of England considered with reference to their Medicinal Topography with mention of the springs at Dinsdale (near Neasham) and Croft near Darlington plus the springs at Askern; his work also presents analyses of the waters. Some of these very early analyses give what we would now consider very strange combinations of elements, but in general they show high concentrations of calcium sulfate and the weights of dry residue of the analyses can be considered fairly accurate. In the late 19th century, studies of the mineral waters became based more on their geological origins with papers by Bothamley (1894) on the mineral waters of Askern and by Burrell (1896) on the water of the Dropping Well at Knaresborough. springs, but also to compare with our modern analyses. In some cases springs have disappeared due to groundwater abstraction or piping away in culverts, and these old references are the only records. under artesian pressure, it can be warmer than surface temperatures and many of the ponds associated with the springs were noted for not freezing in the winter. This characteristic is also enhanced by the amount of dissolved chemicals in the water. The Mather Pit at water is also favorable to the deposition of calcareous 1998). Planning for these problems and developing such areas requires special engineering mitigation measures (Paukstys et al., 1999; Cooper and Saunders, 2002; Sulfate-rich springs Historical records of sulfate-rich springsThe historical records of springs and wells give us an insight into the locations and approximate characters of particularly sought for their medicinal purposes with long lists of cures attributed to them. We used these early sampled using modern techniques. th century, doctors were investigating the nature of the mineral waters that they prescribed as cures for many ailments. Burrell (1896) records that examination by Dr French in 1654 of the Dropping Well water at Knaresborough found a pint weighed 10 grains more than a pint of common water. solids. In the 18th century Thomas Short described of the Dropping Well at Knaresborough and the waters of Croft Spaw near Darlington and Askern Spaw, both of which he noted as having a white sediment Figure 3. Vertical section through the Permian Zechstein Group se quence in northern England; for lithological legend see Figure 2.

PAGE 161

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE144Neasham Low Springs negligible amounts of sulfate are present. Snape, Snape Mires, and Gruntland springs near Bedale with abundant subsidence features (Cooper, 1986; Powell et al., 1992). The area is a bedrock depression caused by subsidence due to gypsum dissolution. It is fed by water under artesian pressure from the Cadeby Formation dip slope to the west (pale blue on Figure the mire numerous springs well up and are associated with peat mounds, including one about 5 m high called water. Farther east, in the middle of the mire, the sulfate Springs, which emanate from the Triassic Sherwood Sandstone in the east. Ripon, Hall Garth Ponds, near Nunwick 1986) and several other sinkholes that have been landscaped into larger ponds of different shape. The and high sulfate levels also suggest spring activity in the shown in Figure 2, with the Brotherton Formation (dark blue on Figure 2) present in the east wall of the sinkhole. Ripon springs and ponds at Ripon Parks Ripon Parks includes a number of ponds occupying very large sinkholes and a number of springs that emanate from the gypsum of the Edlington Formation and the glacial deposits. Despite the springs emerging next to small gypsum outcrops, very little sulfate was present. Similarly, the ponds situated in collapsed glacial deposits also had low sulfate levels. Ripon town Spa Field and road bridge springs These springs are located in a low part of Ripon where there are numerous peat bogs and abundant active subsidence features (Cooper, 1986, 1989, 1998, 2008a; Cooper and Saunders, 2002). Recent site investigation part of the former Spa Field near the former Spa Well. tufa, the best example of which is the Dropping Well tufa screen at Knaresborough (Burrell, 1896; Cooper activity with cemented gravels at Ripon (Thompson et al., 1996; Cooper, 1998).Modern sampling and analyses of sulfate-rich springs springs are shown on Figure 1 and detailed in Table 1. The analyses were undertaken by three of the authors Brown, 2010). Hell Kettles into Double Kettle and one by itself a little to the south of sulfate occur in the southern sinkhole, reaching 1225 to that shown in Figure 2, being in the alluvial tract of the River Tees where it cuts down through the Permian sequence. Oxen le Fields Farm borehole is located about 400 m northeast of the kettles and yielded water less rich Croft Sweet Well and Spa is near the base of the Cadeby Formation dip slope and formerly fed the spa hotel at Croft. The water is rich in Figure 4. Hell Kettles near Croft, Darlington. Double Kettle on left, Croft Kettle on right; north is to the left and the field is 190m wide; oblique view on digital terrane model. Air Photography copyright UKP/Getmapping reproduced under licence No UKP2006/01.

PAGE 162

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2145 Spring Name E N Temp tivity Analyst or reference oC Ca2+SO4 2 TDS Croft, Darlington: Croft Kettle (CK2) at 5m 428151 510844 7.75 11.2 1.680 480.9 256.9 2148.2 Miller 2006 Croft, Darlington: Double Kettle (DK1) at 4.2m 428091 510912 7.76 14.7 1.500 1097.8 Miller 2006 le Fields Borehole 511210 15.8 1.250 257.4 Env. Agen cy (Miller 2006) Neasham; Low Neasham Springs 510709 7.2 12.7 12.0 2008 Croft: Sweet Well (Spa) 427879 509186 6.7 10.7 428.0 867.0 229.0 2008 Snape Mires; Pud 427815 484458 7.6 0.161 58.7 416.4 Brown 2010 Snape Mires; The 428767 12.0 548.9 952.5 Brown 2010 Snape Mires; The 428794 20.6 0.720 189.7 Brown 2010 Snape Mires east; 1 488409 7.26 10.0 211.4 71.2 698.2 Brown 2010 Snape Mires east; 2 488409 7.26 10.0 247.9 948.4 Brown 2010 Snape Village; 2 427408 484065 7.46 9.9 0.202 92.2 45.4 24.4 408.5 Brown 2010 474707 10.2 1.700 490.7 1172.4 227.4 2001.0 Miller 2006 Ponds; east pond 474681 7.82 1.210 549.6 1146.6 Miller 2006 Ripon Parks, Dubb at 4m 474828 7.41 14.9 0.072 9.5 50.8 128.4 Brown 2010 Ripon Parks, at 1.4m 474959 14.9 0.147 66.4 42.8 214.4 Brown 2010 Ripon Parks spring 4 475472 8.27 11.6 0.220 122.9 55.7 91.5 Brown 2010Table 1. Names, locations and main sulfate, calcium and bicarbonate compositions of springs in the study area. (Continued on following page.)

PAGE 163

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE146 Spring Name E Temp tivity Analyst or reference oC Ca2+SO4 2 TDS Ripon, Spa Field (Nr Spa Well at at 1.6m 471746 2015.0 Site investi gation report Ripon, Bridge foundations Thomson et al 1996 Ripon, Racecourse gravel pit (approx locality) 469500 289.0 Thomson et al 1996 Ripon, Sharow spring 7.54 9.5 247.0 409.0 Simon War wick pers. comm. Knaresborough, Dropping Well spring 456498 6.9 10.7 681.0 298.0 2008 Knaresborough, Dropping Well tufa screen top 456518 7.6 11.7 709.0 1420.0 264.0 2008 Brotherton; Burton 448979 427269 178.7 65.8 P Murphy (1998, unpub.; high Cl.) Brotherton: Byram 448626 425162 256.1 P Mur phy (1998 unpub.; high Cl.) Askern; main lake inlet 7.1 610.0 1400.0 166.0 2008 Askern; main lake deep sinkhole (A1N) at 5m 456286 17.5 561.0 122.0 2008 Askern: Manor Bath (well) 947.0 Bothamley, 1894 (page Askern; Lake borehole (location 697.0 1172.0 Env. Agency 2008) Askern; Colliery borehole 455690 585.0 1422.0 279.0 Env. Agency 2008) Askern; Manor Well at 0.9 m 9 to 17 594.0 1085.0 Brewerton wood, 2008) Askern; Manor Well Lankester wood, 2008) Askern; Charity Well 456197 584.0 1251.0 Lankester wood, 2008)Table 1. Continued from previous page.

PAGE 164

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2147Just to the east of this the new Ripon Bridge over the in the foundation excavations and some of the local Quaternary deposits of sand and gravel were cemented with calcareous tufa, as are the gravels that occupy this part of the buried valley at Ripon (Cooper, 1986; Thompson et al., 1996). Nearby artesian springs were recorded in the bed of the river, and boreholes drilled for 1 m above river level. This water is fed from the higher ground to both the west and east as indicated in Figure 2, which is drawn through this area. Ripon has a spa hotel, but the water for this was piped in from Carboniferous strata several kilometers to the west. Sharow Spring Situated a little to the southeast of Ripon in a former related to both gypsum dissolution in the Permian Sherwood Sandstone to the east. Ripon Racecourse Water collected from the gravel pit ponds at Ripon Racecourse is high in sulfate and attests to spring activity into the gravel pit that is also fed by percolation from the nearby river. The location of the gravel pit is over the buried valley of the River Ure (Figure 2). Knaresborough Dropping Well (petrifying) The Dropping Well spring emerges from the Edlington Formation and its associated gypsum strata lying to the west of Knaresborough close to the contact with the underlying Cadeby Formation. The water is high in both sulfate and carbonate, the latter being actively deposited as a tufa ramp and screen below, in which the catchment to the spring. The spring above the tufa screen was sampled on 19th June 2008, it had rained at 5040 liters an hour. Brotherton area springs In the Brotherton area there are a number of sinkholes and several springs that feed into a small lake (Murphy, 2000): however, the water emanating from the lake has low sulfate levels (Table 1). Farther south, on similar Permian geology, a spring yields a moderate amount of sulfate, but also high levels of chloride, which may be anthropogenic or possibly related to chloride salts in the Permian sequence. Askern springs and lake Askern today has an ornamental lake next to the road in the middle of the village. The air photographs and the sinkholes in the bed of the lake (Figure 6), but although sulfate levels were high in the lake they appeared to be than artesian water coming up from the sinkholes. This might be because of groundwater pumping for mining that in the past there were more ponds to the south of Pits that had very high sulfate levels and which did not freeze in the winter (Brewerton, 1818; Lankester, 1842; Bothamley, 1894). Figure 5. The Dropping Well (petrifying spring) at Knaresborough.

PAGE 165

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE148ground that the most aggressive dissolution is occurring and where the most active subsidence is happening. Consequently, where the valleys of the rivers Tees, Ure, Nidd, Wharfe, and Aire cut through the Permian springs and sinkholes. The distribution of the sinkholes is controlled by several of the gypsum onto the underlying Cadeby Formation transition from gypsum to anhydrite. This transition is largely controlled by the reduction in groundwater gypsum to anhydrite transition occurs at a depth of about of gypsum is present in the Edlington Formation and another 10 m or so above it in the Roxby Formation and breccia pipes cannot generate enough breccia to features at the surface (Cooper, 1986). more widespread than previously recognized. It also shows that sulfate concentrations similar to those recorded by Cooper (1986) commonly occur. The geology along the Permian escarpment and gypsiferous rocks is similar situation, with buried valleys cutting the sequence, is also repeated in several places (Figure 2). Cooper (1986, 1988) suggested models for the amount of gypsum dissolution in the Ripon area based on groundwater with in Ripon. The calculation presented by Cooper (1986) suggested an annual amount of around 120 cubic meters of gypsum being dissolved every square kilometer. The escarpment, then similar annual amounts of gypsum are being dissolved along most of the gypsum belt. indicate areas where there is very poor groundwater quality due to the high presence of sulfates. They also indicate areas where groundwater abstraction should InterpretationThere is a close relationship between the presence of sinkholes that affect the Permian gypsiferous sequence dominated by the Cadeby Formation dolomite sequence that underlies the gypsum formations and which has a large dipslope to the west of the main gypsiferous units. Water passing through this dolomite sequence is rich in dissolution resulting in numerous sinkholes (Klimchouk, dolomite at its contact with the gypsum (Bischoff et al., 1994) and changes it to a weak calcitic mesh that can also break down and cause subsidence. This process also helps to release calcium carbonate that is redeposited as tufa and tufa cements. The water analyses of the springs show different mixtures of sulfate and carbonate waters that appear to be controlled by the positions of the springs with respect to the Cadeby Formation dip slope, the low ground with the gypsum, and the higher ground formed by the overlying Triassic sandstone. Artesian water is present wherever valleys cut through the Permian sequence. It is in these locations in the lower Figure 6. Askern Lake showing the sinkholes within it as deeper water in the middle, southeast and southwest; west (left) edge of lake is 195 m long. Air Photography copyright UKP/Getmapping reproduced under license No UKP2006/01.

PAGE 166

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2149Brown DS. 2010. Karst and groundwater chemistry associated with the Magnesian Limestone and gypsiferous rocks of the Ripon to Bedale area of North Yorkshire [MSc Thesis]. Leeds (UK): Burrell BA. 1896. A history of the Dropping Well at Knaresborough with the analysis of the water. resulting from the dissolution of Permian gypsum in the Ripon and Bedale areas, North Yorkshire. dissolution of Permian gypsum in the Ripon area; its relevance to mining and water abstraction. In: Bell subsidence caused by Permian gypsum dissolution at Ripon, North Yorkshire. Quarterly Journal of the dissolution of Permian gypsum in England: geology, investigation and remediation. In: databasing and use of information about building damage caused by subsidence and landslides. of karst geomorphology for planning, hazard construction across gypsum karst in England. investigation of groundwater resurgences from gypsiferous Permian strata, N.E. England [MSc Thesis]. Leeds (UK): The University of Leeds. 106 p. be discouraged for purposes such as irrigation, both because of the acceleration of dissolution such as that described by Cooper (1988), but also because of the prone area some farm irrigation boreholes have been implicated as the possible causes of subsidence in the immediate vicinity to them. Currently there is an likelihood of enhanced dissolution mean that within considered (Cooper et al., 2011).AcknowledgementsThe authors thank the following for their help and All the landowners and farmers who have given access, Estate. Adrian Sayers for access to the Dropping Well. access to Ripon Parks. Simon Warwick of the Lower Ure Conservation Trust, for chemical analyses of springwater. Alison Parker is thanked for sailing her canoe at numerous localities and for collecting samples anonymous reviewer are thanked for their editing which Survey (NERC). ReferencesBischoff JL, Julia R, Shanks (III) WC, Rosenbauer RJ. Dermatologist of the Millennium. Journal of the water of Askern, including a description of the village, history of The Spaw. Doncaster: W. Sheardown.

PAGE 167

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE150 Kingdom Croft, near Darlington. London (UK): F. Johnson and W. Brown. Willan R. 1786. Experiments and observations on the Yorkshire. London (UK): F. Johnson and W. Brown. Northumberland and Durham. Transactions of the prediction, and mitigation of sinkhole hazards Klimchouk A. 1996. The dissolution and conversion of gypsum and anhydrite. In: Klimchouk A, Lowe, D, Utilising seasonal variations in hydrogeochemistry for subsidence hazards: an example from Co. Durham, for investigating subsidence problems potentially Lankester E. 1842. An account of Askern and its mineral springs; together with a sketch of the natural history, and a brief topography of the immediate neighbourhood. London (UK): John Churchill. Lee E. 1854. The watering places of England, considered with reference to their medical topography. London (UK): J. Churchill. Miller C. 2006. What can subsidence hollows tell us about hydrogeological processes? A case study based on Darlington and Ripon, NE England [MSc Thesis]. Leeds (UK): University of Leeds. 111 p. strata east of Leeds. Proceedings of the Yorkshire for gypsum geohazards in Lithuania and England. history of the mineral waters of Derbyshire, Lincolnshire, and Yorkshire, particularly those of Scarborough: Wherein, they are carefully examined and compared, their contents discovered and divided, their uses shewn and explained, and an account given of their discovery and alterations; together with the natural history of the earths, minerals and fossils through which the chief of Assessment of subsidence arising from gypsum dissolution. In: Technical Report for the Department

PAGE 168

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2151IntroductionThis study examines the surface and subsurface geology northwestern Oklahoma (Figure 1). The study area embraces nearly 400 km22 Watonga wind turbines would be constructed. The study focuses on the thickness, distribution, structure, and depth of gypsum beds in the Permian Blaine Formation that susceptible to being partially or totally dissolved by Abstract form karst features identical to those associated with limestones and dolomites. Investigations in Blaine County, in northwestern Oklahoma, evaluated potential problems that subsidence due to gypsum karst may collapse of a wind turbine is clearly unacceptable, and minor settlement could also be a risk. Differential foundation could lead to the turbine tilting out of tolerance, requiring remedial repairs. 45 m below ground level. The Blaine Formation here is about 29 m thick: it consists of four gypsum beds, each shales. The Blaine is overlain by the Permian Dog Creek Shale and by unconsolidated Quaternary sands, clays, and gravels that may obscure karst features. Field studies, Although lacking direct evidence of karst in or near for subsidence due to dissolution of shallow gypsum. Additional mitigation of this risk can be achieved by placing wind turbines at sites where the gypsum beds are deepest: we believe that where gypsum is 25 m below ground level, or deeper, the risk related to gypsum karst is low. Placing turbines at sites where gypsum beds are less than 25 m deep would pose a medium or high risk. To minimize this risk, a map was prepared showing areas of low, medium, and high risk, related to potential gypsum karst. GYPSUM KARST AND POTENTIAL RISK IN SITING WIND Kenneth S. JohnsonOklahoma Geological Survey, 1321 Greenbriar Dr., Norman, OK 73072, USA, William J. BangsundBarr Engineering Co., 4700 W. 77 Street, Minneapolis, MN 55435, USA, BBangsund@barr.comNeal A. HinesENVIRON International, Austin, TX, Figure 1. Map of western Oklahoma showing location of Watonga Wind-Power Project and the windenergy-resource potential at 50 m above ground level.

PAGE 169

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE152groundwater, and to developing karst features such as caves, sinkholes, and underground water courses that commonly are also found in limestones and dolomites dip gently to the southwest where they underlie all parts karst in western Oklahoma, the distribution and depth of the various gypsum beds is important in the siting of wind turbines because of the potential for collapse due to subsurface bedrock dissolution. A wind turbine located above a sinkhole, cave, or other karst feature could become unstable if there is any settlement of the ground. The depth to the top of the uppermost gypsum in the Blaine Formation ranges from 10 to 45 m, and in most of where the depth is 25 m or more, such a site would pose a low risk for problems related to gypsum karst, because if there is karst in these Blaine gypsum beds below that depth it is unlikely that subsidence would reach up to and impact the land surface. We believe that karst in the Blaine gypsums at shallower depths would pose a higher risk for a wind turbine. generated energy (Figure 1). According to the Oklahoma Department of Commerce (personal correspondence, operating in western Oklahoma, with a generating 900 megawatts.Geologic SettingOutcrops of Permian rocks in the Watonga area include gypsum and shale beds of the Blaine Formation, overlain in turn, overlain by Quaternary terrace deposits and alluvium. The Blaine Formation is about 29 m thick in the Watonga area, and individual beds of white gypsum range from 0.6 to 4 m thick (Fay and others, 1962; Fay and several thin beds of dolomite. The thickest gypsum formation. The Blaine Formation dips gently to the south The Dog Creek Shale generally is 50 to 58 m thick in Blaine County (Fay and others, 1962; Fay, 1964), but the to 15 m of shale remains. The Dog Creek is principally several thin beds of gypsum (0.5 to 2 m thick) in the Above the Dog Creek Shale are Quaternary terrace deposits and alluvium deposited by the North Canadian River. that generally are 10 to 20 m thick. All of the proposed wind turbines will be sited directly upon these Quaternary deposits where they overlie the Dog Creek Shale.Methods of StudyA comprehensive review of the geologic literature to determine if karst features have been reported in Blaine County and nearby parts of Oklahoma was carried out. possible karst features. Nearby outcrops of gypsum were Figure 2. Geologic map of the Watonga WindPower Project Area in Blaine County (Fay, 2010). Blaine Formation gypsum beds dip gently to the south and southwest (the strike and dip symbols) beneath the Dog Creek Shale and Terrace Deposits that mantle most of the Project Area.

PAGE 170

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2153 categories map (discussed below) to help minimize the possibility that gypsum karst will have an adverse Literature ReviewA State geologic map was prepared by Miser (1954), followed by detailed maps of the surface geology of Blaine County and the Watonga area by Fay and others (1962) and Fay (2010). These later maps show the outcrop area of gypsum beds in the Blaine Formation, as well as outcrops of the overlying Dog Creek Shale and 2). In addition, Fay (1964) discussed the stratigraphy and character of the Blaine Formation and associated strata throughout northwestern Oklahoma. These three studies by Fay were most valuable for understanding the geology beds elsewhere in western Oklahoma by Myers and caves in the Blaine Formation of western Oklahoma have been examined, mapped, and described by John Oklahoma City: these reports have been released in Oklahoma Underground, which is a serial publication Additionally, electrical logs (also called geophysical logs) of many petroleum test wells drilled within and and this was used to determine the depth to the top of the Figure 3. Exposure of the Shimer Gypsum Bed (white) at top of Blaine Formation and the overlying Dog Creek Shale (redbeds). Exposed in US Gypsum quarry about 20 km north of Watonga. Gypsum here is 4.5 m thick. Figure 4. Cross section showing thickness of gypsum beds of the Permian Blaine Formation in outcrops just east of the Watonga Wind-Power Project Area (after Fay and others, 1962). View looking to the southwest.

PAGE 171

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE154collapse. The small ephemeral ponds present in some of the depressions appear to result from local runoff (from precipitation) into areas that have no external drainage.Field ExaminationField studies have discounted the presence of potential sinkholes in terrace deposits, as suggested in the Area were examined, and no evidence of karst features or voids was found. This observation agrees with the (1964), and with recent discussions with John and Sue Bozeman (personal communication, 2010). The one exception is the Foley sink, described above, that is 6.5 depression in the ground.Electric-Log StudyAbout 400 petroleum tests have been drilled in, and wells were examined by Johnson, but only 20 of them contained data about the Blaine Formation or other rock layers present in the top 100 m of the borehole. Recognition of gypsum beds and associated rock types on electric logs is well established (Alger and Crain, 1966), and the senior author has conducted many studies using various types of well logs to identify, correlate, and map gypsum beds in the subsurfacesome of the studies are in public documents (Johnson, 1967, 1985, On each of the 20 useable electric logs in the study area, individual gypsum beds (and interbedded shale units) northeast. Correlation from outcrops to the electric logs shows that the overall thickness of the Blaine Formation, as well as the thickness of individual units, is quite the elevation of the top of the uppermost gypsum (the on gypsum caves of western Oklahoma, John and Sue Bozeman reported that they are not aware of any caves in communication, 2010). gypsum beds here are quite thin and are interbedded with access to the gypsums. The one feature showing possible evidence of gypsum karst is the Foley sink, that formed in 1957 (Fay, 1958). The sink formed in Quaternary Formation. Originally about 15 m wide and 5 m deep, have been carried out to determine the true cause of the sinkhole. Fay also mentioned several older sinks located about 225 m northeast of Foley sink. A statewide, general assessment of potential karst terrains was preliminary, with only general data discussing the various potentially karstic rocks (limestone, dolomite, gypsum, and salt), and showing the general outcrop area of Blaine gypsums in northwest Oklahoma.Aerial-Photo StudyPersonnel at Barr Engineering in Minneapolis, MN, November 1990; and April 1979). A number of linear drainage features are evident that extend in directions features are related to sinkholes or voids. depressions are present in the Quaternary terrace deposits that blanket the area. Inasmuch as these terrace deposits consists of many sand dunes that create a hummocky glance, may appear to be related to sinkhole development or collapse structures, but there is no evidence that any of the small depressions have, in fact, resulted from ground

PAGE 172

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2155 karst can pose risks to wind turbines. If gypsum karst subsidence or settlement of the overlying Dog Creek Shale and Quaternary sediments. This could result in tilting of, or damage to, the foundation of a turbine built cm across a turbine foundation that is 15 m wide could cause the turbine to tilt and require remedial repairs. karst features in outcrops in the Watonga area, or in the to further reduce the potential risk of gypsum karst on at sites where the top of the Blaine gypsum beds is at least 25 m below the land surface.Structure-Contour Map and Cross Section plotting the elevation of the top of the Blaine Formation Results of StudyAs a result of the foregoing studies, it is possible maps to select sites for construction of wind turbines where the risk is low, with regard to possible gypsum karst.Potential Karst RisksEvaporite rocks, mainly gypsum (or anhydrite) and rock salt, are the most soluble of common rocks. They can be dissolved readily to form caves, sinkholes, documented in the Watonga area in the literature, probably because the gypsum beds here are quite thin shales that inhibit groundwater access to the gypsum. Evidence of gypsum karst would include surface and disappearing streams, springs, collapse structures, and while drilling through gypsum beds. None of these features have been found. Figure 5. Electric logs of two wells showing depths (in feet) and log characteristics of gypsum beds in the Blaine Formation in (and near) the Watonga Wind-Power Project.

PAGE 173

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE156shows the approximate elevation of the uppermost from the elevation of the land surface to determine the in the shallowest and thickest gypsumthe Shimer A cross section (Figure 7) shows subsurface relationships between the Blaine gypsums and the overlying Dog Creek Shale and the Quaternary terrace deposits and based on outcrop information from Fay and others (1962 and Figure 4), and on the structure map on top of the Area where turbines might be constructed, they would be located upon Quaternary terrace deposits or alluvium. Immediately beneath these Quaternary deposits, remnants of the lower Dog Creek Shale are present and would separate the Quaternary deposits from the Shimer barrier that provides added protection against karst by sediments into underlying gypsum beds.Risk MapA schematic cross section shows possible karst conditions in gypsum, related to the depth of gypsum below the land on the outcrop and on electric logs of petroleum tests. The Blaine Formation dips gently and uniformly to the 0.1 to 0.2 degree). This map is most useful because it Figure 6. Structure-contour map showing elevation on top of the Blaine Formation gypsums beneath the Watonga Wind-Project Project Area. Contour interval is 6 m (20 feet). Cross section AB shown on Figure 7. Figure 7. Cross section AB through the Watonga Wind-Power Project Area showing the dip of strata and depths of gypsum beds of the Blaine Formation. Cross section based upon: Figure 6; outcrop data; two borings by Barr Engineering Co. (borings #87 and #76); and electric logs of oil-well tests (including Texaco, Horsley Unit B No. 1, shown in Figure 5). Line of cross section shown on Figure 6.

PAGE 174

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2157Being a highly soluble rock, gypsum is readily dissolved to form karst features identical to those associated with limestones and dolomites. Settlement of a turbine karst features could lead to tilting of the turbine out of tolerance and require remedial repairs. Therefore, we examined the local geology to evaluate whether gypsum surface (Figure 8). It is reasonable to expect that placing wind turbines at sites where the gypsum beds are 25 m deep, or deeper, would pose a low risk related to gypsum karst subsidence. Sites where the gypsum is more than 6 m, but less than 25 m, below the surface are sites of medium risk, and sites where the gypsum is 6 m, or less, below the surface are sites of high risk. a risk map (Figure 9) was prepared by Barr Engineering of the land surface with the elevation of the Blaine Formation (Figure 6). This established the depth to the top of the conservative. Although there is no evidence of gypsum karst in the Watonga area, it is possible that dissolution At least 25 m of sand, clay, gravel, and shale between development and collapse structures, and turbine sites in those areas would be at low risk. Using these criteria and the risk map, it will be possible to locate any proposed Therefore, most of the area is considered to pose a low risk for problems related to gypsum karst. Summary and Recommendations Figure 8. Schematic cross section showing low-, medium-, and high-risk karst zones, which are related to the depth of possible karstic gypsum in northwestern Oklahoma (modified from Johnson, 2003c). Figure 9. Risk categories at Watonga Wind-Power Project, based upon depth to the Shimer Gypsum at top of the Blaine Formation. Contour lines from Figure 6.

PAGE 175

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE158Fay RO. 1958. A recent sink hole in central Blaine County, Fay RO. 1964. The Blaine and related formations of northwestern Oklahoma and southern Fay RO. 2010. Preliminary geologic map of the and mineral resources of Blaine County, Bulletin 89. 258 p. Johnson KS. 1967. Stratigraphy of the Permian Blaine Formation and associated strata in southwestern Oklahoma [PhD dissertation]. Champaign (IL): University of Illinois. 247 p. Johnson KS. 1985. Structure contour map and Scale 1:125,000. [Blaine Formation, southwestern Oklahoma]. NSS Blaine gypsumdolomite aquifer, southwestern during Salado time in the Wink area, Winkler County, Texas. In: Love DW, and others, editors. Carlsbad region, New Mexico and west Texas. Annual Field Conference. p. 211. United States, In: Johnson KS, Neal JT, editors. problems in the United States. Oklahoma Blaine Formation and associated strata of western Oklahoma. In: Johnson KS, Neal JT, editors. problems in the United States. Oklahoma The Blaine Formation is about 29 m thick, and consists of four gypsum beds, each 0.6 to 4 m thick. The top of the gypsums is 10 to 45 m below ground level in the Creek Shale and by unconsolidated Quaternary sands, clays, and gravels that may obscure karst features. Field that there is no direct evidence of gypsum karst in the Examination of electric logs of petroleum tests enabled us to prepare a cross section and risk map showing the depth to the top of gypsum, and enables us to determine sites where the risk due to gypsum karst would be low. We believe that those locations where the depth to gypsum the risk would be moderate where the depth to gypsum is at those locations where gypsum is less than 6 m deep. We suggest further evaluation of the potential for gypsum karst in the Watonga area by considering the following actions: 1) examine cores drilled through the Shimer evidence of dissolution in that shallowest and thickest gypsum bed; 2) conduct a survey of landowners within collapse structure is reported by landowners, it should survey of landowners where the Blaine Formation crops gypsum dissolution.References with electrical well logs. In: Rau JL, editor. Second Oklahoma gypsum caves. In: Johnson KS, Neal JT, editors. Evaporite karst and

PAGE 176

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2159 terrains in Oklahoma for avoidance of engineering and environmental problems. In: Johnson KS, Neal JT, editors. Evaporite karst and engineering and environmental problems in Martinez JD, Johnson KS, Neal JT. 1998. Sinkholes in Survey. Scale 1:500,000.

PAGE 177


PAGE 178

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2161Many of the evaporite intervals lie within or between within them are of concern to oil and gas exploitation. In this paper we discuss two areas of interest: Wyoming and South Dakota, and its extension into the subsurface in the Williston Basin of North Dakota, and 2) a paleokarst horizon in the upper Madison Limestone of Wyoming and South Dakota, and its possible extension into the Williston Basin. Oil and gas potential in the Williston Basin of has recently attracted considerable attention, with particular interest in the organic shales of the Bakken Formation of Devonian age. With the use of hydraulic fracturing to increase the release of oil and gas from these source rocks, other stratigraphic units in the and Triassic age, have been touted as possibly being additional unconventional oil plays in the subsurface of North Dakota (LeFever, 2011). In contrast to the reducing environment of deposition of the dark mi) from the nearest outcrops in the northern Black North Dakota, so one of us (Epstein) was requested to discuss the geologic characteristics of that formation of South Dakota and Wyoming.The Spearfish Formation and evaporite karst in the Black Hills is cored by Precambrian age metamorphic rocks that are Abstract been touted as another Bakken Oil Boom for North Dakota. The senior author, totally uninformed about the subsurface geology of North Dakota, was requested by INFOCAST* to discuss petroleum potential in the South Dakota. That request was extended to a discussion four formations ranging from Pennsylvanian to Jurassic age. Dissolution of these rocks, which has resulted in sinkholes, caves, springs, breccia pipes, and subsurface was formed during the Early Tertiary, and continues has been documented to have occurred many times in the Tertiary. Reported subsurface collapse has affected rock characteristics, including local structure, fracturing, of evaporite karst in the Williston Basin, as well as in should be of concern to any oil exploration efforts, as well as for surface infrastructure development such as IntroductionLarge areas of the United States are underlain by evaporite deposits that mainly consist of gypsum, has created abundant karstic features at the surface occurred during intervals in the geologic past, thereby creating many intervals of paleokarst throughout Jack B. Epstein, Daniel H. DoctorU.S. Geological Survey, 12201 Sunrise Valley Drive, MS 926A, Reston, Virginia, 20192,, ________________________________________________________________________ * (accessed 11/5/2012)

PAGE 179

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE162 interbedded gypsum abundant at varying horizons (Figure beds and ripple laminations are not uncommon. Salt casts have been reported (Sabel, 1984). These sedimentary as much as much as 140 m (460 ft) across (Figure 7). The Vore Buffalo Jump shown in Figure 7 is rimmed bottom of the sinkhole, which is probably less than 15 m (50 ft) above the underlying Minnekahta Limestone. The Minnekahta crops out along the service road about one mile to the west where a 1.2 m (4 ft) thick bed of gypsum about 140 m (460 ft) across, occurs at the base of the sedimentary rocks of Paleozoic and Mesozoic age from Mississippian to Jurassic (Figure 4). The Limestone core, comprises rocks of the Madison Limestone (Pahasapa Limestone of other reports) and the Pennsylvania age Minnelusa Formation. Most surface exposures of the Minnelusa contain abundant karst features resulting from dissolution of anhydrite at depth. The Madison contains brecciated rocks resulting partly from similar gypsum Formation. Resistant sandstone of Cretaceous age forms outer physiographic perimeter. Below those sandstones of Jurassic age. Much of the information about the karst slightly more than 240 m (800 ft) in thickness, consisting Figure 1. Map showing the distribution of outcropping and subsurface evaporite rocks in the United States and areas or reported evaporite karst. The 32.5-in. mean-annualprecipitation line approximates a diffuse boundary between eastern and western United States karst terrains (from Epstein and Johnson, 2003).

PAGE 180

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2163 m (4 ft) thick bed of gypsum which in turn overlies the Minnekahta Limestone; both dip gently underneath the sinkhole. The limestone outcrop closest to the sinkhole contains a small collapse opening in a blind valley. A smaller sinkhole (inset of Figure 8) opened in 1985, and was observed by local ranchers who heard running water in a cavern that extended horizontally beyond the limits is deemed to be partly responsible for dissolution and the observed depth of the collapse. Figure 2. Geologic map of parts of North Dakota, South Dakota, Wyoming, and Montana. The outcrop area of the Spearfish Formation in the Black Hills is located approximately 530 km (330 mi) from the Spearfish oil play in Bottineau County, North Dakota. The Black Hills is a domal structure cored by Precambrian rocks in the center (orange). It is encircled by uplifted and erosionally breached Paleozoic to Cretaceous rocks. Red beds and evaporite deposits of the Spearfish Formation (red) form a valley nestled between the hogback of Cretaceous age sandstones and plateau of underlying carbonate rocks, including the Madison Limestone of Mississippian age. The surface rocks in the western part of North Dakota are dominated by Tertiary rocks (solid green) overlying Cretaceous sediments to the west. Figure 3. Generalized diagram showing the geology and geomorphology of the Black Hills, Wyoming and South Dakota. Modified from Strahler and Strahler (1987), with permission. Figure 4. Stratigraphic column showing distribution of gypsum and anhydrite in the northern Black Hills.

PAGE 181

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE164wells. Perched water tables with springs below zones of gypsum in fractured red beds are present in several bedding is not disturbed by gypsum expansion, and the consist of chaotic blocks in a matrix of rocks of the enclosing formations and they commonly extend below as much as 250 m (820 ft) thick and contains several intervals of gypsum in the lower 60 m (200 ft). Anhydrite, which probably was the original form of calcium sulfate percent expansion when hydrated to form gypsum. As a result, the gypsum is commonly highly folded (Figure 9A, B irregular fractures as thin veinlets, generally less than 2 (Figure 9C). Thus, the lower 60 m (200 ft) or so of the Figure 5. Typical exposures of the Spearfish Formation in the Black Hills, South Dakota and Wyoming. (A) White gypsum interbedded with red shale and siltstone about 15 km (10 mi) southeast of Newcastle, Wyoming. (B) Planarbedded red beds overlain by shales of the Stockade Beaver Member of the Sundance Formation (dashed line), one mile northeast of Beulah, Wyoming. A bed of gypsum, about 3 m (10 ft) thick, comprising the entire Gypsum Spring Formation a few miles to the east of this locality, is missing here. This is the closest exposure of the Figure 6. Small solution openings near Beulah, Wyoming. (A) Circular sinkholes about one meter across, with remnant gypsum present in thebottom. (B) Soil collapse along a solution-widened joint that extends downslope towards a creek. Spearfish Formation to the Spearfish oil play in northern North Dakota. (C) Fining-upward sequences of very fine-grained sandstone, siltstone and shale in Devils Tower National Monument, Wyoming. (D) Spearfish mudcracks 13 km (8 mi) northwest of Spearfish, South Dakota.

PAGE 182

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2165 Figure 7. Two sinkholes located 5.6 km (3.5 mi) west of Beulah, Wyoming, and immediately north of Interstate 90. (A) Sinkhole with a laterforming smaller internal sinkhole at arrow. (B} The Vore Buffalo Jump is an 18-m (60 ft)-foot deep sinkhole. More than 300 years ago the Native Americans stampeded bison over the steep rim into the hole. As many as 20,000 beasts were butchered for food. The site is now a major archeological dig by the University of Wyoming (http:// Accessed 11/13/2012/). Figure 8. Karst features located immediately northwest of the Vore Buffalo Jump shown in figure 7. A deep sinkhole (see inset photo) lies within a larger, 140 m (460 ft) wide, flat-bottom sinkhole in center of the photo. Another sinkhole lies in the lower right corner. A conspicuous gypsum bed lies immediately above the Minnekahta Limestone and also underlies the sinkholes. Another gypsum bed lies above and to the north of the sinkholes.

PAGE 183

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE166entirely the result of dissolution of the local gypsum, but additionally due to removal of evaporites at greater depth. The top of the Minnelusa Formation lies about 45 m (150 consists of dolomite, sandstone, and shale with anhydrite being prevalent in the middle part. The anhydrite is mostly absent in outcrops due to removal by solution in the subsurface. The solution of anhydrite and consequent formation of voids in the Minnelusa at depth resulted in foundering and fragmentation of overlying rocks, thereby producing extensive disruption of bedding, a regional collapse breccia, many sinkholes, and breccia pipes and pinnacles (Figure 11). The collapse breccia consists of angular clasts of all the local rock types in a sandy matrix that is generally cemented with calcium carbonate. The vuggy secondary porosity of the collapse breccia, along with the porous sandstone makes the upper half of the Proof that collapse subsidence extended upward from the within the thin intervening Minnekahta Limestone as seen near the sinkholes in Figures 7 and 8. Along Redwater Creek the Minnekahta is exposed in a 600 m (2000 ft) low cliff where numerous sinkholes are present and the unit is extensively brecciated, and the underlying Opeche Shale is also disrupted (Figure 12). This observation suggests that collapse structures in the Minnelusa extend upward through the bottom of the sinkhole is near or at the base of the formation, suggesting that the solution collapse is not Figure. 9. (A) Contorted gypsum resembling tectonic folds in the Spearfish Formation, Red Valley of the southwestern Black Hills, southeast of Newcastle, Wyoming, resembling tectonic folds. (B) Disrupted gypsum near Beulah, Wyoming. (C) Gypsum-filled fractures extending above and below contorted gypsum bed in the uppermost Spearfish Formation, along State Route 71 about 15 km (9 mi) southwest of Hot Springs, South Dakota. Figure 10. Resistant pinnacle-forming breccia pipe in the Spearfish Formation, about 24 km (15 mi) west of Hot Springs, South Dakota. The breccia pipe occurs about 100 m above the dissolved anhydrite in the Minnelusa Formation. Some sandstone clasts were derived from the overlying Sundance Formation, more than 30 m (100 ft) above.

PAGE 184

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2167the Minnelusa, as well as in the underlying Madison Limestone, has caused collapse features in the Minnelusa, commonly seen on canyon walls. Breccia pipes and some sinkholes have extended upwards into extensive disruption in the Minnekahta precedes Karst development evolved continuously after the of rainfall and snow melt at high altitudes of the central the rimming sediments, including the Madison and Minnelusa. Dissolution and removal of the anhydrite in the Minnelusa, progresses in a downdip direction While breccia pipes and sinkholes may extend upward from the Minnelusa Formation into overlying formations, may not necessarily be directly connected to pipes in the underlying Minnelusa. Abundance of sinkholes in the cavernous passageways developed as gypsum dissolved at the base of the formation while collapse extended upwards into the underlying Minnekahta Limestone. and relationships of surface and subsurface karst features within and between the Minnelusa Formation, Figure 11. Collapse structures in the Minnelusa Formation. Top photo is split into two segments at the dashed match line, showing the upper half of the Minnelusa Formation in Redbird Canyon, South Dakota, 16 km (10 mi) southeast of Newcastle, Wyoming. All beds are disrupted and mostly brecciated. Sinkholes and caves dot the canyon wall. Bottom photo shows Cold Brook Canyon, just west of Hot Springs, South Dakota. The upper half of the formation is brecciated with through-going breccia pipes (arrow). The lower half of the formation is not brecciated. Collapse was due to removal of many tens of meters of anhydrite in the subsurface prior to exposure of the canyon wall within the covered slope in the middle of the photo.

PAGE 185

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE168Paleokarst in the Madison GroupThe Charles Formation, which forms the upper part of the largely limestone and dolomite succession of the Madison the canyon of the Bighorn River in eastern Wyoming (Figure 15). It has an areal distribution extending across the Williston Basin in North Dakota and far to the southwest into the Bighorn Basin of Wyoming and Montana (Figure this paleokarst horizon is exposed within numerous cave systems such as Wind Cave and Jewel Cave. This extensive intrastratal paleokarst horizon contains abundant evaporite of this interval show that it formed from groundwater grained silt and clay that is cemented with late calcite cement (Palmer and Palmer, 2008). Characteristics of this paleokarst horizon are similar to those described above for breccias pipes, and resurgent springs that are developed at the site of (Figure 14). Minnelusa moves downdip and radially away from the center of the uplift. As a result, resurgent springs will dry up and new ones will form with the downdip migration of the dissolution front as the geomorphology of the on canyon walls (Figure 11) attest to the former position the anhydrite in the Minnelusa, but it is present in the subsurface, therefore a transition zone must be present where dissolution of anhydrite is currently taking place. Consequences of this model suggest that present springs should develop in a downdip direction within the of anhydrite in the Minnelusa probably began soon after continues today. Figure 12. Collapse sinkhole (between dashed line) in a zone of brecciated Minnekahta Limestone, along Redwater Creek, 1.6 km (1 mi) southeast of the sinkholes seen in Figure 7. Inset shows details of the breccia.

PAGE 186

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2169Figure 13. Generalized cross section showing common karst features in the Beulah area, Wyoming. Features include sinkholes, cavernous gypsum in the basal Spearfish Formation, outcropping sinkholes in the Minnekahta Limestone, breccia pipes and dissolution zone at the top of the Minnelusa Formation, and artesian flow direction from the Minnelusa Formation and the Madison (Pahasapa) Limestone. Where the potentiometric surface is below ground level, sinkholes are dry; where it is above ground level, sinkholes contain emergent springs. Figure 14. Generalized cross section showing the downdip migration of the anhydrite dissolution-front in the Minnelusa formation. Cox Lake and other resurgent springs are near the position of the dissolution front. As the front moves downdip with continued solution of the anhydrite, and as the Black Hills is slowly lowered by erosion, these sinkholes and related collapse structures are left abandoned on canyon walls (see Figure 3).

PAGE 187

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE170Figure 16. Generalized west-east structural stratigraphic cross section from the Bighorn Mountains, Wyoming, to northeastern North Dakota. Modified from Peterson, 1988. Green color indicates units with significant evaporites, especially salt (compare with Figure 17). Blue color indicates the Bakken Formation. Figure 15. Exposure of paleokarstic breccia due to evaporite dissolution in the upper Madison Limestone. Exposure is within the Bighorn Canyon in eastern Wyoming. The outlined reddish layer contains abundant brecciated clasts of dolomite with vugs of anhydrite and gypsum. The red matrix is fine silt and clay cemented with late calcite. Layer thickness varies between 3 to 6 m (10 to 20 ft).

PAGE 188

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2171The modern regional flow path of formation waters is generally towards the east and northeast, emanating from the recharge areas in the Bighorn and Big Snowy Mountains and the Central Montana 1996, Figure 10). The presence of this paleokarst horizon in the Madison today due to evaporite solution at depth have occurred collapse features can provide insight for understanding possible geologic structures and characteristics within deeply buried evaporite horizons that are experiencing solution in the Williston Basin. Salt dissolution in the Williston BasinThe Williston Basin is a large intracratonic structural downwarp in North Dakota and eastern Montana, extending into Canada, and lying immediately north of 1). Deposition was intermittently continuous since the Middle Cambrian, with a total accumulation of about Sedimentation rate kept pace with basin subsidence rate, so deposition mainly occurred in shallow water and in recurring evaporite basins. Transgressive and regressive cycles are numerous; pinchouts and disconformities are abundant (Peterson and MacCary, 1987, Figure 5). More than a dozen periods of evaporite deposition characterize the stratigraphic section (Figure 17). minor salt beds within the basin. In contrast to the rocks The transition from salt in North Dakota to anhydrite, gypsum, and minor interbedded dolomite in many Black in South Dakota and Wyoming. Peterson (1995) presented a stratigraphic chart showing the occurrence of 12 hydrocarbon source beds and 15 reservoir rocks in the Williston Basin. In the Williston Basin evaporites are commonly interbedded with petroleum source rocks. These evaporites may form seals and traps, or their karstic effects may disrupt stratigraphic continuity and disrupt the integrity of a reservoir. Subsurface dissolution features similar to those in the Black waters, due partly to dissolution by fresh water in many of the Paleozoic aquifers. Figure 17. Chart comparing the Devonian-Jurassic stratigraphy in the Black Hills, South Dakota, and Williston Basin, North Dakota. Formations with significant evaporites, especially salt, and some with more than one horizon, are in green (compare with Figure 16). Modified from Fahrenbach and others, 2010. See also: figure 1 in LeFever (2012); figure 2 in Burrus and others (1996); and figure 4 in Peterson (1988).

PAGE 189

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE172was also intermittent, generally with long periods of no perceptible dissolution LeFever and LeFever, 1995). cores in the Prairie Formation in Saskatchewan by the very poor sorting of angular rock fragments. The material in the lower part of the brecciated units are Breccia deposits extend laterally beyond the present limits of salt, suggesting a migrating salt front as Dissolution of evaporites (salt) in the Williston Basin requires a hydrostatic head that allows ground water to dissolve and carry the salt in solution upwards, as well as below the salt (Parker, 1967). Breccia pipes (dissolution) may extending vertically upward from several salt horizons and terminating at various unconformity horizons affect the porosity of surrounding rocks by either creating solution voids or depositing cements. Breccia pipes in ft) upwards from their source beds in the Minnelusa, suggesting considerable vertical interconnection in the sedimentary succession. Subsurface solution collapse can affect important rock characteristics for petroleum development, such as fracturing, porosity, and permeability. Several studies have linked the dissolution of the thick Silurian to Jurassic age salt deposits in the Williston Basin to petroleum reservoir development. Salt and evaporite dissolution can also form stratigraphic and unconformity traps (LeFever, 2011, 2012) and salt collapse, halite plugging, incomplete anhydrite or residue seals, and dolomitization have affected or destroyed potential reservoirs (LeFever and LeFever, 1995). The age of salt dissolution may be determined by the age of overlying sedimentary deposits that have been Salt thickness throughout the Williston Basin locally can be quite variable in thickness, apparently related to uneven dissolution due to several factors. Salt dissolution can be recognized by comparing thinning of salt in isopach maps in relation to overlying thicker compensating sediments of the Dunham Salt in the Jurassic age Piper Formation occurs in isolated lenses due to dissolution (LeFever and southwest corner of North Dakota. No salt is exposed in springs along Salt Creek about 14 KM (8.5 mi) north of Newcastle, Wyoming, along US 85. At this location a roadside sign indicates that Darton, in 1904, determined that salt comprised 5 percent of the water there; about four percent salt was reported by Brobst and Epstein of salt were produced by evaporation. Whether that known, but it does show that evaporite dissolution in the subsurface is ongoing. 190 to 268 m (625 to 879 ft) in water wells within the borders This suggests that the gypsum, and possibly unrecognized salt, has been removed to create the discrepancy. Parker distances in several different units, such as the Middle Devonian Prairie Formation and Permian and Triassic age rocks of the Minnelusa Formation, Opeche Shale, and dissolution front that reduced the original depositional boundary of an individual salt deposit, such as in the Jurassic age Piper Formation (LeFever and LeFever (1995). to delimit the southern margin of evaporite in the Middle Devonian Prairie Formation, which is very complex due to dissolution removal of salt and brecciation. A correlation exists between of areas of salt dissolution and the presence of faults and fractures that extend down into the Precambrian basement, and that were reactivated in the Paleozoic. These faults up into the salt bed from aquifers below. Salt removal

PAGE 190

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2173 will dry up and new ones will form as the geomorphology breccia pipes that are preserved in cross section on canyon walls attest to the former position of the dissolution front. and siltstone, has developed secondary fracture porosity due to considerable expansion during the hydration also been reported or suggested in the subsurface Williston Basin, including collapse breccias, breccia pipes, sinkholes (solution depressions), and progressive geologists contemplating the effects of salt karst in the oil patch in North Dakota would do well to examine the Dakota and Wyoming. As an additional thought, a better helpful in planning interstate pipeline routes from the References of formation waters in the Williston Basin. Burke RB. 2001. Some aspects of salt dissolution in the Williston Basin of North Dakota. North Dakota basin model of Williston Basin hydrocarbon systems: American Association of Petroleum 1:750,000,000 scale. Climate Analysis Service, Oregon State and Climate Center, Portland, Oregon; USDA Center, Fort Worth, Texas; Available from: removal and collapse. Using this criteria Parker (1967) described the age of salt dissolution in different units in the subsurface of Wyoming, North Dakota, and Montana: Late Devonian through Mississippian for the Middle Devonian age Prairie Formation, and a Jurassic age for the Minnelusa Formation, Opeche Shale, and Rasmussen and Bean (1984) concluded that the salt the Powder River and Williston Basins was removed Cretaceous time. Fractures extending from the basement controlled movement of groundwater and solution. Orchard (1987) showed that salts in the Mississippian age Charles Formation pinches out abruptly and are mostly missing over the Poplar Dome in the western part are continuous and exhibit little change in thickness. ascending water along fractures that developed during Tertiary Laramide uplift. LeFever and LeFever (1995) concluded that the irregularities in thickness of the Pine of the basin were caused by dissolution during the Late affected during the Late Cenozoic. LeFever (2011) also noted numerous times of salt removal in the Devonian Prairie Formation, including Middle and Late Devonian, even Late Pleistocene to Recent. ConclusionsDissolution of gypsum and anhydrite in several Wyoming, has resulted in development of a variety of collapse breccias, breccia pipes and pinnacles, sinkholes, and extensive disruption of bedding. Evidence of recent collapse includes fresh scarps surrounding shallow depressions, recently formed sinkholes, and sediment disruption and contamination in water wells and springs. Anhydrite dissolution in the Minnelusa Formation probably dates back to the Early Tertiary when the Black moves downdip and radially away from the center of springs associated with the migrating dissolution front

PAGE 191

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE174 Johnson KS. 1997. Evaporite karst in the United States, carbonates and evaporites, vol. 12, Distribution and geologic characteristics of anhydrite deposits in the United States, in Dean, W.E., and Johnson, K.S., eds., Anhydrite deposits of the United States and characteristics of anhydrite important for storage of radioactive LeFever JA. 2012. Impact of the dissolution of the LeFever JA, LeFever RD. 1995. Relationship of salt patterns to hydrocarbon accumulations, North Schalla RA, editors. 7th International Williston Orchard DM. 1987. Structural history of Poplar Dome and the dissolution of Charles Formation salt, Roosevelt County, Montana. In: Fisher DW, editor. Fifth International Williston Basin Symposium: Sasowsky ID, Feazel CT, Mylroie JE, Palmer AN, Palmer MV, editors. Karst from Recent to reservoirs. Karst Waters Institute Special Parker JM. 1967. Salt solution and subsidence structures, Wyoming, North Dakota, and Montana: American plays, Williston Basin, Montana, North and South Dakota, and Sioux Arch, South Dakota and Peterson JA. 1995. Williston Basin Province, in U. S. Dunrud CR, Nevins BB. 1981. Solution mining and subsidence in evaporite rocks in the United Ege JR. 1985. Maps showing distribution, thickness, and depth hazards, and hydrology. In: Johnson, KS, Neal JT, Epstein JB, Davis AD, Long AJ, Putnam LD, Sawyer JF. Fahrenbach FV, Steece JF, Sawyer KA, McCormick stratigraphic correlation chart: South Dakota geology of the Williston Basin. In: Leighton MW, Kolata DR, Oltz DF, Eidel JJ, editors. American seismic delineation of the southern margin of the Middle Devonian Prairie Evaporite Formation in Summary of Investigations 2004, Volume 1. Saskatchewan Industry Resources, Miscellaneous of Saskatchewan: Saskatchewan Department of

PAGE 192

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2175Peterson JA, MacCary LM. 1987. Regional stratigraphy and general petroleum geology of the U.S. In: Peterson JA, Longman MW, Anderson SB, Basin, anatomy of a cratonic oil province: Rocky Rasmussen DL, Bean DW. 1984. Dissolution of Permian salt and Mesozoic syndepositional trends, central Sabel JM. 1984. Sedimentology and depositional Permian and Pennsylvanian geology of Wyoming, (Mississippian), Williston Basin, Montana and North geography: Wiley & Sons, New York, 544 p.

PAGE 193


PAGE 194

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 playa reservoirs during the past 45 years, collectively 7 m and playa sediment. Drainage occurs through piping into in the bottom and along the margins of these playas. Introduction Arizona occurs above interstratal bedded evaporites, principally halite, in Permian sediments. The karst displays a variety of geomorphic features common and other features (Neal et al., 1998). The karst features increasing encroachment of residential and industrial development, and because of potential groundwater formation is still active, noting that a sinkhole visible on since, with more recent activity in December 1995 on Local ranchers have continued to report periodic karst features are among the lesser known geomorphic curiosities in Arizona, but surely one of the most spectacular displays of evaporite karst in the United AbstractAt least six distinct forms of evaporite karst occur from collapse of overlying Permian and Triassic strata long dissolution front on the southwestern edge of the basin, and is characterized by numerous sinkholes and depressions generally coincident with the axis of the sinks up to 200 m across and 50 m deep, the main karst features along the dissolution front. Westerly along the dissolution front, fewer discrete sinkholes occur, and several breccia pipes are believed to be forming. and broad collapse depressions also occur. km2 occurs in the western part of the basin, northward Chimney Canyon area is some 12 km east of McCauley possibly manifested in at least four other closed depressions. Interferometric Synthetic Aperture Radar (InSAR) data of one depression shows active subsidence Karst formation is ongoing, as shown by repeated drainage of Dry and Twin Lakes into newly opened V ARIATIONS IN EVAPORITE KARST IN THE James T. Neal12 Petroglyph Trail, Placitas, NM 87043, Kenneth S. JohnsonOklahoma Geological Survey, 1321 Greenbriar Dr., Norman, OK 73072, Paul Lindberg205 Paramount Drive, Sedona, AZ 86336, Pandplindberg@aol.com177

PAGE 195

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEAdditional exploration for uranium is known to have occurred earlier in depressions now considered to be Anticline, directly related to evaporite dissolution and 21st Century. With each such interest, new exploration has occurred, revealing more subsurface data regarding the evaporite deposits. Hydrogeology the Coconino Sandstone Formation aquifer (Mann, 1976), but the overlying Kaibab Limestone and the uppermost beds of the underlying Schnebly Coconino Sandstone is conspicuously cross bedded, yellowish gray to tan, and is weakly cemented by quartz, iron oxide, and calcite. It thickens from 120 to 250 m towards the northwest across southern Recharge of the aquifer results mainly from precipitation near the Mogollon Rim, some 50 km to the south, where piping of surface waters downward through Dry Lake, River (Mann, 1976). The water table of the Coconino aquifer typically is 120 several areas near the crest of the anticline the Coconino Sandstone is dry, or nearly dry, and the water table is Formation; these strata do not yield much water. The States! Newly discovered (2012) InSAR anomalies in the Chimney Canyon area reveal active subsidence depressions that are described by Conway and Cook 178Evaporite Deposition and Geologic Setting in the Holbrook Basin part of the Pangean supercontinent and perhaps not so Eagle Basins in the Four Corners area to the northeast to open ocean. The current cycle of plate tectonics did Basin evaporite deposition, while roughly equivalent 1990; Blakey and Ranney, 2008), although the Corduroy as far west as Sedona (Figures 1, 2). in outcrop in Sedona. In time the Pedregosa Sea retreated southeasterly and desert erg conditions existed in northern Arizona and western New Mexicoresulting in Coconino Sandstone deposits. and Sedona) did not extend this far eastward. A later marine transgression at 270 mya resulted in Kaibab Formation thinning rapidly eastward to as little as 5 m Basin and disappearing altogether farther east. Triassic Moenkopi Formation outcrops overlie Coconino Sandstone locally and in turn Chinle Formation in the eastern part of the basin. th century, until being shown unsuccessful; since then, storage of st century intense current interest is seen in potash mining (Rauzi, 2008).

PAGE 196

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2179Figure 2. Principal stratigraphic units associated with evaporite karst in the Holbrook Basin, Arizona. The Corduroy Member of the Schnebly Hill Formation (below Coconino Sandstone and above Hermit Formation) is the principal unit undergoing dissolution. Fault in Precambrian basement and pre-Corduroy Member strata is speculative.Figure 1. Extent of Corduroy evaporite member with overlay of surface areas of karst features referenced in this article.

PAGE 197

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE180flowing in the direction of the hydraulic gradient Karst activity in the area involves lateral and downward percolation of fresh water through the Coconino aquifer until it encounters the uppermost salt layers in the Corduroy Member, about 215 m below the land surface. Salt dissolution is accompanied by development of sinkholes and collapse structures in to the dissolution zone. Thus, evaporite karst in this area water in the Coconino aquifer is good, typically with principal constituents are calcium, magnesium, and bicarbonate (Mann, 1976). In the vicinity of the water is high in sodium chloride, and is present mainly in the lower part of the aquifer; undoubtedly it is part of the brine formed by dissolution of salt in the immediately underlying Corduroy Member of Figure 3. Water quality in Coconino aquifer in western part of Holbrook Basin (from Mann, 1976), showing total dissolved solids in mg/L and direction of hydraulic gradient (6 m/km). Holbrook Anticline, McCauley Sinks (X) and Chimney Canyon (CC) subsidence areas are shown.

PAGE 198

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2181 structures over large areas of the Plateau (Kelley and Clinton, 1960; Wilson et al., 1960; Davis, 1978) is a compelling statement for structural control of dissolution effects. Peirce et al. (1970) also argued that the surface anticline expression is not seen in the subsurface beneath the salt, suggesting that basement faulting at depth, as shown by Brown and Lauth (1958), is speculative. The principal sinkhole occurrences are in the Coconino Sandstone, almost exclusively on the steep, southwestern side of the flexure at six distinct locations. breccia pipes conspicuously are near the outer limits of evaporites (Figure 1).Initiation of Karst DevelopmentKarst development required elevation of the Colorado Plateau, which began during the Late Cretaceous some 75 mya, with intensity increasing during the Neogene combined with northward movement of groundwater integrated Colorado River drainage system. Sandstone aquifer and began salt removal of the Corduroy evaporite member, continuing to the present. The groundwater encroachment upon evaporite beds and its consequent dissolution is particularly manifested in the southwestern part of the basin; the area of Dry Lake Valley resulted from the collapse of overlying strata into includes The Sinks, which contains 250 plus prominent sinkholes, and is perhaps the most conspicuous of karst and the overlying Coconino Sandstone, Kaibab Formation (limestone), and Moenkopi Formation top of the fold and compression at the bottom, creating bottom. The surface expression is locally named the Pink Cliffs, deriving its color from red beds of the Moenkopi Formation. Originally the structure was referred to as the dissolution origin for the structure and argued that the anticline apparently does not extend below the salt. from dissolution and collapse of a narrow portion of the Mogollon Slope. Doeringsfeld et al. (1958) show in the southwestern part of the Colorado Plateau. Figure 4. Mechanism of sinkhole collapse along Holbrook Anticline at The Sinks, showing thinning of Corduroy member evaporites in well records (Dean and Johnson, 1989).

