xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001994155
007 cr mnu|||uuuuu
008 090330s2008 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002435
Spencer, John M.
Comparing a low-volume piezometer to traditional wells in evaluating hydraulic lag caused by low-permeability sediments
h [electronic resource] /
by John M. Spencer.
[Tampa, Fla] :
b University of South Florida,
Title from PDF of title page.
Document formatted into pages; contains 33 pages.
Thesis (M.S.)--University of South Florida, 2008.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
ABSTRACT: Traditionally-constructed wells are commonly used to measure hydraulic head in all saturated systems, even in fine-grained sediments. Previous studies (Hvorslev 1951, Penman 1961) have shown that time lag in response to head changes between traditional wells and the surrounding fine-grained sediments can be a significant source of error. Time lag is caused by the time required for water to flow into or out of the well to reflect the appropriate change in head. A low-volume piezometer was constructed to measure changes in hydraulic head without requiring a change in fluid volume within the piezometer by directly measuring pore pressure in the surrounding sediments. The low-volume piezometer used a commercially-available pressure transducer that is hydraulically connected to the surrounding sediment by a porous-ceramic cylinder.The device is attached to a drive point that allows for quick insertion without creating excessive over-pressure so that equilibrium is achieved rapidly. The low-volume piezometer was inserted near traditionally-constructed wells in 3-4 m thick, saturated clay in west-central Florida. The low-volume piezometer was field tested to compare measured pore pressures with observed levels in traditionally-constructed wells. The comparison highlights any head difference between the two methods, and determines if there is a time lag between the two measurement methods and its magnitude. The low-volume piezometer was installed next to a traditionally-constructed well and heads in both wells were monitored for three months. Results show that the low-volume piezometer can take up to a month to reach equilibrium. Using Hvorslev's equations, traditionally-constructed wells have time lag of roughly 6 orders of magnitude greater than the low-volume piezometer.If this is correct, it could take up to 83,000 years for a traditionally-constructed well to reach equilibrium. However, when a trend analysis is performed on the hydrographs from the low-volume piezometer and the two traditional wells, the correlation coefficients are 0.95 and 0.96. The very strong correlation suggests that the low-volume piezometer and the traditional wells both respond similarly to changes in head. More field data need to be collected, but it appears that contrary to theory, time lag in traditionally-constructed wells may be negligible.
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
t USF Electronic Theses and Dissertations.
Comparing a Low-Volume Piezometer to Traditional Wells in Evaluating Hydraulic Lag Caused by Low-Permeability Sediments by John M. Spencer A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science Department of Geology College of Arts and Sciences University of South Florida Major Professor: Mark Stewart, Ph.D. Charles Connor, Ph.D. Mark Rains, Ph.D. Date of Approval: April 2, 2008 Keywords: Time-Lag, Diaphragm, Clay, Pore Pressure, Response Copyright 2007, John M. Spencer
i Table of Contents List of Tables ii List of Figures iii Abstract iv Chapter One Introduction 1 Description of Field Site 6 Statement of Problem 7 Chapter Two Methods 9 Plotting and Statistics 17 Chapter Three Results 18 Discussion 26 Conclusion 30 References 32
ii List of Tables Table 1 Table of Estimated Time Lags for Various Soil Types (Hvorslev 1951) 5
iii List of Figures Figure 1 A Cross-Section of a Traditionally-Constructed Observation Well 3 Figure 2 Well Nest Locations at Ft. Meade CSA 10 Figure 3 An Example of a Manometer (Moore and Gobdwin 1941) 11 Figure 4a First Low-Volume Piezometer Design 13 Figure 4b Second Low-Volume Piezometer Design 13 Figure 5a Hydrographs of Traditionally-Constructed Wells in the CSA 19 Figure 5b Hydrographs of the Surficial Aquifer Surrounding the CSA 20 Figure 6 Initial Insertion of the Orig inal Low-Volume Piezometer Design 22 Figure 7 Locations of Low-Volume Piezometer Head Measurements 23 Figure 8 Selected Recorded Measurements Using the Second Low-Volume Piezometer Design 24 Figure 9 Hydrographs of Low-Volume Pi ezometer and Traditionally-Constructed Wells within the saturated clay layer. 25 Figure 10 Scatter plot of th e Low-Volume Piezometer and Traditionally-Constructed Well 1W8. 27 Figure 11 Scatter plot of th e Low-Volume Piezometer and Traditionally-Constructed Well 1N15 28
