Influence of climate variability on Tampa Bay, Florida

Citation
Influence of climate variability on Tampa Bay, Florida

Material Information

Title:
Influence of climate variability on Tampa Bay, Florida
Creator:
Schmidt, Nancy Jeanne 1963-
Place of Publication:
Tampa, Florida
Publisher:
University of South Florida
Publication Date:
Language:
English
Physical Description:
ix, 116 leaves : ill., maps ; 29 cm.

Subjects

Subjects / Keywords:
Climatic changes -- Florida -- Tampa Bay ( lcsh )
Southern oscillation ( lcsh )
El Niño Current ( lcsh )
Dissertations, Academic -- Marine Science -- Doctoral -- USF ( FTS )

Notes

General Note:
Includes vita. Thesis (Ph.D.)--University of South Florida, 2001. Includes bibliographical references.

Record Information

Source Institution:
University of South Florida
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
028581315 ( ALEPH )
49419154 ( OCLC )
F51-00213 ( USFLDC DOI )
f51.213 ( USFLDC Handle )

Postcard Information

Format:
Book

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

INFLUENCE OF CLIMATE VARIABILITY ON TAMPA BAY FLORIDA by NANCYJEANNESCHMIDT A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy College of Marine Science University of South Florida August 2001 Major Profe ssor: Mark E. Luther Ph.D.

PAGE 2

Exam inin g Committee: Office of Graduate Studies Un i versi t y of South Flo rid a Tampa, Florida CERTIFICATE OF APPROVAL This i s to certify that the di sser tati on of NANCY JEANNE SCHMIDT in the graduate degree program of Marine Science was approved on May 1 1 2001 for the Doctor of Philo sop h y degree Major Professor: Mark E. Luther Ph.D. Member: G ary Mitchum, Ph.D. Member: E.S. Van Vleet Ph.D. Member: B o ri s Galperin Ph.D. Member: Rebecca John s Ph.D.

PAGE 3

DEDICATION To my grandmo ther Tina Schmidt. Your example h as helped chart the course of my life. And to Marc Reisner, whom I never met but whose writings deeply touc h ed my life.

PAGE 4

ACKNOWLEDGEMENTS I wou ld like to acknowledge m y advisor, Dr. Mark E. Lu t her for his patien ce and s upport over the many years and topic switches that pr e ceded this final product. I would al s o lik e to thank Dr. Gary Mitchum, who has provided me with mu c h guidance and encouragement. In addition I wou ld like to thank the other m e mbers of m y committee, Dr. Ted Van Vleet Dr. Boris Galperin and Dr. Rebecca John s for their support and input. It i s my pleasure to thank Dr. Peter Betzer and the College of Marine Sci e nc e for t h e financial assistance the y have provided throu g h v a rious fellowships and sc hol a r s hips. None of thi s research would have been po ss ibl e w ith out the assis tanc e and e nc o ur age ment o f my friends and family. In particular I would like to thank Brian Donahue, Teresa Greely, and Mark Hafe n for ... everyt hing. With re s pe c t to the s econd chapter ofthis disse rtation I wo uld lik e to acknow ledge the followin g in s ti t uti o n s for their s upport of thi s r esearc h : U.S. E n vi romn ental Protection Agency (EPA), NOAA Office of Global Program s, an d th e Natio n a l Wea th e r S e r v ic e Tampa Ba y Office. In a dditi on, I wou l d lik e to thank Dr. J a m es O Brien a nd Dr. Robert Li vezey as well as the anonymous r ev i ewers for th e ir thoughtfu l s u ggesti ons. I a l so would like to acknow l e d g e and th a nk my co-investigator Dr. E rin K Lipp The r esea r c h presented in Appendi x 1 i s t he res ult of our on-going collaboration. W o rkin g with Erin i s e njoy a ble satisfyi ng, productive, a nd motivational. I look forwar d to many more yea r s of collab ora tion

PAGE 5

TABLE OF CONTENTS List of Tables I V List of Figures VI Abstract IX Introduction Section 1: ENSO, Florida Precipitation and River Flow in South-central Florida 1 Section 2: Modulation of ENSO Impacts by the NAO 2 Section 3: ENSO Impacts on Salinity in Tampa Bay, Florida 2 Section 4: Policy and Management Implications of this Research 2 Appendix 1: Human Health and Climate Variability in the Tampa Bay Area 3 Climate Variability 3 El Nino-Southern Oscillation 3 The North Atlantic Oscillation 4 Impacts of Climate Variability 6 Florida and the Tampa Bay Region 7 Statistical Methods 7 ENSO Influences on Seasonal Rainfall and River Discharge in Florida 8 Introduction 8 Examples of Local-scale Imp acts in Florida 8 Background 1 0 ENSO Teleconnections 10 ENSO and Florida 11 The Focus Area 12 Data and Methods 14 Data 14 ENSO 14 Precipitation and Di scharge 16 Analyses 17 Precipitation Results 17 Winter 20 Spring 20 Summer 20 Fall 21 River Discharge Results 21 Winter 24

PAGE 6

Spring Summer Fall Discussion Seasonal Precipitation in F lorida Seasonal River Flow in South-centra l Florida Summary and Conclusions Modulation ofENSO Impacts in Florida by the NAO Introduction Background Florida and Focus Area ENSO and Florida NAO and Florida Combined Impacts ofNAO and Florida Data and Methods Results ENSO and NAO Data Precipitation and Discharge Data Analyses Rainfall Winter Rainfall Spring Rainfall Summer Rainfall Fall Rainfall Summary of Rainfall Results Stream Flow Winter Stream Flow Spring Stream Flow Summer Stream Flow Fall Stream Flow Discussion and Conclusions Precipitation Stream F low Additional Considerations Implications of Water Resources Management ENSO Impacts on Salinity in Tampa Bay, Florida Introduction Implication s of Variability in Salinity Distribution ENSO and F l orida Material s and Method s Site Characteri s tics Data ENSO 11 24 25 25 25 26 27 29 30 30 31 31 33 35 36 36 36 37 39 40 41 41 46 48 49 50 50 50 50 51 5 1 52 5 2 53 55 56 57 57 58 58 59 59 62 6 2

PAGE 7

Salinity 62 Analyses 62 Results 63 Salinity and ENSO SST As 63 Salinity, E l Nino and La Nina 63 Spatiotemporal Variability and El Nino /La Nina Correlations 64 Discussion 64 Salinity and ENSO 64 Spatiotemporal Variability in ENSO Impact s on Salinity 68 Summary and Conclusions 69 Climate Variability and Estuarine Water Resources 71 Introduction 71 Background 73 Climate Variability and the El Nino-Southern Oscillation 73 Impacts of El Nino -S outhern Oscillation in the Tampa Bay, Florida Area 73 Tampa Bay, F lorid a 74 Eco l ogica l and Econom ic Characterization 76 Proposed F re shwater Diversion 76 Hydrobiological Monitoring Plan 78 ENSO Conditions and the HBMP s Baseline Monitoring 79 ENSO Conditions and the Timing of Freshwater Avai l ability 81 Potent i al Role of ENSO Predictions 85 Implications for Tampa Bay's Estuarine Ecosystem 86 Discussion 87 Summary and Conclusions 87 New Directions 89 Identification of the Research Problem 89 Links with Climate Variability 89 Which Climate Patterns are Important for the Links / Issue of Interest? 89 Links to Climate 90 Survey Availab l e Data Sources for Measuring Links 90 Interactions Between Links 91 Addit ion a l Factors to Consider 91 Ana l yses 91 References 93 Appendices 103 Appendix 1 : Determining the Effe cts of E l Nino-Southern Oscillation Events on Coastal Water Qual ity 104 About the Author End Page Ill

PAGE 8

L I ST OF TABLES Tab l e 1 Classification of seasons based on ENSO phase 16 Tab l e 2 Descript i on of data 17 Table 3 Breakdown of years by season into those charac t erized by a positive(+) or negative() NAO state 38 Tab l e 4 Description of data 3 8 Tab l e 5 For the comparative analyses, ENSO and NAO phases fo r seasons / years from 1 950-1999 40 Table 6 The corre l ation coeffic i ents betwee n seasonal rainfall and climat i c indices for the t hree geographical regions in Florida 46 Table 7 E l Nifio La Nifia and neutra l ENSO mean precipi t ation (inches) for posit i ve neut r al and negative NAO years by seaso n for each geographical region in Florida 47 Table 8 Significant corre l a ti ons between winter st ream flow (all 30 sta ti ons) and seasonal climatic indices 5 1 Table 9 Significant correlations between spring stream flow (all 30 stations) and seasonal climatic indices 5 1 Tab l e 1 0 Significant correlations between fall stream flow (all 30 stations) and seasona l climatic indices 52 Tab l e 11 Mean r va l ue for correlations between monthly sa l inity data and ENSO SSTA for t he period 1974-1999 63 Tab l e 12 Mean r va l ue for corre l at i ons between mid -d epth salinity and E l Nifio or La Nifia mon th s for the period 197 4-1 999 64 Tab l e 1 3 Mean r-va l ue for significant correla t ions betwee n mid-depth sal i nity and El Nino months from Tab l e 2, broken out by each ba y sect ion for the per i od 197 4-1999 67 I V

PAGE 9

Table 14 Table 15 Table 16 Table 17 Mean r-value for significant correlations between mid-depth salinity and La Nina months from Table 2 broken out by each bay section for the period 197 4-1999 Withdrawal schedules for the Hillsborough River, Palm River /Tampa Bypass Canal, and Alafia River Percent of days from 1950-1999 that would have met each withdrawal level for the Hillsborough River site Percent of days from each month of the 1999 water year that would have met the minimum discharge criteria for withdrawal at the Hillsborough river site. v 67 77 84 85

PAGE 10

LIST OF FIGURES Figure 1 Winter precipitation anomali es in the Tampa Bay area and ENSO SST As, 1950-2000 Figure 2 Map of Florida, with precipitation stations indicated by filled squares and river gage stations by open circles Figure 3 Map of south central Florida focus area, with river gage stations (closed circles) and rivers identified as follows F igure 4a-h Seasonal ENSO maps of mean precipitation in Florida, showing the significance level for each stations for the approximate randomized difference of means te s t and the percent deviation from mean neutral season precipitation Figure 5a-h Seasonal ENSO maps o f mean river flow in south central F l orida, showing the significance level for each stations for the approximate randomized difference of means test and the percent deviation from mean neutral season river flow F i g ur e 6 Map of Florida with location s of precipitation stations indicat ed Figure 7 Map of west-central Florida focus area, with river gage stations (closed circles) and rivers identified Figure 8a-d Correlation maps for seasonal ENSO SSTA and precipitation showing the significance l evel for eac h station for the approximate randomized correlation test Figure 9a-d Correlation maps for seasona l NAO indices and precipitation showing the significance lev e l for each station for the app r oximate randomized correlation te s t Figure 1 Oa-h Correlation maps for seaso nal ENSO SST A indices and precipitation split by NAO state Figure 11 Mean winter precipitation in Ce ntral Florida binned by its corresponding NAO value versus ENSO SST A VI 5 14 15 18 22 34 42 43 44 54

PAGE 11

Figure 12 Map of Tampa Bay area, with estuarine salinity stations indicated by circles and tributary stations indicated by triangles 60 Figure 13a-d Maps of Tampa Bay showing, for each season the sign, strength, and significance of correlations between ENSO SST A and middepth salinity for the four bay sections for El Nifio conditions 65 Figure 14a-d Maps of Tampa Bay showing, for each season the sign, strength, and significance of correlations between ENSO SST A and middepth salinity for the four bay sections for La Nifia conditions 66 Figure 15 Map of the Tampa Bay area 75 Figure 16 Mean seasonal precipitation in the Tampa Bay area 80 Figure 17 Mean seasonal daily discharge at the Hillsborough River propo s ed withdrawal site 82 Figure 18 Seasonal precipitation, ENSO SST A, and discharge at the Hillsborough River withdrawal site for the period winter 1997 through summer 1999 83 VII

PAGE 12

INFLUENCE OF CLIMATE VARIABILITY ON TAMPA BAY, FLORIDA b y NANCYJEANNESCHMIDT An Abslract of a dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy College of Marine Scienc e Univer s ity of South Florida August 2001 Major Professor: Mark E. Luth e r Ph D. Vlll

PAGE 13

ABSTRACT Hemispherical and regional analyses of climatic patterns relating to El Nino-Southern O s cillation (ENSO) and the North Atlantic Oscillation (NAO) indicate strong responses in the southeastern United States, especially during the wintertime In this dissertation the relationships between climate variability rainfall river discharge salinity and water quality are examined and discussed with respect to Florida and Tampa Bay in particular. Seasonal precipitation and stream flow both exhibited strong responses to ENSO conditions, as shown by their relationships to Niiio-3.4 sea surface temperature anomalies for the period 1950 98. Salinity in Tampa Bay also showed strong relationships with ENSO conditions during winter for the period 1950 99. The impact of the ENSO on Florida's climate was modulated by the North Atlantic Oscillation (NAO) from 1950-99. Although NAO impacts on precipitation in Florida and stream flow in south-central Florida were not as strong as ENSO impacts, they were significant in winter. When the NAO was negative, there was a robust relationship between winter rainfall and ENSO in Florida. On the other hand, when the NAO was positive, ENSO variability was less closely associated with variability in winter rainfall. ENSO-related impacts on stream flow in west-central Florida responded at lags of up to six months to modulation by the NAO and were significantly related to stream flow during fall, winter, and s prin g Policy and management implications of this research are that ENSO impacts during the winter have the potential to influ e nce both the availability of surface water for water supply withdrawal and the e v aluation of the impacts of surface water withdrawals on the ecosystem Abstract Approved: -------------------Major Professor: Mark E. Luther, Ph.D. Professor, College of Marine Science Date Approved: _--'-7-1--/..:....1 7'--if'--O---'--( ______ lX

PAGE 14

INTRODUCTION This dissertation examines the impacts of climate variability on several parameters of interest in the Tampa Bay Florida area. Relationships between climate variability, rainfall river discharge, salinity and water quality are examined and discussed in the s ub sequent chapters. Although regional and hemispherical studies of climate variability such as El Nino -South ern Oscillation (ENSO) provide a broad picture of potential impacts, they do not adequately address the scale of variability at which decisions are made. Effective plann in g and management of not only natural disasters but also the impacts of climate-related variab ility on agriculture tourism, water resources, and human health occur mostly at the county and community lev e ls. In Flo rid a, w h ere the economy is strongly based in the touri s m and agricultural sectors and where coasta l communities are experiencing unprecedented population growth, linking larger scale re gio nal patterns of climate variability to local impac ts / conditions is particularl y rele vant. The Introduction section of the dissertation presents background information about climate variabi li ty and about Florida and the Tampa Bay area as well as a s hort descrip tion of the statistical techniques used in the analyses. Additional information on these topics is included in the subsequent chapters as needed in order to provide context for the research that is presented in those sections. The research presented in the following sections of this dissertation has been or will be published in peer-reviewed journals. SECfiON 1 : ENSO, FLORIDA PRECIPITATION, AND RIVER FLOW IN SOUTH-cENTRAL FLORIDA Hemispherical and regional analyses of climatic patterns r e latin g to E NS O indicate strong responses in the southeastern United States, especially during the wintertime. U s ing F lori da as an examp l e, this research focus e d on l ocal scale patterns with i n this region in order to exa mine the geographic var ia bility of seasonal rainfall and river discharge as related to ENSO. Forty-eight years ( 1950-98) of precipitation and river di sc harge data in Florida were classified using sea surface temperature anomaly data from the equatorial Pacific Oc ean, as occurring during an El Nifio (wam1 event), La Nifia (cold event), or neith e r (neutral). Seasonal precipitation and stream flow both exhibited s trong re spo n ses to ENSO as s hown by th ei r re l at ionship s to Nifio 3.4 sea surface temperature anomal i es. F lorida does not r espo nd as a uniform r eg ion to ENSO, particular l y with resp ec t to precipitation in the Panhandle and the sou th ernmost areas of Florida. In particular sea-

PAGE 15

sonal river discharge in s outh central Florida responds in a complicated manner to ENSO conditions; however, there are seasonal ENSO patterns. This research links the results of larger regional and h emis pherical research to a focused local-scale approach that demon s trates variability in precipitation and river flow using data and statistical techniques that a re readily available to and interpretable by water re s ource planners and managers. SECTION 2 : MODULATION OF ENSO IMPACTS BY THE NAO The impact of th e ENSO on Florida's climate varies in association with the North Atlantic oscillation (NAO). Although NAO impact s on precipitation and stream flow during the period 1950-99 were not as strong as ENSO impact s, they were significant in winter. When the NAO was negative there was a robu s t r e lation s hip between wi nter rainfall and ENSO in Florida. On th e other hand w h e n the NAO was po s iti v e ENSO variability was les s clo se l y associated with va ria-bilit y in winter rainfall. Pattern s for o ther seasons were l ess clear-cut. Spatial var iability ofNAO and ENSO impact s on winter pre c ipitation included a gradient of s tronger and more s i g nificant relationships f rom North to South Florida. ENSO-related impact s on s tream flow in west-central Florida re s ponded at lags of up to s i x month s to modulation b y th e NAO and were s ignificantl y related to st ream flow during fall, winter, and s prin g. SECTION 3: ENSO IMPACTS ON SALINITY IN TAMPA BAY, FLORIDA Estu ar ine sa lini ty distribut io n s reflect a dynami c balance between th e processes that con tr o l estuarin e circulation. At seasona l a nd l o n ge r tim e scal es, freshwater inputs into estuar i es represent the prima ry control o n salini ty distribution and estuarine circulation. ENSO conditi o n s influ e nc e seasona l rainfall and st ream discharge p atte rn s in the Tampa Bay, Florida re gion. The res ultin g variability in freshwater input to Tampa Bay influences it s seasona l salinity di str ibution. Durin g El Nino eve nt s, ENSO sea surfa c e t emperature a nomalies (SST As) were s i g nificantl y and in verse l y corre lated with salinity in th e bay durin g w inter and sp rin g These patterns reflected th e e l evated rainfall over th e drainage basin a nd the re s ultin g e levated s tream disc h arge and runoff, w hich depr essed salinity level s. Spatially, the correlations were s tron ges t at the head o f th e bay, especia lly in b ay sec tion s w ith long r es id ence t imes. During La Nina conditions, s i g nifi cant in verse correlations between ENSO SST As and sal ini ty occurred during sp rin g. Dry co ndition s and low s tream discharge characterized La Nina w inter s and springs and the hi g her sa linit y l eve l s durin g La Nina s prin gs reflected the lower fres hwater input l eve l s. S EC TION 4: POLICY AND MANAGEMEN T IMPLICATIONS OF T HIS R ESEA RCH Natural variability in the m y riad of phys ical proc esses th a t impa c t and contr o l es tuaries occurs at time scal es that typica lly m ay excee d or partly excee d many monitoring programs. With r espec t to do c umen t in g and monitoring impacts of human influ e n ces o n 2

PAGE 16

estuaries, it is therefore important to frame short-term monitoring program within the context of longer term natural variability in the environment. In the Tampa Bay, Florida area climate variability strong l y influence s seaso nal precipitation, stream flow and sa linity. In particular ENSO imp acts during the winter have the potential to influence both the availability of surface water for water supply withdrawal and the evaluation of the impacts of s urface water withdrawals on the ecosystem. APPENDIX 1: HUMAN HEALTH AND CLIMATE VARIABILITY IN THE TAMPA BAY AREA The importance of the ENSO on re g ional-scale climate variability is well recogni ze d ; however the associated effects on local weather patterns are poorly understood Little work has addressed the ancillary impact s of climate variability at the community level, which require analysis at a local scale. In coastal communities water quality and public health effects are of particular interest. Here the hi s torical influence of ENSO events on coastal water quality in Tampa Bay, Florida i s examined as a test case. Using approx imate randomized stat i st ics, s i gnificant ENSO influences on water quality are s hown particularly during winter months, with s i g nificantl y greater fecal pollution le ve ls during strong El Nifio winters and significant l y lowe r level s during strong La Nifia winters as compared to neutral conditions. Similar sign ificant patterns were also noted for El Nifio and La Nifia falls. The s uccess of the analyses in this test case demonstrates th e feasibil ity of assessing local effects associated with lar ge-sca le climate variability in any area. It also highli g ht s the possibility of u s ing ENSO forecasts to predict periods of poor water quality in urban r eg ion s which local agencies may use to make appropriate preparations. CLIMATE VARIABILITY Climate variability refer s to major, lar ge sca le modes of atmospheri c circulation. These dynamical modes are the imprints of fundamenta l g lobal proce sses, such as instabilities of the climatological mean flow or large -sca l e atmosphere-ocean interaction s (Wallace 2000). Several examples of climate variability with impacts on the North e rn Hemisphere include ENSO, the NAO, the Pacific Decadal Oscillation (PDO), and t he Arctic Oscilla tion (AO). Recent st udi es suggest that the AO encompasses many of the features of both the NAO and PDQ (Thompson and Wallace 1998 ; Thompson et al. 2000; Wallace 200 0) ; howe ve r, these re sul t s generally a re not embraced by all rese a rchers. The re search into climate variability impact s presented in th i s di ssertatio n focuses on ENSO and the NAO. El Nino-Southern Oscillation E l Nino-Southern Oscillation refer s to a g lobal weather pattern that originates in the equatorial Pacific Ocean through largesca l e interaction between th e ocean and atmosphere. During El Nifio (La N ifia) events the waters of the equatorial Pacific are anomalously warm (cool) and sea le ve l pressure lower s (rises) in the eastern Pacific 3

PAGE 17

Ocean an d rises (lowers) in the west. These pressure changes, which are accompan ied by sh ift s in tropical rainfall affect wind patterns over much of the globe (Rasmusson and Carpenter 1982; Trenberth 1991 ). Mid-latitude sy noptic weather patterns s hift equatorward (poleward) across North America during winter during El Nino (La Nina) even ts leadin g to shifts in temperature and precipitation patt erns (Rasmusson and Wallace 1983 ; Ropelewski and Halpert 1 986). E l Nino-Southern Oscillation events can be measured using different parameters, and the applicability of each method varies with the location and impacts that are in question. The Southern Oscillation Index which tracks changes in the average air-pressure difference across the equatorial Pacific traditionally ha s been used to monitor ENSO teleconnections, especially in Australia and the Indo-Pacific. Anomalous sea-s urface temperatures in the equatorial Pacific i s another method of ENSO quantification; this type of index which is calculated and published by the Climate Prediction Center is used in the research presented in this dissertation. Multivariate analyses at various locations across the Pacific bas in are also available for monitoring ENSO. Winter rainfall anomalies in the Tampa Bay area are related closely to ENSO co ndition s, as measured by SST As (Fig. 1 ). Winters w ith anomalously high rainfa ll often correspond to El Nino conditions whereas winters with anomalously low rainfall often correspond to La Nina conditions. Tlze North Atlantic Oscillation The North Atlantic Oscillation (NAO) is a major disturbance of t he atmospheric circulation and climate of the North Atlantic-European region linked t o a waxing and waning of the dominant middle-latitude westerly w ind flo w during winter. During positive (or high) NAO w in ters th e middle-latitude wester li es ar e stronger than during negative (or low) NAO winters, and anomalous southernly flow occurs over the eastern Un ited States (Hurrell 1995). The NAO exerts a strong in fluence on year-to year climate variability and there is e v id ence of lon ger-term trends in this phenomenon. The strength of the NAO typically i s measured as th e pressure difference between various stations to the north (such as Iceland) and so uth (such as the Azores) of the middle latitude westerly flow (e.g. Hurrell1995; Jon es et al. 1997). In addition, various multivariate indices derived from rotated principal components are also used to quantify the NAO (Barnston and Livezey 1987; Portis et al. 2000). The research presented in this dissertation used the C lim ate Prediction Cen ter's multivaria te indices, which are based on the methodology ofBarnston and Li vezey (1987) 4

PAGE 18

14 1 2 c 1 0 >. (ij E 8 0 c <1l 6 c 0 4 o._ 2 '() o._ 0 .... 0,) c 2 '3 c 4 <1l Q) 2 -6 -8 1 4 ----:12 c ;: 10 (ij E 8 0 c <1l 6 c 0 4 o._ 2 () Q) .... o._ 0 .... Q) c 2 '3 c -4 <1l 0,) 2 -6 -8 3 .. 2 .5 2 :E 1 .5 ::I C1) .... z ::II --0.5 9 0 (/) -0.5 1 -1.5 5 0 51 52 53 5 4 55 56 57 58 59 6 0 6 1 62 63 6 4 65 66 67 68 69 7 0 71 72 73 7 4 year 3 2.5 2 :E 1 .5 ::I C1) .... z ::II 0.5 0 I w .;:. --0 (/) (/) -1 : 0 .5 l> -1 -1 .5 2 75 76 77 78 79 80 8 1 82 83 84 8 5 86 87 88 89 9 0 9 1 92 93 94 95 96 97 98 9 9 00 year !'"''"""! Mean w i n t er precipitat i on a n oma l y (in.) -+-Winter Nif\o-3.4 sea surface temperat ur e anomaly (SS TA) Fi g ur e 1. Wint e r pr e cipitation a nomali es in the Ta mpa Ba y area and E NSO SST As, 1950-2000. 5

PAGE 19

IMPACTS OF CLIMATE VARIABILITY El Nino-Southern Oscillation teleconnections that affect precipitation and stream flow in the United States during the w inter include an equatoria l (poleward) displacement of the midlatitude jet, which increases (decreases) frontal precipitation in the southeastern United States during E l Nino (La Nina) events (Ropelewski and Halpert 1986, 1989; Kiladi s and Diaz 1989). Additionally moisture is advected from the tropical Pacific by the subtropica l jet stream into the southeastern United States during El Nino winters (Ropelewski and Halpert 1986). In th e southeastern United States summer precipitation and stream flow are affected by convective and tropical storms; during El Nino (La Nina) years, tropical storm development decreases (increases; Gray 1984 ; Bove et al. 1998). Precipitation and stream flow, consequently, during El Nino (La Nina) summers and falls are more (le ss) likely to be the result of highly localized convective storms. With respect to ENSO te l econnections, it is important to note, however that El Nino and La Nina climate anomalies are not nece ssar ily equa l and opposite. For example, Hoerling et al. ( 1997) find a lar ge nonlinear component in North American climate anomalies that is consistent with a phase shift in ENSO teleconnections. The relationship of the ENSO phenomena to globa l weather patterns is a topic of considerable interest and importance because of its documented effect on so many natural and social di sas ters-drou g ht flood, famine, hurricanes disease, to mention jus t a few. For example Epstein (1998) c ite s costs of over $1 billion for the 1991 EN SO related cho l e ra e pid e mic in Peru, mostly from l osses in sea food export and touri s t income, and th e National Center for Atmospheric Research (1994) report s that the U.S. Gulf of M ex ico r egio n experienced over $1.27 billion in loss e s associated with floodin g during th e 1 982 -83 E l Nino (warm) event. Check l ey e t al. (2000) docum e nted a 200% increase in the number of chi l dren dia g no se d with nonc holera diarrhea in Lima Peru during the 1997 -98 E l Nino event. Variations in wintertime wind patterns, temperature, and precipitation across Europe and North America are related to the NAO (Hurrell 1 995, 1996). The positive (or high) NAO pha se that h as predomin a ted over the l ast 30 years i s associated with milder winters in Europe and more se vere winters over eastern Canada and the northwest At lanti c (Hurrell 1995 1996 ; Thompson and Wallace 1998) changes in storm activity and s hifts in the Atlantic storm track (Hurrelll 1995; Hartley 1999; E l sne r et al. 2000); changes in the production of zoop l ankton and the distribution of fish (Fromentin and Planque 1 996); and changes in the growing season over Europe (Post and Sten set h 1999). Although NAO teleconnections with Europe and the Middle East have been the focu s ofmuch re sea rch (e.g., Hurrell1995 1996; Hurrell and van Loon 1997; Rod6 et al. 1 997; Stockton and G lu eck 1999 ; Cullen and de Menocal 2000; Q i an et al. 2000) less is known about the NAO's association with North American climate. For winter snowfall in New Eng land positive (high) NAO winters are generally warmer and wetter than normal with 6

PAGE 20

below-average snowfall (Hart le y and Keables 1998). In the south and central Appalachian region of the United States, winter snowfall has a significantly inverse association with the NAO (Hartley 1999). The NAO has significant positive association in the southeastern United States with winter temperature (Stephenson et al. 2000). FLORIDA AND THE TAMPA BAY REGION Florida is divided into two meteorologic and hydrologic regimes, based primarily on latitude, with different seaso nal patterns. North Florida and the Panhandle area experience winter maximums in precipitation and streamflow due to the frontal systems that impinge southward into the northern parts of the state. This frontal influence decreases to the south, where summer maximums in precipitation and river flow are generated from convective and tropical storms. Rainfall accounts for nearly all of the precipitation in Florida and averages annually about 127 152 em (50-60 in.). The Tampa Bay region is located in south central Florida on the Gulf of Mexico coast and experiences maximum precipitation and river flows during the summer months in response to convective storms and, less frequently, tropical storms. The Tampa Bay drainage area encompasses 6,583 km2 and contains a surficial aquifer that is recharged by local precipitation as well as deeper aquifers (Southwest Florida Water Management District 1998). A number of springs also contribute to the hydrologic regime and most ofthe rivers are categorized as gaining rivers. Land u s e within the region is variable; residential and commercial use dominate the coastal and estuarine areas and agriculture and phosphate mining, the more rural interior areas. As a result most of the lar ger watersheds experience rural influences in their upstream portions and urban influences downstream. Land-use impacts and diversions dams, and channel hardening all combine to create river flow regimes in Tampa Bay that are highly affected by human influences. STATISTICAL METHODS The hypotheses in this research were statistically tested using approximate computer intensive statistical tests. The significance of a statistic in a hypothesis test can be assessed using computer-intensive methods. In the research presented in this dissertation approximate randomized statistical tests were used to test the hypothesis that one variable i s unrelated to another variable. Computer-intensive tests generate the probability distribution of the test statistic by recomputing it for many ( > 1 00) artificially constructed datasets and can be used to assess significance under minimal assumptions. The ob s ervations that are tested do not need to meet the normal distribution criteria of conventional parametric statistics, nor do they need to constitute a random sample. In addition to avoiding some of the assumptions required by conventional statistical methods approximate randomization methods maximize the ability to discriminate between hypotheses because the sampling di s tribution is known (Noreen 1 989). 7

