GEOLOGIC, HYDROLOGIC, AND WATER-CHEMISTRY DATA FOR A MULTI-AQUIFER SYSTEM IN COASTAL PLAIN SEDIMENTS NEAR GIRARD, BURKE COUNTY, GEORGIA, 1992-95 by David C. Leeth, William F. Falls, Lucy E. Edwards, Norman 0. Frederiksen, and R. Farley Fleming U.S. Geological Survey Prepared in cooperation with the U.S DEPARTMENT OF THE INTERIOR U.S. GEOLOGICAL SURVEY and the U.S. DEPARTMENT OF ENERGY INFORMATION CIRCULAR 100 Any use oftrade, product orfirm names is for descriptive purposes and does not imply endorsement by either the U.S. Government or the Georgia Department ofNatural Resources. Cover photograph: Georgia Forestry Commission fire tower near Girard, Burke County, Georgia. Photograph by Alan M. Cressler, U.S. Geological Survey. GEOLOGIC, HYDROLOGIC, AND WATER-CHEMISTRY DATA FOR A MULTI-AQUIFER SYSTEM IN COASTAL PLAIN SEDIMENTS NEAR GIRARD, BURKE COUNTY, GEORGIA, 1992-95 GEORGIA DEPARTMENT OF NATURAL RESOURCES Joe E. Tanner, Commissioner ENVIRONMENTAL PROTECTION DIVISION Harold F. Reheis, Director GEORGIA GEOLOGIC SURVEY William H. McLemore, State Geologist Prepared in cooperation with the U.S. DEPARTMENT OF THE INTERIOR U.S. GEOLOGICAL SURVEY and the U.S. DEPARTMENT OF ENERGY Atlanta, Georgia 1996 INFORMATION CIRCULAR 100 CONTENTS Page Abstract .. ......... . ...... ... . . ... .. ....... . ............. , ... .. . . .. .. .............. .. ...... . 1 Introduction . , .... . . . ... . ....... . .... . ...... . ...... .. ... . .... . ... .. .......... . ........ . .. . . . 2 Purpose and scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Description of study area................................................. . .. .. . ....... . . . . . 4 Well-numbering system ................................................... ... .. .. . . ... .. . . . 4 Acknowledgments . ...... .. .. . . ... . . ........ ... ... . .... .... . . . .. .. . . .. .. .. .. ............. . 4 Well construction and coring ................................... . .............. ... ..... . . ....... . 4 Geologic data ................................................ . ............. ... .. ............ . 5 Lithology .............................................................. ........... . . .. . . 5 Micropaleontology ..................................................... .......... . . . .... . 7 Hydrologic data ...... .. ... . ......... . ..... . ............ . . . .. .. ... . ... . ..... ........ . ........ . 7 Hydrogeologic units ......... . ........................... . .............. .. . . .. ...... . .. . . . 7 Hydraulic properties ..................................................... . . ......... . .. . . . 8 Hydraulic conductivity estimated from aquifer tests ....................... ...... .... .. . ... . 8 Hydraulic conductivity of core samples ............................... ... ... .. . ........ 10 Hydraulic conductivity estimated from borehole resistivity ............. ....... . ... . ........ 11 Ground-water levels . .. ......... . ........................... . ........ .. . .. . . . . . ....... . .. 12 Water-quality data .......................................................... .......... . ...... 14 Summary and conclusions ..................................................... ... . ..... ..... .. 16 Selected references .......................................................... .......... . . . .. . . 18 Appendix-lithologic description of Girard hydrogeologic test site, Burke County, Georgia .... . ... .. .. . .... 23 Plate ILLUSTRATIONS [Plate is in pocket in back of report] l. Chart showing hydrogeologic units, lithology, geophysical logs, well construction, and hydraulic characteristics for wells at the Girard hydrogeologic test site, Burke County, Georgia Figure 1. Map showing locations of Savannah River Site, Girard hydrogeologic test site, and test wells .... 3 2. Chart showing generalized correlation of units of Late Cretaceous through Eocene age in the southeastern United States ................................................. .... 6 3. Chart showing correlation of hydrogeologic units in the vicinity of the Girard hydrogeologic test site, Girard, Georgia .................................................. ... . 9 4. Hydrograph showing daily mean ground-water levels in Girard test wells TW-1, TW-2, and TW-3; daily mean stream stage, Savannah River at Jackson, South Carolina; and precipitation at Waynesboro, Georgia ........................................ ... 13 5. Trilinear diagram showing percentage composition of major ionic constituents in ground water at the Girard hydrogeologic test site, January 23-27, 1995 ................... .. . 15 TABLES Table 1. Summary of aquifer-test analyses at the Girard hydrogeologic test site, January 1994 and December 1995 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I0 2. Estimates of horizontal hydraulic conductivity determined from formation resistivity logs at the Girard hydrogeologic test site ...................................... ... .. . 11 3. Statistical comparison of precipitation and stream-stage data to ground-water levels at the Girard hydrogeologic test site . .. ............ . ....................... ....... 14 4. Physical and chemical characteristics of ground-water samples collected at the Girard hydrogeologic test site ................................................. ...... 16 iii CONVERSION FACTORS, ACRONYMS AND DEFINITIONS, AND VERTICAL DATUM CONVERSION FACTORS Multiply by to obtain Length inch (in.) foot (ft) mile (mi) 25.4 0.3048 1.609 millimeter meter kilometer square mile (mi2) Area 2.590 square kilometer gallon (gal) Volume 3.785 liter gallon per minute (gal/min) 0.06309 liter per second part per million picocurie per liter (pCi/L) Concentration 1,000 3.19 milligrams per liter (mg/L) micrograms per liter (J..tg/L) tritium unit Specific conductance micromho per centimeter at 25 o Celsius (J..tmhos/cm at 25 o C) 1 microsiemens per centimeter at 25 o Celsius (J..tS/cm at 25 o C) Temperature in degrees Fahrenheit ( F) can be converted to degrees Celsius (0 C) as follows: = o C 5/9 CO F -32) iv ACRONYMS AND ABBREVIATIONS DIC DNR DOE EPA EPD SRCC SRS TOC USGS Dissolved inorganic carbon Georgia Department of Natural Resources U.S. Department of Energy U.S. Environmental Protection Agency Georgia Environmental Protection Division Spearman's rank correlation coefficient Savannah River Site Total organic carbon U.S. Geological Survey VERTICAL DATUM Sea level: In this report "sea level" refers to the National Geodetic Vertical Datum of 1929--a