iv Comparing a Low-Volume Piezometer to Trad itional Wells in Evaluating Hydraulic Lag Caused by Low-Permeability Sediments John Spencer ABSTRACT Traditionally-constructed wells are commonl y used to measure hydraulic head in all saturated systems, even in fine-grained sediments. Previous studies (Hvorslev 1951, Penman 1961) have shown that time lag in re sponse to head changes between traditional wells and the surrounding fine-g rained sediments can be a si gnificant source of error. Time lag is caused by the time required for water to flow into or out of the well to reflect the appropriate change in head. A low-volume piezometer was constructed to measure changes in hydraulic head without requiring a change in fluid volume within the piezometer by directly measuring pore pressure in the surrounding sediment s. The low-volume piezometer used a commercially-available pressure transducer that is hydraulically connected to the surrounding sediment by a porous-ceramic cylinde r. The device is attached to a drive point that allows for quick insertion wit hout creating excessive over-pressure so that equilibrium is achieved rapidly. The low-volume piezometer was inserted n ear traditionally-constructed wells in 3-4 m thick, saturated clay in west-central Florida. The low-volume piezometer was field tested to compare measured pore pressure s with observed levels in traditionally-
v constructed wells. The comparison highlig hts any head difference between the two methods, and determines if there is a time lag between the two measurement methods and its magnitude. The low-volume piezometer was installed ne xt to a traditiona lly-constructed well and heads in both wells were monitored for three months. Results show that the lowvolume piezometer can take up to a month to reach equilibrium. Using HvorslevÂ’s equations, traditionally-constructed wells have time lag of roughly 6 orders of magnitude greater than the low-vo lume piezometer. If this is correct, it could take up to 83,000 years for a traditionally-construc ted well to reach equilibrium. However, when a trend analysis is pe rformed on the hydrographs from the lowvolume piezometer and the two traditional well s, the correlation coefficients are 0.95 and 0.96. The very strong correlation suggests that the low-volume piezometer and the traditional wells both respond similarly to change s in head. More field data need to be collected, but it appears that contrary to theo ry, time lag in traditionally-constructed wells may be negligible.
1 Chapter 1 Introduction Low-permeability units have signific ant influence on the hydrology of aquifer systems, and are common in hydrogeologic sett ings. However, the hydraulic function of low permeability units is rarely examined. The leakance (KÂ’/bÂ’) of a confining unit is usually calculated from aquifer stress test data, but the unstresse d hydrology of the lowpermeability unit is seldom evaluated. Hydrau lic head is rarely measured within finegrained sediments, yet it is a critical piece of information when estimating horizontal and vertical groundwater gradient s and fluxes into or out of the system. Accurate head measurements are essential to understand the fl ow system within the unit of study and to determine its effect on surrounding systems. Hydraulic head within a low-permeability unit is one of the simplest and most useful aspects to study, however, it may be the hardest to accurately measure. Low-permeability units are often treated lik e more permeable units, in that, if they are to be monitored for any period of time, observation wells are inserted into the unit and the corresponding water levels are recorded. However, ma ny studies such as suggest that using traditional observ ation wells and standpipe piez ometers in very low hydraulic conductivity units causes larg e errors in measured heads [Hvorslev 1951, Penman 1961, Hanschke and Baird 2001]. Hvorslev (1951) is the most fre quently-referenced study focusing on the possibility that observed water levels in traditionally-constructed wells are directly
2 influenced by both the construction of th e traditional well and the surrounding waterbearing sediments. Traditional observati on wells and piezometers require moving volumes of water into or out of the well to change water levels within the well (Fig. 1.). For this reason, a large volume of water may n eed to pass into or out of the well and the surrounding well pack for water levels in th e well to accurately reflect changes in hydraulic head within the surround ing matrix. The time required for water to flow into or out of a device until a desired degree of pre ssure equalization with the surrounding matrix is attained is called the hydr ostatic time-lag (Hvorslev 1951). If the permeability of the matrix is sufficiently large, the time-lag is small and little lag effect is observed in the piezometer. However, silts and clays with permeabilities of 10-5 to 10-8 cm/sec can potentially cause significant time-lags in traditional observation wells and standpipe piezometers. Hvorslev derived several an alytical solutions to approximate time-lag error with several different types of sta ndard piezometer construction. For a standard, flush-bottom piezometer, k d T 11 (1) where T is basic time-lag (T), d is the diamet er of the piezometer (L), and k is hydraulic conductivity (L/T) (Hvorslev 1951). Hvorslev (1951) also derived equations on time-lag using diaphragm piezometers instead of standpipe piezometers. Diaphr agm piezometers can m easure the direct pore pressure from surrounding sediments instead of requiring a volume of water to flow into or out of the standpipe. The direct measurement of pore pres sure to measure changes in
3 Figure 1. A cross-section of a traditionallyconstructed well. The water level within a traditionally-constructed observation well is directly influenced by the flux of water moving into and out of the well screen.