PAGE 199

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE182suggesting how some sinkholes are initiated. The sinkholes occur less than one kilometer to the west and southwest of the monoclinal crest; one of these showed draping and overturning of beds in its collapse. Similar features in the Chimney Canyon area suggest common mechanisms of formation. A group of 24 sinkholes, termed here Northwest Sinks, particularly well developed sinkholes are conspicuously larger, deeper, and more regular than the others in this is so prominent at The Sinks is not nearly as evident here. Playa Depressions km2 in the central and western part of the larger collapse zone. Sinkhole development and collapse are ongoing playa (reservoir) in the eastern part of Dry Lake Valley lost more than 6.8 x 105 m o W trend, generally parallel to the regional structural trend (Sergent et al., 1984). The piping occurred along is problematic, but sinkhole formation and collapse are ongoing, as noted previously, and probably began at least by Pliocene time. The uplift and tilting of the environment, but the rates and timing of the orogenic processes are imperfectly known (Lucchitta, 1979). The close association of the dominant regional fractures is conspicuous throughout the region, but this was not understood by early investigators. Accelerated dissolution of halite during pluvial stages of the processes are often noted elsewhere in the arid southwest Figure 5. Aerial view of The Sinks looking south down monoclinal flexure toward collapsed Dry Lake Valley (see Fig. 4). Sinkhole locations are strongly influenced by joint openings and collapse grabens in Coconino Sandstone. Nearly 50 sinkholes are shown in this one view, the largest being 100m across. Principal Varieties of Karst ExpressionThe Sinks and Adjacent Areas topographic expression of what has historically been named to foster interest in petroleum exploration. In fact, monoclinal and now known to result from dissolution of in this part of the Colorado Plateau (Kelley and Clinton, 1960). At many places along the dissolution monocline are collapse grabens that locally form incipient sinkholes, gaping features, some of which purportedly swallowed cattle and possibly two people, and have been described as bottomless by local residents. Field observations suggesting a similar mechanism for the appearance of piping features in the Dry Lake Valley drainage incidents. ground water to penetrate to the relatively shallow (~250 m deep) salt beds below. Near the intersections of some

PAGE 200

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2183contain irrigation water in Dry and Twin Lakes. Bahr (1962) suggested that all of the Dry Lake Valley area could have formed in this manner, noting numerous sinkhole scars along the base of the Pink Cliffs. The many recent drainage events support his hypothesis. The linear northeastern shore of Dry Lake is parallel suggesting a possible structural control as is seen at surface lowering has parallels in New Mexico and West scale or by the same mechanisms.Breccia Pipe StructuresThe McCauley Sinks are comprised of some 50 rings (Figure 6). The outer ring is an apparent tension zone containing ring fractures. The two inner rings are larger depression show local downwarping and possible incipient sinkholes. Permian Kaibab Formation limestone is the principal its easternmost extent. Although surface rillenkarren are present, and the sinks occur within the limestone outcrops, the Kaibab is a passive rock unit that has the water surface dropped 0.6 m, and standing water remained in the surface expression of features within the reservoir (Rucker, personal communication, 1996). The below the land surface in Dry Lake Valley (Mann, 1976). during the 20th century, including nearly 7.4 x 106 m clusters of some 40 sinkholes formed in recent lacustrine sediments over an area of about three square kilometers. Nine sinkholes in this group are aligned along a N 44o W trend and some of the others are oriented orthogonally to them. These were still visible in 1996, as subsequent lowering of the reservoir, combined with the arid climate, has effected little erosion since their formation. Loss of water during piping, or when new openings form and rapid drainage results. A new sinkhole was observed forming in 1996 at the northwestern edge of the basin at an elevation of 1,780 m; apparently it formed in response to normal been maintained below a threshold level of 1,777 m for many years. the downdropped side of a steeply dipping fault oriented N 40o (Rucker, personal communication, 1996). amounts of suspended silt and clay, which in turn seals Comparison of more recent aerial photos taken in 1977 Valley; this area was perennially moist in 1996 because of stock ponds and playas into piping features at other locations in the valley. According to local reports, the abandonment of Zeniff townsite by Mormon settlers in the early 1900s was prompted in part by the inability to Figure 6. McCauley Sinks, a group of 50 sinkholes in a concentric, nested pattern showing three imperfect rings. The outer and uppermost is a ring fracture; the two inner rings are sinkholes with the innermost at the lowest elevation. A compound breccia pipe origin is suggested for this structure.

PAGE 201

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE184that seen in many breccia pipes on the Colorado Plateau, and to the collapse depressions at and near McCauley Sinks. The horst in the depression center is not seen at other breccia pipe structures that overlie evaporites, however; potash thickness beneath the structure may be a developmental factor. The structural framework of this depression may involve concentric ring fractures of unknown attitude. There are no central sinkholes, such as at Richard Lake near McCauley Sinks. The evaporites at shallow depth and similar structure suggest a breccia pipe origin is possible (Colpitts and Neal, 1996), similar to McCauley Sinks and Richard Lake. Other probable Basin at Ortega Sink, The Crater, Deep Lake, and The geometry of all these depressions bears some resemblance to the San Simon Sink, located above thick Permian salt deposits in the Delaware Basin of southeastern New Mexico. At San Simon, a central sinkhole within a broader depression showed renewed growth and concentric fracturing in 1927, and has been interpreted as a possible breccia pipe being formed Reef, the source of many carbonate karst features at the edge of the evaporite basin (Martinez et al., 1998). Dissolution of Redwall Limestone beneath these because of the size of these depressions combined with thinner salt in this part of the basin, its involvement is breccia pipes formed by the Redwall Limestone in the The Chimney Canyon dissolution area includes ~16+ km2 of karst development and occurs coincidentally along the powerline that extends southwesterly from the Cholla Power Plant at Joseph City. Along the western recognized as a dissolution monocline. At the crest of features in folded rocks (Sanford, 1959; Ramsey,1967). Downdip from the crest are numerous collapse depressions (Figure 7) and buckle folds, typical of karst collapsed into solution cavities developed in underlying salt beds. Beneath the Kaibab is Coconino Sandstone, halite in the Corduroy Member. The karst in this part southeast, probably because of the virtual disappearance systems that help channel water down to the salt beds. McCauley Sinks are also near the edge of the evaporite basin, as are the several other broad depressions of unknown origin. The structure at McCauley Sinks suggests a compound breccia pipe, with multiple sinks Richard Lake depression, 5 km southeast of McCauley Sinks, is about 1.6 km wide and with topographic closure and contains only a single, central sinkhole. Richard Lake formerly contained water after heavy rains prior incised, Chevelon Canyon drainage, but the hydrologic connections are unknown. The larger McCauley Sinks provides substantial hydrologic catchment. Because of at shallow depth, runoff water does not pond easily at the surface. There appears to be much greater recharge for groundwater users downgradient from the karst area. the alluvium at the bottom of the central sinkhole, two structure. Two smaller depressions of lesser dimensions occur in tandem immediately west along a N 62o W azimuth. Secondary sinkholes occur within each of these depressions, as at Richard Lake. Breccia pipes are apt to be found beneath all of these structures. Blue Mesa Sink is a semicircular collapse feature about

PAGE 202

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2185Summary and ConclusionsEvaporite karst is displayed over some 9000 km2 of sinkholes, and breccias pipes. Well over 600 such features are readily visible on air photos, and many more of smaller scale exist. Karst expression is manifested compared with the east. The west contains substantially more sinkholes, with more mature development and greater relief along the southwestern part of the basin. And whereas individual collapse depressions in the west are of wider dimension and have fewer discrete sinks, they also are present in the east, but to a lesser extent. Surface lithology, alluvial cover, and evaporite different karst features. The surface conduits through the many areas of karst expression create multiple avenues potential water supplies in downstream communities, requiring continuing vigilance.References Blakey RC. 1990. Stratigraphy and geologic history of Pennsylvanian and Permian rocks, Mogollon Rim Blakey, RC, Ranney W. 2008. Ancient Landscapes of Colpitts R, Neal JT. 1996. Blue Mesa Sink, a possible Basin, Arizona using Interferometric Synthetic Engineering and Environmental Impacts of Karst. (this volume) features and monoclinal folding elsewhere. This area evaporites. Yet the similarity leads one to suspect a similar origin in groundwater movement toward the Little Colorado River drainage system. The prominent at The Sinks gradually is less developed progressing northwestward, virtually disappearing at a point that is only about 15 km south of the Chimney Canyon structures. This close association suggests a possible connection of the two collapse and fold belts that is not apparent at the surface. Surface drainage reversal was noted along several arroyos in the Chimney Canyon drainage, similar to the drainage reversal noticed along Seven Mile Draw at The Sinks to the southeast. Regardless of the uncertainties noted above, it is clear that salt removal and subsidence are ongoing, but perhaps at a slower rate, or more recently than in the thicker part of the salt basin to the southeast. In early 2012 two anomalous subsidence areas were this volume). A reconnaissance of this area revealed a broad depression similar to those at and around Richard Lake (Neal and Colpitts, 1997). Linear Sandstone, along with buckle folding near the low point in the depression. Figure 7. Chimney Canyon cracks and depressions in Coconino Sandstone at top of monocline are similar to those near The Sinks, but at an earlier stage of development. Circular-shape sinkholes were not observed.

PAGE 203

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE186Neal JT, Colpitts R, Johnson KS. 1998. Evaporite karst in Joseph F. Poland Symposium on Land Subsidence: Sudbury, MA, Association of Engineering evaporite karst: McCauley Sinks, Arizona. Norman, gas, helium, and uranium in Arizona: Tucson, AZ: Arizona Bureau of Mines Bulletin 182, 289 p. Rauzi S. 2008. Potash and related resources of the Rucker M. 1996. Personal communication Phoenix, AZ: Sanford AR. 1959. Analytical and experimental study of simple geologic structures. Boulder (CO): Corporation. western United States. Minneapolis (MN): Univ. Minnesota Press, in Porter, S. C., editor, The late Pleistocene, vol. 1, Late Quaternary environments Industries pond containing waste water develops Stone Container Corporation. 1991. Contractor Site Characterization Report. Phoenix, AZ: Contract Arizona Bureau of Mines. Arizona Bureau Mines Bulletin 119. 298 p. Plateau; In Matthews, V. III, Laramide folding associated with basement block faulting in the Dean WE, Johnson KS, editors. 1989. Anhydrite deposits of the United States and characteristics of anhydrite important for the storage of radioactive waste: US dissolution of Permian salt in the Anadarko, Dalhart, and Palo Duro Basins of the Texas Panhandle. Report of Investigations 106. 40 p. features of the Colorado Plateau, Arizona. Tucson State Land Department. 47 p. Jennings JN. 1985. Karst geomorphology. New York Kelley VC, Clinton NJ. 1960. Fracture systems and tectonic elements of the Colorado Plateau. Univ. of New Mexico Press, Albuquerque: University of the Delaware Basin, southeastern New Mexico and West Texas. Albuquerque, NM: Sandia National Lucchitta I. 1979. Late Cenozoic uplift of the Merrill, RB. Plateau Uplift: mode and mechanism: Arizona Water Commission Bulletin 10. 106 p. Martinez JD, Johnson KS, Neal JT. 1998. Sinkholes in Neal JT. 1969. Playa Variation; in Arid Lands in Perspective. Tuscon, AZ: University of Arizona Neal JT, Colpitts R. 1997. Richard Lake, an evaporate

PAGE 204

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2187different collapse morphology than nearby sinkholes, and warrant further study. InSAR will remain a critical Introduction and is predominantly composed of halite with lesser amounts of anhydrite, gypsum and sylvite interbedded with dolomite, sand, and shale in the Corduroy Member body underlies more than 9,000 km2 of eastern Arizona between the towns of Winslow, Sanders, Springerville, 2 (Figure 1). Salt up to 200 m thick has been measured south of salt body has become the focus of much interest from investors and mining companies due to the presence of up to 2.5 billion metric tons of potash (sylvite) near the top of the evaporite deposits. Due to the potential for land subsidence related to possible future potash mining, ADWR has begun collecting InSAR data for the region surrounding the limits of the potash deposit. Through evaluation of InSAR interferograms it was determined subsiding today or have been in the last 20 years. Geologic Setting by Permian Coconino Sandstone, Kaibab Limestone, and Coconino Sandstone is the most laterally extensive exposed bedrock at the surface throughout the study Abstract The Arizona Department of Water Resources (ADWR) recently acquired, processed, and interpreted archived Interferometric Synthetic Aperture Radar (InSAR) data to evaluate historical deformation patterns in the future subsidence related to planned potash mining by ADWR using InSAR data from the European Space Continued subsidence in two of the three features was collected from 2006 to 2011. and ADWR staff visited one of the more prominent InSAR are much more extensive, up to 1,100 m across; are not as deep, up to 15 m; and do not have steep walls. Local subsidence has resulted in broad closed basins with Coconino Sandstone exposed at the surface. A thin sandy expanded depended on location relative to ongoing with other collapse features in the region, these three subsidence features are relatively young, constitute MONITORING EVAPORITE KARST ACTIVITY AND LAND USING INTERFEROMETRIC SYNTHETIC APERTURE RADAR (INSAR) Brian D. ConwayArizona Department of Water Resources, 3550 N .Central Ave, Phoenix, AZ 85012 USA, bdconway@azwater.govJoseph P. CookArizona Geological Survey, 416 W. Congress St, Suite 100, Tucson, AZ 85701 USA,

PAGE 205

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE1881, Neal et al., 1998), but isolated depressions do exist in undeformed beds above evaporite deposits to the northeast. Many of the depressions along the from the dissolution flexure are often much more m deep. Dissolution of Permian salt and associated Basin has likely been occurring since the Pliocene (Neal et al., 1998). Using modern remote sensing monitoring, it is possible to determine whether area. Thin beds of Kaibab Limestone are present near the western boundary of the salt body, but pinch out to the east. Isolated thin exposures of red Moenkopi sands mantle the lighter tan, distinctly crossbedded Coconino Sandstone elsewhere. has resulted in more than 500 sinkholes, expanded geomorphic changes including drainage captures and near the southwest margin of the salt body (Figure Figure 1. Location map of Holbrook Basin, study area extent, and distribution of existing evaporite karst sinks relative to the extent of the Holbrook salt body and anticline.

PAGE 206

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2189evaluate land subsidence associated with any of the existing possible future land subsidence related to proposed potash SAR data were downloaded from the ASF which allowed InSAR data (Figure 2). sink was as high as 5 cm and 26 cm, respectively. This pairs. The spatial extent and magnitude of land subsidence of the new land subsidence features was consistent across all the other interferograms. To better understand the historical activity of the new land subsidence features, ADWR ordered archived SAR data 456 were downloaded from ESA which allowed ADWR to create 27 interferograms. The historical dataset from Land Subsidence Monitoring using InSARADWR has been collecting, processing, and analyzing InSAR data for monitoring land subsidence throughout produced invaluable results and end products that are used not only by ADWR but also other state, county, and local agencies, universities, and private companies for their own monitoring, modeling, mitigation, planning that transmits a pulsed microwave signal towards the earth and records both the amplitude and phase of the is a technique that utilizes interferometric processing that compares the amplitude and phase signals received during geographic area at different times. InSAR techniques, scale vertical resolution. Changes in land elevation are detected through the change in phase of the radar signal. InSAR is used to detect surface motion along active faults, on volcanoes, landslides, sinkholes, and other geologic ADWR has compiled an extensive historical InSAR with InSAR in Arizona. Most data sets cover time periods between 1992 to 2000, 2004 to 2010, 2006 to than 25 land subsidence features in Arizona, collectively ADWR has used InSAR not only for monitoring land subsidence but also seasonal deformation (uplift and tool for geological mapping and investigations, locating mitigation and land subsidence modeling. InSAR ResultsADWR collected InSAR data from the Alaska Satellite Facility (ASF) and the Japanese Aerospace Exploration northeastern Arizona. The InSAR data were collected to Figure 2. ALOS-1 12/06/2006 to 02/01/2011 Interferogram.

PAGE 207

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE190Active Sink MorphologyThe thickness of salt below the western and northern of approximately 240 m and 225 m, respectively. Salt located approximately 5 km north of the eastern feature. dataset (Figure 2). The reasons why the northern land subsidence feature subsidence at the northern feature is episodic or complete. the western, eastern, and northern subsidence features, respectively. Continued InSAR monitoring of this area will help constrain subsidence rates and patterns.Figure 5. Example of a compression ridge within the eastern subsidence feature. Broken bedrock along this ridge follows a sinuous east-west trend. Bedrock slabs in the background are pushed against and above one another.Figure 6. Example of an expanded joint within the eastern subsidence feature.Figure 3. Eastern Land Subsidence Sink and InSAR Profile.Figure 4. ERS-1/2 11/24/1992 to 03/07/1997 Interferogram.

PAGE 208

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2191present at the eastern sink as observed in more mature, portion of the eastern sink near the southern portion exposures exist near the northern and northwestern limits of the actively subsiding sink. The depth of the m relative to the top of surrounding sandstone slopes which is dramatically less than depths observed at many mature, inactive sinks to the west. Drainages near the southern limit of the eastern sink and 60 m thick at a depth of 245 m. For comparison, The overall shape of the eastern sink is that of a broad depression with gently sloping sides that gradually rise Coconino and Moenkopi sandstone outside the InSAR signature. While no vertical walls or steep sides are Figure 7. Subsidence-related geomorphic features of the eastern active sink. Aerial imagery from 2010 National Agricultural Imagery Program (NAIP).

PAGE 209

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE192 feature results in compression between subsiding blocks lower in the landscape. Compressional ridges have been Basin such as the McCauley Sinks, Richard Lake, and at ridges observed here are somewhat smaller than those described by Neal and others but some appear freshly slabs of rock precariously balanced against one another (Figure 5). The ridges observed in the eastern active more abundant in the northern portion of the subsidence feature immediately downslope from the cluster of wide of modern subsidence. InSAR Methodology sensors are processed using the same interferometric methodology. Interferometry is used to process the change in phase between each pair of satellite data. It is important to processed together due to the different sensors. InSAR (DEM) to remove the topography from the phase website () was used to remove the topographic phase component. viewed as deformation contours when examining each feature. ADWR recently started collecting regularly InSAR data will provide ADWR with a critical tool for broken Coconino Sandstone litter the slopes. Portions of No water was present in the lowest point during any of runoff events. Rills, bar and swale channel deposits, and plant matter suspended in and around trees near channels indicate at channels leading to the lowest point of the sink. nearby Moenkopi outcrops, mantles much of the side slopes and bottom of the active sink. Occasionally this nearly continuous cover is broken by collapse wide and up to several meters deep. Joint width tapers Linear depressions in surface sands often parallel or obscured or plugged by overlying sand. Areas of exposed bedrock within the subsidence feature exhibit open to greater depths than those mantled by sediment. that are alternately expanded depending on orientation relative to tension from ongoing subsidence. In some at the northern and northwestern edge of the InSAR across, up to 10 m deep, with vertical offset up to 1 m between blocks. Successive vertical offset across open the edge of the active subsidence feature. Coconino outcrops beyond these exposures do not exhibit this subsidence indicates subsidence may have initiated somewhat farther to the north than indicated by recent InSAR data (Figure 7).

PAGE 210

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2193 opportunity to observe the gradual processes that lead to sinkhole formation is an exciting prospect. In addition to continued subsidence monitoring observations, photos, and have installed eye bolts across to enable repeat measurement in the future. A benchmark near the most rapidly subsiding area within the eastern obtained to supplement future InSAR data and enable repeat surveying of subsidence within the feature. In an attempt to better understand the subsurface geometry monitoring ongoing land subsidence at these two new sinks as well as possible land subsidence associated National Park.Conclusions created a dynamic, geomorphically intriguing landscape. With recent interest in potash mining and continued that is the source of dissolution beneath hundreds of subsidence rates and mechanisms is important. Because Figure 8. Radarsat-2 InSAR frame used to monitor future land subsidence in the Holbrook Basin.

PAGE 211

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE194and lithologic structure deep Refraction Microtremor conducted both within and outside the subsidence feature. Results of this investigation are pending and future visits to the eastern subsidence feature to collect as Richard Lake which most closely resembles the extent and morphology of the eastern active sink, are proposed.References derived ground displacements in hydrogeology. Neal JT, Colpitts R, Johnson KS. 1998. Evaporite Borchers JW, editor. Land Subsidence, Case Studies and Current Research. Belmont (CA):

PAGE 212

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2195dissolution feature. Two other sites were surveyed near (indirectly to Edwards Aquifer) along Cibolo Creek. IntroductionThe Edwards and Trinity Aquifers are critical water region of central Texas, USA. These carbonate aquifers tectonic deformation associated with the Balcones karstic aquifers are managed separately by regional water regulatory entities, and have been historically from a water policy standpoint. Three separate electrical geophysical investigations at Camp Bullis Military Training Site (Camp Bullis) (Figure 1) were performed to characterize the hydrogeologic 2 (28,000 acre) area that helicopter survey of Camp Bullis and nearby areas to map and image AbstractElectrical and electromagnetic geophysical characterization is a proven tool for delineating obscured subterranean karstic features, such as caves, sinkholes, and allow a wide range of deployment scales; airborne methods can accommodate a regional view on the order with focused data on the order of meters. A helicopter Bullis Military Training Site (Camp Bullis) in central Texas have been used to characterize permeability properties of the Edwards and Trinity Aquifers in the zones of high density karst features and characterized and nearby areas to map and image subsurface features locations of mapped and previously unmapped faults and characterized the heterogeneity of the subsurface electrical signature. Karst mapping at Camp Bullis features correspond with areas of high resistivity from used to infer and characterize known and hypothesized karst features. Site 8 suggests an inferred fault and GEOPHYSICAL INVESTIGATIONS OF THE EDWARDS-TRINITY AQUIFER SYSTEM AT MULTIPLE SCALES: INTERPRETING AIRBORNE AND DIRECT-CURRENT RESISTIVITY IN KARSTMarcus O. Gary Edwards Aquifer Authority, 900 E. Quincy, San Antonio, TX 78215, USA, mgary@edwardsaquifer.orgDale F. RuckerHydroGEOPHYSICS, Inc., 2302 N. Forbes Blvd., Tuscon, AZ 85745, USA, druck8240@gmail.comBruce D. Smith, David V. SmithUnited States Geological Survey, Crustal Geophysics and Geochemistry Science Center, P.O. Box 25046, MS973, Denver Federal Center, Denver, CO 80225, USA,, dvsmith@usgs.govKevin BefusThe University of Texas at Austin, Department of Geological Sciences, 1 University Station C9000, Austin, TX 78712, USA,

PAGE 213

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE196 subsurface of Camp Bullis. It is divided into two members. The upper member has been divided into formalized with names by Clark et al. (2009). Figure 2 shows a three dimensional block diagram of Camp Bullis (Zara, 2011). The cavernous member (interval A) is formed by Bullis member (interval B), which is lithologically similar to the cavernous member, but has less karst development and lower permeability. The upper evaporite member (interval C) is a thin layer of highly soluble carbonates and evaporites, characterized by breccia porosity, boxwork permeability and collapse structures. The fossiliferous member (interval D) has low porosity and permeability, with the exception of a caprinid biostrome near the top of the interval, in the center of Camp Bullis, and thins to the north. The lower evaporite member (interval E) is quite similar to the upper evaporate member (interval C), with mostly dissolved evaporites diverting groundwater horizontally. The lower member of and caves that rapidly transmit surface water to the aquifer) along Cibolo Creek. The area in and around Camp Bullis has been conducted over many years, documenting over 1500 karst features (Zara, 2011). Karst feature density was estimated using karst feature locations and the weighted were determined by giving numerical values to each karst feature, using hydrogeological characteristics proportional to potential recharge. Results are shown with results of geophysical studies. subsurface features related to groundwater resources. infer and characterize known and hypothesized karst features and structural features. One of these sites, located near a heavily investigated remediation area (Site 8; Figure 4) possibly indicates an inferred fault and dissolution feature. Two other sites were surveyed along Cibolo Creek on the north side of Camp Bullis the surrounding area and are complemented with two of localized electrical resistivity properties related to dissolution features.Hydrogeologic SettingThe hydrogeologic setting of Camp Bullis has been documented in numerous reports related to the Edwards and Trinity Aquifers, both formed in Cretaceous limestones. Two publications in particular lithologic description of the Edwards in Bexar County Figure 1. Location of Camp Bullis lies within the Edwards and Trinity aquifers (Zara, 2011).

PAGE 214

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2197Figure 2. Block Model of Camp Bullis based on Clark (2003) and Ferrill et al. (2003). Mapped karst features are shown as orange points. Figure modified from Zara, 2011. Geophysical InvestigationsThree independent geophysical investigations are shown here to display different scales of data collected at Camp Bullis. Locations of all the studies are shown Airborne Electromagnetic Survey Bexar County covering the Edwards Aquifer Recharge Zone and the Trinity Aquifer at Camp Bullis, Camp the west), and part of Cibolo Creek east of Camp Bullis (Figure 4).

PAGE 215

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE198Figure 3. Karst feature density map of Camp Bullis showing the spatial distribution and significance number of karst features (Zara, 2011).

PAGE 216

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2199Figure 4. HFDEM survey data at 115 kHz frequency from Camp Bullis. DC-ERI sites are shown as white circles. The Edwards-Trinity contrast is clearly shown in the HFDEM data (Smith et al, 2005) The water table is 30+ meters below the land surface throughout Camp Bullis, thus these resistivity values reflect the vadose zone. five horizontalcoplanar coils and one vertical coaxial Details of the survey specification and digital data aredescribed by Smith et al. (2005). The survey was except as required for safety considerations and FAA part of the survey area to yield an effective flight line spacing of 100 m. The measured electromagnetic fields were converted to apparent resistivity by the contractor for each frequency. The depth of penetration increases with decreasing frequency and with increasing resistivity. The shallowest depth meters for all of the survey. The apparent resistivity is shown in Figure 4. previously unmapped faults, and characterized the heterogeneity of the subsurface electrical signature. In general, the massive limestones of the Edwards Recharge Zone at the southern end of the survey are shown as an area of very high apparent resistivity (100s boundary between the high resistivity on the south and a normal fault boundary between the Edwards and Trinity Aquifers. The Trinity Aquifer is characterized

PAGE 217

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE200 through superposition (Rucker, 2012). Although the data were collected along 2D transects, the spacing between lines allowed the domain to be modeled in three dimensions, to a depth of 45 m below ground data and was opted for additional diagonal smoothing to reduce striping inherent in modeling volumes comprised of individual transects. The results of the resistivity distribution are shown in Figure 5B as an overhead view (which can be observed through the lighter transparent blue in the northwestern portion of the site). those areas devoid of color, i.e., the lowest values have been blanked. Based on the vertical distribution of the upper 11 m of the domain. Below 11 m, the resistivity of increased saturation. The results show that there is an overall trend of high resistivity features that align along an approximate N22E strike to the northeast. A clear banding of the highest values can be observed through the center of the site, which likely represents more competent limestone. The low resistivity material that has been removed from the with higher clayey material and moisture content. Unfortunately, wells drilled in the immediate vicinity of the study did not uncover evidence of sinkholes, as they were placed prior to resistivity acquisition. The sinkholes appear to also align at N22E or perpendicular at N58W. Arrows have been provided to highlight these directions. The spatial density of low resistivity material increases in size and number towards the east (closer scenarios of contaminant transport emerge. Either the existed) integrity has been breached through further sinkhole development.Cibolo Creek Karst Features by alternating mudstones, siltstones (low resisitivities; cooler colors in Figure 4), and limestones (warmer colors) which give the aquifer a variable signature. In general the upper part of the Trinity Aquifer is composed of thin discrete limestone and siltstone layers that give the variegated color pattern in Figure 4. The middle Trinity is composed of more massive reefal structures within mudstone units. The trends in the apparent resistivity map correlate to and greater detail in the lithologic changes than indicated in geologic maps such as the thin limestone units and more detail in structural trends. There is also a strong high resistivity) exist in the subsurface.Site 8 DC-ERI the center of Camp Bullis, to the west of Lewis Creek (Figure 5A). The purpose of the resistivity survey was to gain a better understanding of potential karst features that would help explain contaminant transport through the underlying aquifer. Contaminants were detected in a the area shown in Figure 5. The resistivity data were acquired along 16 transects was used for acquisition, with remote electrodes placed at least 700 m away. Transects were about 95 m long rapid acquisition with high signal to noise ratio and resolution and therefore not optimal for locating small scale features that would provide the best insight into the range of sinkhole sizes. To accommodate a higher (Schlumberger array), and overlapping dipoles according to the procedure outlined in Loke et al. (2010). The

PAGE 218

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2201Figure 5. (A) Study area of surface resistivity south of the Site 8 Landfill showing the survey lines. (B) Overhead view of three dimensional resistivity showing two isopleths: 250 (light blue) and 400 removed to highlight patterns of potential sinkholes filled with soil. The medium blue is from the combined effect of both blues. Electrodes are black dots.

PAGE 219

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE202 anomaly within 1 m of the surface. A more notable high 69 m) that extended to 5 m depth, which was interpreted from the ER line where both the depth (~2 m) and size of the cavity agree between the ER and cave map. Uncertainty dimensionality of electrical properties in the subsurface that are modeled in 2D, and the ER inversion process inherently smooths discrete and abrupt ER features and boundaries resolve any apparent karst features. imaged well by the ER surveys (Figure 9). In Line 1, the cavern depth (10 m) at the correct position along the line (~75 m). The extension of the high ER values m) of voids in the subsurface, as near equal horizontal and vertical smoothing (averaging) was used during the Cibolo Creek at the northern border of Camp Bullis. creek bluff and are mapped to extend below Cibolo Creek (Zara and Veni, 2010). ran across sinkholes associated with the cave entrance, 2D ER lines crossed over the cave location (Figure 9). Electrode spacing ranged from 1.5 to 5 m, depending on the depth required for imaging the karst features and datasets were collected and merged prior to inversion for each line. Line topography was recorded with a total station and included in the inversion. The merged datasets were inverted in RES2DINVx64 with a robust model constraint. the subsurface expression of the cave (Figure 7). The main Figure 6. Cave map shows Bullis Hole cave, which is located on the right bank (south) of Cibolo Creek. This cave extends below the creek bed (Zara and Veni, 2010).

PAGE 220

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2203understanding of recharge heterogeneity across the aquifer system. To accomplish this at Camp Bullis, we mapped geology (Figure 2) indicates the electrical properties imaged closely relate with the different hydrogeologic properties of different formations and primary porosity heterogeneity is one component of electrical resistivity data. This is most clearly observed the southeast section of Camp Bullis. The differences in primary porosity between these two formations are substantial electrical variation, and relate to increased porosity and varied lithology associated with reefal Cibolo Creek. may represent a small cavity.DiscussionIn this paper, we evaluated three geophysical case studies performed at Camp Bullis. They were each conducted independently from one another with different the subsurface electrical properties of the geology beneath Camp Bullis and surrounding areas. The Site 8 investigation imaged the geophysical signature near a contaminant remediation site, characterizing variable zones of resistivity, relating to possible locations of directly targeted known, mapped caves below the creek bed, and these caves have been observed to discretely recharge into the aquifer. A next step would be to link these disparate studies with other known hydrogeologic, hydrologic, and geomorphic data to improve the Figure 7. ER results at Bullis Hole cave. Line 1 imaged a high ER area where the cave crossed the line (~70 m). The second line did not reveal any new potential karst voids.

PAGE 221

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE204 Figure 8. Cave map of Jabbas Giant Sink cave, located on the right bank (south) of Cibolo Creek. The cave extends below the creek bed and has been observed to rapidly recharge the aquifer through an active whirlpool during floods (Zara and Veni, 2010).

PAGE 222

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2205 signature. This is shown on the local scale by the other were conducted in areas of moderate to high resistivity aquifers in this region of central Texas, and is one meticulously surveyed for karst features, possibly in greater detail than any other large, contiguous area in the U.S. This rich dataset (Zara, 2010) provides a unique opportunity to compare known, evaluated karst features with the regional electrical properties (Figures Figure 9. ER results at Jabbas Giant Sink (cave outline shown in orange). Both lines 1 and 2 resolve high ER features where the cave was expected to cross into the surveys (~75 m line 1; ~ 40 m line 2). Line 2 may have resolved a shallower cavity (~50 m).

PAGE 223

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE206this manuscript: Allan Clark, Charles Blome, Ron commented on the material.References Bullis Training Site, Bexar County, Texas. U. S. Clark AR, Blome CD, Faith JR. 2009. Map showing geology and hydrostratigraphy of the Edwards Aquifer catchment area, northern Bexar County, petrophysical models to radar travel time and Ferrill DA, Sims DW, Morris AP, Waiting DJ, Franklin Bullis Quadrangle, Texas. Southwest Research Institute report prepared for the Edwards Aquifer Authority and U.S. Army Corps of Engineers. Rucker DF. 2012. Enhanced resolution for long survey data and maps, northern Bexar County. U.S. hydrogeologic characteristics of the Edwards aquifer recharge zone, Bexar County, Texas. U.S. Zara Environmental, Veni and Associates. 2010. paleontological karst investigations, Camp Bullis, Texas. Prepared for Natural and Cultural Resources Antonio, Texas. Zara Environmental. 2011. Karst hydrogeology of Camp Bullis, Bexar and Comal Counties: A window into Natural and Cultural Resources Environmental rocks and enhanced secondary porosity in the faults, fractures, solutional voids, and karst conduit networks. to improve our understanding of the spatial heterogeneity of aquifer recharge.ConclusionThree different electrical geophysical studies performed at Camp Bullis were evaluated for their characterization of the permeability fabric of the Edwards and Trinity study area closely correlate with mapped geologic outcrops and spatial distribution of karst features. correspond to electrical signatures (high resistivity of potentially identifying karstic voids, or areas with hydrogeologic data that exists for Camp Bullis can be expressed in the electrical signature of the subsurface, Applying this methodology throughout the region to increase the ability of regional groundwater models to simulate aquifer dynamics of the Edwards and Trinity aquifers and their interaction with each other.AcknowledgementsResearch presented in this paper is greatly indebted to at Camp Bullis, particularly Lucas Cooksey and Chris Thibodeaux. Their support of ongoing, integrated all work performed at Camp Bullis. We thank the U.S. Department of Defense and the Edwards Aquifer thank the University of Texas at Austin Applied Karst Edwards Aquifer Conservation District provided Environmental performed years of karst research at the site, providing the rich karst dataset used for analysis. We also thank those who gave thorough reviews of

PAGE 224

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2SUBBOTTOM PROFILING INVESTIGATION OF SINKHOLE LAKE STRUCTURE IN BAY AND Thomas L. Dobecki, Sam B. Upchurch, Thomas M. Scott, Beth Fratesi, Michael C. AlfieriSDII Global Corporation, 4509 George Rd., Tampa, Florida, 33634 USA, are often circular in plan with a striking blue color sand bottoms, and their internally drained basins are the homes of many endangered or threatened floral and faunal species, including bald eagles, gopher species of carnivorous sundews. It has long been common knowledge that the lakes have but after viewing shorelines during periods of drought when lake levels drop, it was suspected that the mode of creation and growth was complex as demonstrated by the string of pearls appearance of new, smaller depressions forming around the perimeter of the lakes (Figure 2). In addition, there were questions regarding the interaction and connection between the lakes, the means of developing an understanding of lake structure and growth mechanisms as well as an understanding of the interconnection between the lakes and the FAS, a program of marine (lake) geophysical imaging was AbstractThe sandhill lakes of Bay and Washington Counties, Florida, are deep, pristine environments which owe their existence to sinkhole activity as observed in limited bathymetric maps and in the appearance of small circular depressions around their perimeters (string of pearls) observed on aerial photography little investigative information exists that shows the internal, deep structure of these lakes and how that structure might affect interaction with groundwater lakes in order to investigate the deep structures of the lakes for purposes of determining the mode of sinkhole development within each lake and their relationships aquifer system (FAS). The SBP provided mapping and subbottom structure up to an additional 18 24 m below the bottom, all with a theoretical resolution of approximately 0.10 m bed thickness. The resulting SBP profiles showed that a) the lakes form through the coalescence of numerous small sinkhole features, b) the sinkhole features penetrate the uppermost FAS, disrupting the overlying sediments, and c) multiple stages of sinkhole development and sediment movement are exhibited in the subbottom strata. IntroductionThe sandhill lakes region (Figure 1) of the Florida Panhandle occupies portions of Washington and Bay Counties (north of Panama City), Florida. It owe their existence to sinkhole activity. The lakes 207 Figure 1. Plan view of sandhill lakes of NW Florida.

PAGE 225

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEdownwards to scan below the boat. Communications system so that positional coordinates can be recorded concurrently with the acoustic echoes. Parameters (e.g. maximum depth range, instrument gains) are optimized during the crossing. Figure 4 is a representative SBP printing) of these data sets was a mute to remove The following principal features can be seen in the example cross section in Figure 4 and in other cross sections included here. Time zero (the water surface, 0 meters depth) is represented by the horizontal line across the top of the section. The total horizontal length of the section is labeled at the top of the section. Actual position can be determined by playing back the SBP The vertical scale is depth in meters. For the SBP Geophysical Methodology technology that measures the depth of water as well as acoustic echo sounding principles. The technology is on boats. The principal difference is that the SBP employs an acoustic source that has increased signal power and somewhat lower frequency than a common from the sediment layers may also be detected as the al., 1996; Reich, et al., 2012). SDII employed the Syqwest StrataBox SBP system laptop computer. The system is battery powered, and With an assumed seismic compression wave velocity the order of one wavelength or less (approximately 0.10 Operationally, the acoustic source is lowered below the water surface to ensure excellent coupling of the acoustic waves into the water column. The source is directed 208Figure 2. Aerial photograph of a sandhill lake showing a string of pearls of new, recent sinkholes along the lake shore. Figure 3. Photograph of SYQWEST SBP system mounted on an aluminum boat.

PAGE 226

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 bathymetry varying between 4.5 and 22.9 meters water depth. This vertical exaggeration emphasizes the features that indicate that Big Blue Lake has not formed as the result of a central, conical depression related to a single sinkhole throat. From the image, we see that structures, and some of these are well developed (i.e., they are deep with steep sides) while others appear to Note in Figure 6 that the depressions have a common the FAS. The small sediment volumes highlighted in yellow on Figure 6 represent accumulations that have developed in the bottoms of the depressions after sinkhole development. Since the lakes have no surface water tributaries, these sediment accumulations were derived from slope wash into the lake from the portion slumps from the sides of the depressions. represents an acoustic wave that traveled down to (down plus up path). downwards from the water surface and represent second, third, fourth, etc. roundtrips of the signal. water depth is shallow, but as depth increases the multiples begin to disappear off the bottom of bottom sediment layers. shown on Figure 5. In this manuscript, we present the down the long axis of Big Blue Lake from north to south extending from southeast to northwest along the southern smaller circular depressions observed on historical air photographs taken when lake levels were lower. afternoon thundershowers typical of this region in the summer months. All geophysical activities are curtailed 209 Figure 4. Example SBP section. Figure 5. Map of Big Blue Lake showing SBP profiles. The results of Profiles BB-3 and BB-4 are presented in this manuscript.

PAGE 227

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEFigures 6 and 7 demonstrate that, even though the lake appears to be a single, large sinkhole depression in plan view, it actually consists of a series of overlapped and coalesced individual sinkhole structures. These overlapped sinkhole structures, which are highlighted in yellow in Figure 7, appear distorted because of a complex image that shows not only the structure of the lake bottom bathymetry (as per Figure 6), but it also shows considerable (22 m) penetration into the lake bottom sediments. 210Figure 7. SBP Profile BB-4 (Big Blue Lake). PRIMARY (BOTTOM) REFLECTIONNNE SSWBB-3 ~2,027 meters MULTIPLES 0m 12.2m 24.8m 36.5m 0ft 40ft 80ft 120ft Figure 6. SBP Profile BB-3 (Big Blue Lake). PRIMARY (BOTTOM) REFLECTIONSoutheast BB-4 Northwest ~945 metersMULTIPLES 0m 12.2m 24.8m 36.5m 0ft 40ft 80ft 120ft

PAGE 228

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 The lakes appear to be growing in surface area sinkholes rather than by suffusion or slump of perimeter sand into the existing depression in the lake bottoms.References Pratt, TR, Richards, CJ, Milla, KA, Wagner, JR, the Northwest Florida Water Management District. Northwest Florida Water Management District Tihansky, AB, Arthur, JD, and DeWitt, DW. 1996. in the Lake Wales Ridge, Central Florida, US investigation of Sentinel Lakes in Lake, Seminole, Orange, and Volusia Counties, Florida, US history of multiple subsidence events and development of slump structures. Figure 7 reveals a smaller (newer?) sinkhole structure forming to the left of the main depression. There are depression suggesting that the larger, basinal features and internal sinkholes are continuing to grow. The maximum observed depth of subbottom penetration is 18 meters MSL. The maximum depth of penetration therefore, would extend down to approximately elevation meters from the water put the elevation of the top of the intersect the FAS, verifying direct connection between the lake and the underlying FAS. The small, relatively new sinkhole on the left hand side that the sinkholes appear to bottom out at the top of the limestone of the Floridan aquifer, it appears that there is a limit as to the available sediment to ravel into the Floridan from the center of the lake. Rather, the lake appears to be growing laterally by development of new, Many of the sandhill lakes in the area show these young, string of pearls (Figure 2).Conclusions SBP is an excellent tool for determination of the origin and geometry of sandhill lakes, which typically have little clay or organic sediment. Penetration is excellent, and sinkhole structures and slump features are readily apparent. The sandhill lakes of Washington and Bay counties, Florida, have formed by coalescence of a complex of smaller sinkholes that have a common apparent depth at the approximate top of the underlying limestone of the FAS. The FAS in the area is characterized by many springs and lakes appear to represent the headwaters of the conduit systems. 211

PAGE 229


PAGE 230

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2IMPROVED IMAGING OF COVERED KARST WITH THE MULTI-ELECTRODE RESISTIVITY IMPLANT TECHNIQUEDavid Harro Geo3Group, 2410 Success Drive, Suite 7, Odessa, FL 33556, david.harro@geo3group.comSarah KruseDept. of Geology, University of South Florida, 4202 E. Fowler Ave., SCA-528, Tampa, FL 33620, resolution of the survey. This is especially true in urban areas where restrictions in the surface array length limit the depth of penetration. In urban areas cultural features can also have a considerable impact on the geophysical results and complicate interpretation of geophysical karst settings is often a critical problem and geophysical resolution imaging between borings. The resolution depth of electrical resistivity imaging (ERI) surveys is limited by the distance between the furthest electrodes involved in any single reading (e.g. address this depth restriction is to place electrodes at depth (e.g. Pidlisecky et al., 2006). To fully exploit the available array length, we install electrodes at uniform Electrode Resistivity Implant Technique (MERIT), deeper features can be imaged. In covered karst, we interface to image epikarst or possible cover collapse development. By combining measurements with surface and deep electrodes we can also improve imaging of the sediment column above the karst development. In cases where sinkholes are stabilized by grouting, this method could be used to help verify sediment stabilization.MERITWith MERIT, the depth of penetration of a resistivity survey can approximately be extended by lowering the electrodes closer to the depth of target horizons (Figure expected to resolve features to approximately 7 meters in depth, with greater depths at the center of the array and shallower depths near the ends of the array. If, for example, the bedrock surface is 10 meters below land surface (bls) then the surface geophysical survey will AbstractElectrical resistivity tomography (ERT, also called ERI) is commonly used to identify geologic features associated with sinkhole formation. In covered karst top of limestone with this method. This is due to the fact rich and highly conductive. The resistivity method has limited sensitivity to resistive zones beneath conductive with electrodes implanted at depths near the top of limestone, in addition to readings at the surface. Deep electrodes are installed with direct push technology, top limestone surface contact. This method, which we are resolution of epikarst and cover collapse development zones at the limestone surface sediment interface in heterogeneous karst environments. The technique could also help reduce the effects of cultural features typically encountered by surface electrical resistivity surveys in urban environment. The results of a case study sinkhole investigation in At this site the resistivity array length is restricted to 60 meters. The depth to the top of the limestone lies at ~15 meters. Electrodes were implanted both at the and deep measurements improves the resolution of the measurements alone. Introduction karst often have had limited success because the depth to 213

PAGE 231

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEground stabilization efforts. Successive images could be acquired during the sinkhole formation process, and pre and post compaction grouting. perhaps at the center of the array. This geophysical much of the array. By lowering the ERI array through the unconsolidated sediment, MERIT can get closer to the electrode array as well as below it, measurements are now responsive to overlying sediments as well as image the underlying contact and voids in the limestone requires that the resistivity of the overlying sediments is simultaneously resolved. This is done by also taking measurements with electrodes implanted at the surface, as in a conventional array. With the MERIT method, electrodes are installed with direct push technology. Upwards of 150 linear meters of implant installation can be performed in a single day. Referring again to our example for the top of limestone surface at 10 meters bls a 28 electrode implant would require 277 linear meters of direct push drilling at a cost of approximately 1.5 days of direct push installation. The additional cost of installation is offset by the enhanced In this example, without the implanted electrodes the the center of the array. Lateral variability and features associated with the development of cover collapse sinkholes could not be imaged. the same location), the MERIT method could be used for imaging sinkhole development and the effects of 214Figure 1. MERIT method schematic. Electrodes are emplaced at the surface and at depth with direct push technology. Figure 2. Bordeaux Apartments sinkhole, Tampa, FL. Resistivity profile location shown with red line. Boring results are shown in Figure 3. Geophysical surveying was limited to the apartment complex grounds; the spatial constraints on survey dimensions are clear from the photo. North is to the lower right of the photo. The sinkhole was filled with sand at the time of the survey.Case Study Bordeaux Apartments The Bordeaux apartments in Tampa, Florida received national news coverage in July of 2010 after a car in the parking lot was swallowed by a 7 meter diameter cover structure. Over several weeks the sinkhole continued to enlarge, further threatening the existing structure. Florida geology and sinkholesSinkhole occurrences such as this one are numerous in Florida, and have resulted in substantial number of insurance claims for damages to structures (Schmidt 2005). The development of karst on the Florida carbonate platform has been related to sea level changes of up to 92 to 109 meters below current sea level (Tihansky 1999). These sea level changes have resulted in carbonate rocks being exposed to karst processes (Beck 1986, 1991). In

PAGE 232

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2meters bls and continuing through the entire casing installation. The two deeper locations were located on the east and west sides respectively of the structure affected by the sinkhole activity.Conceptual ModelThe Bordeaux Apartments test site lies in an area conceptual model of the sinkhole formation (Beck 1988) was developed prior to the geophysical testing. Two possible cover collapse geometries were considered: The sinkhole forms part of a collapse conduit could possibly extend under the affected building. vertical and radial extent. created sinkholes that have affected many structures, irrigation and drinking water wells and farm lands. in geologic areas of the county where the cover is 10 to 65 meters thick (Sinclair et al. 1985). The cover is characteristically comprised of undifferentiated Quaternary sediments that overlie Tertiary clay consists predominantly of limestone with subordinate dolostone, sand and clay (Scott et al. 2001). The area of the test case is known locally for a high development of sinkhole occurrences. Standard Penetration Test borings performed on the entire property of the Bordeaux Apartments. Results of borings B1 and B2 near the of three basic stratums. From the surface, Stratum 1 consists of 7 meters or less of undifferentiated quaternary sediments of mainly sands. Stratum 2 is comprised of clays and sandy clays of thickness ranging from 6 to 10 meters thick. These sediments vary in clay content and contain limestone fragments limestone in the borings varies from 10 to 19 meters depths around the perimeter of the structure averaged from 12 to 15 meters bls, however at one location ,top of limestone bedrock was encountered at 75 meters bls. Additional analysis came from the grouting program (e.g. Sowers, 1996). A total of 62 compaction grouting points were also installed around the perimeter of the structure and ranged from an average of 12 to 15 meters bls with a single location reaching 44 meters bls. Loss of circulation was recorded at all grout except the grout point that extended to 44 meters in 215 Figure 3. Borings B2 and B1 (location shown in Figure 1). The uppermost sand constitutes Stratum 1, the intermediate layers constitute Stratum 2, and the underlying limestone constitutes Stratum 3 as discussed in the text.

PAGE 233

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE216Data were inverted using Res2Dinvx64 version 4.0 software required the use of the borehole geometry algorithms within using only the surface electrodes. With surface electrodes, there is no indication of the more resistive limestone below the clays. Figure 5 illustrates the reason for this, namely meter depth of the limestone contact. Figures 6 and 7 show that when data from the deep electrodes are added, higher resistivities associated with the limestone are imaged (reds and yellows at depth). of interest is increased dramatically. Discussion and Conclusions MERIT ProfileAt the time of approval for the use of MERIT, remedial efforts of underpinning and compaction grouting were clean sands. It was determined the metal underpinning was positioned too close to the structure. Additional power lines and property boundaries. Thus it was edge of the sinkhole (Figure 2). The MERIT array was comprised of 18 surface ERI The MERIT implants were set at 10 meters depth and were recorded for both surface and buried arrays, and an additional set of readings were taken in which surface electrodes were used as current dipoles and potential measurements were recorded with buried dipoles. Figure 4. Resistivity profile inversion using data from surface electrodes only. (See Figure 2 for location of profile and borings B1 and B2.) The rms error on this inversion is 11.1%. Only the central 53 meters of the 59 m-long profile are shown. There is no indication of higher resistivities at depth associated with the limestone.

PAGE 234

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2217Figure 5. Relative sensitivity of the resistivity survey using data from surface electrodes only. The sensitivity is a measure of how well the resistivity in a given part of the model can be resolved by the data collected. Sensitivity values are normalized by dividing by the mean, and are unitless (Geotomo, Inc. 2011). Resistivities in yellow areas are well-resolved, resistivities in dark red areas are poorly resolved. The surface survey has limited sensitivity below 8 meters depth. Figure 6. As for Figure 4, but incorporating readings from electrodes at depth. This inversion includes the traditional surface dipole-dipole array, the equivalent dipole-dipole array at 10 meters depth, and readings with current electrodes at surface and potential electrodes at depth. In this inversion, zones of higher resistivity are observed at depths where the limestone was reached in SPT borings.

PAGE 235

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE218Acknowledgements and an anonymous reviewer improved this manuscript. ReferencesBeck BF. 1988. Environmental and Engineering Effects of Sinkholes The Processes Behind the Problems. Beck BF, Sinclair WC. 1986. Sinkholes in Florida: An Introduction. Florida Sinkhole Research Institute at the University of Central Florida College of Purposes. Florida Sinkhole Research Institute at the University of Central Florida College of Florida Center for Instructional Technology. 2008. Sinkhole Maps of Florida Counties. John Wiley & Sons Ltd. modeled in the 2D inversion process, or to gradational changes in resistivity across zones of changes in lithology that are more gradual than noted in the boring logs, or to problems in the inversion of relatively noisy data. We note the cross data, in which current was at another level, were particularly noisy. Optimal acquisition and inversion procedures are a topic of ongoing investigation. With additional case studies and synthetic modeling in progress, we anticipate better understanding of the sources of noise and uncertainty in the resistivity images. Despite the current limitations, it is clear that at the Bordeaux Apartments the top of the limestone could not be imaged with surface electrodes alone, but could be seen with the addition of the MERIT deep electrodes. The MERIT approach has increased the depth of resolution, and permitted imaging of the target horizon despite the spatial limitation of the site.Figure 7. Relative sensitivity of the resistivity survey data when incorporating readings from electrodes at 10 meters depth. Compare to Figure 5; note the increased sensitivity at depths of 10-13 meters.

PAGE 236

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2219 Investigation Procedures for Evaluation of the Causes of Subsidence Damage in Florida. Florida Sinclair W. C. et al. 1985. Types, Features, and construction of Foundations in Karst terrain. New York (NY): ASCE Press.

PAGE 237


PAGE 238

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2change at depths of roughly 120 to 150 meters. Results of this work, including interpretations and assessments of knowledge gained, practical additional assessment work that could be performed, and some as yet unanswered questions are presented.IntroductionA relatively large region of evaporite karst, primarily developed above bedded halites of Permian age, is described by Neal et al. (1998), over 500 karst features dissolution front coincident with a structure named Synthetic Aperture Radar (InSAR) group within the AbstractThe Arizona Department of Water Resources (ADWR), Radar (InSAR) to assess subsidence in parts of Arizona, potential future sinkholes in an area with several hundred natural evaporite karst depressions or sinkholes. An initial reconnaissance geophysical subsurface evaluation can be obtained using relatively simple, unobtrusive and methods such as Refraction Microtremor (ReMi). ReMi can utilize ambient ground vibrations from natural construction equipment. Shallow ReMi has been applied shallow subsurface material strength as part of assessing the potential for collapse of an evaporate brine cavern into a large sinkhole in southeast New Mexico, but had conditions in the deeper subsurface above and in the vicinity of a possible impending sinkhole. Two deep ReMi surface wave soundings and two resistivity soundings using the Wenner array method were performed, one each within and outside of the extent of current subsidence as derived from the InSAR. Surface wave velocities within the subsidence zone were lower (weaker material) than surface wave velocities outside the zone. The underlying horizon had high surface wave velocities indicating relatively competent rock. Deep resistivity soundings indicated possible lithologic RECONNAISSANCE EVALUATION OF A POTENTIAL FUTURE SINKHOLE USING INTEGRATED SIMPLE SURFACE GEOPHYSICS AND SURFACE MONITORING POINTSMichael L. Rucker, Sean HulburtAMEC Environment & Infrastructure, Inc., 4600 East Washington St., Suite 600, Phoenix, AZ 85034,, Mark D. Edwards Ninyo & Moore, Geotechnical & Environmental Sciences Consultants, 3202 East Harbour Drive, Phoenix, Arizona 85034, mark.edwards@ninyoandmoore.com221 Figure 1. From Conway and Cook (Figure 1, 2013), InSAR study area in relation to the Holbrook salt body, Holbrook anticline and known evaporite karst features.

PAGE 239

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE horizontal ground displacements at various subsidence feature locations. Several eyebolt monument pairs have width can be monitored over time. This paper addresses initial preliminary geophysical characterization of the subsurface at the active Section that might be characterized include depths or thicknesses of the geologic formations and generalized aspects of geologic material strength. Karst activity can commonly Above a carbonate or evaporite dissolution zone, enhanced weathering and fracturing, or even collapse, of the overlying rock mass reduces the overlying rock mass strength. Rock mass strength is related to seismic velocity and can be assessed using seismic methods. In addition to lithologic changes, electrical resistivity may provide information concerning groundwater or changes in rock mass weathering that may be related to karst activity. Evaporite karst may impact or develop within a in depth. Exploratory drilling to such depths can be overall site characterization and coverage over a and characterization of minable potash deposits. Although such methods may provide excellent critical applications can be prohibitive. Refraction microtremor (ReMi) surface wave seismic and Wenner cost reconnaissance surface geophysics capable of providing useful subsurface characterization to depths in excess of 100 meters. have impacted industrial activity in the region. Neal et to identify areas of active subsidence. Incorporating Arizona Department of Water Resources (ADWR) data for the region in preparation for monitoring potential future subsidence related to proposed potash mining in InSAR interferogram indicating active subsidence at this feature is shown is Figure 2. The sink feature is located in a relatively isolated area, with no known potential geohazard impacts, and thus no pressing public or private safety or economic needs for monitoring or engineering further study on a volunteer basis to document observable geologic behaviors or changes that may contribute to knowledge of subsidence and sinkholes. Part of the study initiated involves measurement of information is an interpretation of recent vertical monuments at the feature and have conducted initial 222 Figure 2. L-band InSAR interferogram (courtesy B. Conway, ADWR) of Section 3 feature showing extent of local subsidence, RTK survey, subsidence profile A-A and geophysical sounding locations.

PAGE 240

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2recorded in ADWR databases (ADWR, 2012). A depth 18)02ACB in 1946. Depths to groundwater at Well sea level (amsl) are reported to be about 112, 120 and 121 meters in 1968, 1975 and 2009, respectively.Surface Elevation and SubsidenceThe regional ground surface trend in this portion of the A local ground surface elevation trend is included southern and northern ends are consistent with a ground surface trend in the absence of local subsidence. The RTK surface elevation survey indicates that subsidence has Assuming a uniform grade prior to local subsidence activity, possible subsidence could have been as great 200 meters. Possible subsidence is reduced towards the north, and is typically 2 meters or less compared to the no local subsidence trend. The InSAR current subsidence pattern covering the apparent pattern of historic subsidence derived from indicates a steep increase in apparent subsidence over InSAR with other characterization and monitoring and mitigate some subsidence and sinkhole hazards. Geologic SettingSubsidence, sinkhole and karst activity in the area is presumed to be driven over geologic time by dissolution Colorado Plateau. As summarized by Neal et al. (1998), rocks of Permian Age, including the Coconino sandstone and salt deposit bearing Corduroy member of the Sedona known as the McCauley Sinks, are overlain by younger east of the McCauley Sinks, and Coconino sandstone feature. Johnson (1962) describes the Coconino in the sandstone is tightly cemented with silica, but the degree of cementation varies considerably from place to place as well as vertically in any given section. Without well log control in the vicinity, the local thickness of the Coconino is unknown. Based on limited historic geophysical log data in the region (Figure 4 in Neal et al. 1998), it may be Underlying the Coconino sandstone is the Corduroy shale. Dissolution of salt in the Corduroy Member is considered to be the primary subsidence mechanism for generating karst. Neal et al. (1998) indicate that dissolution has likely been continuing since the Pliocene. bedrock fracturing activity, providing concentrated pathways for groundwater to access soluble formations, tends to be concentrated closer to the anticline. It is also dissolution would tend to be fractured or perhaps even brecciated as the rock column subsides downward while dissolution progresses. 223 Figure 3. Comparison of RTK surface elevation survey with L-band InSAR-derived subsidence on a south to north profile through Section 3 feature (Figure 2).

PAGE 241

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE karst activity. The Coconino sandstone is primarily of Aeolian origin. largely lacking in conductive clay particles, high resistivities can be anticipated in the upper subsurface consisting of the Coconino. Lower resistivities can be anticipated in underlying shales at the top of the underlying Corduroy Member. Thus, deep Wenner array resistivity soundings can be anticipated to encounter a low resistivity horizon. An estimate of thickness of the overlying Coconino might be interpreted from resistivity sounding results. Seismic velocity relates quantitatively to material modulus, and thus is strongly related to geologic an indicator of material strength below the groundwater the Coconino sandstone can be anticipated to have high in a loss of support under the Coconino, distortion, fracturing and perhaps even brecciation and collapse of the otherwise intact formation into the underlying void underlying dissolution collapse has not occurred and the formation is relatively competent. Low measured underlying dissolution has occurred and the overlying formation has been fractured, distorted or brecciated.Initial Geophysical Reconnaissance was performed by a small group of volunteers on 15 September 2012. Two each deep Wenner array resistivity and ReMi surface wave soundings were completed at locations shown in Figure 2. Based on existing data interpretation, Soundings Location 1 was generally assumed to be outside the area of active subsidence to attempt to provide a baseline for the subsurface geologic subsidence in that area is negligible. The area of greatest apparent subsidence, typically about 12 meters, along the InSAR derived subsidence up to about 0.05 meters in decreases at the northern edge of this deep local area. nearing its minimum, the current active subsidence is subsidence is at the northern lip of the local apparent subsidence area. A benchmark was established in this serve as an RTK survey base. Active subsidence then gradually tapers off to the north. If the current rate of local deep subsidence area would increase in size to the north, but the maximum subsidence depth would not Anticipated geophysical measurement results or targets can be developed based on the geologic setting, apparent local subsidence pattern and magnitude, and of anticipated geophysical results would assist the process of understanding and characterizing details of conditions and mechanisms driving continued development of karst in this area. Deep sounding electrical resistivity results are anticipated to be material lithologies; both impact development of 224 Figure 4. Resistivity equipment including meter (yellow box), electrode cable spools and water jugs; water is added at electrodes to improve contact with ground.

PAGE 242

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2typically high for a shale below the water table. It may also indicate a combination of a relatively thin section of saturated Coconino and somewhat deeper underlying low resistivity shale. At a minimum, the top of the Corduroy Member can be interpreted to lie at a depth greater than about 128 meters at Sounding 1. It must be understood that the resistivity interpretations result models such as horizontally uniform layers or horizons. Resistivity Sounding 2, located in the active subsidence area, had similar interpreted results (Figure 5) as Sounding 1. The interpreted top of the lower resistivity horizon, at a depth of about 111 meters, was shallower at Sounding 2 than the 128 meter depth at Sounding possible contributor to explain a shallower depth to lower resistivity. Discarding the measurement at 15 meter data was excellent, as shown in Figure 5. An anomalous reading at the 15 meter electrode spacing could be bedrock near the northern outer current electrode. Such (partially blocking) the current path near that electrode.ReMi Seismic MeasurementsRefraction Microtremor (ReMi) is a seismic surface (Rayleigh) wave method (Louie, 2001; Optim, 2004) that utilizes surface wave dispersion physics to Location 2 was centered near the RTK base in the area of greatest current subsidence magnitude to attempt to provide some initial characterization of geomaterial properties in the area of active subsidence.Resistivity MeasurementsResistivity measurements were made using the Wenner method. Four electrodes are set into the ground, making electrical contact with the earth, along a line at equal spacing. An electrical current is applied to the two outer electrodes, and the voltage difference in the resulting apparent resistivity, a function of voltage, current and earth with radius of the electrode spacing. A sounding consists of a series of progressively larger electrode spacing measurements that sample progressively deeper to assess, but is less than the largest electrode spacing. Interpretation consists of developing mathematical models of layer thicknesses and resistivities that is matched to the measured resistivities. resistivity soundings. Electrical contact between the stainless steel electrodes and earth was established by driving the electrode into the dry to very slightly moist cables and reels were used to set electrode positions at the longer electrode spacings. Resistivity Interpretation ResultsInterpretations of the two resistivity soundings were consistent with high resistivity unsaturated rock lacking the active subsidence area, ranged from about 4,400 above the groundwater table. Below a depth of about 128 meters, the interpreted resistivity dropped to about saturation in the Coconino below the water table. Alternatively, it may indicate the top of the underlying 225 Figure 5. Interpretation of resistivity Sounding 2 showing apparent resistivity measurements, curve fitting and resulting interpretation.

PAGE 243

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEThe interpretation at ReMi Sounding 2 within the area of active subsidence is presented in Figure 7. The relatively low surface wave velocity is in the perspective, it might be considered equivalent to a indicative of competent rock. Thus, a relatively competent rock portion of the Coconino sandstone is interpreted to be present at the Sounding 2 location in the area of most active subsidence. If available, other subsurface information can help velocity reversals and below the water table where compression wave methods (such as standard seismic refraction) can be severely limited. ReMi can also be used in naturally and culturally noisy environments and can use ambient ground vibration as a passive surface wave energy source. Depth of investigation is constrained by the wavelength penetration of the surface waves into the earth; depths of investigation greater than 100 meters are often attainable if sufficiently low frequency surface wave energy is available. Transcontinental several kilometers to the north of the study site provided a ready source of ambient low frequency surface wave energy. The field procedure included laying out cabling for and setting and leveling the 24 low frequency northeast to point towards the general direction of seismograph was used to collect 12 or 24 second ambient surface wave datasets at sampling intervals of 1 or 2 milliseconds. Since railroad traffic was many kilometers away, geophone channel gains were set to high amplification; no filtering was applied during data collection. The interpretation at ReMi Sounding 1 outside the area of active subsidence is presented in Figure 6. An immediate difference of the ReMi interpretation and general concept of the geologic profile is a relatively low surface wave velocity in the upper 50 meters of the subsurface. At a velocity less than 700 Coconino sandstone is not a relatively competent, high strength rock material. At best, it might be considered equivalent to a very soft rock, but may perspective). Below about 50 meters, the greater competent rock.226 Figure 6. Interpretation of ReMi Sounding 1 showing spectral amplitude as a function of surface wave velocity and frequency, dispersion curve fitting and resulting interpretation. Figure 7. Interpretation of ReMi Sounding 2 showing spectral amplitude as a function of surface wave velocity and frequency, dispersion curve fitting and resulting interpretation.

