ML18026A611

Tucci-1992- TN5034--USGS 1992 Melton - Hydrology of Melton Valley at Oak Ridge National Lab

ML18026A611
Person / Time
Site:	Clinch River
Issue date:	01/24/2018
From:	Tucci P US Dept of Interior, Geological Survey (USGS)
To:	Office of New Reactors
Fetter A
References
Download: ML18026A611 (82)

Text

HYDROLOGY OF MELTON VALLEY AT OAK RIDGE NATIONAL LABORATORY, TENNESSEE By Patrick Tucci U.S. GEOLOGICAL SURVEY Water-Resources Investigations Report 92-4131 Prepared in cooperation with the U.S. DEPARTMENT OF ENERGY Nashville, Tennessee 1992

U.S. DEPARTMENT OF THE INTERIOR MANUEL LUJAN, Jr., Secretary U.S. GEOLOGICAL SURVEY Dallas L. Peck, Director For additional information write to:

District Chief U.S. Geological Survey 810 Broadway, Suite 500 Nashville, Tennessee 37203 Copies of this report can be purchased from:

U.S. Geological Survey Books and Open-File Reports Section Federal Center Box 25425 Denver, Colorado 80225

CONTENTS Abstract 1

Introduction 2

Purpose and scope 2

Description of study area 2

History 2

Topography and climate 4

Previous studies 5

Acknowledgments 7

Methods of investigation 7

Well construction and well identification 7

Aquifer tests 9

Geologic setting 10 Bedrock 10 Regolith 12 Alluvium 12 Structure 12 Hydrology of Melton Valley 13 Surface water 13 Stream discharge 15 Base flow 20 Ground water 26 Regolith 26 Hydraulic conductivity 26 Ground-water flow 29 Recharge and discharge 34 Bedrock 37 Hydraulic conductivity 37 Ground-water flow 38 Recharge and discharge 38 Simulation of ground-water flow 38 Cross-sectional model 40 Model construction, assumptions, and boundary conditions 40 Results of simulations 41 Area! model 45 Model construction, assumptions, and boundary conditions 45 Results of simulations 49 Sensitivity analysis 56 Model limitations 58 Water chemistry and radionuclides 62 Surface-water quality 64 Major-ion chemistry in surface water 64 Radionuclides in surface water 64 Ground-water quality 65 pH 66 Major-ion chemistry of ground water 68 Radionuclides in ground water 69 Summary and conclusions 72 Selected references 75 Contents III

FIGURES

1. Map showing location of study area and waste burial grounds 3

2. Graph showing annual precipitation at Oak Ridge and Oak Ridge National Laboratory 4

3. Graph showing cumulative departure from normal monthly precipitation at Oak Ridge and at Oak Ridge National Laboratory 6

4-6. Maps showing:

4.

Location of wells 8

5.

Generalized geology of Melton Valley 11

6.

Location of streamgaging stations on Whiteoak Creek and Melton Branch 14

7. Duration curves of daily-mean discharge at three stations on Whiteoak Creek at Oak Ridge National Laboratory, water year 1988 17

8. Hydrographs of daily-mean discharge at three stations on Whiteoak Creek at Oak Ridge National Laboratory, water year 1988, with estimated average effluent discharge 18

9. Duration curves of daily-mean discharge at stations on Whiteoak Creek and Melton Branch at Oak Ridge National Laboratory, water year 1988, and Poplar Creek near Oak Ridge, water years 1961 through 1988 19

10. Stream-discharge hydrograph and base-flow component at station Ml, Melton Branch near Melton Hill 22

11. Adjusted and unadjusted stream-discharge hydrographs of Whiteoak Creek below Melton Valley Drive and unadjusted stream-discharge hydrograph of Whiteoak Creek near Melton Hill 24 12-14. Graphs showing:

12. Relation between annual precipitation and adjusted annual base flow at stations on Whiteoak Creek and Melton Branch at Oak Ridge National Laboratory 27

13. Relation between adjusted annual stream discharge and adjusted annual base flow at stations on Whiteoak Creek and Melton Branch at Oak Ridge National Laboratory 28

14. Hydraulic-conductivity variation with depth 29

15. Boxplots showing ranges of hydraulic-conductivity values for selected depth intervals 30

16. Boxplots showing ranges of hydraulic-conductivity values for regolith developed over selected geologic units 31

17. Map showing location of wells in which water-table depth was measured on June 28,1988 32

18. Map showing water-table configuration and general directions of ground-water flow, June 28,1988 33

19. Hydrographs for north well cluster and south well cluster, burial ground 5 35

20. Hydrographs of wells U19 and U35, January 1986 through September 1988 36

21. Map showing potentiometric surface of the shallow bedrock zone, June 28,1988 39

22. Cross-sectional model grid and boundaries 40

23. Hydraulic-conductivity distributions, unadjusted for horizontal anisotropy, for cross-sectional model simulations 1 and 2 42

24. Hydrograph of well 474 showing water-level fluctuations from April 1983 through September 1988 43

25. Simulated water levels and general directions of ground-water flow for cross-sectional model simulations 1 and 2 44 26-32. Maps showing:

26. Areal model grid and boundary conditions 47

27. Areal distribution of simulated recharge 48

28. Saturated thickness of layer 1, June 28,1988 50

29. Simulated water levels for layer 1 51 iv Hydrology of Melton Valley at Oak Ridge National Laboratory, Tennessee

30. Simulated water levels for layer 2 52

31. Simulated water levels for layer 3 53

32. Simulated water levels for layer 4 55 33-36. Graphs showing:

33. Changes in root mean square error of simulated water levels with respect to changes in recharge rate and layer 1 hydraulic conductivity 57

34. Changes in root mean square error of simulated water levels with respect to simultaneous changes in recharge and hydraulic conductivity 59

35. Changes in root mean square error of simulated water levels with respect to changes in layer 1 row-to-column anisotropy and layer 2 transmissivity 60

36. Changes in root mean square error of simulated water levels with respect to changes in river and drain conductance values and vertical leakance between layers 1 and 2 61

37. Map showing location of water-quality sampling sites and burial grounds 63

38. Trilinear diagram of principal-ion data of water from 31 wells in Melton Valley 69

39. Graph showing concentrations of dissolved gross beta and tritium in ground-water samples collected at Oak Ridge National Laboratory in August and September 1987 71 TABLES

1. Monthly and annual precipitation at Oak Ridge National Laboratory and normal monthly and annual precipitation at Oak Ridge, Tennessee 5

2. Results of slug tests in wells at Oak Ridge National Laboratory 9

3. Streamgaging stations on Whiteoak Creek and Melton Branch, and periods of record and drainage areas above the stations 15

4. Summary of discharge data for stations on Whiteoak Creek and Melton Branch, water year 1988, and Poplar Creek, water years 1988 and 1961 through 1988 20

5. Summary of annual discharge data, including effluent part, in cubic feet per second, for all complete years of record at stations on Whiteoak Creek and Melton Branch 21

6. Annual base flow at stations on Whiteoak Creek and Melton Branch that are not affected by effluent discharge 23

7. Estimates of annual base flow at stations on Whiteoak Creek and Melton Branch that are affected by effluent discharge 25

8. Model-calculated water-budget components for cross-sectional model simulations 45

9. Adjustments to June 28,1988, water levels used to estimate average-annual water levels 49

10. Percentage of simulated water levels within 20,10, and 5 feet of estimated average-annual water level 54

11. Model-calculated water-budget components for areal model simulations 56

12. Concentrations of major ions in water from streamflow stations in the vicinity of burial grounds 4, 5, and 6 64

13. Activities of radiochemical constituents in water from streamflow stations and wells distant from trenches 65

14. Concentrations of major ions in water from wells in the vicinity of burial grounds 5 and 6 67

15. Activities of radiochemical constituents in water from wells in the vicinity of burial grounds 5 and 6 70 Contents

CONVERSION FACTORS AND VERTICAL DATUM Multiply By To obtain inch (in.)

25.4 millimeter inch per year (in/yr) 25.4 millimeter per year foot (ft) 0.3048 meter foot per day (ft/d) 0.3048 meter per day mile (mi) 1.609 kilometer square foot (ft2) 0.0929 square meter square mile (mi )

2.590 square kilometer gallon (gal) 3.785 liter gallon (gal) 0.003785 cubic meter cubic foot (ft3) 0.02832 cubic meter foot per day (ft/day) 0.3048 meter per day cubic foot per second (ft3/s) 0.02832 cubic meter per second cubic foot per second-day (ft3/s-d) 0.02832 cubic meter per second-day cubic foot per second per square mile 0.01093 cubic meter per second

[(ft3/s)/mi2]

per square kilometer foot squared per day (ft2/d) 0.0929 meter squared per day picocurie (pCi) 0.037 becquerel 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 both the United States and Canada, formerly called Sea Level Datum of 1929.

vi Hydrology of Melton Valley at Oak Ridge National Laboratory, Tennessee

HYDROLOGY OF MELTON VALLEY AT OAK RIDGE NATIONAL LABORATORY, TENNESSEE By Patrick Tucci Abstract A hydrologic investigation of Melton Valley, the site of three low-level, radioactive solid-waste burial grounds, was initiated by the U.S. Geological Survey, U.S. Department of the Interior, in 1975. The investigation included description and analysis of the geology, surface water, ground water, and water qual-ity of the valley. This report focuses on information obtained since 1983.

Melton Valley normally receives abundant rainfall.

At the nearby city of Oak Ridge, normal annual precipitation for the period 1951 through 1988 was about 53 inches.

White oak Creek and its tributary, Melton Branch, are the main streams that drain Melton Valley.

Whiteoak Creek rises in adjacent Bethel Valley and drains the Bethel Valley section of Oak Ridge National Laboratory before entering Melton Valley. During relatively dry years, most streamflow in Melton Valley is effluent that is discharged from a sewage-treatment plant in Bethel Valley; the water is imported from a source outside the drainage basin.

Available data (12 months of record during a relatively dry period) indicate that Melton Branch has a larger natural discharge per unit drainage area and a larger base-flow component of natural discharge per unit drainage area than Whiteoak Creek.

Most ground-water flow in the valley is through regolith developed on six formations of alter-nating shale and limestone lithologies of the Cona-sauga Group.

The regolith consists mainly of clay, silt, and rock fragments.

Regolith is generally less than 50 feet thick, but locally is as much as 86 feet thick. Flow within the bedrock is much less than flow in the regolith. Hydraulic-conductivity values for both regolith and bedrock are generally low, but highly variable. Regolith hydraulic conductivities range from 6.6 x 10~4 to 6.9 feet per day, and the median value is 0.19 foot per day.

Bedrock hydraulic conductivities range from about 1 x 10"5 to 2.4 feet per day, and the values generally decrease with increasing depth.

Recharge to the ground-water system is primarily by precipitation on the ridges and hills. Average annual-recharge rates are estimated to range from about 1 to 8 inches per year. Discharge from the ground-water system is primarily to streams within, or bordering, the valley and to springs.

Cross-sectional and areal ground-water-flow models were used to provide an understanding of the flow system of Melton Valley.

The models indicate that from 91 to 96 percent of the recharge to the water table flows within the upper 50 feet of the ground-water system, and 97 percent of the recharge flows within the upper 100 feet. Less than 1 percent of the total ground-water recharge reaches depths greater than 250 feet.

Simulated recharge rates range between 1.0 and 4.7 inches per year; however, these rates are considered only as approximations because model results are non-unique.