4 head in the surrounding matrix makes a dia phragm piezometer ideal for low-permeability situations. Diaphragm piezometers need such a small volume change to accurately measure fluid pressure that they can ev en be potentially grou ted into a borehole (Mikkelsen, 2003). HvorslevÂ’s (1951) standard time-lag equation for a diaphragm piezometer is, k d h T 22 (2) where d is the diameter of the diaphragm (L), is the deflection of the diaphragm in the pressure cell (L), h is head (L) and k is hydraulic conductivity (L/T) (Hvorslev 1951). Hvorslev (1951) calculated time-lags for various soil types (Table 1). For the estimated time-lag for a so il with a permeability of 10-6 cm/s, a piezometer with a 5.08 cm casing and a flush bottom will have an approximate time-lag of 17 days, where a diaphragm piezometer in direct contact with th e soil will have an approximate time-lag of 4 seconds. Penman (1961) further investigated Hvor slevÂ’s time-lag analytical solutions by performing bench tests on several standpipe piezometers and el ectrical diaphragm piezometers. The experiment consisted of a testing apparatus much like a triaxial cell filled with London Clay with a permeability of ~10-8 cm/s. The cell was filled with clay and the piezometer was inserted into th e center. A pressure of about 0.703 kg/cm2 was then applied to the cell and the piezometer was monitored until the measured equilibrium was reached. The standpipe piezometer consis ted of a vertical cap illary tube 1mm bore connected to a ceramic filter. The standpi pe took 885 minutes to reach equilibrium, by far the longest of the methods tested. The el ectric diaphragm piezometer consisted of a
5 Table 1. Table of Estimated Time-lags for Various Soil Types (Hvorslev 1951). The Basic Time Lag, T column represents es timated time-lags for different head measurement methods in a soil with a coefficient of permeability of 10-6cm/s.
6 vibrating wire connected to a diaphragm with a ceramic filter at the end. The diaphragm piezometer was able detect a pressure change within 6 seconds of pressure alteration and reached equilibrium within 10 minutes. Most test data presented in literatu re written concerning time-lag error in traditionally-constructed wells have been obtained on a bench scale with little or no field data presented (Hvorslev 1951, Penman 1961, Hanschke and Baird 2001, Mikkelsen 2003). The objective of the current test is to conduct a field scale study of 4.5-6.0 m of saturated clay to determine if heads can be accurately measured in a fine-grained sediment system. Two methods of head measurement were chosen: traditionallyconstructed well nests and a newly designed lo w-volume piezometer. The field study also provided data on the effect of time-lag on the traditional wells and the low-volume piezometer. Description of Field Site Florida stratigraphy contains a Mioc ene-age, phosphate-and-clay-rich layer (Hawthorn Formation) that acts as a se mi-confining unit between a sandy, surficial aquifer and a karstic limestone aquifer be low (Miller 1986, Scott 1988). The Bone Valley Member of the Hawthorn Formation has been mined extensively for phosphatic ore since the beginning of the 20th century. In the mining pr ocess, the surficial sands above the Bone Valley Member are removed by large drag lines and set aside for land reclamation. After the surfic ial sands are removed, the drag lines excavate the underlying ore matrix and send it to a screening plant where the phos phatic ore can be separated from the sand and clay matrix. The matrix is made up of approxi mately 1/3 clay, 1/3 sand, and 1/3 pebble phosphate. During benefici ation, the matrix materials are put into
7 suspension. The residual clay slurry after beneficiation contains 3-5% solids and is pumped into clay settling areas (CSA) to dewater and consol idate. The overlying water is drained off and recycled into the phosphate beneficiation process. After 3-5 years the CSA develops a hard crust about 20 Â– 30 cm thick with saturate d clay underneath (Florida Institute of Phosphate Research, 2004.). CSAÂ’s make up about 40-50% of all reclaimed mining land in Florida. The selected field site is a CSA located a bout 8 km west of Ft. Meade, Florida. The CSA covers an area of 2.6 square kilome ters and the clay slu rry has a thickness of 4.5-6.0 m. The clay slurry was deposite d on top of unmined su rficial sand and is contained on all four sides by a berm made of the same sand. After approximately 20 years since formation, the CSA has formed a hummocky surface with a dry crust capable of supporting vegetative cover. The dried clay is predomin antly covered by cogan grass in the uplands and willows in the depressional wetlands. Statement of Problem Traditionally-constructed, 5.08 cm diam eter observation wells were installed to monitor head in the saturated clay. Afte r months of monitori ng water levels, the traditionally-constructed clay wells appeared to possibly be disconnected from each other and the surficial aquifer. Literature and pr evious studies seemed to support the idea that the apparent disconnection between the clay wells was caused by time-lag between the traditionally-constructed wells and the fine-grained sediment system. A low-volume piezometer was designed to measure direct por e pressures from the fine-grained sediment system to greatly reduce or ev en eliminate the possible timelag. The heads measured by the low-volume piezometer were compared with the heads measured in the traditionally-
8 constructed observation wells to test the hypothesis that well cons truction, or more specifically, well volume, can be a significan t source of time-lag when measuring heads in fine-grained sediment systems.