PAGE 244

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2DiscussionInitial geophysical reconnaissance has begun to to understand the active subsidence documented by InSAR and possible previous subsidence indicated results have also raised questions about assumptions of subsidence mechanisms. The interpretation of a high velocity horizon at Sounding 2, consistent with relatively competent Coconino sandstone in the area of highest current subsidence and the edge of greatest underlying salt. Is there an alternative explanation for the apparent 12 south of the highest current subsidence? The shallower low velocity upper Coconino horizon might provide an explanation. If differences in seismic velocity in outside (higher velocity) the active subsidence are correct, changes in the weak rock fabric similar to collapsible soils might explain some of the apparent subsidence. As described previously (Johnson, 1962), the Coconino sandstone has a cemented structure. If that cemented structure is disrupted, the particles could collapse into a more dense structure, resulting in surface subsidence.Geophysical AnomaliesEven though quantitative analysis may require ignoring or discarding some anomalous data, anomalies in geophysical data can provide critical information to assist in understanding subsurface conditions. As was previously noted, a resistivity reading at Sounding 2 was discarded for interpretation due to the presence of Sounding 2 ReMi seismic array was deployed in this same area. Anomalous signal loss in ReMi data from this location is shown in Figure 9. A coherent surface wave blocked between traces 9 and 10 from the right in Figure 9 on the trace printout. That type of attenuation anomaly propagation. From the interpretation shown in Figure resulting in a wavelength of about 227 meters. Since most of the surface wave energy propagates within the upper quarter to half wavelength of the surface, such extend well into the subsurface.227 Figure 8. Open fissure along surface geophysical arrays alignment. Note near-surface bedrock and relatively uneroded adjacent soil surface that indicates recent fissure development. Figure 9. Anomalous surface wave signal loss in ReMi Sounding 2 field data at open fissure. All geophones are set at equal gains (84 db). Note loss of coherent signal from right to left between the 9th and 10th traces.

PAGE 245

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEis small relative to the capability of the geophysical measurement, potentially critical variations within the that is large relative to the capability of the geophysical can be performed to develop some understanding of the If nearby borehole or well logs, or other detailed data, are available at a site, then surface geophysical soundings may be of limited value, or may have value in verifying whether conditions are similar to or change from nearby known conditions. When no other local subsurface level geophysical soundings may be a primary feasible economic means to obtain preliminary subsurface information to begin a site investigation. Conclusions and Recommendations subsidence in a geologic setting of probable geophysical measurements have been applied to initiate a baseline of data to improve understanding of possible subsidence and karst mechanisms. Initial results from these measurements indicate the likelihood of complex, feature, and by inference, other karst areas within the The authors recommend that a continuing, possibly informal program of mapping, monitoring and and modulus based on concepts of Percolation Theory as described by Sahimi (1994). Rucker (1998, 2000, 2008), has developed these relationships for cohesionless points freely move (roll) relative to each other. When limestone, they are not free to move (roll) relative to each other. Relationships of surface wave velocity to data points are discussed in Rucker (2008). The ReMi Sounding 1 horizon with surface wave velocity chemical gel material that, although relatively low density, has an intact cemented rock fabric relatively unaffected by subsidence or karst activity. It might have a mass The same horizon within the subsidence area may have suffered disruption of the rock fabric as the rock mass has been stressed and degraded, perhaps through mass With loss of cementation or bonding, the rock material structure might behave as a physical gel material that has consolidated or collapsed into a more dense structure. As a physical gel material, the analogous ReMi Sounding 2 The change from chemical to physical gel behavior could cause the horizon volume to original volume. Since 71 percent of 50 meters is about to a physical gel behavior is consistent with the difference in the interpreted horizon thickness. Also, the difference the roughly 12 meters of apparent subsidence measured by the RTK elevation survey.Limitations of 1-Dimensional MethodsCharaterization of karst can be a highly complex and may be, in part, a function of other available information about a karst site and the scale of measurements being made compared to the scale of the phenomena being evaluated. Scale is a primary variable. When a target 228 Figure 10. Percolation Theory-based relationships between seismic velocity and mass density for physical gel and chemical gel geo-materials. When bonds in a chemical gel material are disrupted, that material can revert to a physical gel as shown.

PAGE 246

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 100 meters depth from refraction microtremor arrays. Bulletin of the Seismological Society of Neal JT, Colpitts R, Johnson KS. 1998. Evaporite karst in Land Subsidence, Case Studies and Current Research. Rucker ML. 1998. Seismic velocity in cohesionless granular deposits. In: Robertson PK, Mayne PW, International Conference on Site Characterization. Rucker ML. 2000. Estimating earthwork factors for Sahimi M. 1994. Applications of Percolation Theory. Bristol (PA): Taylor & Francis. measurements proceed at this feature. Detailed mapping of of exposed bedrock should be performed and analyzed patterns of current active subsidence. Such tasks could activities and other knowledgeable, technically capable volunteer groups. Further simple surface geophysics might then continue to be strategically deployed based features and continuing InSAR monitoring. Due to cost, land ownership constraints and the absence of an urgent geohazard condition, future exploratory drilling is not Results presented here are an initial reconnaissance information concerning this feature increases through further investigation and monitoring, knowledge and understanding of the feature and the mechanisms acting experience gained by such a program, even if informal, will further demonstrate the value of incorporating multiple tools and approaches, including integrated simple surface geophysics, for characterization and monitoring of subsidence and karst. AcknowledgementsThe authors would like to acknowledge the continuing support of Brian D. Conway and ADWR with the InSAR data and processing, and survey support, with the ReferencesArizona Department of Water Resources (ADWR). 2012 November 4]. Available from: Arizona using Interferometric Synthetic Aperture Carlsbad, NM. In press. Johnson PW. 1962. Water in the Coconino Sandstone 229

PAGE 247


PAGE 248

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 SPONTANEOUS POTENTIAL INVESTIGATIONS OF A Philip J. Carpenter, Ryan F. Adams, Melissa Lenczewski Dept. of Geology and Environmental Geosciences, Northern Illinois University, 312 Davis Hall, DeKalb, IL 60115 USA,,, lenczewski@niu.eduRosa M. Leal-Bautista made near a surface conduit. This transparent zone was at least 1.5 m wide and extended over several meters depth. SP measurements near this conduit during a suggesting SP may also be a viable method for mapping overall implication of this work is that geophysical methods are valuable in delineating recharge points and shallow contaminant pathways, and should be used more extensively in this part of the Yucatn Peninsula to support groundwater investigations.IntroductionThe municipal water supply for Cancn, in the northeastern Yucatn Peninsula of Mexico, has been degraded often by fecal coliform bacteria and other contaminants. Water supplies for the Yucatn are largely derived from highly permeable fractured karstic limestone characterized by geophysical techniques that could be of use in identifying In early January 2012, a team from Northern Illinois University (NIU) traveled to Cancn, Quintana Roo, Mexico de Yucatn (CICY) to perform exploratory geophysical techniques were chosen based on instrumentation traveling that provide rapid recharge and contaminant pathways that lead from the land surface to the aquifer. VES were used to examine the overall vertical electrical structure of the well necessitate 2D resistivity surveys in the future.Abstract Mexico, approximately 20 km southwest of the city of Cancn, in order to identify karst conduits that channel contaminated surface waters into the main aquifer. electrical soundings (VES) and spontaneous potential (SP) surveys were employed to identify these conduits and detect water movement through them. Cancn s municipal water supply has frequently been affected by fecal coliform bacteria and other contaminants. Water supplies are largely derived from highly permeable fractured karst limestone aquifers characterized by rapid transport of microbial and chemical contaminants from the surface to subsurface Tertiary limestone bedrock outcrops across this entire Schlumberger array VES were made at two locations.. limestone), overlying a high resistivity layer 8.2 m thick (massive limestone with some small caves), overlying could not be successfully inverted due to lateral resistivity water table. An unusual transparent zone (absence of 231

PAGE 249

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE232 and collapse features (generally less than 1 m across). visible at the surface. Thus no cave maps exist in the study area.Methodology Three areas were investigated in detail over the southwest were near Well (Pozo) 49A (Area 1) and at the intersection geophysically were essentially targets of opportunity: Well 49A was open and being serviced (Area 1), (2) a visible road, Well 40 and the Ruta de los Cenotes, and a service road intersection with nearby apparent depressions in the 2). In general geophysical surveys were made along roads. Road pavement consists of the limestone bedrock. resolution technique has the potential to identify conduits transmitting contaminants from the surface Geological Setting near the Cancn International Airport, approximately 20 km southwest of the city of Cancn. This area lies within the Yucatan northeastern coastal plain (Isphording, 1975) on a low upland between the eastern coastal ridge the west. Pliocene Upper Miocene limestone of the Carillo Puerto formation outcrops, or is covered by only turn, overlies carbonate Cretaceous and Jurassic units containing evaporites (Perry, 2002). Hydrogeology, Contaminants and Karst Features however, are under considerable strain from extensive pumping and development along the Caribbean coast, study area, discharging at the coast, about 25 km away The survey area is regarded as a contaminated part of the Cancn municipal well system, which consists of several fractures and conduits in the limestone that are directly connected to the shallow drinking water aquifer. There of contaminants. The most common contamination is related to feces pathogens, nutrients, etc. This area scale agriculture, tourism, or residential activities. solution openings. Numerous publications discuss the cave systems of northeast Yucatan, including Thomas (1999), Beddows Riviera Maya cave systems, consisting of several long caves 10s of km long and 10s of meters wide are the most extensively studied and mapped. Navigable caves Figure 1. Map of study areas showing roads and cenotes. Inset shows location of study area within the NE Yucatn (inset after Beddows, 2003).

PAGE 250

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2233geophysical surveys one wellhead had been removed for maintenance (Pozo 49A); the water level in this well was measured at 6.2 m beneath the surface. It is not or not. Most likely it represents the water table, since personnel showed up, and the well was left open. The nearest vertical outcrop is at the Calica quarry, near this area of the Yucatn make this a plausible comparison to the study area. water and small ponds, suggesting the water table is at the base of this massive unit, placing the water table Results GPR zone between depths of 4 and 12 m containing perhaps surveys were along straight lines, although some lines were run parallel to each other to examine the areal 4 to 40 m. Antenna separation depended on antenna frequency and antennas were manually moved along arrival and shifting traces, if necessary, to account for area and color displays were employed, generally using automatic gain control with a maximum gain of wave velocity measured in the CMP surveys or through The SP and VES surveys were both run from the same information about SP and VES surveys in karst areas see Ford and Williams [2007]). The VES utilized four stainless steel electrodes pounded into the thin topsoil SP surveys were made over several intersecting lines near the pumping wells, as well as over apparent karstic conduits (holes visible at the surface) during rainfall. The target was small voltages induced by moving water (streaming potentials) (Reynolds, 2011). Two porous electrode while the other acted as a roving electrode and the unit recorded the difference in potential (in mV) least 10 m from the roving electrode. Ground truth calibration area. The municipal wells for this area were not logged, and any records made during drilling (during the 1970s and 80s) were not immediately available from the Cancn Municipal Water Department. During the Figure 2. Cross-section of limestone at the Calica quarry, near Playa del Carmen, scale bar is in upper left.

PAGE 251

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE234 resistivity model consisting of a 2.1 m upper layer space layer, which probably represents saturated limestone. This structure is consistent with what was observed at Calica (Figure 2), although thicknesses are different. The lower resistivity upper layer is probably highly fractured and weathered limestone, relatively unweathered limestone and may contain as shown in Figure 2, resulting in its elevated resistivity. The lowermost layer probably represents diffractions, possibly generated by conduits, small caves, or other sharp heterogeneities. Figures 5 and 6, directly below a surface conduit (a hole at the surface VESSchlumberger resistivity arrays were used for VES in Areas 1 and 2 (Figure 7). Electrodes were inserted into the thin soil (zero to 5 cm thick) covering bedrock. No electrode conditioning was employed. Figure 3. GPR section across part of Area 1. Antenna frequency was 50 MHz, separation 2 m and the step size between traces 0.1 m.Figure 4. GPR section across part of Area 2, showing disrupted reflections and possible collapse features. Antenna frequency was 100 MHz, separation 1 m and the step size between traces 0.1 m.

PAGE 252

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2235 in the potential was noted. This could be due to the conduit flow nature of the aquifer, i.e. the SP lines might not have passed near the hydraulically active fractures those wells were drawing from (streaming potentials that generate SP are discussed in Reynolds [2011] as well as other geophysical texts). The SP surveys, however, recorded significant changes in potential (about 16 mV) over a conduit where rainwater was infiltrating, as shown in Figure 8. 5 and 6. The VES in Area 2 was severely affected by lateral resistivity variations and could not be interpreted as a layered model with high confidence. This suggests 2D resistivity should be employed in future surveys in Area 2.SP Several profiles were collected along cross shaped, intersecting lines near the pumping wells. While the lines were not very long, the wells were pumping at Figure 5. GPR section across part of Area 3. Antenna frequency was 100 MHz, separation 1 m and the step size between traces 0.1 m.Figure 6. Same GPR section as in Figure 5, but plotted with variable density and in color, to denote areas of signal loss.

PAGE 253

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE236 methods should also be employed that would allow for more expansive and contiguous data sets such as towed antennas, employing very low frequency (VLF) and other systems. This would allow the extent of the conduits to be better characterized and help to understand possible contaminant routes. AcknowledgementsThe authors would like to thank two anonymous reviewers and the Editors whose comments and corrections greatly improved the quality of this manuscript. Travel expenses were paid for, in part by the Mobil Oil Foundation at Northern Illinois University.References Juan A. 2009. Mapping subsurface karst features Analysis of the karst aquifer structure of the Review: The Yucatan Peninsula karst aquifer, (published online) Beddows PA. 2002a. Where does the sewage go? The karst groundwater system of the Municipalidad de Solidaridad, Quintana Roo, Mexico. Association for Mexican Cave Beddows PA, Smart PL, Whitaker FF, Smith SL. 2002b. Density stratified groundwater circulation on the Caribbean coast of the Yucatan Peninsula, Mexico. In Martin JB, Paleozoic carbonate aquifers, Karst Waters Yucatan coast. Bulletin 11, Association of Mexican and Sons. Conclusions and Future WorkThis study evaluates the feasibility of using geophysical conduits in a karstic aquifer utilized as a water source by the City of Cancn. Three techniques were evaluated: model with a moderate resistivity upper weathered layer overlying a high resistivity layer perhaps representing lowermost layer was much lower resistivity, suggesting it is surface, where an anomaly of about 16 mV was recorded, presumably due to streaming potentials. Future work should concentrate surveys over known voids or high permeability zones, so that geophysical Figure 8. SP profile across a flowing conduit in Area 3. Figure 7. Resistivity sounding made with a Schlumberger array in Area 1, along with the layered model inverted from the sounding curve.

PAGE 254

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2237Isphording WC. 1975. The physical geology of Yucatan. hydrogeochemistry of the karst aquifer system of the northern Yucatan Peninsula, Mexico. Reynolds JM. 2011. An Introduction to Applied and Smart, PL, Beddows PA, Coke J, Doerr S, Smith S, Whitaker FF. 2006. Cave development on the Caribbean coast of the Yucatan Peninsula, C, editors. Perspectives on karst geomorphology, hydrology and geochemistry a tribute volume to

PAGE 255


PAGE 256

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 since the 1960s to solve karst geology questions including karst collapse. Direct current (DC) resistivity techniques have been used in cave detection because of its low costs, (Vincenz, 1968; Smith, 1986; Panno et al, 1994; Batayneh geophysical tool for identifying and locating subsurface karst features, such as cavities, conduits and fractures (Arzi, 1975; Blizkovsky, 1979; Butler, 1984) and the methods are effective for forecasting karst collapse, but the precision, continuity and depth of the exploration are limited Monitoring for trigger factors based on the groundwater pressure has become more common and can be an effective method for forecasting karst collapse (Lei et al, 2002; Li et al, 2005; Meng et al, 2006). The greatest laboratory studies of the seepage deformation test have indicated that errors often appear due to discrepancies the soil, even within a single layer. Thus, the smallest experimental marginal hydraulic gradient usually is adopted as a threshold for engineering safety, but this leads to a very low forecasting accuracy. latest method for predict karst collapse. Model tests show that there is a very good relationship between the techniques need to be implemented to understand the AbstractThe aim of this paper is to describe improvements in the accuracy of forecasting karst collapse by summarizing the methods and analyzing their advantages and as geophysical surveys, monitoring of triggering but the precision and depth of exploration are limited by equipment parameters and geological conditions. For example, ground penetrating radar can discover a soil cavity when the overburden layer is less than 15 m thick, and frequent scanning can determine changes in the soil cavity and predict sinkhole collapse when combined with a balance arch model. Monitoring of triggering factors is widely used to forecast karst collapse when the opening is caused by pumping, as the dynamic groundwater conditions can be acquired a method based on the relationship between the times when anomalous monitoring data appear and the time collapse is the most advanced technology currently available in China. The location and time of sinkhole opening can be forecast by this method in theory, but some key issues have yet to be resolved. These issues and the soil, the effect of temperature on the optical the soil. Finally, some proposals are suggested in the hope that they will generate public discussion, reducing the damage caused by karst collapse.IntroductionIn most studies of sinkholes, forecasting has been mainly based on geophysical surveys, monitoring of triggering factors, and strain measurements of the soil TYPICAL METHODS FOR FORECASTING KARST COLLAPSE IN CHINAYan Meng, Jianling Dai, Long Jia, Mingtang Lei Institute of Karst Geology, Guilin, 541004, P.R. China,, Feng JiState Key Laboratory of Geohazard Prevention and Geoenvironment Protection, Chengdu University of Technology, Chengdu, 610000, P.R. China, 54589998@qq.com239

PAGE 257

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE2 2 develop further when h>(b2 Frequent scanning to monitor changes in the arch is very important, but can have high costs when the survey area soil cavities because of equipment limitations and the complicated geological conditions. Monitoring of Triggering Factors Following a geological survey and risk assessment, boreholes may be constructed in high risk locations to monitor changes in the groundwater through the karst Determining the forecasting threshold is a key part of the method, and laboratory testing is the most common way. First, some undisturbed soil samples are obtained, and these in the lab. If the hydraulic gradient in the laboratory tests is the soil, and a karst collapse will open at the ground surface. mechanical properties of soil even in the same layer. In the paper, the principles underlying each method forecasting results. Finally, the merits and disadvantages of each method are analyzed and some suggestions are provided for improving the forecasting precision and reducing the harm from karst collapse.Ground Penetrating Radar (GPR) SurveysThe simplest and most popular method for forecasting geophysical surveys. This technique can directly identify is very important to predict sinkhole openings at the ground surface. If changes in the soil cavity arch from the on an idealized balance arch as shown in Figure 1, where: v is the natural vertical stress; is the natural horizontal stress; h is the height of the arch; b is the half v (Protodyakonov, 1962). The coordinates x and y describe the location of A on the arch LOM in the Figure 2. The 240Figure 1. Ground penetrating radar image of a highway in Guilin, China and the idealized model of its balance arch.

PAGE 258

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2241Figure 3. Monitoring of groundwater pressure.Figure 2. The idealized GPR modeling of its balance arch (Protodyakonov, 1962). v is the natural vertical stress; is the natural horizontal stress; LOM is the arch; h is the height of the arch; b is the half span; f is the Protodyakonov (1962) coefficient; x and y are the coordinates of A on the arch LOM; Rv is the horizontal thrust at the top of arch; P is the thrust at the arch springing and T is horizontal component, N is vertical component.

PAGE 259

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEtime of the sinkholes opening have a linear correlation. The equation describing the relationship between the time of the maximum anomaly appearing and the time the equation describing the relationship between the time of the minimum anomaly appearing and the time of the relationship between the time of the most anomalous To improve the forecasting accuracy, a new method based on the residual analysis of groundwater pressure was developed (Table1). The monitoring data is anomalous based on relationships between the time that anomalous monitoring data were recorded and the time of karst collapse (Table 2). The data analysis shows that the times of the maximum, minimum and most anomalous values appearing and the 242Figure 4. Relationship between timing of maximum anomalies and collapses. ID The monitoring time Monitoring data(kpa) Simulating data(kpa) residual errors 1 21.061 2 21.078 21.078 4 5 21.111 .... .... .... .... .... Table 1. Monitoring simulation data and residual errors.

PAGE 260

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 respectively. The feasibility of this linear relationship can be seen by the model test (Figure 5). First, optical will show the changes in the soil, and thus can predict sinkhole opening at the surface. The results show that the position of peak strain in the optical fibers corresponds to the area of disturbance and cavern formation in the soil, and the change in optical fiber strain in different soil layers indicates the vertical boundary of the area of disturbed soil (Figure 6). The time series of optical fiber strain shows the ongoing formation of areas of disturbed domain reflectometer (BOTDR) will become a reliable method for monitoring and predicting sinkhole collapse or subsidence, especially along highways and railways. value appearing and the time of the sinkhole opening 0.998. Consequently, the time of a sinkhole opening may will be very important in the future.Strain measurement using optical fibers the newest technologies. It has been found that there is a linear relationship between the Brillouin frequency shift (BFS, described by VB( expressed as: (Eq 1) VB( ) is the BFS including strain and temperature; VB(0,T0) is the BFS at the initial temperature, T0, without strain; and C1 and C2 243 Time of sinkhole opening Time of anomalies Number of anomalies Minimum anomalies (kTime of minimum anomaliesMaximum Anomalies (kpa) Time of maximum anomalies 175 107 41 186 17 26 52 84 11 11 72 22 5Table 2. Relationship between the timing of anomalous values and karst collapses in Guangzhou, China. (,)(0,)() 0120 VTVTCCTT BB

PAGE 261

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE forecasting karst collapse can have very high costs, and the precision and depth of the exploration are limited by equipment parameters and geological conditions. Integrating geophysical surveys with other methods to identify and explore sinkholes may provide better results. When monitoring trigger factors using groundwater, it Discussion Sinkholes are one of the main geological hazards in karst concealed nature and sudden appearance. Research from characterizing karst collapse formation and monitoring collapse, such as geophysical surveys, monitoring trigger 244Figure 5. Model sinkhole tests using optical fibers in Guilin, China.Figure 6. Strain curve from optical fiber during sinkhole opening

PAGE 262

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 collapse research in China. Environmental hazard assessment in evaporitic materials from ground penetrating radar: a case study. Environ Meng Y,Yin KL, Lei MT. 2006. Probabilistic analysis on technical problem in monitoring and predicting Protodyakonov MM (1962) Mechanical properties and drillability of rocks. In: Proceedings of the 5th symposium on rock mechanics, University of Minnesota. Smith D. 1986. Application of the poledipole resistivity technique to the detection of solution cavities Ulriksen C. 1982. Application of impulse radar to civil engineering. PhD Thesis, University Lund, Sweden. Vincenz A. 1968. Resistivity investigations of limestone location of the sinkhole opening cannot be determined. of groundwater withdrawal can be controlled by this method. promising new technology. The position and time of sinkhole opening can be forecast by this method in deformation; the effect of temperature on the optical immature, this method is very promising. These observations above are provided in the hope that they will generate more public discussion. Acknowledgements survey (No. 1212011220192), the special foundation of public service of land and resources in 2012 (the Key Technologies of the Sinkhole Monitoring and the National Natural Science Foundation (No. 40572164 ReferencesArzi A. 1975. Microgravimetry for engineering using electrical resistivity methods. Environ Eng Blizkovsky M. 1979. Processing and application in Butler K. 1984. Microgravimetric and gravity gradient techniques for detection of subsurface cavities. 61:1050. study of monitoring sinkhole collapse by Kaspar M, Pecen J. 1975. Detection of caves in a karst formation by means of electromagnetic waves. 245

PAGE 263


PAGE 264

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2IntroductionSinkholes are a common cause of damage to residential buildings and other infrastructure in the Frank and Beck, 1991). Sinkhole activity can be manifested as recognizable topographic depressions sinkhole potential in the absence of such surface subsidence features is a challenge. Schmidt (2005) states that sinkhole investigations should be done in an integrated way that involves desk and site reconnaissance study, geophysical investigation, and geological interpretation, laboratory analysis and structural analysis of the site. Standard penetration tests (SPTs) and cone penetration tests (CPTs) are the investigation. SPTs are made by repeatedly vertically total penetration of 45cm is reached. The number of density of granular soils or stiffness of cohesive soils. raveling into a sinkhole cavity or a dissolution cavity blind drilling on sites would have low probability of intercepting a raveling zone and may instigate ground useful in identifying stratigraphic indicators of subsidence. These indicators include downwarping or, discontinuities total number of borings could be minimized, decreasing total cost and minimizing unnecessary ground collapse incidents (Dobecki et al., 2006). Abstract test (SPT) borings, which are located, in part, based discontinuities, or sudden increases in amplitudes. We analyze the degree to which the shallow features imaged used to examine (1) which SPT indicators show the regression analysis shows that the strongest correlations minimum SPT value increases from 0 to 20. Boreholes more likely to show zones of anomalously low SPT values Several statistical results suggest that raveling zones that connect voids to the surface may be inclined, so that STATISTICAL ANALYSIS OF GPR AND SPT METHODS FOR SINKHOLE INVESTIGATION IN COVERED KARST Henok Kiflu, Sarah Kruse Dept. of Geology, University of South Florida, 4202 E. Fowler Ave., Tampa, Florida 33620 USA, Michael Wightman Geoview, Inc., 4610 Central Ave., St. Petersburg, Florida 33711 USA, mwightman@geoviewinc.com247

PAGE 265

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE in dielectric permittivity of subsurface materials. In this residential sites respectively. The depth of penetration if ever image subsurface cavities. This is related to the penetration limitation and to the fact that the underlying cavity may not be directly below the site or survey raveling activity at depth which may result in downward migration of granular sediments from the shallow soil layers. This movement can make near surface granular soils less dense and result in downward deformation of cohesive layers. These processes result in recognizable features in the radar images, if they are within the range Neither ground penetrating radar nor SPTs are in active sinkhole (e.g. Schmidt, 2005). This argument has also been made by others, including Zisman (2001), who developed a scoring method for characterizing sinkhole potential of a site using geological and geotechnical factors. Zisman (2001) developed his criteria based on sinkhole anomalies and SPT records, we can address questions about the strengths and limitations of each method for detecting sinkholes. We can also test hypotheses about sinkhole structure. For our study characteristics of SPTs show the strongest correlation about the structure of sinkholes at these sites. To do this, sinkhole activity was suspected (Figure 1). Across these sites a total of 299 SPTs were run (Figure 1), or about strongest for SPT values from shallow depths.Study areaMost of the study area is characterized as lowland area, with Quaternary sediments overlying Tertiary carbonate rocks (Scott, 1988). Most of these sediments are unconsolidated sand, silty sand and sandy clay deposits that range in thickness from 0 to >60 meters (Figure 1). Other morphological features in the study area include plains, uplands, ridges and swamps. Carbonate rocks are but not at any of the residential sites studied. At the study sites, the mean depth to groundwater was 2.6m, with a minimum depth of 0.46 meters.Field MethodsConsistent methodology was used for ground penetrating radar surveys and geotechnical tests at the works by emitting high frequency electromagnetic waves into the ground with a transmitting antenna and while both antennas are pulled across the ground. The 248 Figure 1. Study sites (red dots), soil type, and overburden thickness in west-central Florida. Capitalized labels show counties. GPR surveys and SPT borings were conducted at each study site. Soil types and overburden thickness from Sinclair et al. (1985). County boundaries from FDEP (2013). GPR profiles were collected by GeoView, Inc and SPT borings by Ground Down Engineering, Inc.

PAGE 266

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2Binary logistic regression is a method for describing the relationship between an independent variable that can take on a range of values (e.g. SPT) and a yes or This method was applied to the entire data set of 299 minimum SPT value observed in each of the three depth zones, and the average SPT value observed in each of the three depth zones (Table 1). For each of these six at the SPT site was computed as a function of the SPT criteria value. If there were a perfect SPT threshold predictor of the features, recognized as locally downwarping layers, lateral discontinuities, and abrupt increases in amplitude, were then delineated. investigations, including drilling, soil sampling, of three boreholes were drilled in 95% of the sites. Borehole sites were chosen to include both areas within boreholes was 17.4m. Standard penetration tests (SPT) were conducted in all boreholes. SPTs were usually started at 1.8m depth and continued downward at 1.5m A number of methods exist to characterize the strength of soil based on SPT values (e.g. Carter et al., 1989). Following Meyerhof (1956) and Peck et al. (1974), granular soils are considered loose if they have an 249Statistical Analyses: Results and Discussion indirect and imperfect indicators of sinkhole processes, we can use neither as a direct proxy for the presence of a sinkhole. Thus it is valid to examine SPT data we do both, but we use different statistical methods because we are addressing different questions with each analysis. as a function of depth at each boring location, at intervals of no, no anomaly is observed at that point. SPT data from are considered per SPT site. These measurements typically some sites clays are encountered. To examine at least indirectly the role of the stratigraphy in the SPT readings, the SPT records are divided into three depth zones, as shown in Figure 2. For the analysis below, data were treated with using codes written in Perl and Matlab R2010a.SPT values as predictors of GPR anomalies Figure 2. SPT zones defined for use in the statistical analysis. For each boring, the average N-value and the minimum N-value are found for each of the three depth ranges.

PAGE 267

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE general. This implies that SPT criteria from shallow zones anomalies could be explained by a variety of phenomena. include (a) partially saturated unconsolidated sediment may be naturally loose without being disturbed by penetration in the presence of a shallow clay layer or threshold criteria. The probability would then decrease abruptly to zero at the threshold SPT value and remain sharper the plunge in the probability curve, the better the values, i.e. those with high minimum SPT values, indeed anomaly for 5 of the 6 SPT criteria. (The 6th criteria did anomaly will decrease by 84% (from 70% to 11%). Minimum SPT values in the intermediate zone are less anomaly drops by 68% (from 59% to 19%). Minimum SPT values in the deep zone show the weakest correlation: from 0 to 20 the probability drops by only 2. The SPT criteria for the shallow zones show much 250 SPT criteria Description sinkhole potential used in odds ratio analysis ShallowA Average SPT < 10 ShallowM Minimum SPT< 4 IntermediateA Average SPT < 15 IntermediateM Minimum SPT< 4 DeepA Average SPT < 20 DeepM Minimum SPT< 5Table 1. SPT criteria used in statistical analyses for Figure 3, Table 2, and Table 4. Figure 3. Logistic regression results for the 5 SPT categories (Table 1) that satisfy the 90% confidence level criteria. The horizontal axis shows the average of or minimum N-value over the defined depth range. The vertical axis shows the probability of observing a GPR anomaly coincident with the SPT location. Zero correlation would appear as a horizontal line. A perfect SPT threshold criterion would appear as a vertical line that would drop from 1 to 0 at the threshold N-value.

PAGE 268

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 To use odds ratio analysis a threshold SPT value must value, then the SPT is considered low. The threshold values are listed in Table 1. For example, when the SPT procedure. First, an optimization code searched for the threshold that showed the strongest correlation between to facilitate comparison with other studies. For example, for residential sites or for individual boreholes. Table odds ratios were computed for the six SPT criteria. The results are shown in Table 4. For inclusion in Table 4, we require that the null hypothesis (an odds ratio of 1) in the shallow and intermediate depth zones, in effect in most cases the odds ratios computed using minimum are also generally highest for shallow zones, lower for intermediate zones, and lowest or statistically is made using group 4 (Table 4). Several aspects of this statistical analyses support the with, but laterally offset from low SPT borings. Odds sinkholes that are no longer active; or (b) active sinkholes with shallow cohesive soil layers that gradually deform downward as one unit without disturbing its overall stiffness or density. Finally, one phenomenon that could common sinkhole are nevertheless spatially offset from each other. For example, material migrating into a cavity may migrate laterally or along an inclined path, contrary to the simple assumption of a vertical path. 251 SPT Criteria Square ShallowM 11.18 <0.01 IntermediateM 8.19 <0.01 DeepM 2.64 0.10 ShallowA 12.42 <0.01 IntermediateA 2.66 0.10 DeepA 0.58 0.45*Table 2. Model fit statistics for logistic regression shown in Figure 3. *=not significant at 90% confidence. GPR anomalies as predictors of low N-value SPT results independent variable. One applicable statistical method is odds ratio (OR) analysis. The odds ratio is simply the ratio of the probability of SPT are equal for boreholes drilled inside and outside this purpose, the six SPT criteria of Table 1 are used, and

PAGE 269

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE252 Boreholes drilled outside Boreholes drilled on residential sites Boreholes drilled on residential sites with anomalies Boreholes located on sites anomalies anomalies Boreholes drilled on residential sites with Table 3. Classifications for spatial correlation between SPTs and GPR anomalies. Data SPT Criteria SPT Zone Depth Range (m) SPT Threshold value Odds Ratio for ob serving SPT below threshold based on Average Shallow 10 2.22 0.0001 Minimum Shallow 4 2.27 0.0001 Intermediate 4 0.0156 Average Shallow 10 2.00 0.0017 Intermediate 15 1.76 0.0064 Minimum Shallow 4 0.0001 Intermediate 4 Average Shallow 10 1.41 0.0641 Intermediate 15 1.84 0.0147 Deep 20 1.95 0.107 Minimum Shallow 4 1.77 0.0188 Intermediate 4 2.24 0.0062 Deep 5 0.0709 Average Shallow 10 2.57 0.0001 Intermediate 15 1.70 0.0159 Minimum Shallow 4 2.89 0.0001 Intermediate 4 0.0019Table 4. Odds ratio analysis results for SPT categories with ratios significantly different from 1. The odds ratio is the ratio of the probability of observing a SPT value below the threshold on boreholes drilled on GPR anomalies to that for boreholes drilled outside GPR anomalies An odds ratio >1 implies that GPR data add value, in that SPTs on GPR anomalies are more likely to encounter zones with N-values below the threshold.

PAGE 270

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 values rather than average SPT values over a given depth range. Taken together, these observations suggest that raveling zones that connect voids to the surface Future analysis of this data set will seek to account for the effects of soil type, shallow clay layers, overburden sediment thickness, geology, and geomorphology on the and SPT values.ReferencesCarter M, Symons MV, 1989. Site Investigations and Foundations Explained. London: Pentech Press limited. applications to detect sinkholes and ground "Florida Department of Environmental Protection Frank EF, Beck BF. 1991. An Analysis of the Cause of Subsidence Damage in the Dunedin, Florida Area University of Central Florida, Orland, Florida. capacity of cohesionless soils, Proceedings ASCE engineering.2nd ed. New York(NY):John Wiley and Sons. Investigation Procedures For Evaluation of the Causes of subsidence Damage In Florida. Florida Miller, RL. 1985. Types, features and occurrence of Report. 85 (4126): 81 p. Terzaghi, K, and Peck RB. 1948. Soil Mechanics in engineering practice. New York (NY): John Wiley & Sons. link between surface water and ground water. United Zisman ED. 2001. A Standard Method for Sinkhole detection in the Tampa, Florida, Area: Environmental grouping (group 4) excludes boreholes drilled outside anomalies. The higher odds ratio for group 4 suggests Another result supporting the above hypothesis is the criteria, for both logistic regression analysis and odds ratio given vertical borehole may only encounter a portion of Finally, a third result supports the hypothesis that inclined both the logistic regression and odd ratio measures. Conclusions and SPT values within three depth ranges were used Florida. Logistic regression analysis was used to show the strength of the correlation decreases with depth. The strongest correlations are observed when 253

PAGE 271


PAGE 272

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2INTEGRATED GEOPHYSICAL METHODS FOR GROUNDWATER EXPLORATION IN A KARST AREA WITH OR WITHOUT Fuping Gan, Yixiang Chen, Wei Zhao, Yuling Chen, Wei Liu Institute of Karst Geology, CAGS, Guilin, Guangxi 541004, China Karst Dynamics Laboratory, MLR&GZAR, Guilin, Guangxi 541004, China the audio frequency telluric method (AFTM) has been successfully employed to detect water in karst settings of water in combination with the induced polarization (IP) method (Li, 2009). The typical advantages of this method this is countered by disadvantages, including a shallow investigation depth and a lack of anomaly variations with depth. Electrical resistivity tomography (ERT) to delineate aquifers (Kumar, 2012), to search for karst of the resistivity variation with depth. Compared with using a single geophysical method alone, the integrated approach usually provides more reliable information, and as a result has been widely applied for groundwater investigations. Alexopoulos (2011) employed both the very low frequency (VLF) electromagnetic method and ERT to map water pathways. Vargemezis (2011) even together to optimize locations for the construction of hydro wells. Two examples presented here to trace karst water show that the AFTM and the ERT methods have the advantages Geophysical methods Audio frequency telluric method (AFTM)This method, which takes advantage of natural telluric current variations with frequencies induced in the earth by phenomena such as solar emission and thunderstorms, detects conductive difference distributions underground and interprets them in terms of geology or hydrogeology. The AFTM is a preferred method for rapid ground reconnaissance and shallow exploration where AbstractBecause of heterogeneity and anisotropy, it is very locations in karst areas using only hydrogeological audio frequency telluric method and electrical resistivity exploration in karst areas with or without thin cover. In the case studies presented here, audio frequency location and strike of a karsted or fractured zone where is designed to reconstruct the resistivity structure with a Schlumberger array. By combining the geophysical results with available hydrogeological data, an optimal drilling site can be established. This integrated geophysical approach for karst water exploration has the method is reasonable and useful.IntroductionSince 2009, extreme climate events, persistent drought and low rainfall, has made drinking water scarce for human consumption and agricultural purposes in some parts of China, especially in karst regions. At the beginning of 2012, a groundwater exploration team was drought disaster relief. The team traveled to Tai'an city, Shandong Province, in the east of China (Figure 1) to search for promising exploratory sites for karst water using hydrogeological investigations and geophysical optimally positioned and drilled. Subsequently, 21 were tested by pumping, and produced an abundance of water. environments, suitable geophysical methods needed to be 255

PAGE 273

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE Apparent resistivity data were collected and stored automatically using 60 electrodes with a 10 m spacing. because of its suitability for the study of horizontal and vertical structures. The results are produced using Res2dinv software, which calculates resistivity and depth values for geological investigation depth mainly depends on the total length of the spread. Case study 1 Geological setting Shangdong Province, China. The widely distributed overburden in the region, with a thickness of about 10 m, is conducted in situ to determine potential gradient measurements along a line, which is usually set up perpendicular to the strike of geological structures. The equipment used to measure the potential gradients was stations were usually 10 m apart. Because telluric currents change with time, the potential gradient between within an hour). In the end, the potential gradient was plotted against the midpoint of the potential electrodes. This method is often used to map shallow subsurface karst features showing relatively low potential values if Electric Resistivity Tomography (ERT) (manufactured by Chongqing Benteng Digital Control 256Figure 1. Location and geological map of the project area in Taian City, Shandong Province, China 1. Granite, 2-4. Limestone (1, 2, and 3), 5. Dolomite and Limestone (O1) 6. Sandstone and breccia (E), 7. Quaternary, 8. Taian City, 9. Wen River, 10. Fault, 11. Guanlu, 12. Momoshan.

PAGE 274

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2formed locally, are targeted as karst water reservoirs. A hydrogeological map shows their distribution in Figure 2.Geophysical methods horizontal trend of zones of lower potential and then the ERT method was used to vertically probe lower resistivity variations with depth. somewhere underground. In the end, the regions of lowest potential from all two main branches of the low potential anomalies in wide in the south. These lower potential anomaly zones, because of their wider distribution, are inferred to be consists of Quaternary deposits of sand and gravel. The bedded limestone or argillaceous limestone embedded Cambrian), as shown in Figure 2. eastward about 600 m to the north, coincides with the south to north following two discharge mechanisms. The scattering discharge as a result of the shallow incision of the valley, the low hydraulic gradient of groundwater, and the wide areal distribution of the groundwater. The second involves abundant volumes of groundwater The condition of the landforms favors groundwater one disadvantage is that contaminated water may intrude into the system from the Wen River as a result of the extraction of karst groundwater. Regional tectonic structures trend from northwest to 257Figure 2. Hydrogeological map of survey site and geophysical profiles (1. Survey site, 2. Surface water direction, 3. Drill hole, 4. Fault, 5. Quaternary, 6-10. Cambrian Limestone, 11. Recommended borehole, 12. Fracture zone, 13. Profile direction, 14. Station/profile No., 15. Dip direction and dip).

PAGE 275

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEkarst and valley topography. The maximum elevation difference in the region is about 150 m and the valley positioned with only small changes in elevation, are located across a hill slope showing bare karst rocks. argillaceous limestones or edgewise limestones embedded with thin shale (upper Cambrian). Karst moves from north to south in a direction that is nearly opposite to the dip direction of the beds. A map of the hydrogeology is shown in Figure 5. a depth of about 100 m and the location of the survey to search for groundwater resources. Since 1949, three dry wells have been drilled nearby. From a landform perspective, high in the north and low in the south, it is likely that groundwater in this region moves southward Therefore, appropriate geophysical methods can be chosen to outline these favorable zones.Geophysical methods constrain the position and orientation of groundwater runoff zones, then the ERT method was applied to Figure 5. The zone of lower potential anomaly is (see Figure 6). Because of site limitations, only one ERT array resistivity contours in Figure 7 (upper) readily Regions of lower resistivity determined by the integrated favorable well positions. Because of site limitations, the the similar characteristics of the anomaly there. Subsequently, the ERT line was set up along the strike of the lower potential anomaly. In Figure 4 (upper), Schlumberger array apparent resistivity contours, taken from the ERT for line 5, show a clear contact between In particular, the striped low resistivity zone between in Figure 4 (lower) was predicted from the Schlumberger resistivity section in the ERT to consist of a subsoil of alluvial clay, sand and gravel that is about 26 m thick. 26 70 m and 90 110 m, respectively. These predictions were tested by hammer drilling. The resulting well is 150 m deep, with a static water level highly fractured nature of the rock that caused the pipe to stick often. After pumping tests, the pumping capacity was set at 480 m Case study 2 Geological setting at Momoshan village, approximately 20 km west 258Figure 3. Potential curves of the AFTM for lines 1, 2, 3, 4 and 5.

PAGE 276

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2259Figure 4. Resistivity imaging (upper), Geological interpretation for profile 5 (lower) 1. Fine sand with gravel, 2. Gravel with fine sand, 3. Limestone, 4. Karst fissures, 5. Interbedded limestone and shale. Figure 5. Hydrogeological map of the survey site and geophysical profiles 1. Survey site, 2. Surface water direction, 3. Fault, 4. Quaternary, 5-8. Cambrian Limestone, 9. Fissure zone, 10. Survey line direction, 11. Station/ Line No., 12. Recommended borehole, 13. Dip direction and dip.

PAGE 277

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEConclusions shown that an integration of AFTM and ERT to optimize data acquisition and provide crosschecks can produce promising results. Fieldwork procedures usually consist of the following steps. 1. collected primarily to roughly establish the that may focus karst development and result in depth estimates are not generally possible with this procedure alone. 2. Second, to constrain the geometry of horizontal and vertical geological structures, ERT data low potential anomalies, thereby enabling the detailed interpretation of karst geological structures by using information about resistivity variations with depth. For a Schlumberger array consisting of 60 electrodes spaced at 10 m, the maximum depth of penetration is about 150 m. The well depth is 190 m with a static water level at a depth of about 80 m. After pumping tests, the yield was determined to be 280 m drawdown of 40 m. The flow comes mainly from units at depths of about 127 and 161 m.260Figure 7. Resistivity imaging (upper), Geological interpretation for line 1 (lower) 1. Clay, 2. Limestone, 3. Karst fissures, 4. Interbedded limestone and shale.Figure 6. Potential curves of the AFTM for lines 2, 3, and 4.

PAGE 278

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 Metwaly M, Elawadi E, Moustafal SR. 2012. Saudi Arabia. International Journal of Physical Sun JL.2002.Applicaton of natural audio frequency islandsthe island of Olib, Croatia.Environ Monit Assess.184:6211. Available from: karst cave probing. Site Investigation Science and The two methods taken together constrain lower potentials distributed horizontally (along strike) and similar changes in resistivity varying vertically. When combined with hydrogeological information, this procedure provides important and effective geophysical indications of the geological setting, and enables the optimal determination of well positions.ReferencesChen S. 1988. The use of selected frequency of natural exploration in the arid and semiarid areas tomography technique in mapping shallow Zaidi FK, Kassem OMK. 2012. Use of electrical resistivity tomography in delineating zones of groundwater potential in arid regions: a case study from Diriyah region of Saudi Arabia. Arabian focusing approach to ground water detection by means of electrical and EM methods ,the geophysical surveys to assess the structural conditions of a karstic cave of archaeological Sciences 5:17. Alexopoulos JD, Dilalos S, Vassilakis E.2011. Meteora). Advances in the Research of Aquatic Environment Vol. 2. Available from: prospecting in granite region by comprehensive geophysical methods. Site Investigation Science 261

PAGE 279


PAGE 280

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 by such type of risk. In particular, Marsala (western end of Sicily, Figure 1) was affected in the past decades by several collapse phenomena which caused extensive damage to infrastructures and population. These phenomena are related to the presence of underground quarries for the extraction of calcarenites used as building materials. These cavities are carved at depth varying from several meters to roughly 25 m on a single or superimposed layer, following the excavation techniques of chambers and typically occurred without any planning. With time, the underground quarries were progressively abandoned for several reasons (i.e., interaction with the excavation work). Nowadays in many cavities signs of instability can be clearly recognized along ceilings, walls and pillars. These instabilities easily propagate upwards causing sinkhole and subsidence phenomena according to the mechanisms described in Parise & Lollino (2011), thus creating consistent damages to buildings and infrastructures. A sinkhole, occurred in June 2011 and related to an underground quarry in the eastern sector of Marsala, is described in this paper as a case study (Figure 2). The site was selected for the availability of topographic data of the underground quarry, prior to the formation of the Abstract of Sicily (southern Italy) representing a great risk to people, buildings, and infrastructures. These phenomena are generally associated with the presence of ancient underground quarries for the extraction of calcarenite rock, used for building or ornamental materials. These quarries were poorly constructed and abandoned throughout history. in the rock mass, and numerical modeling were factors responsible of the instability processes of underground quarries. The genetic mechanisms of anthropogenic sinkholes have also been investigated. Jointing and saturation conditions of the calcarenite, along with indiscriminate enlargement of voids, can reduce the available strength. This strength is needed to maintain the stability of the rock mass above the underground quarry, the lack of which can cause the sinkholes formation. Finally, a comparison between the cases of Sicily and Apulia regions, where similar anthropogenic sinkholes are widespread, was carried.IntroductionAnthropogenic sinkholes and subsidence phenomena are very widespread and generally linked to the including water, oil, and gas (Waltham, 2002; Waltham et al., 2005). EXAMPLES OF ANTHROPOGENIC SINKHOLES IN SICILY AND COMPARISON WITH SIMILAR PHENOMENA IN SOUTHERN ITALYMarco Vattano, Cipriano Di Maggio, Giuliana Madonia marco.vattano@unipa.itMario Parise, Piernicola LollinoCNR IRPI, Via Amendola, 122/I, Bari 70126, Italy, m.parise@ba.irpi.cnr.itMarco BonaminiVia Trinacria 8, Palermo 90144, Italy, 263

PAGE 281

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE264Geological settingIn the Marsala area Lower Pleistocene calcarenites, Unti, 1974; Ruggieri et al., 1975) and referred to the Marsala syntheme, crop out (Di Maggio et al., 2009). They are formed by three main lithofacies with b) sands; and c) gray sandy clays. Both b and c types are rich in microfossils. The calcarenites and the sands show a gradual lateral transition to sandy clays and silts which generally can be found at the base of the entire sequence. sinkhole, which were useful for defining the genetic mechanisms of this depression. The quarry was surveyed in 2000. During this survey various signs of instability, such as open fractures and detachment of large blocks, were observed on the vaults and at the pillars. After the June 2011 sinkhole, new surveys were conducted inside the quarry aimed at collecting additional data to better understand the factors that have played a role in generating the sinkhole (Figure 2). Since the collapse occluded many of the remaining rooms, the 2011 surveys were limited to the western surface topographic surveys were conducted in the following months to detect new signs of collapse or movements of the topographic surface. In this article, we intend to describe the geological and structural features of the site study, and illustrate the preliminary results about the genesis of sinkholes modeling was implemented following a similar procedure applied in other site studies in southern Italy (Parise & Lollino, 2011; Parise, 2012; Lollino et al. Sicily and similar anthropogenic sinkholes of Apulia region was made, in order to gain a better understanding of the mechanisms causing the sinkholes.Figure 1. Planimetric view of the studied underground quarry. In the inset, location of the Marsala area. Figure 2. The sinkhole formed in June 2011, in correspondence of the quarry in Figure1.

PAGE 282

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2265 perimeter, shape and depth of the sinkhole. The survey was carried out using a topographic station, measuring at points with a grid spacing of 2.5 m, because of the presence of thick vegetation and the pronounced topography (Figure 7). The sinkhole has an elliptical shape with a maximum depth of 2.4 m. It is asymmetric with the deepest point both the underground galleries and the main tectonic discontinuities of the area. During the months following the formation of the sinkhole a total of 4 survey campaigns were performed in order to evaluate the likely progressive evolution at According to Ruggieri et al. (1975) and Arces et al. (2000), the calcarenitic lithofacies can be divided into three lithotypes: i) coarse calcarenites and calcirudites calcarenites with thickened grains; iii) medium to coarse calcarenites in irregular strata with intercalations of thin beds of silt. Usually the ii) type was carved in the Marsala Calcarenite is covered by Middle and Upper Pleistocene marine terraced deposits. In the underground quarry, chosen as a study case, the mining activity Figure 3. Sinkhole formed in June 2011. A: Breakdown deposits within the quarry in the proximity of the area affected by the sinkhole. B: Damage caused by the sinkhole to the overlying buildings. Figure 4. Stereographic projection (Schmidt, lower hemisphere) of the poles of the fracture plans in the calcarenite rock mass, as from the survey within the quarry (data elaboration by means of Daisy 3; Salvini, 2011). B: Main development directions of the underground passages.The sinkhole and the underground quarryIn June 2011 a sinkhole occurred in correspondence of an underground quarry, damaging several constructions, including a recreation centre and the

PAGE 283

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE266addition to falls. This sector is located in the area where the sinkhole occurred. Along these fractures, saturation and chemical alteration due to the percolating water are to the walls, with presence of both incipient fractures and existent to excavation of the quarry, and fractures caused by the local instability of rocks inside the quarry. ceilings of different rooms, and are mainly tensile fractures with visible signs of rock alteration due to the water circulation. In the second type, along the not show any consistent variation of the topographic The underlying quarry consists of a series of rooms with galleries, separated by thin rock walls, are present in the central sector of the quarry (Figures 1 and 4B). The room size is highly variable: the average height is 2.7 m, ranging between 1.1 and 7.5 m.; the average width (2.6 40 m). The walls are 20 cm to 4 m thick, whilst the length between 70 cm and 6 m. Overall, the quarry is 480 m long, with overburden varying from 8.2 to 11.8 m. The calcarenite was extracted along variable directions. The of the rock and the perimeter of the land properties. Shape and distributions of the rooms display very different directions in the western sector of the quarry. The central and eastern sector of the quarry, instead, show preferential excavation direction as inferred by the presence of rooms and galleries tens of meters long discontinuities (Figure 4). The enlargement of the quarry following the principal discontinuity directions was necessary to extract entire blocks of calcarenite. Figure 5. Instability phenomena in the quarry vaults. Falls (photo A) and movements of large volumes of rock along sub-vertical fractures (photo B) surveyed in 2000, in correspondence of the sector where the sinkhole formed. Instability phenomena inside the quarryInside the quarry, signs of instability at different stages of evolution are visible along the ceilings, the walls and the pillars. Instabilities are generally due to the fracture system of the calcarenite body, the dense and irregular distribution of rooms and galleries, and the rock alteration caused parallel or coincident with the bedding planes caused detachment of large volumes of rock from the vaults (Figure 5). According to the 2000 survey, falls were mainly located in the eastern sector of the quarry where rooms are larger thick. In this sector, movements of large portions of rocks

PAGE 284

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2267 to the recommendations provided by the International Society for Rock Mechanics (ISRM). In particular, Arces saturation ratio on the uniaxial compression strength of the calcarenite samples, with strength values under c corresponding values for saturation degree equal to zero c also observed for the modulus of elasticity, measured at 50), with values ranging 44 MPa. In the following model, the value of the tensile strength t t conditions, the shear strength parameters according to to simulate conditions of low saturation degree of the calcarenite rock mass. These parameters have been obtained as a linear approximation of the shear strength envelopes according equal to 95. in Table 1. discontinuities, the deformations of the calcarenitic rock are visible with outward protrusion of wedges. Both types of fractures can isolate slices of rock of variable thickness which at later stages can cause collapse. systems with different inclinations and aperture ratio. the quarry, while others appear to have been caused by the high level of stress of the calcarenitic rock pushing the pillars. These fractures can isolate blocks of metric size, causing failures from the corner of the pillars. In some cases, the fractures can run along the entire length of the pillars, causing a reduction in thickness and, eventually collapse (Figure 6).Finite element analysisA detailed geotechnical characterization of the Pleistocene calcarenite of Marsala is described in Arces et al. (2000). Figure 6. Sub-vertical joint crossing a quarry pillar. Dry Calcarenite Saturated Calcarenite (S Saturated Calcar enite ) 15 15 15 50 (MPa) 600 165 110 75 t (kPa) 200 c (kPa) 2000 Table 1. Parameters adapted in the FEM analysis.

PAGE 285

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE268 implementing mechanical properties representative of out (Analysis 1); thereafter, an analysis implementing parameters representative of saturated conditions for the calcarenite, in the hypothesis of a rock mass characterized been performed. Such analysis is representative of conditions of calcarenite saturation at depth as an effect as typically observed during in situ surveys. Moreover, an analysis assuming saturated conditions of the calcarenite and strength parameters representative of a moderate used accordingly. Aimed at the reconstruction of the conditions that formed the sinkhole in the eastern sector of Marsala, a been performed. Indeed, this method allows to simulate rock masses affected by the presence of underground analysis has been implemented with the numerical code 2D section 6 shown in Figure 7. The constitutive model The analyses have been carried out in drained conditions and assuming the whole domain above the water table, Figure 7. A: Detailed topographic survey of the June 2011 sinkhole. Dashed lines indicate traces of the cross sections shown in figure 7B; crosses mark the topographic measurement stations. B: Cross sections of the sinkhole and the corresponding underground quarry sectors. Dashed line indicates the topography before the occurrence of the sinkhole. Figure 8. Analysis 1 (calcarenite with low degree of saturation; GSI = 95): plastic points in the pillars of the underground quarry (above), and shear deformations in the underground quarry (below). Vertical scale is meters above sea level.

PAGE 286

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2269ground surface as a consequence of the failure process failure mechanism which is in good agreement with the surveys (see Figure 7). implies a lack of convergence of the numerical model and consequently unstable conditions of the rock mass fracturing to enhance the failure process that led to the occurrence of sinkhole. to K0 state of the model. Afterwards, the excavation stages have been simulated in the model up to the current geometry of the cavity rooms. When stable conditions resulted from the model, a reduction analysis (Matsui & San, 1992), reproducing a gradual reduction of the strength properties of the rock mass up to the global failure, has been performed in order to derive the safety factor. results show stable conditions for the rock mass. and the accumulation of shear strains (Figure 8) within the pillars delimiting the excavation rooms. Based on the has been deduced in this case. rooms, with inclined plastic zones that propagate from the associated failure mechanism is shown in terms of contours of incremental shear strains and in terms of Figure 10 suggests a failure process characterized by an Figure 9. Analysis 2 (S = 1; GSI = 95): plastic points in the pillars of the underground quarry. Vertical scale is meters above sea level. Figure 10. Analysis 2 (S = 1; GSI = 95): shear strains in the region overlying the underground quarry (above), and final profile of the ground surface as a consequence of the collapse (below). Vertical scale is meters above sea level. A comparison with similar situations in southern ItalyThe situation described in the present paper for the underground quarry at Marsala is certainly not limited to this territory, or to Sicily. Presence of ancient

PAGE 287

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE270such subterranean activity are represented by both volcanic rocks and recent calcarenite deposits (as in the Marsala case study). Vulnerability of the sites is greatly increased by loss of precise locations of underground quarries, and by unavailability of a survey depicting the real directions of development of the subterranean passages, their sizes, These issues should be addressed in order to understand the instability conditions underground, and to reduce the risk of sinkhole collapses in the future.Conclusions In situ surveys, and numerical analyses carried out with hypothesize the formation conditions of anthropogenic sinkholes in Marsala (Italy). Preliminary results indicate that the genesis of the sinkhole can be attributed to the saturation of the calcarenite, along with the presence of discontinuity systems. These conditions can decrease the strength needed for the stability of the rock mass above the underground quarry. In the case study the use of 2D modeling creates resolution limits. To obtain a more realistic representation of the investigations from our working group. In addition, it is important to notice that even if the effective quarry area is 2000 m2, only 1500 m2 are pillars, walls) over the entire area. As a consequence, after the formation of the quarry, the weight of the rock portion above a chamber is distributed over a surface reduced of of the initial area. The stress on the pillars, which are often not thick enough, increases 4 times. The case study here presented is one example of many situations that exist in Sicily, as well as in several other regions of Italy, due to the presence of underground quarries. Caving exploration and surveys, structural analyses (including evaluation of instability features), topographic monitoring at the surface, and geotechnical underground sites used for the extraction of building and ornamental materials is extremely common in Italy, especially in the southern regions, and in those areas (i.e. Tuscany, Latium, Campania) where local geology is characterized by volcanic rocks. It is worth here to recall the many experiences carried out during the last decades in Apulia, probably one of the regions with the highest number, and typological variety Among many types of cavities, underground quarries have produced many subsidences and sinkholes (Parise, 2010, 2012). This because these sites, that were originally located at the outskirts of towns, have been progressively expansion of the urbanization, so that in many cases the constructions realized in the last decades have been built above the quarries. The cases of Altamura, documented (Parise, 2012; Pepe et al., this volume). A very high percentage of the historical parts of Apulian towns are built above a complex subterranean network of quarries. Looking at the chronology of sinkhole events so far reconstructed for Apulia (Parise & temporal reference) can be attributed to underground quarries.There is therefore a very critical situation in terms of civil protection, as in Sicily; in this latter region, however, the problems are still greater, due to the fact that the rocks potentially interested by Figure 11. Analysis 3 (S = 1; GSI = 80): plastic points in the pillars of the underground quarry. Vertical scale is meters above sea level.

PAGE 288

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2271 southern Italy. These proceedings. calcarenite di Marsala (Pleistocene inferiore) e i Universit di Roma Tre, Roma. Waltham T, Bell F, Culshaw M. 2005. Sinkholes and Subsidence: Karst and Cavernous Rocks in Engineering and Construction. Berlin: Springer. modeling, represent some of the standard procedures used to evaluate the real possibility of propagation of the underground failures toward the surface. Further, it allows researchers to obtain precious information to design rehabilitation works, or to help decision makers in the choice of the most proper action for the involved sites.AcknowledgmentsThe authors would like to thank Francesco Patti of the Municipality of Marsala for the access to the study suggestions greatly improved the quality of the article. Many thanks to the reviewers Thomas L. Dobecki and References Pleistocene Calcareniti di Marsala: geotechnical and sedimentological characterisation. Mem. Soc. Di Maggio C, Agate M, Contino A, Basilone L, Catalano R. 2009. Unit a limiti inconformi utilizzate per la survey and numerical modeling of the potential failure mechanisms of underground caves. Matsui T, San KC. 1992. Finite element slope stability analysis by shear strength reduction technique. Parise M. 2010. The impacts of quarrying in the Apulian karst. In: Carrasco F, La Moreaux JW, Duran Valsero JJ, Andreo B. editors. Advances in research Parise M. 2012. A present risk from past activities: sinkhole occurrence above underground quarries. Parise M, Fiore A. 2011. Chronology of sinkhole Parise M, Lollino P. 2011. A preliminary analysis of failure

PAGE 289


PAGE 290

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2DEVELOPMENT OF SINKHOLES IN A THICKLY COVERED KARST TERRANE Sam B. Upchurch, Thomas L. Dobecki, Thomas M. Scott, Steven H. Meiggs, Sarah E. Fratesi, Michael C. AlfieriSDII Global Corporation, 4509 George Road, Tampa, Florida, 33634, USA, duration hydraulic head stresses can result in rapid sinkhole development. The three investigations relate to sinkhole activity in different stages of development, but with apparently similar origins. From these three investigations, a case can be made for (1) rapid dewatering and consolidation or form sinkholes.The Three InvestigationsThe three investigations are discussed below in chronological order of occurrence.Tampa Bay Regional Reservoir InvestigationIn 1999, we completed site characterization for construction 1999; Dobecki and Upchurch 2010). The site is located on the Polk Upland Physiographic Province (White 1970) and is underlain by a thick (up to 120 m) sequence of clay, sand, and limestone and dolostone of the Peace River and (Scott 1988; Arthur et al. 2008). The Miocene strata form an effective aquitard for the underlying limestone of the upper Floridan aquifer (UFA). discovered as part of the reservoir investigation (Figure 2). The reservoir embankment footprint was altered to avoid one feature that was of concern because of loose sediments in the subsurface; the second feature sandy clay typical of the Miocene of Florida. This unconformity.Abstract in 2010. The sinkholes resulted in millions of dollars in losses and caused us to revise our thinking about how sinkholes form in a terrain normally considered to have expansive clay, sand and sandy clay, and carbonate primarily carbonate and is up to 90 m thick. The upper clay and sand with minor amounts of carbonate and is effective aquitard for the underlying upper Floridan aquifer (UFA), which is composed of karstic, Oligocene and Eocene limestone and dolostone. A rapid drawdown of up to 20 m in the potentiometric surface of the underlying UFA resulted in mobilization consolidation resulting from dewatering and loss had developed as early as the Miocene Epoch. Stable, discovered in an earlier investigation in the same area sinkholes in the area. Combining information gathered from study of these modern and relict sinkholes presents evidence of sinkhole development mechanisms in the IntroductionThis paper synthesizes evidence from three karst investigations suggesting that migration or consolidation 273

PAGE 291

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEThe latter two features were geotechnically stable and levels at the reservoir are regularly monitored and it is clear that these relict sinkholes are isolated and stable. The latter two features are deemed to be safe for impounding water within the reservoir because they do not react to sudden stresses caused by drawdown in the underlying Floridan aquifer or to changes in reservoir stage. This is not the case for the two case studies that are described below.The Plant City Sinkhole Cluster InvestigationIn January 2010 a hard freeze with overnight temperatures UFA for irrigation to protect crops from freezing. The potentiometric surface of the UFA declined up to 18 m, with up to 9 m daily excursions in potentials. As a result, Testing of many of these sinkholes by the authors revealed a persistent pattern: there was a thick (up to 274 Figure 1. The Tampa Bay area of west-central Florida. Locations of the three investigations are shown in red, and known sinkholes that have formed since 1984 are shown in violet. Figure 2. North-south cross section through the reservoir to illustrate the locations and stratigraphic context of the Polk Upland sinkholes. Source: Dobecki and Upchurch 2010.