Surface-water chemical data collected in 1985 and 1987 indicate principally calcium bicarbon-ate and calcium bicarbonate sulfate waters in which concentrations of dissolved solids range from 200 to 450 milligrams per liter. Surface-water flow during the sampling period was principally effluent discharged from Oak Ridge National Laboratory. Radiochemical constituents included strontium-90, gross beta, and tritium. Strontium-90 ranged from less than 0.4 to 420 picocuries per liter (pCi/L), gross beta ranged from 2.2 to 880 pCi/L, and tritium ranged from 3,900 to 2,900,000 pCi/L at five surface-water sites.

Principal ions in ground water are calcium, magnesium, and bicarbonate in most wells less than 100 feet deep, and sodium and bicarbonate in most wells greater than 100 feet deep. This indicates that two flow patterns are present in the subsurface. Shal-low ground-water circulation, at depths of less than 100 feet, is in regolith and upper bedrock.

Deeper ground-water circulation, at depths of greater than Abstract

100 feet, is in bedrock only, and has undergone greater geochemical evolution. The distribution of chemical data are insufficient to relate water chemistry to geologic unit.

Radiochemical constitu-ents were prevalent in water from wells near waste disposal trenches, and activity levels of radionuclides were considerably less in water from wells distant from disposal trenches.

Radiochemical constituents from water in wells included strontium-90 (ranging from 0 to 11,100 pCi/L), gross beta (ranging from 1.6 to 20,000 pCi/L), and tritium (ranging from 170 pCi/L to 360,000,000 pCi/L).

Apparent background activ-ities of gross beta and tritium were 5 and 500 pCi/L, respectively. Elevated activity levels of radiochemical constituents indicate mixing of ground water with water that has passed through disposal trenches.

INTRODUCTION Low-level radioactive waste has been generated at Oak Ridge National Laboratory (ORNL) since 1943. Most of the waste has been buried in trenches in Melton Valley. Water flow through the trenches resulted in the transport of radionuclides from the burial grounds to local streams (Cerling and Spalding, 1981; Steuber and others, 1981).

In order to plan remedial actions that might be taken to reduce water-transported radionuclides, information on and an understanding of the hydrology of the area is neces-sary. In an effort to fulfill this need, the U.S. Geo-logical Survey (USGS), in cooperation with the U.S.

Department of Energy (DOE), undertook a study of the hydrologic environment of the waste-disposal sites and nearby area.

Purpose and Scope The purpose of this report is to describe the hydrology of Melton Valley at Oak Ridge National Laboratory (ORNL), where three radioactive-waste burial grounds are located. Remedial actions being undertaken by the DOE and their principal contractor at ORNL, Martin-Marietta Energy Systems (MMES),

to reduce water-transported radionuclides may be evaluated in light of the hydrologic data and interpre-tations presented here. The information may be useful at other waste-burial sites located in similar hydro-geologic environments; that is, sites that receive abundant rainfall, and are underlain by rocks of low hydraulic conductivity in which part of the ground-water flow is through fractures.

This report includes a discussion of the geology of Melton Valley, analyses of the surface-water and ground-water systems of that area, a discussion of the construction and use of a two-dimensional cross-sectional model and a three-dimensional area! model, and interpretation of water-quality analyses to aid in defining ground-water flow. Many types of data were collected for use in this study. They include precipi-tation data, streamflow data, geologic data, surface geophysical data, subsurface geophysical logs, water-level data, hydraulic conductivities, and water-quality data. Data from previous studies and other ongoing studies also were used.

The investigation began in 1975 and ended in 1989. Emphasis prior to 1984 was on the develop-ment of conceptual models of ground-water flow through regolith at burial grounds 4, 5, and 6, and through regolith and bedrock at burial ground 5, and is described by Webster and Bradley (1988). This report is based primarily on work completed since 1983.

Description of Study Area The ORNL is located about 25 miles west of Knoxville, Tennessee, and 6 miles southwest of Oak Ridge, Tennessee. Melton Valley is in the southern part of ORNL, and three radioactive-waste burial sites are in the valley (fig. 1).

History The history of ORNL and the radioactive-waste burial areas has been described by, among others, Evaluation Research Corporation (written communica-tion prepared for ORNL, 1982), Webster (1976),

Francis and Stansfield (1986), Haase and others (1987), and Webster and Bradley (1988). Only a brief summary from these reports is given here.

Operation of ORNL began in 1943 for the purpose of nuclear-weapons development during World War II. The ORNL is now (1992) a research facility.

Most research is in nuclear-power development and other energy-related fields. Radio-active waste has been generated as a byproduct of the research and development. Solid waste is buried near Hydrology of Melton Valley at Oak Ridge National Laboratory, Tennessee

uol)onpoj)Ul

<Qc c

CL Q3 CL V

C (Q

O C

3 CL

the ORNL facility in unlined trenches and auger holes at waste burial sites (called Solid Waste Storage Areas by MMES personnel).

The trenches are of wide-ranging length and width, and most are less than 15 feet deep. They are covered with about 2 to 3 feet of soil. Waste burial began in 1944 at burial ground 1, and has continued at burial sites numbered sequen-tially through burial ground 6. Burial operations were transferred to Melton Valley in 1951, after burial grounds 1, 2, and 3 in Bethel Valley (fig. 1) were closed. Burial ground 4 is closed, and most of burial ground 5 is closed. Burial ground 6 is in operation.

Topography and Climate The ORNL is located in the southern half of the Ridge and Valley province. This province is an area of northeast-trending ridges and valleys which extend from the Saint Lawrence Valley in New York to the Gulf Coastal Plain in Alabama, and has an average width of about 40 miles in eastern Tennessee (Fenneman, 1938). The bottom of Melton Valley ranges from 740 to 860 feet in altitude. Haw Ridge, which forms the northwest perimeter of the valley, has crests at an altitude of about 1,000 feet. Copper Ridge forms the southeast perimeter, has crests at about 1,200 feet, and has the highest altitude in the ORNL area of 1,356 feet at the top of Melton Hill. Burial grounds 4, 5, and 6 are located on or near several small hills south of Haw Ridge.

Melton Valley is about 3 miles long from the headwaters to the Clinch River, and about 1.2 miles wide from the crests of Haw Ridge to the crests of Copper Ridge.

Most facilities of the ORNL laboratories, offices, and support units are located in adjacent Bethel Valley (fig. 1).

Normal annual precipitation at the U.S.

Weather Bureau station in Oak Ridge, about 6 miles northeast of ORNL, is 53.27 inches for the period 1951 through 1988, as computed from data given in a report by U.S. Department of Commerce (1972), and U.S. Department of Commerce Annual Climatological Summaries for 1972 through 1987, and U.S. Depart-ment of Commerce Monthly Climatological Data for 1988 (fig. 2). Normal is defined here as the long-term annual mean of 38 years of record.

The mean at ORNL, computed from data collected by the USGS at a continuous-recording station in burial ground 5 (table 1), is 47.03 inches for the period 1977 through 80 eo -

40 20 n

Normal at Period 19 0

51 ok Ridae 53.2 through 1988 l\\

7 inches 1 [

£ Zl 7A ANNUAL PRECIPIT ANNUAL PRECIPIT Mean Period

-; j \\

w.

ATION AT OAK RIDGE ATION AT ORNL at ORNL 47.03 inches 1977 through 1988 HIT

' , f'< \\\\

', /

';//

1950 1955 1960 1965 1970 1975 1980 1985 1990 Figure 2.--Annual precipitation at Oak Ridge and Oak Ridge National Laboratory (ORNL).

(Oak Ridge data from U.S. Department of Commerce, 1972, Annual Climatological Summaries for 1972-87, and Monthly Climatological Data for 1988.)

Hydrology of Melton Valley at Oak Ridge National Laboratory, Tennessee

1988. The mean at Oak Ridge for the same period is 50.00 inches.

Cumulative departure from normal monthly precipitation was within +/-30 inches of normal during most of the period 1951 through 1972 (fig. 3). The wet period from the beginning of 1973 through 1979 had a cumulative departure from normal of 54 inches (80 inches at the end of 1979 minus 26 inches at the beginning of 1973), and 24 inches of this amount occurred in 1973. The dry period from the beginning of 1980 through mid-1988 had a cumulative departure from normal of 79 inches (79 inches at the end of 1979 and 0 inches in mid-1988), and 40 inches of the total occurred during the period 1986 through mid-1988. The 1987 total precipitation of 34.14 inches at ORNL was the lowest ever recorded at ORNL and was less than the record low reported at the weather bureau station at the city of Oak Ridge. Much of the water-level and streamflow data were collected during the period 1986 through 1988, and these data reflect the abnormally dry periods.

Previous Studies Previous hydrogeologic studies conducted at ORNL, and history of ORNL operations regarding waste burial, were described in hydrologic reports by Webster (1976) and Webster and Bradley (1988), and are not repeated here. The information given below pertains to reports of hydrologic and geologic investi-gations of Melton Valley, all of which have been published since 1985.

Tucci (1986) used a three-dimensional, finite-difference model in a preliminary analysis of ground-water flow in Melton Valley. That study identified additional data needs and provided a basis for the flow modeling discussed in this report. Tucci (1987) described results of direct-current resistivity surveys, which were used to determine depth to bedrock, and terrain-conductivity surveys, which were used to deter-mine positions of geologic contacts between shale and limestone in Melton Valley.

Table 1. Monthly and annual precipitation at Oak Ridge National Laboratory and normal monthly and annual precipitation at Oak Ridge, Tennessee

[Values in inches]

Year Jan Feb Mar Apr May June July Aug Sept Oct Nov Dec Total 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 2.33 6.30 7.18 5.31 0.87 6.36 1.53 2.97 2.40 0.87 4.89 5.43 1.73 0.93 4.32 1.66 4.69 5.17 4.47 3.43 3.74 4.70 4.95 2.72 5.21 4.41 4.43 8.22 2.93 6.29 2.22 4.10 2.28 2.71 2.44 3.58 7.20 4.19 5.23 4.04 4.17 2.42 6.36 4.09 1.93 1.87 2.47 2.83 2.91 4.35 8.78 3.72 3.38 2.43 5.52 10.29 2.64 2.59 3.03 1.93 6.87 5.44 3.90 0.91 5.82 3.16 2.72 3.71 4.97 1.34 2.84 1.61 2.45 5.71 11.47 2.44 2.58 6.21 2.92 6.85 4.09 3.52 2.03 6.97 2.48 7.60 4.66 1.35 3.07 4.99 0.80 2.78 9.69 5.08 1.62 2.14 8.06 1.79 3.83 2.56 2.81 2.65 2.07 1.16 2.02 2.51 3.83 4.83 4.30 0.47 2.07 1.72 3.84 2.41 4.50 6.10 3.25 4.71 0.69 1.92 7.83 5.38 6.26 4.13 3.12 6.17 4.98 4.97 4.21 4.24 2.10 5.92 3.65 6.87 2.09 1.73 4.05 6.85 6.57 2.05 1.89 4.67 3.25 4.33 55.02 53.44 64.22 37.79 41.33 55.11 44.66 52.50 43.11 38.81 34.14 44.21 Mean 3.87 3.54 4.07 3.90 4.30 3.61 4.77 3.85 3.18 3.00 4.94 4.00 47.03 Normal 1 4.84 4.59 5.67 4.23 4.24 4.03 5.20 3.66 3.80 3.01 4.55 5.44 53.26 1 Normal for period 1951 through 1988 at Oak Ridge, Tennessee. Oak Ridge data computed from U.S.

Department of Commerce (1972), and U.S. Department of Commerce Annual Climatological Summaries for 1971 through 1987 and Monthly Climatological Data for 1988.