PAGE 21

ENSO INFLUENCES ON SEASONAL RAINFALL AND RIVER DISCHARGE IN FLORIDA INTRODUCTION The re lationship of the El Nino-Southern Oscillation (ENSO) phenomena to g l oba l climate patterns is a topic of considerable interest and importance because of its documented effect on so man y natural and soc ial disa sters-drought, flood famine, hurricanes disease to mention just a few. For example, Epstein ( 1 998) cites costs of over $1 billion for the 1991 ENSO-related cholera epidemic in Peru mostl y from losses in s eafood export and touri st income and the Na tional Center for Atmo s pheric Research (1994) reports that the U.S. Gulf of Mexico re g ion experienced o ver $1.27 billion in lo sse s associated with flooding during th e 1982-83 El Nifio (warm) event. Check l ey et al. (2000) document a 200% increa se in the number of children diagnosed with non cholera diarrhea in Lima Peru during the 1997 -98 El Nifio eve nt. Although regional and hemi sp herical studies of climate var i ab ili ty s uch as ENSO provide a broad picture of potential impacts th ey do not adequat e l y address the scale of varia bility at which decision s are made. Effec ti ve planning a nd mana ge ment of not only natural di saste rs but also the impacts of climate-rela t ed var i ability on agric u lture, tour ism, water resources, and human h ea lth occur mostly at the county and community le ve l s In Florida where the economy is strongly based in the tourism and agric ul tura l sectors and where coastal communitie s are experiencing unpr ecede nt ed popu l ation growth linkin g la rger scale regional patterns of climate variability to local impacts / conditions is particularly rele vant. Examples of Local-scale Impacts in Florida Understanding local ENSO patterns on the scale of counties or drainage basins has appl ic atio n for Florida's economy. Florida's two most important economic sectors are tourism and agriculture, both of which are hi g hl y affected by rainf all and river dischar ge i n coastal areas. For example, Hansen et al. (199 9) find that durin g the winter season in F lorid a, quarterly yields, pric es, production and value for c r ops such as tomato bell pepper sweet com, and sna p bean are re l ated to ENSO phase (and it s relationship to rainfall temperature, and so lar radiati o n). G iv en the fact that Florida veget a ble gro wers are the dominant and a t times only source of fres h winter vege t ab l es grown in th e U nited States characterizing the variability of precipitation in Florida benefits produc e rs and 8

PAGE 22

consumers alike. Improvements in the forecasting accuracy of ENSO coupled with an understanding of how the local area responds to ENSO-related phenomena hold forth the possibility of tailoring crop management to anticipated conditions thereby improving or maintaining yields (Brown et al. 1986; Hansen et al. 1998) This approach has been demonstrated reasonably well for other areas of the United States particularly for long range streamflow forecasting (Piechota and Dracup 1999). Florida's other economic mainstay, tourism, is also affected by climatic anomalies associated with ENSO. For example, during the 1997-98 El Nifio, elevated rainfall and cloudiness during the prime winter tourist months led to depressed revenues from tourism in the St. Petersburg Florida area as evidenced by hotel occupancy rates (Albright 1998). Water resources and human health are also potentially influenced by climate variability such as ENSO Florida has experienced a 600% increase in population since 1940 and now has over 15 million residents most of whom are concentrated in southern Florida's coastal counties. Hand-in-hand with burgeoning coastal populations comes the necessity of managing and maintaining coastal waters that are increasingly stressed by human impacts. For example, increa s ed wastewater originating from treatment plants and septic tanks increased biosolid s loads at treatment plants, and higher volumes of urban nonpoint runoff all result from population growth in coastal communities (the National Oceanic and Atmospheric Admini stra tion 1998). Furthermore, urbanization will continue to alter coastal watersheds and freshwater flows to estuaries such as F lorida's Tampa Bay and Charlotte Harbor as rural lands are converted to hou s ing developments and river flows are div e rted to meet the freshwater needs of the growing population. Human impacts on the estuarine environment often are associated with d e terioratin g water quality and increased risks to human health. Public health issues have been highlighted by the Environmental Protection Agency's (EPA) Clean Water Initiative as a r es ult of poor environmental conditions and microbial contaminants in coastal waters due to increa se d population growth and urbani za tion (EPA 1999). The vuln e rability of coastal waters to conditions associated with human health risk s also is exacerbated by unfavorable weather conditions such as increased precipitation and river flow. High levels of enteric pathogens are associated with heavy rain s and elevated river flow (Fergu so n et al. 1996; Goyal et al. 1978; Wyer et al. 1995; Barbe and Francis 1995) By understanding how climatic variability influences precipitation and river flow managers and planner s can anticipate ri sk of exposure and/or illness and provide appropriate countermea s ures (i .e., b eac h closings or warnings). Climate variability, such as ENSO and the North Atlantic oscillation plays an important role in moderating precipitation (e.g., Ropelewski and Halpert 1986 ; Gershunov and Barnett 1998; McCabe and Dettinger 1999) Although the relationship between precipitation and 9

PAGE 23

ENSO events has been broadly demonstrated within Florida and for the southeastern United States, detailed local analyses are not available for most areas. In addition impacts ofENSO-related variability in precipitation on river flow in urban watershed s are not widely known but potentially impact water quality and human health. The objectives of this paper are three-fold. 1) To examine how seasonal rainfall in Florida responds to ENSO. Much research has focused on the relationship between precipitation and ENSO at global, hemispherical, and regional scales. Many of these analyses have included Florida (e.g., Ropelewski and Halpert 1986, 1996; Livezey et al. 1997; Gershuno v and Barnett 1998; Livezey and Smith 1999); however, little has been done to examine in detail the patterns of rainfall within Florida, which may reflect both large-scale and local effects due to its subtropical location and proximity to both the Gulf of Mexico and the Atlantic Ocean. Both drought and prolonged hea vy rainfall potentially have negative impacts on many aspects of Florida's economy and quality of life. 2) To examine how river flow in Florida responds to ENSO with respect to seasonal precipitation patterns In particular, the relationships between ENSO rainfall, and river discharge in south-central Florida are assessed The numerous rivers and streams in this area discharge not just water which affects salinity regimes in coastal areas, but also urban rural, and industrial pollutants into Florida's bays and coastal beach areas. This has implications for Florida's coastal water quality, particularly with regards to public swimming areas and she llfish harvesting (Lipp et al. 2001 ). 3) To document local climate-related variability in precipitation and river flow using data and techniques that are accessible to local plann e rs and mana g ers. Because decisions regarding strategies for responding to adverse conditions associated with climate variability are made at the local level information pertinent to regional and local-scale impacts are necessary. We focus on the extreme phases ofENSO, based on Nifio-3 4 sea surface temperature anomalies (SST As) from the equatorial Pacific. The selected methodology statistically examines the relationship between seasonal ENSO SST A values and precipitation in Florida. For south-central Florida, the r e lationship between seasonal ENSO SSTA value s and river flow is evaluated at lags of up to three months. The background section provides pertinent information about ENSO and Florida. The data and methods section presents the data used and describes the methodology for analysis. Results discussion, and a brief summary follow. BACKGROUND ENSO Teleconnections ENSO teleconnections with precipitat i on a nd stre a mflow in the Unit e d States during the winter include an equatorial (poleward) displacement of the midlatitude jet, which 10

PAGE 24

increases (decreases) frontal precipitation in the southeastern United States during El Nino (La Nina) events (Ropelewski and Halpert 1986, 1989; Ki1adis and Diaz 1989). Additionally, moisture is advected from the tropical Pacific by the subtropical jet stream into the southeastern United States during El Nino winters (Ropelewski and Halpert 1986) Hoerling et al. ( 1997) however, find a large nonlinear component in North American climate anomalies with a phase shift in El Nino anomaly patterns that is not present (or weak) in La Nina patterns. In the southeastern United States, summer precipitation and streamflow are affected by convectional and tropical storms; during El Nino (La Nina) years, tropical s torm development decreases (increases; Gray 1984 ; Bove et al. 1998). Precipitation and streamflow, consequently, during El Niiio (La Nina) summers and falls are more (less) likely to be the result of highl y locali ze d convective sto rm s. ENSO and Florida In general, winters in the southeastern United States tend to be cooler and wetter during El Nino years and warmer and drier during La Nifia years (Ropelewski and Halpert 1 986, 1996; Kiladis and Diaz 1989; Sittel 1994a,b ; Livezey et al. 1997; Gershunov and Barnett 1998; Livezey and Smith 1999) These patterns are strongest in the fall and winter. There are several studies that report as part of larger regional or hemispherical analyses, associations between precipitation and ENSO for Florida (Sittel 1994a ; Livezey et al. 1997).1 Hanson and Maul (1991) focus specifically on Florida ; however, their methodology for classifying El Nifio episodes is substantially different from that used in this research. Although all the se analyses agree broadly for seasons such as winter, there are substantial differences in their interpretations of the geographic characteristics of seasona l precipitation in Florida with respect to ENSO. Complicating comparisons with their results is the var iability in their approaches to classifying ENSO events, to the number of stations examined within Florida, to dividing Florida up geographically, and to categorizing changes in precipitation (e.g., "anomalously high" vs. "very wet"). For example, Sittel (1994a) bases his analyses on precipitation data from only eight stations within Florida whereas Livezey et al. (1997) composite many stations into Florida's four Climate Districts. By focusing on precipitation within Florida, this research clarifies many of the conflicting interpretations presented in past studies. A summary of the results of these studies follows. Hanson and Maul (1991) rep ort that precipitati on during El Nino events is anomalously high during the winter and spring in Florida. During the summer, precipitation anomalies tend to be small and during the fall, inconsistent. Sittel (1994a) finds that for El Nino years, all but Florida's southernmost regions are wetter than normal during the fall. In contrast, Livezey et al. (1997) find that during November-December 11

PAGE 25

all of Florida is wetter than normal and southernmost Florida is very wet. El Nino winters are unambiguously wetter than normal in Florida, with southern Florida showing the strongest response (Sittel 1994a ; Livezey et al. 1997) During El Nino spring, Sittel (1994a) finds that north and central Florida are wetter than normal and southern Florida is drier. Florida experiences drier than normal conditions during El Nino summers (Sittel 1994a). Florida's La Nina precipitation is not a symmetric counterpart to El Nino except during the winter, when Florida is drier overall than normal (Sittel 1994a ; Livezey et al. 1997). During La Nina spring, Livezey et al. (1997) find southernmost Florida to be drier than normal whereas central and north Florida experience normal precipitation. Sittel (1994a), however, reports that Florida receives less precipitation than normal, with south Florida having a high probability of a dry spring. La Nifia summer is characterized in a similar fashion by Sittel (1994a): Florida is drier than normal, with the pattern strongest in south Florida. Both Livezey et al. (1997) and Sittel (1994a) find Florida to be drier than normal during La Nina fall; however, Sittel (1994a) reports that south Florida experiences normal precipitation. Streamflow, which integrates precipitation over drainage basins, responds to precipitation by a temporally variable combination of runoff and groundwater inputs. Analyses of the relationship between El Nino and regional streamflow in the southeastern United States demonstrate a similar, but lagged, response to the precipitation (Kahya and Dracup 1993). Similar relationships between ENSO and streamflow in South America have also been explored (Depetris et al. 1996; Compagnucci and Vargas 1998). Likewise, Zorn and Way len (1997) find wintertime responses in their analyses of mean monthly streamflow and ENSO in north-central Florida, with lags of one to two months The duration of the response is longer during El Nifio than during La Nina events. In addition, they document a late summer ENSO response in streamflow which they tentatively relate to the increased (decreased) frequency and intensity of tropical storms impacting north-central Florida under La Nina (El Nino) conditions. Sun and Furbish ( 1997) examine annual precipitation and river discharge patterns in Florida in response to ENSO; they find wet (dry) conditions and higher (lower) stream discharge in El Nino (La Nina) years. In addition, they conclude that reservoir effects might cause a six month to 1-year delay in annual stream discharge response to SST signals. The lag reflects the relationship between streamflow and the groundwater table-in Florida most streams gain water from aquifers, and their discharge directly reflects groundwater levels (Sun and Furbish 1997) Tlze Focus Area Florida is divided into two meteorologic and hydrologic regimes based primarily on latitude, with different seasonal patterns. North Florida and the Panhandle area (Fig. 2) 12

PAGE 26

Panhandle Central South Central Focus Area South Southernmost ,..., .... ,.,.., Figure 2. Map of Florida, with precipitation stations indicated by filled squares and river gage stations by open circles. Geographic designations used in this paper are also shown. 13

PAGE 27

experience winter maximums in precipitation and streamflow due to the frontal systems that impinge southward into the northern parts of the state This frontal influence decreases to the south, where summer maximums in precipitation and river flow are generated from convective and tropical storms. Rainfall accounts for nearly all of the precipitation in Florida and averages annually about 127 152 em (50-60 in.). The Tampa Bay and Charlotte Harbor regions are located in south central F lorida on the Gulf of Mexico coast (Fig 2) and experience maximum precipitation and river flows during the summer months in response to convective storms and, less frequently tropical storms. In this study, river flow data are examined from 11 of the major drainage basins in Tampa Bay and Charlotte Harbor (Fig. 3). The Tampa Bay drainage area encompasses 3550 km2 and contains a surficial aquifer that is recharged by local precipitation as well as deeper aquifers (Southwest Florida Water Management District 1998). A number of springs also contribute to the hydrologic regime, and most of the rivers are categorized as gaining rivers. Land use within the region is variable; residential and commercial use dominate the coastal and estuarine areas and agriculture and phosphate mining the more rural, interior areas. As a result, most of the larger watersheds experience rural influences in their upstream portions and more urban influences downstream. Land-use impacts and diversions, dams, and channel hardening all combine to create river flow regimes in both Tampa Bay and Charlotte Harbor that are highly affected by human influences. DATA AND METHODS The approach used in this study was to analyze the s tati s tical dependency of monthly and seaso nal precipitation and river discharge levels on ENSO phases using historical records for over 100 stations in Florida focusing on so uth-central Florida's river di sc harge (Fig. 2). Data ENSO Using the Climate Prediction Center's Nino-3.4 sea surface temperature (SST) monthly anomaly indices seasons from 1950 to 1998 were classified as extreme or neutral. ENSO extreme seasons are defined to occur when the 5-month running average, centered around the season, of the Nino-3.4 SST As exceed 10.71C. Neutral ENSO seasons are defined to occur when the 5 month running average, centered around th e season, falls between+/0.4 C (Table 1). For all seasons except La Nina spring, these thresholds exclude questionable ENSO events while providing an adequate number of cases for analyses for all ENSO pha s e seasons. This is in good agreement with Gershunov and Barnett ( 1998) who define El Nino (La Nina) years as tho se years in which the SST A from the same region were mor e than one standard deviation above (below) the mean. 14

PAGE 28

Figure 3. Map of south central Florida focus area, with river gage stations (closed circles) and rivers identified as follows: BC-Brooker Creek; RC-Rocky Creek; SC-Sweetwater Creek; HR-Hillsborough River; CC-Cypress Creek ; AR-Aiafia River; LMR-Little Manatee River; MR-Manatee River; MYR-Myakka River; PR-Peace River; and PC-Praire Creek. 15

PAGE 29

Table 1. Season Winter JFM S prin g AMJ Summer JAS Fall OND Classification of seasons based on ENSO phase. E l Nino Neutral 1958 1966, 1969 1970, 1 952, 1 953, 1 954, 1957 1973 1983 1987, 199 2, 1960 1961, 1962, 1963 1995 1998 1965 1967, 1972, 1979, 1981, 1 982, 1990, 1991, 1994 1997 1958 1969 1982, 1983, 1951, 1952 1954 1959 1987 1992 1993 199 7 1960 1961, 1962, 1963 1967, 1968, 1970, 1973 1976, 1977 1978, 1979 1980 1981, 1984, 1986 1990, 1994, 1 995, 1996 1998 1957 1963, 1965, 1969 1952 1958, 1959 1960, 1972 1982 1 987, 1991, 1961, 196 2, 1966 1978, 199 7 19 79, 1 980, 1981, 1983 1984, 1985, 1989, 1990, 1992 1 995, 1 996 1951, 1957, 1963 1965 1952, 1956, 1 958, 1959, 1968 1969, 1972, 1 976, 1960 1961, 1966, 1967 1 977, 1982, 1 986, 1987 1978 1980 1981, 1985 1991, 1994 1997 1989 1992 1996 La Nina 1955 1 956, 1971, 1974 1976 1985, 1989 1950 1 988 (1955 1956, 1965, 1971, 1974, 1975 1985, 1989)2 1955 1970 1973, 1975 1988 1998 1950, 1954, 1955 1 964, 1970 1973 1975, 1984 1988, 199 5, 1998 This i s also in good agreement with Sittel (1994b ) who defines E l N ifio (La Nifia) years based on the 5-month running mean of the Japanese Meteorological Agency SST A. The mean must be greater (less) than 0 5C for six consecutive month s s tartin g in the fall for a yea r to b e considered an E l Nifio (La N ifia) year. This approach to cla ss ifyin g ENSO events was chosen for seve ral reason s: there is no one generally accepted cl ass ification scheme, this sc heme captures the most widely r ecogn ized and accepted ENSO events and this scheme is simple to apply to th e SST A dataset. Precipitation and Discharge Sources for precipitation and river discharge data in Florida are summarized in Table 2. Only stations with more than 300 months (approximately 25 yr) of data were used from each data se t ; most of the stations analyzed had over 500 month s of data Monthly m ean values were calculated from daily observations only if valid d a ta were available for every day of the month. Similarly, seaso n a l precipitation totals and mean discharges were calculated for seasons that had valid monthly values for all thr ee months of the season. 16

PAGE 30

Table 2. Description of data. Source National Weather Service Global Surface Summary of Day National Climate Data Center Daily Surface Data South Florida Water Management District U.S. Geological Surve y Analyses Type monthly precipitation monthly precipitation monthly precipitation monthly mean daily discharge Number Maximum Mean number of period of of months/ stations record station 22 1950-98 513 42 1950-98 528 30 1950-98 523 30 1950-98 442 To test the hypothesis that ENSO phase affects seasonal rainfall and river flow in Florida, precipitation totals and mean daily river discharge during El Nifio (EN) and La Nifia (LN) seasons were compared to neutral conditions using an approximate randomized difference of means test (Noreen 1989). The significance of each randomized analysis was eva lu ated at the 95% confidence level (a= 0.05) after 10 000 iterations. Due to the paucity ofLN springs from 1950 to 1998 (Table 1; only two in almost 50 years) this ENSO season was expanded to include weaker ENSO events. This is indicated in parenthesis in Tab le 1 and includes springs with 5-month running average SST A less t han or equal to -0.4 C. Additionally, if a s tation was missing monthly data such that less than three cases of a particular ENSO season were recorded, that season was not examined. Thi s restriction tended to impact LN seasonal ana l yses, which typically had smaller numbers of cases than th e EN seasons. PRECIPITATION RESULTS Figures 4a-h illustrat es the ENSO-re lated rainfall patterns in F l orida for winter, spring, summer, and fall. In the following discussion of these patterns the significance of the difference between mean rainfall for a se ason and rainfall during the neutral season is considered, as is the overall rainfall pattern (greater than or less than neutral rainfall). Variability in rainfall patterns between closely located st ations is attributed to heterogeneous nature of precipitation, which is somewhat smoothed out by using mean rainfall over the period of record but not as much as it would be if data were combined over geograp hic regions (such as Climate District s). 17

PAGE 31

C. IEli!Mintol swmtrrw
PAGE 32

E. El Nino Sum m e r F. L a Nina Summer G. E l Nino Fa ll H. L a N ina Fa ll % Deviatio n from Ne ut ra l Pr ecipi t a t ion Sig ni f icance L eve l 0 0 75-so ... 0.01 24 < 0.05 0 6. 0 -49--25 .A 25-49 50-74 0 /::,. 0 -24-o o 1 <0.1 0 & > 0 .05 * 0 .A 75-99 A 100129 no t s i gni f icant 125 150 F i gu r e 4ah (contin u ed). Seaso n a l ENSO m aps of mean precipi t a t ion i n F lorid a showing the s i g n ifica n ce l eve l for eac h stations for t h e appr ox imat e ran do m ize d differe n ce of m ea n s tes t and the percen t d ev i atio n fro m m ean n eutral seaso n precip i ta tion. 19

PAGE 33

Winter For winter months total seasonal precipitation showed strong re sponses to ENSO phase. Statewide, EN (LN) winter precipitation totals were higher (lower) than during neutral winters. During neutral winter s, Florida experienced a SE-NW trend in pre c ipita tion, with the southern portions of the state receiving less precipitation (10 30 em; 4-12 in. ) than the northern portions ( 40-51 em; 16 20 in.). This pattern was rever sed dur i ng EN winters, when stations in southern Florida experienced as much as 50%-150% more rain and accentuated during LN winters when so uthern Florida receive d 50o/o-1 00% less rain (F igs. 4a,b). Overall 75% of the sta tion s experienced EN winter rainfall tot als that were signi ficantly greater than neutral winter rainfall and 92% received significantly less rainfall during LN winters. Spring The re spo n se of precipitation to s pring ENSO conditions was va riable During neutral sp rings Florida's precipitation was fairly uniform with total precipitation durin g April May and June typically 20--40 em (8-16 in.). This pattern did not change dramaticall y during EN springs; the majority of stations exhibited -50 to 50% change in precipitation levels (Fig. 4c ). Only 10 stations ( 11%) had significant l y higher level s of precipitation during EN springs compared to neutral springs. Southernmost Florida s howed th e s tron ges t response both in terms of percentage change (up to 100 % ) and significance ( 6 of the 1 0 sta tions exhibiting a significant relationship between precipitation and ENSO SST A are in southernmost Florida) The Panhandle and northern Florida also had higher l eve l s of precipitation although not sign ific ant l y higher than durin g neutral sp rin gs. Due to the paucity of L N springs the cla ss ification requir ements were relaxed to include sp rin gs whose 5-month runnin g mean ofNifio-3.4 SSTA were < 0.4 C (instea d of 0.7 C). During LN sp rings Florida was overall drier than n eu tr a l but rarel y significantly so (Fig. 4d). Thirteen percent of the s tation s were si g nifican tly drier at the 95% confidence level ; this increa se d to 22% at the 90% confidence l evel. All stations with s i g nificant differences from neutral precipitation levels are found in southern Florida. The Panhandle was wetter than during neutral springs; however this pattern may not be real: u s ing the stricter classification requirements 90% of the stations experienced drier conditions during LN springs including all but one s tation in the Panhandle. Summer Florida experienced its hi g hest precipitation leve l s in the summer as a result of tropical s torms and local convective thunderst o rms. Although precipitation le vels tended to be l owest in the Panhandle overall l eve l s ranged from 40 to 61 em ( 1 6 to 24 in). Durin g EN and LN summers thi s pattern did not appreciably change (Figs. 4e,f), except that during LN s ummer s the Panhandle and the we s t e rn part of northern Florida r e c e ived 20

PAGE 34

significantly more precipitation Of the 12 stations in the Panhandle regio n, 6 r e ceived significantly more rain during LN summers than during neutral summers; this number increased to 7 at the 90% confidence level. Fall Precipitation le vels were uniform statewide during neutral fa ll s with most of the st a te receiving between 10 and 20 em ( 4 and 8 in.) of rainfall. There was a s li ght tendency for th e Panhandle to be wetter. Precipitation during EN fa ll s wa s higher statewide, with both the greatest and the most significant deviations from neutral precipitation occurring in the central and southern parts ofthe state (Fig. 4g). Fifty-seven percent of the stations in central Florida received significantly higher rainfall durin g EN falls than during neutra l falls; thi s went up to 69% if stat ion s that were significant at a= 0.10 were included Precipitation in the Panhandle did not vary much or s i gnificantly from neutral fall leve l s. In contrast, 65% of the stations received less precipitation duri n g LN fa ll s than during neutral fall s. Howe ve r LN fall precipitation levels were almost indi st i ng u ishable fro m tho se durin g neutral fall s; the var iabili ty was low (-50 to 50% deviation from n eutra l precipitation) and few s tations ( 1 2%) showed a significant change from neutral precipit a tion levels (Fig. 4h) RIVER DISCHARGE RESULTS The r i ver discharge dataset compr i ses 30 s tations in 11 drainage b as in s in south-central F lorida : Brook e r Creek, Rocky Creek, Sweetwater Creek, Cypress C reek, Hill s bor o u gh River Alafia River Little Manatee River Manatee River, Peace Ri ve r M yakka River, and Praire C r eek (Fig. 3). These drain ag e basin s vary substantially in s i z e and characteristics, with some being spring -fed so me damm ed, some c hanneli ze d in places and so me relatively pristine. Land use ran ges from agricultur a l to f ull y urban to mixed drainages Because the character istic s of an individual drain age impacts the timin g of i ts response to precipit a tion analyses were la gge d up to 2 m on th s in order to accommodate the uniquene ss of each basin. Fi g ure s 5a-h illu s trate th e ENSO-re l ated river flow patt erns in the focus area for winter spr in g, summer and fall In the followin g d i sc us s io n of th ese patterns the sig ni ficance of the difference in mean river flo w for an El N ino or La N i na season ver sus river flow during th e neutral season i s considered, as i s the ove r all ri ver flow pattern ( hi gher th a n or lower than neutral river flo w l evels). Variability in river flow patterns betwee n stat ions w ithin the same drainage basin is attribu te d to both the spat i a ll y h ete r ogeneous n ature of pr ec ipit a tion eve nt s and human impacts s u c h as dan1s di vers io n s, an d di scharges Rive r flow patterns in ad jacent sub-b asin s ma y show opposite re spo n se s to ENSO condition s ; see for examp le the two s tations a long the Littl e Manatee River, wh i ch are separated b y only a few mil es. T he up s tr eam s tation 21

PAGE 35

C. El Nino C: D. La Nina Spring Spring % Deviat i on from Neutral Precipitation D 6 0 20o Signif ica nce Level .. <0.05 <0.1 0 & >0. 05 not significant Figure Sa-d. Seasonal ENSO maps of mean river flow in south -central Florida s howing the significance leve l for each station for the approximate randomized difference of means test and the percent deviation from mean neutral season river flow. (Continued on next page) 22

PAGE 36

E El Nino Summer G. El Nino Fall H. La Nina Fall % Deviation from Neutral Precipitation D .6. 0 <-2oo o .o1 49 0 l::. 0 -199-150 .... 50-99 0 6 0 149 -100 .... 100149 o e:. o -99 -50 1501 99 o 1:>. o -49 -o .o1 Significance Level < 0.05 <0.10 & > 0 0 5 not sig nif icant Figure Se-h (continued). Seasonal ENSO maps of mean river flow in southcentral Florida, showing the significance level for each station for the approximate randomized difference of means test and the percent deviation from mean neutral season river flow. 23

PAGE 37

maintains elevated ri ve r flow le ve l s durin g La N ina fa lls, winters, and s prin gs whereas the downs tream s tation experiences lower flo w levels (Figs. Sc,d,g, and h) The upstr eam s tation is in a su b-ba sin that i s dominated by phosphate mining and in dry seasons actually experiences increa se d runoff from groundwater di sc har ges (Flannery et a!. 1992) Similarly, base flows in s ub-ba si n s dominated by agri culture have increased during dry se a so n s over th e past few decad e s because of runoff associated w ith g roundwater-ba se d irri gat ion (M. S. Flannery 2000 p ersona l communication) In these exam p l es, local river flow re s pon ses to ENSO conditions are m ed i ate d b y human impacts. It i s important to n ote that our analyses cannot differ e nti ate un eq ui voca ll y b etwee n local ri ve r flow responses to ENSO conditions and to human impact s. It appears that the ENSO influence i s st ron g enough in winter t o overwhelm any others but e ven this i s unknown in a formal se n se, g i ven the "wild card ofhuman int erve ntion Therefo r e, human impacts in our urban and suburban focu s area may mask the i nflu e nc e of climate varia bility on seaso nal streamfl ow. Winter Durin g winter months, mean ri ver disc har ge l eve l s in the Tampa Bay and C harl otte Harbor r egions showed s tron g r espo n ses to ENSO phase (Figs. Sa b ). Si xty-seve n percent o f the stat ion s had sig nific antly higher mean discharg e level s during EN w i nters compared to neutral wint e r s; t h e percentage increases to 83% a t the 90% confidence interval. Discharge l evels were typ i ca ll y over 200% g r ea t e r than neutral during EN winte r s and were lagged 1-2 month s re l ative to the ENSO seas on. In contrast, mean d i scha r ge l evels during LN w inters were typica ll y 70% l ower than durin g neutral winter s, a lth ough the difference in mean di sc harge during LN and n e utr a l w int e r s was significant in only 69% of the stat ion s (73% a t the 90% co nfi dence interval) Again most s tation s s h owed th e stro n gest response at l ags of 1-2 months. Spring For EN sprin gs, mean di scharge levels we r e 136% great e r co mpar e d to neutral springs; h owever, the diffe rence in mean discharge l eve l s for EN and neutral sprin gs was rarely s i gn ificant (23% of s tati ons; Fig. 5c). T y pical l ags were 0 1 months r e l a tive to the ENSO season. Overall 87% of the s t ations exh i b it ed grea t e r mean d i sc h a r ge l eve l s during EN sprin gs than durin g neutral springs. Although almost all ( 90%) of th e s t atio n s exhibit lo wer than neutral flo w levels during extreme LN s prin gs, there were n ot eno u g h extreme LN spr ings for s tati s ti ca l analyses After subs titutin g th e broader classification of L N springs (Tabl e I), mean river discharge was also depre ssed (58% lower than during n eutra l sprin gs; Fig. Sd). This decrease in river flo w was significan t for 50% of th e stat ion s; thi s in creased to 63% at the 24

PAGE 38

90% confidence interval. Lags of 0 1 month r ela ti ve to the ENSO season were m os t common. Summer Re s ponse was typically not s ignificant for either EN or LN s ummers; however mean di s charge levels were lower than during neutr a l s ummers (Fi gs. 5e,f). During EN s ummers, 67% of the stations and during LN 57% of the stations experienced depre ssed mean river flow level s. La gs from ENSO seaso n were typically 0 1 month for both EN and LN summers. Fall Higher mean river discharge levels occurred during EN falls; flow s averaged 1 3 0% above leve ls during neutral falls (Fig. 5g). Sixtyseve n p e rcent of the stations experienced s ignificantly high e r flow levels; thi s increa ses to 83% at the 90% confidence level. Most s tations showed the s tronge s t response at lags of 1 2 month s. During LN fall s, 20% of the sta tions experienced s ignificantly higher river discharge level s; s ome of which were greater than 500 % above neutral fall levels (Fig. 5h). At th e 90% confidenc e le ve l 30% of the sta tion s experienced hi g her d isc harge le vels. Overall ho wever, the stations were almost equall y divided between tho s e with higher and th ose wit h lo wer di scharge leve l s, and t h e majori ty of stations experie nced level s within 50% of neutral levels. Stations w ith hi g her discharge l eve l s durin g L N falls tend ed to h ave 0-1 month la gs from th e ENSO seas on whereas those wi th lowe r le ve l s tended t o h ave 1 2 month l ags. DISCUSSION Seasonal pr eci pitation and ri ver dis c harg e b ot h exhibit s tron g responses to ENSO, as s hown by their relation s hip s t o Nifio-3.4 SST A. In Florida pr ec ipitation leve l s are ove rall higher (lower) during EN (LN) events, a lthou g h th ese responses d o not necessarily mirror one another durin g seaso n s s uch as s prin g and fall. In sout h-c e ntr a l Florida, river flo w re spo nds in a more complicated fashion to ENSO events, with indi v idu a l s tation s ch arac t er i zed b y varyi ng l ags and st ren gt h s of re sponse. Discharg e leve l s are elevated durin g EN falls and wi nter s; thi s p a tt ern may per s i s t in EN s prin gs. Extre me La Nifia winters a nd s prings typically ha ve lower di sc har ge l eve l s and this pattern extends into th e s ummer for m a n y s t a ti o ns. Becau se our analyses te st only loca l s i g nifican ce, we limit the di sc u ssio n to sugges tion s abo ut geog raphic a l patterns in Flor ida with respec t to E NSO-r ela ted r a in fall an d river flow s imilarities and dif ference s between our results and tho se di sc u sse d in the introduction and po ss ibl e mechanisms that mi g ht acco unt for th ese p atte rn s 25