9 Chapter Two Methods Traditionally-constructed observation wells were installed in the CSA itself and in the surficial aquifer both dire ctly beneath and surrounding the CSA. The wells installed in the CSA were augered to 2.5 and 4.0 m below land surface at three different sites within the CSA. Two wells at each CSA site were installed at a depth of 12.0 and 30.5 m below land surface to monitor heads in the surfic ial aquifer, beneath the CSA. Nine other wells were installed to the nor th, east, and west of the CS A. The nine wells surrounding the CSA were installed at a depth of 4.0 m below land surface to monitor the surficial aquifer around the CSA. (Fig. 2.). Each well was constructed of 5.08 cm PV C with a 1.5 m slotted well screen. The screen was surrounded with a 15.25 cm diameter well pack consisting of medium sand. A drawback of traditionally-constructed wells is that the well screen is not directly in contact with the surrounding material due to the sand pack buffer. A device that can measure pore pressure and can be placed direc tly in contact with the clay pores would provide a much more accurate measurement of hydraulic heads than traditional wells and piezometers. A manometer was first considered (Fig 3.). A manometer is generally constructed out of a hollow st eel tube with a hard, porous membrane at one end, and a clear, flexible tube to view head changes at the other. The manometer is filled with water and driven into the soil to the determined depth. The porous membrane allows pore
10 Figure 2. Well Nest Locations at Ft Meade CSA. 1E, 1N, and 1W nests contain wells of 2.5 m, 4.0 m, 12.0 m, and 30.5 m BLS. 2E, 2N, and 2W nests contain wells of 7.75 m and 13.5 m BLS.
11 Figure 3. An example of a manometer (M oore and Gobdwin 1941). Section A is the soil probe, B is the manometer consisting of flexible tubing, capillary tubing, a mercury meniscus, and a manometer bulb. Section C is the probe point consisting of a porous ceramic cell, synthetic r ubber gaskets, and a hardened point.
12 pressures from the soil to be transmitted thr ough the membrane and cause the water level to rise or fall accordingly in the manomete r. The manometer doesnÂ’t need a large volume change of water to measure a head change in the soil, and it can be driven directly into the ground so no borehole with a surrounding sand pack is needed. However, the manometer is mainly designed to help iden tify rough vertical gr adients in a shallow groundwater system by recording the rise or fa ll of the water level in the viewing tube. All of the data would need to be manually recorded, hindering long term monitoring, and the actual pressure changes would be difficult to calculate. A manometer also is a large and heavy piece of equipment which would ma ke quick measurements at a field site difficult. Ideally, a device that has the qualities of both a diaphragm piezometer and a manometer would be ideal for head measurements It could be driven into the sediment with little disturbance and could quickly reach equilibrium with the surrounding pore pressure and give an accurate head measuremen t. After the needed data are collected, the device can be quickly removed and reused, unlike current diaphrag m piezometer designs. A low-volume piezometer was designed specifi cally for monitoring heads within fine sediments, using a commercially-available pressure transducer, and constructed to function like a manometer. The low-volume piezometer is ideal for field situations, lightweight, easily transportable, and can log measurements for as long as needed (Figs. 4a & 4b). The original prototype (Fig 4a) featured a hard, porous, ceramic cup threaded on the end of the pressure transducer. The pr essure transducer was fixed to 1.9 cm