PAGE 292

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 moisture contents that were at or near the liquid limits of the clays. Under standard penetration (SPT) testing, surface sediments, there was little evidence of suffosion; resulting in sinkhole depths of up to 5 m and diameters The Southeast Hillsborough Landfill InvestigationIn December 2010 there was a second, but less severe, episode of rapid drawdown of the UFA as a result of pumping for freeze protection. Shortly after this event a 45 m wide and nearly 60 m deep sinkhole (Figure 5) km south of the reservoir site. sand, clay and limestone strata (Figure 6). The unique feature of this sinkhole is that there is no evidence of a bottom of the aven. void space where the materials had been removed from the stratigraphic column to form the aven and The sediments in this larger void space had also been previously removed. While the waste mass or the shallow limestone bed (Figure 6) may have bridged the void, a mechanism was required for the removal of the missing siliciclastic sediment. Refusal strength (N > 50 275Table 1. Sample results from a standard penetration test boring adjacent to a sinkhole in Plant City. Lithology N Values* clayey sand seams were approx. 1.5 m thick clay for 5.5 m) chert Figure 3. A large sinkhole that developed by vertical movement near Frostproof, Florida. Note the intact trees in the down-thrown sediment plug.Figure 4. Sinkhole that developed in a suburb of Plant City. This was one of seven on the street. Sand has been placed in the foreground to reduce risk of additional damage to the street.Figure 5. The landfill sinkhole in southeastern Hillsborough County, Florida. Note the slump features developed within the waste mass.

PAGE 293

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEclay, sand, and carbonate units (Figures 2 and 6). the underlying UFA. When extremely heavy withdrawals of groundwater from the UFA occur, lack of concomitant leakage from overlying water surface. 2. Karst features, including sinkholes, began unconformity; Scott 1988). Some sinkholes features thought to have formed this way at sediments that appear unaffected by modern hydraulic stresses. Elsewhere, some of these features appear to have been well consolidated, resulting in localized surrounds most of the void. As such, the walls of the the perplexing question is: where did the collapse material go? It appears that the sediments that occupied the void space had been washed out of the space at some earlier time, perhaps during the freeze event in January 2010 or even earlier. This would explain the absence of breakdown debris at the bottom of the void. The void, therefore, was either bridged over by the upper limestone, which was not detected as rubble at the base of the void, or perhaps the waste mass itself.276Figure 6. North-south cross section of the landfill sinkhole. (Note that depths are presented in feet and that the cross section was developed from standard penetration testing after initial stabilization of the sinkhole and backfilling to allow safe access for testing. CLSM is cementitious Controlled Low Strength Material used to cap the aven and stabilize the site) Source: on-going investigation by the authors. EpilogueBased on comparisons of the results from the three investigation areas, we suggest that the Polk Upland sinkholes developed as follows (Figure 7): 1. The Polk Uplands are underlain by the Miocene

PAGE 294

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2that links these sinkholes genetically with the Southeast The Polk Uplands remain a low sinkhole probability area, but when sinkholes develop they can be locally common, large, and catastrophic. The ones we investigated hydraulic gradients caused by groundwater withdrawals Florida Water Management District, has taken steps to water caution area and adopting additional regulations related to water use and sources. ReferencesArthur JD, Fischler C, Kromhout C, Clayton JM, Kelley, the Southwest Florida Water Management District. 66, 102 p. 59 plates. to site characterization C.W. Bill Young Upchurch SB, Dobecki TL, Daigle DM. 1999. investigation. In: Law Engineering Services and White WA. 1970. The geomorphology of the Florida pockets of soft clay with natural moisture contents near or exceeding their liquid limits. 4. caused by groundwater withdrawals occur in the potentiometric surface of the UFA, some of these may fail, either by simple dewatering and rapid consolidation or by migration of the clay and associated sandy sediments into the voids of the As shown in Figure 7, our current concept as to how these sinkholes form is by migration of clay and clayey sand or the underlying Oligocene and Eocene limestone of the and clayey sands allow them to migrate farther laterally than might be otherwise expected into void space not directly beneath the sinkhole. This migration explains why there was no evidence of breakdown or collapse Clay and clayey sand remain under the sinkholes we tested near Plant City. In this case, either the clay and clayey sand were simply dewatered and consolidated or migration into nearby void space was incomplete. While this clay and clayey sand sediment is poorly sudden migration of water out of the clay mass would be hindered by the low intrinsic permeability of the material. Therefore, partial physical migration of the 277 Figure 7. Potential mode of sinkhole formation within the clay-rich strata of the Hawthorn Group on the Polk Uplands of Florida.

PAGE 295


PAGE 296

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2PALEOKARST CRUST OF ORDOVICIAN LIMESTONE AND ITS CAPABILITY IN RESISTING WATER INRUSHES IN COAL MINES OF NORTH CHINALin Mou, Gongyu LiXian Research Institute of China, Coal Technology & Engineering Group Corp., Xian, Shaanxi 710054, China,, pressure is typically measured at monitoring wells in the Middle Ordovician limestone while the limestone has equally important parameter. Similar approaches are used in other countries (Verbovsek and Veselic 2008; Permian coal seams, the potentiometric pressure of the Ordovician groundwater becomes higher, indicating a higher risk of water inrush. It is particularly important to study the heterogeneity of karst development in the underlying Middle Ordovician limestone and determine any impermeable strata that may prevent the pressurized karst water from bursting into coal mines. Detailed analysis of the exploratory borehole data suggests presence of a paleokarst crust (hereinafter referred to as crust) at the top of Middle Ordovician, which is often referred to as the Fengfeng Formation capacity makes recognition and application of the crust less straightforward in coal mining.Studies on Paleokarst CrustStudies by Li and Wang(1997)indicate that the Early years (Ma) of the Middle Ordovician, which resulted in the ubiquitous Bauxite mudstone. The top Middle Ordovician was further compacted and any fractures of weight of the overlying strata (Li and Wang 1997; Li to illustrate this concept. Between 2009 and 2011, Bai and others (2009a; 2009b; 2010; Miu and Bai 2011) controlling water hazards and utilizing water resources. Abstract Permian coal seams in North China, it is particularly important to study the heterogeneity of karst development in the underlying Middle Ordovician limestone and determine any impermeable strata that may prevent the pressurized karst water from bursting into coal mines. Detailed analysis of the exploratory borehole data suggests presence of a paleokarst crust at the top of Middle Ordovician Fengfeng Formation. Because of its mechanical strength and low permeability to water, resisting layer. This paper takes Sihe Mine of Shanxi Province as an example to study the geotechnical and hydrogeological characteristics of the paleokarst crust. Incorporation of this additional hydrological barrier led to more minable coal seams in the coalmine.Introduction Permian coal seams are widespread in North China, as shown in Figure 1. Mining of the coal seams in the past few decades has led to multiple theories and techniques of evaluating the pressurized water from the limestone into underground mining areas. These theories emphasize the potentiometric pressure in the Middle Ordovician limestone and the aquifuge between the coal seams and the Middle Ordovician limestone. The term aquifuge which contains no interconnected openings or interstices and therefore neither absorbs nor transmits water and it is interchangeable with the term aquiclude. The most potentiometric pressure; and M is the thickness of more likely occurs a water inrush. The potentiometric 279

PAGE 297

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE Weathered and leached zone Upper section with gypsum Thick micrite section Lower section with gypsum Of the four zones, the lower section appears to have a analysis, Wang and Bai (2009) analyzed the lithology, karst pore structure and distribution in the Fengfeng Shanxi Province. Their results show approximately 140 meters (m) aquifuge in total at the top and bottom of the Fengfeng Formation. Bai and others (2009) proved the presence of the aquifuge by statistic analysis of data The presence of aquifuge may explain why the water inrush did not occur in the theoretically predicted water bursting areas, while it occurred in the anticlinal development areas (Wang and Bai 2009). Fu and others (2010) took samples from the Fengfeng Formation in Wang Jialing Mine in Shanxi Province and divided the Fengfeng Formation into four zones: 280Figure 1. Distribution of coal resource and aquifuge in North China. Characteristics of Paleokarst Crust at Sihe MineSihe Mine is in Shanxi Province (Figure 1), which is a large anthracite base with an annual production of 1,200 million tons. The average thickness of the Fengfeng Formation is approximately 100 m. The distance between the lower coal seams and the Middle Ordovician limestone is approximately 20 m. The

PAGE 298

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2water inrush has become a key factor restricting its sustainable production at the mine.Lithology of Fengfeng FormationFigure 2 shows the relationship between the coal seams and the underlying limestone formations. The Fengfeng Formation at Sihe Mine consists of cyclic deposition of pure limestone and argillaceous limestone with variable in its unaltered state. This cyclic depositional structure is common in the middle of the Fengfeng Formation with a thickness between 20 and 50 m. The top 20 m of the Fengfeng Formation is pure limestone which is the with calcite and weathering residuum. This altered top section of the Fengfeng Formation provides a geologic Fractures in Fengfeng FormationThree types of fractures were observed in the Fengfeng structural fractured fractures (Figure 4). The upper part as obvious as the top part. The main form of fractures is Filling materialsPhysical and chemical weathering over 150 Ma has produced a large amount of breccia cemented with shown in Figure 5, the high extent of cementation and 281 Figure 2 Figure 2. Typical geological column in North China coal fields. Figure 3. Palaeokarst profile in Fengfeng Formation.

PAGE 299

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE1 m than 1 m in the other eight boreholes in the middle segment of the Formation suggests that the fractures in the Fengfeng Pumping test resultsTable 1 lists the thirteen boreholes in which pumping tests were conducted in both the Fengfeng Formation and the Those values are two to three orders of magnitude smaller the water level of the Fengfeng Formation is apart from the water abundance of the Fengfeng Formation is much Hydrogeochemical test resultsFigure 6 shows the Piper diagrams based on the water samples collected in these thirteen boreholes at which the pumping tests were conducted. The main anion resisting capacity. The characteristics of geologic barriers in preventing water inrushes are discussed in Zhou and Li (2001).Water-resisting capacity distribution of the aquifuge in the Fengfeng Formation. The hydrogeological characteristics of the Fengfeng Formation were then compared with those of the is often considered as a strong aquifer.Results of water-pressure tests boreholes. The technique of conducting the pressure tests is detailed in Zhou and Li (2001). The test results an extremely poor permeability ranging from 0.00002 to in the Ordovician limestone gradually decreases in the underlying stratum. The thickness of this part is 8.6 m and 282Figure 5. Cemented fractures in Fengfeng Formation. Figure 4. Three different types of fractures in Fengfeng Formation. Hydrogeogical Properties of Fengfeng FormationDrilling fluid lossOf seventeen Ordovician hydrogeological boreholes drilled in the Fengfeng Formation in Sihe Mine, fourteen

PAGE 300

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 the top zone.Thickness of aquifuge in Fengfeng FormationIn North China, the water abundance of an aquifer capacity: in the Fengfeng Formation are Ca2+ and Mg2+. The composition of mixed samples of the Fengfeng and These observations indicate that the Fengfeng Formation Formation. The water in the Fengfeng Formation has poor Flow rate logging results because of its poor water yield. In comparison, obvious 283 Table 1. Result of Ordovician limestone pumping tests Borehole Comparison of water level Comparison of Specific Capacity Fengfeng Formation Formation Difference Fengfeng Formation and Fengfeng Formations C ombined Ratio of Combined and Fengfeng Formation SWY2 555.8 521.2 0.000942 0.000859 0.9 SC2 449.7 0.001 0.188 188 450.9 481.5 0.00588 6.2 SWY1 482.4 10.6 0.0026 0.144 55.4 SB1001 511.4 488.79 22.61 0.000477 292.4 SB1002 546.54 508.94 0.00299 544.92 0.004489 0.06624 14.8 SB1004 495.46 478.75 16.71 0.004721 0.044552 9.4 SB1005 466.84 477.75 10.91 0.0004 0.00052 SW2 497.7 15.4 1.1 488.7 487.6 1.1 0.00478 0.594 489.2 487.1 2.1 0.0268 6.2 SW1 512 488.4 74.9 Figure 6. Piper diagrams for groundwater in Fengfeng Formation (a) and Majiagou Formation (b).

PAGE 301

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE in Ordovician karst. Chinese Journal of Rock Tang Junhua. 2009. Research on simultaneous limestone aquifuge in top Ordovician and its application in roadway excavated through graben faults. Journal of Ji Qing, Ding Zhuo. 2010. Impermeability analysis of hydrologic characteristics of mine spoil aquifer. on Ordovician Limestone Karst Study in area of characteristics and distribution rule of carbonate strata in the top of Ordovician in North China. Research on Porosity of Ordovician Fengfeng aquifuge. The thickness of relative aquifuge ranges from Conclusions limestone for mining above an aquifer 1. Problems of mining above an aquifer are increasingly serious in North China. In Sihe Mine for example, after considering the relative aquifuge, more than 90 million tons of coal resources are considered to be not threatened by the underlying pressurized water. Careful analysis Fengfeng Formation helps understand the real conditions when mining above the aquifer. The aquifer system but consist of protective barriers. Recognition of the heterogeneous nature provides us with a new approach for evaluating water inrushes when mining above an aquifer. 2. of the paleokarst crust in top part of the Middle Ordovician limestone, the paleokarst crust can be in Sihe Mine. When doing a study on the hydrogeological features of the Ordovician limestone, a multidisciplinary approach, such as water abundance, hydrochemistry characteristics, permeability of strata, and strata combination, is 4. be variable in different areas. In some areas may increase the coal production or lengthen the operation of a mine.References strata impermeability in Ordovician top and risk 284

PAGE 302

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2DEEP TIME ORIGINS OF SINKHOLE COLLAPSE F AILURES IN SEWAGE LAGOONS IN SOUTHEAST MINNESOTAE. Calvin Alexander, Jr.Earth Sciences, University of Minnesota, 310 Pillsbury Dr. SE, Minneapolis, MN 55455, USA, Anthony C. Runkel, Robert G. TippingMinnesota Geological Survey, 2642 University Ave., St. Paul, MN 55114, USA,, tippi001@umn.eduJeffrey A. GreenMnDNR Ecological & Water Resources, 3555 9th St. NW, Suite 350, Rochester, MN 55901, USA, soil. Over eighty years of soil conservation efforts have those stored soils out of their spring outlets. Finally, the is reactivating conduit segments that have been clogged and inactive for millions of years. The karst solution voids into which the lagoons collapsed have formed over 480 million years. The recognition and accurate karst hazard maps to be constructed and used in sustainable resource management decisions.Introduction wastewater treatment facility (WWTF) lagoons constructed in the 1970s and 1980s in southeastern through sinkhole collapse (Alexander and Book, 1984; fail, the Altura WWTF lagoon, collapsed twice. All four collapses occurred in the same stratigraphic position. That stratigraphic interval is centered on the unconformable contact between the Shakopee and Oneota Formations of transmissivity zone in the Paleozoic hydrostratigraphy et al., 2006). The interval has long been known as a productive zone by the water well drilling community.Abstract wastewater treatment lagoons constructed in the 1970s have failed through sinkhole collapse. Those collapses occurred between 1974 and 1992. All three failures occurred at almost exactly the same stratigraphic position. contact between the Shakopee and Oneota Formations of hydrostratigraphy of southeastern Minnesota. carbonates about 480 million years ago. The existence of syndepositional interstratal karst unconformities between the Oneota and Shakopee Formations and between the Shakopee and St. Peter Formations, were recognized in the 1800s. About 270 million years ago galena, sphalerite has been above sea level since the Cretaceous and huge Pleistocene. The incision of the Mississippi River and its tributaries has and is profoundly rearranging the ground during glacial cycles. The Pleistocene glacial cycles erosion rates from row crop agriculture between the 285

PAGE 303

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE Mossler (1987, 2008). This column nicely demonstrates that siliciclastic sandstones, mudstones and shales dominate in the upper Cambrian column of SE Minnesota, that carbonates dominate in the Ordovician section and that the two fundamentally different rock A conventional geologic column is one of the most basic conceptual tools used by hydrogeologists and geoengineers to understand and describe the region. geologic concepts to the public. From a karst perspective, however, they have an important failing. They and importance of unconformities formed by erosion. It was during these erosion periods that karst features could and did develop. The unconformities have been emphasized with bold black lines in Figure 2. column displaying the same information on a linear time scale. This geologic column emphasizes the relatively rocks. The Upper Cambrian Mt. Simon through the Jordan, i.e. roughly 250 m (820 ft) of siliciclastic rocks, were deposited in about 10 million years. The remaining construction and operation. Were these collapses random, from the interactions between deep time geology with recent human activities? Can we improve the safety of lagoons? These questions are the focus of this paper.Hydrogeologic SettingMossler (2008) and references listed therein are the basis for much of this section. To construct the current and Middle Paleozoic sedimentary rocks, Mossler sedimentological framework. karst aquifers occur in southeastern (SE) Minnesota the area roughly south and east of the Twin Cities Metropolitan area. SE Minnesota forms the east limb of The regional dip is about two meters per kilometer to the southwest but local structures with several meters of amplitude are hydrogeologically important. Figure 1 shows the structural context of the Middle and Lower Paleozoic rocks in the upper Mississippi River Wisconsin, northwestern Illinois and northeastern Iowa, with southeastern Minnesota, comprise the UMV Karst. The Paleozoic sedimentary units are relatively thin, regionally extensive, siliciclastic and carbonate rocks that can be correlated across the UMV Karst. These rocks were deposited during three episodes of complex sea transgressions during the Paleozoic. The tops of each episode are characterized by interregional unconformities. Because SE Minnesota was near the center of the sea withdrew. The tops and bottoms of many of the units are marked by erosional unconformities. Conventional geologic columns show the sequence of rocks exposed as the rocks appear in outcrop or drill core. The columns diagram the rock types and linear thicknesses. The geologic ages of the rocks are indicated 286 Figure 1. Regional geologic setting of the Lower and Middle Paleozoic rocks in SE Minnesota (Mossler, 2008, Figure 2B).

PAGE 304

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2287 Figure 2. Lithostratigraphic column for the Lower and Middle Paleozoic rocks in southeast Minnesota. The bold black, wavy lines emphasize the unconformities. Modified from (Mossler 1987, 2008).

PAGE 305

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE288Figure 3. Chronostratigraphic column for Lower to Middle Paleozoic rocks in southeast Minnesota. Modified from (Mossler, 2008, Figure 1).

PAGE 306

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 is a very old feature. It started as solution weathering to 5 million year unconformity between the Oneota and the overlying Shakopee. The geometry of the overlying Peter large magnitude unconformity, insured that the stratigraphic interval would localize what is a regional Smith (1997) documented that solution removal of anhydrite and the subsequent brecciation of the eroding top of the Oneota played an important role in the subaerial weathering before the deposition of the Shakopee. Smith et al. (1997) further emphasize that authigenic silica fabrics present in this unconformity syndepositional. About 270 million years ago galena, sphalerite and This includes precipitation of these mineral into macropores associated with the paleokarsts feature in the Karst processes from the Mesozoic to present The UMV region has been above sea level since the Mesozoic and huge volumes of fresh water have The climate was warm to tropical during much of this period and karst process must have been an important incision of the Mississippi River and its tributaries has systems as it changes the regional base level during glacial cycles. 50 to 150 meters of glacial sediments with the internal du Chien carbonates were deposited in two roughly 5 million year episodes with a 5 million year gap between. An unconformity representing a gap of over 15 million years separates the Prairie du Chien from the Upper Ordovician St. Peter through Maquoketa sequence, which was deposited in about 10 million years. The Devonian Spillville through Lithograph City carbonates represent about 15 million years of time, including million years of the Cambrian through Devonian time left no depositional record in SE Minnesota. During those enormous unconformity time gaps the lower Paleozoic 289Karst Processes in SE Minnesota and the mid Prairie du Chien karst high-transmissivity zonePaleozoic karst processes features in the UMV region gathered references to paleokarsts of Ordovician, Devonian, Pennsylvanian between the Shakopee dolomite and the overlying St. Peter sandstone, between the Oneota dolomite and the overlying New Richmond sandstone and between the Oneota dolomite and the underlying Blue Earth in the UMV date back at least to Barris (1880) and Farnsworth (1888). Farnsworth (1888) refers to an Whitney, 1858). Karstification of the top of the Oneota Formation during the 4 to 5 million year subaerial erosional unconformity before the deposition of the Shakopee Formation and subsequent regional groundwater flow systems, extending to the present day, have produced arguably the most transmissivity zone in the Paleozoic hydrogeology of SE Minnesota. documented the hydrogeologic importance of this feature. Figure 4 diagrams the position and extent of this feature in southeastern Minnesota.

PAGE 307

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEcoincides with the water table position (Tipping et al., 2001). Finally, that the four sinkhole collapses in WWTF lagoons were constructed where this zone is the uppermost bedrock (Alexander and Book, 1984; Jannik the geoengineering and environmental management importance of this zone in SE Minnesota. ConclusionThe collapse of the Altura, Lewiston and Bellchester, Minnesota, WWTF lagoons were not random, unpredictable events. The collapses occurred as a result most developed regional karst solution zone in SE Minnesota.The opportunity and the challengeWhen these WWTF lagoons were designed and built in the 1970s and 1980s, the current understanding of the hydrogeologic and environmental importance of the did not exist. The basic geologic information existed but had not been assimilated into the hydrogeologic or geoengineering communities who designed and built the many of the conduit systems with soil. Soil conservation reduced the amount of soil moving into the conduits. out of the conduits. events over the past few decades may be reactivating portions of conduit segments that have been clogged and inactive for millions of years. There may be many currently inactive conduits (plugged by ancient, glacial and modern sediments) that are being reactivated as hydrologic The contact zone between the Prairie du Chien and the Shakopee has long been known as a productive zone by the local water well drillers. This cavernous zone is easily recognized in hydrophysical logging of area water wells. documented the dominance of this stratigraphic interval on the distribution of sinkholes in Winona County, where three of the four collapses occurred. In Wabasha County, which contains the fourth collapse site, sinkholes are more likely to occur where this stratigraphic interval 290Figure 4. Southeast Minnesota hydrostatigraphic cross section, schematically showing solutionally-enhanced fractures and voids (not to scale). The karst high-transmissivity zone near the middle of the Lower Ordovician Prairie du Chien Group is highlighted in red.

PAGE 308

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 Vicinity. Proceedings of the Davenport Academy Dalgleish JD, Alexander EC Jr. 1984. Sinkhole distribution in Winona County, Minnesota. Engineering and Environmental Impact, Proceedings of the First Multidisciplinary on Carbonaceous Material in the Niagara Limestone at the State of Iowa: Embracing the Results of Investigations Made During Portions of the of the Upper Mississippi Valley Region. Studies in Jannik NO, Alexander EC Jr, Landherr LJ. 1992. The sinkhole collapse of the Lewiston, MN waste water treatment facility lagoon. In: Quinlan J, Stanley A, editors. Proceedings of the Nomenclature for Minnesota. Minnesota Nomenclature for Minnesota. Minnesota with evaporite and carbonate dissolution in the Lower Ordovician Oneota Formation (Prairie lagoons. That technical information had also not been that a problem existed, much less how to prevent it. We now have the tools and knowledge to incorporate more focused, nuanced karst hydrogeology and geoengineering concepts and practices into evaluating and practices. We can upgrade the design and operation of new infrastructure and practices. The use of karst hydrogeologic and geoengineering concepts to collect, interpret and apply new tools to gather much larger quantities of higher quality, more detailed data and information to infrastructure siting and management issues is very promising. One example is DEMs and surface feature maps, growing data bases of local subsurface information from well drillers and other sources, and a variety of geophysical tools. Such more traditional bedrock, structural and karst mapping techniques. We have the ability to create accurate, the existing infrastructure to be prioritized according to relative sinkhole failure probability, and to guide siting decisions for future infrastructure. These maps can also be Acknowledgments Environment and Natural Resources Trust Fund as on Minnesota Resources (LCCMR).ReferencesAlexander EC Jr, Book PR. 1984. Altura Minnesota lagoon Engineering and Environmental Impact, Proceedings lagoon collapses. In: Beck BF, editor. Applied Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst; 291

PAGE 309

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE Atlas of Wabasha County, Minnesota: County Atlas St. Paul (MN): plate 5. intergranular and secondary porosity in part of the Cambrian Sandstone aquifer system of the cratonic interior of North America: Improving predictability 292

PAGE 310

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2property damage is estimated to be at least 20 million the geologic background of the study, mechanism of sinkhole collapses, and prevention and treatment of sinkhole hazards in the study area. Geologic Setting Maohe village is located in a highly active karst area and many karst features such as sinkholes, springs, karst windows, caves, and conduit systems are widely distributed in this area. The area is a typical karst plain with isolated karst towers. The altitude of the ground sediments above bedrock are Quaternary alluvium and colluvium deposits with thicknesses ranging between comprised of colluvium clay, silty clay, rounded gravels, belonging to the middle Carboniferous Dapu formation (C2d). The dolostone is brittle with many fractures and dipping angles less than 5 degree. The altitude of the top The study area is a covered karst area. Bedrock is highly fractured due to faulting and tectonic movements. which causes more relief of bedrock topography. Voids and caves are commonly encountered during drilling processes. Thirty three out of 71 (46.5 %) drilling cores showed that caves exist within bedrock. Sixty six caves are detected from drilling with heights ranging from 0.20 to 14.60 m. The number of caves decreases drastically with depth and almost all caves are distributed within with clay and dolostone debris, especially caves close AbstractA series of sinkholes collapsed at Maohe village, areas and 68 fractures in May of 2012. Many ground failures and fractures formed and an area of 40000 m2 was impacted by the collapsing event. The collapsing event was caused by large scale soil piping and soil void collapses. Preliminary investigations revealed bedrock roof collapse of underground streams may trigger a water hammer effect in the karst conduits. The water hammer effect caused severe soil damage and subsequent collapses in Maohe Village. Soil disturbance may cause a change in hydraulic gradient, resulted in sinkhole collapses. By monitoring pressure changes of karst water, turbidity of groundwater, locating soil voids and soil disturbances using ground sinkhole collapses.IntroductionOn May 10 2012, a series of sinkholes collapsed at diameter and 2 to 10 m in depth within one hour after 10:20 am. By May 15, 41 sinkholes collapsed. Eleven the area. Many ground failures occurred and an area of 40,000 m2 was impacted by the collapsing event and an area of 10,000 m2 suffered severe damage. Because the collapsing events occurred in areas with high population density, an initial investigation revealed severe damage Two elementary schools and one middle school were EMERGENCY INVESTIGATION OF EXTREMELY LARGE Mingtang Lei, Xiaozhen Jiang, Zhende GuanInstitute of Karst Geology, CAGS, Guilin, China, Yongli GaoDepartment of Geological Sciences, Center for Water Research, University of Texas at San Antonio, TX 78249, USA, yongli.gao@utsa.edu293

PAGE 311

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE surface when this sinkhole collapsed and all of the water disappeared through the bottom of the sinkhole after a few minutes. Subsidence area no. 2 is the largest with an deposits and karst water within bedrock matrix, fracture, and conduits. A limited amount of Quaternary water is mainly stored within a silty clay aquifer with a depth conduits of the dolostone bedrock. The depth of the water table ranges between 1.60 and 5.95 m and the altitude of the water level is 88.25 90.02 m. 294Figure 2. Subsurface cave distribution based on drilling records.Vertical axis indicates elevation (m) above sea level.Figure 1. Geologic Map of the study area. Water Pumping and Sinkhole DistributionThere is only one deep well that pumps in this area. The well was drilled in 1990 and is located in the northeastern corner of the study area. The well depth is 95 m and the average pumping rate is approximately 150 m for laundry. Most domestic water usage is from city water systems. No heavy pumping of groundwater ever occurred and a cone of depression does not exist in the study area. sinkholes are circular or elliptical in shape on the surface and cone shape below surface. The largest sinkhole is no. 11 with an irregular shape on the surface and a long axis

PAGE 312

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2295 Figure 3. Sinkhole distribution in the study area. Figure 4. Typical sinkhole collapses in the study area.

PAGE 313

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEThe karst water level is recovering to normal conditions after the collapsing event. It is unlikely for large scale collapsing events to occur in the near future because sinkholes and subsidence may continue due to the disturbance of sediments. Large scale sinkhole collapses may occur again during monsoon seasons. It may take several years to stabilize the soil to normal hydraulic conditions.ConclusionsRapid water level rise after a storm event caused a series of collapsing events in Maohe village. Several collapsing an extreme weather event and sinkhole collapses need to be conducted to prevent such large scale collapsing events in the future. dolostone was traditionally thought to be relatively weak and the study area has long been listed as a low risk sinkhole area. This study demonstrates that severe damage can be caused by large scale sinkhole collapses in a dolostone karst area. By monitoring pressure changes of karst water and turbidity of groundwater and locating soil voids and soil is possible to predict future sinkhole collapses (Jiang et 2010). These approaches are being used to prevent future sinkhole collapses and to reduce the damage caused by the collapsing event. center of the subsidence. Four fractures formed around this subsidence area, which caused building no. 4 to be tilted to the north. Most sinkholes formed at the Shangmuzhao section of Maohe village and the overall distribution is along zones dug well and many air bubbles were released from the well on May 10. Total suspend solids were about 10% from pumped groundwater. By May 14, the well had already collapsed (Figure 5). Results and DiscussionCollapsing events were caused by soil piping and deformation due to highly fluctuating hydraulic conditions within the karst water system. Karst features such as cave streams, large karst springs, karst windows, caves, and blue holes exist in the study area. occurred during heavy rainfall. Rapid water level rise caused cave roof collapse, which may trigger a water hammer effect in the karst conduit system. The water hammer effect can release a pressure surge to the karst conduit system, causing severe soil damage and subsequent collapses (Lei et al., 2010; the hydraulic gradient, which can cause water level fluctuations and eventually result in new sinkholes. Red clay near the bedrock could be disturbed with seepage deformations which may trigger sinkhole collapses.296Figure 5. Hand-dug well with high turbidity and air bubbles (L) collapsed on May 14, 2012 (R).

PAGE 314

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2References of Large Scale Sinkhole Collapses, Laibin, soil void formation along highway subgrade using Yuhr L, Alexander E, Beck B, editors. Proceedings of the Eleventh Multidisciplinary Conference on Sinkholes and the Engineering and Environmental and treatment of sinkholes and subsurface voids In: Yuhr L, Alexander E, Beck B, editors. Proceedings of the Eleventh Multidisciplinary Conference on Sinkholes and the Engineering and Environmental monitoring of karst water pressure in Zhemu, 297

PAGE 315


PAGE 316

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2The area is underlain by the limestone of the Saraburi chain of hills, ridges, and occasionally as mounds which create classic 'tower karst' scenery. The rainforests, excessive rainfall and widely variable forming environment to develop, with streams draining into the limestones from mountain catchments. In this area, the mature karst is locally fringed by tall cliffs that overlook valleys and closed basins. The area underlain by limestone is extensive and rainfall is AbstractAgricultural development in the Saraburi Province of Central Thailand has increased the demand investigations have been undertaken by Department of Saraburi group. The area is located 120 km north of Bangkok between the cities of Saraburi and Pak Chong in the south, Lopburi to the west, Chai Badan and Nong Pong to the north, and Nakhon Ratchasima to the east. It covers the following districts: Amphoe Pack Chong, Nakhon Ratchasima Province; Amphoe Muang Muak Lek, Khangnoi, Phaputabat, Wong Muang and Chalormphrakiat in Saraburi Province; and Amphoe Moung, Lamsonthi, Phatananikom, Thaluang and Chaibadan in Lopburi Province (Figure 1). The topography is characterized by mountain ranges, karstic plateaus, and rolling hills of low to medium relief, with low lands in between. The mountainous ridge elevation reaches over 800 m above sea level 500 m ASL, and the low lands are at about 100 m ASL. In the karstic plateaus and mountains areas, springs, caves, and dry stream beds exist. In dry periods, some streams in low land areas are dry, but the large rivers Tropical climate (Monsoon type) with two distinct seasons is characteristic of this area. The dry season begins in October and ends in May, followed by a monsoon season between June and September. Annual rainfall ranges between 1,500 and 2,000 mm and temperature ranges between 20.00C and 40.70C.KARST LANDFORMS IN THE SARABURI GROUP Gheorghe Ponta, Bashir Memon, James LaMoreauxPELA GeoEnvironmental, 1009 23rd Avenue, Tuscaloosa, Alabama, 35401, USA,,, jlamoreaux@pela.comJade Julawong, Somchai WongsawatWater Resource Engineering Co., Ltd., 52 Soi Ladphrao-wanghin 26, Ladphrao-wanghin Road, Ladphrao, Bangkok 10230, Thailand,, 299 Figure 1. Saraburi Province of Thailand (shown in red).

PAGE 317

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEGeology (Figure 2). The carbonate rocks are exposed as a chain of hills, ridges and occasionally as mounds which create classic 'tower karst' scenery. In stratigraphic order from oldest to youngest, the rock units are the: Phu Phe, Khao Khwang, Nong Pong, Pang Asok, Khao Khad and Sap Bon formations. The carbonate limestone, nodular and tabular chert bands, partly intercalated with slaty shale. Fusulinids and crinoids are present. The Phu Phe formation is located in the central approximately 68 km2 as a ridge with vertical cliffs. The Phu Phe formation is thrusted over Sap Bon and Upper Permian Age. The Phu Phe formation outcrops 25 km East of Saraburi. It is divided in two sections the cities of Sap Bon and Khao Phu Phe. The Siam Cement Plant quarry is located in this formation The formation is widely distributed in the northern part of the Saraburi area. The formation was deposited 2011). From the border between Nakhon Sawan and Lop Buri provinces (the Tak FaBan Mi area) to southern Phetchabun Province (the Nong Phai area), Nakornsri (1976, 1981) established the Tak Fa Formation which is The outcrop area is approximately 650 km2. It is a karstic plateau with rolling hills of low to medium relief, which continue north with a narrow strip about 25 km long and 4 to 1 km wide (~100 km2 Tak Fa Formation) ridge strip belongs to the Tak Fa Formation (Amphoe Ban Mi geological map 1:250,000). m vertically. Exokarst landforms are well represented. sinkholes, sinking streams, and closed depressions were A dye study performed in October 2012 indicated the hydraulic connection between a sinking stream and Tham the dye study, the protection area for the Well 114 and the spring also includes the closed depressions.IntroductionDevelopment of land in the Saraburi Province of Central Thailand has increased the demand for groundwater The area is located 120 km North of Bangkok between the cities of Saraburi and Pat Chong in the south, Lopburi to the west, Chai Badan and Nong Pong to the north, and Nakhon Ratchasima to the east. It covers the following districts: Amphoe Pack Chong, Nakhon Ratchasima Province; Amphoe Muang Muak Lek, Khangnoi, Phaputabat, Wong Muang and Chalormphrakiat in Saraburi Province; and Amphoe Moung, Lamsonthi, Phatananikom, Thaluang and Chaibadan in Lopburi Province. The topography is characterized by mountain ranges, karstic plateaus, and rolling hills of low to medium relief, with low lands in between. The mountainous ridge elevation can reach over 800 m above sea level 500 m ASL, and the low lands are at about 100 m ASL. In the karstic plateaus and mountains areas, springs, caves, and dry streams exists. In dry periods, some streams in low land areas are dry, but the large rivers Tropical climate (monsoon type) with two distinct dry season begins in October and ends in May, followed by a rainy season between June and September. Annual rainfall ranges between 1,500 and 2,000 mm and temperature ranges between 20.00C and 40.70C.300

PAGE 318

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2301Figure 2. Geological Map of the Project Area (from Ridd et al., 2011). Figure 3. Geological map with location of karst features.

PAGE 319

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEHydrogeologyThe availability of ground water in karst aquifer system due to complex geology. This aquifer system has an fractures serve as conduits, leading water from the top of aquifer system of the Saraburi domestic, industrial, and agriculture uses. At the interface of noncalcareous formations with karstic formations, streams are sinking underground to discharge to spring(s) and cave(s). The availability of water varies during two distinct seasons. During the dry season (October to May) the flow of the In the mountains area all the sinking streams were dry. During our field visits in December 2011, 11 springs, one perennial sinking stream, and one perennial cave stream were identified. A network of manmade ponds and lakes has been developed to collect and store surface runoff and the rain water from the open ponds and lakes is lost due to the high rate of evaporation. Air temperatures are above 0C for most of the time. During the rainy season (June to September) flooding occurs periodically, most of the annual rainfall of 1,500 to 2,000 mm has been recorded in this time frame. Karst LandformsThe karst landforms had been developed by widely variable climate conditions during geologic times. The karstification process is ongoing in the area because of the excessive rainfall and rainforest is occasionally fringed by tall cliffs that overlook valleys and closed basins. The area underlain by limestone is extensive. Rainfall is abundant (field observations). Exokarst landforms are well closed depressions are present. The Nong Pong Formation is dominated by intercalations of limestone with locally developed chert. It occurs widely in the central part of the Saraburi area and it is approximately 670 m thick (Ridd et al., 2011). The Nong Pong formation is located in the central part of 700 km2. It is a karstic plateau with rolling hills of low to medium relief. crops out south of the Nong Pong Formation. It is local sandstone lenses (Ridd et al., 2011). This locally interbedded argillite, dolomitic shale, siltstone, sandstone and conglomerate. It is distributed extensively in an almost WNWESE direction in the Saraburi area. (Ridd et al., 2011). and 15 km wide. The area of outcrop is approximately 1,875 km2. It is a karstic plateau with rolling hills of high relief and some cliffs and ridges. This grey limestone with chert nodules in the upper part. It crops out along the southern margin of the Saraburi metamorphism and altered into slate, phyllite and schist (Ridd et al., 2011). sedimentary domain. After the latest Carboniferous, the area. These units are mostly folded, fractured, and faulted across the study area.302

PAGE 320

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2Karrenfields extensive ones being located in the north of the study area (Figure 4). The second one is two kilometer east of Mu Si Spring, next to a temporary sinking stream (Figure 5).TsingiTsingi is characterized by vertical rock blades fretted sharp by dissolution (Figure 6). This feature was found on the limestone hills surrounding the karst margin depression, Closed Depressions in the study area: Rain pits Sap River (Figure 7).303Figure 7. Rain Pits in the Vicinity of the Waterfall in Khao Khwang LimestoneHuai Nam Sap River.Figure 6. Tsingi in the Karst Margin Depression (Recharge area of Tham Lumphini Spring).Figure 5. Karren Field Two km East of Mu Si Spring (Ban Khao Chan Hom). Figure 4. Karrenfield with Pinnacle (Wat Tham Phrommaloc). Kamenitza Kamenitza in the vicinity of Quarry Spring Khao Khad limestone (Figure 8). Sinkholes (Dolines) 0 sinkhole is about 10 m in diameter (Figure 9). Closed depressions Three large closed depressions and one karst margin of Ban Khao Takhaeng village, about 15 km south of the area of geothermal anomaly (Figure 10). It is over 2 km long and 1 km wide. A swallet is located at the north end of the depression. During the rainy season, the disappearing stream recharges a temporary karst spring at the base of limestone massif, generating a karst system

PAGE 321

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCESprings the geological map, the spring is located in Pleistocene deposit that lie on the top of the Khao Khvang limestone of Lower Permian Age. The Dug Pond and the Tham Lumphini Suan springs Both springs are used locally for public water supply. A karst margin depression exists in the recharge area discussed in Dye Study section. A third location with closed depressions is situated between the villages of Ban Khao Phra and Ban Khao 304Figure 12. Mu Si Spring.Figure 11. Hot Well (Bat Nong Nun Wat Nam Sut).Figure 10. Closed Depression in the vicinity of Ban Khao Takhaeng Village.Figure 9. Sinkhole with Thermal Water in Vicinity of Wat Tham Phrommaloc Temple.Figure 8. Kamenitza in the vicinity of Quarry Spring Khao Khad limestone.

PAGE 322

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 in the water samples collected during the dye study, as a (powdered Uranine 40% concentration mixed with a total Temperature (24.80 (225.7 us), and salinity (0.1ppt) were determined at the During the dye study, charcoal bags were installed and water samples were collected at the spring, 687 m away from the sinking stream. Also a charcoal bag The Uranine was detected in the elutant from all 17 charcoal bags installed and recovered during the dye and monitoring points (spring and Well 114) are shown are shown on Figure 14. The peak dye concentration of elutant from charcoal 989 Fluorescence Intensity Units (IU), and was detected in the charcoal bag collected 2 days (47 hours) after Two water samples were collected during the dye study. Dye was detected in the water sample collected on which is the peak dye concentration. A second sample was collected on October 18, 2012, the dye was detected These results suggest that the groundwater is rapidly traveling between the sinking stream and spring through movement is to the northwest. Dye StudyTo characterize groundwater resources in the Saraburi province, a dye study was performed in October 2012. and 1 km wide) is located in the recharge area of the of the depression is bounded by igneous rock and the north side by a limestone ridge. This limestone ridge is the boundary between the karst margin depression and the closed depression developed along a parallel stream, both being part of the Tham Lumphini Suan the rainy season, based on the size of the stream bed by the sinking stream is linked to the Tham Lumphini The primary focus of the dye study was to illustrate the potential hydraulic connectivity of the sinking stream to detergents, bathroom cleaners, pigments for inks and A passive dye detector (charcoal bag) was placed at charcoal bag for the background portion of the dye was installed on October 6, 2012. The detector was removed for evaluation for background readings one day prior to temple (Well 114) and the right side tributary located monitored during the dye study. No dye was detected in samples collected at the Tham Lumphini Suan spring respectively. The average of those concentrations 0.18 305

PAGE 323

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE306Figure 13. Topographic map showing the dye injection location and Tham Lumphini Suan Hin Spring.

PAGE 324

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2the presence of water, as well as the salinity of the water. In general, the electrical resistivity of carbonate rock is The electrical resistivity of soil is on the order of ranges are general estimates, but illustrate the relative difference in electrical resistivity of earth materials. generally situated along roadsides and were located within various discharge and recharge zones throughout the Province. Data were acquired using electrode spacing of 20 meters (m) was used for a total Based on this dye study, the recharge area of The Tham Lumphini Suan Spring and Well 114 is extended to 4.5 square kilometers towards the dye input location, which also include the closed depression, west to the spring. Geophysical (Resistivity) SurveyTo identify favorable locations for groundwater system and characterize groundwater resources, an extensive resistivity geophysical survey was performed. rock under saturated conditions (i.e. all of the pores, conditions. It can also exist as underground rivers and lakes in karst environments. Since electricity can move more easily through water than soil or rock, the bulk electrical resistivity of the earth is highly dependent on 307Figure 15. Interpreted DC resistivity profile.Figure 14. Breakthrough of Uranine at Tham Lumphini Suan Spring and Well 114.

PAGE 325

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEaccumulations of groundwater in karst features features exist in the profiles indicating that there are extensive groundwater reserves in the area. Based on the interpretation of the various datasets, the potential for recoverable groundwater was mapped on the geologic map. These zones are marked production potential exist along the edge of the Khorat Plateau. With respect to elevation there is not a strong regional correlation between elevation and groundwater potential, which suggests that the are hydraulically discontinuous, because of geologic structures. This is typical and to be expected in karst terrain due to lack of hydraulic communication between the various subsurface water bearing zones in the region. used so that lateral variations in electrical resistivity indicative of karst features could be resolved. that may be indicative of groundwater were interpreted as potential locations where large amounts of groundwater accumulations, possibly associated with subsurface karst features could be exploited. were found along various inverted resistivity profiles. seven profiles. One profile (Figure 15) is included herein for illustration purpose. Figure 16 is geologic map showing areas of groundwater potential, with 308Figure 16. Geologic map of groundwater production potential.

PAGE 326

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 London. ConclusionsThe area is underlain by the limestone of the Ratburi as a chain of hills, ridges and occasionally as mounds which create classic 'tower karst' scenery. Exokarst landforms are well represented. Various types during site investigation. A dye study indicated that there is hydraulic connection between a sinking stream and Tham Lumphini Suan the protection area for the Well 114 and the spring also includes the closed depression. Based on the interpretation of the various geophysical datasets, the potential for recoverable groundwater was mapped on geologic map (Figure 16). These zones are potential exist along the edge of the Khorat Plateau. With respect to elevation there is not a strong regional correlation between elevation and groundwater potential, throughout the region are hydraulically discontinuous, because of geologic structure. This is typical and to be expected in karst terrain due to lack of hydraulic communication between the various subsurface water bearing zones in the region.References 1:250,000. Dunkley JR. 1997. The Caves of Thailand Addendum Council. Ellis M. Caves and Caving in Thailand [Internet]. 2012. Available from:. Department of Mineral Resources; 1977. Sheet ND 309

PAGE 327


PAGE 328

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2CLASTIC SINKHOLE AND PSEUDOKARST DEVELOPMENT IN EAST TEXASKevin W. Stafford, Melinda G. Shaw-Faulkner, Wesley A. BrownDepartment of Geology, Stephen F. Austin State University, P.O. Box 13011, SFA Station, Nacogdoches, Texas, 75962-3011, staffordk@sfasu.eduIntroductionTraditionally, East Texas has never been considered a dominant karst region because of the lack of soluble dominated by Cenozoic clastic sediments associated and evaporite strata in the region are deeply buried units that are heavily exploited for hydrocarbon resources (Nichols et al., 1968). These deeper strata inevitably host hypogene karst associated with hydrocarbon maturation strata are limited to pseudokarst development, where paths creates void space. Pseudokarst includes any geomorphic features that exhibit morphologies similar to true karst features but have not been formed from solutional processes (Palmer, 2007). Pseudokarst traditionally includes caves, sinkholes and springs but generally lacks karren development. In East Texas, these traditional characteristics have been documented as widespread occurrences, but never in small, ephemeral suffusion features commonly occur recharge of shallow groundwater systems. Springs systems discharge along low permeability horizons discharge vertically to the land surface along preferential While pseudokarst development is relatively rare in East Texas compared to true karst development in other regions of the state (Elliot and Veni, 1994), these features do provide unique ecological and culture resources for reports that thirty seven pseudokarst caves and karst features exist in sixty one counties that cover the greater AbstractPseudokarst development in East Texas is controlled primarily by a combination of suffosion and preferential but occasionally persistent features develop in more indurated facies. Pseudokarst occurs in Claiborne (Eocene) strata in Angelina, Cherokee, Nacogdoches, Panola, Rusk, San Augustine and Shelby counties. clastics with variable cementation and associated fractures and bedding planes create local voids through suffosion that stope upward to create sinkholes and incised collapse valleys often associated with persistent and ephemeral springs. systems and extensive anthropogenic overprinting regionally, which create numerous constructional closed depressions. Sinkhole densities coupled with slope analyses indicate clustered regions of pseudokarst development within Carrizo, Queen City and Sparta sandstones. Known pseudokarst caves within the region include features developed along low permeability boundaries where discharge member product of natural suffosion processes in East Texas with more than 160 meters of surveyed passage and a collapse sinkhole covering approximately a hectare. Smaller suffosion sinkholes occur along steep gradients but generally remain associated with features or enlarged regions associated with spring discharge. Anthropogenic pseudokarst sinkholes are generally associated with leaky pipelines and focused groundwater recharge from impermeable surfaces and produce local geohazards. Traditionally East Texas is not known for extensive pseudokarst development; however, isolated caves and sinkholes can be locally 311

PAGE 329

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE2010), which consist of the Carrizo, Reklaw, Queen City, Weches, Sparta, Cook Mountain and Yegua formations, in ascending order (Shelby et al., 1968). Carrizo Sand ferruginous. The Reklaw Formation is composed of lignitic. The Weches Formation is primarily glauconitic sand with clay interbeds that is often lenticular with local grained massive sandstone that is locally carbonaceous and commonly contains interbeds of silty or sandy clay. The Cook Mountain Formation is primarily clay or marly sand, but locally grades into sheet clays and glauconitic sands. The Yegua Formation is dominated by clay with minor sandstone beds and local concretionary limestone (Shelby et al., 1968). on seven counties (Angelina, Cherokee, Nacogdoches, Panola, Rusk, San Augustine and Shelby) (Figure 1), where pseudokarst cave development is primarily limited to coarser grained, sandstone facies of the Claiborne and Sparta formations, while springs and seeps can be found throughout all facies of the Wilcox, Claiborne and Jackson groups where permeability horizons intersect the land surface. lentils and deltaic silts, with sand abundance increasing series of marine and continental sediment deposits as the Eocene strandline migrated in response to sea level seven formations (Figure 2) detailed below. Jackson forming thin beds mixed with argillaceous clays and lentils of coarse sands. Tuffaceous material derived from Eocene pyroclastic eruptions is common throughout the Pseudokarst development in East Texas has only been 312Figure 2. Simplified stratigraphic section of East Texas Eocene units. Yellow indicates relative abundance of sandstone dominate facies, while grey represents relative abundance of mudstone facies. Claiborne Group divided into formations.Figure 1. Digital elevation model of seven East Texas counties encompassing study area. Note the extensive drainage development associated with low gradient fluvial systems in East Texas clastic strata (data from Texas Natural Resources Information System).

PAGE 330

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2313to early Cenozoic through buckling induced by the large, low amplitude, anticlinal feature in East Texas Fault Zone borders the eastern and northern portions Zone dissecting the study area (Figure 4), associated Lawrence shear system respectively (Adams, 1990). These fault systems have produced abundant near vertical fractures throughout the study area primarily planes for fluid migration. Stratigraphically the region is dominated by Eocene clastics (Figure 2). Wilcox strata are largely undifferentiated in the region because of heterogeneous and lignite lenses, with total thicknesses exceeding 500 Angelina, Cherokee, Nacogdoches, Panola, Rusk, San Augustine and Shelby counties cover 15,146 km2 within the Interior Coastal Plains (Wermund, 1996), Climate in the region is subtropical humid with annual mm, respectively (Estaville and Earl, 2008). Average precipitation increases slightly in late fall and spring with slight decreases in late summer. Temperature averages Earl, 2008). The region is dominated by mixed pine and hardwood forests with numerous low gradient streams.East Texas GeologyEast Texas is dominated by the deposition of Cenozoic clastic sediments associated with the transgression and regression of coastal strandlines that deposited extensive fluvial, deltaic, lagoonal including strata of the Wilcox, Claiborne and Jackson groups as well as overlying quaternary alluvium region is dominated by the Sabine Arch (Figure 4), a basement uplift formed during the late Mesozoic Figure 3. Simplified geologic map of East Texas study area, with Jackson and Wilcox Groups undivided. Cook Mountain, Reklaw, Yequa and Weches formations of the Claborne Group are undivided. Carizzo, Queen City and Sparta sandstones are presented in contrast because these formations are known to host pseudokarst development (data from Texas Natural Resources Information System). Figure 4. Simplified structural map of East Texas, including the Sabine Arch, Mexia-Talco Fault System and the Elkhart-Mt. Enterprise Fault Zone (adapted from Martin, 1978). Enlarged area shown on inset map by dashed box outline.

PAGE 331

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE314East Texas PseudokarstPseudokarst development in East Texas occurs as sinkholes, springs and isolated caves in competent facies, while loose, unconsolidated sediments host numerous small suffosion features (Stafford et al., 2010). As with any environment where competent bedrock is overlain by loose unconsolidated material, suffosion features occur as both natural and anthropogenically enhanced structures. Many of these suffosion features act as macropores and systems. True pseudokarst development does occur in the Carrizo, Queen City and Sparta sandstones; some of these features are directly associated with sinkholes, are associated with spring discharge along permeability horizons. Cave development is limited to competent facies and occurs most commonly in variably cemented zones, where heterogeneous cementation promotes both stable cave development and mechanical disaggregation of clastic grains. Boatman Cave Complex (Figure 5B) in northern Nacogdoches County represents typical pseudokarst development in East Texas, where a series of springs discharge from the Carrizo Sandstone at a low permeability contact. While only one of the three these small caves are each developed along a vertical fracture plane where laterally migrating groundwater has physically disaggregated sandstone grains near the land surface interface, resulting in three distinct springs converging and discharging into an incised valley. Each with the largest feature actively developing an upward stoping chimney. Most pseudokarst features in the East Texas region exhibit this typical morphology and are not associated with sinkhole development. At a slightly larger scale, Tonkawa Springs in northern Nacogdoches County is associated with Camp Tonkawa Cave (Figure 5C) which consists of cave development along an enlarged vertical fracture in the Carrizo sandstone. The cave is primarily developed along an through; a secondary spring inlet converges in the western portion of the cave before discharging to the not known for pseudokarst development in the region, the units in the region with recharge occurring proximal to Claiborne strata in the East Texas unconformably overlie Wilcox strata and consist of alternating and complexly and claystones of lagoonal, shallow marine and deltaic facies are more common towards the northern portion abundantly to the south. Claiborne strata consist of a typical rhythmic series of continental and shallow marine facies and ranges up to 450 m thick. Conformably deposits with zones of argillaceous and tuffaceous clays and tuff, which can reach total thicknesses of 90 m Within East Texas, previous pseudokarst development Recklaw, Queen City, Weches, Sparta, Cook Mountain and Yegua formations from oldest to youngest (Figure sand and 10% sandy clay with the lower portions more heterogeneously cemented by ferruginous sandy silty clay, 7% bentonite and glauconite clay and Sparta sandstones were all deposited in continental to littoral environments and are known to host pseudokarst features; however, the more indurated and heterogeneously cemented Carrizo strata appear to be the most favorable for zones for pseudokarst development. Other formations within the Claiborne promote pseudokarst development in the region.

PAGE 332

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2315meters long with over 160 m of surveyed passage and a depth of 12 meters. The cave consists of a linear dominant spring and two secondary springs. This suite of springs has formed a lower, northern passage and a higher, southern passage that converge into a single large chamber in the western portion of the cave. The chamber is approximately eight meters tall, almost ten meters wide and encompasses the central quarter of the cave. Throughout the cave, numerous small alcoves and ceiling structures occur suggesting a similar speleogenetic origin as Tonkawa Cave but on a much grander scale. The cave opens to the west, where spring discharge forms an incised valley with additional small alcove caves along its margins, while the eastern portion approximately 15 m wide and 50 m long that gently slopes into a watershed covering almost one hectare. the spring was previously exploited for natural spring water bottling and once powered a grist mill (Brune, 1981); however, much of the original morphology can still be discerned. While some of the four meter tall cave chamber appears to be the result of upward stoping processes, much of it appears smooth and indicative of that most of the cave was originally formed as pressurized morphology, suggesting that the spring associated with cave formation may have an artesian component to it. in the Carrizo Sandstone in Shelby County with the most extensive pseudokarst cave currently documented Figure 5. Simplified cave maps of representative pseudokarst caves in the study area, including (A) Bridges Cave, (B) Boatman Cave Complex, (C) Camp Tonkawa Cave and (D) Gunnels Cave.

PAGE 333

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE316 closed depressions within the 15,146 square kilometer not being actual pseudokarst features, based on similar by karst processes (Bryant, 2012). Closed depressions that overlapped or that occurred within ten meters of streams and rivers were removed as these features may collapse or suffusion pseudokarst features; however, it is probable that this process also removed some true Closed depressions that intersected or occurred within ten meters of ponds, lakes and surface impoundments were removed because they represent anthropogenic closed water resources. Closed depressions within ten meters of roads and highways were removed because it is probable that most of these features are the result of anthropogenic activity associated with infrastructure construction, based removed because pseudokarst development has not been documented in these strata within the study area. After City and Sparta sandstones; however, these features are limited to those closed depressions that cover at least one delineation of closed depressions is consistent with studies Density analyses of delineated closed depressions and probable pseudokarst sinkholes indicate clustered trends of development. The highest concentrations of closed depressions occur in Wilcox strata and Quaternary are primarily associated with abundant meandering, from the late 19th century. In contrast to the pseudokarst caves described above, small bypass caves do develop in variably cemented and in the western portion of the study area is developed in the Sparta Sandstone where heavily hematite cemented horizons provide both permeability and structural breached a heavily indurated zone approximately one Flow continues to traverse laterally on top of a second indurated layer, where the cave formed over a distance of approximately ten meters. Small alcoves exist within the during intense storm events that rapidly increased the GIS Analysis of East Texas PseudokarstAnalyses of karst terrains have been greatly aided in the past decade by improved digital resources that enable the precision of the results are directly proportional to the quality of available data and is no replacement for high precision data, extensive vegetation and abundant state, LiDAR (Light Detection and Ranging) data does not occur for most of East Texas, with the exception of limited data recently collected through TNRIS (Texas Natural Resource and Information System) in the proximity of the are limited to low resolution (10 meter) digital elevation models derived from digitized 1:24,000 quadrangle maps densely forested regions has very limited application in A sinkhole analysis was conducted on the seven counties of interest in East Texas as an assessment of the feasibility for pseudokarst delineation across the region. Closed depressions were delineated across the study area through

PAGE 334

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2317 cave in East Texas, is truly a unique anomaly occurring in an extremely low sinkhole density region. While density analyses does provide indications of regions of more probable pseudokarst development, the nature of the original data creates a distinct bias that eliminates the level of complexity in evaluating whether features are the result of pseudokarst development or are constructional In addition to density analyses, slope analyses can be used development; however, the same limitations of data are beyond the angle of repose for loose, unconsolidated sediments. Therefore, these regions are likely areas where collapse structures or incised valleys occur in more competent facies. By comparing regions where slopes greater than thirty degree occur with regions of of probable pseudokarst development can be delineated (Figure 8). While these high gradient regions continue to indicate the western portion of the study area likely has low gradient streams and oxbow lake environs in the northeastern portion of the study area (see DEM on Figure 1), with secondary abundant densities occurring in area. Density analyses of probable pseudokarst sinkholes (Figure 7) indicate that the greatest concentrations occur within the western portion of the study area in Sparta and Queen City sandstones and in the northern portion of Figure 6. Closed depression density identified through DEM analyses (data from Texas Natural Resources Information System). Figure 7. Probable pseudokarst sinkhole density identified through DEM analyses after filtering (data from Texas Natural Resources Information System).Figure 8. Comparison of steep scarps (slopes >30 degrees) with probably pseudokarst sinkhole density identified through DEM analyses after filtering (data from Texas Natural Resources Information System).

PAGE 335

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE318through unconsolidated sediments throughout the economic loss, attesting to the need for greater public education within the region. In spring 2012, several small earthquakes, up to 4.8 in magnitude, occurred near Timpson, Texas in the study area, which were reported to have induced collapse and sinkhole formation. shallow suffusion features and were likely the result of water from leaky pipelines affected by the ground movement and are not associated with true pseudokarst development in the region.References Speleological Society. Estaville LE, Earl RA. 2008. Texas water atlas. College continental margin stratigraphic and structural characteristics of the upper continental margin. Tulsa (OK): American Association of Petroleum Summary of subsurface geology of northeast Texas. In: Beebe BW, Curtis BF, editors. Natural gases of North America. Tulsa (OK): American Association Cave Books. geology of Texas, Volume I stratigraphy, The the greatest pseudokarst development and warrants more should be investigated in more detail. Numerous high angle slopes occur within this region which are likely associated with entrenched valleys and potential spring discharge points that may have formed pseudokarst features in the Carrizo Sandstone.ConclusionPseudokarst development is limited in East Texas and City and Sparta sandstones. Pseudokarst features include sinkholes, springs and caves, which are largely associated with the lateral and vertical migration of as variable ferruginous cementation primarily control the lateral development of pseudokarst caves. Most caves appear to be largely the result of lateral migration of shallow groundwater; however, speleogens in some caves suggest that an artesian component of groundwater pseudokarst features. Density analyses of digital elevation models of the region indicate that pseudokarst development is most extensive in Cherokee and Rusk counties (Figure 7); however, when coupled with slope analyses to identify entrenched valleys, other trends are discernible suggesting that northern Nacogdoches, southern Shelby and northwestern San Augustine counties are also probable sites of more intense pseudokarst development (Figure 8). Although these data are promising, the limitations of quadrangle maps presents a large sampling bias based on data quality, but these data do provide preliminary East Texas region. Most sinkholes and collapse structures in East Texas are the result of suffusion processes; however, true pseudokarst features are common within the region. Because suffusion features are more common, they provide greater infrastructure and economic concern in the region. Leaky pipelines, poor placement of storm runoff and building construction often focus water

PAGE 336

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2319Shelby CA, Pieper MK, Wright AC. 1968. Palestine Plain, Eddy County, New Mexico and Culberson of East Texas: geomorphic features and evolution. Programs 42 (2): 99.