geodetic datum derived from a general adjustment of the first-order level nets of the United States and Canada, formerly called "Sea Level Datum of 1929" v GEOLOGIC, HYDROLOGIC, AND WATER-QUALITY DATA FOR A MULTI-AQUIFER SYSTEM IN COASTAL PLAIN SEDIMENTS NEAR GIRARD, BURKE COUNTY, GEORGIA, 1992-95 By David C. Leeth 11, William F. Falls 11, Lucy E. Edwards 11, Norman 0. Frederiksen 11, and R. Farley Fleming 11 ABSTRACT The Girard hydrogeologic test site (Girard site), in southeastern Burke County, Ga., was constructed during 1992-95 to better characterize the geologic, hydrogeologic, and water-quality characteristics of a multi-aquifer system in Coastal Plain sediments. Data presented herein include the depth, thickness, geologic properties, hydrologic properties, paleontology, and water quality of the Coastal Plain aquifers at the Girard site. During March and April 1992, continuous, 2-inch diameter core was collected from land surface to a total depth of 1,385 feet using the wire-line coring method. The core penetrated Coastal Plain sediments and bottomed in pre-Cretaceous basement rock. Paleontologic data provide geologic age, and when combined with lithologic data indicate the environment of deposition for several geologic units at the Girard site. Twenty-five samples were examined for dinoflagellates, pollen, benthic foraminifera, and calcareous nanofossils. Eleven of the samples yielded age-diagnostic assemblages, ranging from middle Eocene to the Late Cretaceous. Water-bearing units at the Girard site were related to previously defined hydrogeologic units by comparing borehole data collected at this site to interpreted borehole data from nearby sites. This comparison indicates that several equivalent hydrogeologic units are present at the Girard site. In descending order, these are the 11U.S. Geological Survey. Upper Three Runs aquifer, Gordon aquifer, Millers Pond aquifer, the upper and lower Dublin aquifers, and the upper and lower Midville aquifers. Selected core samples were analyzed to determine vertical hydraulic conductivity and porosity. Laboratory analyses indicate that the vertical hydraulic conductivity of confining units is less than 2 X 1o-9 feet per day. Three test wells were completed at the Girard site. Clemson University personnel conducted aquifer tests in two of the wells to determine transmissivity, horizontal hydraulic conductivity, and to detect any interaquifer leakage. Horizontal hydraulic conductivity estimated from these tests indicates that the lower Dublin aquifer is much more productive than the lower Midville aquifer and that both aquifers are confined at the Girard site. Horizontal hydraulic conductivities for several of the aquifers at the Girard site were estimated by applying logarithmic regression models to borehole resistivity data. These values are comparable to those derived fromaquifer pumping tests and are within the range of error reported for the method. Average hydraulic conductivity values from TW-2 ranged from 26.93 feet per day using the logarithmic regression model to 177 feet per day using TW-2 aquifer pumping tests. In addition, average horizontal hydraulic conductivity values for TW-3 ranged from 24.74 feet per day using the logarithmic regression model to a value of 8.9 feet per day using TW-3 aquifer pumping tests. Continuous water-level recorders were installed in each test well to monitor water-level fluctuations and trends. Water-level data also were used to determine the vertical distribution of hydraulic head in the waterbearing units. A statistical comparison of ground-water levels, stream stage, and precipitation was performed using Spearman's rank correlation coefficient. Based on this statistical analysis, two significant correlations are apparent. A mass loading water-level response in the lower Dublin aquifer occurs in response to recharge of the Upper Three Runs aquifer. In addition, some interaquifer leakage from pumping of the lower Dublin aquifer may be affecting the water level in the Upper Three Runs aquifer. Water samples were collected from two of the test wells to determine the physical and chemical characteristics ofwater from the screened water-bearing zones. Trace-element chemistry shows significant differences in water quality between the lower Midville and lower Dublin aquifers. Of the 11 trace elements tested, barium, iron, and strontium have disparate values between the two aquifers. Values for these elements are significantly higher in the lower Dublin aquifer than in the lower Midville aquifer; the value for iron (1,600 micrograms per liter) exceeds the U.S. Environmental Protection Agency drinking-water standards by 1,300 micrograms per liter. This high value for iron would limit the usefulness of the lower Dublin aquifer to agricultural purposes unless extensive pretreatment is utilized. INTRODUCTION The U.S. Department of Energy (DOE), Savannah River Site (SRS) has manufactured nuclear materials for the National defense since the early 1950's. A variety of hazardous materials including, radionuclides, volatile organic compounds, and heavy metals, are either disposed of or stored at several locations at the SRS. Contamination of ground water has been detected at several locations within the SRS. Concern has been raised by State of Georgia officials over the possible migration of ground water contaminated with hazardous materials through aquifers underlying the Savannah River into Georgia (trans-river flow). The U.S. Geological Survey (USGS), in cooperation with the DOE and the Georgia Department ofNatural Resources (DNR), Environmental Protection Division (EPD), Georgia Geologic Survey (GGS), is conducting a study to describe ground-water flow and ground-water quality near the Savannah River. The overall objectives of this study are to identify groundwater flow paths, quantitatively describe ground-water flow, and evaluate stream-aquifer relations between the Savannah River and underlying aquifers. The geologic, hydrologic, and water-quality characteristics of aquifers and confining units are being characterized to support the analysis. Accordingly, a test-drilling program was initiated to establish the geologic, hydrologic, and water-quality characteristics of Coastal Plain sediments near the Savannah River (fig. 1). Test-well cluster sites are being constructed in major aquifers at several locations along the Savannah River in Georgia (fig. 1). Purpose and Scope This report presents