13 Figure 4a. First low-volume Figure 4b. Second low-volume piezometer design. piezometer design.
14 galvanized pipe with a set screw. The transducer cable was run through the galvanized pipe to the surface where it could be used for real-time monitoring or for data download. Two setbacks occurred with the first prototype. Firs t, the rounded end of the ceramic cup allowed a large over-pressuri zation of fluid within the low-volume piezometer as the device was driven, making quick readings impossi ble. Overpressures from driving the device created heads 2-3 m above equilibrium heads. Equilibrium pressures were normally not reached until four to six hours after insertion. Second, the set screw could not withstand the stress of being pushed into the sediment, and multiple failures occurred. The set screw also required the transducer to be inserted and removed from the galvanized pipe before and after every field use. This made an accurate measurement of the correct length of the devi ce time consuming because it needed to be re-measured every time or the correct depth of the pressure meas urement would not be known. The second design incorporates several im provements (Fig 4b). The ceramic cup is replaced with a ceramic cylinder. A meta l tube runs from the end of the pressure transducer, through the ceramic cup, and is thr eaded into a drive point. The metal tube is slotted where it meets the ceramic cylinder to allow horizontal flow between the saturated ceramic cylinder and the pressure transducer. The pressure transducer can be directly threaded onto the end of the galvanized stee l pipe, creating a much stronger attachment and a standard length. The drive point pr events the large over pressurization from occurring which allows for more rapid equa lization and allows the piezometer to be driven into stiffer sediments.
15 Before deployment in the field, the lo w-volume piezometer was initially tested on a bench-scale water column. The ceramic cylinder was saturated and attached to the pressure transducer before being placed in to the water. The pressure transducer measured the water level at 1 second interval s and was monitored on a computer screen. The low-volume piezometer immediately recorded level changes as soon as the ceramic cylinder entered the water and continue d recording changes until the low-volume piezometer reached the bottom of the water column. There was no apparent time-lags. The low-volume piezometer was used to compare pore pressures within the clay to the water levels in the traditionally-c onstructed clay wells. The low-volume piezometer was placed at 2.5 and 4.0 m depths below land surface at the 1N, 1E, and 1W well clusters. The piezometer was then used to measure horizontal and vertical head gradients within the clay. The low-volume piezometer gathered point measurements at depths of 2 m, 3 m, and 4 m bls at three locations between the 1N well cluster and the northern berm of the CSA, as well as three lo cations between the 1E well clusters and the eastern edge of the CSA. The low-volume pi ezometer was left in the clay on average 5 hrs to reach what was then estimated to be 90% of equilibration. It was moved after 90% equilibration because Hvorslev (1951) found th at equilibration time before and after the 90% threshold were nearly equal. The lo w-volume piezometer was also left to run overnight when possible to record an accurate equilibrium. After the point measurements were taken near the clay well clusters, the low-volume piezometer was installed near the 1N8 well for approximately 3 months to mon itor seasonal changes in head within the CSA.
16 Clay samples were taken at various depths near the well clusters to calculate an average gravimetric moisture content and hydraulic conductivity from falling-head permeability tests. The clay samples were taken at two locations near the 1N cluster as well as two locations near the 1E cluster. At each location two samples were gathered at depths of 0.3 m and 1 m bls. Falling-head permeability tests were perfor med on the clay samples to estimate hydraulic conductivity (ASTM D5856-95). Each clay sample was placed into a rigid compaction-mold permeameter and connected to a falling head hydraulic system. The permeameter is designed to prevent swelling of the clay by using porous disks to contain the clay. The tests were run for over 24 hour s and results were collected when four values of hydraulic conductiv ity were equal (Cira 2008). Atterburg limits were obtained from clay samples along with gravimetric moisture content measurements to determine the physical properties of the clay If the saturated clay is acting as a liquid then the density of the clay would have to be added when calculating the total heads for the CSA; if the clay is acting as a plastic solid, it can be treated like any other water-be aring sediment when determining the heads. The plastic limit was determined using a standard ro lling method, while the liquid limit was determined by a cone penetration method (ASTM D4318-05). The bench tests were performed on samples taken at 10 cm, 38 cm and at 1.5 m below land surface. Each sample was tested five times and the mean of the results was used to ascertain the limits Cira (2008).
17Plotting and Statistics A trend analysis was utilized to quantify how closely the heads from traditionallyconstructed wells 1N15 and 1W8 correlate wi th the low-volume piezometer. Scatter plots were created for the recorded heads from June 2007, until September 2007 for each traditionally-constructed well versus the low-volume piezometer. A least-squares regression line with a high residual sum of squares would confirm a close correlation between the traditionally-constructed wells and the low-volume piezometer. If the wells and the piezometer do have a close correlation, then the probability that time-lag is affecting the traditional wells is low.