PAGE 337


PAGE 338

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 potentiometric surface and water level changes were used for regional scale assessment of sinkhole hazards. A relatively simple monitoring method is needed to In the past decade, sinkhole collapses occurred every the ongoing construction of a new natural gas pipeline assess sinkhole hazards along the proposed natural gas pipeline in this area.Study Area province (Figure 1). This area belongs to the central during the raining season between April and September. This is a typical fenglin and tower karst plain. The altitude is mainly agriculture for crops and small patches of rice paddies. The unconsolidated sediments above bedrock are Quaternary alluvium and colluvium with a thickness ranging between 2.0m and 10.0m. Quaternary deposits are clay and silty clay containing gravels. The bedrock Devonian Donggangling formation. Limestone is massive AbstractIn the past decade, sinkhole collapses occurred every was established to assess sinkhole hazards in this area. Monitoring of water levels of residential and community wells indicate a water shortage in the shallow karst the study area. Domestic and irrigation water usage could result in frequent and dramatic changes of water level in monitoring of the groundwater level in a referenced area with no sinkhole collapsing event, a characterization process was developed to assess sinkhole hazards in the Qingyun village area. Characterization criteria recovery of water level in the karst aquifer, maximum declining rate of water level in the karst aquifer, and sinkhole distribution within 500 m of water pumping activity. The characterization process was then applied to the study area to identify and prioritize areas that are most likely to be affected by human activities. This characterization process could be used by engineers induced sinkholes in active karst areas.IntroductionSinkhole hazard assessment has been focused on structural controls, bedrock geology, depth to bedrock, epikarst characteristics, geomorphology, distribution of karst features especially sinkhole distribution, water level to the top of bedrock, pumping rate and duration, and water level changes associated with drilling (Lei CHARACTERIZATION OF KARST COLLAPSE HAZARD BASED ON GROUNDWATER FLUCTUATIONS IN Xiaozhen Jiang, Mingtang Lei Zhende GuanInstitute of Karst Geology, CAGS, Guilin, China, jiangxiaozhen2005@hotmail.comYongli GaoDepartment of Geological Sciences, Center for Water Research, University of Texas at San Antonio, TX 78249 USA, yongli.gao@utsa.edu321

PAGE 339

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEFour monitoring sites were established at existing water wells. Sites 1 and 2 are located in Qingyun village where located in Jitang village where no sinkhole collapses occurred in the past. The distance between Qingyun village and Jitang village is approximately 8 km. with highly fractured exposures. The study area is also MethodologyNo surface water resources exist in the study area. Shallow karst water is scarce and directly affected by for agriculture, industry, and domestic water supplies. Pumping groundwater has been thought to induce many sinkholes in this area (Figures 1 and 2). A comparative study of hydrodynamic changes of karst water is conducted in active karst areas with sinkhole collapses and inactive karst areas without sinkholes. on previous model experiments on sinkhole and soil void formations, measurement intervals need to be less sinkhole collapses. Measurement intervals were set at 10 322Figure 2. A collapsed pit caused by groundwater pumping.Figure 1. Google Earth Map of the study area showing sinkhole distribution and natural gas pipeline.

PAGE 340

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 groundwater is affected mainly by precipitation. For example, rainfall started in the morning on January 25, associated with the natural gas pipeline construction pumped water out of the well during December 10 and by water pumping. Only a limited amount of water was pumped out of this well and water supply is scarce at this the monitoring period. Site 3 Hydrodynamic Changes supply with a depth of 90 m. Water levels were monitored between February 24 and August 24, 2011 with a measurement Changes of Hydrodynamic ConditionsSite 1 Hydrodynamic Changes dug well with a depth of 11.4 m. Water levels were monitored between February 28 and September 2, 2011 measurements were collected. the monitoring study. The lowest water level is 10.47 m below surface at dry season and the highest water level is 0.41 m during monsoon season. The range between the highest water level and lowest water level is 10.06 m. The thickness of Quaternary sediment is between 6 and 10 m in this area. Therefore, karst water has been but leakage of irrigation water for agriculture as well. For level at site 1 remains at a higher level similar to the level of monsoon season. Further investigation reveals that Figure 4 illustrates water level changes caused by pumped out 4 times between 8:00 and 15:00. A decline of 5.2 m was caused by initial pumping between 8:00 and 10:00. The duration of the subsequent pumping was of the water level at approximately 1 m. It took more than 6 hours for the recovery of the water level after the last pumping. In addition, the rate of water level changes is an important factor triggering sinkhole collapses. Site2 Hydrodynamic Changes with a depth of 6.75 m. Water levels were monitored between December 5, 2007 and January 28, 2008 with a measurement interval of 10 minutes and 144 measurements were recorded each day. Monitoring study occurred during a very dry season with the maximum decline of water level and frequent water level changes caused by water pumping.323Figure 4. The variation of groundwater level of monitoring site 1 caused by pumping activities on August 21, 2011.Figure 3. The variation of groundwater level of monitoring site 1. Results and Discussion

PAGE 341

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE dug wells to shallow karst water. Pumping activity in slow recovery due to scarce water supply in shallow karst aquifers, which may cause soil piping and soil void formation and eventually trigger sinkhole collapses. By comparing hydrodynamic conditions between Qingyun village and Jitang village, characterization criteria of sinkhole hazards are as follows: Daily water level change and the recovery of water level after pumping: of groundwater withdrawn, and the recovery of water level after pumping is related to the amount of water storage and supply in the karst hole hazard areas: Low risk area: daily water level change is less than 1.0 m and the recovery of water level is less than 5.0 hours after pumping. The lowest water level is 8.48 m below surface during dry season and the highest water level is 1.11 m below surface during monsoon season in the duration of the monitoring study. The range between the highest water level change is normally less than 4.57 m. Figure 5 shows water level changes on May 21, 2011. Pumping activity is controlled by the storage of water in the water tower. Water level declines drastically during pumping and recovers relatively fast after pumping has stopped (Figure 5). at this site because of a higher pumping rate for water times during the monitoring period (Figure 6). Site 4 Hydrodynamic Changes is 191 m. Water levels were monitored between February 27 and September 2, 2011 with a measurement interval The lowest water level is 4.95 m below surface at dry season and the highest water level is 0.26 m during monsoon season. The range between the highest water level and lowest water level is less than 4.70 m. Daily water level change is less than 0.19 m. The maximum Characterization of Sinkhole Hazards Based on Hydrodynamic ConditionsSinkhole collapses have occurred in Qingyun village every year in the past decade. No sinkhole cases were though heavy pumping is a common practice due to limited water supplies in shallow karst aquifers. A comparative study of characterization of sinkhole hazard is based on hydrodynamic conditions observed in these two study areas. Soil cover thickness, hydrodynamic conditions and sinkhole occurrences were listed in Table 1 based on 324Figure 6. The variation of groundwater level of monitoring site 3.Figure 5. The variation of groundwater level of monitoring site 3 caused by pumping activities on May 21, 2011. (1)

PAGE 342

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2of groundwater levels in a referenced area with no sinkhole collapsing event, a characterization process was developed to assess sinkhole hazards in the Qingyun village area. Characterization criteria include of water level in karst aquifer, and maximum declining rate of water level in karst aquifer. The characterization process was then applied to the study area to identify and prioritize areas that are most likely to be affected sinkhole hazard areas based on daily water level changes time monitoring of daily water level changes, water level recovery, and the rate of water level changes will provide guidelines and limit water pumping activities to reduce potential sinkhole collapses due to increased water demand caused by the construction of a natural gas pipeline through Qingyun village.References assessment in Minnesota using a decision tree Intermediate risk area: daily water level change 10.0 hours after pumping. greater than 1.0 m and the recovery of water level is greater than 10.0 hours after pumping. Maximum declining rate of karst water level: Sudden water and air pressure changes within karst fractures and conduits have been associ ated with sinkhole collapses in many cases. Model experiment based on geologic veals that severe soil damage would occur when the declining rate of water level is above 180 study area is way below the critical value of 180 of sinkhole hazard. More quantitative assessment is needed based on the above criteria. Pumping activities along with ongoing monitoring of daily water level changes, water level recovery, and the rate of water level changes will provide guidelines and limit water pumping activities to reduce potential sinkhole collapses. Conclusions established to assess sinkhole hazards in this area. Monitoring of water levels of residential and community wells indicate a water shortage in the shallow karst 325 Site Thickness of soil coverage (m) Rate of declining water level Time of water level recovery (hour) Daily water level change (m) Maximum water level change (m) # of sinkholes within 500 m Connection between karst and Quaternary 1 8.0 6.58 10 4.4 10.06 Y 2 6.0 6.85 12 2.6 4 Y 1.5 15.26 0.5 4.57 0 N 4 1.8 0.65 5.0 0.8 4.7 0 NTable 1. Soil coverage and hydrodynamic conditions in the study area. (2)

PAGE 343

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE fuzzy Prediction and evaluation of karst collapse in Zaozhuang City, Shangdong Province. The Waltham T, Bell F, Culshaw M. 2005. Sinkholes and Engineering and Construction. Chichester (UK): Praxis Publishing. Zisman ED. 2001. A standard method for sinkhole detection in the Tampa Florida area. Environmental 326

PAGE 344

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2Geologic Settings highly active karst area containing many karst features such as sinkholes, springs, karst windows, caves, and conduit systems (Figure 1). The study area is a typical fengcong and fenglin karst area with isolated and dissolved hills and valleys. The unconsolidated sediments above bedrock are Quaternary alluvium and colluvium. Quaternary deposits consist of silty clay, clay containing gravels, and clay. Karst bedrock units belong to the middle Carboniferous 2h) and Nandan Formation (C2n fossiliferous limestone, dolomite limestone, siliceous and fossiliferous limestone, limestone containing gravels, AbstractA series of sinkholes collapsed at Jili village and transect of the study area and passes the collapse site. Preliminary investigations revealed that extremely have triggered this collapse event. The precipitation, as high as 469.8 mm within one day, was a record high in the study area. A long period of drought in 2009 followed by extremely heavy rainfall along with cave roof collapse may have caused the collapse event on triggered earthquakes caused severe ground failure and fractures in residential houses and Jili Dam. Several collapse events were caused by extreme weather studies of the relationship between extreme weather events and sinkhole collapses will help minimize the damage or impact to human infrastructure by avoiding areas susceptible to collapse or by designing infrastructure to better withstand subsidence. Introduction collapsed at Jili village and Shanbei village, Laibin merged to form a 200 m long collapse zone. Many ground failures and fractures occurred in the area. An area of 0.4 km2 was impacted by the collapse event. Because the collapses occurred in areas with a high density of population, initial investigation results reveal severe damage to residential houses. A total highway were impacted by the collapse event. This paper discusses the geologic background, possible mechanism of sinkhole collapses, and future studies of sinkhole hazard assessment in the study area. INVESTIGATIONS OF LARGE SCALE SINKHOLE Yongli GaoDepartment of Geological Sciences, Center for Water Research, University of Texas at San Antonio, TX 78249, USA, Weiquan Luo, Xiaozhen Jiang, Mingtang Lei, Jianling DaiInstitute of Karst Geology, CAGS, Guilin, China, 327Figure 1. Geographic location of Laibin.

PAGE 345

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCESinkhole DistributionPreliminary investigations revealed that extremely have triggered this collapse event (Figure 2). The precipitation, as high as 469.8 mm within one day, was a expanded and merged to form a 200 m long collapse zone (Figure 4). Walls of these sinkholes are not stable and these sinkholes kept growing after the initial collapse (Figure 5). Earthquakes which occurred on June 1 were probably caused by cave roof collapses. The water hammer effect caused by cave roof collapse can release a pressure surge to the karst conduit system and sediments overlying the karst conduit, causing severe soil damage and subsequent collapses (Lei et al., 2010; lower Cretaceous Yongfu Formation (K2y). Rock types include reddish arkosic sandstone and mudstone. Surface water and Quaternary groundwater are scarce water within bedrock matrix, fracture, and conduits and a limited amount of Quaternary water in porous sediments. Fairly large springs exist in the area, which are recharged through sinkholes, active karst fractures and conduits. m River, located 16 km north of the study area, marks the Preliminary investigations of sinkholes, subsidence areas, and large springs reveal that a large cave stream exists in the may serve as discharge outlets of the cave stream. The cave stream passes through the sinkhole plain and discharges to 328Figure 2. Daily precipitation between September 2009 and August 2012 in the study area.

PAGE 346

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2329Figure 4. Sinkhole distribution in the study area. Figure 3. Hourly precipitation between May 1 and June 4 2012 in the study area.

PAGE 347

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE water hammer effect in the karst conduits. The water hammer effect can cause severe soil damage and trigger Soil disturbance may change hydraulic gradient, which in sinkhole collapses. Recommended techniques for assessing sinkhole hazards include: potentiometric mapping, locating areas of disturbed soil and soil voids using ground penetrating changes, sinkhole inventories, and tracer test of surface water and groundwater interaction. These approaches are being conducted in other areas to prevent or forecast sinkhole collapses and to minimize the damage caused by sinkhole collapses (Jiang et al., 2008; Jiang et al., Karst water levels are recovering to normal conditions after the collapse event. It may take several years to effect may also have caused fractures in many houses and the Jili Dam (Figures 6 and 7).ConclusionsRapid water level rise after the storm event caused a series of sinkhole collapses in the study area. Several collapse events were associated with extreme weather extreme weather events and sinkhole collapses will help minimize the damage or impact to human infrastructure by avoiding areas susceptible to collapse or by designing infrastructure to better withstand subsidence. occurred during heavy rainfall. The precipitation, as high as 469.8 mm within one day, was a record high amount in the study area. A long period of drought in 2009 followed by extremely heavy rainfall along with cave roof collapse may have caused the collapse event 330Figure 7. Fracturing through ground and houses.Figure 6. Ground failure and damage to a residential house.Figure 5. Sinkhole no. 3 with unstable walls.

PAGE 348

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2stabilize the soil to normal conditions. A 10 m x 10 m soil collapse area was recently discovered in the south portion of the study area. A donut shaped subsidence area formed around the new collapse site. The diameter of the subsidence area is 100 m with ground failures and fractures formed inside the subsidence area. This area is located directly above the subterranean stream. Large scale sinkhole collapses may occur again during monsoon seasons. Residents in Jili village and Shanbei village may need to be relocated to a safer place. Jili further damage.AcknowledgementsField assistance and cooperation of local residents in Jili village and Shanbei village are highly appreciated. helpful reviews.References soil void formation along highway subgrade using Yuhr L, Alexander E, Beck B, editors. Proceedings of the Eleventh Multidisciplinary Conference on Sinkholes and the Engineering and Environmental and treatment of sinkholes and subsurface voids In: Yuhr L, Alexander E, Beck B, editors. Proceedings of the Eleventh Multidisciplinary Conference on Sinkholes and the Engineering and Environmental monitoring of karst water pressure in Zhemu, China. Investigation of Extremely Large Sinkholes, 331

PAGE 349


PAGE 350

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 events. A relationship could be established between corresponding elevation of the groundwater level within the karst conduits (or respectively, the relevant spring groundwater is expected to reach the ground surface areas have been delineated and mapped according to the rates at the main outlets have been assessed. Introduction 1st 1804 in Porrentruy (JU) is a the largest known event Associated discharges reaching 100 m recorded for this karst system. Similar events were recorded in 1901 and 1910 (Figure 2). In addition to 9th approached 20 m The local authority (administration of the Jura canton) has to plan protective measures to diminish the potential understanding and prediction of such extreme situation AbstractThe city of Porrentruy (JU, Switzerland) is vulnerable 1901 led to heavy damages throughout the city and its in substantial costs. by a perennial outlet (the Beuchire spring) and several rate often exceeds 15 m not closely correlated. Therefore, the discharge rates active could not be assessed without a comprehensive understanding of the karst system behavior. Thus, area has been built in order to reproduce the aquifer base geometry, the extension of its expected saturated part(s) axes. This approach enabled the catchment area delineation by combination of subterraneous drainage axes. The comparison of the discharge time series of the consequent reproducing of threshold functionality of MAPPING FLOOD-RELATED HAZARDS IN KARST USING THE KARSYS APPROACH: APPLICATION TO THE BEUCHIRE-CREUGENAT KARST SYSTEM Jonathan Vouillamoz, Arnauld Malard, Eric Weber, Pierre-Yves JeanninSwiss Institute for Speleology and Karst Studies, Rue de la Serre 68, CH-2301, La Chaux-de-Fonds, info@isska.chGabrielle Schwab RougeBG Ingnieurs Conseils SA Avenue de Cour 61, CH-1007, Lausanne, Gabrielle.SCHWABROUGE@bg-21.com333

PAGE 351

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEm a.s.l. It becomes active only at high water stage. The global discharges in the city of Porrentruy may Further upstream, at an elevation of 465 m a.s.l., lies second temporary outlet of the system and becomes active only during very high water stage (see Figure 4). The discharge series of the Beuchire spring and have been measured at hourly time steps, respectively between 1998 to 2004 and 1998 to 2008. In addition to these three main springs, a series of minor temporary outlets do exist. Unfortunately they are badly documented due to infrequent activity (Les sources, Libecourt, etc.). is then required for assessing the probable occurrence the study region. The assessment was conducted applying the KARSYS approach (Jeannin et al. 2012) expanded with some hydraulic considerations. The aim was: karst water table for various recharge scenarios. Recharge events with return periods ranging To determine and map areas which are the most The locations where karst groundwater is expected to reach the most forward ground surface are the most catchment areas of the underground tributaries, discharge ContextThe geological context refers to the Tabular Jura which is slightly folded and intersected by numerous strike perennial karst spring emerges in the center of Porrentruy which is composed of alternating units of Upper Jurassic limestone and thin layers of marls. The Malm aquifer is underlied by a thick marl formation (Astartes marls, water exchanges with the lower aquifer are highly likely through discontinuities in the marls. Upstream of the Beuchire spring, the Creugenat 334Figure 2. The flood event of 1910 in Porrentruy. Inundation is the result of extreme karst aquifer discharge through the Creugenat (Archives de la Bourgeoisie de Porrentruy, JU).Figure 1.The Beuchire karst spring is located in the city of Porrentruy (JU) in Northwest Switzerland.

PAGE 352

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2was applied to estimate the geometry of the aquifer(s) boundaries, to delineate groundwater body(ies), and to assess the functioning of the Beuchire spring and This approach was assessed for low, medium, high and extremely high water conditions. Although the Beuchire spring and the Creugenat Schweizer 1970, Monbaron 1975, etc.) available data a clear solution. None of these studies describes in details the potential catchment boundaries and their possible changes in relation to the water stage. In this context the KARSYS approach (Jeannin et al. 2012) 335Figure 4. SW-NE profile of the Beuchire catchment and projection of the overflow outlets. The real distance between the Beuchire spring and the Creugenat temporary overflow is 4,3 km. The Creugenat overflow and the Creux-des-Prs temporary overflow are 1.45 km appart.Figure 3. The Creugenat overflow in 1934. First pumping test to dry up the siphon (picture A. Perronne). A 3D model to assess the aquifer geometry geological model focusing on the aquifer basement (i.e. Astartes marls) was established for the area of interest 5) to meet the requirements of a pragmatic issue. This was possible thanks to an extensive compilation of all existing data relative to geological information (borehole which provided a strong basis of documentation. Once the geological model has been established and the data checked, the hydrological features have been temporary ones. Then, the extension of the saturated part of the aquifer was assessed by following the KARSYS approach. This approach assumes that at low water stage, the top of the saturated part of a karst aquifer is close to horizontal and can be represented within the model by a horizontal plane at the main perennial spring elevation (the Beuchire spring in the present case). The portion of the aquifer located underneath that horizontal plane should be close to the volume extension of the karst phreatic zone.

PAGE 353

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE respective drainage axes were rendered. Drainage axes are recognized as vadose ones if they are located above the saturated zones. They are assumed to be developed at the bottom of the aquifer along the dip of the basement. Phreatic passages located within the saturated zones are drainage axes linking input points into the phreatic zone to the main drainage horizontal and a priori follow the shortest hydraulic distance to the outlet(s). The model result for low water stage is presented in Figure 7. The total groundwater catchment area in low water situations is thus estimated at about 79 km2 Assessing the system at high water stage (hydraulic gradients within the conduits)In the next step, the hydrology of the system is assessed for high water conditions, i.e. active according to the rise of the groundwater head in the conduits. The discharge data from the Beuchire spring are compared with the head data recorded at the Creugenat This process was applied to the model and gives the following results (Figure 6). threshold which top elevation is at 440 m a.s.l. This fact also predetermined the volume extension and 2 (A) and 6 km2 (B) and taking into account the assumed the water content could be estimated; respectively as 6.4 Mm and 2.9 Mm at the low water stage.Knowing the location and volume extension of the saturated parts of the aquifer, the catchment area feeding each of these groundwater bodies could be delineated. Then, following the shape of the 336Figure 6. Model identification of two main groundwater bodies (GWB), A = Beuchire GWB (elevation 423 m a.s.l), B = Bonnefontaine GWB (elevation 438 m a.s.l).Figure 5. Illustration of the 3D geological model of the aquifer basement from the Beuchire karst system intersected by geological cross-sections. This was modeled using GeoModeller (Intrepid Geophysic). The main spring is on the front of the model. Figure 7. Model identification of the main underground vadose (yellow) and phreatic (blue) flowpaths of the Beuchire-Creugenat karst system at low water stage.

PAGE 354

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 type), the relation can be simulated using the following equation: (Eq. 1) With Q [m2) [m plotted on the chart (models 1 & 2, Figure 8). Model 1 suggests that the hydraulic connection between when the water elevation in the conduits ranges from and 451 m a.s.l. Therefore, model 2 depicts a hypothetical outlet at is valid for an outlet located at a distance of 4,000 m and diameter of 2.7 m) to reproduce the observed trend. Previous observations led to the hydraulic schemes of the karst system presented in Figure 9 that depends on the groundwater level elevation. This provides a set of displayed on Figure 10 where the gradients correspond to event) at high water stage. Upstream of the Beuchire spring the hydraulic gradient strongly increases until it is close to 0.7%). If the water level still increases the mapped in yellow. Similar scenarios could be established for two larger remains comparable to the value encountered above The graphic indicates that: As the Beuchire spring discharge remains lower indicates that both systems are disconnected by a threshold (a) situated at this elevation. level progressively rises up and shows a level at the Creugenat rises up a second time until it reaches the output elevation (c) at 451 m a.s.l threshold (b) as a first activation of the Creugenat overflow, this analysis indicates that the Creugenat becomes active only when the Beuchire spring indicates that an intermediate overflow (or large a.s.l). This could be a karst conduit or an outlet to the ground surface.337 Figure 8. Comparison of the hourly pressure data recorded at the Creugenat overflow and the hourly discharge values of the Beuchire spring during flood (grey) and recession (red) events (2002-2004). Model 1 simulates a threshold discharge at an elevation of 437 m a.s.l at a suggested distance of 1,300 m downstream from the Creugenat. Model 2 simulates an ideal function of the spring discharge using a k*S of 28 m3/s (conduits diameter of ~2.7 m) and a straight distance of 4,000 m.

PAGE 355

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEthe gradient): its slope is approximately 1% extending the groundwater bodies as pictured in Figure 11. The areas the hydraulic gradient within the conduits does not change between the Beuchire spring and the Creugenat of the more elevated outlets (higher than 500 m a.s.l) and the shape of the versants. The gradient is therefore approximately 1.5%. Surfaces that are here vulnerable bottom of the valley (several outlets were active during 338Figure 9. Sequential evolution of the hydraulic gradient within the Beuchire-Creugenat karst system (i.e. the conduits) for an average annual flood event reaching the Creugenat overflow. The profile of the conduits is here supposed. Processes are the following: 1. The groundwater level at the Creugenat overflow is independent of the Beuchire spring discharge oscillations; 2. The water level at the Creugenat overflow is controlled by the threshold (a); 3. At 443 m a.s.l. the activation of an additional conduit (or a perched spring) show a lag in the water level elevation rise at the Creugenat overflow; 4. The rise of the groundwater level at the Creugenat overflow depends on the Beuchire spring discharge. At 451 m a.s.l the Creugenat overflow is now flowing! Figure 10. Model prediction of the extension of groundwater bodies (GWB) A and B during a flood event reaching the Creugenat overflow (=multiannual occurrence =case 4 in Figure 9). Water from GWB B overflows over two passes and contributes to the discharge of the Beuchire spring. Figure 11. Model prediction of the saturated groundwater bodies extension in the BeuchireCreugenat karst system during a flood event reaching the Creux-des-Prs overflow (~30-year flood event).

PAGE 356

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 about 0,5% (in the swiss Jura, according to Bauer et al. 1980, Burger and Pasquier 1984). This value may be Mapping the flooded areasAccording to the previous model results it is possible to map surface areas which could be affected by the potential In addition to these gradients some further temporary by Bouvier 2006. They were used as controls for the Considering the respective values of the hydraulic gradients during these events, it is possible to estimate the 339 Figure 12. Model predicted storage in the aquifer (i.e. the karst conduits) and its development due to groundwater increase within the karst system for the respective flood events (average annual, 30-year flood events and ~300-year flood events). The associated volumes refer to water potentially involved in the floods (1.7 Mm3 for an average annual flood event, 4.4 Mm3 for a 30-year flood event and more than 6 Mm3 for a 300-year flood event).Figure 13. Flood hazard map of the Beuchire-Creugenat catchment area. Color code refers to the considered occurrence: average annual flood event, 30-year flood event, 300-year flood event). Filled areas have to be considered as potentially exposed to flooding or at least as potentially impacted by an overflow from a temporary outlet. The interpreted drainage axes (both vadose and phreatic) are also displayed on this map.

PAGE 357

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEdeepening of the collapse. The recorded oscillations of the groundwater are consistent with the previous interpretation related on Figure 9. provides new calibration elements that improve the model functionality. Currently, a series of simulations are being conducted using the actual release of SWMM event. This is possible once the catchments of these have been clearly delineated as well as their respective recharges. This assessment will be conducted in the second part of model. In the present paper only the maximum discharge of the respective outlets (permanent and temporary ones) was construction or hydraulic works.Maximum expected discharges at the outletsDischarge of springs usually increases as the level of the groundwater increases within the system until groundwater level still rises in the conduit, the discharge on that principle, it is possible to estimate the maximum expected discharges at the Beuchire spring and the Regarding the Beuchire spring, the maximum discharge (Figure 8). Even if the groundwater level in the karst system still increases the Beuchire spring cannot The maximum discharge rate at the Creugenat outlet can be estimated from the discharge of the stream in the city was estimated that the maximum discharge rate at the DiscussionIn June 2012 a collapse appeared in the middle of a road, on a straight line between the Beuchire spring and the of 8 m and at that time its bottom was dry. This collapse gave the SISKA the opportunity to install a pressure sensor with the aim to survey an eventual presence and oscillations of the groundwater within the pit. The recorded data (only few weeks available) show periodic rises of the groundwater with a maximal elevation of 340 Figure 14. The new collapse of the community of Courtedoux gave access to a 8 m deep pit. The bottom of the pit was monitored with a pressure sensor. The recent data allowed validation and improvement of the Beuchire-Creugenat karst system model functionality.

PAGE 358

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2in conduit geometry characterization and to improve the hydrological model. The last could be applied in further hydraulic planning, especially in estimating the groundwater discharge contribution for each basin unit. Applications to assist the design of future construction and hydraulic works could also be envisaged.Acknowledgments The study of the characterization and prediction of the and caving clubs for providing information and data. References Bauer F, Benischke R, Bub F, Burger A, Dombrowski mit natrlichen und knstlichen Tracern im Neuenburger Jura .Schweiz.. Steirische Beitrge du Doubs .France. et du canton du Jura .Suisse.. Burger A, Pasquier F. 1984. Prospection et captage Suisse. 219 p. (Storm Water Model Management, Rossman 2004) to the related storage within the aquifer. recharges will be assessed and extrapolated to estimate the maximum discharges which could be expected within the conduits. These simulations are expected to produce relevant results that will improve the modeling of the karst system. They will also bring quantitative elements to design future construction and hydraulic works. Conclusion geological model depicting the aquifer basement has data and literature documentation. By following the hydraulic principles in karst hydrology, it was possible to sketch the vadose and the phreatic zones as well as aiming distinguishing their recharge contribution at the next stages of the study. By using the available discharge data of the Beuchire spring and head measurements of a thresholds functioning of the system and to approach the geometry of the groundwater hydraulic within the conduits in the high water stage and to delineate where water is susceptible to reach the ground surface and to enhance the risk of inundation. Areas on ground occurrence of the mapped considered events (average event). Furthermore successive activation of the outlets and their associated discharges are now predictable. Recent integration of piezometric data from a borehole in the vicinity of the main temporary outlet (Creugenat) and the more recent instrumentation and observations in the collapse at Courtedoux (which appeared in June 2012) brought new indicators to control and improve the established model. Current simulations using SWMM and based on these new data may provide new elements 341

PAGE 359

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE l'effondrement de Courtedoux. Institut Suisse de Jeannin PY, Eichenberger U, Sinreich M, Vouillamoz J, Malard A et al. 2012. KARSYS: a pragmatic approach to karst hydrogeological system conceptualisation. Assessment of groundwater reserves and resources in Switzerland. Environmental Earth Sciences. Available from: karst aquifers: A hydrodynamic modeling approach circulation souterraine en terrains calcaires. Actes de Prudhomme LM. 1804. Dictionnaire universel, Rossman LA. 2004. Storm water management model. user's manual version 5.0. United States Environmental Protection Agency. foreland fold and thrust belt [PhD dissertation]. 342

PAGE 360

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2hypotheses relating a much broader suite of calibration data, including water levels, water budgets, and spring discharges. Introduction in the world, providing more than 950 million liters of water to more than 2 million people on an average day. In addition, this aquifer is home to more than 40 aquatic subterranean species, several of which are endangered, agricultural, industrial, recreational, and domestic water with some depths extending up to 1,000 m. The north south extent of the aquifer ranges between 10 to 60 kilometers (Figure 1). Recharge to the Edwards aquifer occurs from the capture of surface water originating from the contributing zone (allogenic recharge), direct precipitation on the recharge zone (autogenic recharge), both above and below the Edwards Limestone. Discharge Regionally, the structure of the aquifer is exceedingly complex, owing to the extensive faulting associated with the Balcones Fault Zone. The faulting in the Balcones Fault Zone is primarily en echelon normal faulting that AbstractThe Balcones Fault Zone Edwards aquifer (Edwards United States, with an average withdrawal of 950 million between the Uvalde pool and the San Antonio pool of the Edwards aquifer, west of the San Antonio metropolitan area in Uvalde County, Texas. This area is known as the The San Antonio pool is the primary source of water for the greater San Antonio water supply. The Knippa width estimated to be approximately 4 km, is bounded by northeast trending faults of the Balcones Fault Zone on the north, and uplift from the Uvalde salient for numerous years, yet the exact location, nature of improve the aquifer water balance. This study integrates recent research by other scientists 2012 as part of an M.S. thesis. This paper is limited to a discussion of the water quality as it relates to quality data constrain a revised conceptual model which are being tested in the more comprehensive M.S. thesis. Although current interpretations are tentative, it appears this conceptual model will be readily convertible into a digital model that can test CONCEPTUALIZATION OF GROUNDWATER FLOW IN THE EDWARDS AQUIFER THROUGH THE Jennifer Adkins5695 W. Michael Cole Dr., Fayetteville, Arkansas, 72704, USA, jadkins@uark.edu343

PAGE 361

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE aerial exposure, burial (middle Cretaceous), faulting In the catchment area of the aquifer (Figure 1), dominant karst processes are epigenic. This means dissolution is produced primarily by descending recharge and horizontal groundwater movement. morphological forms such as vertical shafts, scallops, thought to delineate the northwestern boundary of the activity occurred, which includes igneous intrusions and uplift. This event bowed the overlying sediments, much shallower depths (Mosher et al., 2006), and resulted in the structural feature known as the Uvalde salient of the Devils River Trend. This feature dips into the Maverick Basin toward the southwest (Figure 2). Lithologically, the Edwards aquifer in the area of the San Antonio pool comprises as many as 8 members carbonates and evaporates that were deposited in the latter part of the Early Cretaceous period (Clark, 344Figure 1. Location of the major hydrogeologic zones of the Edwards aquifer in south-central Texas, by county. The study area of this project is the Knippa Gap in Uvalde County, shown in purple between the towns of Uvalde and Sabinal on the map [Modified from Edwards Aquifer Authority webpage].

PAGE 362

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2Problem StatementThe Edwards aquifer has been intensively studied, but many important questions remain unanswered. One (Figure 4). This part of the Edwards aquifer represents northeast trending faults of the Balcones Fault Zone. Water discharges from the Uvalde Pool on the west into the San Antonio Pool in the east. Southeast of the Knippa and minimal well yields in this part of the aquifer and cupulas, many researchers conclude that hypogenic an important role in the karst development of the systems in the Edwards aquifer resurge as large springs where groundwater is returned to the surface from depth. Permeability derived by this upward water hydrocarbon storage within the rock unit (Schindel et al., 2008). into three regional zones, the recharge zone, the contributing zone, and the artesian zone (Figure 1). The aquifer to occur. This zone is also where contamination of the aquifer is most likely to occur, primarily as a little to no soil cover. The recharge zone is dominated by vertical faulting of the Balcones Fault Zone, and is way to the artesian zone. Entryways are predominantly point and line sources where streams and rivers cut across this zone of faulting. The artesian zone is the southern and easternmost part of the aquifer where 345Figure 2. Location of the Devils River Trend, Maverick Basin, Uvalde Salient and the San Marcos Platform [Adapted from Green et al., 2006]. Figure 3. Stratigraphic column from southwest to northeast across Uvalde County in the study area, Knippa Gap. Karstified focused flow zones of the Edwards are shown in blue [Modified from Green, 2009].

PAGE 363

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEgravels of the Leona River, and the Trinity aquifers are features have been studied throughout Uvalde County, such as the Uvalde salient (resulting from crustal uplift, faulting, and igneous activity that elevates the Edwards aquifer to the surface across the central region of the county), and the Balcones fault zone (a tensional structure area aligned southwest to northeast across the study area). Preliminary interpretation of that acts as a barrier, separating the Uvalde pool from the San Antonio pool under Medina, Bexar, and Comal Counties. It is described as being a narrow opening in an extensive system of barrier faults. (McClay and Land, 1988) Although 2.4 x 1011 annually, the constriction causes water levels to build up conclude that the Uvalde salient has several prominent through topographic saddles between the high points. They also note the large amounts of recharge from the Frio and Dry Frio Rivers that are contributing to the cities of San Antonio, New Braunfels, and San Marcos. Recharge of the aquifer is greatly impacted by periodic Methods and ApproachThe hydrogeology and eleven wells described herein boundaries in the conceptual model. The conceptual model (Figure 7) incorporates samples compositions of these samples. These data allow visualization of geochemically related waters, and the an understanding of the geochemical processes evolution of water type in the aquifer. These should the south through subcrops to the Leona gravels. More the water budget for the San Antonio Pool and more the Uvalde pool.Objective and Scope constriction in Uvalde County, Texas, known as the related to water quality; however, this paper is only a small part of a much broader M.S. study of the karst hydrogeology of the region. Study AreaThe study area is shown in the shaded region of Figure 4. An expanded but secondary area of interest surrounds the main study area, encompassing contiguous portions Edwards aquifer in Uvalde County is predominantly composed of Lower Cretaceous carbonate (dolomitic limestone) of the Devils River Formation within the Devils River trend in the northeast, transitioning into the West Nueces, McKnight, and Salmon Peak Formations in the Maverick Basin in the southwest. These carbonate rocks were formed in evolving environments that ranged across a variety of tectonic and depositional conditions. The Devils River Trend current energy, whereas the West Nueces, McKnight, and Salmon Peak Formations were restricted to open The upper units of the Devils River Trend along with the upper unit of the Salmon Peak Formation are the most Throughout the study area there are numerous Upper Cretaceous or Lower Tertiary igneous rocks that intrude through the stratigraphic units composing the Edwards minor groundwater resources from a thick sequence of sedimentary rocks. The Edwards is by far the most of the county from west to east. The Buda, Austin Chalk, 346

PAGE 364

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 m temperatures (26.6 to 24.7 oC) that occur in wells within the Uvalde salient (QW Sites 2, 5, 7). Waters west (QW oC). QW Site 8 represents the least mineralized of all wells sampled, not of the lowest concentrations of dissolved chloride and dissolved sulfate. Various degrees of mixing of waters from different sources are present in these latter wells, ConclusionsThe conceptual model (Figure 7) allows visualization not be used alone to delineate the gap, but they are a good conceptual start to test alternative hypotheses. Considering the complex faulting in the immediate area, they are consistent with a structural basis for ResultsTable 1 shows the water quality and dissolved constituents in water from wells located within the study area. The Well ID in Table1 is referenced to Figure 5, and the QW Sites to Figure 7. Figure 5 includes 2 sample sites (QW range of 5% error. Table1 and Figures 6 and 7 indicate the presence of high sulfate and high chloride waters with higher 347 QW Site Well ID Date TDS Temp oC Cond pH Ca+2Mg+2Na+K+ALK Cl-SO4 -21 606 481 7.20 82.1 10.4 10.8 1.07 19.9 2 1210 26.6 1605 6.98 277 27 28.1 2.91 169 72.9 260 471 7.25 79.9 0.974 188 20.6 11 4 477 7.49 86.8 9.28 11.1 1.1 20.2 11.7 5 877 24.7 1274 7.24 168 21.9 77.8 5.62 241 158 196 6 24 502 7.27 11.7 206 12 7 24.7 701 7.16 17.8 25.6 200 55.5 55.9 8 25.1 428 6.9 0.097 179 14.1 11.5 9 6942606 502 7.21 8.22 11.7 199 10.6 10 601 88.8 9.19 0.962 212 42.8 18 11 448 7.19 9.57 24 1.09 215 51.1 19.2Table 1. Selected water quality and dissolved constituents in water from wells in the study area. QW Site number is referenced to Figure 4. Chemical parameters are in mg/L. [QW, water quality; TDS, total dissolved solids, in mg/L; Cond, specific conductance, in S/cm]

PAGE 365

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE348Figure 5. Geology of the Edwards aquifer in the study area, including areal geology, faulting associated with the Balcones fault zone (red lines), exposures of igneous intrusives associated with the Devils River Trend of the Uvalde salient (in red), and sampling sites of wells used to measure water levels and collect groundwater samples. The numbers refer to the sampled wells discussed in Table 1. [Map modified from multiple sources, including Clark, 2003; Green, 2006, and personal communications with Vic Hilderbran, Uvalde County Water Conservation District and Rob Esquilin, Edwards Aquifer Authority].Figure 4. Location of key components of the Knippa Gap, the expanded study area, and other relevant hydrogeologic features in Uvalde County [Modified from Green, 2009]. NCKRI SYMPOSIUM 2 Figure 4. Location of key components of the Knippa Gap, the expanded study area, and other relevant hydrogeologic features in Uvalde County [Modified from Green, 2009]. Figure 5.Geology of the Edwards aquifer in the study area, including areal geology, faulting associated with the Balcones fault zone (red lines), exposures of igneous intrusives associat ed with the Devils River Trend of the Uvalde salient (in red), and sampling sites of wells used to measure water levels and collect groundwater samples. The numbers refer to the sampled wells discussed in Table 1. [Map modified from multiple sources, inc luding Clark, 2003; Green, 2006, and personal communications with Vic Hilderbran, Uvalde County Water Conservation District and Rob Esquilin, Edwards Aquifer Authority]. Expanded area of this study including Uvalde Pool on the west, and San Antonio Pool on the east

PAGE 366

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2secondary permeability than the freshwater portions of the aquifer. These factors result in greater salinity levels and distinctive Stiff diagrams. The conceptual model (Figure 7) shows that (QW) sites 5 and 7 plot along a mixing line of meteoric water and down gradient water similar in chemical composition to well 2. As indicated by the curved blue lines on the model, these QW sites have mixing components that are inconsistent with Most of the wells with these attributes overlie the Uvalde salient, and because of the structural uplift, the aquifers proximity may contribute to slightly higher temperatures although this needs to be investigated further. Well yields dissolution of the highly soluble evaporates) through this part of the aquifer. Well 11 is an exception to this, but inasmuch as it lies on the boundary of this study and its explanation at this point is not obvious. Data from the remaining QW sites have Stiff diagrams as the main geochemical process. These QW sites diagram (Figure 6) as well, and have calculated TDS Future WorkIn addition to the geochemical analysis discussed in this paper, the larger M.S. study will incorporate the compilation of a complete table of wells, geophysical aquifer tests within the study area. The completed table of wells represents sites with multiple names and aliases, and will aid in future investigations in accessible digital format. The table will involve historic published well data, and unpublished records from drillers, water managers, and hydrogeologists in of wells which will be conducted during the summer superimposed on Figure 5, (which includes the Balcones Fault Zone and outcrops of intrusive igneous rocks that Stiff (Figure 7) diagrams from sites designated as Knippa m oC. QW Site 8, the least mineralized well sampled, is the only exception to the temperature range listed, with a value of 25 o from the Uvalde Pool to the Leona gravels. Stiff diagrams for QW Site 2 is thought to lie near the bad boundary the freshwater portions of the Edwards aquifer. Increased mineralization is a result in increased contact with gypsum and has more limited development of 349Figure 6. Piper diagram of groundwater in the study area showing water quality types ranging from waters within the Knippa Gap (within black circle) to waters derived from mixing of high sulfate and chloride waters associated with residual evaporites in less dynamic flow zones (see wells 2, 5, and 7 in Table 1).

PAGE 367

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE study area. with the Uvalde salient (star on Figure 5). Tracer testing is one of the most effective ways of quantifying groundwater movement in karst aquifers, and will provide empirical data that will aid in the determination of the groundwater components for this region (Schindel et al., 2008). A synoptic potentiometric map of the study area conditions during the summer of 2012. This effort will the Edwards Aquifer Authority (EAA), and the results will be used to evaluate potential boundaries, assess variability of aquifer hydraulic properties, and indicate A hydrostratigraphic analysis, incorporating a conceptual boundaries), aid in determining physical constraints 350Figure 7. Conceptual model of the Edwards aquifer in the study area, showing Stiff diagrams that reflect major element concentrations dissolved in groundwater (in green), approximate locations of boundaries of flow through the Knippa Gap (curved blue lines), major flow directions through the Knippa Gap constriction (blue arrows), subsurface overflow from the Uvalde Pool to the Leona gravels (black arrow), and exposures of igneous intrusives associated with the Devils River Trend of the Uvalde salient (in red). Sampling sites of wells for which chemical analyses are reported are shown by black dots; the numbers refer to the sampled wells discussed in Table1.

PAGE 368

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 Klar R.V., Morris, A.P., 2006, Evaluation of the Edwards Aquifer in Kinney and Uvalde Counties, Texas. Southwest Research Institute. Conducted for the Edwards Aquifer Authority. Analysis of the Water Resources of the Area Centered Within and Near Uvalde and Zavala Counties and Review of a Report on a Potential Well Field and Pipeline in Uvalde County, Texas Final Report Prepared for City of Uvalde and Wintergarden Investigating the Secondary Aquifers of the Uvalde County Underground water Conservation District Division Southwest Research Institute. Seco Creek and Medina River where They Overlie Engineering Division Southwest Research Institute. Data Report for 2010. geologic controls on porosity development in platform carbonates, South Texas: The University karst, fractures, and permeability in the Cretaceous Edwards aquifer: analogs for fractured carbonate reservoirs: Society of Petroleum Engineers Annual Permeability structure of the Edwards Aquifer, south Texasimplications for aquifer management: The University of Texas at Austin, Bureau of Economic Acknowledgments general manager for the Uvalde County Underground Water Conservation District, for all the help he gave his knowledge about Uvalde County groundwater was outstanding, and he shared his knowledge with me freely. I am truly grateful to him for his assistance. of the Southwest Research Institute for providing me with an internship, for guidance, and for encouragement throughout the summer. Their insight was invaluable. Edwards Aquifer Authority, not only provided me with an internship and invaluable geochemical and hydrogeologic insight, but he made resources and personnel available to assist in this endeavor. The staff of the Edwards Aquifer Authority, especially Rob Esquilin, incredibly generous in sharing their knowledge and data from the Edwards aquifer. Finally, I would like to thank all the well owners in Uvalde County who allowed me access to their property and facilitated sampling and measuring hydrogeologic parameters of their wells. I know their water is precious to them, and I thank them for trusting me and facilitating my study.References hydrogeologic characteristics of the Edwards Survey, Water Resources Investigations Report J., 2010, Late Triassic Texas uplift preceding Society of America Ferrill, David A., Sims, Darrell W., Franklin Nathan, Morris, Alan P., Waiting, Deborah J.,2004, Science, University of Texas at San Antonio.351

PAGE 369

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE New York. Young, S.C., Doherty, J., Budge, R., Deeds, N., 2009, Application of PEST to recalibrate the groundwater (Plateau) and Pecos Valley Aquifers DRAFT. Contract report for the Texas Water Development Board by URS Corporation, Watermark Numerical Pty, Ltd, and Intera Inc. Flow in the Edwards AquiferCharacterizing the Role of Fractures and Conduits in the Balcones Fault Zone Segment: Final Contract, Bureau of Special Paper no. 1, National Cave and Karst Research Institute, Carlsbad, NM, 106 pp. in the Edwards aquifer, San Antonio region, Texas, Painter, Scott., Jiang, Yefang, Woodbury, Allan., 2002, report: Southwest Research Institute, variably paginated. Palmer, A. N., 1991, Origin and morphology of limestone subsurface, central Texas. Bureau of Economic Alexander, E. C., Jr., Alexander, changes during a recharge event in the karstic Society of America Abstracts with Program. Available from: to explain squifer dynamics: Adapted from oral Schultz, A. L., 1994, Review and update of the position interface from Uvalde to Kyle, Texas: Edwards Trace Results. Northwestern Arkansas: unpublished M.S. thesis, University of Arkansas, Fayetteville. and Palmer, M.V., eds., Caves and Karst of the USA: 352

PAGE 370

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 hydrostratigraphic units and minimum hydraulic gradients of 12% across higher permeable units. The updip outcrop of the top of a perching lower permeable unit, as well as caves that breach the lower permeable catchment area. Through deeper investigation of the caves using various methods together, the mapped constrain the boundaries, the subsurface catchment area should always be conservatively overestimated.IntroductionRare cave species in Travis County, Texas, are protected by a federal permit awarded to the City of Austin and suitable habitat for listed endangered karst species and ensure that other rare species of concern would not require federal listing as endangered in the future. sustenance, so hydrogeological studies are conducted to identify water catchments that provide direct runoff to the cave entrance (surface catchment area) as well as the overlying area that supports cave drips and cave streams (subsurface catchment area) through subsurface is considerable more direct and is described elsewhere of subsurface catchment areas, except where surface that supply the studied drips. Basic methodology for delineating source areas are and Drew (2007). For delineating subsurface catchment areas for cave drips and cave streams of the Edwards Aquifer, it is important to have an understanding of AbstractDelineating the source area of cave drips and streams cave ecosystems. In order to focus protection for cave ecosystems, particularly those containing federally listed species, it is necessary to accurately delineate the potential Escarpment sites (Buttercup Creek, Barker Ranch #1 Cave, McNeil Drive, and Davis Lane). The methods drips, speleothems, pools, and streams under wet and dry conditions to characterize drips as discrete or seepage, 2) cave mapping surveys to determine spatial relations and hydrostratigraphic characterization (dip of beds, faulting, and the rock tendency to perch vadose groundwater quality characterization and comparison with potential sources and 5) dye and chemical tracing. Steps 4 and 5 provide the most direct delineation of source areas at all four study sites and some catchment areas are so large that they were not completely delineated without additional investigation. Mapping the highest elevation of a drip source in a cave Direct tracing of vadose groundwater illuminates hydrostratigraphic units in perching groundwater and DELINEATING SOURCE AREAS TO CAVE DRIPS AND CAVE Nico HauwertCity of Austin Watershed Protection Dept., PO Box 1088, Austin, Texas 78767, nico.hauwert@austintexas.govBrian CowanZara Environmental LLC, 1707 FM1626, Manchaca, Texas 78652, 353

PAGE 371

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE29% recharged the aquifer from upland slopes. Using gauging station data distinguished by traced groundwater as evapotranspiration, 22% recharged over the Edwards creeks and ran off downstream of the recharge zone with permeability contrast and aquiclude thickness, as Veni (1992) summarized the effects of geology on cave development described by White (1988) and Ford and Williams (1989) including: vertical cave shafts generally develop above the water table and are associated with beds of lower permeability or lower solubility; horizontal cave passages develop in high permeability beds; caves typically become impassible at common lower contrast and their discharge is proportional to the size of its catchment area. Rose (1972), Maclay and Small (1986), and Small et al. (1996) described the general characteristics of the hydrostratigraphic members of the greater detail how the hydrostratigraphic properties of the Edwards Aquifer. and Palmer, 2002; Veni, 1992). Downdropped faults portion of Edwards Aquifer that simulates the effects of dip nearly parallel to scissor fault directions within ramp structures (Collins, 1995). MethodologyThe subsurface catchment area for a cave drip or simply by mapping the surface extent of any connected recharge and storage through the soils and epikarst. In the 1980s, studies were conducted that examined recharge within the Barton Springs Segment, how groundwater moves through the aquifer. Using a water budget approach based on stream gauging and rainfall measurements. Assuming that all recharge in Barton Creek channel of the recharge zone discharges from Barton Springs, it was calculated that 85% of rainfall was lost to evapotransporation, 5% of rainfall the aquifer within the intervening areas between the 1980s water balance was invalidated in 1996, when direct groundwater tracing and water level mapping revealed that the entire portion of Barton Creek where not actually within the groundwater basin contributing That error alone comprised 28% of total recharge to Barton Springs in the water budget, and explains why the Edwards Aquifer was erroneously attributed a recharge value similar to those measured over the studies commonly fail when based on an incomplete understanding of groundwater source areas, as well transport generally has high dispersion and attenuation Karst aquifers typically show recharge values of 20 to measurements from Central Texas used climate towers in Uvalde County discovered that of measured rainfall, 65% was lost to evapotranspiration, 5% to runoff, measurement of recharge since they quantify roughly 70% of the rainfall budget as opposed to roughly 5 to Barton Springs Segment used an eddy covariance tower, drain of an internal drainage basin, and the remaining 354

PAGE 372

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2Cave Surveys The depth and lateral extent of the cave constrains the subsurface catchment area. The lateral extent of the cave beneath the ground is known as the cave footprint. Without the completion of additional study beyond mapping the cave, the lowest elevation of the cave can be used to eliminate areas of lower elevation as being outside the subsurface catchment area. The cave is mapped from an entrance survey point using and inclination as described by Dasher (1994), Jeannin measured using a nylon tape or Bosch laser survey. Suunto tandem compass. Magnetic declination was set land surface elevation than the cave. This approach is appealing because it requires very little data from the cave other than the total cave depth and surface topography, and includes the entire actual source area to for long distances from the cave, greatly overestimating the actual subsurface catchment area. The purpose of deeper hydrogeological study of the cave is to allow the catchment area, but include all of the source area for The methods utilized in studies to map subsurface catchment areas involve a combination of cave surveys, and introduced tracers. Note that in each of the study areas, emphasis is placed on some of the methods that 355Table 1. Hydrostratigraphic units of the study areas. Modified from Small et al. (1996) and Hauwert (2009). While the Basal Nodular Member appears equivalent to the Comanche Peak and Walnut Formations of North Austin, it has not been formally correlated.

PAGE 373

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE interval, and later quantifying the drip rate of the entire waterproof cave pack of measured volume by capturing Speleothem types may be used to characterize sources. and bacon rinds may be associated with discrete drips. Soda straws are typically associated with seepage drips. pores, although it is possible that a discrete source is transmitted through a porous media, such as a pulverulite. When accessing caves for studies it is recommended and possibly required by local permitting to have trained cave specialists and cave biologists. A cave specialist can ensure the cave is entered safely and determine where specialized techniques such as negotiating vertical techniques or tight crawls are required. Cave biologists frequently accompanied trips into caves during this study to minimize impacts to the cave ecosystem. Hydrostratigraphic Mapping and CharacterizationThe general properties of the rocks and the geological framework, such as rock dip and fracturing of the rocks, can be used to understand the basis for groundwater surface and subsurface geology is an important step for delineating subsurface catchments. Hydraulic gradient One criterion for delineating subsurface catchment areas is based on the properties of the rocks between the surface and cave drip. Using tracing and direct observation of cave passages through various hydrostratigraphic in terms of a minimum vertical hydraulic gradient. This criterion can be applied only to sites where the hydrostratigraphic units are accurately mapped on the surface and subsurface. In highly permeable and soluble rocks, water will tend to descend relatively steeply even where fractures, with a Tandem inclinometer and Brunton compass inclinometer. Where possible, forward and back shots were taken between each station and any discrepancies were resolved through repeat measurement. Existing cave maps provide valuable information but Drips in the caves were rarely mapped on existing cave maps for the study area. Most caves referenced in this report were resurveyed even though existing maps were available. The difference between two cave surveys was used as a fair indication of where the cave footprint lies. Cave radio location was used on longer caves to locate the surface position and depth of several stations within the cave. The elevation of the cave drip is derived from criteria, the subsurface catchment area should extend at least100 m beyond the cave footprint.Characterization of drips and cave streamsIf the highest point of origin of a cave drip or cave stream can be established, then that drip horizon, rather than the bottom of the cave, can be used to delineate a source area. Note that drips within the cave may originate from different sources unless associated by physically summary of cave drip characterization is provided by Jeannin et al (2007). Cave drips are characterized as discrete or seepage. A one or a few locations. A seepage drip has discharge distributed from many formations across a ceiling. The persistence of a drip or cave stream is characterized at various times under wet and dry conditions, particularly during or shortly after an intense storm where the soils are saturated. Methods used to quantify cave drip rates include using a graduated plastic cylinder to measure the drip volume over a measured time interval. Plastic Rainwise tipping buckets with Onset Microstation data loggers are used on some drips to measure drip rates continuously, allowing changes in drip rate to be correlated with rain event cycles or anthropogenic sources such as swimming pool draining or utility line leak. In a cave room with widely dispersed cave drips, one cave drip rate volume is measured to estimate drip volume per drip, and other drips in the room are 356

PAGE 374

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2size of the subsurface catchment area, particularly in the updip direction of a cave drip, where a persistent discrete drip suggests a larger source area, and obviously where Geologic Framework: Stratigraphic Dip and FaultingIn order to examine geological controls of vadose including stratigraphic dip and faults. In areas between mapped faults, the location and elevation of distinctive marker beds are located using global positioning systems and maximum dip magnitude are calculated (Compton, 1962). The stratigraphic dip directly measured at the site scale, and the measured local dips rarely coincided with regional dip reported in the literature. Single outcrop vary more with local collapse and generally were not representative of overall dip within fault blocks. The subsurface catchment area typically extends in the stratigraphic updip direction from the cave discharge. Despite the general rule of downdip vadose spring site, sometimes perpendicular to down dip and generally poorly exposed in the Austin area and are most commonly mapped using abrupt change in surface hydrostratigraphic units. Detailed site geology mapping is generally necessary to distinguish elevation declines in marker beds due to faulting, stratigraphic dips or other geologic structures. If a lower permeable hydrostratigraphic unit is discovered extent of subsurface source area, even where minimum catchment areas. Other structures that may limit the extent of the subsurface catchment area are caves that breach lower permeable perching units or descend below the drip horizon in the studied cave. These are known as breach structures in dissolution is strongly enhanced along fractures, Within the highest permeability hydrostratigraphic units, such as the Leached and Collapsed Members, Kirschberg, extend horizontally through these units, the passages were formed in the phreatic zone, and under unsaturated 2009). Note that where cave passages in overlying permeable units overlie relatively lower permeability beds, such as cave passages within the Leached and Collapsed Member over the Regional Dense Member, the hydraulic gradient. Lower permeability units within Dense Member, and the Dolomitic Member tend to perch groundwater for some distance until breached by shaft. So the lower permeability units have both very low and The property of hydrostratigraphic units to perch gradient that is the distance that a tracer travels divided by vertical depth above or through that unit. Based on the mapped hydrostratigraphic units between the surface and cave discharge, the mapped subsurface catchment area should extend at least beyond the minimum hydraulic gradient measured for those rock units unless other criteria exist, such as direct tracing used to indicate a higher hydraulic gradient and smaller source area. All units potentially have a high hydraulic gradient drip to the surface provides a criterion to limit the lateral extent of potential source area. members through which the groundwater travels. For the sites traced, the hydrostratigraphic units are mapped across the surface, in caves, in logged wells, and from and entire cave depth. It is possible that a lower gradient exists across a hydrostratigraphic unit than we tested, so this criteria should be used with caution to limit the 357

PAGE 375

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE disadvantage of chemical tracing is that if the sampling intervals selected are too wide, a short breakthrough pulse might be missed and a composite sample may dilute the pulse to the extent that it is undistinguishable. A frequent sampling cycle can also be expensive and alkalinity, and iron are conducted by the Lower Colorado River Authority lab in Austin, Texas. All tracers in this study into cave streams. Where tracers targeted cave species preserves, biological surveys of the caves were conducted by permitted cave biologists to verify that the amount and type of tracer were not visibly affecting the cave ecosystem. The considered so as not to create a nuisance or health hazard. At the concentration of tracers in the phreatic zone, the tracers were relatively benign, especially compared to actual contamination sources that have affected or could potentially affect these water supplies. The risk of potential impacts to the species or water supply users can be considered in light of potential impact sources that may be much worse than the tracers used. The information gleaned by direct tracing can help focus disperse and resource intensive efforts over a large area that may offer limited protection. Study Sites Four study sites include Buttercup Creek, Barker Ranch, Buttercup Creek Study, Northern Segment streams of Marigold Cave and Whitewater Cave by Mike Warton & Associates of Cedar Park, Texas. This study was funded by Lumbermans Investment Corporation and the report submitted to US Fish and Wildlife Service charcoal receptors and grab samples. The caves included horizontal passages and shafts through the Comanche Peak Formation and Walnut Formations that underlie the Edwards Formation in the Northern Segment of the Edwards Aquifer. This study differs from the other three this study. In cases where narrow topographic saddles were connected by continuous higher elevation to the would follow a ridge or perhaps take erratic turns rather Water-Quality CharacterizationWater quality similarities help associate drips within the same cave or characterize sources to those cave quality characterization is necessary where tracing cannot be conducted to associate drips with a source discharges. The source water sampling parameters magnesium, potassium, sodium, sulfate, bromide; trace metals: aluminum, arsenic, boron, cadmium, chromium, copper, iron, lead, nickel, strontium, zinc; nutients: nitrate + nitrite, ammonia, phosphorus; and total suspended most parameters. Bacteria samples are collected using both grab and autosamplers. When collected using autosamplers, blank bottles are tested for total coliform and E. coli to test for bottle contamination. Dye and Chemical Tracing sulforhodamine b, and phloxine b is nearly continuously being conducted within the Barton Springs Segment, those tracers could not be used in our short vadose tracing. Organic tracers are also notoriously sorbed by organic debris and sediment and are most effective when tracers is that they can be monitored continuously using charcoal receptors for dyes and cotton receptors for optical brightners. For soil tracing we frequently use the optical brightners tinopal and direct yellow 96, as well as the dye pyranine, even though they are not ideally suited for soil tracing, and frequently not recovered alongside for dyes and optical brighteners was conducted by Ozark Underground Laboratory in Protem, Missouri. Chemical tracers commonly used include potassium 2 CO), 358

PAGE 376

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2brightener direct yellow 96 (DY96). Even though the drips in Barker Ranch #1 are relatively persistent for months after rain, the onset of drought following the third round of simultaneous tracing eventually led to the drips in Barker Ranch #1 drying for several years, resulting in 2009 Study along McNeil Drive Several caves near McNeil Drive contain listed endangered species, including McNeil Bat Cave, Weldon Cave, No surface mapping of the area, examining local quarries and drilling bores and geotechnical borings to gain 2010). Three of the four caves were remapped relative to professionally surveyed surface monuments at the cave entrance. The drips were observed under varying climatic 2012 Study along Davis Lane Austin and Brian Cowan and support staff from Zara Environmental in 2012, funded in part by City of Austin Public Works Department and by a spill simulation Protection Department. The study involved surface and subsurface geological mapping, cave and drip mapping, tracers (potassium bromide, ammonium carbonate, and iron) were used, along with optical brightners tinopal, direct yellow 96, and pyranine. A cave radio survey of Maple Run Cave was conducted.ResultsIn the three cave drip studies, the cave drips were found throughout the cave. Both seepage and discrete drips were encountered. The localization of drips within the through small pores within the entire rock column. 1997 Buttercup Creek and discharged from Blizzard Springs, which discharged on the west (opposite) side of Cypress Creek. Blizzard Barker Ranch #1 Cave Barker Ranch #1 Cave is a relatively shallow upland cave to the entrance has multiple drips that are persistent (Figure 1). tracing study funded by the City of Austin Watershed Protection Department in 2007 involved pouring tracers at six surface locations across the site and monitoring Barker Ranch #1 Cave drips for any breakthrough. with 15 liters of water solvent for powder tracers (Cowan et al., 2007). The chemical tracers used were potassium 2 CO), iron standard (Fe), sodium chloride (NaCl), and potassium iodide (KI). Three traces were repeated to verify results or replace a failed trace. Supplemental organic tracers were used alongside chemical tracers, including dyes sulforhodamine b (SRB) and pyranine, as well as optical 359Figure 1. Location of four study sites, Austin, Texas, USA.