geologic, hydrologic, and water-quality data collected at the Girard hydrogeologic test site (Girard site) in southeastern Burke County, Ga. Data include the depth, thickness, geologic properties, hydrologic properties, paleontology, and water chemistry of the Coastal Plain aquifers at the site. These data, presented in graphs, tables, and diagrams, provide correlation of stratigraphy and ground-water flowsystem characteristics. Data collected at the site are on file at the U.S. Geological Survey, Atlanta, Ga. Test-drilling activities were designed to obtain various data. These included: recovery of core samples for geologic testing and paleontologic examination; acquisition of geophysical logs to aid in the description and definition of the lithology and physical characteristics of the sediments penetrated; collection of water-quality samples from discrete water-bearing zones; measurement of water levels in selected water-bearing intervals; and determination of hydraulic properties of water-bearing zones and confining units. This is the third in a series of reports that present results of the project test-drilling program. The previous reports described the results of test drilling at the Millers Pond site in northern Burke County, Ga. (Clarke and others, 1994) and at the Millhaven Plantation site in northeastern Screven County, Ga. (Clarke and others, 1996). 2 8230' 82 8130' 3330' fly d ~f~P'-Ri~ch~mo \6 / ' Augusta n+d \ g~_ rvv ~ / '/'-<,\ ~ -Miller'sPo~d \ __ ~:x--- - ,/ / Thompson Oak - 1\ __\ tf f 'v Site 1 \ Je erson ~Waynesboro+ \. ,\ 33 I \ <2,1 Burke GirardSite 3230' 0 10 20 1,,,,,1 'I I 0 10 20 EXPLANATION TEST SITE 2 TW- TEST WELL AND IDENTIFI- CATION NUMBER .... STREAM-GAGING STATION AT JACKSON, SOUTH "0 eell CAROLINA PRECIPITATION- t:: 0 MONITORING STATION . " .. . . @r ] Trees Field Trees 0 10 20 30 40 50 FEET ~IW'~ I I 'I I It P Pi 10 15 MElERS TW-2 TW-1 TW-3 Corehole Figure 1. Locations of Savannah River Site, Girard hydrogeologic test site, and test wells . 3 Description of Study Area Sediments at the Girard test site in the Coastal Plain physiographic province in Georgia consist of alternating layers of sand, silt, clay, and lesser amounts of limestone that dip southeastward forming a series of aquifers and confining units. Although data in South Carolina are plentiful, limited geologic, hydrologic, and waterquality data are available in Georgia to determine the characteristics of these aquifers and confining units adjacent to the Savannah River. The Girard hydrogeologic test site is in southeastern Burke County, about 22 miles (mi) south of Augusta, Ga.; about 12 mi southeast of Waynesboro, Ga.; and about 3.8 mi west of the Savannah River on the site of a Georgia Forestry Commission fire tower (fig. 1). The site is about 250ft above sea level with altitudes varying less than one foot between the wells. Well-Numbering System Each of the test wells at the Girard site were numbered according to the order of drilling; that is, test well 1 (TW-1) was the first well completed, TW-2 was the second, and so on. In addition to these project well numbers, wells in Georgia also are numbered according to a system based on the USGS index of topographic maps. Each 7 1/2-minute topographic quadrangle in the State has been given a number and letter designation beginning at the southwest comer ofthe State. Numbers increase eastward and letters increase alphabetically northward. Quadrangles in the northern part of the area are designated by double letters. The letters "I", "II", "0", and "00" are omitted. Wells inventoried in each quadrangle are numbered consecutively beginning ~ith 1. Thus, the 17th well numbered on the 30Z quadrangle is designated 30ZO17. Acknowledgments The authors gratefully acknowledge the cooperation and assistance of the Georgia Forestry Commission, for providing access to the Girard fire tower property. Thanks also is extended to David S. Snipes, Rex Hodges and Peter G. Luetkehans, Clemson University, for conducting and analyzing aquifer tests at the site. WELL CONSTRUCTION AND CORING During March and April 1992, a corehole was completed at the Girard site penetrating Coastal Plain sediments and bottoming in pre-Cretaceous basement rock. During drilling of the corehole, surface casing was emplaced to a depth of 140 ft to alleviate problems with caving sand zones. Continuous, 2-in. diameter core samples were collected from land surface to a depth of 1,385 ft using the wire-line coring method. The core samples were used to determine lithology, grain size, sand/clay ratio, and depositional environment. The paleontology of selected core samples provides age control for the time of deposition. Borehole geophysical logs were collected for the total depth ofthe corehole to aid correlation of units and determine waterbearing properties of the sediments. Geophysical logs from the corehole are shown on plate 1. After geophysical logging, the corehole was capped. To characterize the vertical distribution ofhydraulic head and water chemistry of Coastal Plain sediments at the Girard site, test wells were installed at three waterbearing intervals at depths ranging from 72 to 1,122 ft, using standard mud-rotary drilling practices. Screened intervals for each well were positioned in layers having relatively high sand content overlain and underlain by clay beds of relatively low permeability, as was determined by examination of core and geophysical logs. Screened intervals were made as large as possible (up to a maximum of 50ft) to allow adequate pumping rates during aquifer tests. Well-construction diagrams are shown along with geophysical logs and lithology on plate 1. Test well TW-1, completed in March 1992, originally was constructed as a water-supply well for the corehole. The well is constructed of threaded 2-in. diameter polyvinyl chloride (PVC) casing from a depth of0.5 to 48ft, coupled with a 24-ft section of2-in. diameter, 0.012-in. slotted PVC screen set from 48 to 72 ft. A 5-ft section of PVC was added below the screen as a sediment trap, making the total completed well depth 77ft. To alleviate problems with caving sand, a 3-in. diameter PVC surface casing was installed to a depth of 48 ft. The well was completed by allowing the formation sand to collapse around the screen and grout was emplaced from the top of the screened interval to approximately 2 ft below land surface (plate 1). The two other wells (TW-2, TW-3) were constructed using a double-string technique by