Introduction

Haase and others (1987) described the radio-chemical and major-ion water quality in rocks of the lower part of the Conasauga Group in Melton Valley, at depths ranging from 600 to 1,500 feet. Amano and others (1987) reported concentrations of tritium at ORNL, particularly in areas around burial grounds 4 and 5, with samples taken of air, vegetation, surface water, and ground water from depths of less than 5 feet. Solomon and others (1988) described concen-trations of radionuclides, organic compounds, and other water-quality data of leachates from trenches and wells located in burial ground 6.

Moore (1988a) described hydraulic-conductivity values obtained from aquifer tests of wells at ORNL. Moore (1988b) also presented con-cepts of ground-water flow at ORNL, the relative magnitudes of flow in shallow and deep zones, and possible remedial actions to reduce radionuclide trans-port.

Both of these studies by Moore (1988a and 1988b) included statistical analyses of data from Bethel Valley, as well as Melton Valley. Concepts of ground-water flow in the ORNL area presented by Moore (1988b) were later updated and refined with the inclusion of new data in the analyses (Moore, 1989).

Dreier and Toran (1989) presented the construction data from ten 3-well clusters in Melton Valley, and the geological, geophysical, geochemical, and hydraulic data collected at these wells. Tucci and Hanchar (1989) described lithologic, geophysical, and well-construction data from 19 wells drilled in Melton Valley for the present study. Zehner (1989) presented 100 gs UJ O

a. u UJ UJ act:

a.

Figure 3.- Cumulative departure from normal monthly precipitation at Oak Ridge (period 1951 through 1976) and at Oak Ridge National Laboratory (1977 through 1988).

(Oak Ridge data from U.S. Department of Commerce, 1972, and U.S. Department of Commerce Annual Climatalogical Summaries for years 1972 through 1976.)

Hydrology of Melton Valley at Oak Ridge National Laboratory, Tennessee

supplementary construction data for these 19 wells and 9 additional wells.

Dreier and others (1987) presented a comprehensive summary of geologic data, including a geologic map and several geologic sec-tions, for Melton Valley.

Acknowledgments Parts of the surface-water and water-chemistry sections of this report were written by Harold H.

ZehneroftheUSGS.

Thanks are expressed to the following personnel of Martin-Marietta Energy Systems, Incor-porated for their assistance in conducting this project:

Dale Huff, Leroy Stratton, Thomas Row, and Timothy Myrick provided administrative support and informa-tion on ORNL hydrology and burial-site operations.

Laura Toran, RaNaye Dreier, Richard Ketelle, Gerald Moore, D.K. Solomon, and Robert Hatcher provided information from hydrologic and geologic investiga-tions they are conducting. Larry Voorhees provided water-level and well-construction data from the ORNL data base. James Stokely and Thomas Scott performed preliminary radiochemical analyses of water-quality samples as a safety procedure. Person-nel with the Health Physics Division did radiological safety work during drilling operations.

C. Steven Haase provided invaluable information and insights into the geology of Melton Valley and the Oak Ridge area.

METHODS OF INVESTIGATION The hydrologic investigation of Melton Valley included construction of wells and collection of precipitation, streamflow,

geologic, water-level, aquifer-test, and water-quality data. Precipitation data were used for observing ground-water responses to precipitation, and for base-flow analysis. Streamflow data were used to determine the base-flow component of total streamflow, and to estimate infiltration rates from precipitation.

Geologic information was obtained from rock cores and drill cuttings. Surface geophysical data (Tucci, 1987) and subsurface geophysical logs (Tucci and Hanchar, 1989) provided information on subsurface stratigraphy. Water-level data were used to determine the position of the water table, directions of ground-water flow, and magnitude of hydraulic gradients. Hydraulic conductivities were determined from aquifer tests.

Water-quality data were collected and used as a supplement to water-level data in describing ground-water flow. These data were used to help define the hydrology of the valley and to develop ground-water-flow models.

Much of the interpretation is dependent upon the water-level data and data on hydraulic conductivity of the aquifer.

For this reason, a short discussion follows regarding the wells at which data were obtained and the method used for obtaining hydraulic-conductivity values.

Well Construction and Well Identification Approximately 400 wells were drilled at ORNL prior to 1985.

Most are at single-well sites, completed in regolith at less than a 100-foot depth, and open from or near ground level to the bottoms of the wells. Data from these shallow wells were used to determine the position and configuration of the water table.

Five 4-well clusters were completed in burial ground 5 during this period.

The "clusters" referred to in this report are sets of wells about 30 feet apart. Most of the cluster wells at a site have open intervals of 10 to 20 feet at the bottom of the well and were completed at different depths.

Data collected from clusters provided information on the position of the water table and flow at depth, at what is considered one point on a map. The pre-1985 clusters are finished at various depths to a maximum of about 200 feet, and are open in the lower approximately 10 feet. Location and con-struction data for many of these wells are given by Webster and others (1980, 1981, and 1982), and well-construction diagrams are given by Webster and Bradley (1988).

ORNL personnel completed 31 wells at 11 sites in Melton Valley during the period 1986 through 1987. They are called "hydraulic-head measuring stations" or "HHMS wells" by ORNL personnel. Sites are named HHMS1 sequentially through HHMS 11 (fig. 4). Each site has a 3-well cluster, except HHMS 10 which has only one well of 400-foot depth. Approximate well depths at each site are 400 feet, 200 feet, and less than 100 feet. The deepest well has an "A" designation Methods of Investigation

I o

^>>f 9jr 3JI 0.5 1 MILE 0.5 1 KILOMETER EXPLANATION BURIAL GROUND AND NUMBER WELL-CLUSTER SITE UA WELLS DRILLED BY U.S. GEOLOGICAL SURVEY Cluster contains 2 or 3 wells U26 WELLS AUGERED BY U.S. GEOLOGICAL SURVEY HHMS1 HYDRAULIC HEAD-MEASURING STATION DRILLED BY OAK RIDGE NATIONAL LABORATORY Each cluster contains three wells. HHMS10 contains only one well Figure 4. Location of walls.

(HHMS2A or HHMS3A, table 2); the intermediate-depth well has a "B" designation; and the shallowest well has a "C" designation. All are completed open-hole in the lower approximately 20 feet. Specific well-construction measurements and hydrogeologic information for wells HHMS1A through HHMS11A are given by Dreier and Toran (1989). The ORNL has also completed about 400 additional wells during the period 1986 through 1987, most of which are less than 100 feet deep and open to narrow intervals at the well bottoms.

The USGS completed 28 wells during the period 1985 through 1987. They are located at 18 sites in Melton Valley and near the Clinch River (fig. 4). Construction data for the wells are given by Zehner (1989). Nine wells were installed at single-well sites by augering, and are designated U16, U18, U19, U26, U27, U30, U35, U40, and U41. Each of the nine wells is completed in the regolith and has a 3-foot length of screen at its base. The screens are set a few feet below the water table.

The other 19 wells completed by the USGS were drilled by the air-rotary method, and are in eight 2-well clusters (UA, UB, UC, UD, UE, UF, UH, and UI) and one 3-well cluster (UG). Water spray was usually injected during air-rotary drilling. The drilling water was spiked with a benzoate tracer prior to injection into the borehole so that removal of drilling fluid from the well could be verified when water-quality samples were collected after well completion.

Aquifer Tests Most wells at ORNL can be completely emptied of water in a short period by pumping, even at low pumping rates, because of the low hydraulic conductivity of the rocks. Therefore, slug tests were used to estimate transmissivity. The method of analy-sis used in this study is described by Cooper and others (1967), and Papadopulos and others (1973), for response of a finite-diameter well to an instantaneous charge of water.

Conditions assumed for the analysis are: the medium tested is confined, homogeneous, and iso-tropic; the well must fully penetrate the aquifer; recovery of head in the well must be a function of head in the aquifer, and recovery of head must follow according to the differential equation governing nonsteady, radial flow of confined ground water.

Table 2.

Results of slug tests in wells at Oak Ridge National Laboratory Well 1 UA1 UA2 UB1 UB2 UC1 UC2 UD2 UE1 UE2 UF1 UG1 UG2 UG3 UH1 UH2 UH UI2 HHMS1B HHMS1C HHMS2A HHMS2B HHMS2C HHMS3A HHMS3B HHMS3C HHMS4B HHMS5B HHMS5C HHMS6B HHMS6C Depth interval tested (feet) 41.6-50.5 142.0- 169.0 25.9-35.5 101.0-126.1 77.0- 86.2 188.2-206.7 180.0-205.0 69.2-76.7 175.7-197.7 16.5-23.5 25.0- 32.0 242.0 - 300.0 180.0-200.0 19.0- 26.0 231.0-288.0 18.0- 25.0 188.0-210.0 182.3-201.2 63.7-101.0 380.0 - 400.6 180.6-200.6 62.3-81.1 380.5 - 399.1 189.7-211.6 62.0- 80.6 174.3-215.3 196.1 -219.5 42.1 - 63.0 145.0-165.4 40.8-60.8 Transmissivity2 (feet squared per day) 0.57E+00 0.30E-03 0.40E+01 0.37E-01 0.24E+01 0.42E+00 0.23E-03 0.34E+01 0.23E-02 0.17E+02 0.61 E+01 0.28E-03 0.63E-01 0.21 E+01 0.59E-03 0.52E+01 0.34E-02 0.11 E+01 0.30E+01 0.80E-02 0.13E+00 0.72E+00 0.10E-01 0.15E-02 0.78E+00 0.13E+01 0.29E+00 0.34E+01 0.32E+00 0.27E+01 Hydraulic conductivity2 (feet per day) 0.64E-01 0.11E-04 0.42E+00 0.15E-02 0.26E+00 0.23E-01 0.94E-05 0.45E+00 0.11E-03 0.24E+01 0.86E+00 0.49E-05 0.31 E-02 0.30E+00 0.10E-04 0.75E+00 0.15E-03 0.57E-01 0.80E-01 0.39E-03 0.65E-02 0.38E-01 0.54E-03 0.67E-03 0.42E-01 0.32E-01 0.12E-01 0.16E+00 0.16E-01 0.13E+00 1Well locations shown on figure 4.

2E is exponent to base 10.

Probably none of the test conditions are fully met at ORNL, except that recovery of head in the well is a function of head in the aquifer.

For purposes of this study, it was assumed that water is transmitted uniformly through the rocks, and hydraulic conductivity (K) is determined by dividing transmissivity (T) by the length of saturated, open interval in the well. Actually, all of the water may flow into a well at ORNL from one or two narrow fractures exposed for only a few feet along the borehole wall, and flow is in the same direction as the fracture orientation. Moreover, a slug test is sensitive to the condition of the borehole wall, and stresses only a small volume of the aquifer near the well.

The Methods of Investigation

actual aquifer characteristics may be different for an undisturbed, larger volume of rock. Values of T and K obtained from tests at ORNL are considered to be estimates, due to the lack of agreement between field conditions and method assumptions, and to the limitations of the method.

The method is insensitive for determining storage coefficient (S) because the shapes of the curves are similar, particularly at small values of S.

Values of S obtained from slug tests differed as much as three orders of magnitude for the same test because of the inherent uncertainty in matching curves, so these values are not used in the report.

Similar slug-test results were obtained by using a method of analysis described by Hvorslev (1951) for which a confined ground-water system and complete penetration of the aquifer by the well are not required. This method does not account for expansion of the water, however, as does the method described by Cooper and others (1967). Results of the latter method are therefore considered more representative of actual transmissivity values.

Slug tests were conducted by the USGS at 30 wells in Melton Valley subsequent to 1985. Results of these tests (table 2), and of 294 tests conducted by ORNL and by the USGS prior to 1985, were used to model ground-water flow.

GEOLOGIC SETTING Melton Valley and adjacent ridges are underlain by three major stratigraphic units (fig. 5).