18 Chapter Three Results The clay is a light grey color with moisture content increasing with depth. In some areas of the CSA, a small percentage of silt or sand is present (10-30%), but otherwise the clay is homogeneous. The surface of the cl ay is hard with many fractures running through it. The hard, fractured layer begins at land surface and continues down to about 0.3 m below land surface. From 0.3 m to 0.6 m the fractures end and the clay becomes plastic and stiff. Below 0.6 m the lay becomes totally saturated, sticky, and very plastic with little shear strength. Little down pressure is needed when coring through the clay below 0.6 m. Generally the clay has a gravimetric wa ter content of 90% to 100% 30-38 cm below land surface, increasing to 120% to 150% at 0.5 m to 1.5 m below land surface. Because the clay is so homogeneous, it is assu med that the clay contains no more than a 150% gravimetric water content. Atterburg limits for the clay were 67% and 228% for the plastic limit and the liquid limit respectiv ely. Falling head permeability test results indicate an average hyd raulic conductivity of 10-8 cm/s, which is common for clay (Fetter 1980). During a year of monitoring, the water le vels within the traditionally-constructed clay wells had very different water levels when compared between themselves and also when compared with wells in the surrounding su rficial aquifer wells (F igs. 5a and 5b.). Wells in the clay open to the same depth with in 4-5 meters from each other, had water
19 37.00 38.00 39.00 40.00 41.00 42.00 43.00 44.00 45.00 46.00 47.00 Jun 06Aug 06Oct 06Jan 07Mar 07Jun 07 TimeHydraulic head (NGVD 1929m) 1E8 1N8 1N15a 1N15b 1W8 1W15 Figure 5a. Hydrographs of traditionally-constructed wells in the CSA.
20 Figure 5b. Hydrographs of traditionally-con structed wells in the Surficial aquifer surrounding the CSA. 35.00 36.00 37.00 38.00 39.00 40.00 41.00 42.00 43.00 44.00 45.00 Jun 06Jul 06Sep 06Oct 06Dec 06Feb 07Mar 07May 07 TimeHydraulic head (NGVD 1929 m) 1E40 1N40 1W40 3E15 3N35 3W15
21 levels that consistently differed by 1.5 to 2.0 m. Water levels in the shallow wells in the clay at a depth of 2.5 m, (1E8, 1W 8 and 1N8; Fig. 2.) rose by ~1.5 m from July Â’06 to Oct Â’06 while water levels in wells in the surficial aquifer outside the CSA rose 0.5 m during the same span of time. Water levels in the shallow wells in the clay also rose during February and March of 2007, a period when the wells in the surficial aquifer were at their lowest levels. The first low-volume piezometer design was inserted into the clay near the 1N cluster at 2.5 m below land surface. An overpressure occurred immediately after insertion and did not reach equilibrium until 3.5 hours later (Fig.6.). The low-volume piezometer was driven further to 4.0 m below land surface, insertion overpressures occurred again with equilibrium reached after 10 hours. The second piezometer design was inserted near the 1E well clusters, the 1W well clusters, and at various points near the 1N a nd 1E well clusters (Fig. 7.). The drive point for the new design was able to minimize the initial overpressure that occurred with the first design. The head rises on a logarithmi c curve until an apparent equilibration is reached (Fig. 8.). However, equilibrium time -lags are highly variable, with the required time for the low-volume piezometer to reach th e apex of the apparent equilibrium curve taking between 1 hr and 15 hrs, and full equili bration taking between 2 hrs and 46 hrs. Equilibration for the long term head monitoring event appears to reach a stable level after 27 days (Figure 9). While the pr essure in the low-volume piezometer is rising to reach equilibrium, the water levels in the traditionally-constructed wells remain steady. After the peak in the pressure of the low-vol ume piezometer, all three levels in the low-
22 Figure 6. Initial insertion for the origin al low-volume piezometer design. A large immediate overpressure was measured by the pressure transducer. The overpressure took between 3 and 15 hours to fully dissipate. 43.00 44.00 45.00 46.00 47.00 48.00 49.00 50.00 51.00 52.00 53.00 0.001.002.003.004.005.006.007.008.009.0010.00 Elapsed Time (hrs)Hydraulic head (NGVD 1929 m) 1N @ 2.5 m 1N @ 4.0m
23 Figure 7. Location at the Ft. Meade CSA where the low-volume piezometer was used to measure hydraulic heads in the clay.