PAGE 39

Seasonal Precipitation in Florida Comparison of our results with those of two other studies that include detailed seasonal analyses of rainfall in Florida demonstrates broad agreement, but considerable discrepancies in terms of geographical variation, especially during La Nifia conditions Similar to our results, Livezey et al. (1997) find a NWSE trend in the elevated precipitation associated with El Nifio falls and winters South and central Florida, with their normally lower winter precipitation levels experience greater increases in rainfall during El Nifio winters than does northern Florida and the Panhandle region. This broad geographical pattern is most likely the result of the southward displacement of frontal systems into southern Florida during El Nifio winters (Ropelewski and Halpert 1986 1989; Kiladis and Diaz 1989). For El Nifio spring precipitation, Sittel ( 1994a) fmds that central and northern Florida experience wet conditions and Hanson and Maul (1991) report that Florida is overall wetter. Our results in contrast, suggest that the Panhandle region and southernmost Florida experience elevated precipitation levels whereas precipitation in central Florida is not affected. Our analyses show no clearcut geographic variability in precipitation levels during El Nifi.o summers; Sittel (1994a) however finds that Florida experiences drier than normal conditions. Precipitation during El Nifio falls as indicated by our results and those of Livezey et al. (1997) indicate that precipitation follows a similar, but weaker trend as in El Nifio winters Sittel (1994a) finds the opposite trend for El Nifio falls ; his analyses indicate that southernmost Florida experiences normal levels of precipitation. During La Nifia winters there is broad agreement that conditions in Florida are drier. Our results and those of Livezey et al. (1997) document a NWSE trend, with rainfall in southernmost Florida more strongly depressed than rainfall in northern Florida and the Panhandle. This NWSE trend, however, is not as pronounced as it is during El Nifio winters. Our results suggest a different version of La Nifia spring precipitation than do either Livezey et al. ( 1997) or Sittel (1994a). Statewide, Florida is drier than normal but not significantly drier. Livezey et al. ( 1997) find southern Florida drier but northern Florida about normal. Sittel ( 1994a) indicates that south Florida is drier than normal and has a high probability of being drier than normal during La Nifia springs. Conditions during La Nifia summers do not vary significantly from neutral summers, except in the western part of northern Florida and in Panhandle area which may be slightly wetter. Sittel (1994a) finds that north and central Florida are wetter and southern Florida is drier During La Nifia falls our results suggest that there is a tendency for most of Florida to be slightly drier, with no strong geographical pattern. These depressed precipitation levels however, are not significant at most stations. Livezey et al. (1997) find precipitation levels lower throughout Florida, with the driest conditions in southern Florida. S i ttel (1994a) also finds conditions in Florida to be drier than normal during La Nifia falls 26

PAGE 40

Seusomtf River Flow in South-central Florida Scasona I river flow patterns represent a complex interaction between precipitation and river basin characteristics such as the groundwater table, aquifers natural and human diver s ions such as swamps and dams, evapotranspiration and soil moisture. These interaction s while filtering out the noisier, local aspects of rainfall, may create a more s ubtle RNSO signal that is lagged relative to rainfall. Our analyses represent a first attempt to document the relationship between ENSO and seasonal river flow in south central Florida. Our result s for south-central Florida support previously reported relationships between El Nino (La Nina) and elevated (depressed) mean seasonal river flow in the winter in northern Florida (Zorn and Waylen 1997). River flow during El Nino springs is also generally elevated but more variably and less significantly; during La Nina springs river flow levels are depressed. River flow levels during the summer are lower for both El Nino and La Nina; however the number of stations with significant relationships is low. During the fall river flow levels are elevated for both El Nino and La Nina, but more stations have significant relationships during El Nino falls (Figs. 5g,h). We tentatively explain the elevated river flow levels during fall as relating to increased tropical storm activity during La Nina and increased late fall frontal system precipitation during El Nino. Delays in the response of streamflow to sea surface temperature anomalies support this explanation; lags during El Nino falls are typically 1-2 months and during La Nina falls, 0-1 months. We would expect a faster response to the more intense tropical storm precipitation than to the frontal system rainfall. The monthly rainfall data from the Charlotte Harbor and Tampa Bay focus area also support this hypothesis. During El Nino falls rainfall is elevated relative to neutral falls in N o v ember and December. The proposed explanation for the La Nina fall river flow pattern is more complicated. In the northern part of the focus area, rainfall levels tend to be elevated although not typically significantly so, during La Nina springs, summers and falls (Figs 4d,f, and h). During persistent La Nina conditions, the drainage basins would receive elevated levels of rainfall for more than half ofthe year, cumulating in the significantly elevated river flow levels documented for La Nina falls (Fig. 4h). Also, rainfall during November, typically one of the driest months of the year in central Florida is elevated in the northern part of the focus area and is similar to neutral levels in the southern part of the focus area during La Nina falls. Focusing on south-central Florida and the Tampa Bay and Charlotte Harbor areas in particular. a plausible story can be constructed for the relationship between E N SO precipitation and river flow that uses the results of this research to add detail to the general scenarios for ENSO teleconnections already developed by others (i.e., Ropele\vski and Halpert 1986 1989) In fall and especially winter, the equatorial 27

PAGE 41

displacement of the subtropical jet stream under El Nifio conditions increases frontal precipitation in the central and south portions of the state (Kiladis and Diaz 1989 ; Ropelew sk i and Halpert 1989) Although both winter and fall preci pitat ion levels are low compa r ed to summertime level s, during extreme E l Nino falls and winters, precipitation le ve l s may double. Locall y, s tr ea m s and riv e r s respond to the higher precipitation level s s l owly, with a 1-2month lag before flow leve l s rise to their hi g h est. Levels rem ain high into the E l Nino spri n g, despite precipitation levels that are average. B y s umm er, when loc alized convective rainfall and tropical storms raise precipitation levels to their yearly high s, regardless ofENSO s t a te river flow levels also drop back to their neutral summer levels. Although the opposite "story can be constructed for La N in a teleconnections the timing of the events tend s to be different. Decrea se d frontal precipitation, due to the poleward displacement of the midlatitude jet, occurs during the winter and sometimes in to the spring (Ki ladis and Diaz 1989; Ropelewski and Halpert 1 989). Ri ver flo w le v el s are depressed wi th lags of 1-2 months, durin g the winter and spr in g; this trend continue s through the summer w ith almost no lag. Summertime precipi tation levels are indi sti n g ui s hable from neutral s ummer le ve ls; that is to say, they are high. Fall river flow levels tend to be elevated relati ve to neutral fall levels in th e n o rth e rn part of the Tampa Bay and C harlotte Harbor foc us area due to persistently but no t significantly hi ghe r rainf all le ve l s during the preceding seasons. In the sou thern part of the focus area, riv er flow le ve l s are depres sed durin g the La Nina fall, and spri n g, summer and fall pr ec ipitation levels are l ower than or close to normal. It is inter esting and important to note that our "story does not (an d cannot except in the broade st se n se) take into account the impacts of humans on urban river flow patterns. Intere s tingl y, t wo areas in Florida sta nd o ut because the y often do not respo nd in the same manner as the re st of the state to ENSO conditions: the Panhand l e area, in extreme northwestern Florid a and so uth ernmost Florida. The Panhandl e area is not part of the Florida peninsula and precipitation patterns there corres po nd more closel y to those of the Gulf of Mexico coasta l s tate s (Louisiana, Mississippi, an d Texas). E l Nino precipitation remain s elevated in the spring but i s not s i gnifica ntl y affected durin g th e fa ll. La Nina patterns in the Panhandl e also stando ut: summer precipitation is elevated pe rh aps reflectin g an increase in rainfall associated w ith tropica l storms. Southernmost F lorida which includes the Eve r gla des the Miami D ade area and the Florida Keys, experiences mor e tropical influences than th e rest of Florida. This area also exhibits anomalou s E NSO-precipitation responses. The g r eates t incre ase i n rainfall co mpared to neutral conditions occ urs in so uthernmo s t Florida durin g El Nino winter and sp rin g whereas durin g E l Nino fall the in c rease in rainfall is not as pronounce d as i t is in cent ral Florida. Durin g La Nina co ndition s on l y win ter rainf all is typically and 28

PAGE 42

significantly depressed in southernmost Florida; during the other seasons, levels are depressed but generally not significantly so. These results for Florida confirm the temporal asymmetries in rainfall and river discharge associated with ENSO (Ropelewski and Halpert 1989; Kahya and Dracup 1993; Livezey eta!. 1997; Montroy et al. 1998). Although Montroy et al. (1998) report ess entially linear (symmetric) associations of precipitation and ENSO phases during the winter season, they also discuss nonlinear teleconnections during other months and seasons between the tropical Pacific Ocean sea s urface temperature anomalies and precipitation in the southeastern United States. SUMMARY AND CONCLUSIONS This research examined seasonal precipitation in Florida and river discharge in south central Florida and have analyzed their spatial and temporal relationships with extreme phases ofENSO. The results are in broad agreement with those obtained in previous s tudies that have included Florida. However, this research pro v ides a more detailed description of these relationships regarding their spatial characteristics and temporal development. Also, datasets and methods used in this research are easily replicable by regional and local planners responsible for assessing local impacts of climate variability on water r eso urces agriculture, tourism, an d other economically and societally important topics. The results suggest that the interannual variability in the tropical Pacific Ocean, as manife ste d b y monthly Nifio-3.4 sea surface temperature anomalies, i s an importan t contributor to the seasona l variability of precipitation and river discharge in Florida at interannual scales. Also, the re su lts s ugg es t that Florida do es not respond as a uniform re g ion particularly with respect to pre c ipitati on in the Panhandle and the so uthernmost areas of Florida. Seasonal river discharge in south-central Florida r esponds in a complicated manner to ENSO conditions; howe ver this research documents significant seasona l ENSO patterns. 29

PAGE 43

MODULATION OF ENSO IMPACTS IN FLORIDA BY THE NAO INTRODUCTION In this research regional and local-scale relationships between ENSO and the NAO and water resources (precipitation and stream flow) are examined within the State of Florida. ENSO teleconnections may be modulated by other patterns of climate variability such as the North Atlantic oscillation (NAO) various climatic patterns in the Pacific and th e Arctic Oscillation (AO) For example, ENSO patterns are more closel y associated w ith rainfall surface temperature stream flow, and domestic wheat crop yield in Australia when the Inter-decal Pacific Oscillation lower s temperature in the s ame region (Power et al. 1999). In the contiguous United Stat e s ENSO rainfall signal s tend to be stron g er and more stable during preferred phases of the North Pacific oscillation (N PO ), such that consistent precipitation patterns occur during La Nina winters w hen the N PO is in its low phase whereas consistent El Nino winter precipitation is a ss ociated with the high NPO pha s e ( Gershunov and Barnett 1998b). McCabe and Dettinger (1999) also documented Pacific Decadal oscillation modulation of ENSO winter teleconnection s in the western United States. In the southeastern United States, land-falling hurricane abundance during any given year is strongly influenced by ENSO phase ; however the spatial variability in hurricane tracks can be related to NAO phase (Elsner et al 2000) The impacts of climate variability on social and socio-economic i s sues such as agriculture natural disasters and human health are related indirectly to large-scale climate variability through teleconnections; it is the resulting effects on precipitation. temperature s tream flow and other factors that affect societ y The int1uence of El Nino-Southern Oscillation (ENSO) on climate and society is well documented (e g .. Bove et al. 1998 ; Epstein 1998 ; Hansen eta!. 1998 1999 ; Solow et a!. 1998 ; Piechota and Dracup 1999; Checkley eta!. 2000; Lipp et al. in press). Although re gional and hemispherical studies of climate var i ability such as ENSO and the N AO proYide a broad picture of potential impacts they do not address adequately the s cale of variability at which many societal decisions are made (Wilbanks and Kates 1999). The research presented in this paper is an outgrowth of pre v ious explorat ion s of the connections between ENSO and precipitation, stream flow and human health (a s inf e rred from fecal coliform levels) in coastal Florida (Schmidt et a!. 200 I: Lipp et al.. in pre ss). Significant geographical variability exists in ENSO-related seasonal rainfall in Florida and lagged seasonal river flow and fecal coliform levels within the Tan1pa Bay area on the 30

PAGE 44

west-central coast of Florida. Similar to many coastal areas Tampa Bay's population growth has lead to increased impacts on the area's coastal and fresh water resources. Understanding the effects of natural climate variability on the area's water resources in order to formulate better management strategies is the motivation for our research. This chapter examines the modulation ofENSO teleconnections to Florida's precipitation and stream flow by the North Atlantic oscillation. Correlation s between climate variability (specifically ENSO and the NAO) from 1950-1999 and both seasonal precipitation from 99 stations in Florida and seasonal stream flow from 30 stations in coastal west-central Florida are analyzed. Specifically I explore the following questions: 1) Is Florida's seasonal precipitation and stream flow related to the N AO ? 2) Does the NAO modulate the impact ofENSO on Florida's seasonal precipitation? 3) Does the NAO modulate the impact ofENSO on seasonal stream flow in the Tampa Bay-Charlotte Harbor region of west-central Florida ? BACKGROUND Florida and Focus Area Florida is strongly influenced by the western side of the Bermuda/ Azores hi g h pressure system, which resides (on average) over the subtropical North Atlantic Ocean. Prevailing winds blow from the east and southeast during most months, especially over the peninsula, but local winds resulting from the uneven heating of land and w ater can reinforce or oppose this large-scale flow (Winsberg 1990) Rainfall, which accounts for nearly all of the precipitation in Florida, varies from an annual average of about 1500 mm (60 inches) in North Florida to about 1300 mm (51 inches) in South Florida. Seasonal variability in rainfall and temperature are the basi s for dividing Florida into three spatial regimes: North Florida (31 stations) Central Florida (39 stations), and South Florida (29 stations; Fig 6). North Florida experiences winter maxima in precipitation and stream flow due to frontal systems that impinge southward into the northern parts of the state. This frontal influence decreases to the south, where s ummer maxima in precipitation and stream flow are generated from convective and tropical storms. The boundary chosen between North and Central Florida reflects th e division between peninsular and continental winter climates, based on mean annual frost occurrences (Chen and Gerber 1990). South and Central Florida experience maximum precipitation and stream flow s durin g the summer months in response to convective storms and, less frequently, tropical storms. A second maximum in precipitation and stream flow occurs in the winter in response to frontal systems pushing into the state from the north The boundary between South and Central Florida roughly corresponds to th e division betwe e n 31

PAGE 45

0 0 0 0 0 North Florida Central Florida Focus ar stream flo analyses 0 0 0 0 South Florida Figure 6. Map of Florida, with locati o n s of precipitation stations indicated. Geographic designations u se d in this paper are also shown. Boxes ( ) indicate s tations with greater than 95% d ata coverage, and circles ( ) indicate those with less than 95% data coverage. South Florida stations are in black (); Central Florida stations in grey (o o ); and North Florida stations in white ( o o ). 32

PAGE 46

subtropical and tropical Florida and also conforms to a major biogeographical boundary which is marked by changes in biota and flora (e.g., Coconut palms are found only along the coast in South Florida, where the lowest temperatures during the year are above freezing) that are associated with the divergence of the Gulf Stream offshore from the Atlantic coast of Florida (Wins berg I990) Charlotte Harbor and Tampa Bay are two of the largest estuaries in Florida and precipitation is the dominant fre s hwater input controlling river flow in their drainage basins. Stream flow data are examined from 30 stations in II of the major drainage basins in Tampa Bay and Charlotte Harbor which are located along the Gulf of Mexico in south central Florida (Fig 7). The Tampa Bay drainage area encompasses 6 583 km2 and contains a surficial aquifer that is recharged by local precipitation as well as deeper aquifers (Southwest Florida Water Management District 1998). A number of spring s also contribute to the hydrologic regime, and most of the rivers are categorized as gaining rivers. The Charlotte Harbor drainage area comprises over I2, 653 km2 Land use within the region is variable; residential and commercial use are most prevalent in the coastal and estuarine areas with agriculture and phosphate mining dominating the more rural, interior areas. Several populous cities (e g., Tampa, Clearwater, Port Charlotte) are located in the downstream portions of the drainage basins. As a result most of the larger watershed s experience rural influences in their upstream portions and more urban influences downstream. Land-use impacts and diversions dams, and channel hardening all combine to create stream flow regimes in both Tampa Bay and Charlotte Harbor that are highly affected by human influences. ENSO and Florida El Nino-Southern Oscillation (ENSO) refers to a global weather pattern that originates in the equatorial Pacific Ocean through large-scale interaction between the ocean and atmosphere. During El Nifio (La Nifia) events the waters of the equatorial Pacific are anomalously warm (cool) and sea level pressure lowers (rises) in the eastern Pacific Ocean and rises (lowers) in the west. ENSO teleconnections that affect precipitation and stream flow in the United States during the winter include an equatorial (poleward) displacement of the midlatitude jet, which increases (decreases) frontal precipitation in the southeastern United States during El Nifio (La Nifia) events (Ropelewski and Halpert I986, I989; Kiladis and Diaz 1989). Additionally, moisture is advected from the tropical Pacific by the subtropical jet stream into the southeastern United States during El Nifio winters (Ropelewski and Halpert I986). In the southeastern United States, summer precipitation and stream flow are affected by convective and tropical storms ; during El Nifio (La Nifia) years, tropical storm development decreases (increases; Gray I984; Bove eta!. I998) Precipitation and stream flow consequently during El Nifio (La Nifia) summers and f all s are more (less) likely to be the result o f highly localized con v ective storms With respect 33

PAGE 47

to ENSO teleconnections, it is important to note, however, that El Nino and La Nina climate anomalies are not necessarily equal and opposite. For example, Hoerling et al. (1997) find a large nonlinear component in North American climate anomalies that is consistent with a phase shift in ENSO teleconnections. In Florida, precipitation le ve l s are overall higher (lower) during El Nino (La Nina) events, although these responses do not necessarily mirror one another during the fall and spring seasons. Summer rainfall is the least influenced by ENSO state and winter rainfall has the strongest ENSO influences (Schmidt et al. 200 I) In west -central Florida stream flow re spon ds in a more complicated fashion to ENSO events with individual stations characterized by varying lags and strengths of response. The effects of human impacts s uch as withdrawals and industrial discharges, can explain some of the complexity in the stream flow patterns. Discharge levels are elevated during El Nino falls and winters; this pattern may persist in E l Nino springs. Extreme La Nina winters and springs typically have lower discharge levels and this pattern extends into the summer for man y stations (Schmidt et al. 2001 ) NAO and Florida The North Atlantic oscillation (NAO) is a major disturbance of the atmospheric circulation and climate of the North Atlantic-European region, linked to a waxing and waning of the dominant middle-latitude westerly wind flow durin g winter. During positive (or high) NAO winters, the middle-latitude westerlies are stronger than durin g negative (or low) NAO winters and anomalous so uth ernly flow occurs over the eastern United States (Hurrell 1995). The strength of the NAO is typically measured as the pressure difference between various stations to the north (such as Iceland) and south (such as the Azores) of the middle latitude westerly flow. The NAO exerts a strong influence on year-to-year climate variability, and there is evidence of longer-term trend s in this phenomenon. Although NAO teleconnections with Europe and the Middle East have been the focu s of much research (e.g., Hurrell 1995, 1996; Hurrell and van Loon 1997 ; Rod6 et al. 1997; Stockton and Glueck 1 999; Cullen and de Menoca l 2000; Qian et al. 2000), less i s known about the NAO's association with North American climate. For winter snowfall in New England positive (high) NAO winters are generally warmer and wetter than normal with below-average snowfall (Hartley and Keables 1998). In the south and central Appalachian region of the United States, winter snowfall has a s ignificantly inverse association with the NAO (Hartley 1999) The NAO has significant positive association in the southeastern United States with winter temperature (Stephenson et al. 2000). Relationships between the NAO and Florida's climate have not been specifically documented However because the position of the Azores / Bermuda High plays a role in moderating Caribbean climate (Malmgren et al. 1998 ; Giannini et al. 2000), it i s probable 35

PAGE 48

Charlotte Harbor Figure 7. Map of west-central Florida focus area, with river gage stations (closed circles) and rivers identified. BC-Brooker Creek; RC-Rocky Creek; SC-Sweetwater Creek; CC-Cypress Creek; HR-Hillsborough River; AR-Aiafia River; LMR-Little Manatee River; MR-Manatee River; PR-Peace River; MYR-Myakka River; and PC-Prairie Creek. Stations with >95% monthly data coverage from 1950-1999 are shown with double circles. 34

PAGE 49

that the NAO affects rainfall and stream flow in the Florida. In Puerto Rico, annual precipitation is negatively correlated with winter NAO indice s (Malmgren et al. 1998). Giannini et al. (2000) find that early equatorial displacement of the Azores/Bermuda High during s ummers when the NAO is positive (high) is associated with s tronger trade winds cooler sea surface temperature s, and less precipitation in the Caribbean. Combined Impacts of NAO and ENSO Much research has explored the relationships between different climate variability patterns as well as their underlying associations wit h atmospheric and oceanic interactions that manifest themselves at different time scales and with diffe r ent areas of influence (e.g., Hurrell 1996; Huang et al. 1998 ; Kapala et al. 1998; Mache l et al. 1998; Qian et al. 2000; Wallace 2000). Although these topics currently are under investigation, research into the impacts of interrelationship s between various modes of climate variability has revealed a variety of significant modulations. Research in Europe, Australia, Caribbean, and North and South America docu ments modulation ofENSO impacts by other forms of climate var iability (NAO PDO, etc.; e.g. Rod6 1997; Power et al. 1999; Zhou and Lau 1999; Giannini et al. 2000). In North America, research has concentrated on contributions of Pacific climate variability to EN SO related cl i mate variability (Enfield 1996; Gershunov and Barnett 1998b ; McCabe and Dettinger 1999). For the southeastern United States, research demonstrates the modulation of ENSO r e lated impacts on climate by the NAO (Elsner eta!. 2000). The geographical coverage of this research, however, excludes Florida. DATA AND METHODS The approach used in this st udy was to analyze the statistical correlation of seasonal precipitation and s tream discharge levels w ith ENSO and the NAO, using historical records from over 100 stations in Florida. For stream flow, we focussed on coastal west central Florida (Fig. 6). Comparisons of mean seaso nal precipitation for var i ous combinations of ENSO and NAO were also included. ENSO and NAO Data For the correlation analyses, seasonal ENSO values from 1950-1999 were c r eated from the Climate Prediction Center's (CPC) monthl y Nifio-3.4 sea surface temperature anomaly (SST A) ind ex by applying a five-month running mean, centered around the middle of th e season (e.g., winter ENSO values were formed from December Janu ary February, March, and April SSTA data). Additionally, for comparisons, seasons from 1950-1999 were classified as El Nifio, La Nifia, or neutral. El Nifio (La Nifia) seasons were defined as those whose five-month running mean exceeds 0.7 oc (is less than -0.7 C). Seasons were considered neutral when the five-month running mean falls between +/ 0.4 C. For all seasons except La Nifia sp rin g, the se thresholds excluded qu estionable 36

PAGE 50

ENSO events while providing an adeq uate number of cases for analyses for all ENSO ph ase/seasons. This is in good agreement with Ge r s hun ov and Barnett (1998a) who define El N i no (La Nina) years as those years in which the SST A from the same region were more than one standard deviation above (below) the mean. It is also in good agreement with Hoerling et al. (1997) who define El Nino (La Nina) winters as the nine winters having the largest positive (negative) index values from an index of standardized SSTA from the region 1 60E-120W and 5N-5S. Our approach to class i fying ENSO even t s was chose n for severa l reasons: There is no sing l e generally accepted classification scheme, our sc heme captures the most widely recognized and accepted ENSO events, and applicat ion of this classification s cheme to the SSTA data is straight forward. NAO seasona l indices for the period 1950-1999 were created from the CPC's monthly multivariate ind ex, whic h is deri ved from rota ted principal components obtained from NCEP analyses and i s based on the dia g n os tic procedure established by B arnston and Livezey ( 1987). Indi ces b ased on record s from paired stations are less sensitive to seasonal di sp la ce ment s in the north-sout h dipole of the sea l evel pre ssure anomalies in the North Atlant ic (Barnston and Livezey 1 987; Livezey and Smith 1999). NAO seasona l indices we r e created from the three-month running mean, centered on the middle of th e seaso n of the CPC's multivariate index. For compari sons, NAO seasons (either negative o r p osi ti ve) were defi n ed as those whose three-month running mean index va lu e exceeds 10. 11. Seasons were considered neutral when the three -m onth running average, centered on the middle of the season, ofthe CPC's multivariate index falls between -0.1 and 0.1. This is in good agreement wit h Hurrell and van Loon ( 1997) for high NAO winters. NAO and ENSO seasona l indices were u se d in combination to examine ifNAO modu la tes seasona l ENSO impacts on rainfall and stream flow. For each seaso n two groups were formed for the period 1950-1999: those where the NAO was positive and those where the NAO was negative. The rainfall and stream flow data for these two g roup s (NAO positive and NAO negat i ve) were then separately correlated with ENSO SST A for eac h season (see sec ti on 3 .3). C l assification for of seasons by NAO state, by year, is s hown in Table 3. Preci pitation and Discharge Data Sources for pr ecipitation and stream discharge data i n F l orida are summarized in Table 4. Only s t atio n s with more than 300 months (25 years) of data from 1950-1999 were u sed from eac h data set; mo st of the statio ns analyzed h ad over 500 months of d ata. Mon t hl y mean values were calculated from daily observations only if valid data were available for every day of the month. Similarly seasona l precipitation t otals and mean discharge s were calculated for seasons that had valid monthly values for all three months of the season. 37

PAGE 51

Tabl e 3. B reakd own of years, by seaso n i n t o those character ized by a positive(+) o r nega t ive(-) NAO state Winter (JFM) Spring (AMJ) Summer (JAS) Fall (OND) NAO+ NAONAO+ NAONAO+ NAONAO+ NAO-1950 1951 1954 1950 1955 1950 1951 1950 1953 1592 1956 1951 1959 1951 1953 1952 195 4 1955 1960 1952 1961 1952 1954 1955 1959 1956 1962 1953 1963 1953 1956 1960 1961 1957 1963 1955 1964 1954 1957 1961 1967 1958 1964 1957 1967 1956 1958 1962 1968 1960 1965 1958 1969 1957 1959 1963 1972 1962 1966 1959 1970 1958 1964 1965 1973 1963 1970 1961 197 1 1960 1966 1968 1974 1964 1972 1967 1972 1962 1967 1970 1976 1965 1974 1968 1973 1965 1969 1973 1981 1966 1976 1969 1975 1966 1971 1976 1982 1969 1981 1971 1976 1968 1972 1977 1983 1970 1986 1973 1982 1974 1974 1980 1984 1971 1987 1975 1983 1977 1975 1981 1986 1975 1989 1977 1989 1978 1978 1983 1988 1977 1990 1979 1990 1979 1979 1985 1989 1979 1991 1980 1991 1980 1982 1987 1990 1980 1992 1982 1992 1981 1984 1988 1991 1985 1994 1983 1994 1984 1986 1989 1992 1987 1999 1984 1995 1985 1990 1992 1993 1985 1997 1986 1991 1995 1994 1988 1999 1987 1993 1996 1995 1993 1988 1994 1997 1996 1995 1993 1998 1997 1996 1996 1999 1998 1997 1998 1999 1998 Table 4 Descrip t ion of data. Source Type N u mber Mean number of of months / station statio n s 1950--1999 National Weather Service monthly precipitation 2 1 531 Global Surface Summary of Day National C limate D ata Center monthly precipitation 46 541 Daily Surface Data South F l orida Water monthl y precipitation 32 482 Management District U.S. Geol ogical Survey monthly mean daily 30 442* discharge Discharge data were available from the 1150 thro u g h 9 /99 38

PAGE 52

Analyses To test the hypothesis that climate variability (NAO, ENSO, or combined NAO-ENSO) affects seasonal rainfall and stream flow in Florida, precipitation totals and mean da i l y stream di sc har ge for all season/years from 1950 through 1999 were correlated with the appropriate seasonal indices using an approximate randomized correlation te s t (Noreen 1989) The significance of each randomized analysis was evaluated at the 95% confidence level (a. = 0.05) after I 0,000 iterations Due to the paucity of La Nina springs from 1950-1999 (in bold in Table 5; only three in 50 years), this ENSO season was expanded to include weaker ENSO events These are indicated in regular type in Table 3 and include springs with 5-month running average SSTA less than or equal to -0.4 C. For river flow correlation analyses were performed at lags of up to six months in order to accommodate the time-varying response of river flow to precipitation over watersheds. Results are presented for the lags that had the most sig n ificant correlations. Correlation analyses will work best for seasons that have a symmetrical ENSO relationship with precipitation; for example, for winter when precipitation anomalies are of approximately equal magnitude but opposite s ign for El Nino and La Nina conditions. Other seasons do not have thi s clear-cut line ar ity For example, El Nino falls tend to be wetter than normal and La Nina s pring s generally exhibi t patt erns with the opposite sign but not typically the same magnitude (Schmidt et al. 200 I). A se parat e comparative approach examines the potential i nfluence of nonlinear ENSO impacts. For each seaso n El Nino, La Nina, and neutral ENSO rainfall are compared for positive and negative NAO years. Due to the small number of cases for many of these events (Table 5), thi s is a descriptive attempt to illustrate differences in rainfall and to support the results of the correlation analyses. Comparisons were not made for ENSO-NAO seasons that occurred less than three times during the period 1950 through 1999. Only stat ions with 95% or better monthly coverage were used for the se comparisons. Thirty-two precipitation stations met this criterion ; ten each in North and South Florida and 12 in Central Florida (Fig. 6). 39

PAGE 53

Table 5. For the comparative analyses, ENSO and NAO phases for seaso ns /y ears from 1950-1999 La Nina Spring was expanded to include 5-month running averages of Nino-3.4 SST As with values l ess than or equal to -0 .40. This was done to create a subsample size large enough for analysis. The three years whic h meet the "original" criteria for La Nina Spring are highlighted in bold La Nina Neutral ENSO E l Niiio WINTER (JFM) NAO1955, 1971, 1985 1957, 1962 1963 1958, 1966, 196 9, 1970, 1965 1979 1987 NAOn 1956 1972 1981, 1982 1998 NAO+ 1950 1974 1976, 1989 1953 1954 1961, 1967 1973' 1983, 1992, 1995 1999 1990, 1991, 1994 1997 SPRJNG (AMJ) NAO1950 1955 1971, 1975 1951, 1952 1 96 1 1967 1958, 1969 1982, 1 983, 1985 1988 1968 1973, 1977 197 9, 1993, 1997 1980, 1984, 1995 1996 1 998 NAOn 1959 1978 1 987 NAO + 1956 1965, 1974 1989, 1954 1960 1962 1963, 1992 1 999 1970 1976 1981, 1986 1990 1994 SUMMER (JAS) NAO1988, 1998 1952 1958 1960 1962 1957 1965, 1987 1966 1988, 1974, 1978 1979 1980 1981, 1985 NAOn 1959 1 984, 1996 NAO+ 1955, 1970 1 973 1975, 1961, 1967 1983 1989 1 963, 1969 1972 1982, 1999 1990, 1992, 1995 1991 1997 FALL(OND) NAO1 950 1955 1970, 1973 1952 1960 1961, 1980 1963, 1965, 1968, 197 6, 1988 1995 1981, 1985 1989 1 992 1977 1987 1997 1996 NAOn 1975 NAO+ 1954 1964 1984 1998 1956 1958 1 959 1966, 1951, 1 957 1 969, 1972, 1999 1967 1978 1 982, 1986 1991, 1994 RESULTS Results for rainfall and stream flow are discussed separately, wi th emphasis on the winter seaso n patterns. Except where specifica ll y noted in th e Results section, correlation values (r values) for the approximate randomized correlations are considered significant i f th e probability of the correlation arising by chance is less than or equal to 5%. 40