telescoping a smaller diameter screen and blank casing into a larger diameter outer casing, that was previously grouted into place. A neoprene packer was used to form a seal between the inner and outer casing. Test wells TW-2 and TW-3 were developed using air surging and jetting techniques. Development in each of the wells continued until the return water was free of drilling mud and sand. 4 Test well TW-2, completed in May 1994, was constructed of a 6-in. diameter threaded-and-coupled low-carbon steel outer casing, emplaced from land surface to a depth of730 ft. Using a neoprene rubber "K" packer as a seal, a 4-in. diameter string of lowcarbon steel inner casing is telescoped into the outer casing from a depth of720 ft to the top of the screen at 743ft. Thirty feet of 4-in. diameter, 0.010-in., wire wrapped, continuous slot, type 304 stainless-steel screen continues the inner string down to a depth of 773 ft. The bottom of the well string is completed with a nominal 10 ft, low-carbon steel sump bringing the total depth of the well to about 784 ft (plate 1). Test well TW-3, completed in December 1995, was constructed of a 6-in. diameter threaded-and-coupled, low-carbon steel outer casing, emplaced from land surface to a depth of 1,000 ft. Using a neoprene rubber "K" packer a~ a seal, a 4-in. diameter string of lowcarbon steel inner casing is telescoped into the outer casing from a depth of942 ft to the top of the screen at 1,070 ft. Fifty-two feet of4-in. diameter, 0.0 16-in., wire wrapped, continuous slot, type 304 stainless-steel screen continues the inner string down to a depth of 1,122 ft bringing the total depth of the well to about 1,142 ft (plate 1). GEOLOGIC DATA An understanding of the lithology and micropaleontology of the Girard hydrogeologic test site is necessary to delineate the age and depositional environments of the sediments. Sediments underlying Burke County range in age from Mesozoic to Holocene and consist of units of sand, silt, clay, and minor amounts of limestone. Lithologic and paleontologic evidence from the Girard fire tower core suggests that at least 13 distinct lithologic units (fig. 2) are present in the vicinity of the site. These lithologic units have a total thickness of about 1,385 ft (plate 1). A generalized correlation of units of Late Cretaceous through Eocene age in the southeastern United States is shown in figure 2. These sediments unconformably overlie consolidated red beds of Mesozoic age (Chowns and Williams, 1983). Lithology The sediments underlying the Girard site consist predominantly of deltaic and marine sand and clay. The lithology and geophysical characteristics of sediments at the Girard site are shown graphically on plate 1. A detailed description of the lithology, grain size and sorting, induration, texture, contact relations, and physical and biogenic sedimentary structure of core collected at the Girard site is listed in the Appendix. Textural classification of siliciclastic sediments shown in the appendix was adapted from a standard grain-size scale (Wentworth, 1922) and includes: clay (less than 0.020 millimeters (mm)), silt (0.020 to 0.065 mm), sand (0.065 to 2.00 mm), granules (2.00 to 4.00 mm), and pebbles (4.00 to 64.00 mm). Sand-size grains are further subdivided into five classes: very fine (0.065 to 0.125 mm), fine (0.125 to 0.250 mm), medium (0.250 to 0.500 mm), coarse (0.500 to 1.000 mm), and very coarse (1.000 to 2.000 mm). Grain-size distribution and sorting of siliciclastic framework grains were based on visual classification of sand grains, granules, and pebbles. In this report, granules and pebbles are considered to be grain-size classes in estimates of sorting. Categories of sorting are based on the number of grain-size classes observed in a sediment sample and are herein defined as: well sorted (one grain-size class), moderately sorted (two grain-size classes), poorly sorted (three or four grain-size classes), and very-poorl; sorted (five or more grain-size classes). The size of heavy minerals, mica grains, clasts, lignite, and carbonate grains, and the abundance of matrix were not considered in sorting estimates. Categories of induration for siliciclastic sediment depends on the amount of matrix and cement present. Samples from this core are categorized as: loose (grains are not bound by cement or clay matrix); clay-bound (framework grains are bound in a soft clay matrix); and friable (framework grains are bound in a hardened clay matrix and cement). The textural classification of carbonates is based on the distribution and abundance of carbonate matrix and grains (Dunham, 1962). A mudstone includes less than 10-percent carbonate grains in a matrix-supported texture. A wackestone includes greater than 10-percent carbonate grains in a matrix-supported texture. A packstone has a grain-supported texture with carbonate matrix between the grains. A grainstone consists of carbonate grains without a matrix. In addition to this textural based classification, this report uses the compositional terminology of Folk (1959, 1962) whose four major compositional types are based on allochemical constituents (bioclasts, made of broken and whole skeletal parts; ooids; intraclasts; and fecal pellets). Carbonates in the core are predominantly calcite with some aragonite. Carbonates are described as either loose, partially lithified, or lithified. 5 I System/ European !Provincial Series Stage . ~ Alabama Western Georgia I l Eastern Georgia Utholo9ic Georgia Geologic 2 Unit SurveyNomenclature South Carolina North E Carolina ~ I Priabonian ,.f-- - - "~ ' Bartonian .~~t.l ~'6 .... ... Lutetian ~ ~ j Yp resian ="' ~~ba~ian t.l "'"'0 " Selandian I ~I ~ j Danian Yazoo Jacksonian Clay E8 Ocala limestone E7 Moad~Br.an~hformatioo Fl\ G.11ml!tlSand Usbon Formation Usbon Formation ..... _. Claibornian I . E4 EJ Sabinian I Midwayan Tallahatta Formation Tallahatta Formation -E2 Hatchetigbee/BashiFm Hatchetigbee/BashiFm "ttL-~;~ i Nuflw r au.~ruiu .Form~rodon SMmJn f?rmt!)pn ) Po~ rters C~eek Clavton For. Jation I Pur!t:nl Cnek F'Q[ffiltla.a I Clayton Formation E , -P2 I PI Barnwell Group UsbonF Snapp Formation Undifferentiated Black Mingo Formation Barnwell Group McBean Formation Unnamed Q. I~ Santee Castle Hayne Formation f!l Formation E~ ~-----------; 0 Beaufort Formation "' Maastrichtian Navarroan Prairie BluffChalk Providence Sand 1--- Rinln..