24 Figure 8. Selected recorded measurements using the second low-volume piezometer design. Total equalization times as well as 90% equalization times varied greatly with no known cause. 0 0.2 0.4 0.6 0.8 1 1.2 1.4 0102030405060708090100 Elapsed TimeHydraulic Head (m) 1E 3.75 m 2007-05-21 30 m east of 1E @ 3m 2007-05-22 60 m east of 1E @ 3m 95 m east of 1E @ 3m 2007-05-25 1W @ 2.5 m 2007-05-27 1W @ 4 m 2007-05-29 30 m North of 1N @ 4 m 2001-05-30 60 m north of 1N @ 4 m 2007-05-31
25 Figure 9. Hydrographs of Low-Volume Piezometer and Traditionally-Constructed Wells within the saturated clay layer. 43 43.5 44 44.5 45 45.5 46 46.5 May 07Jun 07Jul 07Aug 07Sep 07 Time Elevation (m NGVD) Piezometer Traditional Well 1N15 Traditional Well 1W8
26 volume piezometer and traditionally-construc ted wells 1W8 and 1N15 steadily decrease in tandem until the end of the recorded period. The results from the trend analysis show a tare during a time when the traditionally-constructed well measured a rain event, but the low-volume piezometer did not. A trend line was plotted for the initial arm and the final arm of each scatter plot. The residual sum of the squares for the piezomete r vs. 1W8 is 0.992 and 0.998 with a p-value of 0.98. For the piezometer vs. 1N15, the R2 is 0.996 for the first arm and 0.99 for the second, the p-value for the scatter plot is 0.99 (Figs 10 and 11). Discussion One concern in measuring the heads within the clay is that the clay contains a very high water content (150%) and is loos ely consolidated only 30-50 cm below land surface. Atterburg limits were measured to de termine if the clay unit is a plastic solid or a viscous liquid. Atterburg limits for the clay are 67% for th e plastic limit, and 228% for the liquid limit. The estimated gravimetric wate r content for the clay is 150%, and is well below the liquid limit of 228%, which means that the clay can be treated as a saturated solid and not a viscous liquid. The permeability of 10-8 cm/s established in the falling-head permeability test was used with HvorslevÂ’s equations to determine an analytical estimation of time-lag for the traditionally-constructed observation wells (e q. 1) and an estimated time-lag for the low-volume piezometer (eq. 2). The estimated time-lag for a 5.08 cm standard well in direct contact with the surroundi ng soil with a permeability of 10-8 cm/s is 1679 days. The estimated time-lag for a diaphragm piezometer with a diameter of 1.9 cm and in contact with the same soil is 2.15 min.
27 Figure 10. A scatter plot of the LowVolume Piezometer and the TraditionallyConstructed Well 1W8. The graph shows a very strong correlation between the two. y = 1.3132x 15.164 R2 = 0.992 y = 1.3318x 16.174 R2 = 0.9975 44.30 44.40 44.50 44.60 44.70 44.80 44.90 45.00 45.10 45.3045.4045.5045.6045.7045.8045.9046.00 Piezometer1W8
28 Figure 11. A scatter plot of the Low-Vo lume Piezometer and the TraditionallyConstructed Well 1N15. The graph shows a v ery strong correlation between the two also. y = 1.0776x 5.2096 R2 = 0.9959 y = 1.4143x 20.742 R2 = 0.9905 43.50 43.60 43.70 43.80 43.90 44.00 44.10 44.20 44.30 44.40 45.2045.3045.4045.5045.6045.7045.8045.9046.00 Piezometer1N15
29 The calculated time-lag for a low-volume pi ezometer, as well as the results of laboratory bench test for low-volume piezome ter response times, differed greatly from the field response times observed at the CSA. The low-volume piezometer is calculated to have a time-lag of 2.15 min using Hvor slevÂ’s equation and Penman observed a timelag of ~ 10 min for total equalization in a bench test using homogeneous clay of 10-8 cm/s. The initial field test using the second low-volume piezometer design indicated equilibrium times varied anywhere between 1.5 hours and 2 days. Each curve seemed to reach some sort of equilibrium, yet the l ong term field test indicates that the short equilibrium time may not be a true equilibrium The long term test shows that it might take up to 1 month before a piezometer that is directly measuring pore pressures can reach equilibrium. If traditionally-constructed wells have a time-lag error of roughly 6 orders of magnitude greater than the lowvolume piezometer it could take up to 83,000 years for a traditionally-constructed well to reach equilibrium. However, when a trend analysis is performed on the hydrographs fr om the low-volume piezometer, traditional wells 1N15, and 1W8, the correlation coeffici ents are 0.95 and 0.96, respectively. The very strong correlation shows that the lowvolume piezometer and the traditional wells have very similar responses to changes in head in the clay. One hypothesis to explain the fast reacti on times of both the wells and the lowvolume piezometer to head change s within the wells is that the clay could be a dual flow system. Fractures were observed in the cl ay up to 1 m below land surface. Flow in a fracture system could allow the levels in the wells to respond rapidly to changes in head