PAGE 54

Rainfall Winter rainfall Winter season precipitation in Florida is related to ENSO state and this rel ationship i s moderated by the NAO. There is, however considerable spatial variability within Florida with respect to winter rainfall variability. Overall correlation between ENSO SST A and rainfall is positi v e and significant for all of F lorida during winter (Fig 8a; Table 6) reflecting the linear relationship of low rainfall during La Nifia events and high rainfall during El Nifio events. In Figure 8 the filled (unfilled) symbols represent positive (negative) correlations between seasona l rainfall and ENSO SST A. Different symbols are used to represent the significance level of the correlations, and the size of the symbol corresponds to the strength of the correlation value. Correlation values are greatest, on average in South Florida and range from 0.45 in North Florida to 0.58 in South Florida Greater than 90% of all stations in Florida have correlations that are significant. The mean r-value statewide is 0.52. Correlations between winter rainfall and NAO indices indicate a genera ll y negative relationship within Florida (Fig. 9a; Table 6). This relationship is most apparent in Centra l Florida, where 50% of the stations have negative correlat ion s that are significant. R-values, however average only 0.24 in Centra l Florida Althou g h correlations also t end to be negative in North and South Florida they are l ess lik ely to be significant and have even lower r-values than do correlations within Central Florida. The s trength and sig nificance of correlations between ENSO SST A and winter rainfall ar e enhanced only for winters during which the NAO is negative (Fig. 1 Oa and b ; Table 6). A north-to-south pattern is evident. In South F l orida a ll 29 stat ions have correlations that are significant and 28 have correlation s that are significant at the 99% confidence level. The mean r-value is 0. 71. In Centra l Florida 87% of stations are significant ( 62% at the 99% confidence l eve l ) and the mean r-value is 0.62. In North F lorid a 93% of stations are s i gnificant ( 45% at the 99% confidence level) and the mean r-value is 0.5 2. Statewide the mean r-value for the corre l ation between rainfall and winter ENSO SST As is 0.63 (0.45) when the NAO is negative (po s itive). Comparisons of mean precipitation for various combination s ofNAO-ENSO winters (Table 7) sugges t that La Nina rainfall is not influenced by NAO state in South F l orida. In Central and North Florida rainfall i s elevated during La Nina winters when the NAO i s negative. During neutral ENSO winter conditions rainfall is hi g her in Florida when the NAO is also neutral. Rainfall during El Nino winters in Florida is e l evated when th e NAO is negative. 41

PAGE 55

A. Winter B Spring C Summer D. Fall Correlation ranges Significance level 0 0 /:::,. 0 -1.00--0.75 0 .$. 0.01 0 0 .6 0 -o .74--o.5o 0 0 t>. 0 -o.49 o.25 0 .$. 0 .05 & > 0.0 1 0 0 IJ. 0 -0.24 0 .01 * * > o.o1 & 0.05 4 0.01-0. 24 4 0.25-0.49 0 not signif ican t ... o.5oo.74 ... 0 .75-1.00 Figure Sa-d. Correlation maps for seasonal ENSO SST A and precipitation, showing the significance level for each station for the approximate randomized correlation test. 42

PAGE 56

A. Winter B Spring C Summer D. Fall Correlation ranges Significance l eve l 0 0 6. D -1.00--0 .75 0 .$. 0 .01 0 0 6 0 -o.74--o 5o 0 0 6 0 -o.49 o.25 0 .$. 0 .05 & > 0 .01 0 0 6 0 o.24 o.o1 * * > -o .o1 & < 0 .01 .6. 6. .s.0.1 0 & > 0 .05 0 .01 0 .24 "' 0 .25-0.49 0 n o t si gn i f icant .... o 5oo.74 .6. 0.75 1.00 Figure 9a-d Correlation maps for seasonal NAO indices and precipitation, showing the significance level for each station for the approximate randomized correlation test. 4 3

PAGE 57

0 0 0 0 0 0 0 0 * .,..., A N AQ + W inter s C NAQ + S p r i n g s Corr e lat io n ranges 6 D 1.00--0 75 6. 0 o .74--o 5o !!>. 0 -0 49 -0.25 0 -o .24 o .o1 * > -o .o1 & <0 .01 .. 0 .01 0 .2 4 A 0 .250 49 0 .500 74 0 .7 5 1 00 B NAo-Winters D NAo-Sp r i n g s Sig nif ica n ce level 0 5 0 .01 0 5 0.05 & > 0.01 6 50. 1 0 & >0 .05 0 not significant Fig ure l O a -h Correlati o n maps for seasona l ENSO SSTA indices a n d ..-..,. precipita tion s plit by NAO state. The m aps show the sig n ificance l eve l for each statio n for the approxi mate r a nd omize d corre l ation test. (Conti n ued o n next page) 44

PAGE 58

0 0 0 0 0 0 0 0 * d"'>"' E. NAO+ Summers G. NAQ+ Falls Correlation ranges 6 0 -1.00--0 .75 t::. 0 "0.74--0.50 t:> 0 o.49 -o 25 6 0 "0.240.01 * > "0.01 & < 0 .01 ... 0 .01-0.24 0.25-0 .49 ... 0 .500.74 ... 0 .751.00 0 0 ... 6 0 F. NAoSummers _,... .,.. H. NAOFalls Significance level 5 0 .01 5 0 0 5 & > 0 .01 5 0 10 & > 0 05 not significant Figure lOe-b (continued). Correlation maps for seasonal ENSO SSTA indices and precipitation, split by NAO state. The maps show the s ignificance level for each station for the approximate randomized correlation test. 45

PAGE 59

Table 6. The correlation coefficients between seasona l rainfall and climatic indices for the three geographica l regions in F lorid a. WINTER (JFM) Mean correlation coefficient ENSO when between rainfall and ENSO NAO NAO positive Nort h Florida 0.45tt t -0.09 0.41 tt Central Florida OAitt -0.24tt 0.39t t South Florida 0.58ttt 0.08 0.54t tt SPRiNG (AMJ) Mean correlation coefficient ENSO when between rainfall and ENSO NAO NAO posi tive North Florida 0.16 0.03 0.05 Centra l Florida 0.13 0.08 0.05 South Florida 0.20 0.07 0.17 SUMMER (JAS) Mean correlation coefficient E NSO when between rainfall and ENSO NAO NAO positive North Florida 0.16 -0.03 0.20 Central Florida -0.01 -0.07 0.00 South Florida -0 03 -0 .17 0.02 FALL(OND) Mean correlation coefficient ENSO when between rainfall and ENSO NAO NAO positive North Florida 0.18 0.15 0.20 Central Florida 0.24 -0.05 0.17 South Florida 0.28 tt 0.27 tt 0.20 tt greater than 50% of sta t ions have sig nificant correlations at the 95% confidence level ttt greater than 50% of stations have significant correlat i ons at the 99% confidence level Spring rainfall ENSO when NAO negative 0.52 tt 0 62 t tt 0. 71 ttt E NSO when NAO negative 0.27 0.1 8 0.22 E NSO when NAO negative 0. 1 2 0.01 0.08 E NSO when NAO negat i ve 0.15 0.32 0.34 Relation ships between climate variability (ENSO and th e NAO) and rainfall in Florida are le ss distinctive in seasons other than winter. Overall spring rainfall in Florida exhibits a positive correlation w ith ENSO SST As (Fig. 8b; 88% of all s tations have positive correlations). The relationships, h owever, are not as strong or as significant in any part of F lorida as they are during the winter (Tab le 6). South Florida has the strongest relationships, with a mean r-va l ue of 0.32 for the 38% of stat i ons wi th significant correlations. Only I 0 and 21 percent, respectively, of stations in North and Centra l Florida have significant correlations. Twenty-seven percent of stations in Central Florida have negative (but not significantly so) correlations between spring rainfall and ENSO SST A. Most of these stations are in the interior of Central Florida. 46

PAGE 60

T able 7 El N ino, La N i n a and n e utral ENS O m ean p r ecipitatio n (in ch es) for posit i v e n e utral and nega tive NAO years, by season for each geograp h ica l regio n i n Florida. W INTER La Nina Neu tra l ENSO El Nino NAO+ NAOn* NAO-NAO+ NAOn NAO-NAO+ NAOn NAONorth Florida 8.39 ** 9 .21 12.62 13.71 12.16 1 5.81 16.08 Central F lorida 4.31 -6.2 2 7.05 9.11 11.07 12. 26 -1 2 52 South F lorida 3.79 3.06 7.1 5 7.45 6.33 1 0.71 -11.6 2 (max. n)*** 3 -5 8 3 6 4 -5 SPRlNG La Nina Neutral ENSO El Niiio NAO+ NAOn NAO-NAO+ NAOn NAO-NAO+ N AOn NAONorth Florida 14.60 -11.78 11. 86 1 2.76 --15.18 Central Florid a 12.38 10.46 12.60 11.20 --13.32 South Florida 13.79 12.60 15.82 15.40 --18.4 7 (max. n) 5 -7 1 0 1 3 -6 SUMME R La Nina Neutral ENSO El Nino NAO+ NAOn NAO-NAO+ NAOn NAO-NAO+ NAOn N AONorth F l orida 1 9.28 -18.43 18.32 19.88 17.92 -22. 1 4 Centra l Florida 20 .36 -20.73 20.64 22.11 21.25 22 .40 South Flor ida 20.39 -1 9.9 1 21.01 23.60 20.69 2 0.4 1 (max. n) 5 -7 3 1 2 6 -3 FALL La Nina Neutra l ENSO El Nino NAO+ NAOn NAO-NAO+ NAOn NAO-NAO+ NAO n NAONorth Florida 7 .86 9.06 8.94 -9.86 11.11 1 0.73 Cent ra l Florida 6.00 -7.95 6.95 -6.96 8.48 -9.41 South Florida 1 0 .38 -7.88 11.41 -7.5 1 1 2 32 10.78 (max. n) 5 -6 6 -9 8 -7 n =neutral ** -=no mean was formed b eca u se th e numb er of cases was fewe r than three. *** The max imum numbe r of cases for each seaso n's ENSO/NAO cases used t o form the mean. Correlations between spring rainfall and NAO indices indicate a gene rall y po si ti ve bu t weak relationship within Florida (Fig 9b). Very few of the s tation s (5%) h ave significant correlations and overall r -val ues are l ow (with a mean value of approximate 0.13) No obviou s geographical patterns stand out. 47

PAGE 61

The strength and significance of correlations between ENSO SST A and spring rainfall are elevated during spring seasons when the NAO is negative compared to when the NAO is positive (Fig. I Oc and d, Table 6) Ninety percent of all stations have positive correlations between rainfall and ENSO SST As when the NAO is negative, and 26% of those correlations are significant. In comparison 66% of all stations have positive correlations when the NAO is positive and 3% of those correlations are significant. There may be a north-to-south pattern; 32% and 26% of stations in North and Central Florida have significant correlations and mean r-values of0.42 and 0.41, respectively, compared to 17% in South Florida and a mean r-value of 0.49. R-values for the correlations when the NAO is positive are much lower (about 0.24 statewide). Comparisons of mean precipitation for various combinations ofNAO-ENSO springs (Table 7) suggest that in North and Central Florida the driest (La Nifi.a) and wettest springs (El Nifi.o) are associated with negative NAO conditions. Comparisons in South Florida are complicated; the driest springs occur when ENSO SST As are low and the NAO is high (La Nifi.a-NAO positive). When the both ENSO and the NAO are neutral, spring rainfall is high. Due to the paucity of cases for two of the three El Nifi.o-NAO cases, it is difficult to conjecture about how the NAO impacts rainfall during El Nifi.o springs. Summer rainfall Summer rainfall in Florida is highest in South Florida and lowe s t in North Florida, although the amount of geographical variability is much less than it is during the winter. Overall in Florida correlations between summer rainfall and ENSO SST As typically are negative, not significant, and have low r-values (Fig. 8c). In Northern Florida, correlations are overwhelmingly negative (94%) and r-values average 0.20. Sixteen percent of the stations in North Florida have significant negative corr e lations; this increases to 48% if the confidence level is reduced to 90%). However both Central and South Florida have no significant or notable correlation patterns. Similar to the results for ENSO correlations with summer rainfall correlations between NAO and rainfall during the summer tend to be negative, not significant, and have low r-values. South Florida has the highest percentage of stations with negative correlations (83%), the highest mean r-value (0.24), and 24% of stations are significantly negative correlations. North and Central Florida have no significant or notable correlation patterns. In North Florida, the typically negative correlations between ENSO SST As and summer rainfall are not particularly sensitive to NAO state (Fig. I Oe and f Table 6). In Central Florida neither ENSO nor NAO appears to have a strong influence on summer rainfall. However in South Florida the negative phase of the NAO is associated with predominantly negative (but not typically significant) correlations between ENSO SST A s 48

PAGE 62

and rainfall ; this association i s not a p pare n t w hen the NAO i s positive. Interestingly, the east (A tlantic ) coast of South F lorida ha s positive ( but not typically s i gnificant) corre l at i ons b etwee n r ainfall and ENSO SST As w h e n the NAO is positive (Fig. 1 Oe). Due to the pauci ty of cases for tw o of the thre e La N ina-NAO cases it is difficult to conjecture about how the NAO impact s r a infall during La Nina summe r s (Tab l e 7). For n e utral ENSO ca s e s, rainfall i s elevate d ( d ep re sse d) when the NAO i s n egat i ve (posit i ve) for all of Florida In South Florida ra in fall i s the sa me for b o th NAO sta t es. In North Florida, rainfall is elevated when the NAO i s n egativ e. Fall rainfall Fall rainfall in Florida is typically le ss than spring and summer rainfall and similar to winter r ainfall l evels Variability in rainfall levels during the fall is not as great as it is during the winte r Overall, co rrelation be tween ENSO SST A and rainfall i s positi v e and s ignifi cant for all of Flo rid a during fall (F i g 8d), but the s trength of the correlations i s lower than for the winter seaso n (Fig. 8a). Ninety s i x p erce n t of statio ns have positive corre l a tion s and the mean r-value for the 47% of stat io n s wit h s ignificant positi v e correlations is 0.3 3 A tren d of increasing sig n ificance ( b o th in te rm s of percent of statio n s with s i g nificant corre l atio n s and mean r-va lu e) from n ort h to so uth exists. The Panhandle region (the wes t ern p o rtion ofNorth Florida) ha s the weakest relationship between fall rainfall and EN SO w h e rea s South Florida has th e strongest. In all but northwest Central Florida corre l atio n s between NAO and fall r a infall are po s itive (Fig. 9d) In northwest Cent ra l Flori da there i s a cluster of negativ e correlations that are significan t and stand in s h arp contras t to the po sitive bu t not typically sig nificant corre l at i o n s in the rest of Central Florida. The mean r-v alue for these negative s i g nifi cant corre lation s is 0.40. In South F lorida, 52% of s tation s have positive and significan t co rrel a tion s between rainfall and the NAO and have a m e an r-value of0.38. In N orth Florida correlations are positive but not gene r a lly s ignificant (o nl y 1 6% of stations). In North Florida the typically po s iti ve corre lations be tween ENSO SST As and fall rainfall are not particul arly se n s itiv e to NAO sta te (Fig. 1 Og and h ; Tab le 6). Both Ce ntral and South Florida exhibit stro n ge r more s i g nificant correlations betw ee n fall rainfall and ENSO SST As when th e NAO is negative. Forty one and 45% of stations h ave s i gni ficant positi ve correlation s and mean r-values of0.46 and 0.47 re spectively in Centra l and South Florida The unu s u a l NAO corre lations in northwe st Ce ntral Florida do not seem to be reflected in t h e comb ine d NAO -E N SO correlation analyses for the fall season Comparisons of mean precipitation for various combination s ofNAOENSO indices (Ta ble 7) s ugge s t that La Ni na falls are wetter s tate wi de when the N AO i s positive and drier when the NAO i s negative. NAO state does not appear to influence rainfall during 49

PAGE 63

neutral or El Nifio ENSO cases in North or Central Florida In South Florida, rainfall levels are higher when the NAO is positive, regardless ofENSO state Summary of rainfall results Winter rainfall is significantly correlated with winter ENSO conditions in Florida, with strongest relationships in South Florida Relationships during other seasons are weaker Seasonal rainfall is significantly correlated with NAO state only in Central Florida during the winter and South Florida during the fall; however, precipitation has an overall inverse (but not significant) relationship with the NAO state during winter. Correlations between winter rainfall and ENSO are greater and more significant when the NAO state is negative. Modulations during other seasons are weaker StreamFlow During winter, spring, and fall, stream flow levels in west-central Florida typically are low (Bendis 1999), reflecting the low rainfall levels during these seasons Stream flow peaks during the sununer, usually in August, and a second peak may occur during the winter months. Although discharge varies from stream to stream, sununer levels typicall y are 3-4 times higher than during the rest of the year However there can be great interannual variation in seasonal stream flow especially during winter The results within this section are all with respect to lagged correlations. For example, winter ENSO SST As are correlated with stream flow contemporaneously and with lags up to 6 months in the future The results for all thirty stations are reasonably homogenous for each lag and therefore we have combined the results for each lag in order to e x plore the response of the region as a whole to ENSO and the NAO. Winter stream flow Wint e r stream flow levels in both the Charlotte Harbor and Tampa Bay areas are positively correlated with both the previous sununer and fall ENSO SST A indices at the 99% confidence level (Table 8) NAO seasonal indices are typically negatively correlat e d with winter stream flow but not significantly so at any lag For NAO positive winter s, stream flow levels are positively and significantly correlated with ENSO SST A sununer and fall indices. For NAO negative winters stream flow levels are positively correlated with sununer E NSO SSTA indices (at the 90% confidence level), positively correlated with fall ENSO SSTA indices (at the 99% confidence level), and positively correlated with contemporaneous winter ENSO SSTA indices (at the 95% confidence level) Spring stream flow Spring stream flow levels in both the Charlotte Harbor and Tampa Bay areas are positively correlated with contemporaneous ENSO SST A indices at the 95% confidence level and with the previous winter ENSO SSTA at the 90% confidence level (Table 9) NAO seasonal indices are typically negatively correlated with spring stream flow but not 50

PAGE 64

Table 8. Significant correlations between winter stream flow (all 30 stations) and seasonal climatic indices. ENSO SSTA Mean r-value For NAO positive conditions ENSO SSTA Mean r-value For NAO negative conditions ENSO SSTA Mean rvalue Summer 0.57*** Summer 0 68*** Summer 0.4 1 * = s ignificant at the 90% confidence level ** = significant at the 95% confidence level *** = significant at th e 99% confidence level Fall 0 57*** Fall 0.55** Fall 0 64*** Winter 0 57** Table 9. Significant correlations between spring stream flow (all 30 stations) and seasonal climatic indices. ENSO SSTA Mean r-value For NAO n e gative condition s ENSO SSTA Mean r-value Winter 0 30* Winter 0.59** = s ignificant at the 90% confidence level **= significant a t the 95% confidence level * = s ignificant at the 99% confidence level Spring 0 37** Spring 0 39* significantly so at any lag For NAO positive winters stream flow levels are not sign ifi cantly correlated with ENSO SSTA seasonal indices at any lag For NAO negati v e springs, stream flow levels are positively correlated with contemporaneous ENSO SST A indices (at the 90% confidence leve l ) and positively correlated with the preceding winter ENSO SSTA indices (at the 95% confidence l evel). Summer stream flow Summer stream flow levels in both the Charlotte Harbor and Tampa Bay areas are not significantly correlated seasona l ENSO SST A indices any lag NAO seasona l indices are typically negatively correlated with summer stream flow but not significant l y so at any lag Stream flow levels during summer are not significantly correlated with ENSO SST A indices at any lag for either NAO positive or NAO negative conditions. Fall stream flow Fall stream flow levels in both the Charlotte Harbor and Tampa Bay areas a r e positively corre lated with both the previous spring and summer ENSO SST A indices and the contemporaneous fall ENSO SST A indices (Table 1 0). Spring NAO seasonal indices are negatively and significantly correlated with fall stream flow. For NAO positive falls, stream flow l evels are positively corre l ated with ENSO SSTA summer indices For NAO negative falls, s tream flow levels are positively correlated with spring ENSO SST A 51

PAGE 65

indices (at the 90% confidence level) and positively correlated with contemporaneous fall ENSO SSTA indices (at the 90% confidence level) Table 10. Significant correlations between fall stream flow (all 30 stations) and seasonal climatic indices. ENSO SSTA Mean r -v alue NAO Mean r-value For NAO positive conditions ENSO SSTA Mean r-va lue For NAO negative conditions ENSO SSTA Mean r-value Spring 0.37** Spring -0.44* Summer 0.42* Spring 0.35* = significant at th e 90% confidence level **=si g nificant at the 95% confidence level *** =significant at the 99% confidence level DISCUSSION AND CONCLUSIONS Precipitation Summer 0 29* Fall 0 .38* Fall 0 28* Overall, ENSO-related seasonal influence on precipitation is greater than is influence associated with the NAO (Table 6). Relationships between seasonal rainfall and ENSO are strongest and most significant in the winter in Florida and weaker during sp rin g and fall. Summer precipitation is not significantly related to ENSO conditions and i s dominated by the locali zed convective storms characteristic of this seaso n in Florida Northern Florida is an exception to this summer pattern; it does not have the same maritime influence as peninsular Florida and has weaker convective precipitation Northern Florida has a weakly negative correlation with ENSO during the summer. Spatial variability within Florida with re s pect to ENSO-related precipitation pattern s typically exhibits a north-to-south trend of increasingl y significant and larger correlations between ENSO and seasonal rainfall. The re s ult s of the randomized correlation analyses of ENSO and seasonal precipitation in Florida agree with the results of previous studies of ENSO teleconnections to precipitation in the southeastern United States and Florida (Ropelewski and Halpert 1986, 1989; Kiladis and Diaz 1989 ; Schmidt et al. 2001 ). ENSO teleconnections with position of the rnidlatitude jet influence how far frontal sys tems penetrate into Florida from the north and therefore the spatial variability of rainfall during seasons except summer. Relationships between ENSO and rainfall during fall winter, and spring seasons are enhanced when the NAO is negative. This modulation is strongest during winter and also typically exhibits a north-to-south trend of increasingly significant and larger correlations between ENSO and seasonal rainfall. A low or negative NAO state durin g winter is 52

PAGE 66

associated with a more meridional circulation regime and a southward displacement of storm tracks across the eastern United States. As a result, there are more frequent incursions of polar cold air masses (Hartley 1999). This might act to reinforce (enhance) fronta l penetration into Florida during El Nifio winters and to destabilize the opposite pattern during La Nifia winters. The comparative analyses of rainfall (Table 7) do not have an adequate number of cases to satisfactorily address this question, but suggest that precipitation levels in North and Central Florida are higher for both El Nifio and La Nifia winters when the NAO is negative and lower when the NAO is positive. Based on this idea, an interesting case can be made for NAO modulation of ENSO rainfall, using winter rainfall in Central Florida (where the patterns are strongest) as an example. Winter ENSO and NAO correlations with rainfall in this area are strongly positive and strongly negative, respectively, although the ENSO correlation values are higher. El Nifio events (high values of the ENSO SSTA indices) are associated with abundant rainfall and La Nifia events, with low rainfall levels (Ropelewski and Halpert 1986 1989; Kiladis and Diaz 1989; Sittel 1994; Schmidt et al. 2001 ) The inverse relationship ofNAO with rainfall in Central Florida during winter may modulate these ENSO patterns such that when the NAO is negative the El Nifio rainfall pattern (elevated levels) is enhanced. When the NAO is positive the La Nifia rainfall pattern should be enhanced (it should be even drier) The comparative analyses presented in Table 7 support this interpretation as does Figure 11, which graphically presents the mean winter rainfall data for Central Florida. However, the small number of cases involved make a definitive statement impossible. StreamFlow The relationships between climate variability such as ENSO and the NAO, rainfall and lagged stream flow reflect a complex suite of interactions operating at various spatial and temporal scales. Stream flow integrates rainfall and smooths out much of the temporal and spatial patchiness of rainfall data. It is not surprising, then, that stream flow exhibits a stronger and more significant relationship with climate variability. The lagged relationships between climatic variability and stream flow levels in the focus area are most likely caused by a mixture of natural and anthropogenic factors that are specific to each watershed and which may vary along its length. For example, the Green Swamp at the headwaters of the Hillsborough River serves as a sponge and filter that accumulates rainfall and slows its downstream passage (Estevez et al. 1991 ) Along its downstream reaches, the Hillsborough River is located in a highly urbanized area Rainfall runs off much more quickly in such areas due to the preponderance of concrete surfaces; this speeds up the river's response to rainfall (Zarbock et al. 1995). Despite these natural and anthropogenic influences, stream flow in the study area does demonstrate impacts related to climatic variability. These impacts are generally uniform throughout the focus area and 53

PAGE 67

-E 0 c 0 :;:: ctl :t::: a. 0 Q) a.. a. a.. Q) c 45 40 35 30 25 20 15 10 5 0 2 .0 ,..., 0 -1.5 1.0 <> p 0 0 I 0 I 0 0 0 0 0 o 0 -0 5 0 0 0.5 1 0 1.5 2.0 ENSO SSTA Figure 11. Mean winter precipi t ation in Central Florida, binned by its correspond ing NAO value, versus ENSO SSTA. Values associated with NAOconditions are show n with open symbols; those values associated with NAO+ conditions, w ith filled symbo ls. Diamonds(+) represent NAO va lues> !LSI; circles ( e), 11.49-1.01; and squares (), j0.99-0.5j. Mean precipitation for stations with the most comp l ete (>95%) winter records from 1950-1999. 54

PAGE 68

with re spec t to time lags, and for some seasons they are highly significant confirming that the expected integrating effect for stream flow is in fact occurring. Similar to ENSO-related variability in rainfall stream flow during fall, winter, and spring demonstrates significant relationships w ith ENSO (Tab les 8, 9, and 1 0). The st r ongest relationships occur w ith ENSO co nditi ons in the preceding one to two seasons Winter stream flow has the stro n gest and most significant relationship. Correlations between seasona l stream flow and the NAO are typically negative and are not sig nifi cant for any season except fall, which has a significant, negative correlation with the NAO conditions during the pr eceding spr ing. The NAO s tate influ ences co rr elatio n s between stream flow and ENSO in a complicate d manner. Summer ENSO conditions are more highly corre l ated with stream flow in the fall and winter when the NAO is positive; however, fall winter and spring ENSO co nditi ons are m ore highly corre l ated w i th stream flow in the fall, winter, and spr in g when th e NAO is negative. Additional Considerations The goa l of this research is to document relationships between climate variability and meteorological and h ydrological phenomena th at are pertinent to wate r-u se issues in Florida. Our research cannot specifically identify teleconnection patterns that create the relationships that we have documented We have suggested some mechanisms that can be further investigated, but wou ld be remiss however, if we did not note some of the limit at ion s inherent in our research. For examp le, the lagged relationships between stream flow and climate indices reflect not only the time delay between precipitation and resulting st ream flow variability but al s o the evolution and persistence of clima t e states (such as an El Nifio event) through a series of seasons. Also the recent trend towards the positive phase of the NAO may be part of the natural variability of the climate system or it may be related to ex t ernal forcing such as anthropogenic greenhouse gas and aerosol inputs t o the atmosphere (Osborn et al. 1999). With this in mind the patterns and spatia l variability in precipitation and stream flow discussed in this paper may in corporate ex t erna l influ ences that are not accounted for solely by ENSO and the NAO. Our r esearc h demonstrates that NAO modulates ENSO-related variability in seasonal rainfall and stream flow in F lo r ida This raises an interesting question-By what mechanism(s) does th e NAO modulate ENSO-re lated variabi l ity ? The NAO may influence F lo rida's response to ENSO forcing by modulating the underlying or background c limatic cond iti ons in F lor i da However, this is not the only possibility. The NAO may directly modulate ENSO or it may affect the way ENSO variability is communica t ed to F lorida by modulating the state of the atmosphere through which the communication occurs. 55

PAGE 69

Implications for Water Resources Management Similar to many urban coa s tal areas, the Tampa Bay area is dealing with the environmental impacts associated with rapid population growth. Water resource planner s in the Tampa Bay area have long known that natural variability in rainfall and the resulting variability in stream flow in the Tampa Bay ar e a lead to large interannual variability in freshwater that is available for harve s ting The re s earch pre sented in this paper provides a mechanism for understandin g variability and po t entiall y for plannin g in advance for extreme condition s, w hich will be applied in a quantitative manner as part of the future research 56

PAGE 70

ENSO IMPACTS ON SALINITY IN TAMPA BAY FLORIDA INTRODUCTION Tampa Bay is the largest estuary in Florida and is one of th e most biologically diver s e su btropical estuarine areas in the United States. It su pports a wide var i ety of marine organisms including over 250 macroalgae species, 250 fish species, and 1,200 species of macro-invertebrates (e.g., sca llop s, sponges, crabs and shrimp; Harwell eta!. 1995). This diversity, combined with its role as nursery habitat for many species, makes Tampa Bay a v it a l habitat for G ulf of Mexico fish and shellfish populations. Recreational and commercial u se of the Tampa Bay estuary system contributes greatly to the local economy and its ecological health is an issue of great concern (DelCharco 1998). Because many regulatory and management decisions that impa ct estuaries and their watersheds (such as Tampa Bay) are based on circulation patterns and flow it is essential to understand and expand knowledge of estuarine d ynamics. Estuarine circu l ation i s controlled by a suite of factors that vary over a range of time sca les from tidal (diurnal to semi-diurnal) to annual and longer. The primary controls include density differences astronomical tides, and winds. Large-scale weather patterns as well as synoptic wind events ma y increase or decrease flows out of or into estuaries. For example, on time scales of days, persistent w inds associated with winter frontal pa ssages have a strong impact on residence time in Tampa Bay ( Burwell 2001 ). Density driven circulation w ith in estuaries is controlled by horizontal salinity g radient s and is influenced by the net freshwater supply into estuaries. Freshwater input is the sum of runoff, stream flow, groundwater di sc harge, and direct precipitation minus evaporation from the estuary. Salinity patterns in an estuary result from a dynamic steady state in which the advec tive flux of sa lt into or out of the estuary, which is driven by the net freshwater supp l y, is balanced by the dispersive flux from the exchange of water by tides and other hydrodynamic mixing processes (Pritchard 1956) Freshwater inputs into estuaries vary on many time scales. For example seasonal precipitation patterns such as wet summer and dry winter seasons may have a pronounced influence on estuarine salinity distributions. At interannual time scales, climate varia bility suc h as E l N ino-S outhern O sc illation (ENSO) h as well-documented influence s on precipitation and stream discharge wh i ch dominate freshwater inputs into estuaries. This chapter describes the impa ct ofENSO on salinity distribution in Tampa Bay, Florida. 57