Copper Ridge is underlain by the Knox Group of Late Cambrian and Early Ordovician age, and Haw Ridge is underlain by the Rome Formation of Early and Middle Cambrian age. Melton Valley is underlain by the Conasauga Group of Middle and Late Cambrian age (McMaster, 1962). Regolith has developed over these rock units, and alluvial deposits exist near the streams. The geologic structure is complex, and is characterized by major thrust faults and many smaller scale structural features.

Bedrock The Rome Formation consists of interbedded massive sandstone, thinly bedded siltstone, shale, and mudstone (Haase and others, 1985). The lower part of the Rome is more shale rich than the upper part. Total thickness of the Rome is not known because of struc-tural deformation; however, reported thicknesses range from 358 to 540 feet.

The Conasauga Group consists of a complex sequence of clastic-rich units alternating with carbonate-rich units.

Six formations constitute the Conasauga Group (in ascending order): Pumpkin Valley Shale, Rutledge Limestone, Rogersville Shale, Maryville Limestone, Nolichucky

Shale, and Maynardville Limestone (fig. 5). Detailed descriptions of each of these units are given by Haase and others (1985), and descriptions of cores, obtained as a part of this study, from the Pumpkin Valley Shale, Maryville Limestone, and Nolichucky Shale are given by Tucci and Hanchar (1989). The following summary descrip-tion of the formations of the Conasauga Group is from those two sources and from recent geologic mapping by Dreier and others (1987).

The Pumpkin Valley Shale generally consists of thinly interbedded mudstone, shale, and siltstone.

The Pumpkin Valley is informally divided into upper and lower units that are generally lithologically similar, however, the lower unit contains a greater percentage of siltstone than the upper unit. Thickness of the Pumpkin Valley ranges from 310 to 375 feet (Dreier and Toran, 1989, p. 188).

The Rutledge Limestone consists of upper and lower limestone-rich intervals and a central clastic-rich interval that consists mainly of mudstone and shale. The Rutledge ranges from 102 to 147 feet thick in the study area.

The Rogersville Shale is mainly composed of mudstone and siltstone, and ranges from 90 to 158 feet thick. A 3-to 9-foot thick limestone-rich unit is present near the top of the Rogersville throughout the study area.

In the Oak Ridge area, the Maryville Limestone is informally divided into a lower clastic-rich unit and an upper carbonate-rich unit. The lower Maryville Limestone consists mainly of calcareous mudstone with interbedded calcareous siltstone, shale, and lime-stone, and ranges in thickness from 154 to 309 feet.

The upper range in thickness may be overestimated, however, because of repetition of beds from structural deformation (Haase and others, 1985, p. 30).

The upper Maryville Limestone is more carbonate rich than the lower Maryville, and contains some of the purest limestone beds of any of the formations of the Conasauga Group.

Much of the upper Maryville 10 Hydrology of Melton Valley at Oak Ridge National Laboratory, Tennessee

oe oe ? ?

UJ O

o

>0 o o O OQ oe ctz O O <

rh oe m

o I

oI z

a. x z

o UJ o

^

o Xo UJz o

a.il O

I-I 8lggg z ooo 0 <*5£sl -^

x zo^ui^a:

oz oz o

oe 00£ o

o

§ 1 5 > £ c>>

I 3 1.1 ^ >>-

" U'"l 15^

XUJ.>>

o>>- c Q£ II

^"5-o "o I!

m 0.0 2 fco.

t-a.

U) 3 s s X ° o

o o

o UJo*°

°p

UjP, Zc/)

c a>

O I

I in a>

3 O>>

t Geologic Setting 11

Limestone is characterized by a distinctive "flat-pebble" conglomerate (Haase and others, 1985,

p. 30). The thickness of the upper Maryville Lime-stone ranges from 150 to 242 feet.

Most of the Nolichucky Shale consists of repeated cycles of mudstone or shale interbedded with limestone. The formation is divided into upper-shale and middle-carbonate (Bradley Creek Member) units of about equal thickness (35 to 50 feet), and a lower shale unit that becomes more carbonate rich in the lowermost third of the unit. Haase and others (1985,

p. 24) report a thickness of about 460 feet for the lower shale unit, which underlies the southern part of burial ground 6.

Only one well penetrates the Maynardville Limestone in the study area; however, data for this formation from Bear Creek Valley (King and Haase, 1987; Hoos and Bailey, 1986),

approximately 1.5 miles northwest of the study area, provide addi-tional information. The Maynardville is divided into upper and lower carbonate-rich units. The lower unit consists primarily of limestone with interbedded shaley and silty limestone, and the upper unit consists of dolomite and dolomitic limestone.

Haase and others (1985, p. 16) reported the total thickness of the Maynardville beneath Copper Ridge to be 324 feet.

In Bear Creek Valley, the Maynardville ranges from 250 to 450 feet thick.

The Knox Group consists mainly of dolomite with some interbedded limestone and sandstone, and ranges in thickness from about 1,970 to 2,950 feet in the Oak Ridge area (Haase and others, 1985, p. 13).

Data from a deep corehole completed at the north end of Copper Ridge indicate that, in the study area, the Knox Group consists of massive dolomite and lesser amounts of limestone that constitutes the Copper Ridge Dolomite of the lower Knox Group (Haase and others, 1985, p. 15).

Regolith Most of the unconsolidated material that over-lies bedrock consists of moderately to highly weath-ered rock. This weathered rock, commonly referred to as "regolith," consists mainly of clay, silt, and rock fragments.

Webster and Bradley (1988, p. 18-19) report that regolith composition is highly variable, and may locally consist of wet, heavy clay with small pebbles; thin, fissile, weathered shale beds with no clay; and alternating beds of clay and shale. Regolith developed over the Rome Formation consists of sandy, silty clay, and regolith developed over the Knox Group contains abundant chert (Moore, 1988b,

p. 16-17).

The transition from soil to regolith to bedrock is usually gradual, although the transition can be abrupt in carbonate beds (Webster and Bradley, 1988,

p. 19). Competent carbonate beds may be found inter-bedded with weathered shale and clay, particularly in deep regolith. Because of the generally gradual transi-tion from regolith to bedrock, determination of rego-lith thickness (depth to bedrock) is often difficult.

Regolith is generally thinnest in low-lying areas and thickest on the ridges. Webster and Bradley (1988,

p. 19), citing earlier reports, stated that regolith in the waste-disposal areas varies in thickness from a few feet to 40 feet. Well UC1, drilled on Haw Ridge for this investigation, encountered the greatest thickness of regolith yet reported, 86 feet. Regolith developed on the Conasauga Group and Chickamauga Lime-stone, in the valley areas, is generally less than 50 feet thick (Moore, 1988b, p. 20). Near Whiteoak Creek and Melton Branch, regolith thickness is generally less than 10 feet. The geometric mean of regolith thickness, obtained from data for 326 wells in both Melton and Bethel Valleys, is about 13 feet (Moore, 1988b, p. 19). Regolith is generally thicker in Melton Valley than Bethel Valley (Moore, 1988b, p. 19).

Alluvium Alluvium deposited by the Clinch River under-lies the relatively flat area at the southwestern part of the study area (fig. 5). The alluvium consists primar-ily of silty to sandy clay and sand. Average alluvium thickness, estimated from the depths of 15 wells com-pleted in areas underlain by alluvium, is 26.5 feet.

The maximum reported thickness of alluvium in this area is 31 feet (Tucci, 1987, p. 9).

Alluvium also occurs along Whiteoak Creek and Melton Branch, but is not areally extensive. In these areas the alluvium is less than 3 feet thick, and is difficult to distinguish from regolith (Moore, 1988b,

p. 18).

Structure Geologic structure is an important controlling factor in the occurrence and movement of ground 12 Hydrology of Melton Valley at Oak Ridge National Laboratory, Tennessee

water and radionuclides in Melton Valley.

For example, zones of increased hydraulic conductivity associated with faults are pathways for the migration of radionuclides from a former liquid-waste-disposal trench south of burial ground 4 (Olsen and others, 1983, p. 65). Faults also are reported to act as barriers to ground-water flow in the study area (Webster, 1976, p. 16). Open fractures can provide pathways for ground water through otherwise impermeable bedrock.

Several investigators have reported that hydraulic conductivity is greatest in a direction paral-lel to strike in the Oak Ridge area (Webster, 1976; Davis and others, 1984; Rothschild and others, 1984; Smith and Vaughan, 1985; and Tucci, 1986).

The rock units generally strike between 50 and 60 degrees northeast, and average about 56 degrees northeast. Rock units generally dip to the southeast.

Dip is steepest (45 to 90 degrees) near the Copper Creek fault that underlies Haw Ridge (fig. 5), but becomes more gentle (10 to 20 degrees) with distance away from the fault. Dips also become more gentle, and in many cases become horizontal with depth.

Wide local variations in both strike and dip are common in the study area because of small-scale structural complexities.

The study area is located within the Ridge and Valley province of the Appalachian erogenic belt.

This region is characterized by a series of sub-parallel, northeast-trending thrust faults that have broken the bedrock into a series of thrust blocks. The Copper Creek thrust block underlies Melton Valley and adjacent ridges (McMaster, 1962). The dominant geo-logic structure of the study area is the Copper Creek thrust fault that is just below the northwest face of Haw Ridge and forms the base of the Copper Creek thrust block. The fault emplaced the Rome Formation above the younger Chickamauga Limestone.

The Copper Creek fault is a complex structural feature, and is accompanied by wide zones of deformation (Haase and others, 1985, p. 67). Imbricate splays of the fault are common (Dreier and others, 1987; Dreier and Toran, 1989), and one is probably penetrated by well UC2 on Haw Ridge (Tucci and Hanchar, 1989).

The strike of the Copper Creek fault generally follows the strike of the rock units. The fault dips at relatively shallow angles, 5 to 25 degrees, to the southeast (Dreier and Toran, 1989, p. 34).

Several tear faults cross Melton Valley. The most prominent tear fault, the Whiteoak Creek fault (fig. 5), underlies the Whiteoak Creek area and extends through the water gap in Haw Ridge across the valley towards Copper Ridge (Webster and Bradley, 1988, p. 17). Movement along this fault was complex, and displacement of the formations indicates rotation of the rock units beneath burial ground 5 (Dreier and others, 1987, p. 12).

In addition to major thrust and tear faults, other structural features such as bedding-plane faults, intraformational thrust faults, high-angle faults, small-scale folds, and fractures are common. These features have been observed in cores (Haase and others, 1985; Tucci and Hanchar, 1989), trench cuts, and outcrops (Webster and Bradley, 1988, p. 16-17).

Fractures occur throughout the study area, and provide the primary pathways for ground-water flow and contaminant transport through bedrock. Small-scale fractures are particularly common in thicker car-bonate units. The fractures are, in general, partly to completely filled with secondary mineralization within limestone and siltstone, but are more open within mudstone (Haase and others, 1985).

HYDROLOGY OF MELTON VALLEY The hydrology of Melton Valley was studied by analyses of data that relate to the flow of water into and out of the valley. Analysis of the surface-water system is considered first, for from this is derived the rate of ground-water discharge to the streams.

Surface Water Surface-water drainage from ORNL and Mel-ton Valley is by Whiteoak Creek and its tributary Melton Branch (fig. 6). Whiteoak Creek flows into Whiteoak Lake, which is impounded by Whiteoak Dam, then discharges into the Clinch River.

The upper approximately one-half of Whiteoak Creek is in Bethel Valley, most of which is underlain by the Chickamauga Limestone (McMaster, 1962).

The lower half is in Melton Valley, which is underlain by the Conasauga Group.

All of Melton Branch is in Melton Valley.