30 in the clay. The clay below 1 m however, is saturated and extremely plastic. Flow in open fractures beneath 1 m seems improbable. Another hypothesis is that th e current estimation of timelag is wrong. If the lowvolume piezometer was seated correctly into the clay, and the wells were installed correctly, then the observed heads are true. If the field data ar e accurate, the error in head measurements caused by time-lag of recorded water levels to heads within traditionallyconstructed wells in fine-grain ed sediments could be small. Conclusion Traditionally-constructed wells were installe d in a fine-grained system to measure the hydraulic heads within the system and estimate groundwater movement. After months of monitoring, the water levels within adjacent wells appeared to be disconnected, and a solution was sought. An alytical estimates from Hvorslev (1951) show that hydrologic time-lag can be a si gnificant source of error by causing head changes in the sediments to be delayed by up to four years before being reflected in the monitoring wells. A low-volume piezometer was designed to directly measure pore pressures from the clay and effectively eliminat e the time-lag error derived by Hvorslev. Initially, the low-volume piezometer a ppeared to reach equilibrium levels between 4 hours and 2 days. Point measur ements between the low-volume piezometer and the traditionally-constructed wells also indicated a difference in measured heads by an average of 0.6 m. However, long te rm monitoring of heads by the low-volume piezometer and a traditionally-c onstructed well suggest that the time required for the low-
31 volume piezometer to reach equilibrium is much longer than first anticipated. Also, the observed heads from both the low-volume piezometer and the traditionally-constructed wells correlate so closely that they can c onsidered almost identical. The comparison between the low-volume piezometer and two tr aditionally-constructed wells point to the possibility that time-lag in the response of tr aditional wells to head changes is small, even in systems with extremelylow hydraulic conductivities. Time-lag is a largely accepted theory in the hydrogeologic community. The theory is often referenced in literature, especi ally in reference to fine-grained sediments; however, there appears to be no conclusive fi eld data that confirms the idea. Further research at the present field site can quantify how head measurements in traditionallyconstructed wells are affected by fine-grained systems. Accu rate field data will be the key in deciding whether time-lag is a real so urce of error in trad itionally-constructed wells, or if it is not significant.
32 References Brown, Mark T., Carstenn, Susan M., Baker, John, Bukata, B.J., Gysan, Tim, Jackson, Kristina, Chinners Reiss, Kelly, and Sloan, Mellini. Successional Development of Forested Wetlands on Reclaimed Phosphat e Mined Lands in Florida. Volumes 1 & 2. University of Florida, Cent er for Wetlands; August 2002 Cira, Aidee, Results from Doctoral Di ssertation Research, Un iversity of South Florida, College of Engineering, 2008 Florida Institute of Phosphate Research Home Page. 1855 W. Main St. Bartow, FL 33830, 2004 Hanschke, Thomas, and Baird, A.J., Time-Lag Errors Associated with the use of Simple Standpipe Piez ometers in Wetland Soils The Society of Wetland Scientists (Vol. 21, No. 3, ) pp. 412-421; 2001 Hvorslev, M.J., Time-lag and Soil Permeability in Groundwater Observations. Bulletin No. 36, U.S. Waterways Experi mentation Station, Vicksburg, MS; 1951 Lewelling, B.R. and Wylie, R.W. Hydrology and Water Quality of Unmined and Reclaimed Basins in Phosphate-Mini ng Areas, West-Central Florida. Tallahassee, Fl; 1993 Mikkelsen, P.E., and Green, G.E., Piezometers in Fully Grouted Boreholes Symposium on Field Measurements in Geomechanics, Oslo, Norway; 2003 Miller, J.A., Hydrogeologic Framework of the U pper Floridan Aquifer system in Florida, and parts of Georgia, Alabama, and South Carolina, U.S. Geological Survey Professional Paper 1403-B; 1986 Moore, R.E., and Gobdwin, K.R., Hydraulic Head Measurements in Soils with High Water Tables American Society of Agricultural Engineers (vol.22 no.7), St. Joseph, MI; 1941 Reigner, Walter R. and Winkler III, Cornelis. Reclaimed Phosphate Clay Settling Area Investigation: Hydrologic Model Calibration and Ultimate Clay Elevation Prediction. Final Report. BCI Engineers & Scientists, Inc., Lakeland, Florida; August 2001
33 Ross, Mark A. et al. Hydrogeology of Clay Settling Areas. University of South Florida Tampa, FL; 2003 Scott, T.M., Lithostratigraphy of the Hawt horn Group (Miocene) of Florida, Florida Geological Survey Bulletin No. 59, 1988