PAGE 71

Implications of Variability in Salinity Distribution Variability in salinity le v els within an estuary ha s implications for both its physical and biological components. The longitudinal salinity di s tribution controls the resi dual or density-driven circu lation in an estuary ; therefore, anything that affects the sa linity gradient from the head to the mouth of an estuary impacts estuarine circulation and flushing. Because the distribution of water quality parameters such as chlorophyll a and dissolved oxygen are often controlled by the same processes that determine sali nity distribution their levels are related to salinity in many estuaries (Rutherford et al. 1995; Bendis 1999). Many organisms in estuaries have optimal salinity range s, and changes in sa linity di st ributions due to both natural and human causes ha ve the poten tia l to impact biologica l resources (Zarbock et al. 1995 ; Boler 1998). Coastal and estuarine area s in the United States are experiencing unprecedented popu l ation growth. Hand-in-hand with burgeoning coastal populations comes the nece ss ity of managing and maintaining coastal waters that are increasingl y stressed by human impact s For example increa se d wastewater originating from treatment plant s and septic tanks and higher volumes of urban non point runoff both result from population growth in coastal communities (NOAA 1998). Furthermore, urbanization will continue to alter coastal watersheds and freshwater flows to estuaries, such as Tampa Bay as rural lands are converted to housing developments and stream flows are diverted to meet the freshwater needs of the growing population. Within this context, it is important to under stan d the role of natural climate variability, through its impacts on precipitation an d stream discharge, on salinity distributions in order to assess effectively and accurately the impacts of human alterations. ENSO and Florida El Nino-Southern Oscillation (ENSO) refers to a g lobal climate fluctuation that originates in the equatorial Pacific Ocean through lar ge-sca le interaction between the ocean and atmosphere. During El Nino (warm) events, the waters of the eastern equatorial Pacific are anoma lou s l y warm and sea level pressure l owers in the eastern Pacific Ocean and rises to the west. This is accompanied by a weakening of the l owl atitude easterly trade wind s and increased heating of the tropical atmosphere over the central and eastern Pacific Ocean. The associated impact s on atmospheric conditions include strengthening of jet s treams and steering of extratropical storms and frontal systems along paths th at are significantly di fferen t from normal. During La Nina (co ld) events, the waters of the equatoria l Pacific are anomalously cool with s tren g thened easterly trade winds and hi g her sea l evel pressure in the eastern Pacific Ocean ( Ra smusson and Carpen ter 1982; Trenberth 1991; Dawson and O'Hare 2000). Te l econnections for both El Nino and La Nina events include temperature and precipitation anoma l ies in man y regions of the world (Rasmusson and Wallac e 1983; Ropelewsk i and Halpert 1986). In Fl orida ENSO 58

PAGE 72

teleconnections include elevated rainfall during El Nifio fall and winter seasons and low rainfall levels during La Nifia winter and s pring seasons. Impacts on stream dischar ge demonstrate similar patterns, although s treamflow response may be delayed by seve ral months depending on drainage ba s in characteristics (Schmidt et al. 2001 ). MA TERJALS AND METHODS Site Characteristics The largest estuary in Florida, Tampa Bay covers about 1,031 km2 stretches about 53 km in length (Fig. 12), and has a watershed of 6,483 km2 (Southwest Florida Water Management District 1998). Mo s t of the bay is shallow, with an average depth of only 3. 7 m but the navigational channels reach depths of up to 13 to 14 m (Zervas 1993). The tides in Tampa Bay are small in amplitude-the diurnal range is 70 em (Goodwin and Michaelis 1976)-and are mixed sem i-diurnal/diurnal. Tidal constituents account for only 52% of sea-level fluctuation s in Tampa Bay (Zhang 1994), with coastal set -up and sy noptic s cale set-up contributin g s ignificantly to error in sea-level prediction. The s trongest tidal currents occur in the deepest, dredged parts of the bay and are on the order of 1 rnls. Residual circulation in Tampa Bay varies from about 0.05 to 0.1 rnls in the navigational channel in the middle of the bay. The principal rainy season is from June to September, the result of local afternoon thundersto rms and the occasional tropical storm. Due to synoptic-scale winter storn1s February and March may also show secondary precipitation peaks (Winsberg 1990). Maximum rainfall level s occur in the sun1mer (averaging about 50 em) and minimum levels in the fall (averaging about 15 e m) Interannual variations in precipitation are common in the Tampa Bay area and may be related to climate variability such as ENSO and the North Atlantic Oscillation (Schmidt and Luther 2001). Annual rainfall averages 140 em. During El Nifio years, there may be an additional 40 50 em of pr ecipitation while during La Nifia years there may be as much as 90 em less pr ec ipitation (Schmidt et al. 2001). Freshwater input to the bay comprises direct precipitation (43%; Zarbock et al. 1 995) and surface water sources (discharge from several rivers, mo stly a long the east si d e of th e bay, and direct runoff ; 41 %; Zarbock et al. 1995) with smaller contribut ion s from domestic point sources, groun dwater s pring s and indu s trial point sources Peak stream flow occurs in August-September, with a seco ndary peak in February-March This seasona l stre am discharge patt ern corresponds to the seasonal rainfall di strib uti on with lags up to a month or two. Due to the small drainage area, mean annual freshwater flow is s m a ll averaging only 95 m3/ s for the period 1985-91 (Zarbock et al. 1 995) w ith ap proximately 35-39 m3/ s from stream flow (Flannery 1989 ; ba sed on gauged riv e rs ; Zarbock et al. 1995). There i s considerable int erann ual var iabili ty in st ream flow: 59

PAGE 73

0 0 0 co C\1 0 1.() 0 """' C\1 0 v 0 """' C\1 0 ('t) 0 """' C\1 Figure 12. Map of the Tampa Bay area, with estuarine salinity stations indicated by circles and tributary salinity stations indicated by triangles. The geographic subdivisions in Tampa Bay are Old Tampa Bay (OTB), Hillsborough Bay (HB), middle Tampa Bay (MTB), and lower Tampa Bay (LTB). 60

PAGE 74

During the period 1974-94, Bendis (1999) finds that annual combined flows for Tampa Bay's four largest drainages (Hillsborough, Alafia, Little Manatee, and Manatee River s) are up to five times higher during wet years suc h as 1979 and 1983 than during dry years s uch as 1990 These four rivers drain approximately 75% ofthe bay's watershed and account for up to 82% of the total stream flow into the bay Mos t of Tampa Bay's freshwater inflow is into Hillsborough Bay, which is located in the northern more industrial and urban part ofthe bay. Flannery (1989) catalogues 44 minor tributaries, most of which are ungauged and less than 28 km long. Many of the small tidal creeks have been substantially modified by channelization, bank hardening, urban runoff industrial discharge s, and flow alteration during the past 50 years. Similarly, the Hillsborough River has been dammed 18 km from its mouth to create a reservoir for the City of Tampa's drinking water supply. Salinities in Tampa Bay vary from highs of 35 or more at the mouth to lows of 22 or less in the upper portions of the ba y, and this gradient occurs during both relatively dry and wet years (Squires et al. 1995) Squires et al. ( 1995) and Galperin et al. ( 1991) both conclude that horizontal density gradients are significant in driving the circulation of the bay. Because of its shallow depth relatively small freshwater inflows small tidal range and winds Tampa Bay is typically vertically well mixed (DelCharco 1998) Salinity throughout the bay typically is inversely related to stream flow, with high values baywide during the winter and late spring and low values in the summer (Bendis 1999) In terms of freshwater inputs and estuarine circulation, freshwater diversions and concentrate discharges from a desalination plant are two pertinent alterations to Tampa Bay's fresh water inputs and circulation. Both of these have as yet unknown interactions with climate variability. Based on both analyses of water quality data (Lewis and Whitman 1985 ; Rutherford et al. 1995; Bendis 1999) and of residence time (Burwell 2001 ) Tampa Ba y can be divided into four bay segments that are delineated by the combined effects of fresh water input morphology and tidal forcing (Fig. 12) Lower Tampa Bay flushes very quickly and is dominated by tidal influences especially along the deep navigational channel. The Sunshine Skyway Bridge causeway and irregularities in the bay's morphology cause the circulation to be slower on the sides of the bay in this section. Middle Tampa Bay typically has the largest salinity gradients and is not dominated by any single mechanism; both freshwater input and wind stress dominate in this section (Burwell 2001 ). Most of the bay's freshwater input enters into Hillsborough Bay and this section flushes relatively quickly. Burwell (200 1) finds Old Tampa Bay, which is shallow and ha s restricted flows due to constrictions and causeways to have very long residence time s and circulation that is dominated by tidal forcing. Both Burwell (200 1) and Bend is' (1999) find that the northernmost parts of Old Tampa Bay are influenced by freshwater flow from several small tidal creeks in this area. 61

PAGE 75

Data ENSO Monthly Nino-3.4 sea surface temperature anomaly (SSTA) values for the period 1974-1999 were obtained from the Climate Prediction Center's (CPC) and u sed to evaluate ENSO conditions. Po sit ive SST As are associated with El Nino (warm) events and negative SST As are associated with La Nina or cold events. El Nino (La Nina) months were defined as those whose SSTA exceeds 0.4 oc (is less than 0.4 C). Month s were considered neutral when SSTA fell between+ / 0.4 C. This methodolo gy is in agreement with prev io u s research on the impact s of climate varia bility in Florida (Schmidt et al. 2001 ; Lipp et al., in press). This approach to classifying ENSO events was chosen for several reason s: There is no s ingle generally accepted classification scheme, this scheme captures the most widely recogni ze d and accepted ENSO events, and application of this cla ss ification scheme to the SST A data is straight-forward. Salinity Monthly salinity data (1974-1999) were obtained for Tampa Bay and its tidal tributarie s from th e Hill s borough County Environmental Protection Commission (HEPC), which maintains th e most exten sive database of its kind for Tampa Bay. Over the peri o d of record sa linity was mea s ured once a month at mid-depth for 11 tributary and 52 estuarine sta tion s (Fig. 12). Tributary stations average 78% data coverage and estuarine s tations average 99% data coverage for mid-depth sa linity over the period of reco rd Seventeen estuarine stations are located in Old Tampa Bay, 12 in Hill s borough Bay 12 in middle Tampa Bay, and 11 in lower Tampa Bay Four tidal tributary stations are locat e d in Old Tampa Bay, 5 in Hillsborough Bay and 2 in middle Tampa Bay. The HEPC sam ple s on a monthl y ba sis over a three-week period durin g which roughly one-third of th e s t a tion s are sam pled on one day in each of three consecutive weeks (Boler et al. 1991 ). This s ampling regime is not synchronized to tidal cycle. In addition depths in Tampa Bay and it s tributaries vary from over 10 meters in the shipping channel to le ss than 2 meters over se agr ass flat s. Therefore, mid-depth is not the same water depth at different stations. Interpr e tation of th ese data mu s t take into account that they were collected under variable conditions However this research demonstrates that, de s pite the asynoptic monthly samp lin g strategy, salinity d a ta from the Tampa Bay area are suitable for answering the que s tions po se d in thi s research Analyses The approach used in this s tudy was to analyze the s tatistical correlation between sa l inity and ENSO SST As for 63 tidal tributary and estuarine s t a tion s in Tampa Bay, Florida from 197 4 through 1999 For each sta tion midd l e sa linity was correlated with ENSO SST As in several different wa ys: over the entire p e riod of record; for all months 62

PAGE 76

corresponding to El Nifio conditions; for all months corresponding to La Nifia conditions; and seasonally for months corresponding to El Nifio and La Nifia ENSO SSTA values. Approximate randomized correlations were used in all analyses to test the null hypothesis that there is no relationship between ENSO SST As and salinity in Tampa Bay. This computer-intensive test generates the probability distribution of the test statistic by recomputing it for many (10,000) artificially constructed data sets and is used to assess significance under minimal assumptions. The observations that are tested do not need to meet the normal distribution criteria of conventional parametric statistics and do not need to be a random sample Additionally, approximate randomized tests maximize the ability to discriminate between hypotheses because the sampling distribution is known (Noreen 1989). RESULTS Results for the correlation analyses are discussed in sections corresponding to ENSO state. Correlation values (r-values) for the approximate randomized correlations are considered significant if the probability of the correlation arising by chance is less than or equal to 5%. Salinity and ENSO SST As The relationship between monthly ENSO SST As and salinity in the Tampa Bay area was examined in several ways, the results of which are summarized in Table 11. No significant correlations exist between salinity and ENSO SST As over the entire period of record (all months) or when only seasonal values are considered. Table 11. Mean r-value for correlations between monthly salinity data and ENSO SSTA for the period 1974-1999. Correlations that are signification at the 95% level are indicated by an asterix (*); those that are significant at the 99% l evel, by a double asterix (**). Middle salinity Tidal Tributary All months Winter (JFM) Spring (AMJ) Summer (JAS) Fall (OND) Salinity, El Nifw, and La Niiia -0.095 -0.154 -0.194 0 020 -0.057 Estuarine -0.061 -0.073 -0.249 0.013 0.014 The relationship between mid-depth salinity and El Nifio and La Nifia conditions was examined over the entire period of record (all months corresponding to either El Nifio or 63

PAGE 77

La Niiia conditions) as well as s easonally. Significant correlation s are document e d for estuarine stations during E l Niiio conditions o ve r the entire period of re co rd and for all seasons except s ummer (Table 12). Significant correlations for La N iiia conditions are found only during s pring months. Overall mid-depth salinity varies negat i vely with El Niiio and La Niiia SST A v alue s. Table 12. Mean r-value for correlations between mid-depth salinity and El Nino or La Nina months for the period 1974-1999. Correlations that are signification at the 95% level are indicated by an asterix (*); those that are significant at the 99% level, by a double asterix (**). El Niiio La Niiia Tributary Estuarine Tributary Estuarine All months -0 .156t -0.311** -0 .082 0.032 Win ter ( JFM) 0.105 -0.401 0. 1 23 0.223 S pring (AMJ) 0.000 -0.407** -0.444* 0.384 S ummer (JAS) -0 057 -0 .23 5 -0.109 0.059 Fall (OND) 0 .324t 0 .341 0.020 0.110 t Correl atio n value s that are sig nificant at the 90% lev el. Spatiotemporal Variability of Salinity and EL Niiio/La Niiia Correlations Correlation re s ults for El Niiio and mid depth sa linity exhibit s p atia l variabi li ty wi th decrea s in g correlation value and s ignifican ce from head to mouth for a ll seasons except sp rin g (Table 1 3; Fig .13 a-d). Stations in Old Tampa Bay (OTB) have th e largest mean correlation value and s i gn ificance l evel, whereas stations in lo wer Tampa Bay (L TB) have the lowe s t mean co rrelation value and significance l evel. This s p a tial tre nd is re ve rsed for s prin g month s. For La Niiia conditions only the s prin g s eason mid depth sa linity has sig nificant corre lations with ENSO SST As. This inverse rela t ionship is s tronge st at the head of the ba y and weakens t owards the mouth (Tab le 14 ; F ig. 14a-d) DISCUSSION Salinity and ENSO The conne c tion between sa lin ity a nd EN SO i s a complicated chain of impact s f rom ENSO SST As to g lobal weather patterns to local precipitation effects to spa ti ally var i able dischar ge and runof f pattern s within the Tampa B ay draina ge area t o salinity distribution. Various factors influen ce thi s cha in of impact s, includin g n on linearitie s in ENSO t e leconn ec ti o ns. Recent research i n t o the differences in timin g and impacts of E l N iiio and La Niiia teleconnections h as s ho wn th a t they are n ot n ecessari l y e qu al an d opposite ( Hoerlin g e t al. 1 997). Thi s h as b ee n doc um e nt e d for ENSO imp ac t s on F lo rida's 64

PAGE 78

Correlation ranges -0.4500 649 0 -0.300--0.449 0 -0.150--0.299 0 0 000--0.149 o e Correlation ranges 0.450-0.649 0 [] 0.300-0.449 D 8 0.150-0.299 0 D 0.000 0 149 0 [] Significance levels not signif i cant o 0 0.95 0.99 Iii@) <0.99 Figure 13a-d. Maps of Tampa Bay showing the four bay sections and for each season, the sign strength, and significance of correlations between ENSO SST As and mid-depth salinity for El Nifio conditions. 6 5

PAGE 79

Correlation ranges -0.450--0.649 0 -0.300--0.449 0 -0.150--0.299 0 0 000--0.149 0 0 Correlation ranges 0.450-0.649 0 0 0.300-0.449 0 0 0.150-0.299 0 lEI 0.000-0.149 0 0 Significance levels not significant o 0 0.95 0.99 EJ@) <0.99 Figure 14a-d. Maps of Tampa Bay showing the four bay sections and, for each season, the sign, strength and significance of correlations between ENSO SST As and mid-depth salinity for La Nifia conditions. 66

PAGE 80

Table 13. Mean r-value for significant correlations between mid-depth salinity and El Nifio months from Table 2, broken out by each bay section for the period 1974-1999. Correlations that are signification at the 95% level are indicated by an asterix (*); those that are significant at the 99% level, by a double asterix (**). Winter Spring Summer Fall (JFM) (AMJ) (JAS) (OND) All estuarine stations -0.401 -0.407** -0 235 -0 .341 OTB -0.513** -0.353* -0.341 -0.418** HB -0 369* -0 .381 -0.071 -0 324* MTB 0 395* 0.444** -0 23 1 -0 348* LTB -0 272* -0.480** -0.255 -0 23 1 All tidal tributary stations -0 .105 0 000 -0 057 -0 .324t OTB 0 04 7 0 039 -0 .148 -0.429** HB -0.085 -0.040 0 023 -0.301 ** MTB -0 275* 0 022 -0 075 -0 .171 t Correlation values that are significant at the 90 % level. Table 14. Mean r-value for significant correlations between mid-depth salinity and La Nifia months from Table 2, broken out by each bay section for the period 1974-1999. Correlations that are signification at the 95% level are indicated by an asterix (*); those that are significant at the 99% level by a double asterix (**). Winter Spring Summer Fall (JFM) (AMJ) (JAS) (OND) All estuarine stations 0.223 -0.385* 0 059 0 110 OTB 0 266 -0.405** -0 034 0 .301 HB 0.218 -0.434** 0 .105 0.052 MTB 0 217 -0.392* 0 102 0 078 LTB 0 169 -0. 292 0 106 -0 085 All tidal tributary stations -0 .123 -0.444* -0.109 -0 020 OTB -0. 156 -0.4 35* -0 .102 -0.027 HB -0 062 -0.439* -0.085 0.018 MTB -0 212 -0.47** -0.184 0 .101 seasona l precipitation, with increased rainfall associated w ith E l Nino falls and winters and drier conditions associated with La Nina winters and springs (Schmidt e t a!. 2001 ). With respect to the results of the correlation analyses for salinity and ENSO SST As (Table 11 ), it is th erefore not surprising that th e correlation val ue s are low and insignificant. 67

PAGE 81

In contrast, mid-depth salinity in Tampa Ba y is s ignificantl y and inve r se l y correlated to El Niiio SST As (Table 12; Fig 13ah). This is a consequence of the increa sed precipitation and elevated discharge levels associated with El Niiio events in Florida. This relationship is strongest in the winter and spring reflecting the pro l onged respon se of discharge to the elevated precipitation levels in fall and winter. Summer salinity patterns are not influenced by ENSO state but instead are dominated b y the influence of highly localized convective storms and their runoff. Salinity during fall exhibits significant relationships with El Niiio conditions but not as strong as during the winter season. This is consistent with the weaker and less significant relationships found between El Niiio conditions and both precipitation and stream discharge (Schmidt et al. 2001). La Niiia impacts on the Tampa Bay Florida area include lower precipitation lev el s and depressed stream discharge rates during winter and spring (Schmidt et al. 2001 ). Mid-depth salinity in Tampa Bay during spring is elevated significantly in response to these local La Niiia conditions (Table 12). Other seasons do not exhibi t s ignifi cant relation s hips between salinity and La Niiia SST As. Salinity level s in the tidall y influenced tributaries of Tampa Bay exhibit no s i g nificant relationships to either El Niiio or La Niiia ENSO SST As with the exception of La N iii a sp ring (Tables 12 and 14). In the tidal portion s of the riv ers, hi gher salini ty wa ter is locat e d farther upstream during drier La Nina conditions accounting for the in v erse relationship. Spatiotemporal Variability in ENSO Impacts on Salinity Tampa Bay is divided into several, distinctive geographical sec tions based on analyse s of water quality data (Lewis and Whitman 1985 ; Bendis 1999 ) and residence time s (Burwell 2001 ), whose characteristics are determined by different proce sses and influences. The results of the analyses of El Niiio conditions and mid-d ep th sa linity in Tampa Bay corroborate these divisions. The spatial patterns are most evident in winter and spring when the relationships between E l Niiio conditions and sa lini ty are strongest (Table 13; Fig. 13a-h) Freshwater inputs from small creeks and runoffhave th e greatest influence on salinity distribution in Old Tampa Bay due to i ts restricted ti dal exchange and lon g residence times. As a result Old Tampa Ba y respond s quickl y to elevated precipitation and discharge l evels associated with El Niiio in th e fall and winter and has significant correlations between El Niiio SST As and mid-depth salinity in all seasons (F i g 13a ,b ) Hill sborough Bay receiv es most of Tampa Bay's stream input and has residence tin1es on the order of weeks (Burwell 200 1 ). Increased freshwater input into Hillsborough Ba y durin g El Niiio falls and winters is reflected in the overall ne gative and significant relationship between mid-depth salinity and ENSO SST As (Fig. 13a b ). 68

PAGE 82

Middle and lower Tampa Bay are dominated by tidal exchange and mixing due to wind events ; overall, salinity in these sections is not significantly correlated with El Nifio conditions (Fig 13a-d). However, in both these bay s ections, mid-depth salinity is negatively and significantly correlated with El Nifio ENSO SST As at all stations. This pattern may represent the flushing of low salinity waters from Old Tampa Bay, which ha s re s idence times on the order of several months (Burwell 2001 ). La Nifia conditions are correlated significantly and inversely with mid-depth sa linitie s at most stations in Tampa Bay only during spring (Table 14; Fig 14ad). During spring larger and more highly significant correlations are found at stations near the head of Tampa Bay (in Old Tampa Bay, Hillsborough Bay, and middle Tampa Bay). Stations in lower Tampa Bay have lower correlations, and only half of the stations have significant correlations Both rainfall and stream discharge are depre sse d during La Nifia winters and springs; this is also reflected in the inverse relationship between salinity and ENSO SST As. SUMMARY AND CONCLUSIONS The research has demonstrated s ignificant correlations between ENSO state and mid depth salinity in Tampa Bay, Florida Salinity in an estuary is determined by a suite of dynamically variable influences, and climate variability such as ENSO may impact salinity distributions through teleconnections with precipitation and stream discharge. Documented El Nifio (La Nifia) impacts in the Tampa Bay area include elevated (depressed) rainfall and stream discharge during fall and winter (winter and spring) with resulting depres se d (elevated) estuarine salinity levels In term s of assessing the impacts of human influences on estuarine areas, it is important to understand the natural variability of estuarine circulation and salinity distribution. For example in Tampa Bay imminent changes in freshwater withdrawal schedules and concentrated discharge generated by desalination have the potential to impact the estuarine system, with possible consequences to both natural and human estuarine a sse t s The environmental impacts associated with increasing groundwater use and a growing population have resulted in the adoption of a Master Water Plan which has been charged with developing new water supply sources (Tampa Bay Water 2000). As a result a desalination plant was built on Tampa Bay and surface water withdrawal schedules have been developed for the Hillsborough and Alafia Rivers (Fig. 12) to supplement existing water supplies and to decrease groundwater withdrawals. In order to detect and monitor potential impacts of surface withdrawals on the hydrology and ecology of the associated tidal river segments, a comprehensive Hydrobiological Monitoring Program (HBMP) was developed. Inherent in the HBMP is the recognition that surface water withdrawals are linked to potential changes in salinity patterns as well as associated water quality 69

PAGE 83

constituents and biological communities (Coastal EnvironrnentaVPBS&J, Inc 1998). The HBMP began in spring 2000 and will operate for three years before initiation of new surface water withdrawals and for three years afterwards. Howe ver documenting impacts on sal inity pattern s based data from a 3-year base period are possible only with respect to an accurate understanding of the role of natural variability, including interannual influences such as ENSO, on the bay's dynamics. The results presented in this paper have direct relevance to the ability of the HBMP to meet its goal of determining whether or not significant post-withdrawal change s in hydrology, water quality biota, habitat and vegetation constitute an unaccepta ble adverse impact. 70

PAGE 84

CLIMATE VARIABILITY AND ESTUARINE WATER RESOURCES INTRODUC TION Ab out 23% of es tuarine area i n th e continental U nited States i s d eg rad e d for anim a l a n d plant communities and for human u ses such a s drinking water ag ricultur e, sw immin g, and boatin g (U.S EPA 2001) Adverse imp ac t s ofte n are due to anthropogenic influ ences in coastal areas an d are exacerbated b y an eve r-growin g influ x of re s id ents. The trend s in population growth in the southeastern United State s are alarmingly lar ge For exam ple in the Gulf of Mexico the popul at ion of coastal coun ties ha s increa se d by 52% f r om 1 970 to 1990 (U.S. Bureau of the Census 1996). Florida has experienced a 600% increa se in popul a ti o n si nce 1940 and no w h as over 15 milli o n r es idents. Estimates indicate that b y the year 2010 Florida's population will reach 16 million making it th e third mo st populous s tate Growth in Florida's so uth ern coastal counti es has b ee n and w ill continue to be particularl y s i gnifi cant with 10 of the 15 fastest g r ow in g Gulf Coas t co unties in southwest Flor ida. Hand in-hand wi th bur geoning coastal populations comes the nec essity of mana ging and maintaining coasta l wate r s th a t a re incr eas in g l y s tre ssed b y human i mp ac ts For examp l e, poor water quality in c r eased levels of co ntaminant s modifications to habit ats, increased inc id ence of marine pathogens and changes in relativ e abundance of estua rine organism s are all w id e l y r epo rt e d eco lo g ical effec t s link e d to human ac ti vities in and around es tu aries (NOAA 1 998; U.S. EPA 2 001 ; Wenner and Geist 2001). Furthe rm o r e, u r b an ization wi ll continu e t o a lt e r coasta l watersheds and freshwater flow s to estuaries s uch as Tampa Bay, F lorida as rur a l land s are con ve rt ed to h o u si ng de velop m en t s and rive r flows ar e di v erted t o mee t t h e fres hw a t er needs of th e grow in g popul ation At ris k in Gulf of M e xic o es tuarie s are hundred s o f spec i es of birds recreational and commercial fish and s hellfi s h spec ies native cypres s and man groves, an d t hre atened a n d end angere d s pecie s such as sea turtl es and manatees. Gulf of Mexico estuar i es such as Tampa Bay pr ovi de critical feeding s pawnin g, and nursery h a bi tats fo r a diver se asse mbl age of fis h, wildlife, an d plant s p ecies s upp ort submerged aqua tic vegetation communitie s that stabilize s hor e lin es from erosio n reduce n onpoint so urce loadin gs, and improv e water clarity (U.S. EPA 2 001) Eva lu a t ing the impact s and risks associated w ith growth an d developme n t are an important part of prot ec tin g the natural resources in est u a rin e areas. Monitoring pro g ram s s u c h a s the Joint G ul f States Co mpr e h ensive Monitorin g Program (U.S. EPA 71

PAGE 85

2000), EPA's Coastal2000 Program and Nat i o n a l Es tuarine Research Reserves program (Wenner and Geist 200 I), as well as many sma ller programs focused on local i ssues, provi d e valuable inform a ti o n about the condition of estuarine re sources and form the ba s i s for evaluatin g ri s k s and im pacts. However, b ecause of the s hort-term nature of mos t m o nitoring pro g ram s, w hi c h typically are act i ve for a year to a few yea r s, var i ab ili ty at time scales lon ger than seas onal in the n atura l e n viro nment often i s not taken into account (Wenner and Geist 20 01) Failure t o thoroughly examine climate variability and incorporate it into plannin g s trategie s and re source u se may le ad to expens ive and sometimes disastrous events. An example from r ecen t North American hi s tory ill ustrates thi s point: t h e water allotment s chedule for the Co l orado River (Reisner 1 986). Rights to the Colorado River's flow are divided among seve n sta t es (Ar i zona New Mexico W yom in g, Utah, California Co lorad o and Nevada) and Me x ic o. The Co lorado River Compact, w hich divided up the river's water rights in 1922 was b ased on only eighteen yea r s of s treamflow mea s urement s from the 1900 s-19 1 Os--eighteen years that were characteri ze d by average or above -avera ge flow s in three o ut of every four years (Reisner 1986) Actual mean river flow i s a bout thirty percent less than th e 17.5 million-acre-foot estima t e that was u sed to di v i de up the r ive r's water. This lack of u nde r s tandin g ofth e natural sys t e m's clim ate variability ha s l ed t o over 75 yea r s of di s pute and liti gat ion b e t ween the seve n s t ates and b etwee n th e United States and Mexico as eac h h as sc rambled to protect it s river ri g h ts. In order to examine th e impact of climate variabi l ity on coastal water resources management thi s case study focuse s o n the Tampa Bay, Florida area and effo rt s to assess a nd monitor, through a short -t erm monitoring pro gram, a d verse environmenta l imp ac t s assoc i ate d w ith propo sed f r es h wa ter diver s i o n s f r om several r iv e r s th a t drain into t h e b ay. The impacts of one type of climate variability E l N ino-Sou t hern O sc illation (ENSO) o n t h i s region's rainfall stream discharge water qua lity, and sa linity distribution are known (Schmidt et al. 200 1 ; Schmidt a nd Luth er 2001a b; Lipp et al. 200 1 ). Spec ifi ca ll y, thi s case s tudy examines the following questions: 1 How may ENSO conditions affec t th e d ata collected by the baseline m onitor i ng program ? 2. How may ENSO conditio n s affect the availability of freshwate r for withdrawa l a t seaso nal and l onger tim e scales? 3. Can ENSO prediction s improve manage ment of water re sources in the Tampa Bay area? 4 What are the implicati ons of thi s case st ud y both for protecting th e natur a l es tuarin e eco syste m and for providing a r eliab l e source of freshwater? 72