Data from four gaging stations (table 3) on Whiteoak Creek (Wl through W4) and two gaging stations on Melton Branch (Ml and M2) were used to interpret surface-water characteristics, including base flow from ground water.

Hydrology of Melton Valley 13

o a>

Sewage Treaiment Plant 0.5 1 MILE 0.5 1 KILOMETER EXPLANATION BURIAL GROUND AND NUMBER A M1 STREAMGAGING STATION Figure 6.--Location of streamgagfng stations on Whiteoak Creek and Melton Branch.

Table 3. Streamgaging stations on Whiteoak Creek and Melton Branch, and periods of record and drainage areas above the stations Station desig-nation on map1 W1 W2 W3 W4 W4 M1 M2 Station number 03536320 03536380 03536550 03537000 03537000 03537100 03537500 Station name Whiteoak Creek near Melton Hill Whiteoak Creek near Wheat Whiteoak Cr bl Melton Valley Drive Whiteoak Cr below ORNL Whiteoak Cr below ORNL Melton Branch near Melton Hill Melton Branch near Oak Ridge Period of record (month/year) 4/1987-9/1988 12/1986-9/1988 4/1985-9/1989 6/1950- 6/1953 8/1955-6/1964 4/1985-9/1988 9/1955-6/1964 Drainage area (miles2) 1.31 2.10 3.28 3.62 3.62 0.52 1.48 1 Location shown on figure 6.

Stream Discharge A streamflow-duration curve illustrates the percentage of time a specified stream discharge was equaled or exceeded during a given period. In areas such as the southeastern United States, the curves are usually meaningful only for complete years of record.

Concurrent periods should generally be used for comparing different stations, particularly if the period of record is short and therefore possibly not repre-sentative of flow conditions over a long period.

Logarithmic probability paper is usually used for duration curves because the log scale is suited for the normally large discharge range, and the probability scale expands the distance between data points generally clustered at the ends of the percent scale.

The points are sufficiently spaced for the duration plots in this report, so a linear scale is used for percent exceedance.

Data points are connected by straight lines because most records used here are short term, and interpolated curves probably would not have represented the long-term discharge characteristics at a station.

Data used for duration curves are from stations Wl, W2, W3, and Ml (hereafter referred to simply by the letter and number designations) because they have recent, concurrent record. Data from the 1988 water year are used because it is the only complete year of record for Wl. The duration curves are based on daily-mean discharge and the same class limits as those suggested by Searcy (1959, p. 7). The curves probably are not representative of long-term flow because the periods of record are short and are from a dry period (see precipitation data in table 1 and figure 3). Longer term data from a gage on Poplar Creek (station number 03538225) are used, including that from the 1988 water year, for comparison to the shorter term data from ORNL stations. The Poplar Creek station is located in Poplar Creek Valley about 5 miles north of ORNL, has a drainage area of 82.5 mi2, and a period of record from August 1960 through September 1988. The lower approximately one-half of Poplar Creek is in Poplar Creek Valley, which is underlain mostly by the Conasauga Group (McMaster, 1962).

Considerable flow in the middle and lower reaches of Whiteoak Creek and Melton Branch is effluent discharged from ORNL.

Water used at ORNL is obtained from the city of Oak Ridge, and is from the Clinch River outside the drainage basins of Whiteoak Creek and Melton Branch. This discharge is routed through sewer lines and other discharge lines, and is termed "effluent" in this report. No data are available on effluent flow at ORNL, except what may be estimated from analysis of total stream discharge.

McMaster (1967, p. N5 and N19) esti-mated that about 3.5 ft3/s of effluent from ORNL entered Whiteoak Creek. This value is 48 percent of the approximate 7.3 ft3/s median discharge (daily mean flow exceeded 50 percent of the time) that he illustrated with the duration curve for Whiteoak Creek at Whiteoak Dam for the periods 1953 through 1955 and 1960 through 1963.

Hydrology of Melton Valley 15

The flow duration curves for stations Wl, W2, and W3 on Whiteoak Creek (fig. 7) show the gain in discharge from effluent near the ORNL plant during 1988. Little or no effluent flows past station Wl, where median flow was 0.096 ft3/s. Daily-mean flow at Wl was 0.02 ft3/s during less than 2 percent of the year, and less than 0.04 ft /s during 10 percent of the year. Station W2 has a significantly greater discharge than Wl, particularly during periods of low flow. The median discharge at W2 during the 1988 water year was 3.1 ft3/s, and the minimum was 1.7ft3/s. The duration curve for W2 is flatter than that for Wl, and 20 to 99 percent of the exceedance is within the corre-sponding narrow range of 4 to 2 ft /s discharge. The flat slope indicates effluent discharge is sustaining streamflow at W2.

Most of the effluent discharged to Whiteoak Creek is apparently from the sewage treatment plant, located on the north side of Haw Ridge between stations W2 and W3 (fig. 6). Station W3 is located about 3,000 feet downstream of W2. The median daily-mean discharge at W3 was 6.5 ft3/s, and the minimum was 4.3 ft3/s during the 1988 water year (fig. 7a). The slope of the duration curve from station W3 is similar to that for W2, and 20 to 97 percent of the daily-mean flows are in the range of 8.5 to 5 ft3/s discharge. The data for W3 also indicate that much of the streamflow is from effluent discharge, and the effluent accounts for nearly 100 percent of the stream-flow during periods of low streamflow. The differ-ence in median discharge for stations W2 and W3 is 3.4 ft3/s. Most of the difference represents a gain in streamflow from effluent between these stations, and most of the gain is probably from the sewage treat-ment plant.

Natural runoff contributes to the gain in dis-charge between stations Wl and W3 because the drainage area increases downstream. Natural runoff per unit drainage area should be about the same at station Wl as that at station W3, which is about 6,500 feet downstream from Wl, because the drainage areas for the stations are within Bethel Valley and have virtually the same rock strata.

The duration curves of discharge per unit drainage area for stations W2 and W3 are significantly different from the curve for Wl (fig. 7b).

The differences are due to the effluent discharged from ORNL between the two stations.

The average effluent discharge at stations W2 and W3 can be visually estimated from discharge hydrographs. The discharge hydrograph for the 1988 water year at station Wl (fig. 8) illustrates a normal pattern of discharge without effluent. Rapid increases in discharge correspond to precipitation events. Grad-ual recessions, and generally lower discharge, occur during the growing season (April to November) as evapotranspiration increases. The natural part of the recessions at W2 and W3 are incomplete, due to effluent sustaining streamflow at about 2.6 ft^/s at W2 and 5.4 ft3/s at W3. Effluent discharge is not con-stant, and values below the estimated average often correspond to weekends when less water is used.

The percentage of time all discharge was efflu-ent, and percentage of time effluent exceeded natural discharge at stations W2 and W3 during the 1988 water year, were determined from the visually esti-mated average effluent discharge (fig. 8) and the flow-duration curves (fig. 7). At station W2, discharge was all effluent (equal to or less than (<) 2.6 ft3/s) about 30 percent of the time and more than half effluent

(<5.2 ft3/s) about 90 percent of the time. At station W3, discharge was all effluent (<5.4 ft3/s) about 16 percent of the time, and more than half effluent

(<10.8 ft3/s) about 93 percent of the time.

Stations Ml and M2 are located on Melton Branch (fig. 6). Station Ml is upstream of the affects of effluent discharged from the ORNL facilities.

Station M2 receives effluent from two upstream tributaries.

The tributaries are adjacent to three ORNL facilities (not shown on fig. 6) located on the southeast side of Melton Valley Drive (fig. 1). Strict comparisons cannot be made of duration curves or hydrographs for stations Ml and M2 because the periods of record are not concurrent, and the period of record for station Ml is short, encompassing only three complete water years.

Estimates of average annual effluent discharge above station M2 were made visually from hydrographs, however, as was described for stations W2 and W3 on Whiteoak Creek.

The duration curve for the 1988 water year at Ml (fig. 9) on Melton Branch shows a median dis-charge per unit drainage area about three times greater than the median discharge for Wl on Whiteoak Creek. These are the only stations in the ORNL area that do not reflect effluent discharge. Data for these stations represent natural conditions. Data from the station in Poplar Creek Valley near Oak Ridge also represent natural conditions. The duration curves for the station in Poplar Creek Valley and station Ml in Melton Valley have similar values for the 1988 water 16 Hydrology of Melton Valley at Oak Ridge National Laboratory, Tennessee

100 r:

0.1 0.01 A. Duration curves in cubic feet per second

+

STATION W1 A

STATION W2 O

STATION W3 I_____l_____l 10 QC UJ Q_ UJ

_J

£]2 UJ U. UJ QC 0.1 0.01 0.001 B. Duration curves in cubic feet per second per square mile of drainage area 10 20 30 40 50 60 70 80 90 PERCENT OF TIME DAILY-MEAN DISCHARGE WAS EQUALED OR EXCEEDED 100 Figure 7.--Duration curves of daily-mean discharge at three stations on Whiteoak Creek at Oak Ridge National Laboratory, water year 1988.

Station W1 is Whiteoak Creek near Melton Hill, W2 is Whiteoak Creek near Wheat, and W3 is Whiteoak Creek below Melton Valley Drive.

Hydrology of Melton Valley 17

100 STATION W1 STATION W2 STATION W3 Estimated average effluent discharge 0.01 1

15 29 12 26 10 24 7

21 4

18 3

17 31 14 28 12 26 9

23 7

21 4

18 1

15 29 8

22 5

19 3

17 31 14 28 11 25 10 24 7

21 5

19 2

16 30 14 28 11 25 8

22 OCT NOV 1987 DEC JAN FEB MAR APR MAY 1988 JUNE JULY AUG SEPT Figure 8. Hydrographs of daily-mean discharge at three stations on Whiteoak Creek (Wl, near Melton Hill; W2, near Wheat; and W3, below Melton Valley Drive) at Oak Ridge National Laboratory, water year 1988, with estimated average effluent discharge.

year, possibly because both valleys are underlain mostly by the Conasauga Group. The drainage area above station Wl is underlain mostly by the Chicka-mauga Limestone. Discharge during the 1988 water year was probably less than normal at all three stations, as shown by the duration curves for the station on Poplar Creek. The median discharge during the 1988 water year at Poplar Creek was 0.24 (tf/s)/

mi2, 70 percent less than that during the period of record from 1961 through 1988.

The reasons for differences in discharge per unit drainage area between stations Ml and Wl are not known. Topography does not seem to be a major factor, drainage areas above both stations contain about equal percentages of hillside and valley bottom.

The differences may be due to more ground-water discharge from the regolith in Melton Valley than from the regolith in Bethel Valley or to differences in ground-water discharge from bedrock in the two valleys.

The differences may also reflect different periods of record for the two stations. Discussions of these possibilities follow.

Regolith generally is more permeable than bedrock at ORNL. A greater thickness of the regolith in Melton Valley could therefore cause more ground-water discharge to Melton Branch than is discharged to Whiteoak Creek from the thinner regolith and bedrock in Bethel Valley. If hydraulic conductivity and thickness of the regolith were the reasons for the differences in streamflow, however, a uniform 18 Hydrology of Melton Valley at Oak Ridge National Laboratory, Tennessee

100 O

POPLAR CREEK, WATER YEAR 1988 A

POPLAR CREEK, WATER YEARS 1961-88

+

STATION W1, WATER YEAR 1988 X

STATION M1, WATER YEAR 1988 0.01 10 20 30 40 50 60 70 80 PERCENT OF TIME DAILY-MEAN DISCHARGE WAS EQUALED OR EXCEEDED 90 100 Figure 9.--Duration curves of daily-mean discharge at stations on Whiteoak Creek and Melton Branch at Oak Ridge National Laboratory, water year 1988, and Poplar Creek near Oak Ridge, water years 1961 through 1988.