PAGE 377

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE 2007, pyranine was measured in a charcoal receptor a reasonable hit since that well has had periodic tracer hits from upgradient traces. The well is mapped to be placed in the well since March 12, 2007, did not detect the tracer. Based on the tracing results, local breach structures, drip characterization, and drip elevation, an area encompassing the subsurface catchment area was mapped (Figure 2). This area could potentially be further constrained by additional traces.2009 McNeil Drive StudySurface geology mapping, subsurface mapping in caves, two borings, and one core allowed subsurface mapping of measured to be northwest and no faults were mapped between the caves (Figure 4). The subsurface catchment area was mapped based on the cave footprint, a minimum hydraulic gradient of 10% from the mapped cave drips. The presence of one persistent and relatively deep discrete drip in the Rhadine Room of McNeil Bat Cave necessitated including a possible subsurface catchment area to the outcrop of overlying rock in the updip area as far as 600 m (2,000 ft) to the southeast. In this case, the application of area to a smaller area. The subsurface catchment areas be adequately delineated to a reasonable area based on outcrop of overlying rock in the updip area and hydraulic gradient from the mapped cave drips. Three borings encountered groundwater within both the Dolomitic Member and Walnut Formation, beneath the elevation of the studied caves. Although the study was beyond the scope of the study, the elevation of groundwater within the Dolomitic Member and presence of the study site suggest that the perched groundwater the measured updip direction (Figure 4). This evaluation Springs discharges from the base of a bluff, about 5 m (15 ft) below the contact of the Walnut Formation in is generally less soluble than the overlying Edwards Formation, but hosts extensive cave development along not included in this study, mapping elevation changes several kilometers south have shown a relatively consistent eastward decline in the contact of the Edwards within cave streams that likely have large contributing source areas, only a portion of the subsurface catchment 2007 Barker Ranch #1 Cave to drips in Barker Ranch #1 Cave (Figure 2). The initial a later breakthrough of potassium following a second below the elevation of drips within Barker Ranch #1 Cave, and are expected to serve as breach structures solution cavities where runoff naturally localizes and a soil site devoid of obvious macropores or depressions. The drip horizon outcrops within 120 m (400 ft) to the east, north, and south of Barker Ranch #1 Cave. Two chemical tracers tested on Barker Ranch #1, potassium iodide and sodium chloride, were found unsuitable for tracing here and were not reapplied. Background concentrations of chloride in the drips were too high to be able to distinguish breakthrough to locate a local laboratory to analyze for iodide. The organic tracers were not detected in Barker Ranch associated chemical tracers were not detected above background concentrations, indicating that those area to Barker Ranch #1. In the remaining two cases 360

PAGE 378

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2361 Figure 2. Tracer injection sites associated with Barker Ranch #1 Cave drip study. Note the subsurface catchment area interpretation is constrained by the surface elevation corresponding to the cave drips, by 6 vadose trace sites, nearby caves that serve as breach structures, and 10% minimum hydraulic gradient for the Grainstone and Kirschberg Members overlying the drips. Figure 3. Concentration breakthrough in Barker Ranch #1 Cave drip of potassium bromide (KBr) tracer injected in a small soil-filled depression (site 1, Figure 3) about 30 m from the entrance. The injection was flushed by a natural rain event. Bromide peaked three to seven hours after injection. Sampling resumed for a second event after a nine-day pause and sampling detected a late potassium pulse breakthrough.

PAGE 379

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE362 Figure 4. Geologic cross section of McNeil Drive Site, through No Rent Cave. Since the persistent drips could potentially be supplied in part by a water-quality pond that captures storm-water runoff, the subsurface catchment area was extended to include the entire contributing surface catchment area for the pond. Figure 5. Geologic cross section of Davis Lane and Subsurface Catchment Area to Balcony Room Drip of Blowing Sink Cave.

PAGE 380

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2363Hydraulic GradientBased on the vadose traces included in this study, minimum hydraulic gradients for various hydrostratigraphic units of the Edwards Aquifer can be characterized. Table 2 below shows the measured results through the tested units. Based on the testing thus far, minimum hydraulic Members, and lower minimum hydraulic gradients of 0.4% were measured across the Walnut Formation. We use 10% in place of 12% for high permeable units both for ease of calculation and because it is more conservative. Although the Dolomitic Member has not been tested yet, in Flint Ridge Cave over the rhythmic beds of the Dolomitic Member, and overall perching of groundwater observed in Midnight Cave, Blowing Sink, Backdoor Springs, and Bee Springs, suggest that this unit can have a potentially caves developed within the Dolomitic Member For low permeable units, the source can potentially Buttercup Creek to Blizzard Springs. ConclusionThe source area to cave drips and cave streams can be delineated using a combination of drip characterization, cave drip elevation and spatial mapping, geologic and direct groundwater tracing. Drip characterization study. It is possible the groundwater encountered in the borings is not hydraulically connected, but may suggest exceptions to the general rule of downdip perched 2012 Davis Lane StudyThe location and characterization of cave drips in initially accomplished by observing the inside of the in the study area near Davis Lane was underlain by the permeable Leached and Collapsed Members. A shallow depth below the surface, the Regional Dense Member (RDM) was present (Figure 5). The RDM was observed in Maple Run Cave and at a lower elevation tends to locally perch groundwater, a number of local caves were mapped to descend through and below the RDM, effectively acting as drains to allow the perched groundwater to descend toward the phreatic zone. Seven surface locations were selected for tracing. caves were both detected in the phreatic cave stream vadose Balcony Drip of Blowing Sink Cave three that groundwater perched over the RDM, descending through a breach in the RDM at Blowing Sink Cave.Table 2. Measured hydraulic gradients across various hydrostratigraphic units associated with the Edwards Aquifer. Year Trace Distance Depth (m) (m) (%) 1997 Marigold Walnut Formation 6,116 0.4% 1997 Whitewater 5,472 56 1.0% 2007 Flat Depression 26 7 26% 2007 Sister Depression 7 19% 2007 Snakehole 67 8 12% 2007 Fieldsoil 65 8 12% 2010 Wade Sink 852 25 2010 929 26

PAGE 381

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE364ReferencesCollins EW. 1995. Structural framework of the Edwards Aquifer, Balcones Fault Zone, Central Texas: (NY): John Wiley & Sons. response in the vadose zone of the Edwards karst Society publication. of Juniperus ashei on evapotranspiration and runoff in the Seco Creek watershed. Water Resources Ford DC, Williams, P. 1992. Karst geomorphology and drainage pattern in the alpine karst system & Francis. protection areas in karst aquifers of the Ridge and Valley and Appalachian Plateaus physiographic provinces, rules of thumb for estimating the capture study of Buttercup Creek area, Cedar Park, south Williamson County, Texas. Report by Mike Warton & Assoc. to USFW Service. Segment of the Edwards Aquifer, southern Travis Edwards Aquifer Conservation District and City of Austin Watershed Protection Department. within the Barton Springs Segment of the Edwards Aquifer, Southern Travis County and Northern includes measurements in variation of drip rate and spelothem type associated with drip to categorize discrete soda straw drips may be associated with approximately overlying soil and epikarst drainage. Deeper within the cave systems, discrete drips commonly represent the Speleothems associated with discrete drips tend to be and rimstone dams. Even where only widely distributed, seepage drips are present, the subsurface catchment area can be assumed to extend at least 100 m beyond the cave footprint, unless direct data suggests a smaller subsurface catchment. Subsurface catchment areas over highly permeable Kirschberg Members tend to be smaller than subsurface catchment areas over low permeable units such as the Regional Dense Member, Dolomitic Member, and Basal Nodular Member (identical to the thicker Walnut Formation north of the Colorado River). For the purposes of estimating subsurface catchment areas, minimum hydraulic gradients of 10% can be used to estimate surface extent from cave drips overlain by only high permeable units while hydraulic gradients as low lower permeable units. Vadose groundwater is generally, but not always, directed in the downdip direction or down faulted permeable units to local discharge sites. In general, it is advisable to extend the subsurface catchment area in the updip or upthrown fault direction to the full outcrop of rocks overlying the drip, even if it exceeds the characteristics can be used to relate drips of similar tracing is the most direct method to associate a surface site and cave drip. Because of the complexities of through the epikarst, vadose, and phreatic zones, a quality sampling, and cave drip surveys delineates the

PAGE 382

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2365United States Fish and Wildlife Service. 1996. Final Conservation Plan for Proposed Issuance of a Karst Invertebrates in Travis County, Texas. terrains. New York (NY): Oxford University Press. Woodruff CM. 1984. Water budget analysis for the area contributing recharge to the Edwards Aquifer, Barton Springs Segment. In: C. Woodruff and MoPac South into the Barton Springs Segment of the Edwards Aquifer Travis County, Texas: City of Austin Springs Segment of the Edwards Aquifer, Southern Austin Watershed Protection Dept. short report. Drew, D. editors. Methods in karst hydrogeology. Maclay RW, Small TA. 1984. Carbonate geology and hydrology of the Edwards Aquifer in the San Mammoth Cave National Park, Kentucky, USA: Books. limitations of standard criteria and methods for (springhead and wellhead protection areas) in central Texas. The University of Texas Bureau of quality of the Edwards Aquifer associated with framework and hydrogeologic characteristics of the Edwards Aquifer outcrop (Barton Springs and the distribution of cave fauna in the Austin, Texas Region. Report to U.S. Fish and Wildlife Service.

PAGE 383


PAGE 384

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2IntroductionKarst aquifers develop in soluble rock where groundwater drainage basins that rapidly recharge water beyond the depth of transpiration and evaporation (Jennings, 1985). fractured rock is also an important, and often overlooked, occurred through the soils and macropores than through microscopic pore spaces within the limestone matrix, but refers to other processes including soil drainage, pool storage in open conduits and possibly urban leakage that may cause physical and chemical responses in drips to describe this process hereafter. Soils are often highly heterogeneous, which makes to characterize soil moisture. Other studies have found high variability in recharge over small areas (De Silva, 2004) and variability with topography (Li et al., 2008). This evidence suggests that the portion of recharge to karst aquifers through the soil zone is an important aquifers. several factors including the conductivity of the soil and underlying rock, thickness and structures, initial water AbstractCave drips are useful for characterizing recharge and transport through soils, particularly in upland karst settings. Estimation of upland recharge is important for the Barton Springs Segment (BSS) of the Edwards Aquifer, but discrepancies between previous and recent studies indicate how little is known about it in the BSS. We outline a methodology for using cave drips to from a study of drips in four BSS caves. characterize their hydraulic properties over larger areas, particularly with methods that only yield information sensors). This is particularly true in the BSS where thin, for rapid recharge through the soil zone. Cave drips are well suited for characterizing recharge in upland areas as they often have large source areas. Drip responses to storm events were monitored at drips surface water reaches many of the drips within a few hours of the onset of storm events, even though reported soil Ksat recharge was observed within the subsurface drainage basin of three of the caves. These results indicate that upland recharge may contribute a greater portion of total recharge in the BSS than previously estimated, and that rapid recharge can occur in the absence or discrete recharge USE OF PHYSICAL AND CHEMICAL RESPONSE IN CAVE DRIPS TO CHARACTERIZE UPLAND RECHARGE IN THE BARTON SPRINGS SEGMENT OF THE EDWARDS Brian Cowan Zara Environmental LLC, 1707 FM1626, Manchaca, Texas, 78652, brian@zaraenvironmental.comNico Hauwert City of Austin Watershed Protection Department, PO Box 1088, Austin, Texas, 78767, nico.hauwert@austintexas.gov367

PAGE 385

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCElimestone and dolomite units in which many sinkholes and caves have formed. Four caves were monitored during two separate studies: Barker Ranch Cave in 2007 and three caves located near Deer Lane in south Austin Kirschberg, and Dolomitic members of the Edwards group. The Leached and Collapsed, Kirschberg and connected units that contain many of the known caves and recharge features in the area. The Regional Dense laterally across the top of these beds until a breach is The caves selected for this study are located on City of Austin Water Quality Protection Lands. Because we focus on upland recharge through soils, Blowing Sink Cave will not be discussed further as several nearby sinkholes content from previous storms and time since onset of the rate, but when soils contain structures such as root holes the structures at a rate much faster than would be expected based on the permeability of its limiting layer. Soil tracer than expected (Quisenberry and Phillips, 1976; Jarvis et al., 1987; Dekker and Ritsema, 1994; Flury et al., 1994; Kelly and Pomes, 1998). Furthermore, the National Research Council (2001) states that, There exists a fractured rocks and structured soils does not always occur as a wetting front advancing at a uniform rate. upland recharge. During a study from July 1979 to December 1982 it was estimated that 85% of rainfall left as evapotranspiration, 9% left as surface runoff, creek channels (Slade et al., 1986). The estimated values measured in other karst areas around the world, these estimates including lack of continuous discharge measurements upstream and downstream of the recharge zone, no direct evapotranspiration measurements, measurement) for the intervening outcrop area between the creek channels, and an incorrect assumption of the size of the recharge area draining to Barton Springs. A multiple year water balance performed at two upland sinkholes in the BSS indicated that the percentage of barriers to groundwater recharge. Study areaThe study area is located in the Recharge Zone of the BSS and is underlain by the Edwards group that 368 Figure 1. Overview of study area.

PAGE 386

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 dyes in cave drips were detected using activated carbon receptors and optical brighteners were detected using unbleached organic cotton receptors. Chemical tracers were sampled using automatic samplers set at intervals ranging from 4 to 6 hours. Surface runoff samples were collected for comparison to cave drips to help determine dyes and optical brighteners were performed at the Ozark Underground Laboratory in Protem, Missouri and chemical analyses were performed at the Lower Colorado River Authority Environmental Laboratory Services facility in Austin, Texas. Driprate was measured continuously at three drip sites at that site greatly exceeds the capacity of the tipping buckets during rain events. To collect enough drip water sink holes or open apertures in the bedrock, are known delineated catchment areas are overlain by soils, and at cover (roads and houses). MethodsChemical and dye tracer arrivals times and the natural geochemical response of several drips were monitored in the caves during two studies. A total of nine traces were 4CO, and an aqueous Fe solution), optical brightener (Tinopal) were also used to determine recharge source areas and travel times to cave drips. Tracers were applied to soils overlying the caves and 369 Figure 2. Summary of tracer injections at BR.

PAGE 387

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE370Table 1. Tracer injection and detection summary. Injection locations correspond to locations shown on Figure 2 and Figure 3. Figure 3. Summary of tracer injections at GC and MR. Location Detection Arrival Time 1 KBr 2 NaCl Inconclusive N ~ 48 hours 4 Pyranine and KBr None 5 4CO and KI 4CO< 26 hours 6 NaCl Inconclusive 2 Direct Yellow 96 and Fe 2 hours 6 4CONone Pyranine and KBr Inconclusive 7 Tinopal and Fe 12 days* 8 Pyranine and KBr None 7 Tinopal and Fe Waterfall <8 days**Dye detected on passive cotton sampler deployed for 1 or more weeks. In process of analyzing water samples taken at more frequent intervals to determine arrival time.

PAGE 388

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2Driprate Response Driprate responded rapidly (within a few hours of delayed drip response to storm events (Figure 6) indicating that that drip has a more slow flow source than other monitored drips. An automatic sampler collected samples at one site in MR where driprate large volume of discharging at that site. Based on our observations, MR Waterfall appears to be connected to a water quality pond located approximately 10 meters away as responds within a couple hours of water entering the pond. There are no discrete recharge features (i.e. sinkholes) within the source features within source area of BR; however, tracers applied to the soils in the BR source area were for chemical analyses, small clusters of active drips were monitored at each of the sites. Plastic tarps were used to direct drip water into a collection apparatus cell (physiochemical parameters), that discharged into sampling container from which automatic samplers so that an automatic sampler could collect samples and a sonde could record physiochemical parameters. All components coming into contact with drip water were laboratory soap, and deionized water. Some gaps in drip rate data exist due to logger failures. arrival times were used to determine when breakthrough of recharging water occurred at various drip sites. Chemograph separation was conducted using changes incongruent dissolution of dolomite and precipitation of calcite (Figure 4, Musgrove and Banner, 2004). In the vadose zone, residence time is controlled by is characterized by long residence time with more Results Tracer ArrivalSeveral tracer breakthroughs were detected whereas some tracers were never detected (Table 1). At BR, tracers typically arrived at the monitoring sites within a few hours of application to the soils (Figure intervals (typically daily during tracing) than the cotton receptors are changed (typically weekly during tracing). Water samples that were collected during the interval that cotton receptors testing positive for Tinopal were deployed will be analyzed to further constrain tracer arrival times. 371 Figure 4. Geochemical changes caused by increased residence time as measured in vadose and phreatic waters of the Edwards Aquifer. From Musgrove and Banner 2004. Figure 5. Breakthrough of Bromide tracer measured at BR-Main within a few hours of application to the soil.

PAGE 389

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEa recent two year drought period. Also recorded response to a single day storm. Prior to, and during the dripwater emerging at that time had experienced values, whereas samples taken 4 days after the storm model in that waters sampled previous to and during the early part of the storm event underwent more These temporal variations suggest that Barker Ranch Similar but somewhat muted geochemical responses that were applied to the soil near BR were detected arrivals typically occurred concurrently with decreases in trace element ratios. Tinopal detections occurred at tracers will require further analyses to better constrain arrival times (Table 1). Optical brighteners and dyes applied simultaneously with chemical tracers were never detected in BR, even Geochemical Response Ca) to storm events was observed at all sites monitored. This response was more pronounced at Barker Ranch was observed several hours after the onset of the rain events and subsequent increase in driprate. The decrease in trace element ration is interpreted shorter residence time water (related to the rain event) reaching the monitoring sites. DiscussionThe lack of discrete recharge features within the mapped caves is likely occurring through the soil and epikarst and not through discrete features such as sinkholes or solution cavities. The rapid drip rate response observed through the epikarst is occurring. It is well documented that pressure waves can rapidly propagate through the vadose zone and cause a rapid increase in driprate that is not associated with the arrival of recently recharged groundwater. To gain a better understanding of how rapidly recharge is occurring one must examine driprate tracer arrival times. precipitation events, BR is characteristic of a system hours (Figure 6). Anecdotal evidence suggests that there as the site has continued to drip, albeit slowly, through 372Figure 6. Representative driprate responses to rainfall events at MR-Fissure, GC-Main and BR-Main.

PAGE 390

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2The rapid driprate, geochemical response, and tracer arrival times indicate that water is rapidly recharging through the soil zone. Soils in the study area are clay recharge is not possible. Because the soils are clay rich, they shrink and swell with continual wetting and drying. This shrinking and swelling can creates macropores, which are visible on the surface during dry periods (Figure 8). When soils contain structures such as root holes and structures at a rate much faster than would be expected based on the permeability of its limiting layer. Soil tracer than expected (Quisenberry and Phillips, 1976; Jarvis et al., 1987; Dekker and Ritsema, 1994; Flury et al., 2994; when chemical tracers were. The optical brighteners and dyes have a high affinity for clay particles and were likely adsorbed by the clay rich soils overlying the study area. Optical brighteners that were applied adsorbed by clay particles in the soil similar to dyes applied to the soils near Barker Ranch; however, a was previously applied at BR, resulting in some of the Tinopal penetrating through the soil zone and washed down a drainage and into a retention pond that overlays MR (Tinopal could be seen in the water flowing into the retention pond). The pond is lined by thick, locally sourced soils, yet Tinopal infiltration through the soils. 373Figure 7. Geochemical response to rainfall events at BR-Main and GC-Main.

PAGE 391

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE374 model from New Zealand karst. Journal of within the Barton Springs Segment of the Edwardsa Counites, Texas [doctoral dissertation]. Austin Rose) and Edwards Aquifers along the Balcones Fault Zone and Related Topics. Proceedings of the Karst Conservation Initiative; Austin, Texas. areas to cave drips and cave streams in Austin Texas, USA. Proceedings of the Thirteenth Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst; Carlsbad, New Mexico. (CA): Academic Press. Kelly and Pomes, 1998). It is likely that macropore observed in the study area despite the lack of discrete recharge features within the mapped source areas of drips is an effective method for characterizing recharge through soils in karst settings. ReferencesDe Silva, RP. 2004. Spatial variability of groundwater Dekker, LW, Ritsema, CJ. 1994. Variation in water content and wetting patterns in Dutch water repellent peaty clay and clayey peat soils: Catena content under the no tillage cropping system. Figure 8. Macropore observed near Barker Ranch. Representative of macropores found throughout study area when soils are dry. The macropore was not enlarged but vegetation was removed to better photograph. Note cell phone is approximately two inches tall by one inch wide.

PAGE 392

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2375 transport of nitrate and bromide in claypan soil. hydraulic conductivity in limestone dolines, Musgrove, M, Banner, JL. 2004. Controls on the spatial and temporal variability of vadose dripwater geochemistry: Edwards Aquifer, central Texas. National Research Council. 2001. Conceptual Models of Flow and Transport in the Fractured Vadose Zone. Washington (DC): National Academy Press. Quisenberry, VL, Philips, R.E. 1976. Percolation of water quality of the Edwards Aquifer associated with WebSoilSurvey.aspx

PAGE 393


PAGE 394

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2caves generally provide only a small area of access to a much larger groundwater system. Presumptive habitat is the concept that all groundwater with subsurface hydrological interconnections should be presumed to contain the species of concern that is found in accessible in this concept is the assumption that the conditions groundwater are compatible with the survival of the species in question. This concept allows for more realistic evaluation of the habitat for subterranean aquatic species of concern. It is important to distinguish the recharge area from presumptive habitat. In most cases, the entire recharge area should not be included as presumptive habitat for the aquatic species. In many cases characteristics of the species in question indicate that certain portions of the recharge area would not be expected to contain the species of concern. Although the entire recharge area for aquatic species should be managed to maintain good water quality of the probable extent of species habitat beyond an area of direct observation is important for the effective management of subterranean aquatic species of concern. AbstractDelineating habitats for aquatic species of concern present in groundwater systems, and especially those in karst groundwater systems, presents challenges. It is not reasonable to limit the delineated habitat to those portions of a groundwater system that can be directly observed. case histories in the Ozarks region of the central USA illustrate differing approaches for identifying presumptive habitat in recharge area delineations for the Tumbling Creek Cavesnail explores the reasoning for designating presumptive habitat downgradient of observed habitat in a cave stream. The second case study presumptive habitat in a complex distributary spring the case for expanding the presumptive habitat in both IntroductionDelineating habitats for aquatic species of concern present in groundwater systems, and especially those in karst groundwater systems, presents challenges. It is not reasonable to limit the delineated habitat to those portions of a groundwater system that can be delineations? This paper presents three case histories of the use of presumptive habitat designations in recharge area delineations. Presumptive Habitat observation or capture. This method is severely limited when dealing with subterranean aquatic species. Even THE NEED FOR PRESUMPTIVE HABITAT CONSIDERATIONS IN WORKING WITH SUBTERRANEAN AQUATIC SPECIES OF CONCERN: THREE OZARK Shiloh L. Beeman, Thomas J. Aley Ozark Underground Laboratory, Inc., 1572 Aley Lane, Protem, Missouri, USA, Shiloh@ozarkundergroundlab.comMichael Slay377 The Tumbling Creek Cavesnail (Antrobia culveri) is federally listed as endangered under the provisions of the Endangered Species Act. The only known habitat for this aquatic cavesnail is the stream in Tumbling Creek Cave in As a result of litigation, the U.S. Fish and Wildlife Service was required to designate critical habitat for this species. A hydrogeologically based determination of the area that constituted critical habitat for this species was needed.

PAGE 395

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE 20 springs depending upon how one counts springs that are relatively close to one another. These springs are located along 1,585 meters (5,200 feet) of surface stream channels. The range in elevation of these springs is 15.7 introduced into Tumbling Creek in areas near the Big Room will subsequently discharge from all of the 15 to 20 documented. The relatively insoluble nature of the chert is the reason for the large number of springs. Under high cms) that passes the weir in the Big Room. Flow beneath the chert unit between the cave and the springs is through a matrix of solutionally widened and interconnected openings localized immediately beneath large number of springs draining the cave and the rapid tracer tests. This area provides presumptive habitat for the cavesnail. The karst groundwater system beneath the chert unit is not humanly accessible so no survey can be conducted in the hydrologic and biologic conditions present for snails beneath the chert unit and hydrologically down gradient of the accessible portions of Tumbling Creek are essentially identical with the conditions found in the areas of Tumbling Creek that are known habitat for the Bat guano that is an important energy source for the cave area are rapid and, like the accessible portions of the cave stream, capable of transporting sediment and organic matter in suspension. As a consequence, during storm events Tumbling Creek and all of the springs are turbid. The designated critical habitat for the cavesnail (Federal Register 2011) includes all accessible portions of Tumbling Creek within the cave. It also includes as presumptive habitat the springs known to drain Tumbling Creek Cave and most of the lands immediately underlain by the chert unit in locations tributary to the springs. portions of the cave are in the lower members of the dolomitic Cotter Formation of Ordovician age. A locally massive chert unit with a typical thickness cave passages. While dolomite is relatively soluble in groundwater, chert is relatively insoluble. Water passing through a fracture in dolomite will, through time, enlarge the opening by solution. In contrast, the passing water will be much less effective in enlarging a similar fracture in the chert. The distance from the bed of the cave stream to the top the cave stream the water rises through a solutionally water is rising under pressure up through the chert bed rises through a pool that apparently overlies another fracture in the chert unit. The weir at the stream gauging station in the cave is about 595 meters (1,952 feet) upstream of the Bear Cave entrance to the Tumbling Creek Cave System. Flow rates of about 0.12 cubic meters per second (cms) at the weir are needed before any flow in the cave stream will discharge through the Bear Cave entrance. Water lost from the channel of Tumbling Creek moves downward through fractures in the chert into dolomitic units that underlie the chert unit. The sinking segments of the cave stream are highly localized and most are within 150 meters of the Bear Cave entrance. The largest single flow loss zone is about 60 meters (197 feet) upstream of this entrance. With the exception of the Bear Cave entrance to the cave (which is above the chert unit), all of the springs that drain the cave and the accessible portions of the snail habitat derive almost all of their water from present in a karst window located between the Bear Cave entrance and most of the springs.378

PAGE 396

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2some habitat in epikarstic zones. The epikarstic zone is the weathered upper portion of soluble bedrock units. Its thickness is highly variable, but road cuts near Ozark thickness of the epikarstic zone in valleys, and especially likely that typical thicknesses equal and probably exceed those observed in highway road cuts passing through stream valleys are largely saturated with water and, as a result, provide substantial habitat for cave fauna Dye tracing can yield detections at two or more springs. Distributary spring systems can be common in areas is known habitat for a species of concern, then it is reasonable to conclude that all of the hydrologically linked spring system should be viewed as presumptive springs are draining a saturated epikarstic system. This hydrogeologic condition further supports presumptive habitat within the other springs with hydrologic interactions. Credible hydrogeologic and biologic data, including the uniformity of conditions, support recognition of this downgradient area as presumptive habitat for the cavesnail and warrant its designation as critical habitat.Amblyopsis rosae Ozarks into northwest Arkansas and northeast Oklahoma. species by the U.S. Fish and Wildlife Service. Several small community of Neosho, Missouri USA. the threatened species. Five groundwater traces were Spring (Aley and Aley 1997). A recharge area of 14.7 square kilometers (5.67 square miles) was delineated for the spring and consisted of lands in two different Spring were also detected at South Big Spring, located approximately 1220 meters (4000 feet) to the northwest. previously considered to be presumptive habitat based important hydrobiological concept. shared with the recharge areas of four other springs east. This distributary groundwater system is draining al. (2007) summarized recharge area delineation results Oklahoma. They found that 79% of the sites had at least 379 Cambarus zophonastes) was discovered in a small spring located along a perennial stream that bisects the small town of Yellville, Arkansas. This troglobitic species is federally listed as endangered and previously known only from two caves located approximately 65 kilometers (40 miles) to the southeast. tributary. The recharge area delineation was performed several good rainfall events did occur that allowed for the introduction of tracer dyes, the study was largely performed under regional drought conditions. basins, including East Prong and Town Branch upstream

PAGE 397

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE380 conditions (12.6 square kilometers or 4.9 square miles). The expanded presumptive habitat designation and subsequently the larger recharge area provide a more reasonable area for conservation management for species protection.SummaryFrom a conservation perspective, the entire recharge area must be managed for subterranean, aquatic species recharge area, the presumptive habitat must be considered in addition to the known habitats of observed populations. In most cases it is more reasonable to extend presumptive habitat downgradient from a known site than upgradient. This was the case with Tumbling Creek Cavesnail. an expanded area downgradient from the known habitat within the cavestream even though direct observation in this area is not available. Other cases of important presumptive habitat designation include distributary spring systems that drain saturated it is reasonable that if a known population exists in one spring, populations could exist in other areas with hydrogeologic connections and similar hydroglogical conditions. Presumptive habitat designations in these laterally expansive, and can therefore result in larger connections between observed population sites and other important to be established during the recharge area delineation of a known population of concern. In some settings, it is reasonable to extend the presumptive habitat upgradient from a known population site. The hydrologic connections through seasonal groundwater subterranean species of concern has led to a presumptive habitat designation that was expanded to include some presumptive habitat is important to identify an adequate recharge area for effective species protection, especially considering the population growth and ongoing development in this area. present at the time of the study, only one of these dye introductions (East Prong) was detected in Legion Spring (Kirkland and Aley 2012). Although gaining Spring, groundwater tracing results indicated water from Town Branch downstream of Legion Spring was sinking into the subsurface and recharging two small springs in Dye tracing indicated the recharge area for the spring conditions of the study included 18.4 square kilometers spring was reasonable. Several lines of evidence pointed to a larger presumptive well developed epikarst was observed in road cuts and encountered in borings at a nearby petroleum release site. The spring was located on the main branch of a small creek only water from the tributary valley was traced to the to be present within the epikarst on the across the small creek. Upstream of Legion Spring on Town Branch is a gaining segment of the creek that drains the epikarst, as evidenced by a healthy population of watercress in the creek immediately downstream of the point where water is reasonable that the areas of saturated epikarst also be included within the presumptive habitat. Areas downstream of Legion Spring along Town Branch were also included within the presumptive habitat due to the presence of perennial springs with hydrologic interactions with Town Branch below Legion Spring. These downgradient areas extensively developed and saturated epikarst. A larger presumptive habitat was reasonable for the designated presumptive habitat area consisted of 2.2 square kilometers (0.85 square miles) in areas both upgradient and downgradient of the observed population site. In consideration of the larger presumptive habitat, a larger recharge area (18.4 square kilometers or 7.1 square miles)

PAGE 398

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2381References delineation, hydrobiological assessment, and (Amblyopsis rosae) populations in Missouri. Contract study for U.S. Fish and Wildlife Service by Ozark Underground Laboratory, Inc. 115 p. plus appendices. Aley T, Aley C. 2001. Delineation of the recharge areas for Tumbling Creek Cave and Millrace Springs, Taney County, Missouri. Contract study for U.S. Fish and Wildlife Service by Ozark Underground characteristics of delineated recharge areas for 40 in Missouri, Arkansas, Oklahoma, and Illinois. Proc. 18th National Cave and Karst Management Aley T, Coonrod D, Kirkland S. 2011. Walbridge Spring delineation study and vulnerability mapping, City of Neosho, Newton County, Missouri. Contract study for U.S. Fish and Wildlife Service by Ozark Underground Laboratory, Inc. 48 p. plus appendices. area delineation and vulnerability assessment of Arkansas. Contract study for The Nature Conservancy by Ozark Underground Laboratory, Inc. 66 p. plus appendices. (Antrobia culveri) recovery plan. U.S. Fish and Wildlife Service, Columbia, MO. 97p. for subterranean aquatic species of concern. Proc. 16th National Cave and Karst Management

PAGE 399


PAGE 400

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2of attention by scientists, practitioners and authorities on this subtle hazard.IntroductionItaly is affected by a high number of geological hazards that cause severe losses every year and are at the origin of many casualties, thus producing a very high toll to the society. Earthquakes, volcanic eruptions, slope 2000). To these, further phenomena, at least in part with the consequent effects on loss of vegetational cover and soil erosion. Within such a framework, sinkholes have often been under estimated if not entirely neglected, and only in some sectors of Italy have they been effectively considered in hazard and risk analysis. Nevertheless, sinkholes affect large parts of the country, covering the whole territory with a variety of typologies, and showing both natural and anthropogenic origin. The latter clearly derives from the long history of Italy, with the complex historical vicissitudes that have characterized this territory, during realized underground for different purposes in different epochs (Parise, 2012). Over time, many of these cavities have been abandoned, and the loss of memory about such features resulted in their inclusion in newly developed environment above (Pepe et al., this volume). When unstable underground caves are located below buildings or infrastructures, their collapse may result evolution of the failure processes (in particular during the catastrophic phase of rupture, and the resulting sudden collapse) can also represent in some cases a risk to human life. In this regard, the high rate of evolution is related to either the propagation of fractures through AbstractAmong the many different types of geological hazards affecting the Italian territory, sinkholes have often been underestimated (if not neglected entirely), and only in some sectors of the country have they been effectively considered in hazard and risk analysis. Nevertheless, sinkholes affect large parts of Italy, covering the whole territory with a variety of typologies, and exhibit both natural and anthropogenic origin. The latter clearly originate from the long history of Italy, with the complex historical vicissitudes that have characterized cavities have been realized underground for different purposes in different epochs. Over time, many of these cavities have been abandoned, and the consequent loss of memory resulted in their inclusion in newly developed environment above. Starting from these considerations, an archival research was started to collect information about the occurrence of sinkholes in Italy, with particular attention to their precise site and date of occurrence, in order to make an effort in assessing, respectively, the susceptibility and the hazard related to the particular phenomenon under study. As concerns date of occurrence, the accuracy of the information is provided (depending upon the amount of available data), with the highest quality when hour, day, month and year of occurrence are indicated, and a decrease in quality when one or more of these data are lacking. In order to be included in the database, at least some kind of temporal reference (even if general) of the sinkhole has to be known. describing the catalogue obtained so far which consists of more than 650 sinkhole events for which at least some information about temporal occurrence of the event represent a good starting point for analysis of the sinkhole hazard at a national scale, aimed at increasing the level A CHRONOLOGICAL CATALOGUE OF SINKHOLES IN ITALY : THE FIRST STEP TOWARD A REAL EVALUATION OF THE SINKHOLE HAZARD Mario Parise, Carmela Vennari CNR-IRPI, Via Amendola 122-I, 70126, Bari, Italy, m.parise@ba.irpi.cnr.it383

PAGE 401

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEIn the case of natural sinkholes, the related surface effects may go unnoticed, especially when they occur in rural areas; further, land owners often prefer not to spread the news, in order to avoid loss of value of the land. In Italy, some databases about sinkholes are already and by University of Rome Tre. These databases, however, do not focus in particular neither on time of occurrence of the sinkholes, nor in the distinction between the natural or anthropogenic origin. They are hazard, but rather to indicate the areas of the territory where sinkholes do occur (at the national or regional information, or a number of cases related to leakage from water pipelines, which may bring to wrong or incorrect conclusions. To fully evaluate the hazard related to natural danger, the return time of the phenomena needs to be estimated. Lacking such information, there is no possibility to effectively determine the hazard, and the research has generally to stop at the stage of susceptibility assessment (Varnes, 1984). Thus, availability of documentation about past events is crucial. Unfortunately, very often the record of what occurred when some documentation is available, most of the information show low reliability, and are based on memories, or are not supported by real documents. This represents a very important drawback that often has to be faced in the search for historical information on different types of hazards (see, for instance, Calcaterra & Parise, Actually, it would be useful to better understand the conditions that are likely to cause sinkhole formation research is to evaluate the timing of sinkholes in Italy, (climatologic, meteorologic, seismic, etc.) under which they formed. There is a great variety of sinkhole types in Italy (Buchignani et al., 2008; Nisio, 2008; Del Prete et al., 2010; Margiotta et al., 2012), with events triggered by very different factors, including rainfall, seismic shocks the intact rock surrounding the caves (Kowalski, 1991; Liu et al., 2000) or the slip movement of rock of these being generally controlled by progressive failure conditions, that sometimes induce catastrophic collapse (Diederichs and Kaiser, 1999a, b; Lanaro, 2000; Starzec and Tsang, 2002; Pine et al., 2006; Parise, 2008; Parise and Lollino, 2011). Several aspects make sinkholes differ from other types of is a phenomenon which at the surface appears as punctual generally of limited dimensions. This, however, does not mean that the affected area is limited to that point, since underground the caves responsible for the event may also show a wide extension. Consequently, sinkholes are rarely taken into account in the analysis of hazard and risk assessment, even though they may represent the With respect to natural sinkholes, these typically are found in areas with soluble rocks affected by karst processes, or alluvial sediments in peculiar geological may be done for natural sinkholes based solely upon The situation is quite different for anthropogenic depends upon many other factors, such as historical, cultural, and social issues. In this regard, it is important to highlight that we excluded from the catalogue all those events of anthropogenic origin which were reported as due to water leakage from pipelines, with the consequent erosion in the subsoil, without any evidence of an anthropogenic sinkhole means a sinkhole caused by the presence of a cavity created by man underground, regardless of the cavity typology. In general, there is more detailed information about anthropogenic sinkholes than for those of natural origin. areas, and causes direct damage and negative effects to society (blockage of roads and communication routes, disrupting lifelines, etc.). At least some of these effects and site of occurrence of the phenomenon.384

PAGE 402

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2The catalogue structureBased upon the considerations presented in the previous section, the catalogue on Italian sinkholes (Figure 2) has been structured giving a crucial role to the information on time of occurrence of the sinkhole, and, in addition, clearly covering hour, day, month and year. Ideally, availability of all data represents the best condition (high category class). also considered the possibility to have available only a part of the timing information, which in many cases is likely limited to month and year (medium category class). This in particular occurs for sinkholes located in rural areas, where generally no record or observation is recorded soon after the occurrence of the event. Lacking a temporal reference of the sinkhole, it is range (low category class) for occurrence of the event, as and human actions. In fact, even though most of the cases seem to occur in response to meteoric events, and the deriving runoff and groundwater circulation, reactivation of sinkholes may be a consequence of earthquakes too, as recently experienced at the Sinizzo Lake after the Abruzzo earthquake in 2009 (Figure 1; Parise et al., 2010); or be originated by construction from putting in operation or maintaining the network of pipelines. Further, the interest on sinkholes is increasing sinkhole features also in geological settings never before considered, such as offshore (see at this regard the work by Taviani et al., 2012). In the following sections, we will describe the structure presenting the future perspectives of this research activity on sinkholes.385Figure 1. Development of cracks along the shores of Sinizzo Lake (Abruzzo), in the aftermath of the Mw 6.3 LAquila earthquake, on April, 6, 2009. Figure 2. Distribution of the sinkholes listed in the catalogue over the Italian territory (n = 652). The different colors indicate the degree of certainty in the location: green means certain location, yellow uncertain location, and red very generic indication about the site of occurrence.

PAGE 403

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEthe growing public interest in sinkhole formation in the region (Brinkmann and Parise, 2010). refer to origin of the sinkhole (divided into natural, anthropogenic, and of unknown origin), information on the triggering factor, the main morphometric parameters of the sinkhole (diameter, depth), and an indication of already known phenomenon. Regarding origin, sinkholes were included in the unknown category when no clear information about the nature of the underground cavities was available. Damage recorded as a consequence of the occurrence of sinkholes provides very important information. Evaluating damage allows, as a matter of fact, assessment hazard, in this case represented by a sinkhole. It is not cases the information are qualitative, especially for the oldest events. In any case, having the possibility to make a general framework of the damage occurred is a good starting point in the effort to assess the vulnerability linked to sinkholes. With respect to location, the sites affected by sinkholes also mapped at 1:25,000 scale, using the topographic The sinkholes are represented (Figure 2) with different colors which indicate the level of uncertainty in location of the event: green is for certainty in the location, yellow for uncertainty, red for very generic indication of the site affected by the sinkhole.Preliminary evaluation of the catalogue outcomesIn total, 652 sinkholes, for which availability of temporal references has been found, are included so far in the catalogue. Of these, about 100 are anthropogenic sinkholes in the town of Palermo, Sicily, and have been extracted from the work by Sottile (2010), whilst about dealing with anthropogenic sinkholes in the town of Naples, Campania, one of the most well known sites in idea of the time evolution at a particular site. Availability mandatory to include an event in the database. When the event is included, a degree of accuracy is also attributed, move the degree of reliability toward the lowest values. This is a very important aspect that strongly makes our catalogue differ from other sinkhole databases in Italy. As aforementioned, we consider of crucial importance the knowledge of the time of occurrence, as a fundamental element to allow an estimate of the likely return time of the events, and therefore to assess the sinkhole hazard. The sources of historical information include newspaper available database on sinkholes. Furthermore, regional and local history books, and transcriptions or translations of old chronicles have also been considered. Some additional information derived from reports prepared effects of single events and their consequences, and from unpublished technical reports of practitioners. last 15 years provided additional information for some regions of southern Italy (namely, Apulia, Campania, and Calabria). A few bachelor degree theses also dealt areas (for instance, the town of Palermo, with the work by Sottile, 2010). For some recent events, new or updated information was obtained through the Internet, line newspaper. This is certainly one of the most useful sources nowadays, and has become increasingly used to collect data on a variety of natural hazards. Regarding sinkholes, one of the few published examples of the use of internet information was recently presented by Brinkmann and Parise (2010). They used two sources of data for recently formed sinkholes in Florida, in the attempt to determine the timing history of sinkholes at database and the LexisNexis database of newspaper articles. From such sources, a good number of information was extracted on several tens of sinkholes, which allowed associated with seasonality, at the same time highlighting 386

PAGE 404

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2The events are not equally distributed over time. There is an overall increase in the last decades, which, on the other hand, is not always accompanied by adequate completeness in the temporal information. In other words, it was expected that, moving toward recent times, a higher degree of precision in the temporal information could be reached. Nevertheless, it is very common that, even for recent occurrence of sinkholes, the accuracy in the timing of occurrence may be low. In practice, if the sinkhole is recorded or witnessed soon after its reports or chronicles, timing is typically accurate; when, on the other hand, the news is obtained later on, only generic information, generally limited to month and year, have to be registered. As noted by previous scholars dealing with natural events. Lack of occurrence of events in a given period to changes in the conditions that led to trigger the event Information on the location (either precise or approximate) of the sinkhole is available for most of the events in the catalogue. Namely, location of 72% is certain, whilst 21% of the events has some degree of uncertainty, and only the remaining 7% (47 sinkholes) presents a very generic indication about the site of occurrence of the sinkhole. Sites affected by sinkholes with a chronological reference are not distributed equally in Italy. Their distribution is Italy for sinkhole problems. Overall, more than half of the sinkholes in the catalogue (precisely, 54%) have an anthropogenic origin, whilst a quarter of the entries are due to natural caves, and for the remaining cases no clear In the catalogue, information on the timing and documents and chronicles are typically more accurate natural hazards, rather than on their precise location and date (Salvati et al., 2010). As previously stated, timing of occurrence represents the main information needed for inclusion of a sinkhole in the catalogue. The best situation, represented by availability of all the data concerning time of occurrence (high the day, month, and year is known, an overall percentage chart in Figure 4). Lower percentages characterize the other categories, where the availability of information about time of occurrence of the sinkhole progressively With respect to the chronological distribution of the sinkholes, the oldest information goes back to historical times for a few cases; some sinkhole events th century, even though most of the information concentrates starting from the 19th century. For anthropogenic sinkholes, the oldest documented event goes back to the end of the 19th century in Calabria. 387 0 20 40 60 N A U 24 % 54 % 22 % Figure 3. Histogram showing the origin of sinkholes in the catalogue: N=natural sinkholes; A=anthropogenic sinkholes; U=sinkholes of unknown origin. Figure 4. Pie chart showing the percentage of accuracy of date of occurrence of sinkholes.

PAGE 405

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCELazio, Sicily and Apulia will be by far the most affected by sinkhole problems, especially due to presence of It has, however, to be noted that small numbers in Table sinkholes, but rather to unavailability of a temporal reference (which, as before stated, is mandatory for inclusion of an event in the catalogue). In particular, the very low number of anthropogenic sinkholes in Latium (5) has to be noted, given the high frequency of events affecting many towns of this region, including Rome. are available about the damage produced. In 17 cases, sinkholes have caused fatalities (the highest toll has been 4 victims in a single event), whilst casualties (deaths and is certainly an issue which needs to be better examined, since damage analysis is crucial for assessing the effects the resulting data might be used for delineating different scenarios in case of occurrence of further sinkholes, or Apulia (Puglia) region (Figure 5) is here presented as an example of analysis of the catalogue at a higher degree of detail. Choice of the region was dictated by our activity in this territory about sinkholes, and the research consequence of the good knowledge of this territory, and of our direct experience in many sinkhole events, is the high certainty in sinkhole location (70 % of the sample; Figure 6); analogously, the accuracy in date of occurrence essentially dependent upon the presence of soluble rocks or alluvial deposits as regards the natural sinkholes, whilst the location of the anthropogenic sinkholes cavities. Thus, some towns presents a very high number of phenomena, which is also dependent on a better record of the sinkhole events (this is the case, for instance, of towns as Naples, Campania, and Palermo, Sicily; see Concerning the spatial distribution of sinkholes (Figure 2 and Table 1), some regions show very high numbers, whilst others are much less represented. Two regions subdivided so far do not have any event in the catalogue. This inhomogeneity derives from greater availability of from direct experience in some others. Nevertheless, we expect that the future data entry of other sinkholes will in some ways reduce this gap. In any case, it is well known that regions such as Campania, 388Table 1. Regional distribution of the sinkholes in the catalogue. The origin of sinkholes is also shown (N=natural; A=anthropogenic; U=unknown). regions tot N A U Abruzzo 22 8 14 Basilicata 6 6 Calabria 8 6 1 1 Campania 246 174 49 Emilia Romagna 6 1 2 1 2 Lazio 107 67 5 Liguria 4 1 Lombardia 1 1 Molise Piemonte 1 1 Puglia 88 29 6 Sardegna 9 6 Sicilia 117 1 111 5 Toscana 6 2 5 Trentino Alto Adige 1 1 Umbria 10 10 Veneto 10 10 tot 652 157 350 145 Figure 5. Sinkhole distribution in Apulia region (n=88).

PAGE 406

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 at Altamura and, especially, that of March 29, 2007, at and to a more careful record of the sinkhole occurrence. In fact, following the above cited events, the Basin Authority of Apulia (that is, the Regional Body in charge related regulations for land management) issued some to high accuracy (where medium means knowledge of month + year of occurrence, and high availability also of the hour of the event; Figure 6). Looking at the chronological distribution of the events, the histogram a in Figure 7 shows that, starting from 1925 (date of occurrence of the oldest documented event in the region, registered at Canosa di Puglia), an almost continuous increase in the number of sinkholes with temporal reference has been observed in Apulia, with slight decreases in this general pattern during the 1960s and 1980s It has also to be noted that the last bar in the histogram covers a period of less than two years (2011 b is the detail for the 389 Figure 6. Pie charts illustrating for the Apulia sample (n=88) the percentage of accuracy of date of occurrence of sinkholes (above), and the certainty in location of the sinkhole (below). Figure 7. Chronological distribution of sinkholes in Apulia: a) histogram covering the time span from 1925 (oldest documented date) to the present day; b) histogram showing the detailed distribution in the time span 2000-2012. Figure 8. The March 29, 2007 sinkhole at Gallipoli (Apulia), a recent example of anthropogenic event due to an ancient underground quarry.

PAGE 407

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEbetween sinkholes related to natural karst caves (natural sinkholes) and those linked to cavities realized by man two elements above are of crucial importance for a better understanding of the sinkhole hazard and risk. Knowing dates of the events represents the necessary element for anthropogenic sinkholes may determine very different scenarios to be faced. On the other hand, we are well aware that in terms of civil protection issues, it is very important to take into account the analysis of all those events which have caused damage to the society, regardless of their origin as data about the intensity of the consequences of sinkholes, with particular regard to number of fatalities and number of casualties appears to be the main goal, since these data are a direct, quantitative measure of the intensity of a disaster, and can be used to evaluate individual possibility of sinkhole occurrence in the case of presence of underground caves, of both natural and anthropogenic origin (Fiore, 2006). In addition, the increase in attention on sinkholes was also a consequence of the severe crisis occurring Apulia) where, starting from the 1990s, a high number of cover collapse and cover suffosion sinkholes (sensu Waltham et al., 2005) were recorded in the gypsum local hydrogeological changes caused by maintenance works of the canal in the coastal evaporite aquifer (Fidelibus et al., 2011).Future perspectivesThe catalogue of sinkholes here presented represents At the same time, it differs from other databases on the same topic, for providing a clear discrimination 390 Figure 9. Examples of sinkholes, of natural (upper pictures) and anthropogenic (lower pictures) origin. Above left: Pianelle sinkhole (Campania region), triggered by the November 23, 1980 earthquake. Above right: one of the many sinkholes in the gypsum deposits of Marina di Lesina (Apulia). Below: sinkholes due to presence of underground quarries at Altamura (left; photo courtesy of CARS) and Cutrofiano (right), both in Apulia.

PAGE 408

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2Calcaterra D, Parise M. 2001. The contribution of historical information in the assessment of the historical and geological data for the assessment of the landslide hazard: a case study from and distribution of different types of sinkholes in Diederichs MS, Kaiser PK. 1999a. Stability of large excavations in laminated hard rock masses: the voussoir analogue revisited. Int J Rock Mechanics Diederichs MS, Kaiser PK. 1999b. Tensile strength and abutment relaxation as failure control mechanisms in underground excavations. Int J Rock Mechanics Festa V, Fiore A, Parise M, Siniscalchi A. 2012. Sinkhole evolution in the Apulian karst of southern Italy: a case study, with some considerations on sinkhole hazards. induced hydrogeological changes and sinkholes in the coastal gypsum karst of Lesina Marina area (Foggia Fiore A. 2006. Pericolosit geologica connessa alla presenza di cavit sotterranee. Atto di indirizzo to earthquake sources: extending the contribution of historical seismology to seismotectonic studies. Research: Kluwer Academic Publishers, 17. territory of the city of Naples. Journal of Physics As next steps, we intend therefore to complete the collection of data of the catalogue, since some regions are evidently less covered than others so far (see Figure 1 entries. After that, we will perform statistical analyses on the database, to establish the rate of occurrence of regions or areas that might appear as particularly susceptible to this type of phenomena. Spatial persistence of sinkholes, that is the occurrence of multiple events at the same site, will also be examined, as a very important element in the effort to evaluate the sinkhole hazard. At the same time, the effects of surface and ground waters, will represent one of the main issues to investigate, as a likely factor inducing the occurrence of sinkholes. that appear to be the most affected by anthropogenic sinkholes; this is the case, for instance, of Rome, Naples of many other minor towns in several regions of Italy. the outcomes of the research will be managed in strict collaboration with the Civil Protection Department, aimed at taking into account the sinkhole hazard in a geologically fragile and highly vulnerable territory such as Italy.References the risk of karst subsidence and collapse. Proc. 9th Multidisciplinary Conf. on Sinkholes and the Engineering and Environmental Impact of Karst, italian earthquakes from 461 b.C. to 1997. Annali Brinkmann R, Parise M. 2010. The timing of sinkhole formation in Tampa and Orlando, Florida. The Puccinelli A. 2008. Evaporite karst and sinkholes: a synthesis on the case of Camaiore (Italy). 391

PAGE 409

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEPine RJ, Coggan JS, Flynn ZN, Elmo D. 2006. The development of a new numerical modelling approach for naturally fractured rock masses. Rock Sottile R. 2010. Problematiche geologiche legate della citt di Palermo [bachelor thesis]. Palermo: Palermo University. 248 p. density for predicting the volume of unstable blocks in underground openings. Int J Rock Mechanics Taviani M, Angeletti L, Campiani E, Ceregato A, Foglini F, Maselli V, Morsilli M, Parise M, Trincardi F. 2012. Drowned karst landscapes offshore the Apulian Margin (Southern Adriatic Sea, Italy). J. Varnes, DJ. 1984. Landslide hazard zonation: a review Waltham T, Bell F, Culshaw M. 2005. Sinkholes and subsidence. Karst and cavernous rocks in risk to the population posed by natural hazards in E, editors. Landslide risk management. Taylor & In: Cruden DM, Fell R, editors. Landslide risk Italian Web Sinkhole Database [Internet]. 2012. Rome: University of Rome Tre; [cited 2012 Oct 26]. Available from: Kowalski WC. 1991. Engineering geological aspects of different types of karst corrosion and fracture roughness and aperture of rock fractures. Int J Liu D, Wang S, Li L. 2000. Investigation of fracture behaviour during rock mass failure. Int J Rock Margiotta S, Negri S, Parise M, Valloni R. 2012. Mapping the susceptibility to sinkholes in coastal areas, based on stratigraphy, geomorphology and Nisio S, editor. 2008. I fenomeni naturali di sinkhole nelle aree di pianura italiane. Memorie Descrittive Parise M. 2008. Rock failures in karst. In: Cheng Z, Engineered Slopes. Proceedings 10th International Parise M. 2012. A present risk from past activities: sinkhole occurrence above underground quarries. Parise M, Fiore A. 2011. Chronology of sinkhole events Parise M, Lollino P. 2011. A preliminary analysis of failure Parise M, Perrone A, Violante C, Stewart JP, Simonelli Research Council in the aftermath of the 6 April 2009 Abruzzo earthquake: the Sinizzo Lake case study. Proc. 2nd International Workshop Catastrophic sinkholes in the natural and southern Italy. These proceedings.392

PAGE 410

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2LESSONS LEARNED FROM OCCURRENCE OF SINKHOLES RELATED TO MAN-MADE CAVITIES IN A TOWN OF SOUTHERN ITALYPietro Pepe, Nunzia Pentimone, Giuditta GarzianoApogeo s.r.l., Via Caduti di Nassiriya 170 70022, Altamura, Italy, http://www.apogeo.bizVincenzo MartimucciComune di Altamura, Italy, PariseCNR-IRPI, Via Amendola 122-I, 70126 Bari, Italy, in recent years, which has been organized and managed All activities used to identify the underground quarries, surface, survey the cavities, produce detailed maps, and reclaim the sites in order to allow future development, are described in this paper, as an example of how to properly manage a territory characterized by sinkhole problems.IntroductionAltamura (Figure 1) is one of the largest towns in the a foreland area consisting of carbonate rocks where karst is the main agent shaping the landscape (Parise, 2011). In addition to the many karst caves and surface landforms, including the Pulo, one the largest dolines in the region with a diameter of over 500 m and depth of 92 m, there are several cavities of anthropogenic origin at the outskirts of town, in a sector undergoing urban expansion. population. Until 2006, it was commonly believed that the subterranean voids were limited to a few areas near developed in recent years causing safety concerns. has increased greatly in recent years, with over 16 km of subterranean passages explored and mapped. The present paper describes the methodological approach used to increase the understanding of the caves and procedures have been undertaken to mitigate safety concerns. Abstract parts of towns is a common feature in large portions of Italy. Different typologies of anthropogenic cavities have been excavated in different epochs for many purposes, including research and collection of potable water, establishment of underground working sites for olive oil production, worship sites, etc. Underground quarries are probably the most diffuse typology of subterranean cavities, especially the largest ones. Originally located at the outskirts of towns, quarries are increasingly found in built up areas due to urban expansion that has characterized the last century. This paper describes the recent occurrence of sinkholes related to underground quarries in the town of Altamura, in the Murge plateau of inland Apulia, where since 2006 a number of sinkholes have formed above subterranean calcarenite quarries, the local rock mostly used for building purposes. These quarries developed below ground because the calcarenite is generally located covered by clays (ranging in thickness from a few to 15 meters). Their abandonment, and the progressive weathering of the rock, has caused failures in the underground quarries. Eventually, such instabilities propagated upward until reaching the surface, and producing sinkholes. Many sinkholes in Altamura have occurred within the constructions. As a result, in 2008 the local Authority established a new building code, requiring detailed geological studies in areas determined to be at risk in order to verify and mitigate any hazardous situations. A great amount of data has been collected in these studies 393

PAGE 411

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEBradano Foredeep (Azzaroli et al., 1968; Iannone & Pieri, al., 1996; Tropeano & Sabato, 2000). The municipality the Bradano River catchment from those of several karst valleys to the east, toward the Adriatic Sea. The location of this topographic feature divides the Basin Authority of Apulia, covering the northern municipality, and the Basin Authority of Basilicata, urban area of Altamura. The town, located at the top of a ridge at 485 m above MSL, rests on a morphological high of tectonic origin in the Cretaceous bedrock. The areas with underground quarries, on the other hand, are located in a topographically depressed sector where limestones are covered by Miocene quarrying activity. In 2008, following some sinkhole events, the Altamura all building owners considered at risk to undertake a detailed geological study to verify, and eventually mitigate, all dangerous situations for both public and private properties. At the same time, the Basin already existing in the areas at risk (Fiore, 2006; Berardi et al., 2010). These regulations included direct and indirect surveys aimed at ascertaining underground conditions, particularly the presence of voids and instability problems.Geological setting and legal constraints sinkhole formation at Altamura consist of underground quarries located at the north eastern and eastern outskirts of town (Figure 2). Their development may be hazardous in this recently urbanized area, where many buildings have been constructed in the last ten years. Only a small number of the buildings date back before 1990, and have been built using into account the underground cavities until sinkholes began to develop. When cavities were found during construction, foundation piles were used to cross the cavities which often produced rock failures within the quarries. The Altamura area rests between the southern portion of 394Figure 1. Location of the study area.Figure 2. NE outskirts of town, and the underground quarries. Red numbers indicate the documented sinkholes (see also Table 1). Figure 3. Geological map of study area.

PAGE 412

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2Sinkhole eventsDocumentation and temporal references have been found for six sinkholes in the Altamura territory (Martimucci et al., 2010; Spilotro et al., 2010; Fiore & Parise, in press). The sinkhole locations are shown in Figure 2, whilst the dates of occurrence are listed in Table 1. analysis of aerial photographs. They are located in the same areas where recent sinkhole events have occurred, standing problem going back at least several decades. Sinkhole no. 1 was registered at locality Chiancone (Via Di Vagno) in 2006 (Figure 4). It occurred near a small building built upon large diameter pilings. The May 2007 sinkhole at Via Barcellona (no.2; Figure 5) greatly increased concern about sinkholes, focusing the attention of both authorities and the general population. It was triggered by the roof collapse of an underground calcarenite quarry. There was 15 m of clay overburden, which sustained a shallow aquifer. The sinkhole, of circular shape and with a diameter of 2 m at the surface, was 24 m deep, and exposed an area where the underground galleries are closely distributed. In 2008, two sinkholes were recorded. No precise date is near Via Copenaghen, and originated due to collapse of a gallery vault which had already experienced previous material of an open clay quarry dropped into an old entrance of the underground calcarenite quarry.395Table 1. List of documented sinkholes. id date locality 1 March 2006 Via Di Vagno 2 May 07, 2007 Via Barcellona 2008 Via Copenaghen 4 Via Fornaci 5 before 1947 Via Fornaci 6 before 1947 Via Praga Figure 4. Sinkhole no. 1, occurred in March 2006 at Via Di Vagno.Figure 5. The sinkhole at Via Barcellona (May 07, 2007; no. 2): this was the main event at Altamura, directly affecting an area of recently-built houses. Figure 6. The sinkhole at Via Copenaghen (no. 3).

PAGE 413

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCECauses of the origin of the sinkholesThe main causes of the sinkholes at Altamura are both natural and anthropogenic. Anthropogenic activities have often accelerated processes that were occurring naturally. On the other hand, recent construction activities in the area where the underground quarries are located has highlighted a phenomenon that otherwise would have been less noticed. Lack of ventilationIn most of the cases, access to the underground quarries water also played a role by transporting and depositing material at the entrances, which in turn produced additional loadings over gallery vaults. Whatever the reason, access closure impeded air circulation in the underground quarries, leading to degradation of the physical and mechanical properties of the calcarenite rock mass. Water percolation and infiltration producing instability in the underground quarries. It is well documented that runoff from the urban area of Altamura is discharged into the old entrances of underground quarries (Figure 7), or in topographically depressed areas. Transport of solid materials by these waters is erosive and may, over time, change the permeability characteristics of the rock mass. This phenomenon is particularly evident along the stretch of Via la Carrera from Via Bari to Via Cassano, and in nearby areas.Deep foundationsPile foundations (Figure 8) have likely contributed to the development of cracks and failures in the calcarenite rock mass; at the same time, they also represent preferential Boreholes focusing runoff waterIn a few cases, boreholes below houses have been found to discharge rainfall runoff collecting in areas around the buildings.Other factorsThere are several other factors that can contribute to sinkhole formation, although generally subordinate to 396 Figure 7. Flooded passage in a calcarenite underground quarry. Figure 8. Foundation piles within the quarries.