PAGE 86

BACKGROUND Climate Variability and El Nino-Southern Oscillation Climate variability refers to major largesca l e modes of atmosphe ric circulation. These dynamical modes are the imprints of fundamental global processes, s uch as in stab i lities of the climatological mean flow or large-scale atmosphere-ocean interactions (Wallace 2000). Several examples of climate variability w ith impacts on the Northe rn Hem isphere include El Nino-Southern Oscillation (ENSO), the North Atlantic O sci llation (NAO), the Pacific Decadal Oscillation (PDO), and the Arctic Oscillation (AO). The research into the impacts of climate variability pre se nted in this case study focuses on ENSO. El Nino-Southern Oscillation refer s to a global weather pattern that originates in the equatorial Pacific Ocean through largescale interaction between the ocean and atmosphere. During El Nino (La Nina) events, the waters of the equatorial Pacific are anomalously warm (cool) and sea level pressure lowers (rises) in th e eastern Pa c ific Ocean and ri ses (lowers) in the west. These pressure changes which are accompanied b y s hifts in tropical rainfall affect wind patterns over much of the globe ( Ra smusson and Carpenter 1982; Trenberth 1991 ). Mid-latitude synoptic winter weather patterns shift equatorward (poleward) across North America during El Nino (La Nina) events, lead in g to sh ifts in temperature and precipitation p atterns (Rasmusson and Wallace 198 3; Ropelewski and Halpert 1 9 86). El Nino-Southern Oscillation events can be mea sured u si n g differen t parameters and the applicability of each measurement method varies with the impacts and location that are in question. The Southern Oscillation Index w hi ch tracks c h anges in th e a v era ge air pressure difference across the equatorial Pacific, traditionally has been used t o monitor ENSO t e leconn e ctions especially in Australia and the Indo-Pacific. Multivariate indi ces at various locations across the Pacific ba si n are also available for monitoring ENSO. T rackin g anomalous sea s urface temp era tures (SST) in th e e qu atorial Pacific i s another method of ENSO quan tifica tion; this type of inde x, whic h is calcu l ate d and pub li shed b y the C limate Prediction Ce nter (CPC) is u sed in the research presented in this case study. El Nino (La Nina) seasons are defined as those with five-month runnin g mean of the CPC's monthly Nino-3.4 SST anomalies, centered around the m iddl e of the seaso n that exceed 0.7 oc (fall belo w -0.7 C). Impacts of El Nino-Sou thern Oscillation in the Tampa Bay, Florida area Teleconnections for both E l Nino and La Nina events inclu de temperature and precipitation anomalies in many re gions of the world. For examp le, the Atlantic hurricane season is more active during La Nina events than during E l Nino events (Gra y 1984 ; Bove et al. 1998) In the so uthea s tern United States and in Florida in particul a r ENSO teleconnections include elevated rainfall durin g E l Nino fall and win ter seasons 73

PAGE 87

and low rainfall levels during La Nina winter and spring seasons (Ropelewski and Halpert 1986, 1989; Kiladis and Diaz 1989; Schmidt et al. 2001 ). Impacts on stream discharge in the Tampa Bay region demonstrate similar patterns, although streamflow response may persist for several months depending on drainage basin characteristics (Schmidt et al. 2001 ). The resulting variability in freshwater input to Tampa Bay influences its seasonal salinity distribution (Schmidt and Luther 200la). During El Nino events, ENSO sea surface temperature anomalies (SST As) are significantly and inversely correlated with salinity in the bay during winter and spring. These patterns reflect the elevated rainfall over the drainage basin and the resulting elevated s tream discharge and runoff, which depress salinity levels. Spatially, the correlations are strongest at the head of the bay especially in bay sections with long residence times. During La Nina conditions significant inverse correlations between ENSO SST As and salinity occur during spring Dry conditions and depressed stream discharge characterize La Nina winters and springs and the higher salinity levels during La Nina springs reflect the lower freshwater input levels (Schmidt and Luther 2001 a). Tampa Bay, Florida Tampa Bay is the second largest Gulf coast estuary and the largest estuary in Florida. The bay covers about 398 mi2 (1 ,031 km2), stretches about 33 mi (53 km) in length, and receives freshwater from a 2,542 mi2 (6,583 km2 ) wa tershed (Fig 15). Within the Tampa Bay watershed, 56% of the land is developed, 46 % of the built-up areas are urban and 16% include agricultural and pasture lands (Southwest F lorid a Water Management District 1998). The major rivers are located on the east and south sides of the bay and the four largest rivers (the Hillsborough Alafia Little Manatee, and Manatee Rivers) account for 82% of the freshwater input to the bay, with the remainder from smaller rivers, nonpoint sources and rainfall. Depths typically are 10-13 ft (3--4 m) in Tampa Bay, although the shipping channels which are dredged, are considerably deeper. Salinity levels at the bay's mouth typically are 31-36 ppt and range from the high teens to low twenties near the mouths of the major rivers (Burwell 2001 ). The principal rainy season is from June to September the re s ult of local afternoon thunderstorms and the occasional tropical storm Due to synoptic-scale winter storms February and March may also show secondary precipitation peaks (Winsberg 1990) Maximum rainfall levels occur in the summer (averaging about 20 in (50 em)) and minimum levels in the fall (averaging about 6 in. (15 em)). Interannual variations in precipitation are common in the Tampa Bay area and may be related to climate variability such as ENSO and the North Atlantic Oscillation (Schmidt and Luther 2001b). Annual rainfall averages 52 in. (130 em). Durin g El Nino ye ars there may be an additional 74

PAGE 88

Gulf of Mexico N 10 0 10 20 Miles ----Figure 15. Map of the Tampa Bay area. Surface water withdrawal loca t ion s and geographic designations used in this paper are also shown. 75

PAGE 89

16-20 in ( 40-50 em) of precipitation while during La Niiia ye ars there may be as much as 35 in. ( 90 em) le ss precipitation (Schmidt e t al. 2001 ). Ecologi cal and Economic Charac t eriza t ion An ecologically se n s itiv e and important area, Tampa Bay contains extensive seag ra ss beds, sa lt mar s he s, and man g rove s tand s th at se rve as fis h a nd s hellfi s h nur se rie s, breeding and nesting areas for colonial waterbirds, and protection for its over 900 mi (1,450-k.m) shoreline (Tampa Bay Nationa l Estuary Program 1993 ) One of the mo s t b io logically div e r se s ubtropical estuarine areas in the U nited States, Tampa Bay supports a w id e variety of marine organi s m s, includin g ove r 250 macroalgae s pecie s, 25 0 fis h spe cie s, and 1 ,200 species of macro-invertebrates (e.g. scallo p s, sponges, crabs and s hrimp ; Harwell e t al. 1995) Tampa Bay is also the largest port in Florida and i s the thirteenth larg es t p ort i n th e United State s in terms of total tonna ge handled and the ninth largest port with respect to do m estic trade (Ti ffan y and Wilkinson 1989; Ameri can A ssocia tion of Port Authorities 1 998) In addition, recreational uses an d commercial fishing add at least $7 million annually to th e economy. In s umma ry, the n atura l and recreational features of Tampa Bay coexist with comme rcial s hippin g and fishing indu s tri es c ombin i n g to make Tampa Bay a hi g h-ri s k area in terms of climatic variability and it s imp act on es tu a r i n e c i rcula ti on. Effo rt s over the past t we nty or so yea r s (spearheaded b y the Tampa B ay Nation a l Estuary Program and the Florida Legislature s Surface Wa ter Imp rovement and Management (SWIM) Program) to im p r ove wa t e r quality a nd the environmental h ea l th of Tampa Bay appear to have paid off-water quality h as impr ove d and seagrass beds are expa ndin g, as are upland sa lt marsh communities. H oweve r as hi ghlighted by th e recent discovery of non-native g r een mu ssels a nd decline s in seag rass bed cove rage in certain areas, Ta mpa Bay i s s till very sens i t i ve to p otential st r esses on its marine environment. In term s of fre s hwater input s and estuarine c irculation freshwater div ers ions and brin e re leases from desalin ation are two pertin e nt alterat i o n s to Tampa Ba y th a t may ac t as s tr esso rs. Both of th ese will hav e as ye t unkno w n but pot e ntiall y deleterious (a nd / or ben i g n) interacti o n s with climate var ia bil ity. Proposed F r eshwa t er Diversions S i mi lar t o man y urban coastal areas, th e Tampa Ba y area i s dealin g with th e e n v ironment a l impacts associated with ra pid population g rowth. In order to both ameliorate the overuse of it s gro und water resources and t o m ainta in an adequate fres h wa ter s uppl y, the re g ion ha s de ve lop ed a multif ace t ed Master Water Plan (Tampa Bay Water 200 0) T he Master Water Plan includ es freshwater div ers io n s from loc a l river s ( th e Palm River / Tampa Bypass Canal the Hill sboro u g h River, and the Alafia River) for both immediate consumption and l o n ge r term reservoir sto r age. In add ition new wate r 76

PAGE 90

sources from desalination are being developed These sources will supp l y at le ast 85 million gallons of water per day (mgd) by 2007. Both withdrawa l s and desalination ha v e the potential to affect water quality within Tampa Bay, which is related to flushing associated with natural cycles of flooding and drought. Therefore the Master Wat e r P lan incorporates a range of maximum withdrawa l or diver s ion rates based on ranges of di s charges in order to minimize impact s on the overall riverine and estuarine system s (Tab le 15). During low flow condition s no water i s withdrawn and reservoir storage groundwater pumping, and desalination make up the supply shortfall Table 15. Withdrawal schedules for the Hillsborough River, Palm Riverffampa Bypass Canal, and Alalia River. (Note 1 m3s-1 = 22,839,519.43 ga llons /day) River Discharge River Discharge Maximum Withdrawal Rate (mgd) (m3s -1 ) Pa lm Riverffampa < 7 mgd < 0.31 m3s1 0 mgd / m3s 1 (no withdrawal) Bypass Canal 7 -81 mgd 0.31-3.55 m3s '1 80% of tota l flow ; flow not to be l ess than 7 m g d / 0.31 m3s'1 > 81 mgd > 3.55 m3s '1 6 5 m g d/2 .85 m3s '1 (peak withdraw a l r ate ) Hillsborough River River Discharge River Discharge Maximum Withdrawal Rate < 65 mgd <2.8 5 m3s '1 0 m g d / m3s 1 (no withdrawa l ) 65-9 7 m g d 2 .85-4.25 m3s '1 10 % oftotal flow 9 7-139 m g d 4.25-6.09 m3s 1 I 0 % of to t a l flo w inc r easin g pr o p o rti o n ally t o 3 0 % 139--{)47 mgd 6.09-2 8 3 3 m3s '1 3 0 % of t o tal flow > 6 4 7 mgd > 28. 33 m \ 1 194 m g d / 8.4 9 m3s1 (pe a k w ithdrawal rate) A l afia River River Discharge River Discharge Maximum Withdrawal Rate <80 m g d < 3 5 m3s 1 0 mgd/ m3s 1 (no w ithdrawal) 80--5 1 7 mgd 3.5-22.6 1 m3s '1 I 0 % of t otal flo w > 5 1 7 m g d > 22.61 m3s '1 52 m gd/2.2 8 m3s '1 (pe a k w ithdr awa l r ate ) B ase d on Tampa Bay Water 2000. Water resourc e planners in the Tampa Bay area have l ong known that natural variability in rainfall and the re s ulting variability in s tream flow in the Tampa Bay a r ea lead to large interannual variability in fre s h w at e r that is available for harvestin g How ev er, asse ss ment s of impact s of the propo s ed withdrawal s chedules are based o n climato logie s of monthl y a v erage di s char ge derived from dail y m e an stream flo w (Coastal Environrnenta l /PBS&J Inc. 1 998). Thi s approach doe s not incorporate interannual variability Part of the research pr es ented in thi s ca se s tud y provid es an ex a mple of the impacts of recent ENSO e vent s on water a v a ilability and discu ss e s the 77

PAGE 91

potential for planning in advance for extreme conditions associated w ith climate variability using climate predictions. Hydrobiological Monitoring Plan Permits for the freshwater wit hdra wa l s from the Hillsborough River, Palm River/Tampa Bypass Canal, and Alafia River r equired the development of a comprehensive Hydrobiological Monitoring Program (HBMP) to detect and assess the potential impact s of the withdrawals on the hydrology and ecology of the associated tidal river segments. Inherent in the permitting rules is the recognition that surface water withdrawals are linked to potential changes in salin ity patterns, associated water quality constituents, and biological communities. The goal of the HBMP is to ensure that post-implementation flows do not deviate from the normal rate and range of fluctuation to the extent that wate r quality, vegetation, animal populations salinity patterns or recreational / aesthetic qualities are impacted adversely (PBS&J, Inc. 2000). In order to document existing pre-operation and appropriate baseline conditions and to detect sign ificant post operational changes the HBMP began monitoring h y drolog y, water quality biota, and habitat/vegetation in the downstream portions of the Hillsborough River Alafia River, and the Palm River/Tampa Bypass Canal in the sp rin g of 2000. This monitoring effort is s uppl emented by on-going water quality monitoring in Tampa Bay by th e Environmental Protection Comm i ss ion o f Hill sbo rough County, which ha s been co nduc ting routine sampling si nce 1972 (Boler 1998). The HBMP will operate for three years before initiation of new surface water withdrawa l s and for three years afterwards. Salinity and water quality indicators respond more or less in stantaneous l y to changing freshwater flows. However, biological indi cators such as species distributions of fauna and vegetation have indirect relationships w ith changes in freshwater flows and are mediated by physical and chemical changes. T h ese biolo g ical indicators respond on a s lower time scale (e.g. days months, seasons, ye ars). The criteria that will be used to determine unacceptable adve r se impacts caused by th e water withdrawals are biological in nature: "A detected change supported by statistical inference or a preponderance of evidence, from the pre-operational abundance, distribution, spec ies composition or species richness of biological communities of concern in the Lower Hillsborough River Lower Palm River / Tampa Bypass Canal, McKay Bay, or Lower Alafia River reporting units t hat can b e attributed to reductions in freshwater inflows caused by t he permitt e d surface wa t er withdrawals" (PBS&J Inc. 2000). However documenting adverse imp acts using data from a 3year b ase period i s po ss ibl e only if the data are eva luat ed within the context of lon ge r term climate variability, includin g in terannua l influences such as ENSO, and its influence on th e bay s 78

PAGE 92

dynamics. The results presented in this case study have direct relevance to the ab ility of t he HBMP to meet its goa l of determining whether or not sign ifi cant post-withdrawal changes in hydrology, water quality biota habitat and vegeta ti on constitute an unacceptable adverse impact. ENSO CONDITIONS AND THE HBMP's BASELINE MONITORING Seasonal precipitation an d stream discharge i n the Tampa Bay area exhibit int erannua l var i abi lity which can be related to ENSO conditions. The ENSO impacts are s t rongest in w int er (Schmidt et al. 2001 ). For example, mean rainfall during El Nifio winters was 2.6 time s greater than during La Nina winters for the period 1950-1999 (Fig. 16; based on 5 stations in the Tampa Bay watershed with greater than 95% data coverage from 1950 1999). During the most recent E l Nino (spring 1 997-winter 1998) and La Nina (summer 1999-sprin g 2000) events winter rainfall averaged over the same five stations in the Tampa Bay waters h ed was 21.9 and 3.2 inches (55 6 and 8.2 em), respectively. Although EN SO-related rainfall patterns during other seaso n s are not as dramatic (or s tati s tically significant) as they are during winter, the elevated or depressed levels during spring, summer, and fall contribute to th e cumulative impact on annual rainfa ll l evels. In the Tampa Bay area, mean annua l rainfall was 88.0 in. (223.6 em) for the 1997 1 998 E l Nifio and 41. 8 inches (1 06 .2 em) fo r th e 1999-2000 La Nifi.a eve nts. Discharge records for the three withdrawal s ite s vary in duration and qua li ty. Long term r ecor d s, which capture ENSO-related var iability in discharge, are available on l y for the Hillsborough River and th e Tampa Bypass Canal. Both of these records have been im pacted by channe l alterations. At the Hillsborough River withdrawal s ite Tampa Dam was constructed in 1 945. This st ructure stores water for use by the City of Tampa and during low flow periods negligible amounts of water are released over the dam The Tampa Bay Byp ass Cana l was built in 1979 altering the pre-exi s tin g Palm Ri ve r to the point where it no lon ger functions as a natural tid a l river system (F l annery 2001 ). In addition to diver s ion s from the Hillsborough River during high flow periods, di scharge at t he Tampa Bypa ss Canal withdrawal site may be augmented b y groundwater discharge during low flow conditions because the cana l breac h es th e aquif er. Therefore, although long term records ex i s t for two of the three withdrawal site s, these records are highly impact e d by human influences through time. However these records r e present the best data avai l ab l e not only for thi s case s tudy, but a l so for formulat i ng t he w i thdrawal sc hedules. Daily discharge records for the period 1 950 1 999 at the U.S. Geologica l Surv ey gage s tation on the Hillsborough River near Tampa (a propo sed diversion site) show seaso n a l variability in mean daily discharge re l ated to seaso nal ENSO conditions (Fig. 1 7). Overall, di scharge is depressed during La Nina seasons and elevated during E l Nino 79

PAGE 93

30 25 :C 20 0 :e. 15 0 0.... 10 5 0 W int e r (JFM) Spring (AMJ) Summer (JAS) so-y mean 50-y El N i fio mean 1 IJ 1 997-98 El N ino pre cipi t a tion levels IJ 50-y La Nina mean D 19 99-2000 L a N in a precipitation leve ls Fall ( OND) Figure 16. Mean seasonal precipitation in the Tampa Bay area. Mean seasonal precipitation is shown for the period 1950-1999, for El Nino periods during 19501999, for La Nina periods during 1950-1999, for the 1997-1998 El Nino, and for the 1999-2000 La Niiia. Error bars are +1one standard deviation. 80

PAGE 94

seasons compared to the 50-year mean, and these differences are significant during the winter (Sclunidt et al. 2001 ). Mean daily discharge during La Nifia winters and springs is close to or below the "no withdrawal" (2.85 m3/s) discharge limit (Table 15). Variability in stream discharge is large for all three cases and often equal to the seasonal mean daily discharge (Fig. 17). Examination of seasonal Tampa Bay precipitation discharge at the Hillsborough River withdrawal site, and ENSO SST A from winter 1997 through summer 1999 (Fig 18) provides a useful example of how baseline data collected during periods of extremes with respect to ENSO conditions might jeopardize the usefulness of HBMP' s baseline data for evaluating adverse impacts. Precipitation levels follow the typical summer seasonal pattern with elevated precipitation levels compared to other seasons. However, anomalously high levels of precipitation during fall 1997-winter 1998 and low levels during fall 1998 and winter and spring 1999 are both associated with extreme ENSO conditions. As a result discharge during fall 1997 winter 1998 was elevated in respon se to high rainfall levels associated with extreme El Nifio condition s; discharge during winter spring, and summer 1999 was depressed in response to low rainfall levels associated with extreme La Nifia conditions Basing a 3-year long baseline data collection effort like the HBMP during either (or both) of these periods potentially would skew the baseline "norm" towards drier or wetter than normal discharge conditions. This would also affect both physical and biological monitoring parameters of the HBMP as changes in freshwater supply impacted the salinity distribution The HBMP began in the spring of2000; La Nifia conditions have characterized the HBMP's monitoring efforts to date. With two more years of monitoring before the implementation ofthe withdrawal schedule, ENSO conditions may change dramatically. Any conclusions drawn by the HBMP concerning the baseline or "ty pical conditions in Tampa Bay need to take into account the ENSO state during the monitoring program. ENSO CONDITIONS AND THE TIMING OF FRESHWATER AVAILABILITY Based on the daily discharge records from the Hillsborough River withdrawal site from 1950 1999, no withdrawals would have been made (river discharge was lower than 2.85 m3/ s) on 8,472 days or over 47 % of the record (Table 16). Most of the days that met the criteria for no withdrawal were in the spring when 65% of the months from 1950 to 1999 had monthly mean daily flows below 2.85 m3/s. During La Nifia periods there were often greater than 6 continuous months of monthly mean daily flows that met the no withdrawal criteria. La Nifia conditions during the 1999 and 2000 water years (October 1998-September 2000) are associated with an extended period of low discharge for both the Hillsborough River and Tampa Bypass Canal proposed withdrawal sites (Table 17) Eight and nine of the 24 months had no days when the discharge exceeded the minimum 81

PAGE 95

45 40 35 30 1 0 5 0 -5 Summer (JAS) I 50-y o 50 -y El Nin o mean 50 -y La Nina mean Fall (O NO) Figure 17. Mean seasonal dail y discharge at the proposed Hill s borough River withdrawal site. Mean seasonal dail y discharge is shown for the period 19501999 for El Nifio periods during 1950-1999 and for La N iiia periods during 19501999 Error bars are + 1 one s t a ndard deviation. 82

PAGE 96

-60 0 T'""" -1< 0 50 ..!._.
PAGE 97

level for withdrawals for the Tampa Bypass Canal and Hillsborou gh Ri ver withdrawal sites, respective l y For the Hillsborough River wit h drawal site, the minimum level was met less than half of the da ys of the all the remaining months except July and October of 1999. This may be partially explained by diversions from the Hillsborough River to the Tampa Bypass Canal durin g th i s period of time. Table 16. Percent of days that wou ld have met each withdrawal l eve l for the Tampa Bypass Canal and Hillsborough River withdrawal sites. (Note 1 m3s-1 = 22,839,519.43 ga llon s/day) River Discharge River Discharge Percentage Maximum Withdrawal Rate (mgd) ( m3s-1 ) of days Tamp a Bypass Canal ( 1 956-1999) < 7 mgd <0.31 m3s1 13.8 0 mgd/ m3s1 ( no withdrawal) 7-81 mgd 0.31 3.55 m3s 1 50.7 80% of total flow; flow not to be less than 7 mgd/0.31 m3s-1 >81 mgd >3.55 m3s 1 35.5 65 mgd/2.85 m3s1 (peak withdrawal rate) Hillsborough River (1950-1999) < 65 mgd <2.85 m3s1 46.6 0 mgd / m3s1 (no withdrawal) 65-97 mgd 2.85-4.25 m3s-1 6.4 10% oftotal flow 97-139 mgd 4.256.09 m3s1 7.3 0% of total flow increasing pr opo r t i onally to 3 0% 139-647 mgd 6.09 28.33 m3s1 27.3 30% of total flow >6 47 mgd > 28.33 m3s1 12.4 194 mgd/8.49 m3s 1 (peak withdrawal rate) The La Nina -r elated drought i n 1999-2000 resulted in a prolonged period (8 months or l onger) when no withdrawals would have been made from either site under the current withdrawal schedule (Table 17). The obv iou s question that follows is: From where would Tampa Bay ge t i ts drinking water in this situation? The implication of these results is that the freshwater supply will need to be managed using a flexible p l an that maximizes supply from a variety of sources and minimizes jeopardizing any one source or the future supply. For example, alternate freshwate r sources are necessary for those periods of time with ex t ended l ow-flow conditions. The Master Water Plan projects include not only freshwater withdrawals but a lso desalination additiona l g roundw ater wells, and a reservoir. Of all of these additional sources of drinkin g water sup pl y, the re servo ir has the most potential to augment sources during periods of no or little s urfa ce wate r withdrawals due to drough t conditions. The re se r voir will provide storage during hig h flow periods and its anticipated storage capacity is approximate l y 15 billion gallons (Tampa Bay Water 2000) This prop osed reservoir i s not schedule d for completion until 2005, and its long-term mean supply is anticipated to provide 60-66 mgd, providing an estimated 7.5 months of drinking water supply. 84

PAGE 98

Table 17. Percent of days from each month of the 1999 and 2000 water years that would have met the minimum discharge for withdrawal at the Tampa Bypass Canal and Hillsborough River withdrawal sites Month/Year ENSO SSTA p ercentage of Days Meeting Minimum C Discharge Leve l for Withdrawal Hillsborough TBC Oct. 1998 -1.1 I 19 100 Nov. 1998 -1.18 90 100 Dec. 1998 -1.61 100 100 Jan. 1999 -1.45 77 100 Feb. 1999 -1.20 100 100 Mar. 1999 -0.83 100 100 Apr. 1999 -0.74 100 100 May 1999 -0.56 100 1 00 Jun. 1999 -0.81 100 10 0 Jul. 1999 0.62 42 100 Aug. 1999 -1.0 5 74 55 Sep. 1 999 -0.70 90 20 Oct. 1999 0 82 16 9 7 Nov. 1999 -1.20 0 14 Dec. 1999 1.50 0 0 Jan.2000 -1.79 0 0 Feb.2000 -1.49 0 0 Mar. 2000 -1.03 0 0 Apr. 2000 0.06 0 0 May 2000 -0.50 0 0 Jun.2000 -0.45 0 0 July 2000 -0.30 0 0 Aug. 2000 -0.20 3 64 Sep.2000 -0 .36 47 1 00 POTENTIAL ROLE OF ENSO PREDICTIONS ENSO prediction s are not accurate enough ( nor are ENSO teleconnections to local precipitation levels clear cut enough) to be us ed as a short-term (daily or weekly) management tool. However, as a longer term mana geme nt tool such predictions may be useful. Predictions of ENSO conditions up to one year into the future are available from many sources; the U.S. Nationa l Weather Service offers "Nifio-3.4 SST Outlooks" through the Climate Prediction Center's web s ite (CPC 2001). The foreca s ts are generated by CMP 14 a coupled ocean-atmosphere g lobal climate model and consist of 85

PAGE 99

16 forecasts made from a combination of four weekly initial conditions and four atmospheric initial conditions within one calendar month The most recent Nifio-3.4 SST Outlook shows warming in the equatorial Pacific and a change over to El Nifio conditions by the end of2001 (CPC 2001). Climate predictions are not infallible ; however, they do provide advance warning of extreme conditions such as drought and floo ding Such predictions may be useful to help Tampa Bay Water to balance its water supply sources in order to maximize reservoir storage prior to anticipated La Nina drought conditions. IMPLICATIONS FOR TAMPA BAY'S ESTUARINE ECOSYSTEM The most direct impact of the proposed surface water withdrawals from the Hillsborough River, Palm River / Tampa Bypa ss Canal, and Alafia Rivers will be changes in the amount of freshwater entering Tampa Bay. This will affect the salinity distribution within the downstream portions of the riv e r and perhaps in Hillsborough Bay These changes potentially have implications for the circulation d yna mic s of the bay, which are determined in part by a combination of d ens ity difference s astronomical tides and winds. The density-driven circulation within estuaries is controlled by horizontal salinity gradients and is influenced by the net freshwater supply into estuaries The strength of the density driven circulation plays a large role in the residence time of Tampa Bay particularly in Hillsbor ough Bay (Burwell 200 I). Changing s alinity distribution s within the tidal portions of Tampa Bay's rivers impacts its ecosystem In particular, Tampa Bay's function as a nursery for fish and invertebrates is tied strongly to the tidal s tretches of its rivers For example recent analyses of juvenile and adult bay anchovy (Anchoa mitchilli (Valenciennes, 1848)) indicate that abundance and distribution in the tidal portion of the Little Manatee River are influenced strongly and significantly by variations in freshwater inflow and their impacts on salinity (Matheson et al. 2001 ). Peebles (1999) notes that the distributions of freshwater zooplankton larvae of freshwater insects larval stages of some estuarine dependent fishes and invertebrates planktonic fish eggs, marine-derived fish larvae and marine-deri ve d planktonic invertebrates respond to variations in freshwater inflows or associated changes in salinity di s tributions in Florida s tidal rivers. Studies such as these demonstrate the biological importance of freshwater inflows and salinity distribution in Tampa Bay s tidal rivers and document that natural variability in freshwater inflows and salinity distribution impact the bay's fauna These organisms are adapted to these variable conditions, which are ultimately controlled by climate variability such as El Nino-Southern Oscillation. However it is d i fficult to speculate on what will be the combined impacts of natural variability and the proposed withdrawals on the bay's fauna and flora. The HEMP's role in detecting and assessing the potential 86

PAGE 100

impacts of the withdrawals on the hydrology and ecology of the associated tidal river segments of Tampa Bay will only be effective if it is framed within the context of natural climate variability and its impacts on freshwater in the Tampa Bay area. DISCUSSION The alteration of the timing and duration of low flow conditions associated with La Nina related drought has the potential to disrupt or adversely impact Tampa Bay's ecosystems which are adapted to natural variability in freshwater inflows and their impacts on salinity distribution circulation, and water quality. Consider the following hypothetical (but plausible) scenario: During low flow conditions, no water is withdrawn from the rivers and the bay's salinity and other parameters adjust naturall y. However, once drought conditions cease and the river discharge increases, there will be an immediate diversion for drinking water (to take burden off groundwater resources, the reservoir system, etc.). By changing the scale and timing of flow conditions this will prolong artificially the effects of the drought conditions in the bay and this may result in deleterious environmental impacts Given the short-term nature of the HBMP, it is unlikely that the impacts of this artificial extension of low flow conditions on natural systems can be modeled or anticipated Inherent in the hypothetical example presented above is that maintaining minimum flow levels is not sufficient for maintaining the environmental integrity of the bay's ecosystems. The withdrawal schedules must also maintain natural variability in discharge at many time scales (days, weeks, seasons, years ... ) In addition, as documented by this case study, the proposed withdrawal schedules will not provide adequate supplies during prolonged droughts. The proposed w ithdrawal schedules have been devised so that in this situation, the e stuari ne ecosystem comes first (no withdrawal). However, it remains to be seen if during an actual drought these schedules are enforced or if political pressure s override environmental stewardship. SUMMARY AND CONCLUSIONS Natural variability in the myriad of physical processes that impact and control estuaries occurs at time scales that typically may exceed or p art l y exceed many monitoring programs. With respect to documenting and monit o ring imp ac ts of human influence s on estuaries, it is therefore important to frame short-term monitoring programs within the context of longer term natural variability in the environment. This case study uses Tampa Bay, Florida and its proposed s urface water withdrawals to examine the issu e of how documented impacts of climate variability affect the monitoring program design e d to detect and assess th e potential impacts of the withdrawals on the hyd ro logy and ecology of the associated tidal river segments. 87

PAGE 101

Specifically this case study finds that EN SO-related variability in precipitation and river discharge in the Tampa Ba y area occur at time scales that may interfere with th e characterization of normal" river and estuarine conditions by the 3-year baseline monitoring program. In addition, ENSO conditions through teleconnections to prolonged drought in Florida during La Nifi.a events, may make river withdrawals for drinking water supply unavailable for more than 6 consecutive months. Prediction s of ENSO conditions may allow planners to optimize water supply from multiple so urce s over long time periods (up to a year). However, it remains to be seen if the current Master Water Plan can provide an adequate water supply for Tampa Bay's inhabitants while protecting a natural estuarine ecosystem that is dependent on seasonally and interannually variable freshwater inflow. 88