-nrmntifth Ripley Formation !J(<; St~IC rc ek - F.nrmntirm Peedee Formation Black Creek Peedee Formation Tayloran Campanian Demopolis Chalk Cusseta Sand UK4 GaFiUmard~J-aF1oarcmkaCtrieoenk Group Black Creek ...="'"' 0 t.l ell ... iccll.. ~"' ~ Santonian Austinian I I Coniacian Mooreville Chalk Blufftown Formation _ Eutaw Formation 1-McShan ~mation _ Eutaw Formation "Tuscaloosa Fm" ~ UK2 UKI - - ?- Caddin Formation Sb"JlherdGro~-el1m Group Middendorf Formation Middendorf Formation r- Cape Fear Formation ~pe Fear Fm ---4--- -~~ . - ? Turonian r---? ~Eaglefordian I Tuscaloosa Group !Tuscaloosa Formation Cenomanian I Woodbinian Cape Fear Formation Clubhouse Formation - ?--- + - Beech Hill Formation 1 ModifiedfromProwellandotheJ>, 1985 2 FromHuddlestun and Summerour, 1995. Figure 2. Generalized correlation of units of Late Cretaceous through Eocene age in the southeastern United States. Gray areas indicate missing stratigraphic interval. Abbreviation used: Fm, formation; Modified from Clarke and others, 1994. Micropaleontology Paleontologic data provide geologic age for several geologic units at the Girard site. Paleontologic samples were examined for dinoflagellates, pollen, benthic foraminifera, and calcareous nanofossils. Eleven of the paleontologic samples yielded age-diagnostic assemblages for Tertiary sediments. The locations of all samples are shown on plate 1 as small triangles adjacent to the lithologic column. Paleocene palynomorphs were taken from four paleontologic samples ranging in depths from 484.3 to 532.7 ft. The sample from 532.5 to 532.7 ft contains a nondiagnostic dinoflora consisting of small peridiniacean cysts, Spiniferites spp., and a few specimens of the Areoligera group. The sample from 521.0 to 521.2 ft contains lower Paleocene, Midwayan dinocysts, including Carpatella cornuta Grigorovich and a new species of Alterbidinium. The assemblage is dominated by small peridiniacean cysts. Pollen in this paleontologic sample included Pseudoplicapollis serenus Tschudy, Caryapollenites prodromus group of Frederiksen (1991 ), and an unnamed species of Sparganiaceaepollenites in addition to long-ranging forms. Osculapollis? colporatus Frederiksen was also tentatively identified in the sample. The overlap of age ranges represented by this assemblage suggests a late Paleocene age with reworked material of early Paleocene age. The sample from 514.0 to 514.3 ft contains a sparse dinoflora. The paleontologic sample from 484.1 to 484.3 ft contains a typical late Paleocene (late Midwayan or early Sabinian) dinoflora, again dominated by small peridiniacean cysts and containing Damassadinium californicum (Drugg) Fensome et al. and Phelodinium sp. of Edwards. Paleontologic samples examined between 327 and 415 ft indicate a range in age from early Eocene to the early part of the middle Eocene. The age of the sample at 423 ft is Eocene, but attempts to determine whether the age is early or middle Eocene were inconclusive. A paleontologic sample at 415 ft recovered a single specimen of Wetzeliella or Dracodinium. A second sample from 415.2 to 415.5 ft, tentatively identified; Homotryblium tenuispinosum of Davey and Williams. This paleontologic sample contains very few pollen grains, but two of the specimens were of Platycaryapollenites sp. cf. P swasticoidus; and therefore, are of Eocene age, (probably early Eocene or possibly early middle Eocene age). The next productive paleontologic sample was from 362 to 362.3 ft and contained a variety of dinocysts suggesting correlation to the lower part of the Lisbon Fom1ation in Alabama. Important species include Pentadinium favatum Edwards, Cerebrocysta bartonensis Bujak, Phthanoperidinium echinatum Eaton, Samlandia chlamydophora Eisenack, Wetzeliella articulata Eisenack, Glaphyrocysta? vicina (Eaton) Stover and Evitt and a new species of Eocladopyxis. The paleontologic sample from 327.3 to 327.5 ft contains a similar dinocyst assemblage. Paleontologic samples examined between 258ft and 322 ft indicate that these sediments are characteristic of the upper part of the middle Eocene (correlative to the upper Lisbon and Gosport Formations). The paleontologic sample at 322.3 to 322.5 ft contains a very sparse and relatively nondiagnostic dinocyst assemblage, but includes Pentadinium goniferum Edwards. The paleontologic sample from 321.4 to 321.6 contains the lowest occurrence of Cordosphaeridium cantharellus (Brosius) Gocht. The paleontologic sample from 257.8 to 258ft contains a well preserved, diverse dinocyst assemblage which includes Pentadinium polypodum Edwards, Samlandia sp. and Enneadocysta arcuata (Eaton) Stover and Williams. The paleontologic sample from 211.1 to 211.3 ft contains the species Rhombodinium perforatum (Jan du Chene & Chateauneut) Lentin & Williams and Dapsilidinium pseudocolligerum (Stover) Bujak et al. These species suggest a late Eocene age, but are not wholly diagnostic. The paleontologic sample at 146.7 ft contains a very sparse and nondiagnostic dinocyst assemblage. Samples at 64 and 104 ft were barren of dinocysts. HYDROLOGIC DATA Hydrologic data are presented for 1992-95 and include the hydrogeologic units encountered, aquifer properties, and ground-water levels at the Girard site. Hydrogeologic units are correlated to other named units. Aquifer-property data include core analyses, resistivity estimates, and aquifer-test results. Ground-water levels were collected on a continuous basis from each test well upon completion of construction and development. Hydrogeologic Units Hydrogeologic units at the Girard site were related to previously named hydrogeologic units by comparing core and geophysical data collected at the site to interpreted borehole data from nearby sites reported by Miller (1986); Clarke and others (1985, 1994, 1996); Brooks and others (1985); and Aadland and others (1992). This comparison indicates that several 7 equivalents to hydrogeologic units described in previous reports also are present at the Girard site. A generalized correlation chart of hydrogeologic and timestratigraphic units in the study area is shown in figure 3. Based on the nomenclature of Clarke and others (1996), they are, in descending order: loosely consolidated sand and calcareous sand of Eocene age of the Upper Three Runs aquifer of Aadland and others ( 1992); the Gordon aquifer of Brooks and others (1985); and Aadland and others (1992); the Dublin aquifer system (Clarke and others, 1985)--comprised of the upper Paleocene clay of the Millers Pond confining unit; and the predominantly sand, upper Paleocene Millers Pond aquifer (Clarke and others, 1996); the lower Paleocene calcareous sand and clay of the upper Dublin confining unit combined with the Upper Cretaceous sand of the upper Dublin aquifer; and the Upper Cretaceous clay of the lower Dublin confining unit and the sand of the lower Dublin aquifer (Clarke and others, 1996); and the Midville aquifer system (Clarke and others, 1985, 1996)--comprised of the interbedded sand and clay of the upper Midville confining unit that overlies the clayey sand of the upper Midville aquifer; this, in tum, overlies the Upper Cretaceous clayey sand and clay of the lower Midville confining unit and the predominantly sand lower Midville aquifer. At the Girard site, TW-1 is screened in the Upper Three Runs aquifer and is used only to monitor water levels. Wells TW-2 and TW-3 are screened, respectively, in the lower Dublin and lower Midville aquifers and were selected because of the sparsity of hydrologic data for these two aquifers in the study area. Hydraulic Properties Hydraulic properties described for the Girard site include horizontal hydraulic conductivity in aquifers and vertical hydraulic conductivity in confining units. Clemson University, Clemson, S.C., personnel conducted aquifer tests in wells TW-2 and TW-3 to determine transmissivity, horizontal hydraulic conductivity and to detect any interaquifer leakage. Selected low-permeability core samples were analyzed in the laboratory, by Core Laboratories Inc., to determine vertical hydraulic conductivity and porosity. Horizontal hydraulic conductivity for the upper Midville and upper Dublin aquifers also was estimated from borehole formation resistivity logs using the method described by Faye and Smith (1994). Hydraulic Conductivity Estimated from Aquifer Tests Horizontal hydraulic conductivity, estimated from aquifer pumping tests for the lower Midville and lower Dublin aquifers, indicates that the lower Dublin aquifer is much more productive than the lower Midville at the Girard site. During December 1994 and January 1995, aquifer tests were conducted and analyzed by Clemson University at two of the three test wells at the Girard Site (Luetkehans, 1995a,b). The Jacob straight-line method (Cooper and Jacob, 1946) was used to estimate transmissivity values for TW-2 and TW-3 test wells. Aquifer-test results at the Girard site are listed in table 1. Barometric efficiency was determined by calculating the ratio of change in hydraulic head in a well (because of atmospheric changes) to the actual change in atmospheric pressure. A barometric efficiency of 1 indicates that 100 percent of the atmospheric pressure changes have been transmitted to an aquifer; whereas, a barometric efficiency of zero indicates atmospheric pressure changes have not been transmitted to an aquifer. Prior to computation of hydraulic properties, water-level readings were corrected for barometric fluctuations by subtracting atmospheric pressure changes multiplied by the barometric efficiency (Luetkehans, 1995a,b). Well-efficiency was determined by dividing the theoretical drawdown by the actual drawdown after 24 hours of pumping (Luetkehans, 1995a,b). Partial penetration of both of the aquifers was not considered in these estimates, so actual well efficiencies could be higher than those reported by Leutkehans (1995a,b). The lower Dublin well (TW-2) had a well efficiency of 26 percent and the lower Midville well (TW-3) had a well efficiency of 49 percent. The horizontal hydraulic conductivity of 177ft per day (ft!d) for the lower Dublin aquifer was almost 20 times that of 8.9 ft/d for the lower Midville aquifer. In addition, the transmissivity of 5,300 ft2/day for the lower Dublin aquifer is almost five times greater than the 1,130 ft2/day of the lower Midville aquifer. Finally, the specific capacity of 3.52 gallons per minute per foot (gaVmin/ft) in the lower Dublin aquifer is almost twice the 1.85 gal/min/ft of the lower Midville aquifer. 8 TIME-STRATIGRAPHIC UNIT SERIES Georgia HYDROGEOLOGIC UNIT This Study South Carolina, EOCENE -------------- PALEOCENE ---- ----- --- -- (/) ::J 0 w () aw~: () aaaw..:.. ::J 2 JACKSONIAN/ 3 UPPER FLORIDAN AQUIFER UPPER THREE RUNS AQUIFER CONFINING UNIT GORDON AQUIFER 4 SYSTEM CONFINING DUBLIN AQUIFER 5 SYSTEM CONFINING UNIT MIDVILLE AQUIFER 5 SYSTEM GORDON CONFINING UNIT GORDON AQUIFER MILLERS POND CONFINING UNIT" MILLERS POND AQUIFER" UPPER DUBLIN CONFINING UNIT UPPER DUBLIN AQUIFER LOWER DUBLIN CONFINING UNIT LOWER DUBLIN AQUIFER UPPER MIDVILLE CONFINING UNIT UPPER MIDVILLE AQUIFER LOWER MIDVILLE CONFINING UNIT LOWER MIDVILLE AQUIFER UPPER THREE RUNS AQUIFER GORDON CONFINING UNIT GORDON AQUIFER CROUCH BRANCH CONFINING UNIT CROUCH BRANCH AQUIFER MCQUEENS BRANCH CONFINING UNIT MCQUEENS BRANCH AQUIFER CONFINING UNIT BASAL CONFINING UNIT APPLETON CONFINING SYSTEM Figure 3. Correlation of hydrogeologic units in the vicinity of the Girard hydrogeologic test site, Girard, Georgia. 9 Table 1. Summary of aquifer-test analyses at the Girard hydrogeologic test site, January 1994 and December 1995 [Analyses by Clemson University (Luetkehans, 1995a,b)] Well Water-bearing number unit Dates of test Pumping period (hours) Average discharge (gallons per minute) Maximum drawdown (feet) Specific capacity (gallons per minute per foot) Well Barometric Transmissivity Horizontal hydraulic efficiency efficiency (feet squared conductivity (percent) (percent) per day) (feet per day) TW-2 lower Dublin 12/27/94 72 77.67 22.01 3.52 26 58 5,300 177 to 01/4/95 TW-3 lower Midville 01/14/95 72 76.50 41.30 1.85 49 55 1,130 8.9 to 01/21/95 Extended pumping during each of the two aquifer tests did not induce leakage between the two aquifers, indicating that these aquifers are hydraulically separated at the Girard test site. However, this lack of response also could be because of the large vertical distance (297 ft) between the screens in wells TW-2 and TW-3-an interval that includes the upper Midville and lower Midville confining units-and does not necessarily reflect the competency of the confining units. Hydraulic Conductivity ofCore Samples Laboratory analyses indicate that the vertical hydraulic conductivity of confining units is low, with all values less than 9.07 X 10"5 ft/day. Laboratory analyses were performed on selected low permeability core samples (table 1, plate 1) by Core Laboratories, Inc., New Orleans, La., to determine vertical hydraulic conductivity and porosity. Vertical hydraulic conductivity was determined using a flexible