PAGE 414

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2the underground spaces (greater than 15 meters) and to the presence of clays (good conductive materials) overlying the calcarenites, the georadar method did not provide good results. Electrical tomography results were much better (Figure 9), especially using systems with 96 electrodes and more than 2000 measurements. Such methods, however, require large spaces that are not always available, as well as particular care during data acquisition along roads with asphalt, waste materials and pipelines. When the necessary space was not available, seismic tomographies in boreholes were carried out using hydrophone chains. This method produced interesting results. Even though it was expensive due to the requirement to drill properly equipped boreholes, the method produces information of greater detail along the the presence of unknown cavities after the reclaiming operations. The large amount of data collected in the last few years has demonstrated that geophysical surveys are useful for obtaining information about the presence of underground cavities. Nevertheless, the possibility of wrong indications or not detecting cavities cannot be excluded. For this reason thus cannot be used to ensure safety in building design. On the other hand, when compared and integrated with data from other sources and methodologies, they may play areas, and to support decisions. the previously discussed factors. Leakage from hydric and sewer systems is common. At several areas the sewers directly enter the underground quarries; it is well known that contact between waste waters and soft rocks such as calcarenites may have serious consequences in terms of reduction in the mechanical properties of the rock mass (Dobereiner & De Freitas, 1986; Coop & amplifying or accelerating the process of degradation in the rock mass. Eventually, failures and falls from the vault progressively reduce the thickness of the overlying stiff calcarenite, and frequently of a size that can damage the piles sustaining the buildings above. Methods of studyThe methods used for the survey at Altamura are similar to those normally used for geologic explorations. Conversely, we stress that the approach followed in this study, consisting of a systematic application of surface and subsurface investigations, can produce detailed understanding of complex geologic phenomena, in exploration. All of the applied methodologies were designed to gain access to the underground quarries. Our started conviction was that the precise survey and location of the above was the main priority, before proceeding with the following phases of study. The surveying activities were especially complex in the urban areas where direct surveys were impossible and indirect surveys were obtained with indirect analyses carried out in boreholes. Indirect surveys electrical tomography, georadar, seismic tomography in boreholes, etc.) are available to identify undergound voids. In the Altamura area, the complex geological setting and Thus, we developed a complex methodology based on those indirect methodologies that, in our opinion, were the most suitable to the studied setting. Due to depth of 397 Figure 9. Example of geophysical survey for the detection of underground cavities.

PAGE 415

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE established aimed at putting all the relevant information layers of the database (Table 2).Caving and topographic surveysCaving activities were mainly designed to explore and the land surface, in order to verify any possible interaction topographic system was established to precisely locate the accesses of the quarries. Starting from this system, the underground network was then mapped together with The caving activities carried out in the territory of Altamura have been crucial for land use planning and control, and in the land management as well. The Register of Underground Cavities has been established at the Altamura Municipality, according to regulations issued by the Basin Authority of Basilicata (Berardi et al., 2010). In particular, surveys carried out by cavers were compared with land use maps, and the zonation in different areas This approach avoids using only geophysical anomalies to locate cavities, which can be inadequate, thus permitting better designs of the stabilization works. Further, survey and graphical representation methods have evolved in the last three years. The documentation on subterranean voids is not simply provided as graphic data, since practitioners and local authorities need extremely precise information to properly work in highly urbanized areas. Therefore, the underground surveys must be accompanied by the surface topographic network.398 Figure 10. Screen views of the GIS system implemented for management of the cavity register.Multi-criteria approach in the definition of the sinkhole spatial hazardIn order to map sinkhole spatial hazards, some criteria of exclusion, repulsion, and attraction (ERA) have been the database. In this way thematic maps were created, to which a weight was then given. Such an approach is not only important for describing and characterizing the territory, but is also aimed at determining those areas that are particularly affected by the phenomena and therefore require particular attention, especially systems.Data management and models of interpretation with the underground quarries (Figure 10), and to highlight the most important data. These include plan views of the cave boundaries, which are basic information for planning any operational activities. The precise locations of narrow passages within the subterranean of underground voids.Data collection and geo-database implementation Using data available at the Technical Office of developed.

PAGE 416

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 carried out using the comparative matrix of coupled informative layers (Saaty, 1980). Through this procedure the following data layers necessary to produce the hazard Presence of cavities Underground failures Cavity depths Land use Elaboration and digital overlaying of the individual data layers have produced the sinkhole spatial hazard map (Figure 11). This map is being used to identify the sectors that most need mitigation measures, taking into account all possible interactions between the spatial characteristics and potential susceptibility. Values with respect to criteria of exclusion, repulsion, and attraction were assigned to each theme with respect to its relative contribution in terms of susceptibility. The coupled themes were put into a comparison matrix, in All the maps are in raster format, and, by assigning them weights, it was possible to sum the pixel values and obtain the sinkhole susceptibility map (Figure 11), with a maximum value of 60 (greatest susceptibility). ConclusionsThe municipality of Altamura, like many towns in Apulia, is underlain by a long underground network of caves that are the cause of sinkhole formation, which threaten the stability of buildings and other infrastructure. Unique to the region, a process of collecting, managing and disseminating data about the underground environment at Altamura has begun, with a goal of obtaining all the information necessary for reclaiming the sites. The data organization is crucial for identifying sites that are most prone to sinkholes. This work provides a starting point to support mitigation measures and appropriate management of a highly urbanized setting. In a region such as Apulia, which is highly affected by sinkhole problems related to anthropogenic cavities (Delle Rose et alii, 2004; Fiore, 2006; Barnaba et al., 2010), and particularly to subterranean quarries (Parise To summarize, the tools used have been: vulnerable areas and develop a zonation based on susceptibility to sinkholes; overlay mapping for spatial analysis. sinkholes are: 1. Presence of cavities; 2. Presence of underground failures; Depth of cavities; 4. Lithotypes; 5. Land use; 6. Surface hydrography; 7. Presence of faults; 8. Seismicity. Some features such as tectonics and seismicity have not been taken into account because they are not available or homogeneous over the whole area.399 thematic groups informative layers topography building cadaster (as of 2011) TERRITORIAL DATA land use lithotypes faults terrace morphology sinkholes endhoreic catchments hydrographic network SURVEYS direct surveys stratigraphies indirect surveys anomalies cave exploration cave mapping recognition of failures excavation direction CONSTRAINTS PAI 2010 cadastral data protocols etc.Table 2. Structure of the thematic groups and related layers.

PAGE 417

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE400 Table 3. Matrix of paired comparison with the weight of the thematic maps (*normalized value on base 1). Figure 11. Extract from the sinkhole susceptibility map.

PAGE 418

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2Fiore A, Martimucci V, Parise M. 2011. Nuove opportunit per la conservazione e valorizzazione delle cavit Iannone A, Pieri P. 1982. Caratteri neotettonici delle Murge. theoretical study of the behavior of a calcarenite in Analysis. New York, John Wiley. Le cave di tufo di Altamura Prime relazioni e Parise M. 2010. The impacts of quarrying in the Apulian karst. In: Carrasco F, La Moreaux JW, Duran Valsero JJ, Andreo B. editors. Advances in research Parise M. 2011. Surface and subsurface karst geomorphology in the Murge (Apulia, southern Parise M. 2012. A present risk from past activities: sinkhole occurrence above underground quarries. Parise M, Lollino P. 2011. A preliminary analysis of failure geodinamico dei caratteri deposizionali e strutturali della Fossa Bradanica nel Pleistocene. Memorie 07.05.2007 di Via Barcellona, Altamura. Il valutazione del rischio connesso (Provincia di Bari, water carbonate systems to forced regressions: the Waltham T, Bell F, Culshaw M. 2005. Sinkholes and subsidence. Karst and cavernous rocks Springer Praxis. & Lollino, 2011; Parise, 2010, 2012), the methodology developed at Altamura provides an example to follow. This is especially true in light of the fact that a new exploitation of the caves, once their stability conditions have been fully ascertained (Fiore et al., 2011).ReferencesAzzaroli A, Perno U, Radina B. 1968. Note illustrative Barnaba F, Caggiano T, Castorani A, Delle Rose M, Di Santo AR, Dragone V, Fiore A, Limoni PP, Parise M, Santaloia F. 2010. Sprofondamenti connessi a cavit antropiche nella regione Puglia. Proc. 2nd interessate da cavit antropiche. Il Catasto delle Cavit Sotterranee (CCS) di Altamura (BA). Un modello litotecnico e di comportamento Delle Rose M, Federico A, Parise M. 2004. Sinkhole genesis and evolution in Apulia, and their interrelations with Fiore A. 2006. Pericolosit geologica connessa alla presenza di cavit sotterranee. Atto di indirizzo Fiore A, Parise M. in press. Cronologia degli eventi di sprofondamento in Puglia, con particolare riferimento 401

PAGE 419


PAGE 420

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2as a rallying point for many environmental issues in Austin. During the early 1990s, at the crescendo of issues surrounding development and the protection of Barton Springs came a call to protect Barton Springs by additional regulations including the Save Our Springs (SOS) Ordinance (Dunn 2007, Smith 2012,). Several years after the SOS ordinance was passed, bonds were proposed to further protect Barton Springs as part of the land over the recharge and contributing zones in fee title or conservation easement.Protecting the LandIn May of 1998 the citizens of Austin voted to support $65 million in bonds that would acquire land including fee title and easements in the Barton Springs contributing and recharge zones to provide for the conservation and to maintain the safety and quality of a part of the bonds, grants and other funds since then have raised the entire contribution toward this goal of land acquisition to approximately $145,000,000. The Water Quality Protection Lands program was created to manage these Fee Simple versus Conservation EasementThe Water Quality Protection Lands (WQPL) Program land would be owned by a private individual, also referred to as fee simple land ownership. In this case the land is owned outright with all rights and obligations intact. On such fee simple lands the City can conduct land management and outreach, provide public access, and perform other activities as needed. Such land also requires the use of City funds to conduct operations and maintenance related to managing and protecting the land, including installing and maintaining fences, vehicle trails, gates and other sundry activities. This land can still be condemned by higher levels of government (county, state, or federal government).AbstractThe Water Quality Protection Lands program was established in 1998 based on a bond proposal passed to protect Barton Springs in the heart of Austin, Texas. Barton Springs is a popular swimming area for citizens and is also home to at least one federally endangered species of salamander. The initial bond called for 6,070 hectares of land to be protected. Land acquisition has hectares at present. Additional cost saving measures such as the use of conservation easements have allowed these dollars to be stretched further. Science has helped guide the acquisition of land into more productive geographic areas (based on recharge) and helped direct quality and quantity. Land management focuses on ecological restoration of vegetation back to native prairie and savanna ecosystems which provide optimal water yield from the land based upon the inverse relationship between woody cover and water yield. These restoration actions combined with proper karst management protects both water quality and water quantity recharging through these lands. IntroductionThe Barton Springs segment of the Edwards Aquifer is a segment of the much larger Edwards Aquifer et al Texas. The aquifer primarily discharges at Barton Springs, which is a collection of four main springs The springs are home to the federally endangered Barton Springs salamander (Eurycea sosorum) and the rare Austin blind salamander (Eurycea waterlooensis), which is a candidate for Federal listing as endangered popular swimming destination for citizens as well RESTORING LAND AND MANAGING KARST TO PROTECT Kevin Thuesen, Ph.D.Wildlands Conservation Division / Water Quality Protection Lands Program, 3621 South FM 620, Austin, Texas, 78738, USA, kevin.thuesen@austintexas.gov403

PAGE 421

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEThe previous assumption that the most proximal creek to of recharge to Barton Springs has been disproven separates water feeding the Barton Springs segment of the Edwards Aquifer to the north and the San Antonio segment of the Edwards Aquifer to the south. Studies and the Barton Springs Edwards Aquifer Conservation rapid from this southern boundary of the recharge zone, travelling up to 11.9 km per day to reach Barton Springs relative to other local watersheds, Onion Creek provides by far the greatest volume of water to the Barton Springs of the total discharge of Barton Springs originating in Springs and near the furthest extent of the recharge zone for Barton Springs. Land Management Owning or otherwise protecting land, such as by conservation easements, provides the greatest measure of protection from impacts such as potential pollutant sources and further allows the natural conditions that feed Barton Springs to continue unimpeded into the cannot curtail the transition or succession of land into ecological states that may produce lower water yields than other ecological states. In the central Texas area grassland and savanna can quickly transition into dense woody canopy following invasion by brush species (Fowler and Simmons 2008). Previously such invasions have been reversed over the evolutionary history of the (Bray 1904, Smeins and Fuhlendorf 1997). The concept of an inverse relationship between woody canopy cover and water yield has been demonstrated in the literature from around the world (Thurow 1998, Wu et al 2001, Le Maitre et al 2002, Davie and Fahey The other mechanism for land ownership is the conservation easement agreement. Under this scenario the City purchases the development rights and other rights that govern the allowed activities on the land in perpetuity. These are always made with willing buyers as are all real estate transactions related to the private property rights required by these conservation easements is the amount of impervious cover allowed on the land (usually between 1 to 2 percent of the net site area). In addition, such easements also have provisions restricting the use of certain pesticides, limits on stocking rates of livestock, a requirement to manage brush on the property and other restrictions. Such conservation easements cost the City about 50 percent of the real value of the land. Further, such lands require no outlay of City funds for operations and maintenance of the land, as these are borne by the annually by WQPL staff to confirm compliance with the easement and provide technical assistance as requested. Occasional legal assistance is also needed to administer this work. hectares protected by conservation easements. These purchases have resulted in protecting over 22 percent of the Barton Springs recharge zone and seven percent of the Barton Springs contributing zone. Figure 1 shows the location and type of land holdings and their locations relative to the contributing or recharge zones.Karst Science Enabling Counterintuitive PurchasesThe purchase of these lands includes a variety of factors that determine the acquisition priority of each potential property. Most relevant of these for this paper, but by no means the only priority, is the karst science that has led to relatively counterintuitive acquisitions of property far from Barton Springs. As shown in Figure 2, the Onion Creek watershed has five different watersheds separating it from where Barton Springs discharges prior to reaching the Colorado River. Yet, the WQPL Program has made significant purchases in this watershed404

PAGE 422

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2Center 2010). These are the same ecosystems that the literature has demonstrated yield the greatest quantity of water. Work conducted in this regard utilizes a number of tools to manage brush and encourage grass restoration, including mechanical thinning, prescribed fire and native grass seeding. The work is conducted to be as low impact as possible to avoid erosion and other negative consequences on the land. Balancing water quality and water quantity can be challenging and at times counterproductive, but again the literature has indicated improved water quality under grassland settings compared to other ecological states (Banta and Slattery 2011). In the case discussed herein, the restoration of native grasslands and savanna ecosystems in the recharge and contributing zones has the potential to further protect or even improve water quantity and water quality at Barton Springs. Further, various studies from Texas have shown additional water yield following brush management 2011). This has not been without controversy (Wilcox et al 2005, Wilcox et al 2008, Wilcox and most ideal for brush management from a water yield standpoint are well represented on the recharge zone lands protected by the WQPL Program: that is, a shallow soil overlaying a highly fractured subsurface where water can quickly be transported underground (Wilcox et al 2006). The WQPL Program conducts ecological restoration activities on land held in fee simple to restore the ecosystems back to or maintain their native ecological states of grasslands and savannas (Land Management 405 Figure 1. Map of land protected by the Water Quality Protection Lands program as of October 2012.

PAGE 423

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEStreams over the recharge zone in central Texas are frequently ephemeral in nature and under such conditions management of karst features in streams frequently has the highest potential for recharging the largest volume of water over the longest time and accordingly receives the bulk of attention on the WQPL. As a case in point, one Swallets can have their function impaired by their success in capturing water as this process also brings in substantial volumes of organic matter, sediment and rocks included in the bed load of the streams in which they are located. Over time this debris can plug swallets and negatively impact their function. Over a period of geologic time, such features are likely to close and open Karst ManagementOnce the land is protected and opportunities for optimizing the quantity and quality of water are implemented by land management, the last integral action is to protect the function of karst features. Locating and identifying karst followed up with prioritizing features in terms of potential to transmit water. Logically, features located in streams beds, such as swallets, would rise above typical upland 2005), but these upland features should not be discounted. For example the WQPL has at least two upland features with internal drainage basins approaching 24 hectares each. Such internal drainage basins can recharge up to 42 et al 2005). A swallet by comparison may have a drainage basin measured in square kilometers. That said, a swallet is unlikely to be able to transmit this total volume due to 406Figure 2. Map of watersheds in the area protected by the Water Quality Protection lands.

PAGE 424

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2necessary to wait for a dry period to enter the caves and remove any debris plugs from deeper inside the feature. These swallets likely owe their origin to dissolution by Onion Creek, as they have a strong vertical that caves carrying water through the vadose zone tend to stair step (i.e. have vertical drops), whereas caves formed at the water table tend to have a strong horizontal component. The humanly explored vertical depths of these features are relatively shallow, reaching at most only 9 to 10 meters as creek alluvial infill is excavated. Most of these swallets become constricted and horizontal in nature at the current limits of human exploration. such areas to positively impact the quality and quantity of water reaching a spring on a human time scale, steps must be taken to keep the function of existing swallets in proper functioning condition rather than waiting for formation of new swallets. This is even more of an acute need when additional demands are made on an aquifer without any offsetting decreases in usage or increases in recharge. The WQPL Program uses a variety of simple techniques to manage such features to maintain their function. Once a swallet is located, it is evaluated to help determine its recharge, a grate will be installed above it to help prevent debris from collecting within the swallet. debris covers attached externally to these grates. Such debris covers are structurally weak, but are supported by the initial grate and removable without affecting 5). The blinding of the grate then keeps the sediment through the grate. Naturally, this also prevents a large amount of water from reaching the feature, however, passed and allow the cleaner portion of the stream of such features deep within the swallet such that maintenance of the grates on the surface is usually conditions. Prior to the use of these grates it would be 407Figure 5. Example of swallet grate with fine debris cover after storm event and prior to manual cleaning.Figure 4. Example of swallet grate with fine debris cover.Figure 3. Photo of a swallet recharging on Onion Creek.

PAGE 425

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEthose associated with ecological restoration are used to restore or maintain the vegetation as native grasslands and savannas, which have been shown to yield greater water than more woody landscapes. Finally, to ensure that water recharging off these lands can continue to swallets, are managed and restored to proper functioning condition and protected from sedimentation that could impede or obstruct recharge. ReferencesBanta JR, Slattery RN. 2011. Effects of brush management on the hydrologic budget and water Bray WL. 1904. The timber of the Edwards Plateau of Texas; its relation to climate, water supply and soil. USDA Bureau of Forestry Bulletin No. 49. Aquifer Conservation District and the City of Austin. Unpublished summary report. City of Austin. 1998. 1998 May Bound Elections. brochure. Davie T, Fahey B. 2005. Forestry and water yield current knowledge and further work. New Zealand Dugas WA, Wright P. 1998 Effects of removal of Juniperus Ashei on evapotranspiration and runoff in the Seco Creek watershed. Water Resources Dunn, L. 2007. The Unforeseen. New Yorker Films Fowler NL, Simmons MT. 2008. Savanna dynamics in Central Texas: Just succession? Applied Vegetation Resources Assessment 2005. Rome. Food and Agricultural Organization of the United Nations. FAO Forestry Paper 155. Segment of the Edwards Aquifer, southern Travis District and City of Austin Watershed Protection Department. 110p and appendices. Once grated, some swallets can then be excavated to remove accumulated sediment with very little accumulation of new sediment. This can allow the or movement of old sediment. Few terrestrial organisms survive the periodic and occasionally long lasting inundations, but contractors doing such excavations are required to have U.S. Fish and Wildlife Service permits for working with endangered karst invertebrates. In one example of this sort of excavation, a former landowner who was raised on the property, likely around the 1950s or 1960s, reported a frequent whirlpool originating at a known swallet. No whirlpool had been dye tracing showed it had a much longer travel time to Barton Springs than did a nearby feature also on born sediment might be preventing this feature from this sediment in combination with the addition of grates will return this swallet to proper functioning condition. sediment and debris has been removed to date. Upland features require much less attention, as they are frequently less prone to becoming plugged by debris, and proper functioning condition is maintained in these features by vegetation management that helps reduce erosion into the features. Frequently, upland features may also be home to karst invertebrates that could be endangered and may require the cave be protected as a refuge for such organisms versus being managed as a recharge feature.ConclusionsThe Water Quality Protection Lands were established namely Barton Springs. The methods of this protection began with the purchase of land and the protection of additional land with conservation easements leading to The WQPL Program went further and is implementing a land management plan to manage the land owned in fee simple to optimize the quality and quantity of water leaving the lands and recharging into the Barton Springs segment of the Edwards Aquifer. Techniques including 408

PAGE 426

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 brush control on evapotranspiration in the North Concho River watershed using the eddy covariance technique. Journal of Soil and Water Conservation Smeins FE, Fuhlendorf SD. 1997. Biology and ecology Agricultural Experiment Station, Texas A&M University System. Smith A. 2012. The SOS Ordinance turns 20: political Environment. The Austin Chronicle, Thurow TL. 1998. Assessment of brush management as a strategy for enhancing water yield. Proceedings of the 25th Water for Texas Conference :Water Planning Strategies for Senate Bill 1. Texas Water Resources Institute. IN: Taylor CA (ed) 1997 Juniper Symposium, 1997 Texas Agricultural Experiment Station, Texas A&M University System. Terrains. Oxford University Press, Oxford. 464 p. Wilcox BP, Owens MK, Knight RW, Lyons RK. 2005 karst rangelands? Ecological Applications paradox: Rivers rebound as degraded grasslands Sharp JM. 2004b. Flow systems of the Edwards Aquifer Barton Springs Segment interpreted from Water Resources In Central Texas,Retrospective And Prospective Symposium Proceedings, San and water balance of internal drainage sinkholes. In: Beck B. ed: Proceedings of the Ninth Multidisciplinary Conference Sinkholes and the within the Barton Springs Segment of the Edwards Aquifer, Southern Travis County and Northern contributions. City of Austin short report (in progress). 2006. Springs on rangeland: Runoff dynamics Johns D. 2005. Dye tracing recharge features under segment of the Edwards Aquifer, hays County, Summary of 2005 groundwater dye tracing, Barton Springs segment of the Edwards Aquifer, hays and Travis Counties, Central Texas. Barton Springs Edwards Aquifer Conservation District Report of land management for the Water Quality Protection Lands, Austin, Texas. 89 p. Recommended land management for the Water Quality Protection Lands, Austin, Texas. 110 p. C, Chapman RA, Nel JA. 2002. Invasive alien trees and water resources in South Africa: Case Mark AF, Dickinson KJM. 2008. Maximizing water Zealand perspective. Frontiers in Ecology and the 409

PAGE 427


PAGE 428

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 of the underlying karst bedrock and associated karst aquifer, Jo Daviess County. The crop lines were consistent with the drainage pattern. Stream patterns like these are well known patterns in the top of bedrock. The lines appear to have been coincide with the occurrence of karst features throughout eastern Jo Daviess County. The pattern observed in the crop lines closely mimics widths and extent of the lines may be used as a surrogate for the karst features present on the bedrock surfaces. in bedrock exposures, yield a three dimensional view could provide a more complete and accurate model of the karst aquifer in the study area and similar karst areas in the Midwestern United States and perhaps in other karst regions of the world. IntroductionCarbonate rock at or near the surface are fractured and typically creviced due to solution enlargement to the point AbstractThe persistent drought of the 2012 summer in the outcome of the extreme drought was that it provided a rare opportunity to examine and characterize the bedrock surface and underlying karst aquifer within the Driftless Area of northwestern Illinois. Complex networks of vegetated lines and polygonal patterns, herein referred to as crop lines, crisscrossed the dry summer landscape of Jo Daviess County. Initially, the crop lines were examined and photographed using a handheld digital camera on the horizontal separations of the lines were measured. Crop lines and their patterns and orientations were compared with those of crevices in outcrops, road cuts and quarries, and with lineaments seen in LiDAR elevation data of Jo Daviess County. extent, soybeans and corn, the crop lines are the result of a combination of extremely dry conditions, and a Dolomite bedrock. The plants forming the lines tend to grow denser, taller (0.5 m vs the deepest of the aforementioned crops extending up soil moisture within the vadose zone present within bedrock fractures and crevices provide the necessary moisture to sustain the overlying healthy plants, while sparse plant growth. TOOL FOR CHARACTERIZATION OF KARST TERRAIN Samuel V. PannoIllinois State Geological Survey, 615 E. Peabody Drive, Champaign, Illinois 61821, USA, s-panno@illinois.eduDonald E. LumanIllinois State Geological Survey, 615 E. Peabody Drive, Champaign, Illinois 61821, USA, luman@illinois.eduWalton R. KellyMatthew B. AlschulerPO Box 325, Warren, Illinois 61087, USA, matthew@cottonexpressions.com411

PAGE 429

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE(Figure 2) occurred in the thin soils of the Driftless Area overlying the carbonate bedrock of the eastern two thirds of Jo Daviess County, the western portion of Stephenson County and well into southern Wisconsin. the crop lines observed in eastern Jo Daviess County, Illinois, and assess their usefulness as indicators of with a karst aquifer. Because the crop lines appear to accurately mimic the creviced surface of the underlying karst bedrock, we explored their usefulness in identifying the extent, character and geometry of the underlying karst aquifer.MethodsThe authors were contacted by a Jo Daviess County resident on July 16, 2012 concerning the appearance of crop lines in the Driftless Area of northwestern Illinois. We conducted a reconnaissance trip to eastern Jo Daviess County on July 19 during which time we examined and photographed abundant crop lines on the ground, as well The initial photography was used to secure funding for additional trips to the site. where the rock body constitutes a karst aquifer (Quinlan et al. 1991). These fractures and crevices are usually fractures and crevices in the bedrock are only observed in excavations, road cuts, quarries, outcrops and rarely where the soil and sediment overlying the bedrock where soils are thin and during extremely dry periods and where crops are planted, lines in the vegetation (hereafter referred to as crop lines) have been observed in the Wisconsin Driftless Area (Maureen Muldoon, University of Wisconsin, personal communications, 2011). It is of the underlying bedrock carbonate aquifer. Unfortunately, the reporting of these occurrences is drought conditions in Illinois and in the surrounding states in the 2012 summer created a rare situation that has resulted in the formation of vivid lines and patterns in crops in Jo Daviess (Figure 1) and surrounding counties 412 Figure 1. Generalized geologic map of Jo Daviess County (modified from McGarry 2000).

PAGE 430

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2photography. The widths, orientations and spacing of the crop lines were documented and compared with those of crevices and fractures in exposures, and with lineaments seen in LiDAR elevation data of Jo Daviess County examined and compared with information gleaned from existing bedrock exposures as described by Bradbury Results and Discussion Geology of Jo Daviess CountyJo Daviess County (Figure 1) lies within the Driftless Area of northwestern Illinois; the county lacks glacial drift that covers the bedrock of most of the upper thin remnants of the Ordovician age Maquoketa shale, conducted one month later to document additional occurrences and estimate the geographic extent of the crop lines in Jo Daviess County. Based upon this information, a request was made to the Illinois Department of Transportation, Aerial Survey Division to acquire vertical aerial photography for 16 selected sites issues, aerial photographs were not taken until August 28 and 29. Soon after the IDOT aerial photography acquisition, the area experienced increased rainfall and harvesting began earlier due to the drought. The combination of these two factors resulted in the crop line features disappearing or being eliminated. probe and a tape measure. Vegetation that created the crop lines was examined on the ground and from aerial 413Figure 2. Alfalfa fields reveal the geometry and character of the underlying karstified Galena Dolomite. Crevice-controlled rectangular stream patterns are also apparent. This oblique aerial photograph was taken 4.0 km south of Warren, IL in Jo Daviess County.

PAGE 431

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEto explain the origin of these deposits (Rowan and de Marsily 2001). Ore mineralization and dolomitization of been dated by several techniques as Early Permian in age ranging between 270 and 280 Ma (Brannon et al. 1992; Pannalal et al. 2004). Recent work by Panno et al. (in review) expanded on our understanding of the karst terrain of Jo Daviess County of the LiDAR elevation data revealed numerous been used since the early 1950s for oil, gas, and mineral exploration. Lattman and Parizek (1964) and Parizek (1976) extended this work to groundwater resources Lattman and Parizek (1964) state that Fracture traces and lineaments appear to be universal in their distribution and will have their greatest utility in rocks where secondary permeability and porosity dominate and where intergranular characteristics combine with water and groundwater movement. The lineaments in Jo Daviess County consisted of the alignment of study area. Many stream valleys are linear, while others have sharply angular meanders. These angular features or rectangular patterns are classic geomorphologic indicators and strongly suggest bedrock control of the streams in the study area (Figure 2). The lineaments in the study area have three distinct trends; approximately N20oW, N70oW, and N70oE. Examination of sinkholes in eastern Jo Daviess County revealed individual sinkholes and en echelon sinkholes, all of which are coincident with lineaments. Examination of mapped lineaments in far eastern Jo Daviess County revealed that every lineament had one or two sinkholes in close proximity to one another and to the streams. Sinkholes in this area were typically 0.5 m to greater than 1 m deep and about 1 m in diameter. Larger sinkholes up to 10 m in diameter were also found. Features that were often found associated with the sinkholes were the stream bank that extended 1.6 to 5 m into the much of the highlands in the county. Tectonic compression and extension occurred in this area during and following the formation of the Wisconsin continued to be active in late Silurian or Devonian time part, separates the Illinois Basin to the south from the Michigan Basin to the east. The Mississippi River Arch separates the Illinois Basin from the Forest City Basin to the west. Jo Daviess County lies on the southwestern As a result of the compression and extension, bedrock the rock formation in the district [most of Jo Daviess o W, and two oo E and N 20oo W. Bradbury (1959) found that crevice orientations in numerous exposures in far eastern Jo Daviess County trend N 85oooo W. mineralization in this area. The geology of the Upper Daviess County and extends into Iowa and Wisconsin, crevices or in solution cavities in carbonate rocks of (PbS2) was the main ore mineral in these deposits, and sphalerite (ZnS2) was the most abundant ore mineral associated with bedding planes and reverse faults and zinc in solution were implicated as the source of the mineralization by various geochemical and isotopic forming solutions originating from evaporative brines associated with the Reelfoot rift system (late Paleozoic time) is one of the more recent hypotheses proposed 414

PAGE 432

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2Crop LinesThe 2012 summer drought that affected the health and vigor of crops of the Midwestern United States and thin soils provided an opportunity to view the geology of the underlying carbonate bedrock in the Driftless Area of Illinois. Less than 8 m of soil and unconsolidated materials overlie the fractured and creviced carbonate bedrock of Jo Daviess County, northwestern Illinois ranged from 0.6 to 1.2 m. Soil thicknesses immediately over the lines were typically greater than 1.5 m. Excavation of the lines revealed that many contained clays that were visually and texturally identical to the weathering product of Maquoketa Shale. The depth of few crevices examined. The wetness and presence of water within the clays suggested that water was moving through the crevices, perhaps along piping channels. During the 2012summer, complex networks of dark green vegetation separated by gray patches of nearly bare bank. In addition, the stream orientation tended to shift abruptly and follow the trend of the lineament where they crossed streams instead of having the more typical gently curved meanders. These angular stream features had similar orientations to the mapped lineaments and reflect bedrock control. On the basis of these data, the lineaments identified on the LiDAR maps are interpreted to be a reflection of open crevices in the underlying carbonate bedrock aquifer that are transmitting groundwater. Examination of the geomorphology of the stream valleys suggests that the lineaments are linear depressions that formed probably form where overlying sediment becomes thinner with proximity to stream valleys where carbonate bedrock is exposed in the stream floor. Lineaments in carbonate rock terrain have been found to be indicative of zones of enhanced well yields; productivity wells in carbonate terrain has had a success rate of 75 to 80% (R. Parizek, Pennsylvania State University, personal communications, 2009).415Figure 3. LiDAR shaded relief image of eastern Jo Daviess County showing lineaments and their orientations (from Panno et al. in review).

PAGE 433

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE rooted perennial. Upon germination, a strong taproot develops rapidly and penetrates almost vertically downward. It often reaches a depth of 1.5 to 1.8 m the and may ultimately extend to depths of 6.1 m or more. It is notably a deep feeder. Consequently, the alfalfa bedrock crevices near the top of the karst aquifer. Corn ground were observed. These lines of vegetation, often forming complex polygonal patterns with dominant sown with alfalfa, and to a lesser extent, soy beans and corn. The alfalfa forming the lines tended to grow more densely, taller (0.5 m vs on relatively steep slopes to retard soil erosion. These in September 2012 (Figure 4).416Figure 4. Google Earth imagery showing the complex geometry of crop lines in an alfalfa field.

PAGE 434

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2crevices. Preliminary results indicate that the crop lines are data in the same area by Panno et al. (in review). The trends springs) throughout eastern Jo Daviess County. and soybean roots tend to have shallower root systems. Corn is a more shallow rooted plant reaching depths of 1.2 m or more. Less than 10% of the water taken up by the corn plant is acquired below 1 m (McWilliams et al. 2004). Soybean plant roots extend to a depth of 1.2 to 2.4 m of soil (North Dakota State University 1997). The bedrock fractures and crevices provide the necessary moisture to sustain the overlying healthy plants, while sparse plant growth. In aerial view, the crop lines create rectangular drainage patterns. Rectangular stream patterns like these are well known and due to control Streams tend to follow the paths of least resistance. The orientation of the stream channel is consistent with the orientation of crop lines and the straight reaches of the stream follow trends seen in crevices exposed in outcrops, road cuts and quarries (Figure 5). This and the similarity of crop lines with lineament patterns in LiDAR elevation data, and crevice occurrence in outcrops, road cuts, and quarries found throughout eastern Jo Daviess County, western Stephenson County, and northern Carroll County indicate that the lines and of the creviced pattern of the underlying karst aquifer. As crevices of the underlying carbonate bedrock and reveal of Jo Daviess County. Finally, desiccation cracks are almost exclusively located along the crop lines and follow their trends (Figure 6). lines where vegetation was stunted and typically about 6 inches high. This phenomenon is probably due to fact that the shallow roots of the crops tend to take up moisture from the soil and create desiccation cracks, but only in those areas were vegetation is thriving.Ongoing ResearchThe authors are currently georeferencing and analyzing sixteen sites in eastern Jo Daviess County, and their 417 Figure 5. Road cut through Galena Dolomite and located along Rt. 20 near Elizabeth, IL in Jo Daviess Co. The solution-enlarged crevices are oriented N-S and are 1 m wide (from Panno et al. in review). Figure 6. Desiccation cracks within and following the trends of the crop lines.

PAGE 435

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE exposed in outcrops, road cuts, and quarries, and of lineaments seen in LiDAR elevation data of the study area. Crop lines, combined with other data from outcrops, road cuts, quarries and LiDAR elevation data, may be used to identify and characterize the degree and extent Area on northwestern Illinois, and other karst areas of similar geology. Further research currently is underway. AcknowledgementsWe thank Jeff Kromer of Mount Carroll, IL for the generous giving of his time, skills and aircraft while in order to acquire reconnaissance aerial photography. Finally, we thank Amy Eller of the Illinois Department of Transportation for contributing the Aerial Survey Division resources to acquire the vertical aerial photography in a timely manner.References Survey Report of Investigation 210. 49 p. Brannon, J.C., F.A. Polosek and R.K. McLimans. 1992. Alleghenian age of the Upper Mississippi Valley In Geology of Illinois (D.R. Kolata and C.K. Nimz, Jr. 1959. The geology of the upper Mississippi U.S. Geological Survey Professional Paper between fracture traces and the occurrence of Soybean growth and management quick guide. from: Nelson, J.W. 1995. Structural features in Illinois. Illinois North Dakota State University. 1997. Corn production Available from: Based on the presence and orientations of the crop lines in road cuts, quarries, and outcrops, and lineaments seen in LiDAR elevation data, it is clear that the patterns of the bedrock surface. The widths and extent of the lines may be used as a surrogate for the karst features present on the bedrock surfaces. Crop lines, coupled with bedrock exposures, yield orientations and spacing carbonate bedrock. Taken together, these features yield system, and ultimately could provide a more complete and accurate model of the karst aquifer in the study area and similar karst areas throughout the world. ConclusionsAn investigation of crop lines that appeared during the extreme drought of the 2012 summer in the Midwestern United States may provide an additional method for investigating and characterizing karst terrains. The crop lines were noticed by local farmers primarily in alfalfa Within days, the crop lines on the ground and from aerial photography. The widths, orientations, and distance of separation of the lines were examined and compared with crevices in outcrops, road cuts and quarries, as well as with lineaments observed in LiDAR elevation data. and composition of the materials within the crevices. The crop lines formed in very shallow soils (less than about 1.5 m thick) and are of similar orientation as those cuts, and quarries, and of lineaments seen in LiDAR beneath the thin soils contained clays similar to those of weathered Maquoketa shale. It is likely that these clays clay suggests there is piping and movement of recharge and groundwater through the clay. We conclude that the crop lines may be used as a the carbonate bedrock surface. The crop lines are of 418

PAGE 436

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 Pannalal, S.J., D.T.A. Symons and D.F. Sangster. 2004. Paleomagnetic dating of Upper Mississippi Valley fracture traces and lineaments in carbonate and Publications, Fort Collins, CO. Jr., A.J. Edwards, A.R. Smith. 1991. Recommended aquifers, practical evaluation of vulnerability of karst aquifers, and determination of optimum sampling frequency at springs. Proceedings of thickness of Quaternary deposits, Jo Daviess Weibel, C.P. and Panno, S.V. 1997. Karst terrains and Survey, Illinois Map Series 8, 1:500,000 Scale.419

PAGE 437


PAGE 438

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2portions of South Knox County in East Tennessee. The proposed corridor is located in a section of the Valley and Ridge Province of East Tennessee where several ridges and valleys will be crossed, as well as creeks, roads, subdivisions and rural lands (Figure 1). The proposed corridor connects the current terminus of SR 71 at Moody Ave. in South Knoxville to John Sevier In an attempt to properly evaluate the potential for effort, a karst map and a map of one cave was completed.General Geology of Proposed CorridorThe proposed corridor is situated in the rolling to hilly topography of the Valley and Ridge Province of East Tennessee and includes a portion of South Knox in terrain underlain by folded and faulted sedimentary limestone, shale, sandstone, and siltstone. The geologic formations that underlie the study area and include the Lenoir FormationOl (argillaceous crystalline limestone), Chapman Ridge FormationOcr (sandstone), and the Ottosee FormationOo (shale, siltstone, and limestone). These formations have geologically normal conformable stratigraphic contacts In general terms, the bedrock has a northeast to southwest trending strike with bedding that dips to the southeast. In addition, there is a southwest plunging anticline, known as the Rocky Valley anticline, located within the center AbstractThe proposed extension of SR 71 (South Knoxville Boulevard) located in Knoxville, Tennessee necessitated a preliminary geologic evaluation of the corridor which was under consideration by the Tennessee Department of Transportation. A geohazards review disclosed the presence of extensive karst terrain located within the corridor being considered. A number of caves were also found during the investigation. The proposed routes within the study corridor were found to cross a series of very large multiple hectare (acre) sinkholes. In addition, a biological investigation of the route revealed the presence of a rare and endangered species of cave salamander called the Berry Cave Salamander. The geologic and geotechnical investigation resulted in the development of a surface karst map of the study corridor. Sinkholes and cave entrances were located and a generalized karst boundary was established. In addition, a survey map of the Meades Quarry Cave was made which provided supporting quantitative data in connecting surface sinkholes and the cave containing the endangered Berry Cave Salamander. A recommendation of the study was a dye tracing of the suspect sinkholes. If be required to mitigate the effects of the highway run off in the karst terrain affecting the Berry Cave Salamander.IntroductionLarge areas of East Tennessee are underlain by carbonate rocks such as limestone and dolostone with some estimates of as much as 50% to 60% of all underlying rock types being carbonates. As a result, these areas are resulting karst landform development, such as sinkholes and cave systems. The Tennessee Department of Transportation (TDOT) is planning a new roadway alignment (SR 71) that crosses MAPPING SURFACE AND SUBSURFACE KARST GEOHAZARDS FOR HIGHWAY PROJECTS: Harry L. MooreGolder Associates/Tennessee Dept. of Transportation Retired, 3730 Chamblee Tucker Rd., Atlanta, Georgia 3034, 865-603-7652, HLMoore@golder.com421

PAGE 439

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE422Figure 1. General location map of the SR 71 study area in Knoxville, Tenn. Figure 2. Location of SR 71 study area in South Knoxville; Noted are Karst Area 1 and 2 and Meades Quarry Cave locations.

PAGE 440

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2The karst landscape of the study corridor as well as the rest of East Tennessee is characterized by features such as sinkholes, caves and cave entrances, sinking streams and outcroppings of weathered carbonate rock (limestone and dolostone). There have been incidences of basins crossed by roads or buildings. The recognition of areas of active karst subsidence and collapse is of considerable importance to the design and construction of the proposed highway. TDOT initiated a karst geohazards inventory and assessment to assist in evaluating and selecting a satisfactory alignment within the corridor study area. This study was designed to locate caves and sinkholes In addition, the caves were visited and preliminarily evaluated as to their geotechnical and environmental importance. The caves that were found to be of surveyed to determine their lateral and vertical extent. Mapping of the caves was performed by both TDOT Caves and sinkholes offer special challenges relative to both the physical and environmental issues of highway development. The karst geohazards inventory study disclosed the presence of both sinkholes and caves, some of which may have detrimental structural, geologic, and environmental issues for the roadway alignment and grade design considerations. A total of eight caves as well as numerous sinkholes were located within the proposed roadway corridor. Two of the caves, Cruze Cave and the Meades Quarry Cave For reasons of sensitivity concerning cave conservation and landowner privacy, the cave locations are not reference when needed. Of recent note, Meades Quarry Center Park to control access to the cave system. of the study area and extends from near the French of the Tennessee River) to the southwest part of Knox County. Due to the anticline structure, some strata located on the northwest limb of the structure dip to the northwest and form a limb of a small syncline, the The rock units are structurally oriented in a northeast to southwest position with the rock layers (bedding) dipping generally to the southeast at varying angles. Numerous fractures and other discontinuities are found throughout the strata along the study corridor and are the result of the intense folding and faulting history of the region. Weathering has also produced knobby terrain and classic karst topography characterized by sinkholes, disappearing streams, and caves. Notable caves found in the study area include Cruze Cave and Meades Quarry Cave.Geohazards Evaluation landslides or areas prone to landslides, the presence of acid producing rock, soft or weak soil, old abandoned mine sites, contaminated ground, caves and sinkhole areas (referred to collectively as karst), and earthquake prone areas. With the exception of karst, abandoned quarry pits, and steep terrain, geologic hazards do not pose an immediate threat to the environment and public along the study corridor is karst, which is located in belts or zones that 423Figure 3. Geologic Map of the study area; Ol Lenoir Fm.; OhHolston Fm.; OoOttosee Fm.

PAGE 441

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE that they were there. Afterwards, the sinkholes and other karst features were boundaries were drawn to encircle these areas. Typically, cave openings. Actual cave entrances were not plotted private owner protection. The geohazard areas were then expressed as outlined patterns on topographic maps to better illustrate the geohazard relative to the surrounding landscape. In addition, the proposed corridor route was overlain on the geohazard map. This map is then used by the roadway In some cases, maps with contour intervals as small as Efforts to avoid karst areas with roads and other good land is already mostly developed, leaving only the geologically undesirable land for current development. highway corridor includes such areas as karst terrain and steep, very rocky terrain. In planning new highway corridors certain constraints Engineering Section mapped the karst features in an effort to identify geohazard areas. In general, 7.5 minute topographic maps (that use a 6 meter (20 foot) contour interval) were used to locate karst features such as sinkholes (closed depressions on the contour map), caves, springs, and sinking streams. Once 424 Figure 4. This sinkhole map was prepared for the geohazard study for the SR 71 project; yellow areas are sinkholes that were found based on a 20-foot contour interval.Mapping Karst Areas within the Proposed Study Corridor

PAGE 442

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2to the proposed roadway. It is anticipated that mapping of cave passages will become increasingly mandatory as society continues encroaching onto and into the karst environment. investigations, the SR 71 Extension corridor was found to be located within several strike belts of karst. Numerous sinkholes and caves were found to be located in these karst areas. The most intensive sinkhole areas were mapped in order to better assess the corridor terrain made in order to locate as many sinkholes as possible. outside of the sinkhole zones (outlined on the attached The possible impacts from the proposed road alignment (or any other structure, building, subdivision, etc.) on the karst environment include sinkhole collapse, sinkhole effects on the cave and subsurface dwelling wildlife. this issue in the karst regions of the Valley and Ridge of on the ground than were depicted on selected 7.5 minute 7.5 minute quadrangle with a 6 meter (20 foot) contour interval that revealed 55 sinkhole features, while ground In addition to the surface mapping of sinkholes, it is becoming increasingly important to locate (step two in the geohazard study) and map the caves(step three in the study of karst geohazards) in close proximity to the proposed roadway. By knowing spatial locations of the cave passages, a more accurate design of proposed roadway cut slopes can be made. This prevents the unnecessary opening of a cave system to the surface, thus protecting the cave biota, such as bat colonies and salamanders. The use of experienced cavers in combination with engineering survey crews provided the best results for locating the cave passages spatially with respect 425 Figure 5. Karst Map of study area (dashed line); blue is active/intensive sinkhole area, red is area of caves, downtown Knoxville is pink area in upper part of map.

PAGE 443

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEKarst Area 2 is located near the center of the study area and encompasses the heavily populated communities Numerous sinkholes are found in this area, some of area and are locally known as Brown Cave, Backyard Cave, and Cruze Cave. Formation and Lenoir Formation) have karst features that are prevalent across much of the area. The area that is underlain by the Lenoir Formation tends to have relatively moderate to low relief with thin soil cover underlie high ridges and hilly terrain, and produces a derived soil is variable in thickness, but can attain thicknesses up to 15 to 18 meters (50 to 60 feet). Cruze Cave was visited by the author. The cave is formations and tends to follow the dip of the bedrock as it plunges downward with the structure of the geologic anticline. The roof of the cave along with some of the upper passages were found to be developed in the drainage conduit is developed in the Lenoir Formation. communities. The following is a brief discussion of these two karst areas. Karst Area 1Karst area 1 is located in the northern part of the study area ( see Figure 2) and is developed along the bedrock the Lenoir Formation. Numerous unique formations and biota, including several species of bats and salamanders, are found in many of the caves in this area. Tennessee Marble in the historic past, as evidenced by several old abandoned quarry pits in the Meades Quarry operations have exposed a cave system and underground stream in several areas of the abandoned quarry pits. Five openings into the cave system were found in the old pit areas. The cave is known as Meades Quarry Cave. area. Entrance into the cave is very limited due to a spring that issues out of the submerged entrance. Several other large sinkholes are also found in this area, mostly A third cave, Cave Spring Caves, is found along the Park. These short double caves, approximately 24 meters (80 feet) in length), are developed in sandstone strata of the Chapman Ridge Formation. Meades Quarry Cave southwest of the Tennessee River. The karst in this area is characterized as being well developed with numerous sinkholes, open solution channels, and solution cavities located well above the zone of saturation ( vadose zone). Naturally developed bedrock which is characteristically carbonates, shale, and sandstone, are also tilted and pinnacled. 426Karst Area 2 in the Lake Forest/ Moreland Heights Communities Mapping Subsurface Karst: Meades Quarry Cave Mapping InitiativeDue to the existence of Meades Quarry Cave within the proposed study area, it was decided to map the of mapping the cave was to determine if the general trend of the cave and cave stream is toward the sinkhole area around Old Sevierville Pike and Red Bud Drive, which is located within the study area of the proposed parkway extension. The presence of the Berry Cave Salamander ( Gyrinophilus gulolineatus, a subterranean amphibian listed as a potential threatened species see Figure 6) in the Meades Quarry Cave stream has made the cave an important issue with respect to the proposed SR 71 extension. The Berry Cave Salamander derives its

PAGE 444

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2From our mapping effort it was determined that the cave trends in an azimuthal direction of 248 degrees (or S 58 W from the cave entrance), which is toward the numerous sinkholes found along Old Sevierville Pike. surveyed (Figure 9 and Figure 10). cave stream in which the salamander lives. Sinkhole the Meades Quarry Cave system. This section has its extensive mud deposits in the cave forced us to switch to using a hypsometer. for the cave passages. The mapping was performed on passage The cave is developed in the vadose zone and of the cave contained numerous dripstone formations (speleothems), some of which were a pure white color (Figure 8). Past wastes from a lime kiln operation at the quarry may have contributed to this white character. inches) thick were found on the stream bed and most exposed rock surfaces. 427 Figure 6. This photo shows the Berry Cave Salamander (Gyrinophilus gulolineatus) as found in Meades Quarry Cave by University of Tennessee researcher Matt Niemiller (photo by Niemiller from: ). Figure 7. Man-made entrance to Meades Quarry Cave. Figure 8. Unusual white soda straw formations in Meades Quarry Cave.

PAGE 445

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEConclusionsThe result of this paper is to show how mapping karst geohazards can be effective in geotechnical studies of proposed roadways. Three main steps are involved in cave passages suspected to be impacted by the proposed roadway. South Knoxville Boulevard Extension (SR 71) corridor is karst. Karst is found in numerous areas of East Tennessee, predominantly in the Valley and Ridge Province. Long linear belts of sinkholes, caves, sinking streams, underground stream drainage, and dry valleys all characterize the karst in East Tennessee. Most counties in the East Tennessee region contain some amount of karst topography. The proposed South Knoxville Boulevard Extension (SR 71) corridor likewise, has numerous karst features that are found within the area topography. Mapping surface and subsurface karst features greatly aided the evaluation of proposed roadway corridors A result of this study was the development of a surface karst map which showed areas of numerous sinkholes, which is interpreted as areas of potential future sinkhole development. Mapping the Meades Quarry Cave System, located feasible and necessary in order to ascertain the location of the Meades Quarry Cave passages which contain the Berry Cave Salamander. As a result, a portion of Meades Quarry Cave was mapped using both conventional tape and compass survey methods and hypsometer and compass methods. sites, shopping malls, and subdivisions will be problematic in karst areas and should be avoided then minimization of the impact of construction on the and minimal alterations in the surface drainage of an area (both surface and subsurface) are recommended. Mitigation of the impacts on the karst areas needs to be included with the roadway design and construction was found to be approximately North 55 degrees East. Meades Quarry Cave is developed in a linear fashion may tend to bend somewhat to the southwest near the upstream portions of the cave. The cave is developed in the northwest limb of a southwest plunging anticline meaning that the bedrock is dipping to the northwest. It the mapping program tend to dip toward the northwest as a result of being on the northwest limb of the anticline As a result of the geohazard investigation and mapping initiative, it was interpreted that the stream passage of Meades Quarry Cave does indeed lay within the study area of the proposed SR 71 extension.428 Figure 9. Mapping stream cave passage in Meades Quarry Cave.

PAGE 446

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2rare Berry Cave Salamander which lives in the Meades Quarry Cave System. Once a groundwater dye trace is completed, then a more appropriate evaluation of the karst groundwater drainage can be made. It is important to understand that there will be an impact on the karst environment (sinkholes, caves, wildlife, and groundwater). In some places, this impact may be drainage is directed into sinkholes that empty into the caves and groundwater systems. These impacts can be already assisted the Planners and Designers in providing recommendations that minimized the impact systems for sinkholes, impermeable lining for ditchlines (to prevent new sinkholes from forming), the use of graded rock embankments (for stability and groundwater impacts), and the use of structural bridges over sensitive sinkholes should phase. (Stephenson and Beck, 1995; Stephenson, et al., 1997). The groundwater contamination issue is an important 429Figure 10. This map of Meades Quarry Cave, made by the TDOT Geotech section, shows the general stream flow direction, which is parallel to the bed rock strike direction.

PAGE 447

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE John Kizer, and Fred Barrell all from the TDOT their help with the cave mapping effort. References Department of Environment and Conservation, Nashville, Tenn. sinkhole mapping for sinkhole susceptibility 122: Sinkholes and the Engineering and Environmental Impacts of Karst, proceedings of the 9th Multidisciplinary Conference on of the alignment on the known cave and karst systems construction should be anticipated due to the unknown and variable nature of karst features.AcknowledgementsThe Author wishes to acknowledge several people for their assistance in the development of this paper. Len section, and David Barker, Manager of the Knoxville both offered their assistance and support for this study (before my retirement from TDOT). It was while I was In addition, I wish to acknowledge Lori McDowell, TDOT 430Figure 11. Map of SR 71 project area showing Karst Areas 1 and 2 and selected routes based on karst.

PAGE 448

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 Development and Use of Karst Maps in the B. F., ed., Sinkholes and the Engineering and the Environmental Impacts of Karst, Proceedings of the 11th Multidisciplinary Conference on Niemiller, Matt. 2008. Photos of Berry Cave Salamander. Available from: Stephenson, J. B. and B. Beck, 1995. Management of the discharge quality of highway runoff in karst areas to control impacts to groundwater A review of relevant literature: in Beck, B.F., ed., Karst Problems in Karst Terrane, Proceeding of the Fifth Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst, 1997. Management of highway storm water runoff in karst areas baseline monitoring and design of interchange in eastern Knoxville, Tennessee. In 431

PAGE 449


PAGE 450

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2maintain this sole source system as a sustainable regional water supply. Partnerships with multiple agencies and volunteers have minimized individual costs, provided more thorough and complete assessment of karst resource issues, and developed public educational programs on the values of karst.Introduction is located within the northwest limits of San Antonio, Texas, USA. It is a karst area that was planned for urban development but purchased by a partnership organizations. This arrangement was unprecedented purchase, research, develop, and manage a property for research and cooperation of several partners can be used for effective karst resource management.Abstract located on the northwest edge of San Antonio, Texas, USA. Ninety percent of the 47.04 km2 property is located on the recharge zone of the karstic Edwards (Balcones Fault Zone) Aquifer. Urban development is encroaching onto the Edwards Aquifer karst and threatening groundwater quality and karst ecosystems. restoring impacted resources; monitoring and protecting groundwater quality and quantity by encompassing 62% of the 2 Edwards Aquifer recharge and contributing zones, 2 preserving the unique cave fauna; infrastructure to minimize water and ecological impacts; monitoring land use conditions for an adaptive resource management plan; and establishing contiguous buffers around the core resource area. This approach was made possible by designating predominantly determine the location, type, magnitude, and cultural resources. Federally listed endangered bat population occur in its caves. Springs and deep These springs and species, along with chert deposits and natural trails through rugged terrain, have supported GOVERNMENT CANYON STATE NATURAL AREA: AN EMERGING MODEL FOR KARST MANAGEMENT George Veni Executive Director, National Cave and Karst Research Institute, 400-1 Cascades Avenue, Carlsbad, New Mexico 88220-6215 USA, 433Natural and Cultural Resources: Description and Setting2 in northwestern Bexar County at the southern edge of the Edwards Plateau. It is comprised of gently sloping karst ridge tops along its north, east, and west borders that slope steeply down to of the property. It ranges in elevation from 469 m above property. The Balcones Escarpment, the topographic expression to the north, and marks the boundary where several geological, biological, and cultural zones meet, resulting in a high diversity of natural and cultural features.

PAGE 451

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEthat occurs in the hydrologically discontinuous, isolated outcrops of Edwards Limestone on hilltops in and near karst features and fractures in the limestone, and locally does not yield enough water to support water wells. Some Rose, and some is discharged from seeps and small springs Limestone is perched on poorly permeable beds. contains enough clay and marl beds to make it the lower aquiclude for much of the Edwards Aquifer, its outcrop exposes enough limestone and dolomite beds to absorb some recharge. Regionally, there is relatively little use low yield and its contact with gypsiferous zones, which results in occasional high sulfate concentrations. Yet locally in the north Bexar County area, most privately of water. In addition to wells, the Upper Trinity Aquifer also discharges through seeps and minor springs. Edwards Aquifer. This hydrologic connection was best demonstrated in northern Bexar County through a series of dye tracing studies (Johnson et al., 2010) roughly 20 The Edwards Aquifer is a complex hydrologic system which is divided into four zones: contributing or the recharge zone where water enters the Edwards of Edwards Limestone within the Balcones Fault Zone. artesian zone is that area where the Edwards Limestone groundwater rises up fractures to discharge in stream valleys that intersect the potentiometric surface. The bad water line is the downgradient boundary of the artesian GCSNA GeologyThree Cretaceous age limestone formations crop out at dolomite beds interbedded with softer beds of clay and marl, tracks are occasionally found in the limestone beds. the most cavernous unit in the study area. Rose (1972) subdivided the Edwards into the Kainer Formation at the base, with ascending Basal Nodular, Dolomitic, Formation at the top, with ascending Regional Dense, Collapsed, Leached, Marine, and Cyclic members. Maclay and Small (1984) included the Basal Nodular Member as the base of the Kainer. The Edwards is a hard, crystalline, and fossiliferous rock that forms most of the The third and youngest formation exposed is the Austin Chalk. It is a relatively soft unit, approximately 60 m thick, of which about the lowermost 20 m are exposed. It to subparallel faults that locally trend northeast to southwest and drop down to the south and southeast. greatest displacement of any fault known in Bexar Balcones Escarpment, a sudden rise in the land where the Austin Chalk meets the Edwards Limestone Formation. below the Austin Chalk. North of the fault, the Austin Chalk has long ago been eroded from above those units. or more of three aquifers: the Edwards Outlier Aquifer, Upper Trinity Aquifer, and Edwards (Balcones Fault Zone) Aquifer (hereafter called the Edwards Aquifer). The Edwards Outlier Aquifer is the highest in elevation 434

PAGE 452

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2 (Wiesema, 1972). This diversity is a direct result of and especially due to the presence of perennial karst springs, which are not found in many other canyons and valleys along the Balcones Escarpment. and proportionally diverse. Cave ecosystems, by species populations and diversity compared to surface ecosystems. But relative to many karst areas, the caves invertebrates are federally listed as endangered species, Bat Cave containing the largest bat colony in the county (Miller and Reddell, 2011). Cave and Karst Cultural Resources an important natural thoroughfare through the rugged hills along the Balcones Escarpment in the San Antonio area. Unlike most canyons in the region that end steeply, the base of the escarpment up to the fringe of the Edwards Plateau. Indians, Spaniards, US military, and ranchers game, and vegetation. A number of archeological sites have been recorded in Native American encampments, one roughly estimated as representing several thousand years of regular occupation; 24 sites are eligible as State Archeological Landmarks. McNatt et al. (2000) provided an evaluation of prehistoric use of its southernmost area, and archeological study of the entire property is needed in and the nature of that use. Doubtless it served as a vital and reliable source of water, as well as for chert for zone with the saline zone, where total dissolved solids in in western Bexar County is complicated but generally Caves and Karst of which seem likely to open into caves with some excavation (Miller, 2012). Most of the caves and karst features are developed in the Edwards Limestone as recharge features for the Edwards Aquifer. Some phreatically formed and predate the modern aquifer, the largest cave chamber in the county at about 90 m long by 20 m wide by 8 m high. While the Austin Chalk is cavernous elsewhere in Bexar County, no caves or the outcropping horizon is less permeable and soluble. shallow. Dancing Rattler Cave is the longest at 225 m and sinkhole collapse occurred within the unit following Turquoise Sink is the largest collapsed sinkhole in the county to occur in historic times. subdued, their collective karstic permeability is high. except where perched for short distances immediately downstream of springs. A few sections of the valley (unpublished data) show only water from the largest storm stream bed without being fully diverted into the aquifer.GCSNA Cave and Karst Biology of mammals, herpetofauna, and invertebrates. Birds 435

PAGE 453

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCElacked the funds. The City of San Antonio and Edwards Underwater Water District (EUWD; now reorganized as the Edwards Aquifer Authority) were approached, but the city did not see the value in owning parts of the recharge zone while EUWD saw the value but did not want to own and manage land. This impasse was breached when and facilitated a deal for its purchase by TPWD. Since EUWD to pay 75% ($1.5 million) of the purchase price, while TPWD maintained title and general management of the 19.09 km2 property. The property was designated (1994) for a detailed history of the property up to the time of this acquisition. watershed, but also for the protection of its other natural and cultural assets. Figure 1 illustrates the acquisitions of properties surrounding the initial purchase as described in the following narrative. and Urban Development committed to adding 4.54 km2 to TPWD 15 months later. This tract was also part of San Antonio Ranch, but the presence of the endangered limited its capacity for development. TPWD made it a sanctuary for the Warbler, with no public access into that area during the months the birds are nesting. The next acquisition occurred in 1999 through TPL km2 Davis RanchUpland Tract due to $1,581,000 in donations from the San Antonio Water System and the Duncan, Frost, Kronkosky, Meadows, Morris Stafford, and USAA charitable foundations and trusts. The next 2 Ranch to TPWD via additional independent fundraising efforts which covered all but $500,000 of the total contract price, which TPWD paid with general operating 2 Kallison Ranch from TPL for approximately $5 million and sold a conservation easement to the City of San Antonio. TPWD used the funds from the city to qualify for a Land Limestone in the region. At least two caves are known to have served as human burial sites (Veni, 1994 and 1996). Also within the Natural Area is a historic purportedly has seen use as a home, a stage coach stop, GCSNA: A Model for Karst Management Partnerships for Land Acquisition and Protection route between San Antonio and military forts to the northwest. This use by government troops gave the canyon its name. In the 1880s, it became a busy stage coach route between the towns of San Antonio and Bandera, and along with surrounding areas, most of the canyon began to be consolidated under the ownership of ranched the property for about 100 years. In the 1970s, the San Antonio Ranch New Town Corporation purchased the ranch to build a community with a proposed population of more than 80,000 residents. They continued to lease most of the property for ranching 16. In the late 1980s, the corporation failed during the nationwide savings and loans collapse, and the property Trust Corporation (RTC). RTC placed the property for auction, where it was watershed over the karstic Edwards Aquifer recharge of the City of San Antonio onto the recharge zone had effective means of aquifer protection. the Texas Parks and Wildlife Department (TPWD), which 436

PAGE 454

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2endangered species, and allow for limited recreation. By 2009, this tax initiative led to the purchase of four 12.14 km2. They are all in the process of being deeded to TPWD for management. During this time the 1.70 km2 donated a small but environmentally important 0.01 km2 for similar aquifer and species protection. These and Water Conservation Fund grant which covered 58% of the purchase price. This creative and cooperative development of funding was instrumental in heightening public awareness of Aquifer and its endangered species. So starting in 2001, increases that in total raised $220 million to buy land to protect the Edwards Aquifer recharge zone and 437 Figure 1. Properties acquired to expand GCSNA and protect the karstic Edwards Aquifer.