PAGE 102

NEW DIRECTIONS The research presented in this dissertation evolved through time as new ideas data and ' directions became available. Because it did not follow a systematic game plan it is fruitfu l to revisit the research methods and results and create such a plan for similar research. This plan is an outline of the different steps and considerations involved in this planning this research. It will be useful for researchers approaching similar problems in other areas and reflects the knowledge and insight gained from this dissertation. It also includes recommendations for improving the approaches and techniques used in this research. IDENTIFICATION OF THE RESEARCH PROBLEM In o rder to maximize the usefulness of the r esearch and its results to managers and planners, it i s important to s tart with the end" of the process-the iss u e of inter es t for local p l anners and managers. This may relate to water r esources (such as the results presented in this dissertation) ecosystems, sal ini ty and ci r culation, water quality, and many other environmenta l i ssues. Some important considerations include the followin g : How is th e issue of interest measured? What standards and / or criteria are used for decision making ? Research should co n form to or exceed these standards Research should incorporate c rit eria What are strengths and weaknesses of the available measurements ? For example fecal coliforms are standard measurement t ool for water quality, even though they are only a proxy measurement and their presence is influenced by other sources and factors. There are standard teclmiques and standards related to fecal coli forms that are used to indicate water quality. LINKS WITH CLIMATE VARIABILITY Wlziclz climate patte rns are important for tlze links / issue of inter es t ? For many localities and i ssues of interest surveying both refereed scien t ific literature and local gray li terature from various management agencies will b e useful in determining t he types of climate variabilit y that ha v e the greatest influences and what the meteoro l o g ical connection i s between climate variabilit y and the l oca l area. 89

PAGE 103

From thi s su rvey it is important to determine what tim e scales of variability are important. There are probably multiple time sc ales involved ; from daily to seasonal to a nnual to in terannua l to interdecadal and l onger. Important considerations include: For which time scales are data availab l e to examine relationships? How are they measured? What spatia l sca les are important? Links to climat e A useful step in formulating a climate variab ility research plan is to create a model of all ofthe factors that link climate variability to the issue of int erest. For example, climate variabi l ity is linked to estuarine salinity through its impacts on precipitation and discharge. There may be multiple links within the model; for example, salinity will be influenced by freshwater discharge, rainfall over the bay evaporation, its connection to residual circulation, the feedback between r esidua l circulation and freshwater input/salinity gradient, etc Freshwater d isch arge w ill be influenced by river discharge, rainfall, surface runoff withdrawals and diversions, sewage discharge and other factors. Survey availabl e data so urc es for m easu rin g links It m ay not be possib l e to find data (or an adequate amount of data) for all of the links between climate variability and the issue of int e re st. The period of record for the different data needs to be evaluated r e l ative to the time scale of the climate variabi li ty. Also, the collection frequency of the data may not be adequate to examine the relationship between the link and climate variability. Another consideration is the adequacy of t h e distribution of the avai l ab l e dat a spatia ll y. An a d dit i ona l consideration is meeting the standards and c riteri a for decision-making. After the characteristics of the available data h ave been ascertained, a decision on which data t o u se can be made by examining the advantage s and disadvantages of the different data sources. Although thi s approach was not used in this research it may be possible to construct proxy measurements of certain types of d ata to provide impro ved data coverage and resolution. For examp l e, Dopp ler radar rainfall estim ates from over Tampa Bay mi ght be combined with data from waters h eds with historical rainfall r eco rds to get an area int egrated rainfall vol um e throu g h time. This would provide an im proved estimate of the amount of freshwater e nterin g Tamp a Bay. A different approach might be to use the output of watershed runoff models for th e Tampa Bay watershed to get watershed-w i de estima te s of freshwater input (rivers and runoff) in to Tampa Bay. 90

PAGE 104

I nte r act i ons b e t wee n link s Although it may not be possib l e examine all of the links (because of data availability and quality), it i s important to attempt to quantify as best as po ssible the inter-rela tionships between various links. For example, rainfall influence s salinity both directly and indirectly (through its link to freshwater disc harge). What are the relative contributions of both the direct and indirect influences ? Feedback s are another type of interaction between links; for example, the feedback between freshwater inpu t and residual circulation in an estuary that has implications for sal inity di str ibution ADDITI ONAL FACTORS TO CONSIDER Human imp ac t s have the potential to influence many of the links between climate variability and the issue of interest and may also influence directly the issue of interest. For example, land-use change both through t ime and spa ti ally in the loca l area impacts river discharge (with respect to the timing and duration of flow events and to withdrawals and diversions), runoff patterns (with respect to magnitude, timing, and duration of flow events), discharge from both sewage treatment plant s and the s torm drain systems, salinity and re si dual circulation (with respect to alteration of the estuary by causeways finger canals, and dred ge and fill related to maintaining channels) and water qua l i ty (because of its relationship to many of the impacts discussed above and because of changing population le ve l s). An additional potential impact related to human activities is global climate change. This may introduce trends i n the data that are superimposed on variability related to climate or that interact with climate variability. The i nteraction between natura l and human influences on climate and its links to other natural systems is a complicated subject. I n addition to examining how natural and human influ e nces interact other factors must be considered. For example, w hat time and spatial sca les are important? These interactions ma y impact analyses of climate variabi l ity and its influ e nce on the issue of interest. ANALYSES Determining the strength of the r e lation s hips between the iss ue of interest and climate variabi l ity involves many considerations. Re l ationships mu s t be shown with appropriate leve l s of uncerta i nty to meet standards of planne r s and must be quantified appropriately for use and consideration by planners. Thi s may limit or guide the choice of statistical m e th o d s used in the analysis, a lthough additional statistical tests may be perfonned after the requirements for planners have been met. For examp l e, "simp l e statistical meth o d s such as means standard deviations, and conventional s tati stical tests, in addition to graphical representations, may be useful to establish patterns and re la tion s hip s between the data. More comp l icated methods, such as time-ser i es analysis, empirical 91

PAGE 105

orthogonal functions (EOF) analysi s, or wavelet analysis may be useful for examining the interactions between variables 92

PAGE 106

REFERENCES Albright, M. 1998. El Nifio came to town and kept tourists away. St. Petersburg Times, 9 April, p. E1. American Association of Port Authorities. 2000. U.S. Port Ranking b y Ca r go Volume, 1998 [Avai l ab le a t http ://www.aapa -port s.o rg] Barbe, D. E ., and J. C. Francis. 1995. An analysis of seasonal fecal coliform levels in the Tchefuncte River. Water Resources Bulletin 31:141-146. Barnston A. G. and R. E Livezey. 1987. Classification, seasonality and persistence of lo w-frequency atmospheric ci r cu l ation patterns. Monthly Weather Review 115:108 3-1126. Bendis Brian. 1999. Water quality trend s in Tampa Bay, Florida. Unpublished master's thesis Uni vers i ty of South Florida Department of Marine Science, 84 pp. Boler, R.N. 1998. Surface water quality: 1995-1997, Hillsborough County Florida. Eni vornmental Protection Commission of Hillsborough County Variously paginated. [Available from EPC, 1 900 9th Avenue Tampa, FL 33605] ---, R. C. Molloy and E. M Lesnett. 1991. Surface water quality monitoring by the E n vironmen tal Protection Commission of Hillsborough County. In S. Treat and P. Clark (eds.), Proceedings of the Tampa Bay Area Scientific Information Symposium 2, Tampa, Florida, February 27-March 1, 1991. Bove, M. C. J. B. Elsner C. W. Landsea X. Niu, and J. J. O Brien. 1998 Effect of E l Nifio on U.S. landfalling hurricanes re v isited. Bull etin of the American Meteorological Society 79:2477-2482. Boyer, J. N., J. W. Fourqurean, and R. D. Jones. 199 9. Sea so nal and long-term trends in the water quality of Florida (1989-1997). Estuaries 22:417-430. Brown, B. G., R. W. Katz, and A. H. Murphy. 198 6. On the economic val u e of seasonal-precipitation forecasts: The fallowing/p l anting problem. Bulletin of the American Meteorological Society 67:833 -841 Burwell, D. B. 2001. Modeling residence times: Eulerian vs. Langrangian. Unpublis hed Ph.D Dissert a tion, University of South Florida, College of Marine Science Tampa, Florida. 101 p. 93

PAGE 107

Checkley, W L. D. Epstein, R H. Gilman D Figueroa R.I. Cama, J. A. Patz and R E. Black 2000. Effects of El N ino and ambient temperature on ho s pital admission s for diarrhoeal di seases in Peru v ian children. Lancet 335 : 442-450 Chen, E., a nd 1. F. Gerber. 1990. Climate. Page s 11-34 in R L. Myer s and J. J. Ewel, eds., Ecosystems of Florida. University of Centra l Florida Press : Orlando 765 pp C hen X., M S Flannery, and D L. Moor e 2000 Response times of salinity in relation to changes in freshwater inflows in the l owe r Hillsborough Riv er, Florida Estuaries 23:735-742. C lim ate Prediction Center. c it ed 20 00. El Nino-Southern Oscillat ion and its as so ciated l inks. [A va ilabl e on-line from http: // www.cpc n cep.noaa. gov / products / analysism oni t o ring / ensostuff /.] Climate Predi c tion Center. Cited 2001. Nino-3.4 SST Outlooks [ Avai l able at http://www nnic .noaa. gov /p roduct s / analysis_ moni torin g/l anin al ensoforecast.html] Coastal Environmental / PBS&J, Inc. 19 98 Cumulative impact analysis for Master Water Plan projects. Final Report prepar e d for the We st Coast Regional Water Supply Authority, Clearwater, FL. [Available fro m PBS&J Inc., 5300 W. Cypress St. Suit e 300, Tampa FL 33607 USA] Compagnucci R. H and W M. Vargas. 1998. Int e r-annual variability ofthe Cuyo Rivers' streamflow in the Argentinean Andean Mount a ins and ENS O events. Int e rnational Journal of Cli matol ogy 18:159 3-1609 Cullen H M. an d P. B d e Menocal. 2000. North At lantic influence on Tigri s Euphrates streamflow. Int er nati o n a l Journal of C limatology 2 0:853-863. D awson, A. G and G. O'Hare. 20 00. Ocean-atmo s phere circu l ation and g l oba l cl i mate: TheEl Nino-Southern Oscillation. Geog r aphy 85:1 93-20 8 DelCharco M J. 1 998. Tidal flow in se l ec ted area s of Tampa Bay and Char l otte Harbor Florida, 1995-96. U.S. Geological Survey Water-Resources Investig a tion s Report 97 4265, T alla h assee, Florida Dep e tris P. J., S. K e mp e, M. Latif and W. G. Mook 1996 EN SO-controlled floodin g in the P arana River ( 1 904-1991). Na turwi ssensc haft e n 83: 1 27-129. E lsn er, J. B., K -B. Liu, and B Kocher. 2 000. Spatial variations in major U.S. hurrican e activity. Statistics and a physical mechanism Jou r nal ofClim a t e 13:229 3-2305. Enfield, D B 1996. Relationships of inter-American rainfall to tropical Atlantic and Pacific SST vari abil i ty Geophys i cal Research Letters 23:3305-3308. E n vironme n ta l Prot ection Agency. 1 999 Action Plan for Beaches and Recreational Waters. EPA Report EPA/600 / R-98 / 079 W as hin g ton DC 1 9 pp. [Available 94

PAGE 108

from Environmental Protection Agency, 1200 Pennsylvania Avenue, NW Washington DC 20460.] Epstein, P. R. 1998 Watching El Nino. Public Health Reports 113: 331-332. Estevez, E. D., L. K Dixon, and M.S. Flannery. 1991. West-coastal rivers of peninsular Florida. Pages 187-221 in R J. Livingston ed., The Rivers of Florida Springer Verlag Ecological Studies 83, New York 289 pp. Ferguson, C. M B. G Coote, N.J. Ashbolt, and I. M. Stevenson. 1996 Relationships between indicators, pathogens and water quality in an estuarine system. Water Resources 30:2,045-2,054. Flannery, M.S. 1989. Tampa and Sarasota Bays' wate rsheds and tributarie s, p. 18-48. In E. D. Estevez (ed.), NOAA Estuary-of-the-Month, Series No. 11. Washington D.C 215 p ---,H. C Downing Jr., G A. McGany, and M. 0 Walters 1992 Increa se d nutrient loading and baseflow supplementation in the Little Manatee River watershed Proceedings of the Tampa Bay Area Scientific Information Symposium (BASIS) 2 Tampa, FL, 369-395. [Available from Tampa Bay Regional Planning Council 9455 Koger Blvd., St. Petersburg, FL 33702 ] Flannery, M S 2001. Personal communication. Fromentin, .J.-M., and B Planque. 1996. Calanus and environment in the eastern North Atlantic. II. Influence of the North Atlantic Oscillation on C. finmarchicus and C. Helgolandicus. Marine Ecology Progress Series 134:111-118 Galperin, B., A F Blumberg and R H Weisberg. 1991. A time-dependent three dimensional model of circulation in Tampa Bay. InS. Treat and P. Clark (eds.), Proceedings ofthe Tampa Bay Area Scientific Information Symposium 2, Tampa Florida February 27-March 1 1991. Gershunov, A., and T P Barnett. 1998a ENSO influence on intraseasonal extreme rainfall and temperature frequencies in the contiguous United States : Observations and model results. Journal of Climate 11:1575-1586 Gershunov A., and T P Barnett 1998b. lnterdecadal modulation ofENSO teleconnections. Bulletin ofthe American Meteorological Society 79:2,715-2 725. Giannini, A., Y Kushnir, and M.A. Cane. 2000. Interannual variability of Caribbean rainfall ENSO, and the Atlantic Ocean. Journal ofClimate 13: 297-311. Goodwin, C. R. and D M Michaelis. 1976 Tides in Tampa Bay, Florida: June 1971 to December 1973 United States Geological Survey Open File Report FL 75004, Tallahassee, Florida. 95

PAGE 109

Goyal, S.M., C. P. Gerba and J. L. Melnick. 1978. Prevalence ofhurnan enteric viruses in coastal canal communities. Journal of the Water Pollution Control Federation 50:2 ,2 47-2,257 Gray W. M. 1984. Atlantic seasonal hurricane frequency Part II : Forecasting its variability. Monthly Weather Review 112:1,669-1,683. Hansen, J. W., A. W. Hodges and J. W. Jones 1998. ENSO influences on agriculture in the southeastern United States Journal of Climate 11 :404-411. --, J. W Jones C. F. Kiker, and A. W. Hodges. 1999. El Nino-Southern Oscillation impacts on winter vegetable production in F l orida. Journal of Climate 12:92-102. Hanson, K and G A. Maul. 1991. Florida precipitation and the Pacific El Nino 1895-1989. Florida Scientist 54 : 160-68. Hartley, S ., and M. J. Keables 1998. Synoptic associations of winter climate and snowfall variability in New England USA, 1950-1992 International Journal of Climatology 18:281-298. Hartley, S. 1999. Winter Atlantic climate and s nowfall in the South and Central Appalachians. Phy s ical Geogeography 20:1-13. Harwe ll, M .. A., J. S. Ault, and J. H. Gentile. 1995 Comparison ofthe ecological risks to the Tampa Bay ecosystem from spills of fuel oil #6 and orimulsion. Comparative Ecological Risk As sess ment Volume 1 Center for Marine and Environmental Analy ses, Rosenstiel Schoo l of Marine and Atmospheric Science, University of Miami FL 177 p Hoerling M.P., A. Kumar, and M. Zhong. 1997. E l Nino, La Nina, and the nonlinearity oftheir teleconnections. Journal ofClimate 10: 1 769-1,786. Huang, J., K. Higuchi and A. Shabbar. 1998. The relationship between the North Atlantic oscillation and El Nino-Southern Oscillation Geophysical Research Letters 25:2,707-2,710. Hurrell, J. W 1995 Decadal trends in th e North Atlantic o sc illation: Regional temperatures and precipitation Science 269:676-679. ---. 1996. Influence of variations in extra tropical wintertime teleconnection s on Northern Hemi sphere temperature. Geophysical Research Letters 23:665-668. ---,and H. van Loon 1997 Decadal variation s in climate associated with the North Atlantic oscillation. Climatic Change 36:301-326. Jones, P. D., T. Jonsson and D. Wheeler. 1997. Extension to t he North Atlantic Oscillation using early instrumental pre ss ure observations from Gibra l tar and SouthWest Iceland. Int erna tional Journal of C limatology 17:1, 433-1 450 96

PAGE 110

Kahya, E., and 1. A. Dracup. 1993. U S. streamflow patterns in relation to the El Nino/ Southern Oscillation. Wat e r R eso urc es Resear c h 29:2,491-2,503. Kapala, A., H. Machel, and H. Flohn. 1998 Behaviour of the centres of action above the Atlantic si nce 1881. Part II: Associations with regional climate anomalies. International Journal of Climatology 18:23-36. Kiladis G N ., and H. F. Diaz. 1989 Global climatic anomalies associated with extremes in the Southern Oscillation. Journal ofClimate 2:1,069-1,090. Lewis, R R. III, and R. L. Whitman, Jr. 1985 A new geographic description of the boundaries and s ubdivisions of Tampa Bay, p 10-18 In S F Treat, 1. L. Simon, R. R Lewis, III, and R L. Whitman Jr. (eds.), Proceedings of the Tampa Bay Area Scientific Information Sympo si um, 1982. Lipp, E. K., N. Schmidt, M E. Luther and 1. B. Rose 2001. Determining the effects of El Nino-Southern Oscillation events on coastal water quality. Estuaries, in press. Live zey, R E., and T. M Smith. 1999 Covariability of aspects ofNorth American climate with g lobal sea surface temperatures on interannual to interdecadal timescales. Journal ofClimate 12:289-302. ---, M. Masutani, A. Leetmaa H. Rui M Ji and A. Kumar. 1997. Teleconnective response of the Pacific-North American region atmosphere to large central equatorial Pacific SST anomalies. J ournal ofClimate 10 : 1 787-1,820. Machel, H., A Kapala, and H. Flohn. 1998. Behaviour of the centres of action above the Atlantic since 1881. Part I: Characteristics of seasonal and interannual variability. International Journal of Climatology 18: 1 -22. Malmg ren, B A., A Winter, and D Chen. 1998. El Nino-Southern Oscillation and North Atlantic oscillation control of climate in Puerto Rico. Journal of Climate 11:2 ,7 13-2 ,71 7 Matheson, R. E., Jr., D. A. R y dene, and B. H Lawrence. 2001. Distribution and abundance of the bay anchovy, Anchoa mitchilli in the Little Manatee River, Tampa Bay, Florida, including the effects of variations in freshwat e r inflow Manuscript in preparation. [Available from Florida Fish and Wildlife Conservation Commission, Florida Marine Research Ins ti tute 100 8th Ave. SE St. Petersburg FL 33 701] McCabe, G. 1., and M D Dettinger. 1999. Decadal variations in the strength of ENSO teleconnections with precipitation in the western United States. International Journal of Climatology 19:1,399-1,410. Montroy, D L. M. B Richn1an, and P. 1. Lamb 1998. Obse rved nonlinearities of monthly teleconnections between tropical Pacific sea s urface temperature 97

PAGE 111

anomalies a nd central and eastern North American precipitati o n Journal of Climate 11: 1,2 18-1 835. National Center for Atmospheric Research. 19 94 Report to the nation: On our changing pl anet. Report pur suant to NOAA Award #NA27GP0232-01, Spring/# 3 23 pp. National Oceanic and Atmospheric Administration (NOAA). 1 998. Population : Distribution den s i ty and g rowth NOAA's State of the Coast Report. URL: http: I I state_of_coast.noaa.gov /bulletins/htrnl/pop_Ol I pop.htrnl ---. cited 2000. Popul atio n : Di s tribution density and growth NOAA's State of the Coast Report 1998. [A vai labl e online at http:/ I state_of_coast.noaa.gov I bulletins/htmllp op 0 1 /pop. html.] Noreen, E. W. 1 989. Compu t er intensive Methods for Testing Hypoth eses. John Wiley and Sons, 229 pp. N uttl e, W. K., J. W. Fourgur ean, B. J. Cosb y, J. C Zieman, and M. B Robblee. 2000. Influence of new fres hwater su p ply o n salinity in Florida Bay Water R esou r ces R esearch 3 6: 1805 -1822 Osborn, T J., K. R. Briffa S F. B Tett P. D. J ones, and R M Trigo. 1 999. Evaluation of the North Atlantic oscillation as s imulated by a coupled climate model. C limat e Dynami cs 15:685 -702. PBS&J, Inc. 2000. Tampa B y p ass Canal/Alafia River Water Supply Projects H ydro Biolog ical Monitoring Program. Draft r eport prepared fo r Tampa Bay W a ter. Variously paginat ed. [ Available from PBS&J, Inc ., 5300 W Cypress St., Suite 300, Tampa, FL 33607 USA] Peebles, E. 1999. An assessmen t of the effec t s of freshwa ter inflows on fish and invertebrate habitat u s e in th e Manatee River, B raden Ri ver, Peace River and Shell Creek estuari es Volume I : Monthl y monitoring. Report prepared for Southwest Florida Water Management. Variously paginated. [Available from SWFWMD, 7601 U.S. Highway JOIN, Tampa, FL 33638.] Piechota, T. C ., and J. A Drac up 19 99. Long-range foreca sting u sing El Nino-Southern Oscill atio n indic a tor s. Journal of H y drological Engineering 4:144-151. Portis, D. H ., J. E Walsh M. E l Haml y, and P J. Lamb. 2000. Seasonality of the North Atlantic Oscillation. Journal ofC/imate, s ubmi tte d Pos t E., and N.C. Stenseth 1 999 Climate change, plan t phenology and n o rth ern ungul a t es. Ecology 80: I ,322-l ,339. Power, S. T. Casey, C. Folland A Colman, and V M e hta. 1999. Int e r-decadal modulation of the imp ac t of ENSO on Australia. Climate Dynamics 15:319-324. 98

PAGE 112

Pritchard, D W. 1 956. The dynamic structure of a coastal plain estuary. Journa l of Marin e Resear c h 1 5:33-42. Qian, B. J. Corte-Real, and H. Xu. 2000. I s the North Atlantic oscillation the most important atmospheric pattern for Europe? Journal of Geophysical Resear ch 105:11 ,901-11, 910. Rasmusson, E M. and T. H Carpenter. 1982 Variations in tropical sea surface temperature and surface wind fields associated with the Southern Oscillatio n/E l Nino. Monthly W eather Review 110:354-384 ---,and J. M. Wallace. 1983. Meteorological aspects of the El Nino/Southern Oscillation. Science 222:1,195-1 202. Reisner, M. 1 986. Cadillac De sert (Second edition). Penguin Books: New York, NY. 582 pp. Reilly F. J., Jr. R. J. Spagnolo, and E. Ambrogio. 1999. Marine and estuarine shallow water sc ience and management: The interrelationship among habitats and t heir management. Estuaries 22:73 1-734. Rod6 X., E. Baert, and F. A. Comin. 1997. Variations in seasonal rainfall in Southern Europe during the pre s ent century: Relationships with the North Atlantic oscillation and the El Nino-Southern Oscillation Climate Dynamics 13:275-284 Ropelewski, C. F and M.S. Halpert 1986 North American precipitation and temperature patterns assoc iated with the E l Nino Southern O sci llat ion (ENSO). Monthly Weather Revi ew 1 14:2,352-2,362. -,and 1989. Precipitation patterns associated with th e high index phase ofthe Southern Oscillation. Journal ofClimate 2 : 268-284. ---, and 1996. Quantifying Southern Oscillation-precipitation relation ships. Journal of Climate 9:1 043-1,059. Rutherford, E. H., B. J. Bendis, G A. Vargo and K. A. Fanning. 1995. Review of historical Tampa Bay water qua lity data Fl orida Scientist 58:67 -81 Schmidt, N., E. K. Lipp, J. R. Rose and M. E. Luther. 2001. ENSO influences on seaso nal rainfall and river discharge in Florida. Journal of Clima t e 42:615-6 27 ---and M. E. Luther. 2001a. ENSO impacts on salinity in Tampa B ay, Florida. Estuaries, submitted. ---, and M E. Luther. 2001. Modulation ofENSO impacts in Florida by the NAO. Journal ofClimate, submitted. Sittel M C. 1 994a. Margina l probabilitie s of the extremes of ENSO events for temper a ture and precipitation in the southeastern United States. Technical Report 941 Center for Ocean-Atmospheric Studies, The Florida State University, 99

PAGE 113

Tallahassee, FL, 155 pp. [Available from Center for Ocean-Atmosphere Prediction Studies, The Florida State University, Tallahassee FL 32306.] ---. 1994b. Differences in the means of ENSO extremes for maximum temperature and precipitation in the United States. Technical Report 94-2 Center for Ocean Atmospheric Studies, The Florida State University, Tallahassee, FL, 50 pp. [Available from Center for Ocean-Atmosphere Prediction Studies The Florida State University, Tallahassee, FL 32306 ] Solow, A. R., R. F. Adams, K. J. Bryant D. M. Legler, J. J. O'Brien, B. A. McCarl W. Nayda, and R. Weiher. 1998. The value of improved ENSO prediction to U.S agriculture. Climatic Change 39:47-60. Southwest Florida Water Management District (SWFWMD). 1996. Surface water / Ground-water interrelationships. Volume 1, Northern Tampa Bay Water Resources Assessment Project. Variously paginated. [Available from SWFWMD, 7601 U.S. Highway 301N, Tampa, FL 33638.] ---. 1998. Comprehensive Watershed Management Plan Tampa Bay/Ancl ote River Vol. I. Assessment, Draft. [Available from SWFWMD, 7601 U.S. Highwa y 301N, Tampa, FL 33638.] Stephen s on, D. B. V. Pavan and R. Bojarui. 2000. Is the North Atlantic oscillation a random wal k? International Journal ofC!imatology 20:1-18. Stockton, C. W., and M. F. Glueck. 1999. Long-term variability of the North Atlantic oscillation (NAO). Pa ges 290 293 in the Proceedings of the American Meteorological Society's 10111 Sympo s ium on Global Change Studies, January 1015, 1999 Dallas TX. 562 pp. Squires, A. P., G. A. Vargo, R. H. Weisberg K. A. Fanning, and B Galperin. 1995. Review and synthesis of historical Tampa Bay water quality data Florida Scientist 58:228-233. Sun, H., and D. J. Furbish. 1997 Annual precipitation and river di scharges in Florida in response to El Nifioand La Nina-sea surface temperature anomalies. Journal of Hydrology 199:74-87. Tampa Bay Nationa l Estuary Program (TBNEP). 1993 Tampa B ay : Status and trends. St. Petersburg, Florida 45 p Tampa Bay National Estuary Program (TBNEP). 1995: Charting th e course: Preliminary action plans for the Tampa Bay National Estuary Program St. Petersburg, Florida, 254 p. Tampa Bay Water. 2000. New Water Plan. Quarterly Report September 2000, prepared for th e Southwe s t Florida W ate r Management Distri ct. [Avai lab l e from 100

PAGE 114

Tampa Ba y Water, 2535 Lan dmark Dri ve, Suite 211 Clearwa ter FL 33761, USA] Thompson, D W. J., and J. M. Wallace. 1998 T h e Arc tic Oscillation s i g n at ure in t he wint e rtime geopotential he i g ht and temp e rature fields. Geophysical Research L e it e r s 25: 1 ,297-1,3 00. Thompson, D W. J., J. M. Wallace and G. C. H ege rl. 2000. Annual modes i n the extratropical circulation, Part II: Trends. Journal ofClimate 13: 1 ,018-1, 036 Tiffany W. J., an d D E. Wilkinson. 1989. P orts and port impacts. NOAA Estuary of the Month 1 1 : 171-185. Tre nberth K. E. 1 991. Gene ral characteristics of El Ni no-S outhern O sci ll ation. P ages 13-4 2 i n M. H Glantz, R. W. Katz and N. Nicholls, eds. Teleconnections Linking Worldwide Climate Anomalies. Cambr id ge Unive r sity Press New York. 535 pp. U. S. Bureau of the Census. 1996. P op ulation in co asta l coun ties: Apri l 1 1990 and Jul y 1 1994. P op ul a ti o n Distribution Br a nch [Av ailable at h tt p ://www.census.gov / popul ation/estimates / county / 9094cstl.txt] U.S. E n v ir onmen ta l Protection Agency. 2000. Gulf o f M exico Pro gram: A framework for an inte grated and comprehensive monitoring plan for the estuaries of the Gulf ofMexico. E PA620/R-00 / 005. Variously p agi n ated U. S. E n v ir onmenta l Pr otect i o n Agency. 2001. C l ean Water Act i on Plan: National Coastal Con diti o n Report. EPA620-R-00-004. Dr aft. Variously p aginated. [Av a il ab l e from the Office of Researc h and Development/Office of Water, Was h i n gton D.C 20 4 60] Wallace J. M. 2000. North A tl antic oscillation/annu lar mode: Two paradigms--Dne phenomeno n Quarterly Journal of the Royal Met eorological Soci e ty 1 26:791 805. Weisberg, R H., and R. G. W i lliams. 1992. Ini tia l finds of the circulation o f Tampa Bay p. 49-66. InS. Treat an d P. Clar k (ed s.) Proceedings of the Tampa Bay Area Scient i fi c In for mati o n Symposiun1 2 Tampa, F l or id a Febru ary 27-March 1 1 991. Wenner, E L., and Maggie Geist. 2001. The Nat i ona l Estuarine Research Reserves Program to m on it or and pres erve estuar in e wate rs. C oastal Manag e m e nt 29 : 1-17. Wilbanks T. J., and R W. Kates. 1 999. G l obal change i n local places: How scale matters. Climati c Change 43:601-628. Winsberg, M.D. 1 990. Florida Weath er. University o f Centra l F l or i da Press: Orlando 1 7 1 pp. 101

PAGE 115

Wong, K.-C. 1994. On the nature of transverse variabi l ity in a coastal plain estuary. Journal ofGeophysical Research 99:14 209-14 222. Wyer, M.D. D. Kay, G. F. Jackson H. M. Dawson J. Yeo, and L. Tanguy 1995. Indicator organism so urces and coastal water quality: a catchment stu dy on the Isle of Jersey. Journal of Applied Bacteriology 78:290 296 Zarbock, H., A. Janicki, D. Wade D. Heimbuch and H. Wilson 1995. Current and historical freshwater inflows to Tampa Bay, Florida. Tampa Bay National Estuary Program Technical Publication # 0 1-94. Various l y pagina t ed. [Available from Tampa Bay Estuary Project 100 8th Avenue SE, St. Petersburg FL 33701, USA] Zervas, C. E., ed. 1993. Tampa Bay Oceanography Project: Ph ys ical oceanographic synthesis. NOAA Technical Report NOS OES 002, 184 p. Zhang, M. 1994 Predictability of sea level and currents of Tampa Bay. Department of Marine Science University of South Florida St. Peter sb urg unpublished Master's The s is. Zhou, J., and K.-M. Lau. 1999 Interannual and decadal variability of summer rainfall over South America. Page s 516-519 in the Proceedings of the American Meteorological Society's 1Oth Symposium on Globa l Change Studies, January 10-15 1999 Dallas, TX. 562 pp. Zorn, M. R., and P.R. Way l en. 1997. Sea so nal r es ponse of mean monthl y strea m flow to E l Nino/southern oscillation in north central Florida. Professional Geographer 49:51-62. 102