W1 is Whiteoak Creek near Melton Hill and M1 is Melton Branch near Melton Hill.

distribution of rainfall in both valleys would produce greater overland runoff and higher flow in Whiteoak Creek than in Melton Branch because infiltration would be less in Bethel Valley. The streamflow data do not show this to be the case, however.

Throughout much of Melton Valley the water table frequently is near, or below, the base of the regolith, so most saturated strata are weathered and unweathered bedrock. Therefore, regolith thickness could be less of a factor than hydraulic conductivity and saturated thickness of the bedrock in causing the differences in ground-water discharge in Bethel and Melton Valleys. If rocks of the Chickamauga Lime-stone yield more water in Bethel Valley than the rocks of the Conasauga Group yield in Melton Valley (Webster and Bradley, 1988, p.97-100), the expected discharge relation would be opposite that shown in figure 9. However, Moore (1988b) stated that most ground-water flow at ORNL is through macropores, and that difference between frequency of macropores in units of the Conasauga Group and units of the Chickamauga Limestone is not significant. If the flow through macropores was not significantly different in the two valleys, ground-water discharge in Melton and Bethel Valleys would be similar.

Although base-flow data indicate that discharge is apparently greater in Melton Valley than in Bethel Valley, the data are insufficient to determine the rela-tion, or the causes of the relation, with certainty. A longer period of record at stations Wl and W2, perhaps having a more uniform distribution of pre-cipitation, might clarify the discharge relation in the two valleys. Discharge data discussed in this section are summarized in table 4. A summary of discharge Hydrology of Melton Valley 19

Table 4. Summary of discharge data for stations on Whiteoak Creek and Melton Branch, water year 1988, and Poplar Creek, water years 1988 and 1961 through 1988

[Q, daily-mean discharge; DA, drainage area; f^/s, cubic feet per second; ftVs/mi2, cubic feet per second per square mile]

Station desig-nation on map1 W1 W2 W3 M1 P

Median Q2 (ftVs) 0.096 3.1 6.5 0.13 19.8 Estimated effluent Q (ft3/s) 0 2.6 5.4 0

0 Median Q/DA for 1988 (ft3/s/mi2) 0.073 1.5 2.0 0.25 0.24 Median Q/DA for 1961 -88 (ft3/s/mi2) 0.81 Estimated time effluent Q exceeded natural Q (percent) 0 90 93 0

0 1W1 is Whiteoak Creek near Melton Hill, W2 is Whiteoak Creek near Wheat, W3 is Whiteoak Creek below Melton Valley Drive, M1 is Melton Branch near Melton Hill, and P is Poplar Creek near Oak Ridge. Locations, except Poplar Creek station, are shown on figure 6.

2 Determined from flow-duration curves.

data by water years at the stations described in this report is presented in table 5. This table includes esti-mates of effluent discharge at the stations as deter-mined by visual inspection of the hydrographs.

Base Flow Seepage runs (series of discharge measure-ments made simultaneously or within a brief period along short reaches of stream length) were made on Whiteoak Creek and Melton Branch in August 1985, November 1985, and June 1988. The purpose of these seepage runs was to locate gaining and losing reaches along the streams, and to measure the base flow in these reaches. The data did not prove useful because most streamflow was effluent from ORNL facilities, and measurement error was larger than the apparently small gains and losses in flow. Base flow in these streams was therefore estimated by use of a hydrograph-separation method.

Most base flow at ORNL is probably from the regolith and shallow bedrock, but some could be from deeper bedrock.

Moore (1988b, p. 35) discussed ground-water flow through the regolith at ORNL, and stated that stormflow, the water that infiltrates and flows through the upper 3 to 6 feet of land surface, has transit times of a few days to a few weeks before discharging to local drainages. Moore (1988b, p. 39) also stated that stormflow in local areas "is about 95 percent of total surface-water discharge at the time a hydrograph peaks."

Base-flow values in this report probably include much of the storm flow described by Moore (1988b). However, when stormflow amounts to as much as 95 percent of total discharge, most of the dis-charge is considered in this report as surface runoff, rather than base flow, because the base-flow reces-sions occur after the flood peaks on hydrographs. No relations are established in this section regarding rapid and slow, or shallow and deep, ground-water dis-charge.

The hydrograph-separation method is described by Rorabaugh (1964), and is based on the following equation for ground-water discharge to a stream:

q = 2T(ho/a)(e -n2Tt/4a2S + e-9n2Tt/4a2S where e -25nTt/4a2S is the ground-water discharge per unit of stream length on one side of the stream; is time after an instantaneous water-table rise of HO at time t$

is transmissivity; is the height of an instantaneous rise in the water table; is storage coefficient, and is distance from the stream to the ground-water divide.

20 Hydrology of Melton Valley at Oak Ridge National Laboratory, Tennessee

Assumptions for the equation are that the drain-age basin has uniform, homogeneous, and isotropic characteristics; distances from the stream to hydrologic boundaries are equal at all places in the basin; and the initial ground-water level is at stream level.

When (7Y)/(fl2S) is greater than 0.2, all terms except the first in the series are negligible, and equation (1) reduces to:

(2)

A semi-log plot of ground-water discharge resulting from an instantaneous rise in ground-water level will become linear when time tc = Q.2a2S/T, where tc is the critical time. At critical time, one-half the ground-water volume from a recharge event will have discharged to the stream (Glover, 1964).

Rorabaugh (1964, p. 440) integrated equation 2 with respect to time from t = tc to t = infinity to obtain the volume V in storage on one side of the stream, as V =

q(4a2S/n2T). When evaluating this equation at critical time, and considering that (a) recharge to the aquifer Table 5. Summary of annual discharge data, including effluent part, in cubic feet per second, for all complete years of record at stations on Whiteoak Creek and Melton Branch

[Q, discharge]

Station desig-nation on map1 W1 W2 W3 W3 W3 W4 W4 W4 W4 W4 W4 W4 W4 W4 W4 M1 M1 M1 M2 M2 M2 M2 M2 M2 M2 M2 Water year 1988 1988 1986 1987 1988 1951 1952 1956 1957 1958 1959 1960 1961 1962 1963 1986 1987 1988 1956 1957 1958 1959 1960 1961 1962 1963 Minimum daily Q

0.01 1.7 5.8 5.3 4.3 2.4 3.0 3.8 3.2 4.0 2.6 3.0 3.2 3.6 4.0 0.0 0.0 0.0 0.0 0.1 0.3 0.3 0.2 0.3 0.0 0.1 Daily mean Q

0.42 3.77 8.12 9.01 7.47 9.79 8.53 9.25 10.2 10.7 7.75 8.54 9.76 11.5 10.2 0.33 0.47 0.25 1.90 2.46 3.07 1.69 2.06 2.96 3.24 2.62 Maximum daily Q

18 36 63 83 58 221 96 93 148 130 121 110 113 121 204 14 23 17 46 55 55 53 23 56 66 99 Daily Total effluent annual component2 of Q

Q 153 1,380 2,960 3,290 2,730 3,570 3,120 3,390 3,730 3,890 2,830 3,120 3,560 4,180 3,710 119 171 90.8 694 898 1,120 618 755 1,080 1,180 955 0

2.6 6.7 6.0 5.4 3.1 4.0 5.0 5.0 5.0 5.1 5.0 4.5 5.0 5.0 0

0 0

0.1 0.2 0.8 0.7 0.3 0.3 0.1 0.3 1W1 is Whiteoak Creek near Melton Hill, W2 is Whiteoak Creek near Wheat, W3 is Whiteoak Creek below Melton Valley Drive, W4 is Whiteoak Creek below ORNL, M1 is Melton Branch near Melton Hill, and M2 is Melton Branch near Oak Ridge. Locations are shown on figure 6.

2Average daily-mean, determined visually from hydrographs.

Hydrology of Melton Valley 21

is discharged to the stream, (b) the natural log is converted to base 10 log, and (c) the volume is doubled to include discharge from both sides of the stream, the equation becomes Q = 2q(0.933a2S/T)/

2.30, where Q is the total discharge from the aquifer and q is the stream discharge at critical time. The quantity 0.933a2S/r represents the inverse slope, as time per log cycle of discharge, of the discharge recession after critical time.

The hydrograph separation for station Ml on Melton Branch near Melton Hill (fig. 10) illustrates how this method is used to determine base flow of streams in the ORNL area. The period shown is the part of the 1986 water year which had most of the annual base flow. No effluent is discharged above this station, and the data represent natural runoff con-ditions.

The recession slope of one log cycle per 18 days is determined from the recessions in late Feb-ruary and early March. Critical time is tc = Q.2a2S/T, which is 0.2/0.933 times the inverse slope of the reces-sion in days per log cycle, or (0.214)(18 days) =

4 days. Other recharge-discharge events occur before completion of a single base-flow recession (fig. 10).

Increments of each recession are cumulated, but the 100 10 TOTAL FLOW BASE FLOW Upper dot denotes base flow value at critical time, and lower dot denotes base-flow at corresponding time on previous recession DISCHARGE Critical time =

4 days n

0.1 Slope =

one log cycle per 18 days Total base flow =

4

[2(0.064 ft3/s - 0 ft3/s)(l8 days)]/2.30 +

[2^0.10 ft3/s - 0.02 ft3/s)(l8 days)]/2.30 +

...... + [2(q n+1 - q n ft3/s)(l8days)]/2.30 0.01 i \\! i i

i i

i i

i i

i i

i i

i I

I I

I I

I I

I I

I I

I 17 24 31 7 14 21 28 5 12 19 26 2 OCTOBER NOVEMBER DECEMBER 1985 9

16 23 30 6 13 20 27 6 JANUARY FEBRUARY 13 20 27 3 MARCH 1986 10 17 24 1

APRIL Figure 10.

Stream-discharge hydrograph and base-flow component at station M1, Melton Branch near Melton Hill. Slope value is for linear parts of base-flow recessions.

22 Hydrology of Melton Valley at Oak Ridge National Laboratory, Tennessee

part of a preceding event that is concurrent with a succeeding event is subtracted before cumulation.

Base flow from the event prior to the first event is less than 0.01 ft3/s (fig. 10), considered to be negli-gible, and set equal to zero. The slope of the first event becomes linear on October 28 at tc - 4 days after the peak discharge on October 24. Base flow from the first event is computed as (2)(<72 - <?i)/2.30 times the inverse slope in days per log cycle of dis-charge, or [(2)(0.064 tf/s-O fr/s)(18 days)]/2.30=

1.00 ft3/s-d.

Base flow from the second event is

[(2)(0.10 ft3/s-0.02 ft3/s)(18 days)]/2.30=1.25 ft3/s-d.

Each base-flow increment is computed in the same manner, and all increments are summed to determine (2, the total base flow.

Base-flow values for ORNL streams are esti-mates because the physical conditions do not meet the simplifying assumptions in the method of analysis, and, more importantly, stream discharge must be adjusted for effluent discharge at most stations, which causes inaccuracy in the analysis.

Regolith and bedrock are not uniform in thickness, homogeneous, or isotropic. Distances from the streams to ground-water divides are not equal at all places. The initial ground-water level is not at stream level, except perhaps in the narrow flood plains adjacent to the streams.

The rise in ground-water level is not instantaneous. Inaccuracy due to effluent discharge will be addressed later in this section.