PAGE 455

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE TPWD mandate that the primary purpose of a State Natural Area is resource protection and management, with recreation being of secondary importance and Therefore, before recreation and public access was possible, TPWD conducted inventories of its natural and cultural resources to determine the important scenic, educational, hazardous, and sensitive areas of the property. Surveys include studies of caves, plants, volunteers were used, and continue to be used to conduct 211) from an initial 19.09 km2 to 47.04 km2, extending into neighboring Medina County, and protecting more than 41 km2 of the Edwards Aquifer recharge zone and its endangered species habitat, and about 62% of the recharge and contributing zones (Figure 2). Partnerships for Multidisciplinary Karst Management opened to the general public until October 2005. The 438 Figure 2. GCSNA boundaries and the Government Canyon drainage area relative to the Edwards Aquifer contributing, recharge, and transition zones.

PAGE 456

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2km22 acquired tract. Study of the more recently acquired properties is underway and their use will be determined and understood. Most cave and karst research and management activity independent research. The most active karst management activity involves monitoring the populations of the endangered karst invertebrates and active management actions such as Fire Ants (Solenopsis invicta) which predate upon the through its Karst Management and Maintenance Plan and assistance by its volunteer partners, the US Fish and Wildlife Service determined that management for the caves and the species in the Natural Area provides adequate special management considerations for the primary constituent elements, and consequently [habitat] units within the Natural Area that we proposed for [critical in its research and management activities, and may lead Conclusions property acquisition, deed restriction, public education research and management actions. The synergistic and the quantity of recharge into the karstic Edwards Aquifer, the primary water supply for nearly 2 million people, and the simultaneous protection of habitat for an endangered bird species that nests mostly in karstic canyons and easy access to a large natural karst environment which educates the citizens of Texas about 1994 through October 2012 to survey, study, and restore resources (Miller, 2012). The focus for much of the resource work, especially before corporation dedicated to organizing and managing support for the Natural Area. On November 15, 1996, a Memorandum of Agreement (MOA) between TPWD promote the preservation and protection of the natural and cultural resources of the natural area. Toward this 1. developing a management plan to ensure the 2. working to preserve and protect the natural and developing educational programs on the natural public; 4. and 5. 2012 has logged over 200,000 volunteer hours in resource study, protection, management, and public education 1998) was adopted. It emphasized protection of the and threatened species and cultural resources. Dictated by principle and enforced by deed restrictions, all development and new park facilities would be restricted to the southernmost sections that are off the Edward is over the recharge zone, currently has almost 68 km emergencies, natural area maintenance, and research. 439

PAGE 457

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCESite Descriptions and Supporting Documentation. Center for Archaeological Research Archaeological San Antonio. Recharge Zone, Panther Springs Creek Basin, northern Bexar County, Texas. Edwards Aquifer development on the Edwards Aquifer recharge zone. Edwards Underground Water District Staff Report. Cultural Resources Program, Texas Parks and Wildlife Department. Association, unpublished report. Miller, M, Reddell J. 2011. Summary of biological collections and observations from caves at unpublished report. hydrogeologic characteristics of the Edwards Aquifer Canyon State Natural Area Master Plan, Texas Parks and Wildlife Department. threatened wildlife and plants; Designation of critical unpublished report. unpublished report. discussed publicly until funds can be allocated for their proper study and management. of karst and highlight it in their public education efforts. understand their complicated resources and their complex relationships, before deciding how to manage them, should be applauded and followed in all karst areas. AcknowledgementsThis paper is a summary culmination of hundreds of My thanks go to all of the volunteers, partners, and the staff at TPWD for their support of my work there, and especially for their efforts to preserve, study, and Estes for their support with graphics and a critical review of the manuscript for accuracy and completeness. Additional thanks go to the staff of the National Cave and Karst Research Institute and reviewers of the Sinkhole Conference for their helpful suggestions.References The University of Texas. Dillehay, TD. 1972. An initial archeological reconnaissance of areas to be affected by the San Antonio New Town, Bexar County, Texas. The University of Texas at Austin Research Report No. State Property, Bexar County, Texas. Austin unpublished report. Canyon [updated 2012 Oct; cited 2012 Oct 20]. Available from: aboutus.html Archaeological Survey of Trail Locations in a Survey Results, and Recommendations. Volume 2: 440

PAGE 458

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2structural controls. The geologic mapping efforts are beyond the scope of this paper. sinkhole depressions eliminated from consideration, IntroductionThere are many different types of karst features found across soluble bedrock terrains. Caves and springs represent some of the most spectacular karst features. Sinkholes on the other hand, while generally more numerous, are much more ephemeral. The appearance of a sinkhole, or its recurrence, is often inconvenient and occasionally catastrophic. Ford and Williams (2007) suggest that sinkholes are a diagnostic feature of karst landscapes. One pundit is quoted as saying that the best predictors of new sinkholes are existing sinkholes (Alexander and Lively, 1995). The mapping of sinkholes is therefore one of the primary tools in the delineation of, and assessment of risk on, karst terrains. Mapping the distribution, size, and shape of sinkholes has to locate points on topographic maps was combined with hours of reviewing aerial photographs into a database for Several new tools are becoming rapidly available that fundamentally change our approach to discerning, locating, and delineating sinkholes across counties and whole states. Chief among these new tools is LiDAR (Light Detection And Ranging). The hillshade LiDAR image, Figure 1, of several large sinkhole complexes illustrates how AbstractThe mapping of karst features has taken on increasing importance in land use planning and zoning regulations across south east and east central Minnesota. The delineation of sinkholes, springs, and other features using topographic maps, and intensive networking with local landowners. The luck of the observer has also landowners, concealed within extensive row crops, or hidden under tree canopies. The application of aerial tools allows mapping across vertical) of small depressions in karst landscapes without interference from vegetation. These features can be visually compared to aerial photography, both visible and infrared, resolution Pictometry imagery allows overhead views from several angles to further identify and verify the genesis of a given depression. In areas with previously mapped karst features, precise locations can be compared to earlier estimates of location, which is particularly useful in applications like nearest neighbor analysis. The improved elevation mapping resulting from LiDAR work has greatly improved geologic mapping efforts geologic mapping allows much better correlation of karst features within stratigraphic units as well as identifying AND PICTOMETRY TOOLS FOR KARST FEATURES DATABASE MANAGEMENTScott C. Alexander, Erik Larson, Cody Bomberger, Brittany Greenwaldt, E. Calvin Alexander, Jr. Department of Earth Sciences, University of Minnesota, 310 Pillsbury Dr. SE, Minneapolis, Minn. 55455, USA, alexa017@umn.eduMina RahimiWater Resources Science, University of Minnesota, 173 McNeal Hall, 1985 Buford Ave., St. Paul, Minn. 55108, USA, rahi0016@umn.edu441

PAGE 459

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE442in the upper left). All of these large depressions show 1:24,000 topological maps. The karst of southeastern covers Ordovician carbonates. These depressions can be approximated as a series of intersecting conical sections as expected by subsurface sediment transport. Airborne LiDAR is used to create Digital Elevation Models (DEMs) and Digital Terrain Models (DTMs). These models can be combined with other geographically landscape features. In particular, bedrock geology, sinkholes into a geologic context. A distinct advantage of LiDAR is its ability to see through vegetation mapping the land surface. This is an improvement over traditional aerial photography especially in densely wooded areas and areas with extensive agricultural cropping. Figure 2 is a conventional air photo taken in the fall, after the leaves have fallen of the trees. Even with the leaf cover off there are no obvious sinkholes, in stark comparison to the numerous area, as Figure 1). While traditional aerial photography looks straight down Pictometry methods are now capturing low altitude, imagery can capture easily interpretable photos across In Minnesota LiDAR mapping has been driven by the event that produced 10 inches of rain across much of southeastern Minnesota (Loesch, 2009). This has been followed up by a statewide effort, the Minnesota legislature as part of the Clean Water Land and Legacy Amendment to the State Constitution. Mapping is being done at 1.5 meter horizontal spacing with 15 centimeter vertical accuracy. In addition to improvements in DEMs due to LiDAR, similar improvements have been occurring with aerial photography. Web Map Services (WMS) compile data Figure 3. LiDAR imagery of sinkholes in Figure 2 woodlot area. Figure 1. LiDAR imagery of large sinkholes east of Utica, Minnesota, with white circle highlighting sinkholes. Figure 2. Color air photo of woodlot west of Utica, MN (Fall 2011).

PAGE 460

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2are indicative of soils that are holding moisture, i.e. are soil. Lastly, there is a light colored area (labeled) that is map does little to help discern whether this is a small rise or if it is a depression. Figure 5 is a hill shade image of LiDAR data for the Figure 4 area. The hill shade image highlights topographic light source, in this case from the northwest. The tile lines now appear as trenches in this shaded relief image and the St. Peter mounds stand up from the surrounding landscape. A small mesa capped by the Platteville Limestone over St. Peter Sandstone is visible in the southwest corner. More subtle are several fence lines that now show up as positive features providing evidence of soil erosion in the numerous state and federal agencies into one easily accessible location. This data includes visible spectrum aerial photography going back to 1991 along with color Data Catalog, 2012). Stepping beyond traditional aerial photography, where the observer is looking straight down from high altitude, oblique imaging methods are becoming much more widely available (Kalinski, 2010). Oblique imaging, commonly referred to as pictometry, is a patented imaging process that collects images looking downward at a 40 degree angle from low altitude aircraft (Pictometry International Corporation, 2002). Typical coverage includes 12 to 20 overlapping images collected from several directions for any given point on the landscape. to what you would see from a tall structure or mountain side, features are more intuitively interpretable and understandable. Traditional aerial photography provides an important Earth [ ]. MethodsThe section of land in Figures 4, 5, and 6 contains downtown Utica, Minnesota in the northwest corner of the images. In this photo a newly installed, at that time, tile line is visible crossing from the lower left to the top center of the image. In addition, two sinkholes (D010 and D295) occur in the top of the Prairie du Chien Limestone. There are several other features of note in Figure 4. First, there is a remnant band of St. Peter Sandstone lying on top of the Prairie du Chien, appearing as a light colored soil north and east of the D295 label. Note that the soils in this area are relatively thin, less than 0.5 meter thick over the St. Peter rises, and up to one meter thick over the Prairie du Chien. Second, there are darker colored soils associated with the tile lines, especially along the 443 Figure 4. Google Earth image Utica, Minnesota area April 1991. Figure 5. Hillshade LiDAR image of Figure 4 area.

PAGE 461

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCEare, however, assorted depressions on the landscape that are not sinkholes; simply identifying depressions will not create a sinkhole inventory. Separating and depressions requires an accounting of these other types of depressions. This is particularly important in glacially Minnesota. building foundations as in Figure 7. Distinguishing characteristics are the relatively rectangular shape and its proximity to an existing farm yard. The combination of pictometry, allowing us to see the farmyard and wood lot, with the shape of the depression from LiDAR A related type of feature would be small quarry pits that were dug to extract building stone for these same barn foundations. Quarries however are usually open on one end to allow for wheeled transport. A second type of agriculturally derived depression is the cattle wallow. Cattle, and their bison predecessors on this landscape, are known to wallow in mud to help ward off Figure 8 shows a Pictometry image of a wood lot grazed by cattle which includes a sinkhole and a cattle wallow. The sinkhole in the LiDAR inset of Figure 8 is a visible depression. The cattle and their associated wallow can be seen on the right side of the image. The slight ridge in the These cattle wallows, and especially bison wallows, can be quite large in size and depth. Figure 9 shows an early 20th century image from Kansas as reported by Darton in up, in this case in a maple forest. A notable feature of they always create an associated mound where the root persist for decades. A series of depressions in the top center of the figure can now be mapped as suspected sinkholes and are indicated by numbered red diamonds. Previously mapped sinkholes are denoted by red xs. Note that the light colored area from Figure 4 is now an obvious depression, suggesting that it can seasonally hold water diverting construction of the tile line. Figure 6 shows a Pictometry image covering most of the same area as in Figures 4 and 5. This image is from 2010 and shows a landscape concurrent with Figure 5. This image is available from BingTM Maps [ using the Birds eye view feature. The more native, or natural view, allows further interpretation of the landscape imagery. Efforts at conservation, in particular the Conservation Reserve Program (CRP) are visible in Figure 6. The creation of berms around sinkholes and permanent As shown in Figures 4, 5, and 6, each aerial imaging method has advantages that can be used in a complimentary fashion. Used together these scales at spatial and temporal scales that were not previously available. Additionally, bedrock geologic maps and surficial sediment thickness maps can aid in the identification of sinkhole prone areas and the distribution of mapped sinkholes can be fed back into geologic maps.Discussion of identifying depressions in the landscape. There 444 Figure 6. Pictometry image of Figure 4 area.

PAGE 462

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2In Minnesota, where most of the state has been recently glaciated, there are numerous depressions formed due glacial processes and particularly due to ice block melting processes. The most recent glaciation however missed southeastern Minnesota passing to the west forming Des Moines Lobe tills. Areas overlain by Des Moines Lobe tills generally have more than 15 meters of unconsolidated sediment over bedrock. The surface expression of karst in these areas is limited (Alexander and Lively, 1995). In areas where the depth to bedrock is less than 15 karst features can become intermixed and examples of composite features are known in Minnesota (Shade, 2002). In portions of Pine County Minnesota, along features is common (Figure 12). The Kettle River valley hydraulic gradients. Sinkholes in this area are the result of dissolution of the highly cemented quartz sandstone Figure 11 shows a LiDAR image of a maple and oak forest on the Kettle River plain in Pine County, Minnesota. This is an area of thin soils over sandstone bedrock. Note the largest trees are vulnerable. While sinkholes are common in the area, sinkholes do not seem to form at this elevation. The low hydraulic gradient to the base level in the Kettle River of the area shown in Figure 11 is apparently not conducive to sinkhole formation.445Figure 8. Sinkhole D904 in woodlot with nearby cattle wallow.Figure 7. Former barn foundation. Figure 9. Buffalo wallow, shallow circular depression in the level surface (from Darton, 1920). Figure 10. Tree tip-up. Photo by Kerry D. Woods.

PAGE 463

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE addition there are numerous wetland depressions of various sizes. The wetland areas are more obvious in the depressions, or ice block melt out features related to retreat of the Superior Lobe ice sheet (Shade et al., 2002). more sinkholes begin to appear. Several of the sinkholes may eventually be entirely captured by the sinkholes. this paper, is the land slide or slump. Figure 14 shows a hill shade LiDAR image of the edge of a Platteville mesa immediately to the west of Utica, Minnesota. The Platteville Limestone provides a cap to the mesa while the steep side slopes are St. Peter Sandstone. Water is transported horizontally along bedding planes in the Platteville. Seeps form across the St. Peter slope creating frequent slumps. Potential sinkhole features, that have a different morphology from the slumps, are highlighted with red and blue circles. Figure 15 is the same area from Figure 14 but the Pictometry image shows slumps in the St. Peter slope where the white sandstone is exposed. The washout from below. The four potential sinkholes are obscured by the wooded slope.446Figure 13. 2010 aerial photo of Figure 12 area.Figure 12. Sinkholes in sandstone karst above Kettle River.Figure 11. Array of tree tip-ups from Kettle River plain, Pine County, MN. Figure 14. LiDAR image of slumps and potential sinkholes (circles) west of Utica, MN.

PAGE 464

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2ResultsOver the past three years summer interns, funded by the employing aerial and Pictometry photos. Efforts to date have been aimed at southeastern Minnesota. In addition, mapping has been supported by the Water Resources Center at the University of Minnesota (see Ramini and Alexander in these proceedings). The results of these mapping efforts are in the process being added to the Karst Features Database for Minnesota Minnesota, the number of mapped sinkholes increased from 5 to 44 after mapping in 2010 by Erik Larson. This seen in the large discussion section of this paper. and Alexander, 1995) a total of 6,199 sinkholes were previously mapped sinkholes with an average correction sinkholes that were not visible in LiDAR had been mapped through discussions with landowners and county sinkholes. Work by Rahimi and Alexander, presented sinkholes. Wabasha, Dodge, Steele, and Washington counties of Minnesota. The Karst Features Database, as managed by the Minnesota Department of Natural Resources, is now land use decisions. A thorough and accurate database of Figure 16 shows the same area as in Figures 14 and 15 but now presents the view looking southward, or up slope. The two at the base of the bluff without light colored wash out fans; they are therefore not slump features. These two features may be sinkholes into the underlying Prairie du Chien. The two red circled areas are still hidden within the forested slope. suspected karst sinkholes. There are occasional sinkholes near the edge of the Platteville into the St. Peter as seen in the southernmost red circle. These Platteville sinkholes, particularly near bluff edges, can be the result of mechanical erosion of the poorly cemented St. Peter Sandstone along bedrock fractures (Barr and Alexander, 2009). The second red circle is a sinkhole into a St. Peter crevice. The two blue circles, based on the visual and elevation data, are likely sinkholes into require excavation as demonstrated by Shade (2002).447Figure 15. Pictometry view looking to the north of slumps and potential sinkholes in Figure 14. Figure 16. Pictometry view looking to the south of slumps and potential sinkholes in Figure 14.

PAGE 465

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE database development in Minnesota: analysis of feature database. Journal of Cave and Karst Studies MW. 2000. Changes in Minnesota forest during 14 years following catastrophic windthrow. Journal of Pictometry International Corporation. 2002. US Patent Shade BL. 2002. The genesis and hydrogeology of a sandstone karst in Pine County, Minnesota. M. Sc. Thesis, University of Minnesota, July 2002. Shade BL, Alexander SC, Alexander EC Jr. 2002. Karst features in Pine County, Minnesota. Chapter 4. in: Boerboom TJ. Contributions to the geology of Pine Witthuhn K, Alexander EC Jr. 1995. Sinkholes and Fillmore County, Minnesota. County Atlas Series, Resources, St. Paul, MN. sinkholes, and other karst features, is helping inform the siting large scale animal operations and municipal waste water facilities, along with many other intense types of land use.ConclusionsModern imaging and elevation tools can provide a wealth of information. Individually, aerial photography, to map many different types of features. Used in in Minnesota the number of mapped sinkholes has been roughly doubled. In addition, the locations of previously mapped sinkholes in these areas have been features, such as with nearest neighbor analysis, will be more robust and meaningful. None of these methods will eliminate the need for that were previously inaccessible and allow mapping effort. The mapping of features using historic air photos ReferencesAlexander EC Jr., Lively RS. 1995. Karst aquifers, caves, and sinkholes. In: Lively RS, Balaban of Fillmore County, Minnesota. County Atlas University of Minnesota, St. Paul, MN. Barr KDL, Alexander EC Jr. 2009. Examples of hypogenic karst collapse structures Twin Cities Metropolitan Area, Minnesota. In: Stafford KW, Land L, and NCKRI Symposium 1. National Cave and Karst Ford D, Williams P. 2007. Karst hydrogeology and geomorphology. Rev. ed. West Sussex (England): John Wiley & Sons. database development in Minnesota: design and data 448

PAGE 466

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2delineation of sinks and depressions. In regions where the DEM had been properly conditioned, the tools for automated delineation performed reasonably well as compared to the manually delineated depressions, but generally overestimated the number of depressions thus from the TPI thresholding analysis were not dependent meaningful depressions depended on careful assessment of analysis scale and TPI thresholding. IntroductionAirborne LiDAR (Light Detection and Ranging) offers enormous potential for mapping sinkholes. LiDAR to map sinkholes or other types of depressions have not been standardized. In the past, contour maps or digital elevation models were visually inspected for the presence of closed depressions, and these depressions were identified or manually interpretation and delineation of depression features of sinkholes and other karst features mapped from interpreters or regions studied. Moreover, complete field verification of individual features is often impractical, thus the reliability of manually digitized sinkhole data produced by even a singular worker may be questionable. Other studies which have examined the use of digital data (including contours derived from LiDAR) for manual interpretation of methodology can result in false positive and false negative identification of karst features (e.g., Seale et al., 2008; Vacher et al., 2008).AbstractLiDAR (Light Detection and Ranging) surveys of models (DEMs) that are particularly useful for mapping sinkholes. In this study, we used automated a 1.0 m resolution LiDAR DEM in order to delineate sinkholes and closed depressions in the Boyce 7.5 minute quadrangle located in the northern Shenandoah Valley of Virginia. The results derived from the use of the automated tools were then compared with depressions manually delineated by a geologist. Manual delineation of closed depressions was conducted using a combination of 1.0 m DEM hillshade, slopeshade, aerial imagery, and Topographic Position Index (TPI) rasters. The most effective means of visualizing depressions in TPI raster atop the slopeshade raster at 1.0 m resolution. checked using aerial imagery to screen for false positives, evaluating, and selecting sinks and depressions as well as thresholding, grouping, and assessing depressions from the TPI raster. Results showed that the automated the DEM to produce hydrologically correct surface generated where roads, railways, or other manmade Additional conditioning of the DEM with drainage paths across these barriers was required prior to automated AN EVALUATION OF AUTOMATED GIS TOOLS FOR DELINEATING KARST SINKHOLES AND CLOSED DEPRESSIONS FROM 1-METER LIDAR-DERIVED DIGITAL ELEVATION DATADaniel H. DoctorU.S. Geological Survey, 12201 Sunrise Valley Drive, Reston, Virginia, 20192, dhdoctor@usgs.govJohn A. YoungU.S. Geological Survey, 11649 Leetown Rd., Kearneysville, W. Virginia, 25430, jyoung@usgs.gov449

PAGE 467

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE450The goal of this study was to compare manual interpretation to automated detection of sinkholes and elevation dataset. Two approaches of automated their hydrologic spill point in order to correctly model DEMs are artifacts of the DEM production process and geomorphic features in areas of karst terrain and on highly accurate LiDARderived DEMs (Zandenberg, 2010; Lindsay and Creed, 2006). By subtracting the original a new difference grid elevation dataset is produced representative of depression location and depths (Anders et al., 2011; Siart et al., 2009; Antonic et al., 2001). The second approach used the Topographic Position calculates the difference between the elevation at each pixel in the DEM and the mean elevation in a neighborhood surrounding the pixel (Jenness et al. 2011). The TPI is similar in concept to other local topographic relief measures that can be calculated McNab (1989) and the difference in mean elevation Wilson (2000). Different feature scales can be assessed by varying the size of the analysis window, and various feature types can be assessed by using square, triangular, circular, or annular window shapes. The of each pixel as being either higher or lower than a localized average (Jenness et al., 2011). Negative TPI values represent topographic lows (concavities, depressions), while positive TPI values represent topographic highs (convexities, ridges). The TPI (or scale, concave (e.g., cave) openings (Weishampel et al., 2011), and convex burial mounds (De Reu et al., 2011), and thresholding the TPI values has been suggested as a means of classifying depressional features on the landscape (Klingseisen et al., 2008). For this study, no effort was made to distinguish among sinkholes and other types of depressions, whether natural or manmade. The goal was simply to test the reliability LiDAR elevation model against a manually produced feature dataset of the area will focus on separating true karst features from manmade features, and sinkholes from other types of natural depressions such as ponded springs, estavelles, or suffosion depressions.Study Area The study area is the northern half of the Boyce 7.5 minute quadrangle in Clarke County, Virginia, a region of approximately 70 km2 (Figure 1). The Boyce quadrangle is located within the Shenandoah River drainage basin, physiographic province of the Appalachian mountain range. The geology of the quadrangle was originally an inventory of karst features was not included in the Figure 1. Map of study area. Red points are larger known sinkholes from Hubbard (1983).

PAGE 468

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2These three rasters were overlain upon a 1 m resolution ArcMap depressions was conducted by close visual inspection each grid tile region was 1.5 km square. Depressions in the elevation surface were manually digitization, each polygon was attributed with LiDAR elevation model, aerial imagery, or some combination of these), and some basic descriptive notes. Description of an individual feature was aided georeferenced aerial imagery, and some aspect of the hydrologic condition was normally noted (e.g., ponded, ephemerally ponded, located within an In general, it was found that the LiDAR rasters were more reliable in terms of an accurately georeferenced dataset than were the aerial images, thus digitization Nonetheless, the georeferenced aerial imagery was critical for identifying false depressions that arose out of processing of the LiDAR bare earth elevation model, as noted by others (e.g., Seale et al., 2008). Digital outlining of polygons created to represent depressions was best facilitated with the TPI raster at 40% transparency draped over the slopeshade raster best illuminated the variations in surface elevation in a manner that could be most readily outlined, avoiding pitfalls of a traditional hillshade surface such as deep shadows, overly highlighted areas, or poor illumination angle and inclination. Although the slopeshade raster high slope (e.g., the edges of a collapsed sinkhole), it does not distinguish between regions of higher or lower elevation; thus, a haystack may appear similar to a sinkhole. Fortunately, the TPI raster permitted distinguishing between areas of higher or lower local elevation (darker areas are local lows, lighter areas are local highs). Thus, the combined use of slopeshade and TPI rasters greatly enhanced the visualization of depressions for manual digitization. number of only the largest depressions in the region from topographic maps. The rocks of the region are composed of Paleozoic sedimentary carbonate and siliciclastic rocks that were faulted and folded during the Alleghanian orogeny, and erosion has left the core of the ancient Appalachian mountain range exposed at the present land surface. dissected karst surface of moderate to low relief, with 90 m total elevation range and a mean elevation of 180 m above sea level. Sinkholes and karstic depressions processes within the residuum overlying the carbonate bedrock; the thickness of residuum varies between 0 to upwards of 10 m. Bedrock structures (folds, faults, and fractures) exert strong control over the locations of karstic depressions in the region (Doctor et al., 2008; Doctor and Doctor, 2012).Methods Manual delineation of depressionsAirborne LiDAR was acquired over the study area between March 01 and March 09, 2011. Flights took one week of complete melting of snow cover. Conditions elevation surface. The vertical accuracy of the delivered LiDAR data was within 9.0 cm root mean square error (RMSE) and the LiDAR point cloud was acquired at breaklines were manually collected based on the LiDAR on areas of known water surfaces, depressions or sinks delivered LiDAR data. derive 1) a hillshade raster, 2) a slopeshade raster, and above the horizon and represented using a grayscale color ramp. The slopeshade raster applied a grayscale color ramp to the surface according to the slope of each pixel. The TPI raster was calculated using a circular annular window with an inner radius of 2 m and outer radius of 10 m (see Jenness et al., 2011 for details), and a grayscale color ramp applied to the resulting raster. 451

PAGE 469

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE452 NCKRI SYMPOSIUM 2 identifying true depressions. This eliminated many elongated depress ions that appear within stream Figure 2. Comparative views of an area with multiple closed depres sions in the Boyce quadrangle. In the lower left of each image is a cluster of manmade percolation pits ( 1) ; in the upper center are two sinkholes, one deeper than the other (2) ; in the upper right is a bedrock ridge surrounded by trees along a fence line ( 3) ( A) Aerial image. ( B) Hillshade of LiDAR DEM. ( C) Slopeshade of LiDAR DEM. ( D) TPI raster of LiDAR DEM (note : dark areas are depressions and light areas are ridges). ( E) TPI draped over the slopeshade. ( F) TPI draped over slopeshade, with manually delineated closed depressions. A r ed X indicate s a fals e depression where bu i lding edges caused errors in the LiDAR elevation model. 1 2 3 NCKRI SYMPOSIUM 2 identifying true depressions. This eliminated many elongated depress ions that appear within stream Figure 2. Comparative views of an area with multiple closed depres sions in the Boyce quadrangle. In the lower left of each image is a cluster of manmade percolation pits ( 1) ; in the upper center are two sinkholes, one deeper than the other (2) ; in the upper right is a bedrock ridge surrounded by trees along a fence line ( 3) ( A) Aerial image. ( B) Hillshade of LiDAR DEM. ( C) Slopeshade of LiDAR DEM. ( D) TPI raster of LiDAR DEM (note : dark areas are depressions and light areas are ridges). ( E) TPI draped over the slopeshade. ( F) TPI draped over slopeshade, with manually delineated closed depressions. A r ed X indicate s a fals e depression where bu i lding edges caused errors in the LiDAR elevation model. 1 2 3 NCKRI SYMPOSIUM 2 identifying true depressions. This eliminated many elongated depress ions that appear within stream Figure 2. Comparative views of an area with multiple closed depres sions in the Boyce quadrangle. In the lower left of each image is a cluster of manmade percolation pits ( 1) ; in the upper center are two sinkholes, one deeper than the other (2) ; in the upper right is a bedrock ridge surrounded by trees along a fence line ( 3) ( A) Aerial image. ( B) Hillshade of LiDAR DEM. ( C) Slopeshade of LiDAR DEM. ( D) TPI raster of LiDAR DEM (note : dark areas are depressions and light areas are ridges). ( E) TPI draped over the slopeshade. ( F) TPI draped over slopeshade, with manually delineated closed depressions. A r ed X indicate s a fals e depression where bu i lding edges caused errors in the LiDAR elevation model. 1 2 3 NCKRI SYMPOSIUM 2 identifying true depressions. This eliminated many elongated depress ions that appear within stream Figure 2. Comparative views of an area with multiple closed depres sions in the Boyce quadrangle. In the lower left of each image is a cluster of manmade percolation pits ( 1) ; in the upper center are two sinkholes, one deeper than the other (2) ; in the upper right is a bedrock ridge surrounded by trees along a fence line ( 3) ( A) Aerial image. ( B) Hillshade of LiDAR DEM. ( C) Slopeshade of LiDAR DEM. ( D) TPI raster of LiDAR DEM (note : dark areas are depressions and light areas are ridges). ( E) TPI draped over the slopeshade. ( F) TPI draped over slopeshade, with manually delineated closed depressions. A r ed X indicate s a fals e depression where bu i lding edges caused errors in the LiDAR elevation model. 1 2 3 NCKRI SYMPOSIUM 2 identifying true depressions. This eliminated many elongated depress ions that appear within stream Figure 2. Comparative views of an area with multiple closed depres sions in the Boyce quadrangle. In the lower left of each image is a cluster of manmade percolation pits ( 1) ; in the upper center are two sinkholes, one deeper than the other (2) ; in the upper right is a bedrock ridge surrounded by trees along a fence line ( 3) ( A) Aerial image. ( B) Hillshade of LiDAR DEM. ( C) Slopeshade of LiDAR DEM. ( D) TPI raster of LiDAR DEM (note : dark areas are depressions and light areas are ridges). ( E) TPI draped over the slopeshade. ( F) TPI draped over slopeshade, with manually delineated closed depressions. A r ed X indicate s a fals e depression where bu i lding edges caused errors in the LiDAR elevation model. 1 2 3 NCKRI SYMPOSIUM 2 identifying true depressions. This eliminated many elongated depress ions that appear within stream Figure 2. Comparative views of an area with multiple closed depres sions in the Boyce quadrangle. In the lower left of each image is a cluster of manmade percolation pits ( 1) ; in the upper center are two sinkholes, one deeper than the other (2) ; in the upper right is a bedrock ridge surrounded by trees along a fence line ( 3) ( A) Aerial image. ( B) Hillshade of LiDAR DEM. ( C) Slopeshade of LiDAR DEM. ( D) TPI raster of LiDAR DEM (note : dark areas are depressions and light areas are ridges). ( E) TPI draped over the slopeshade. ( F) TPI draped over slopeshade, with manually delineated closed depressions. A r ed X indicate s a fals e depression where bu i lding edges caused errors in the LiDAR elevation model. 1 2 3 Figure 2. Comparative views of an area with multiple closed depressions in the Boyce quadrangle. In the lower left of each image is a cluster of manmade percolation pits (1); in the upper center are two sinkholes, one deeper than the other (2); in the upper right is a bedrock ridge surrounded by trees along a fence line (3). (A) Aerial image. (B) Hillshade of LiDAR DEM. (C) Slopeshade of LiDAR DEM. (D) TPI raster of LiDAR DEM (note: dark areas are depressions and light areas are ridges). (E) TPI draped over the slopeshade. (F) TPI draped over slopeshade, with manually delineated closed depressions. A red X indicates a false depression where building edges caused errors in the LiDAR elevation model.

PAGE 470

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 24532. In the following automated analyses, we conservatively chose an area cutoff of a true depression to be greater than or equal to 9.0 m2. Automated generation of depressions by difference raster Several useful tools for processing elevation data exist to preprocess or recondition the DEM in order to hydrologically correct elevation model is one in which every pixel in the surface slopes continually down gradient and out the edges of the elevation model boundaries. If a pixel (or region of pixels) is at a lower elevation than all of the surrounding pixels, the pixel would spill out of the sink. Therefore, reconditioning a direction within each pixel once the sinks in the DEM DEM with this tool results in a new elevation surface larger depression, the process is iterative until all pixels spills over. As a result of this process, stream channels in the DEM that pass under roads through culverts or beneath bridges may become dammed if the culvert or bridge has not been represented (i.e., cut into the elevation surface to stream level) in the DEM (Figure culverts are normally removed from the DEM during processing of the bare earth elevation model; however, in order to create a hydrologically correct DEM, all of any evidence of natural depressions. As requested, the Shenandoah Valley LiDAR was not processed to have by the end user. This presents an opportunity to partially Once the polygon data digitization was complete, the vector data were further attributed with geometric axis length (of an idealized ellipse), circularity index elliptical eccentricity. These geometric characteristics were examined to identify threshold values that could be the automated techniques. The eccentricity of an ellipse (commonly denoted as e) is calculated as: (1) where a and b minor axes, respectively. The eccentricity of an ellipse will be greater than 0 (a perfect circle) and less than 1. Thus, the elliptical eccentricity can be a useful measure of the shape of a possible depression. Based on the manual dataset, we found that an eccentricity of less than or equal to 0.98 worked as a good threshold for identifying true depressions. This eliminated many elongated depressions that appear within stream channels, road ditches, and other features that are unlikely to be true natural depressions. Similarly, the circularity index is a measure of the deviation of a polygon from a perfect circle based upon its perimeter and area. Since a circle has the smallest perimeter to area ratio, a relationship can be established between the expected circular perimeter of a feature (based on area) and its measured perimeter to create an index of circularity (Circi): (2) Where A is area, Pe is the expected perimeter if the feature were a perfect circle, and Po is the observed perimeter. After some experimentation, we used a circularity threshold of less than 1.7 to capture closed depressions while eliminating linear features The smallest area of any of the manually delineated 2, and the smallest area of the e = 1 = 2 = + 1

PAGE 471

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE454 1. The original elevation DEM was then subtracted raster that represented the depth of depressions in the original surface. 2. greater than 0.10 m (10 cm) were extracted to a new raster. The minimum depth threshold of a pixel was conservatively chosen to be 10 cm, or slightly greater than the vertical accuracy of the LiDAR bare earth model which has a RMSE of 9 cm. 4. The remaining pixels were converted into polygons, without simplifying, such that polygon boundaries exactly matched pixel edges. The result is a polygon layer that is representative of the based upon the accuracy of the LiDAR data alone. In represent true depressions, we used a training polygon polygons were calculated, and the results were compared. automate the process of locating natural depressions. It also represents a challenge for determining true sinks. however, this process does not directly discriminate between true depressions and those that are artifacts of the Fill process, thus all depressions in a DEM will be drained unless selected otherwise according to some criteria. Poppenga et al. (2011) suggested using a threshold area of depressions targeted for drainage greater than or equal to 1,000 square meters, and depth greater than or equal to 1 meter and greater than 0.5 standard deviation of the true natural depressions in some karst areas, and exclude a number of depression artifacts. In order to preserve true depressions, the best means is to manually digitize polyline features corresponding to bridges and culverts at lower the elevation of the DEM along the linear feature to approach taken in this study, and was accomplished using (Maidment, 2002). Locating and digitizing actual culverts was facilitated using aerial imagery, and was assisted by the shape of the depression artifacts themselves such that such as a road, railway, or driveway was immediately suspect, and targeted to be drained by addition of a culvert. NCKRI SYMPOSIUM 2 Figure 3. A) Artificial depressions created using the Fill tool in ArcGIS are shown with warm colors atop the LiDAR hillshade. Violet lines represent streams derived from the flow accumulation values applied to the Fill raster. Thick green outlines represent manually delineated depressions. Note that roads, railroads, driveways, and other features in the LiDAR elevation model act as dams when streams pass beneath them through culverts. B ) C ulverts need to be manually digitized (short pink lines) and burned into the DEM in order to allow streams to drain across the obstructions. Light green areas represent the resulting automate d depressions from re-running the Fill routine on the reconditioned DEM. Some culverts were missed in the upper left requiring further refinement through culvert addition. Once the reconditioned DEM was filled, 1. the original elevation DEM was then subtracted from the filled DEM to generate a fill depressions in the original surface. 2. All values in the fill difference raster that were greater than 0.18 m (18 cm) were extracted to a new raster. The minimum depth threshold of a pixel was conservatively chosen to be 18 cm, or twice the vertical accuracy of the LiDAR bare earth model which has a RMSE of 9 cm. This allowed for a great number of possible artificial depressi ons to be culled from the fill difference data, and provides a 95% confidence level that a given pixel is a true depression in the elevation model. The values in this threshold fill difference raster were re classified to an integer type raster such that p ixels in the fill difference raster with values of less than 0.1 m (10 cm) were classified as No 4. The remaining pixels were converted into polygons, without simplifying, such that polygon boundaries exactly matched pixel edges. The result is a poly gon layer that is representative of the possible sinks from the reconditioned fill difference raster based upon the accuracy of the LiDAR data alone In order to refine these polygons down to those which may represent true depressions, we used a training polygon feature class of known depressions that had been i dentified in the field, and digitall y outlined from the LIDAR data. the true and candidate depression polygons were calculated, and the results were compared. The geometric properties were calculated as follows: 1. The Zonal Statistics as Table tool was used to fill difference raster overlain by each polygon. The table also contains the area of the polygon, and a count of the number of pixels per polygon (which are both equal values for 1m square pixels). 2. The results of the zonal statistics table were calculated and each polygon was attributed wi th length of an idealized ellipse that would contain the polygon. 4. The eccentricity of an ellipse was calculated as a measure of elongation of a potential sink, with values between 0 and 1. A threshold value of 0.98 was chosen for the eccentricity based on visual examination of the manually delineated polygons, and any depressions above thi s threshold were removed. NCKRI SYMPOSIUM 2 Figure 3. A) Artificial depressions created using the Fill tool in ArcGIS are shown with warm colors atop the LiDAR hillshade. Violet lines represent streams derived from the flow accumulation values applied to the Fill raster. Thick green outlines represent manually delineated depressions. Note that roads, railroads, driveways, and other features in the LiDAR elevation model act as dams when streams pass beneath them through culverts. B ) C ulverts need to be manually digitized (short pink lines) and burned into the DEM in order to allow streams to drain across the obstructions. Light green areas represent the resulting automate d depressions from re-running the Fill routine on the reconditioned DEM. Some culverts were missed in the upper left requiring further refinement through culvert addition. Once the reconditioned DEM was filled, 1. the original elevation DEM was then subtracted from the filled DEM to generate a fill depressions in the original surface. 2. All values in the fill difference raster that were greater than 0.18 m (18 cm) were extracted to a new raster. The minimum depth threshold of a pixel was conservatively chosen to be 18 cm, or twice the vertical accuracy of the LiDAR bare earth model which has a RMSE of 9 cm. This allowed for a great number of possible artificial depressi ons to be culled from the fill difference data, and provides a 95% confidence level that a given pixel is a true depression in the elevation model. The values in this threshold fill difference raster were re classified to an integer type raster such that p ixels in the fill difference raster with values of less than 0.1 m (10 cm) were classified as No 4. The remaining pixels were converted into polygons, without simplifying, such that polygon boundaries exactly matched pixel edges. The result is a poly gon layer that is representative of the possible sinks from the reconditioned fill difference raster based upon the accuracy of the LiDAR data alone In order to refine these polygons down to those which may represent true depressions, we used a training polygon feature class of known depressions that had been i dentified in the field, and digitall y outlined from the LIDAR data. the true and candidate depression polygons were calculated, and the results were compared. The geometric properties were calculated as follows: 1. The Zonal Statistics as Table tool was used to fill difference raster overlain by each polygon. The table also contains the area of the polygon, and a count of the number of pixels per polygon (which are both equal values for 1m square pixels). 2. The results of the zonal statistics table were calculated and each polygon was attributed wi th length of an idealized ellipse that would contain the polygon. 4. The eccentricity of an ellipse was calculated as a measure of elongation of a potential sink, with values between 0 and 1. A threshold value of 0.98 was chosen for the eccentricity based on visual examination of the manually delineated polygons, and any depressions above thi s threshold were removed. Figure 3. A) Artificial depressions created using the Fill tool in ArcGIS are shown with warm colors atop the LiDAR hillshade. Violet lines represent streams derived from the flow accumulation values applied to the Fill raster. Thick green outlines represent manually delineated depressions. Note that roads, railroads, driveways, and other features in the LiDAR elevation model act as dams when streams pass beneath them through culverts. B) Culverts need to be manually digitized (short pink lines) and burned into the DEM in order to allow streams to drain across the obstructions. Light green areas represent the resulting automated depressions from re-running the Fill routine on the reconditioned DEM. Some culverts were missed in the upper left, requiring further refinement through culvert addition.

PAGE 472

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2455Since the TPI calculates the local relative surface elevation of all pixels in the LiDAR DEM (above or below the local mean elevation), a threshold of the TPI values was used such that only depressional elevations are represented. TPI values on the overall mean of TPI, the TPI threshold was set to distinguish actual depressions as the overall mean TPI Additionally, the grouped TPI values characterize linear (e.g. gullies, road cuts) as well as circular depressional depressions into coherent features, and then assessing the shape of these features by computing an index of circularity using area and perimeter relationships in a manner similar to Seale et al. (2008). By establishing a threshold on circularity of depressional features, elimination of linear or elongated depressions was possible. We also removed from consideration depressions that were located within a 2.0 m buffer zone of building depressions that were likely artifacts of the TPI analysis. Lastly, we set a size threshold of greater than or equal to 9 m2 on features to eliminate small, spurious depressions that elevation model. The geometric properties were calculated as follows: 1. The Zonal Statistics as Table tool was used to The table also contains the area of the polygon, and a count of the number of pixels per polygon (which are both equal values for 1m square pixels). 2. The results of the zonal statistics table were threshold value was used to remove polygons with a maximum depth less than 18 cm. This given pixel is a true depression in the elevation model. was calculated and each polygon was attributed length of an idealized ellipse that would contain the polygon. 4. The eccentricity of an ellipse was calculated as a measure of elongation of a potential sink, with values between 0 and 1. A threshold value of 0.98 was chosen for the eccentricity based on visual examination of the manually delineated polygons, and any depressions above this threshold were removed. 5. The circularity index was calculated. A threshold value of less than 1.7 was used in order to compare to the elliptical eccentricity results. Automated generation of depressions by Topographic Position Index (TPI)While the TPI is useful for guiding visual interpretation, it is also useful for quantitative for mapping closed depressions by thresholding, grouping, and selecting appropriate ranges of TPI values. We calculated TPI across the study area using region. We used an inner radius of 2 meters and an outer radius of 10 meters to capture small closed depressions based on previous analysis in Jefferson County, West Virginia (Young, 2007). Size Class Mean TPI SD TPI Depressions (n) < 20 m2 0.1627 5 11 0.1206 21 0.1698 18 27 14 0.1877 16 > 5000 m2 0.1576 4 Overall 0.1740 116Table 1. Mean and standard deviation of topographic position index (TPI) values found at manually mapped, and field checked depressions.

PAGE 473

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE456 difference grid. The one remaining depression had been drained by the addition of a culvert in the reconditioned DEM, and thus ended up being a false negative. Of the other depressions that were too shallow to be captured by the 18cm threshold, 12 additional depressions would have been captured if the threshold were relaxed to within the RMSE of 9 cm, or one standard error. performed very well in capturing the manually delineated depressions. Before applying area and shape threshold criteria, the total number of depressions depressions. Many of these depressions were unlikely to be true sinks since they were of a very small area. Using 9 m2 as a conservative area cutoff below which depressions occur at the edges of structures due to artifacts introduced in the LiDAR processing. Using a polygon layer of building outlines obtained from the Clarke County, Virginia, the depressions that intersected the building polygons within a 2.0 m buffer distance Finally, a large number of depressions appear within stream channels, and although meeting the area, depth, the elliptical eccentricity and circularity thresholds are for all but the most linear depressions, which are likely to be spurious. Linear depressions often are observed in gullies, ravines, and stream channels and are not true of 410,000 (a threshold visually consistent with stream channels in the LiDAR imagery), depressions that removed. Applying these threshold criteria of 1) more than 18 cm depth, 2) elliptical eccentricity less than 0.98, 2, 4) not intersecting building outlines, and 5) locations outside of known Results difference polygons. Each polygon in the manual dataset Out of the 842 total manually delineated depressions, method thresholded to 18 cm maximum depth (70%). were greater than 18 cm maximum depth totaled 600. depressions it was capable of capturing in the manually created dataset given the 18 cm depth threshold set for removing probable artifacts. In other words, 242 of the manually delineated depressions were shallower than the included in the manual dataset because all depressions that could be seen in aerial imagery were included, regardless of depth. Many of these shallow depressions were likely ponded with water, thus appearing to be (a manmade retention basin) was outside of the 0.98 eccentricity threshold, three were drained during false negatives. Level Depressions (n) 1 air photo 27 2 LIDAR air photo, LIDAR 229 4 LiDAR, air photo 17 5 116 Total Depressions: 842Table 2. Identification method and confidence of manually identified depressions.

PAGE 474

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2457 in the LiDAR DEM. More work is needed to assess mapping performance at multiple TPI window sizes as the single scale analysis presented here may not capture the variation in depression feature area and shape. had not been drained due to the uncertainty of culvert locations. Thus, additional work is needed to cull out would likely prove fruitful for mapping sinkholes from LiDAR, perhaps as much as the effort expended to map the locations of depressions themselves.References geomorphological mapping. Remote Sensing of Environment 115 (12): 2976. Angel JC, Nelson DO, Panno SV. 2004. Comparison of for determining sinkhole density: An example from in sink as an environmental estimator. Ecological De Reu J, Bourgeois J, De Smedt P, Zwertvaegher A, Antrop M, Bats M, De Maeyer P, Finke P, Van Meirvenne M, topographic position of archaeological sites in the landscape, a case study on the Bronze Age barrows in northwest Belgium. Journal of Archaeological Kozar MD, Nelms DL. 2008. Bedrock structural controls on the occurrence of sinkholes and springs West Virginia. In:Yuhr LB, Alexander EC Jr, Beck BF, editors. Sinkholes and the Engineering and Environmental Impacts of Karst. Proceedings of the Eleventh Multidisciplinary Conference; 2008 geologic and hydrologic features relating to sinkhole occurrence in Jefferson County, West 152. Available from: stream channels, the number of depressions resulting By contrast, the thresholding, grouping, and shape coincided with the 842 manually mapped depressions meter annular window, a shape threshold (circularity index) of less than 1.7, and a size greater than 9 m2. circularity of 1.7 met the eccentricity criteria of 0.98. In general, the TPI method resulted in more numerous and smaller depressions than either the manual or the redundancy between the elliptical eccentricity and the circularity threshold measures. While the eccentricity appeared to map more true sinks than the TPI method and circularity threshold, a more detailed comparative analysis is needed to fully evaluate the strengths and weaknesses of these measures. ConclusionsOverall, implementation of the TPI method is much depressions of larger area, and suffered less than the TPI method from identifying artifacts as real depressions Manual mapping method TPI mapped Average Area (m2) TPI sink, no manual equivalent 19.89 12 22.92 29.88 57 10 27.18 Total: Table 3. Results from applying thresholds to the TPI raster for TPI value, shape, and size to identify depressions.

PAGE 475

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE458Young JA. 2007. Lidar data acquisition and processing Quality of the Leetown Area, West Virginia. By Kozar, Zandbergen PA. 2010. Accuracy considerations in the Berryville, Stephenson and Boyce Quadrangles, Virginia. Virginia Division of Mineral Resources northern Valley and Ridge province, Virginia. Virginia Division of Mineral Resources Publication no. 44, one sheet (scale 1:250,000). Jenness J, Brost B, Beier P. 2011. Land Facet Jenness Enterprises. Available from: Lindsay JB, Creed IF. 2006. Distinguishing actual and artefact depressions in digital elevation data. (Vol. 1). Esri Press. 2010. Using selective drainage methods to extract Using ALSM to map sinkholes in the urbanized covered karst of Pinellas County, Florida, methodological considerations. Environmental Siart C, Bubenzer O, Eitel B. 2009. Combining digital case study on Mediterranean karst in Central Crete. ALSM to map sinkholes in the urbanized covered karst of Pinellas County, Florida. Accuracy R. 2011. Detection and Morphologic Analysis the Karst Landscape around the Maya Polity of Caracol using Airborne Lidar. Journal of Cave and

PAGE 476

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2Introduction the Lampasas Cut Plain region of the Edwards Plateau in Bell and Coryell counties (Figure 1). The plateau becomes rolling in areas proximal to streams, exposing Cretaceous carbonates from the Fredericksburg and Stream, prevailing winds are generally from the south controlling factor responsible for this difference in moisture regime (Bradley and Malstaff 2004). Soil development is minimal on the upper plateau over the Abstract and canyons in Bell and Coryell Counties, Texas. The area is located in the Lampasas Cut Plain region of interpolation of 105 km2 Installation provided depression data that were Most of the karst features within the study area are features, creating windows into karst conduits with Information Systems) and accuracy of geographically referenced data has provided the basis for more detailed terrain analysis and modeling. Research on on terrain data collection and the generation of digital models. Traditional methods such as field surveying can yield accurate results; however, they are limited by time and physical constraints. Within the study area, dense vegetation and military land use preclude extensive traditional karst survey inventories. Airborne Light Detection and Ranging data collection. The availability of high density data makes it possible to represent terrain in great detail; however, high density data significantly increases data volume, which can impose challenges with respect to data storage, processing, and manipulation. Although LiDAR analysis can be a powerful tool, natural and anthropogenic terrain modifications resulting from military use, road building and maintenance, and the natural influence of water bodies throughout the study area.DELINEATION AND CLASSIFICATION OF KARST DEPRESSIONS USING LIDAR: FORT HOOD Melinda G. Shaw Faulkner, Kevin W. Stafford, Aaron W. BryantDepartment of Geology, Stephen F. Austin State University, P.O. Box 13011, SFA Station, Nacogdoches, TX, 75962-3011, 459Figure 1. Owl Mountain study area within the Fort Hood Military Installation.

PAGE 477

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE460 over the Comanche Peak and Walnut Clay units that dominate the valleys. Military Installation known as the Owl Mountain province. The area is approximately 105 km2 and is bounded by Owl Creek to the north, Lake Belton to the for military training purposes and grazing, and the This area is home to several endangered species such as the Setophaga chrysoparia, Vireo atricapilla, a nesting songbird; Acer grandidentatum, a rare maple relict from the Pleistocene ice age found in slot canyons within the study area; and Croton alabamensis var. texensis, a rare shrub that has been documented in only a few locations in the United States (Picinich 2011).Geologic SettingThe Owl Mountain province is a karst landscape and canyons with rock outcrops, cliffs, sinkholes, caves, springs, and rock shelters. Strata from the Trinity dolomite, chert and marl alternately crop out at the surface and as scarps along incised stream valleys in the area. These exposures have been described as mounds or shoals that developed in shallow water, high energy environments; possibly as part of the restrictive structure that enabled the deposition of the evaporitic material in the Kirschberg lagoon to the southwest (Fisher and Rodda 1969). The trend of these formations, formed the model presented for Moffatt Mound (Amsbury et al. 1984; Brown 1975) (Figure 4). The Moffatt Mound from the main Edwards reef trend. The eastern section Mountain province, is thought to be a remnant of one of these isolated structures (Amsbury et al. 1984). Early geologic mapping by Barnes (1970) shows the the Comanche Peak Limestone. Most of the units are Figure 2. Stratigraphic column of the Trinity, Fredericksburg, and Washita Groups of the Lower Cretaceous. Figure 3. Location map showing the regional features influencing the depositional environment for the Fredericksburg Group on the Comanche Shelf behind the Stuart City Reef Trend. The Belton High, a smaller, structural high similar to the San Marcos Arch, provided the depositional environment for the mounds and shoals behind the main reef structure.

PAGE 478

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2461The Comanche Peak Formation is a nodular limestone and marl sequence with a maximum thickness in Bell and Coryell counties of approximately 21 m. The Comanche Peak has transitional contacts with the underlying et al.1990). Most of the Comanche Peak is not distinctly bedded, and their transitional contact can be readily distinguished in outcrop from the overlying Edwards. Fossil content and permeabilities within this unit are limestones, dolostones and marls containing mudstone, wackestone, packstone and grainstone facies with chert nodules and rudistid biostromes. These facies form the cap rock of the study area and varies from rudistid rich limestone to vuggy, porous outcrops of peloidal and oolitic wackestone to packstone. The Edwards and although these informal designations have been described and named, most mapping and descriptions differentiated. Epikarst development is spatially limited with some spongework epikarst developing in areas with a thin veneer of soil. limestone and minor marl that have wackestone, packstone and grainstone facies. Pelecypods are well as vuggy porosity present in some of the facies. Although these rocks are included as part of the Northern Edwards Aquifer, none are mapped separately in the study area. In the Owl Mountain province, the Walnut Clay, Comanche Peak and Edwards formations crop out at the surface (Figure 5). The lower valleys along creeks and rivers are covered by thicker soil and vegetative cover developing over the Walnut and lower portion of the Comanche Peak. Comanche Peak outcrops are exposed the uplands stands alone as a positive topographic feature directly coupled to the atmosphere. Precipitation is either directed into short stream segments and drainage basins vugs, sinkholes and smaller conduits. This water will dipping to the east, although the top surface of the Edwards does show some undulation. Deposition of these units began approximately 110 mya on the Comanche Shelf, which was constructed on the tectonically positive Llano and Devils River uplifts in Texas behind the Stuart City Shelf was bounded on the east and south by a relatively and Rodda 1969). formations provide the substrate for the overlying lies underneath the exposed units and is composed of limestone, dolostone and marl that were deposited in a sandstone with partial calcite cement that overlies the the subsurface, no outcrops have been documented by current studies in the study area. represent transgressive facies and are subdivided into six members; in the study area the Keys Valley Marl is the prominent member. This unit lies in gradational contact with the overlying Comanche Peak Limestone in the subsurface mudstones, wackestones and packstones, containing fossil assemblages of pelecypods, echinoderms and gastropods. In can be found near the current base level of stream channels and along the shores of Lake Belton. Figure 4. Conceptual depositional model of the carbonate Moffat Mound structure. (Bryant 2012, adopted from Amsbury et al. 1984).

PAGE 479

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE462between the Walnut Clay, Comanche Peak and Edwards units. To date, surface mapping by Reddell et al. (2011) have been delineated (Figure 5). coupled to the surface and exhibit solutional widening and overprinting by meteoric waters. Sinkholes and cave entrances are often small and associated drainage basins are spatially limited, generally covering less than one hundred square meters in area. In the study area, many sinkholes and cave entrances appear to have formed as 2012). Cave development is commonly associated with the plateaus in the eastern portion.GIS Analyses Information Systems) and accuracy of geographically referenced data has provided the basis for more detailed terrain analyses and modeling (Liu, 2008). Through spatial interpolation of available LiDAR data, depressions associated with karsting can be delineated dependent on terrain data collection and the generation surveying and photogrammetry can yield accurate results; however, they are limited by time and physical constraints. Airborne Light Detection and Ranging collection (Liu, 2008). The availability of high density data makes it possible to represent terrain in great detail; volume, which can impose challenges with respect to data storage, processing, and manipulation. LiDAR The LiDAR data used for this study were captured in March 2009). The raw data collected from the LiDAR surveys permeability unit; where it will then travel laterally to discharge as one of the numerous springs and seeps on the outer edges of the uplands. Karst DevelopmentThe level of karst development within the study area is controlled primarily by lithology; where the Edwards advanced dissolution and karst development (Figure controlled, displaying both local and regional trends (McCann 2012), with karst feature development controlled by lithologic and permeability boundaries within and between the contacts of the Walnut Clay, Comanche Peak and Edwards units. Laramide orogeny resulted in the exposure and partial erosion of these units, increasing secondary porosity and tilting the strata to the southeast (Elliott and Veni 1994). During the Miocene, faulting and subsequent uplift along the Balcones initiated the development of drainage systems and as the stream segments incised exposed rock, the intersection of fracture conduits with stream base level helped widen cavities and develop spring discharge outlets. Some karst development is controlled by bedding planes with springs, seeps, and rock shelters developing along the interface of lithological contacts Figure 5. Geologic map of the study area with mapped karst features including caves, shelter caves, springs, sinks, and seeps identified. Karst features were mapped by Reddell et al. (2011) and are part of the ongoing research at Fort Hood.

PAGE 480

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2463depressions. The original DEM was subtracted from features (Stafford et al., 2002). Depression features were delineated so that their spatial attributes could boundaries for the depression features were delineated polygons; 2) Buffer polygons to incorporate immediately overlapping boundaries; 4) Smooth the polygons to remove hard cell boundaries; and 5) Simplify the polygons to remove any extraneous bends. A single using properties that relate to whether the feature is a naturally occurring karst feature; depressions which intersected or overlapped with natural and anthropogenic the total list of delineated depressions. Depression Classification depression features in the DEM, which means that depressions associated with river channels, roadways, and Wang 2008). In order to identify the depressions associated with karsting, the delineated depressions were depressions as a result of natural and anthropogenic activities, and to identify any inherent spatial relationships that may exist. their spatial relationship to known anthropogenic and determinants. Lake Belton, Cowhouse Creek, and Owl Creek, as well as smaller lakes and ponds, were delineated from aerial imagery and any depressions that Liu 2008). Stream segments were delineated for the study accumulation raster (Figure 6). The cells with the highest 2, were isolated and delineated as being a stream feature. Any depressions brush (second return), and bare earth (third return) in an system (INS) roll, pitch, and heading information recorded by the planes POS (Position Orientation System). The database containing the LiDAR survey data was acquired Terrain Model (DTM) and Digital Elevation Model (DEM) used in this study.DTM and DEM The high density and accuracy of collected elevation points through the airborne LiDAR survey provide the basis for creating a geostatistical model that can then be further analyzed. In order to work within system memory limitations, a Digital Terrain Model (DTM) was built to simplify the management and manipulation of this large dataset (Esri, 2012). In order to perform detailed spatial analyses and utilize the numerous tools available through A 1 m resolution DEM was found to be the best model to adequately resolve karst features in the study area accuracy of the DEM was derived in accordance with for LiDAR Data (Flood, 2004). The expected horizontal LiDAR survey was 2200 m above ground level, resulting in a horizontal accuracy of less than 1.1 m. The vertical accuracy of the DEM was found by calculating the root mean square error (RMSE) between validation points of known elevation and the interpolated elevations from the 1m DEM. The RMSE of the 1m DEM for the study area is 14.04 cm; therefore, the vertical accuracy of interpolated points is less than 27.52 cm. Depression IdentificationThe study area has known karst development and by performing spatial analyses on the high resolution 1m DEM, sinkholes or depression features could be

PAGE 481

NCKRI SYMPOSIUM 2 13TH SINKHOLE CONFERENCE464depending upon where the points were taken, could have been missed by the survey. cover type. Depressions that occurred within the bare earth and disturbed herbaceous cover types were removed from the database. These areas are heavily altered by training exercises and military maneuvers and the resulting depressions would The study area contains roads of all types and sizes from paved roads, tank roads, and pipelines that transect the study geology. The Edwards and Comanche Peak formations are the only geological units in the study area known to support karst development therefore depressions occurring outside that were located within the Walnut or alluvial deposits were The depth or range of elevation values that exist within a Because the vertical accuracy of the interpolated surface was found to be within 0.275 m, depressions that were less resolved. The depression in Figure 8 is a typical karst feature within the study area; this depression would have been within the limits of the LiDAR resolution, however Figure 6. Stream segments were delineated for the study area through the creation and classification of a flow accumulation raster. The cells with the highest flow accumulation, greater than 100,000 m2, were isolated and delineated as being a stream feature and all depressions within 5 m of the stream segments were removed from the depression database. Figure 7. All major roads within the study area were digitized to facilitate classification of depressions that were influenced by roads and infrastructure. All depressions within 20 m of the major roads were removed from the depression database.Figure 8. Typical small depression found within the study area, with major and minor axes measuring 1.56 m and 1.43 m respectively, and a depth of .67 m.

PAGE 482

13TH SINKHOLE CONFERENCE NCKRI SYMPOSIUM 2465a one square kilometer search radius. The density map segments, areas of lower elevation, and within training areas currently used by the military (Figure 9). The density map of depressions that did not intersect or that the densest areas are limited to topographically high regions of the study area (Figure 10), particularly those associated with high plateaus and steeper scarps; however some appear to be associated with anthropogenic Slope AnalysisSlope analysis of the study area was performed representing the slope per cell (Figure 11). Shelter caves previously mapped by Reddell et al. (2011) were plotted on the sl