wall permeameter, following American Society for Testing and Materials (ASTM) standard D-5084-90 (ASTM, 1990). Porosity was determined using procedures described in ASTM standard D-2216-80 (ASTM, 1980). Eight core samples were collected at depths ranging from 305.5 to 850.3 ft (plate 1). The two deepest samples collected from depths of 800.3 and 850.3 ft were unusable because of fracturing. A third, fractured sample, collected from the upper Dublin confining unit at 518.5 ft, yielded vertical hydraulic conductivity and porosity values that appear reasonable but may be suspect because of fracturing. The first core sample, with a reported vertical hydraulic conductivity of3.34 x 104 ft/d and a reported porosity of 43 .7 percent, was collected from within the Gordon confining unit from a layer of yellow green, calcareous, sandy clay at a depth of305.5 ft. Both values fall within expected ranges as reported by Heath (1983) for vertical hydraulic conductivity and Freeze and Cherry (1979) for porosity. The second core sample, with a reported vertical hydraulic conductivity of 1.54 x 1o3 ft/d and a porosity of36.2 percent was collected from near the top of the upper Dublin confining unit from a black, noncalcareous silty clay at a depth of 487.5 ft. The value for the vertical hydraulic conductivity lies at the upper end of the range of vertical hydraulic conductivities of clay reported by Heath (1983); although the porosity value lies at the lower end of the range for porosity of silt reported by Freeze and Cherry (1979). These values (high vertical hydraulic conductivity and low porosity) likely are explained by the 10-percent silt and sand content reported by Core Laboratories, Inc. The third core sample, with a reported vertical hydraulic conductivity of9.41 X 102 ft/d and a porosity of 34.9 percent, was collected from a black, laminated clay at the bottom of the upper Dublin confining unit at a depth of 518.5 ft. This value is suspect because of fracturing during sample collection; however, the value is reported here to provide additional data. The fourth core sample, with a reported vertical hydraulic conductivity of9.07 x 105 and a porosity of 25.4 percent, was collected from a yellow to light gray, sandy clay at a depth of 628.0 ft. This clay lies near the bottom of the upper Dublin confining unit. The value for the vertical hydraulic conductivity lies in the middle of the range of vertical hydraulic conductivity of clay reported by Heath (1983); but the porosity value also lies at the lower end of the porosity values of sand reported by Freeze and Cherry (1979). The porosity value is anomalous and can only partially be explained by the presence of up to 15-percent sand intercalated throughout the sample, but may be due partially to the presence of fine-sand-size mica. The fifth core sample, with a reported vertical hydraulic conductivity of 1.82 x 104 ft/day and a porosity of25.7 percent, was collected from a medium gray, micaceous, silty, laminated, clay within the lower 10 Dublin aquifer at a depth of737.5 ft. The value for the vertical hydraulic conductivity lies in the upper range of vertical hydraulic conductivity value for clay as reported by Heath (1983); whereas the porosity value lies at the lower end of the porosity values of sand reported by Freeze and Cherry (1979). As was the case in the fourth core sample, the porosity value is anomalous and can only partially be explained by the presence ofup to 20-percent silt intercalated throughout the sample, but may also be due partially to the presence of fine-sand-size mica, and fine-sand laminations. The sixth core sample, with a reported vertical hydraulic conductivity of 3.43 x 1o-5 ft/d and a porosity of35.4 percent, was collected at a depth of781.5 ft from a gray black, clay with thin beds (0.05 to 0.2 in.) of very fine to fine micaceous sand that lies at the top of the upper Midville confining unit. The value for the vertical hydraulic conductivity is in the middle range for clay as reported by Heath (1983); however, the porosity value is at the lower end of the values for silt as reported by Freeze and Cherry (1979). The low porosity can be attributed to the thin-bedded sand in the unit. Hydraulic Conductivity Estimated from Borehole Resistivity Horizontal hydraulic conductivity was estimated for several of the aquifers at the Girard site by applying logarithmic regr~ssion models developed by Faye and Smith (1994) to borehole resistivity data. Faye and Smith (1994) developed regression models describing the relation between hydraulic conductivity (as determined by aquifer tests) to aquifer bulk resistivity (as determined from borehole geophysical logs) for clastic aquifers. Using data from boreholes throughout the Coastal Plain ofthe southeastern United States, Faye and Smith (1994) developed regression models based on the age of sediments that comprise aquifers: Late Cretaceous Kh = 3.2R0 0.48 (l) Paleocene and early Eocene Kh = 0.57R0 1.0 (2) middle Eocene Kh =3.8R0 0.67 (3) where: Kh, is the horizontal hydraulic conductivity in ft/d; and Ro, is bulk resistivity in ohm-meters. At the Girard site, bulk resistivity was determined using the long-normal (64-in.) resistivity log (Joan S. Baum, U.S. Geological Survey, written commun., 1995). The maximum, minimum, and mean bulk resistivity for a contributing interval was determined from digital resistivity data collected from the borehole. Estimates were not computed either for the Upper Three Runs aquifer or the Gordon aquifer because the estimates do not fall within one of the age ranges evaluated by Faye and Smith (1994). However, estimates were computed for the Millers Pond, upper Dublin, lower Dublin, upper Midville, and lower Midville aquifers (table 2). The mean horizontal hydraulic conductivity estimate for the lower Dublin aquifer using the Faye and Smith (1994) method (table 2) was compared to the horizontal hydraulic conductivity determined from aquifer testing in TW-2 (table 1), completed in the lower Dublin aquifer. Estimated mean hydraulic conductivity of26.9 ft/d computed using the Faye and Smith (1994) method was about 85 percent lower than the value of 177 ft/d determined from the TW-2 aquifer test. The disparity between these values lies well within the range of absolute error reported by Faye and Smith ( 1994). Table 2. Estimates of horizontal hydraulic conductivity determined from formation resistivity logs at the Girard hydrogeologic test site [Estimated