PAGE 116

APPENDICES 103

PAGE 117

APPENDIX 1 DETERMINING THE EFFECTS OF EL NINO-SOUTHERN OSCILLATION EVENTS ON COASTAL WATER QUALITY PREFACE The re searc h presented in th is Appendix repre sents the results of a collaborative re search effort by the author, Nancy 1. Schmidt, and Dr. Erin K. Lipp of the University of Maryland's Biotechnology Institute, Center of Marine Biotechnology. It is included in thi s di ssertation in order to present fully the bre adt h and depth of climate variability impacts in th e Tampa Bay r egio n INTRODUCTION The link between climate and h ealth was recogni ze d as early as Hippocrates and through the 16th century (Rees 1 996). Without knowledge of disease-causing agents, many believed that strange weather patterns caused a variety of health problems, re s ulting in the adage under the weather" (Rees 1996) While it has long been realized that bad a irs do not cause di sease, in recen t years there has been increa se d scie ntific r ecogni ti on that clim ate variability contributes to the distribution growt h and surviva l of certain pathogenic microorgani s ms and, therefore, impacts pub lic he a lth (Colwe ll 1 996; NOAA 1 999; C h eckley et al. 2000). Both loc a l weat h e r patterns and climate variabi lil y play a role in th e dispersion of pathogenic mi croorganisms. Events such as extreme rainfall and flood s often overburden water treatment facilities and onsite di sposal systems and increase s torm wate r run-off; all of which may r esu lt in the introduction of high level s of ente ric pathogens to nearby surface waters and wells. Inter ann u a l climate varia bility du e to E l Nino-Southern Oscillation (ENSO) even t s and other phenomena can also affect water quality and public health both directly and indirectl y by r es ultin g in poor sanitation due t o flood s (G u er i e t a!. 1 986) or promoting favorable conditions for growth/s urvi va l of certa in pathogen s, i.e. Vibrio cholerae (Co l we ll 1 996). Also given the importance of non-point so ur ces of p o llution in th e Unit ed States and elsewhe r e, heavy or prolonged rains m ay contribute to pollutant l oad in g including pathogenic microorganisms from urban and agricu ltural run-off and on-site sewage disposal (O'Shea and Field 1992; Paul e t al. 19 97) Expos ur e to the public during the se even t s occurs from the contamination of drinking water recreation a l water and s h e llfish (Rose 1997). Short-term predictive mod e l s foreca s t ENSO event s with vary in g success rates and research h as effec ti ve ly demonstrated a strong relationship between regional precipitation patterns and ENSO e ve nts (Ropelewki and Halpert 1986 ; Schmidt et al. 200 1 ). In a ddition to th e importance and utility of regional-scale models an und ers tandin g of 104

PAGE 118

Appendix 1 (Co ntinued) anomalies in local weather patt e rn s i s an important and immed ia te concern. However, there has been little wo r k to d e m onstrate s tati s ticall y s ignifican t weather patterns, o r determine the ancillary effect s of th ese ano m al ie s related to ENSO events at the local scal e. Thi s contribution build s upon previous re s earch b y the authors o n t he topic of ENSO influence on local variability in seaso nal rainfall and ri ver d i scharge in Florida. Using an approximate ran domized di ffe ren ce of m eans tes t Schmidt et al. (200 1) demon s trated s i gnificant seaso nal r espon ses of rainfall and s tream flo w to E l Nino and La Nina conditions in south central Florida The st udy also found s ig n ific ant seaso n a l variabi li ty in rainfall within the sta te of Florida, with di s tinct pattern s n oted p artic ularl y b e tween the panhandle and so uthernmo st Florida. We h y pothe s ize that a statis tical relationship may exist betwee n ENSO events and wate r quality, based on rep o rted relationships between the El Nino -S o u t hern Oscillation precipitation and ri v er dis charge (Sun and Furbish 1997 ; Zorn and Waylen 1 997) and the s ub seq uent relati onship b etween wate r qu al ity and both rainfa ll and di scharge (Barbe and Francis 1 995, Lipp et al. in pr ess). Us in g historical d ata, we analy z ed the relationship b e t wee n microbiologic a l water qualit y and ENS O eve nt s in south central Florida, a r eg ion tha t is known to experience anoma l ous precipi t at ion and river flow associa te d with ENSO phases (El Nino and L a Nina; Schmidt e t al. 2001 ). The s trength of t h ese rel a tionships and seasonal changes were evaluated with analyses of continuous and categorical data. Here we demonstrate a si mple approach to define the role of p art ic u la r modes of climate variability (i.e., ENSO even ts) on coastal water quality by focusing on te mporal and spatial sca le s that are important to public health decisions a t loca l le vels u s ing Tampa Bay, Florida (USA) as a t est case. MATERIALS AND METHODS D esc ription of Study Site Hi s t o rical changes in water quality an d their relation s hip to th e E l N i no -S ou thern O s cillation were a s se ssed in Tampa Bay Florida. Tampa Bay i s the seco nd lar ges t Gulf Coast estuary and the l argest estuary in F l o rida. The entire wate r she d co n tains 35,500 km2 which are drained b y 31 major ba s in s (SWFWMD 1998 ) We studied seven drainage bas ins (Fig. 1 ) which are qualitat i ve l y d esc rib ed in t erms of major land-u se patt e rn s and other sources of poll u ti o n i n Table 1. Currently, withi n th e entire Tampa Bay watershed 56 % of the land i s d eve l ope d ; 40% of the built-up areas are urban and 16% include agricultu ral and p as tur e l an d s (SWFWMD 1998 ). 105

PAGE 119

Appe nd ix 1 (Continued) Tabl e 1. Description of Tampa Bay drainage basins. Basin Drainage Major Land Use Hil l sborough River 43I,742 Agriculture (32% ; north and ce ntra l ) ; Urban & Indu s tri a l (25 % ; south) Alafia River 270,000 Agric ulture (68% ; sou th ) ; Phosphate Minin g (east) ; Urban & Industria l ( north and wes t ) Bull frog Creek 25,758 A g r ic ulture (50%; eas t) ; Urban & R esidentia l (I2%; west) Ro cky C r eek 30,008 Urban (4 I %; south) ; Agriculture (nort h ) Swee t Water C r eek 23,896 Urban ( 69 %) D ela n ey C reek I 5, I 6 I Urban (55%) Littl e Manatee River I 35,046 Fo r est (38% ); A g r ic ultur e (84%) El Nino-Southern Oscillation Indices Notes Springs in u ppe r and lower r eac h es (I eac h ) ; dam an d drinking wa t e r reservoir 2 springs; > 9 1 % of watershed is developed o r altered Severa l lak es up t o 93 ac r es 2 lar ge l akes ( I9I and 283 acres) The sta te of the El N i no-Southern Oscillation was me as ured u s ing the Climate Prediction Center s Nino-3 .4 m onthly Sea Surface Temperature anomaly (SSTA) indices, which are based on recorded temperatures from to 5 S and 1 70W to 120 W in the equatorial Pacific Ocean. Se aso n s from 1 974-1998 were cla ss ified as extreme (El Nino or L a Nina) or neutr al. C lim atology indicates that F lor ida winters b eg in in January (Winsberg 1990), th erefore seaso n s were defi ned as follows: w inter includ ed January, February and March ; s prin g i nc l uded April, May and Jun e ; sum mer incl ude d July, August and September; and fall includ ed October, November and December. Ex tr e m e E NSO seasons were defined t o occur when th e five-month running ave rage centere d on t h e season of th e Nifio-3.4 SSTA exceeded 0.7 C. Neutral EN SO s ea sons were defined to occur when the five month runnin g average cen t ered on the season fell between 0.4 C (Ta ble 2). T here i s c urr en tl y no ge n erally accepte d sc h e me to d efine ENSO events; however, our clas s ifi ca tion ag ree s well with other publi s h ed lit era ture (Gershunov and Barnett 1 998b, Hoerlin g et al. 1 997) and in c lud es the mo s t wid e l y recognized and accepted ENSO events. Furthermore, ap pli ca ti o n of thi s scheme to the SST A data i s s tr a i ghtforward. Our thresholds ex clud e d qu es ti ona ble ENSO events whi le pro v idin g an adequa t e number of cases for analyses for all ENSO ph ase seasons exce p t extreme La Nina spring. Wh e re data were ava ilable ther e was often a s ignificant s i g nal not e d in extreme La Nina sp rin gs. Therefore for those ca ses where insufficien t data ex i ste d for extreme La Nina, all La Nina spr in gs ( 0.40 t o -0.6 C) were tested. 106

PAGE 120

0 00 (\J 0 ll) 0 I' (\J 0 -t r... (\J 0 C') 0 I' (\J 0 (\J-0 I' (\J 0 \ Appendix I (Contin u e d ) N W E s 0 20 Miles 1\) -...j 0 0 Figure 1. Map of Tampa Bay watersheds. Filled circles represent river gage stations, open circles represent water quality stations. BC: Booker Creek, RC: Rocky Creek, SC: Sweetwater Creek, CC: Cypress Creek, HR: Hillborough River, DC: Delaney Creek, AR: Alafia River, BFC: Bullfrog Creek, LMR: Little Manatee River and MR: Manatee River. 107

PAGE 121

Appendix I (Continued) Table 2. Classification of ENSO events between seasons, based on period of record available for water quality (1974-1998) Weak La Nifia events (Nifio -3.4 SSTA -0.4 -0.69) are included in parentheses. Season Extreme El Niiio Wint e r 1 983, 1 987, 1 992, J FM 1 995, 1998 Spring 1982, 1983, 1987 1992 AMJ 1993, 1997 Summe r 1 982, 1 987, 1991, 1997 J AS Fall 1976, 1977, 1982, 1986 OND 1987,1991,1994,1997 Neutral 1 979, 1 98 1 1982, 1990 1991, 1994 1 997 1976, 1977, 1978, 1979, 1980 1981, 1984, 1986, 1 990, 1994, 1995, 1996, 1 998 1 978, 1979, 1 980, 1981, 1 983, 1984 1985 1989 1990 1992 1 995, 1996 1 978, 1980, 1981, 1985 1989, 1 992, 1996 Extreme La Niiia 1974, 1976, 1 985, 1989 1 988 (1974, 1975, 1985, 1989) 1975 1 988, 1998 1 975, 1984 1988, 1995, 1998 Monthly wate r qua li ty data (I 974-1 998) were obt ai ned for Tampa Bay and its tributaries from the Hillsborough Co u nty Environmental Protection Commission (HEPC). W ate r quality was quantitatively asses se d u sing concentrations of f eca l coliform bacteria (co l o n y forming unit s (CFU)/1 00 ml) at 29 stations. B y convention, concentr atio n s for each samp l e we r e transformed by the equation, log10 (CFU/l 00 ml) + 1. When concentratio n s were be low detec t able limits ( < 1 -< 100 CFU/1 00 m l ), a value of zero was used for s t atistica l analyses. Samples we r e collected only o nce pe r m onth from each s t at i on. Consequently, stations were combined i n individual watersheds to mitigate small-scale events that might domi nat e a local area during the monthl y sampling but would not be indi cative of an entire drainage basin Adequate data coverage was available a t seven of the 3 1 drainage basins and includ e d an average of 4.1 sta tions per watershed. A maximum of 7 and a minimum of 2 stat ions were u sed per drainage basin. Appro ximate randomize d sta ti s tics were used i n all analyses. These computer intensive tests gene r ate the probability distribution of the te s t statistic by recomputing i t for many (> 1 00) artificially constructed data sets and can b e used to assess s i gnificance under minimal assump tions. The obse rvations that are tested do not need to meet the normal di st r ibu tion criteria of conventional p arametric statist i cs; likewise they need not constitute a r ando m sam p le. In addition to av oidin g the assumptions required of parametric stat i stics, approximate randomized tests maximize the abi li ty to discriminate between hypotheses beca u se the sampling distribut ion is known (Noreen 1989). Monthly anomali es in fecal coliform level s were obta ined b y s ubtracting th e bas in average d mean monthly value (over the d ata r ecord) from indi v idual bas i naveraged month l y values and were conelated against Nifio-3.4 monthly SSTA. Significance ofthe 108

PAGE 122

Appendix 1 (Continued) Pearson correlation coefficients was determined by comparing the r-value of the observed correlation to that of the distribution of the correlation under the null hypothesis. This distribution was gene r ated by randomly shuffling the SST A values against fecal coliform anomaly valu es and recalculating the r-value 10,000 times. Correlations were run against the entire data record, with a lag of zero to three months between monthly Nino-3 .4 SST A and water quality anomalies to allow the detection of any delayed responses. The correlation test provides information regarding the significance of the relationship between SSTA associated with the El Nino-Southern Oscillation and fecal coliform levels but does not reveal details concerning the relative importance of particular seasons. Therefore the differences in mean fecal coliform concentrations between extreme El Nino, neutral and extreme La N ina events for each season were analyzed u s ing an approximate randomi z ed difference of means test (Noreen 1989). Log-tr a nsformed basin averaged seasonal fecal coliform concentrations for extreme El Nino and La Nina events were tested against the seasonal fecal coliform levels for neutral periods. The observed difference in means was compared to the distribution of the random l y generated difference under the null hypothesis. As in the correlation analysis, recalculating the difference in means 10, 000 times generated the distribution. Given the high l eve l of noise inherent in the fecal coliform data set and the need to average over time and space results were considered to be statistically significant at an a level of 0.10 rather than 0.05. RESULTS Correlation Analysis Work by Schmidt et al. (2001) indicates that both precipitation and streamflow in south central Florida are significantly related to ENSO events. Given that water quality often deteriorates during periods of high precipitation and river discharge we hypothesized that a direct relationship may exist between ENSO events and water quality. Fecal coliform bacteria were used as a proxy for water quality as they are the most commonly used indicator of poor water quality due to fecal pollution and potential health risks wor l d wide. Correlation analyses were used to provide an initial assessment of whether any relationship existed between Nino-3.4 SSTA and changes in fecal coliform levels (water quality) in Tampa Bay. Analysis of monthly anomalies in fecal coliform levels with monthly Nino-3.4 SST A revealed a significant and positive correlation in five of the seven watersheds examined over the 25-year period of record. In general, even for significant correlations, coefficients were low (r = 0.088 to 0 23) However, these values were similar to those obtained for comparisons between Nino-3 .4 SST A and both precipitation and strean1 flow in Florida (Schmidt, N. unpublished data) The majority of basins with strong 109

PAGE 123

Appendix 1 (Continued) corre l a t ions were lo cated in the eastern portion of the Tampa Bay watershed. In this r egion, l and use includes broad areas devoted to agriculture and pasture. Furthermore, there is substantial l and application of sewage sludge in some areas (SWFWMD 1998 ; Table 1 ). Land use appears to be an important factor in relating changes i n water quality to t h e st r e ngth of the ENSO event. Seasonal Analysis The import ance of the E l Nino-Southern Oscillation to variabil i ty in factors such as rainfall and str eam flow in Florida, and elsewhere, varies with season (Schmidt et al. 2001 ). Therefo re given the significant relationship for most of the studied watersheds between fecal coliform levels and Nino 3.4 SSTA, we expanded our examination to assess the seasonal differences in fecal coliform levels between extreme ENSO phase s (E l Nino and La Nina; Table 2) and neutral conditions to better define the relationship between ENSO events and water quality. Basin-averaged fecal coliform value s were compa red between seasons and ENSO phase. Winter Feca l coliform levels were compared between neutral winters and both extreme El Nino and extreme La Nina winters (Fig. 2). For extreme El Nino winters there was an overall increase in feca l co li form levels as compared to neutral. With the exception of Delaney Creek, where the percent deviation was -20.4 (P = 0.087) the deviations in fecal coliform levels from neutral ranged between 7.6% and 18.7%. Howe ver, only at Rocky and Sweetwater Creeks were the fecal coliform levels significantly greater than neutral values (P < 0.1 0). During extreme La Nina winters, there was an overall decrease in fecal coliform levels as compared to neutral winters. With the exception of Rocky Creek where the percent deviation from neutral was 36.8 (P = 0.034), the deviations in fecal coliform levels from neutra l r anged betw e en -16.2% and -45 .5%. Feca l coliform level s at Bullfrog Creek, Delaney Creek, Hillsborough River and Sweetwater Creek were significantly below neutra l values (P < 0.05). Fecal coli form levels at Little Manatee River were significa ntly lower than neutral at P <0.1 0. Although Alafia River fecal coliform values showe d a negative deviation during extreme La Nina winters, the difference from neutral was not significant. Spring For extreme El N in o sp rin gs, the average fecal coliform levels were lower than tho se found during neutral springs (Fig. 2). Percent deviations from neutral ranged between 4.6 and -20.1; however, differences were not s i gnificant in any watershed. For the one case of an ext r eme La Nina spr i ng, fecal coliform levels in all basins were significantly below neutral values (P < 0.10). Deviations from neutral ran ged from 0.9% to -46.0%. 110

PAGE 124

Appe ndi x 1 (C ontinu e d ) Ppt. RF FC RF FC Ppt. RF FC Cl) Q) > Q) -::J 80% 60% 40% 20% 0 % -20 % c co -40 % +-----------il---.......:....::.-----l ..c -Cl) -60 % +------------1---------l Cl) Q) -80 % 2 Winter Spring Summer co 100o/o 0> Cl) c 0 -co -Cl) 0 Q) 80o/o 60 % 40o/o 20o/o g' 0 % -c Q) -20 % -40% -60 % -80% 100 % Winter Spring Summer Fall Figure 2. Chart shows the percenta g e of s tations (or basins [fecal coliforms]) with fecal coliform (FC) precipitation (Ppt.) or discharg e (RF) values that wer e greater than or less than that of neutral seasons for extreme E l Nifio (A) and extreme La Nifia (B ) events. Shad ing indicates significance levels. For spring, all La Nifia e v ents were used as there was only one e xtreme La Nifia event within the period of record. Significance results from Schmidt et al (2001) include statewide precipitation stations and streamflow for south central Florida (Charlotte Harbor and Tampa B ay) for 1950-1998. NS = not significant. 111

PAGE 125

Appendix 1 (Continued) Given the lack of data during extreme La Nifia spring events, deviations from neutral were also examined for all five spring La Nifia events For the more general La Nina spri n gs cases, le vels in all basins were l owe r than during neutral spring ( -2 .1% to -4 7.4% deviation). Ho wever, the differences were not sign ific ant l y different from neutral va lu es. Summer Fecal co l iform levels during extreme E l Nifio sum m ers showed a va ri ed response wi th neutral values (Fig. 2). None of the fecal coliform concentrations were significantly different than le vels found in neutral summers and percent deviations ranged from -13.7 to 11.8. For extreme La Nifia summers, percent deviations were consistently negative ( -0.1 % to -45.1 %). However, the differences from neutral were not significant. Fall Feca l coliform concentrations during extreme El Nino falls were generally greater than that found during neutral periods (Fig. 2). Although feca l coliform values at Bullfrog Creek, Delaney Creek and Little Manatee River were les s than that found during neutral fall the differences were not significant. The remaining stations all showed greater than neutral fecal coliform concentrations, with d ev i at ion s between 8.2% and 25.3%. The difference from neutral was only significant for Hillsborough River (25.3% deviation) and Rocky Creek (22.3% deviation). Patterns during extreme La Nifia falls showed both positive and negative deviations from neutral. The only drainage with s i gnificant differences from neutral was th e Little Manatee River, where levels were 17.5% below neutral values. DISCUSSION Extreme weather conditions including droughts and floods can dramatically affect communities at many levels. Direct effects may include crop damage, property damage, destruction of homes and loss of life. However, even moderate changes in climate can affect water resources in both quantity and qu a lity and thus indirectly affect public health Although research has demon strated regional-scale climate variabi lity during ENSO events, particularly relating to changes in temperature and precipitation (Ropelewski and Halpert 1986; Livezey et al. 1997; Gershunov and Barnett 1998a; Livezey and Smith 1999), it i s at the local le vel where economic and public health impacts are felt and where decisions regarding public policy must be made. Consequently, there is a need to better predict and understand the effects of climate variability at the local scale. To our knowledge this is the firs t study to exam i ne the ancillary or indirect effects of ENSO on fecal pollution in coastal waters as it relates t o recognized changes in precipitation and stream flow. 112

PAGE 126

Appendix I (Continued) The extreme El Nifio conditions observed in 1997 and early 1998 spurred investigations into the effects of climate variability on human health (NOAA 1999). During thi s time, the role of rainfall and stream flow in the introduction and transport of indicators of fecal pollution and human enteroviruses to coastal waters was demonstrated in southwest Florida (Lipp et al. in pre ss). Higher than average precipitation and subsequent river discharge were found in the winter months ( 1997 1998) along with lower water temperatures. Those patterns which are typical of El Nifio winters in Florida, provided a mechani s m for tran spo rt of enteric contaminants by run-off and discharge into coastal waters and incre ased s urvival due to lowe r salinity and temperature (Barbe and Francis 1995 ; Lipp et al. in press Wyer et al. 1995; Weiskel et al. 1996; Sun and Furbi s h 1997). Throughout Florida and in th e Tampa Bay area, tremendou s population growth in the last 20 years has been accompanied b y an increa se d volume of wastewater di s charged to coastal waters. Furthermore, non-point sources (from septic systems and stormwater run off) constitute a major cause of coa s tal pollution. Factors such as agricultural lands, land application of sewag e s lud ge, se ptic sys tems and wildlife contribute to high levels of coastal fecal pollution when transport mechani s m s, such as hi g h precipitation and river discharge are in operation The consequences of such pollution include closures of recr eational and shellfish propagating waters and potential exposure to human pathogens (Lipp and Rose 1 997). The predictable, or "normal," seasonal nature of rainfall in Florida has lead to specific water management strategies. The majority of rainfall in southern and central Florida occurs in the summer months while spring and fall are relativel y dry. In the southern part of the state w int e r s torm s account for l ess than 15% of the average annual precipitation (Nese and Grenci 1996). Similar to the larg e rsca le patterns noted for the so utheastern United States (Ropelewski and Halpert 1986) precipitation and consequently s tream flow along the south central Gulf coast of Florida are stron g l y related to ENSO (Schmidt et al. 2001) The seaso n al ENSO effects result s in precipitation patterns that are s uperimposed upon the normal seasonal tr e nds. Significant correlations between Nifio 3.4 SST A and fecal coliform levels at the majority of basins analy ze d demonstrate that Tampa Bay water quality also may be broadly linked to the ENSO s tate via teleconne c tion s that r es ult s in anomalous precipitation and river di sc harge. Significantly increased fecal pollution, relative to n e utral conditions, was most dramatic for extreme El Nifio winter months. ENSO development tends to peak in th e winter; consequently th e stronges t weather patterns are also noted during that time (Ropelewski and Halpert 1986) Significant increases in wintertime precipitation and discharge in Florida during extreme El Nifio event s (Zorn and Way len 1997 ; Schmidt et al. 200 1) may exacerbate condition s and result in g r ea ter than average level s of indicator organisms or introduction of enteric 113

PAGE 127

Appendix 1 (Continued) pathogens (Li pp et al. in press). Conversely, below average precipitation and discharge (Schmidt et al. 2001) lead to depre sse d fecal coliform levels during extreme La Nina winters A significant ENSO signal was also often noted in the fall, and the signal was generally variable, although not significant, in the sp ring The ENSO s ignal was ambiguous in the summer. In general, these observations follow patterns noted for both precipitation and river discharge in south central Florida (Schmidt et al. 2001; Fig. 2) The lack of s ignificant associations in all basins studied is most likel y rel ated to watershed characteristics, including the type of waste disposal, extent of development and industry, and the presence of dams and diversion s. This type of climate information combined with in depth analyses of watersheds will be useful in proactively tailoring water quality monitoring and control programs in the Tampa Bay region CONCLUSION In this historical assessment changes in water quality (using fecal coliform bacteria) in Tampa Bay, Florida were shown to vary with ENSO phases. We have previously shown a s ignificant relation ship between ENSO events and precipitation and di scharge (Schmidt et al. 2001) and betw ee n water quality and rainfall and di scharge in Florida (Lipp et al. in press) and now report that a direct associ ation between fecal pollution and ENSO events can be measured. De s pite an inherently noisy dat a set, significant trends between fecal coliform levels and Nino-3.4 SSTA were noted for the majority of the Tampa Bay watersheds we examined This study provides a baseline to initiate the development of water quality "fo reca st" models, ultimately using factors such as the El Nino-Southern Oscillation and other climatic variab les combined with land-u se characteristics to predict periods of poor water quality. This will further th e development of public policies for monitoring, assessing and managing important bays and coastlines for recreation, industry and fisheries. REFERENCES Barbe D. E. and J. C. Francis. 1995. An analysis of seasona l fecal coliform levels in the Tchefuncte River. Water Resources Bulletin 31: 141-146. C heckle y, W., L. D. Epstein R. H. Gilman, D. Figueroa R.I. Cama, J. A. Patz and R. E. Black. 2000. Effects of E l Nino and ambient temperature on hospital admissions for diarrhoeal disea s es in Peruvian children. Lancet 335:442-450. Colwell, R R 1996. Global climate and infectious diseases: The cholera paradigm. Science 274:2025 203 1 114

PAGE 128

Appendix 1 (Continued) Gershunov, A. and T. P. Barnett 1998a. ENSO influence on intraseasonal extreme ra i nfall and temperature frequencie s in the contiguous United States: Observations and model result s. Journal ofClimate 11:1575-1586. Gershunov A. and T. P. Barnett. 1998b. Interdecada l modulation of ENSO teleconnection s Bulletin o f the Ame rican M e t e orologi cal Societ y 79:2715-2725. Gueri, M. C Gonzale z and V Morin. 1986 The e f fect ofthe floods caused by El Nifio on health Di s ast e rs 10:118-124 Hoerling M.P., A. Kumar, and M. Zhong. 1997 El Niiio La Nifia and the nonlinearity of their teleconnections. Journal of Climate I 0: 1 769178 6. Lipp E K. and J. B Ros e 1997. The role of seafood in foodborne d i seases in the United State s of America R e vu e Sci e ntifiqu e e t Techni c ale Office lnternationaf e des Epizooties 16:620-640 Lipp E. K. R. Kurz, R Vinc e nt C. Rodri g uez-Palaci os, S. Farrah and 1. B Rose 2001 Seasona l v ariability and weath e r e f fects on microbial fecal pollution and enteric pathog e ns in a s ubtropical e s tuary. E s tuari es, in press Live zey, R E ., M Mas utani A. Leetmaa H. Rui M Ji and A. Kumar. 1997. Tel econnective response of the Pacific -North American region atmosphere to large central equatoria l Pacific SST anomalie s Journal of Climate 10: 1787 1820. Live z ey, R. E ., and T. M. Smith. 1999. Covariabi1ity of aspects of North American climate with global sea s urface temperature s on interarmual to interdecadal timesc a le s Journal ofC /imat e 12:289-302 Nes e J. and L. Grenci. 1 996. The climatology of Florida. In A World of Weath er: Fundamental s of Meteor o l ogy. Kendall /Hunt Publishing Co Dubuque, lA. National Oceanic and Atmos pheric Administration Office of Global Programs 1999 An Experime nt in th e Applications of Climate Fo rec asts: NOAA-OGP Activitie s Related to th e 1997-98 E l Nifio Ev e nt. NOAA, U.S. D epartment of Commerce. Washington D.C Noreen, E. W. 1989 C o mput e r intensive M e thod sfor Testin g Hypothe s es. John Wil ey and Sons 22 9 pp. O Shea, M. L. and R Field 1 992. Detection and di s infec ti on of pathogens in storm generated flows. Canadian Journal o f Mi c robi o logy 38:267-276. Paul J. H. J B. Ro s e S Jian g, X Z hou P. Cochran C. Kellogg J. B. Kang, D Griffin S. Farrah and J. Lukasik. 1997 Evidence for groundwater and surface marine water contamination by w as te dispo s al well s in the Florida Keys. Water Re s earch 3 1:1448-1454. 115

PAGE 129

Appendix 1 (Continued) Rees, R 1996. Under the weather: climate and disease, 1700-1900. History Today 46:35-41. Ropelewski, C. F., and M. S. Halpert. 1986. North American p recipit atio n and temperature patterns associated with the El Nifio Southern Oscillation (ENSO). Monthly Weather Review 114:2352-2362. Rose, J. B. 1997. Cl imat e forecasting, water resources a nd environmental health: Impact of El Nifio assoc iated with climatese n s itive diseases. Florida Journal of Environmental Health August:8-13 Schmidt, Nancy, E K. Lipp, J. B. Rose and M. E. Luther. 2001. ENSO influences on seasona l rainfall and river discharge in Florida. Journal of Climate 14:615-628. Southwest Florida Water Management District. 1998. Comprehensive Water she d Management Plan Tampa Bay / Anclote River. Vol. I. Assessment Draft. [Available from SWFWMD, 7601 U.S. Highway 301N, Tampa, FL 33638.] Sun, H., and D. J. Furbish. 1997. Annual precipitation and river discharges in Florida in response to El Nifio and La Nifia sea surface temperature anomalies. Journal of Hydrology 199:74-87. Weiskel, P. K., B. L. Howes, and G. R. H e ufelder. 1996. Coliform contamination of a coastal embayment: sources and transport pathways. Environmental Science and Technology 30:1872-1SS1. Wyer, M.D., D. Kay G. F. Jack so n, H. M. Dawson J. Yeo, and L. Tanguy. 1995. Indicator organism sources and coastal water quality: a catchment study on the I s l e of Jersey Journal of Applied Bacteriology 78:290-296. Zorn, M. R., and P.R. Waylen. 1997. Seasonal re sponse of mean monthly streamflow to El Nino/southern oscillation in north central Florida. Professional Geographer 49:51-62. 116

PAGE 130

ABOUT THE AUTHOR Nancy Schmidt received a B.A. Degree in Geology from Colgate University in 1985 and a M.S. in Geosciences from the University of Arizona in 1988. After receiving her M .S. she worked as an assistant science editor at the Office of Arid Land Studies and the Arizona Geologica l Survey in Tucson, Arizona, until she entered the Ph.D. program at the University of South Florida in 1992. While in the Ph.D. program in Marine Science Ms. Schmidt was very active in teaching and educationa l outreach. She worked part-time as a science interpreter for an ceo-tourism company on cruises to Baja, Mexico and to Alaska. She worked for nine summers with the Oceanography Camp for Girls, an educational outreach program for girls entering into high school in Pinellas County. Also, since 1999 she has been an adjunct instructor in Natural Sciences at St. Petersburg Junior College (now St. Petersburg College). She a l so has coa uth ored several p ubli cations in refereed scientific journals and has made seve ral presentations at national and international scientific meetings.


printinsert_linkshareget_appmore_horiz

Download Options

close
No images are available for this item.
Cite this item close

APA

Cras ut cursus ante, a fringilla nunc. Mauris lorem nunc, cursus sit amet enim ac, vehicula vestibulum mi. Mauris viverra nisl vel enim faucibus porta. Praesent sit amet ornare diam, non finibus nulla.

MLA

Cras efficitur magna et sapien varius, luctus ullamcorper dolor convallis. Orci varius natoque penatibus et magnis dis parturient montes, nascetur ridiculus mus. Fusce sit amet justo ut erat laoreet congue sed a ante.

CHICAGO

Phasellus ornare in augue eu imperdiet. Donec malesuada sapien ante, at vehicula orci tempor molestie. Proin vitae urna elit. Pellentesque vitae nisi et diam euismod malesuada aliquet non erat.

WIKIPEDIA

Nunc fringilla dolor ut dictum placerat. Proin ac neque rutrum, consectetur ligula id, laoreet ligula. Nulla lorem massa, consectetur vitae consequat in, lobortis at dolor. Nunc sed leo odio.