Table 6. Annual base flow at stations on Whiteoak Creek and Melton Branch that are not affected by effluent discharge

[f^/s-d, cubic feet per second times days; f^/s-d/mi2 qubic feet per second times days per square mile]

Station desig-nation on map1 W1 M1 M1 M1 Water year 1988 1986 1987 1988 Precipi-tation (inches) 38.08 34.54 41.72 38.08 Base flow (ft3/s-d) 32 62 62 20 Base flow per unit drainage area (ft3/s-d/mi2) 24 119 119 38 Base flow to total discharge (percent) 21 52 36 22 Base flow to precipi-tation2 (percent) 2.4 13 11 3.8 M1 Mean 48 92 37 1Stations and drainage areas: W1 is Whiteoak Creek near Melton Hill, 1.31 mi2; M1 is Melton Branch near Melton Hill, 0.52 mi2. Locations are shown on figure 6.

Computed as base flow per unit drainage area expressed as inches per year divided by precipitation.

The best estimates of base flow from records at ORNL stations are those for Wl and Ml (table 6) because no effluent is discharged upstream.

Base flow varies by a factor of about three during the short 3-year period of record at Ml, probably because the rate of base-flow discharge is dependent on the rate, duration, and frequency of precipitation, and time of year of occurrence. Analysis of a longer period, dur-ing which precipitation is nearer the annual 53-inch normal, would probably yield larger base-flow values.

The data must be adjusted to obtain estimates of base flow for stations which are downstream of effluent discharge.

Visual inspection of the hydro-graphs were made to determine an average daily-mean effluent flow for each water year, as explained in the previous section. The effluent component was sub-tracted from each daily-mean value of total discharge, and the adjusted values were plotted. Negative values were set equal to 0.01 when plotting the hydrograph.

The few negative values were due to the average effluent value exceeding total discharge some of the time. Base-flow separation was then completed on the adjusted hydrograph by the method described above. Although the effluent component was differ-ent for each year, the slope of the recession at a sta-tion was held constant for all years.

Inaccuracy is introduced in the base-flow analy-sis by adjusting total stream discharge for effluent flow. This inaccuracy is due to differences between the slope on the pre-adjusted hydro-graph and the slope on the adjusted hydrograph.

The slope on the adjusted hydrograph may or may not represent the natural ground-water recession.

Cumulation of base-flow increments from the adjusted hydro-graph which are much less than the original streamflow can also contrib-ute to the inaccuracy.

The hydrograph of discharge at station Wl on Whiteoak Creek is different than hydrographs of unad-justed and effluent-adjusted discharge at station W3 (fig. 11). Although the drainage-area difference between the stations is less than 2 mi2, the slopes of the ground-water recessions should be similar because both areas are in Bethel Valley, and they drain similar strata.

The discharge recessions on 9.3 Hydrology of Melton Valley 23

100 STATION W1, UNADJUSTED Slope =

1 log cycle per 20 days 1

8 15 22 29 6

13 20 27 3

10 17 24 31 7

14 21 28 6

13 20 27 3

10 17 24 1

8 15 22 29 NOVEMBER DECEMBER JANUARY FEBRUARY MARCH APRIL MAY 1987 1988 Figure 11.--Adjusted and unadjusted stream-discharge hydrographs of Whiteoak Creek below Melton Valley Drive (station W3) and unadjusted stream-discharge hydrograph of Whiteoak Creek near Melton Hill (station W1).

the adjusted hydrograph for station W3 more closely approximate those for station Wl than do the effluent-maintained discharge recessions on the unadjusted hydrograph. However, the linear parts of the recessions (ground-water discharges after critical time) are steeper on the adjusted hydrograph for W3 than for station Wl. The slope is one log cycle per 11 days for station W3 (adjusted), whereas it is one log cycle per 20 days for station Wl. An accurate ground-water recession slope possibly cannot be determined for W3 because it is masked by effluent discharge.

The inability to determine the shape of the natural hydrograph at W3 causes uncertainty in positioning the recession slope at time tc and selection of q\\ and q2 values to be used in the analysis. Most discharge values are only several tenths of a cubic foot per second, and this is small compared to the unadjusted flow of several cubic feet per second. The effluent discharge masks the natural discharge at W3, and other effluent-affected stations, to such an extent that base-flow values from these stations can only be general estimates.

Estimates of base flow were obtained from effluent-adjusted hydrographs for stations W3, W4, and M2 (table 7). Although accuracy of these esti-mates is probably less than the accuracy of values computed from the discharge data for stations Wl and Ml, they are given here because the period of record is longer. Precipitation was greater during these longer periods than during the periods analyzed for stations Wl and Ml. The estimates of base flow at station M2 are probably more accurate than those for stations W3 and W4 because effluent in Melton Branch is a smaller part of total discharge (table 5) and therefore does not mask the recessions as much.

24 Hydrology of Melton Valley at Oak Ridge National Laboratory, Tennessee

Table 7. Estimates of annual base flow at stations on Whiteoak Creek and Melton Branch that are affected by effluent discharge

[ftVs-d, cubic feet per second times days; mi2, square miles]

Base flow, Base flow Base flow as a Station per unit to adjusted percent of desig-Precipi-Adjusted drainage dis-precipi-nation Wate tation base flow area charge2 tation3 on map1 year (inches)

(ft3/s-d)

(ft3/s-d/mi2)

(percent)

(percent)

W3 W3 W3 W4 W4 W4 W4 W4 W4 W4 W4 W4 M2 M2 M2 M2 M2 M2 M2 M2 W3 Mean W4 Mean M2 Mean 1986 1987 1988 1952 1956 1957 1958 1959 1960 1961 1962 1963 1956 1957 1958 1959 1960 1961 1962 1963 34.54 41.72 38.08 47.89 63.03 62.59 53.00 41.40 57.69 55.55 66.24 49.44 63.03 62.59 53.00 41.40 57.69 55.55 66.24 49.44 252 499 308 882 694 605 884 366 531 827 1,010 752 194 253 270 129 298 432 367 247 350 730 270 77 152 94 244 192 167 244 101 147 228 279 208 131 171 182 87 201 292 248 167 110 200 180 49 45 41 53 45 32 43 38 41 43 43 40 30 31 33 36 46 44 32 29 45 42 35 8.3 14 9.2 19 11 9.9 17 9.1 9.5 15 16 16 7.7 10 13 7.8 13 20 14 13 11 14 12 1 Stations and drainage areas: W3 is Whiteoak Creek below Melton Valley Drive, 3.28 mi2; W4 is Whiteoak Creek below ORNL, 3.62 mi2; and M2 is Melton Branch near Oak Ridge, 1.48 mi2. Locations are shown on figure 6.

2Adjusted discharge is total discharge minus effluent discharge. Total annual discharge and effluent discharge are given in table 5.

3Compuled as base flow per unit drainage area expressed as inches per year divided by precipitation.

The approximate error in base flow at Whiteoak Creek stations, due to adjustment for effluent dis-charge, can be calculated by comparison of a station having natural discharge to a station which has been adjusted for effluent flow. A necessary assumption is that the correct base flow per unit drainage area is the same at the stations compared. No effluent discharges upstream from Wl, and base flow is 24 (ftVs-dymi2 for the 1988 water year (table 6).

The effluent-adjusted value at W3 is 94 (ft3/s-d)/mi2 for the same water year (table 7), which is about four times greater than the base flow at Wl.

A comparison can also be made for stations Ml (no effluent discharge) and M2 (includes effluent discharge) on Melton Branch. It is less accurate than the comparison of the stations on Whiteoak Creek, however, because the period of record at Ml is different than the period at M2.

Years of similar precipitation and similar discharge per unit drain-age area are therefore compared for the periods of record on Melton Branch.

The years with the most similar annual precip-itation totals are 1987 (41.72 inches) at Ml (table 6),

and 1959 (41.40 inches) at M2 (table 7). Base flow at Hydrology of Melton Valley 25

Ml was 119(ft3/s-d)Ani2 in 1987, which is about 30 percent more than the base flow of 87 (ft3/s-d)/mi2 at M2 in 1959. Drainage area is 0.52 mi2 at Ml and 1.48 mi2 at M2. As computed from the drainage areas and the data in table 5, the years with the most similar discharges are 1986 at Ml [(118.70 ft3/s-d)/0.52 mi2

= 228 (ft3/s-d)/mi2] and 1959 at M2 [((618 ft3/s-d) -

(0.7 ft3/s)(365 days))/1.48 mi2 = 245 (ftVdVmi2].

Base flow at Ml during 1986 was 119 (ft3/s-d)/mi2 (table 6). The base flow for M2 in 1959 was 87 (f^/s-d)/mi2, so the same 30 percent value also ^applies to the case based on similar discharge.

The base flow on Whiteoak Creek appears to be about four times too large when discharge data are cor-rected for effluent, whereas the base flow on Melton Branch appears to be about 30 percent too small when corrected for effluent. The consistency of the errors is not known because (1) the errors are based only on comparison of data for short periods of record during a dry period and (2) the number of stations not affected by effluent is insufficient to determine if the base flow per unit drainage area is uniform along the streams. As stated earlier in this section, the most accurate base flow values are those given for stations not affected by effluent (table 6), and values from other stations are presented only as general estimates.

The base flow for effluent-affected stations was adjusted by the error factors described above in determining the relation of annual base flow to precipitation at both natural and effluent-affected stations (fig. 12), and the relation of annual base flow to annual stream discharge at these stations (fig. 13).

The relations are considered to be general estimates.

Based on the relation of annual base flow and precipitation (fig. 12), a year of normal precipitation (53 inches) would produce an annual base-flow dis-charge per square mile of drainage area of about 60 ft3/s-d on Whiteoak Creek and about 200 ft3/s-d on Melton Branch.

Base flow increases as precipitation and dis-charge increase, as expected, but variation in these relations is large for different years (figs. 12 and 13).

The data consistently show greater base flow for Melton Branch than for Whiteoak Creek, as do the duration curves of total discharge.

Ground Water The ground-water-flow system of Melton Valley is complex, reflecting the variation of sub-surface lithologies and the structural complexity of the rocks. Tucci (1986, p. 7) and Webster and Bradley (1988, p. 18) conceptualized this complex system as two hydrogeologic units, regolith and bedrock. That concept also is used in this report; however, the bedrock part of the flow system is further subdivided into shallow, intermediate, and deep hydrogeologic units. These units do not represent separate aquifers, but are based primarily on differences in hydraulic conductivity, overall mode of ground-water flow, and amount of ground water in circulation within each unit. Ground-water flow within the "stormflow" zone just below land surface (Moore, 1988b, p. 35) and the vadose zone just above the water table is not considered in this report.

Although ground-water flow in the stormflow zone can be substantial (Moore, 1988b, p. 95), this flow is a relatively transient feature of the ground-water-flow system.

Regolith The regolith hydrogeologic unit, as distin-guished from the regolith geologic unit, consists of both regolith and the uppermost partly weathered bedrock at depths of 50 feet or less below land surface.

Alluvium near the Clinch River also is included in this unit. A depth of 50 feet was chosen as the base of the regolith hydrogeologic unit on the basis of plots of hydraulic conductivity and well depth (figs. 14 and 15), which indicate that hydraulic conductivity tends to be greatest in the upper 50 feet, and tends to decrease with increasing depth.

Hydraulic Conductivity Hydraulic conductivity of the regolith hydro-geologic unit is highly variable and ranges from 6.6 x IO"4 to 6.9 ft/d (fig. 14). The median regolith unit hydraulic-conductivity value, determined from the results of 228 aquifer tests conducted in the valley, is 0.19 ft/d; the mean value is 0.53 ft/d.

Nearly all regolith unit hydraulic-conductivity values are greater than 0.01 ft/d.