using methodology of Faye and Smith (1994); analyses by Joan S. Baum, U.S. Geological Survey, written commun., 1995] Aquifer Gordon Millers Pond Upper Dublin Lower Dublin Upper Midville Lower Midville Lithology fine to very coarse bioclastic sand fine to very coarse sand and gravel fine to very coarse sand fine to very coarse sand fine to very coarse sand very fine to coarse sand Age of sediments early Eocene Paleocene Late Cretaceous Late Cretaceous Late Cretaceous Late Cretaceous Penneable thickness (feet) 38.8 11.8 5.8 33.6 12.8 37.0 Estimated horizontal hydraulic conductivity Minimum Maximum Mean (feet per day) (feet per day) (feet per day) 52.2 94.5 55.2 18.8 48.5 38.0 16.1 23.4 19.0 16.4 33.7 26.9 14.7 29.2 23.3 14.1 31.0 24.7 11 For the lower Midville aquifer, the mean horizontal hydraulic conductivity, using the Faye and Smith (1994) method, was compared to the horizontal hydraulic conductivity determined from the TW-3 aquifer test. The estimated mean hydraulic conductivity of24.74 ft/d computed using the Faye and Smith (1994) method was about 278 percent higher than the value of 8.9 ft/d determined from the TW-3 aquifer test. The disparity between these values also lies within the range of absolute error reported by Faye and Smith (1994), but may be enhanced by a low well efficiency (nearing 50 percent). Runs aquifer is shallow, this decline could be the result ofmicroclimatic changes, such as less localized precipitation or an increase in evapotranspiration because of localized temperature fluctuations that are not recorded by the existing precipitation monitoring network. A change in local agricultural crop type also could increase evapo-transpiration that could, in turn, cause an increase in localized agricultural pumping for irrigation. It is unlikely that evapotranspiration is the causative factor because there should be a relatively strong correlation between evapotranspiration and the summer months; however, this is not the case. In addition to Ground-Water Levels Following well completion and development, continuous water-level recorders were installed in the three test wells to monitor water-level fluctuations and trends. Water-level data were used to determine the vertical distribution of hydraulic head in the waterbearing units (plate l ). these localized factors, the possibility oflong-term deficits in rainfall, causing a general decline for the year, is real and would not be evident from inspection of the short period of record. For the remaining two wells (TW-2 and TW-3), insufficient data prevent a visual assessment of water-level fluctuations; however, a statistical treatment of the data may be useful. Vertical distribution of hydraulic head gives an indication of the potential for vertical ground-water movement and interconnection between adjacent aquifers. Under unstressed conditions, upward gradients occur in discharge areas; downward gradients occur in recharge areas; and minimal vertical gradient exists in areas dominated by lateral flow. Water levels were measured in the three wells on March 13, 1995, yielding head measurements of34.98 ft below land surface for TW-1; 89.85 ft for TW-2; and 74.93 ft for TW-3. These values, when corrected for altitude differences between wells, indicate that there is an upward hydraulic gradient from well TW-3 to TW-2; thus, indicating a potential for discharge from the Midville aquifer system into the Dublin aquifer system. A statistical comparison of stream stage, groundwater levels, and precipitation was performed using Spearman's rank correlation coefficient (SRCC) (lman and Conover, 1983). The SRCC measures the strength of monotonic correlation between two variables. If the two variables increase together, the variables have a positive correlation; however, if one variable increases, while the other decreases, the variables are said to have a negative correlation. The closer the SRCC is to either +1 (a positive correlation) or -1 (a negative correlation), the stronger the relation is between the two variables. Ground-water-level data from the three test wells were compared to stage data from the stream gage at Jackson, S.C., and precipitation data from Waynesboro, Ga. (Raymond Brown, National Oceanic and Atmospheric Administration, Southeastern Region Climatic Data Hydrographs showing relations between ground- Center, oral commun., 1995) (table 3). water levels at the Girard site, stream stage of the Savannah River near Jackson, S.C., and precipitation at the city of Waynesboro, Ga., are shown in figure 4. Locations of the stream gage and precipitationmonitoring station are shown in figure I. Two strong correlative relations are apparent based on the SRCC of0.683 between water levels in TW-1 and TW-2 may be caused by mass loading in the upper Dublin aquifer in response to recharge of the Upper Three Runs aquifer. Also, interaquifer leakage induced Because of the relatively short period of record available for the three wells, few trends are evident in the data. The period of record for TW-1 is somewhat longer than the period of record for TW-1 and TW-2; by pumping the lower Dublin aquifer may be affecting the water level in the Upper Three Runs aquifer. These are possibilities, but other explanations may be plausible but undetected given the short period of record. however, and a few conclusions may be drawn. It is apparent that there is no instantaneous response to rainfall in TW-1 (Upper Three Runs aquifer) and TW-2 (lower Dublin aquifer); and there is no apparent connection to the Savannah River during the period of record. The 4-ft water-level decline observed in well TW-1 between late 1993 and early 1995 is incongruous and not readily explained. Because the Upper Three The SRCC of0.703 between water levels in the deeper TW-2 and TW-3 wells (table 3), indicates that water-level fluctuations for these wells also are correlative based on the SRCC. The underlying causal factor (or factors) for this correlation is uncertain. However, given the correlation of all three wells with precipitation, mass loading from precipitation, is likely a major causative factor. 12 TW-1 Open interval48.7-72.7 feet 36 Blank where data not available L1J ~ 38 a: ::J (/) 0 z I 4o -1~ N ___ D J F M A M J J :) 1993 1994 As 0 N ID J I F M A M J 1995 J A ~ L1J OJ ti:i wu.. TW-2 Open interval 743-773 feet 90 - Blank where data not available ~ _j 92 L1J > L1J _I a: 94 ...1-l--=-!- - 1 L1J s~ : I N - D J F MA M J J 1993 1994 A s F MA M A 1995 z