Most hydraulic-conductivity values were obtained by slug-test analyses conducted by ORNL (O.K. Moore, University of Tennessee, written commun., 1989). Slug tests also were conducted at 60 wells completed in regolith by Webster and Bradley (1988). Hydraulic-conductivity values estimated from these tests range from 2.9 x IO"3 to 6.7 ft/d (Webster and Bradley, 1988, p. 34-49). Slug tests, discussed in 26 Hydrology of Melton Valley at Oak Ridge National Laboratory, Tennessee

the "Aquifer Tests" section, were conducted for this study at eight wells completed in the regolith hydrogeologic unit. Although two of the wells were completed at depths greater than 50 feet, the wells terminated in regolith and hydraulic-conductivity data from those wells are included in values for the regolith. Calculated hydraulic-conductivity values from the eight tests range from 6.4 x 10~2 to 2.4 ft/d (table 2).

Median hydraulic-conductivity values are similar for regolith developed on most geologic units in the study area; however, regolith developed on the Pumpkin Valley Shale generally is less permeable than that developed on other units (fig. 16). Regolith developed on the Rome Formation appears to be more permeable than that developed on other geologic units; however, only two wells are completed in regolith developed on the Rome Formation so that this conclusion may not be valid for the entire valley.

Hydraulic-conductivity data are not available for regolith developed on the Maynardville Limestone or the Knox Group in Melton Valley. Results of slug tests conducted in Bear Creek Valley provide some information on hydraulic-conductivity values for the Rome Formation, Maynardville Limestone, and the Knox Group at depths of less than 50 feet. Median hydraulic-conductivity values are 0.38 ft/d for the Rome Formation (3 tests), 3.02 ft/d for the May-nardville Limestone (12 tests), and 0.17 ft/d for the Copper Ridge Dolomite of the Knox Group (6 tests)

(Z. Bailey, U.S. Geological Survey, written commun.,

1,000 52 100 OD<

<1 18 10 a

X A

30 STATION W1. WHITEOAK CREEK NEAR MELTON HILL, WATER YEAR 1988 STATION W3, WHITEOAK CREEK BELOW MELTON VALLEY DRIVE, WATER YEARS 1986-88 STATION W4, WHITEOAK CREEK BELOW ORNL, WATER YEARS 1952 AND 1956-63 STATION Ml, MELTON BRANCH NEAR MELTON HILL, WATER YEARS 1986-88 STATION M2. MELTON BRANCH NEAR OAK RIDGE, WATER YEARS 1956-63 35 40 45 50 55 PRECIPITATION, IN INCHES 60 65 70 Figure 12. Relation between annual precipitation and adjusted annual base flow at stations on Whiteoak Creek and Melton Branch at Oak Ridge National Laboratory (ORNL).

Base flow adjusted by correction factor (see text).

Water years inclusive.

Hydrology of Melton Valley 27

1989). Because of the small number of tests for each formation, these values may not be representative of the shallow zones of the formations; however, they are probably representative within an order of mag-nitude.

Hydraulic-conductivity values of the May-nardville Limestone are probably greater in Bear Creek Valley than in Melton Valley.

Bear Creek flows directly on the Maynardville Limestone in Bear Creek Valley, so that water movement through that formation has formed many solution openings that increase the permeability of the rocks.

In Melton Valley, the Maynardville Limestone occurs on the flanks of Copper Ridge above Melton Branch, and solution openings within the unit may not be as abundant as in Bear Creek Valley.

The degree of anisotropy, in which hydraulic conductivity is greater parallel to strike than normal to strike, of the regolith hydrogeologic unit is specu-lative. Tucci (1986, p. 7) reported that an anisotropy ratio of 1:3 (strike-normal to strike-parallel) produced the best results in a preliminary ground-water-flow model of Melton Valley, but that simulating hydraulic 1,000 53 3

0£

~a o

§K uj.

V) U)

< V m <

o a

X I-A 100 10 STATION W1. WHITEOAK CREEK NEAR MELTON HILL, WATER YEAR 1988 STATION W3, WHITEOAK CREEK BELOW MELTON VALLEY DRIVE, WATER YEARS 1986-88 STATION W4, WHITEOAK CREEK BELOW ORNL, WATER YEARS 1952 AND 1956-63 STATION Ml, MELTON BRANCH NEAR MELTON HILL WATER YEARS 1986-88 STATION M2, MELTON BRANCH NEAR OAK RIDGE, WATER YEARS 1956-63 100 200 300 400 500 600 700 800 900 1,000 ADJUSTED ANNUAL STREAM DISCHARGE, IN CUBIC FEET PER SECOND-DAYS, PER SQUARE MILE Figure 13.--Relation between adjusted annual stream discharge and adjusted annual base flow at stations on Whiteoak Creek and Melton Branch at Oak Ridge National Laboratory (ORNL).

Streamflow adjusted by subtraction of effluent discharge (0 at stations W1 and M1).

Base flow adjusted by correction factor (see text).

28 Hydrology of Melton Valley at Oak Ridge National Laboratory, Tennessee

10 1 10° 5 io-2

£>

• - o 3 to~3 oz o

(J 10~4 10 10

-5

-6 o

o o

0 O 00 50 100 150 200 250 300 350 WELL DEPTH, IN FEET BELOW LAND SURFACE 400 450 500 Figure 14. Hydraulic conductivity variation with depth.

conductivity as isotropic produced similar model results (Tucci, 1986, p. 16).

Recent work in Bear Creek Valley indicated that the best model results were obtained using isotropic hydraulic-conductivity values in the regolith (Bailey and Lee, 1991). Aniso-tropy may vary locally due to variations in the degree of fracturing or fracture orientation. Because ground-water flow in the regolith hydrogeologic unit has characteristics of both porous-media and fracture-controlled flow (Webster and

Bradley, 1988,

p. 26-30), hydraulic conductivity may be either iso-tropic or anisotropic, depending on local weathering and fracture characteristics.

Anisotropic conditions are probably most prevalent in the deep parts of the regolith hydrogeologic unit where weathering of the bedrock is not as complete as in the shallow parts.

Ground-Water Flow Ground-water flow in the regolith hydro-geologic unit occurs as flow through a porous media, and is influenced by relict bedding and fractures within the regolith, particularly in the deep, weathered-bedrock part. Directions of ground-water flow in the valley are mainly determined by the shape of the water table. On a local scale, directions of ground-water flow also may be influenced by burial trenches, fractures, or geologic structures.

The shape of the water table on June 28, 1988, was determined from measurements of depth to water in more than 200 wells (fig. 17) completed in the regolith hydrogeologic unit (fig. 18). Water levels on that date averaged about 3 feet lower than average-annual water levels in 93 wells for which long-term Hydrology of Melton Valley 29

228 NUMBER OF DATA POINTS 45 39 12

-1 b

-2 o

3)

Oz o

o

-3

-4

-5

-6 X

X X

0-50 50-100 100-250

> 250 GROUPING BY WELL DEPTH, IN FEET EXPLANATION X

EXTREME VALUE HIGHEST DATA POINT WITHIN 1.5 X IQR ABOVE TOP OF BOX 75th PERCENTILE MEDIAN PERCENTILE LOWEST DATA POINT WITHIN 1.5 x IQR BELOW BOTTOM OF BOX Figure 1 5. -Boxplots showing ranges of hydraulic-conductivity values for selected depth intervals.

INTERQUARTILE /

RANGE (IQR) \\

data (more than 3 years) are available.

Although these data were obtained during a relatively dry time of the year, the shape of the water table generally remains the same throughout the year.

General conclusions from these data as to direction of ground-water flow are, therefore, considered valid.

The shape of the water table is similar to that of land surface.

Ground-water flow generally is from ridges and hills to streams (fig. 18), and topographic divides closely correspond to ground-water divides.

Ground water flows southwest to the Whiteoak Creek drainage system and the Clinch River in the south-western and central parts of the valley, and to Melton Hill Lake in the northeastern part of the valley (fig. 18).

Because of this close correspondence between topographic and ground-water divides, ground-water flow across Haw Ridge and Copper Ridge is assumed to be negligible.

However, data 30 Hydrology of Melton Valley at Oak Ridge National Laboratory, Tennessee

NUMBER OF DATA POINTS 13 89 81 11 23 1.0 0.5 Q

0 Of.itt a.

-0.5

-1.0

-1.5

-2.0

-2.5

-3.0

-3.5 X

X ALVM CKMG NCCK MRVL RGSV GEOLOGIC UNIT RTLG PPKV ROME X

INTERQUARTILE./

RANGE (IQR) ~\\

EXPLANATION EXTREME VALUE HIGHEST DATA POINT WITHIN 1.5 X IQR ABOVE TOP OF BOX 75th PERCENTILE MEDIAN 25th PERCENTILE LOWEST DATA POINT WITHIN 1.5 x IQR BELOW BOTTOM OF BOX GEOLOGIC UNITS ALVM Alluvium CKMG NCCK MRVL RGSV RTLG PPKV ROME Chlckomaugo Limestone Nollchucky Shale Maryvllle Limestone Rogersville Shale Rutledge Limestone Pumpkin Valley Shale Rome Formation Figure 16. Boxplots showing ranges of hydraulic-conductivity values for regolith developed over selected geologic units.

Hydrology of Melton Valley 31

• .<<fc.

II o >>

(0 S

• s*.

0.5 1 MILE 0.5 1

KILOMETER EXPLANATION BURIAL GROUND AND NUMBER WELL LOCATION Figure 17. Location of wells in which water table depth was measured on June 28, 1988.

OAK RiDGE NATIONAL LABORATORY 0.5 1 MILE 0.5 1 KILOMETER EXPLANATION o^

2.

BURIAL GROUND AND NUMBER WATER-LEVEL CONTOUR Dashed where approximately located. Contour interval 10 feet; additional contour for 745 feet. Datum Is sea level GROUND-WATER DIVIDE LINE OF SECTION OF CROSS-SECTIONAL MODEL GENERAL DIRECTION OF GROUND-WATER FLOW Figure 18. Water-table configuration and general directions of ground-water flow, June 28, 1988.

from the ridge areas are insufficient to support or disprove this assumption. Recent hydrologic investi-gations in Bear Creek Valley (Bailey and Lee, 1991, and L.E. Toran, Martin-Marietta Energy Systems, written commun., 1990) indicate possible ground-water flow across the ridges.

Differences in potentiometric head between the regolith hydrogeologic unit and deeper units indicate vertical components of ground-water flow.

These components are downward to bedrock over most of the valley, except for areas along Melton Branch and the lower reaches of Whiteoak Creek where an upward component of ground-water flow occurs. The vertical flow components are demonstrated by com-parison of potentiometric heads in clusters of wells completed at different depths.

For example, water levels in wells 469A, 470, 471, and 472 (north well cluster, fig. 17) on a hill at the north end of burial ground 5 indicate a downward component of flow (fig. 19). Well 472, screened from 15 to 20 feet, has the highest potentiometric heads, and well 469A, open from 191 to 201 feet, has the lowest heads. Wells 471 and 470, open from 89 to 99 and from 140 to 151 feet, respectively, have similar heads but at intermediate levels. The overall vertical component of flow at the north cluster is, therefore, down from the regolith hydrogeologic unit.

The rapid water-level declines and subsequent rises reflect removal of water from the wells for water-quality samples and later water-level recovery. Water levels in wells 461, 462, 463, and 464 (south well cluster, fig. 17), which are about 50 feet from Melton Branch at the south end of burial ground 5, show evidence of an upward component of flow (fig. 19).

Potentiometric heads are highest in well 461, which is open from 188 to about 202 feet, and lowest in well 464, which is screened from 6 to 11 feet. Wells 462 and 463, open from 140 to 151 and 88 to 100 feet, respectively, have intermediate heads, and heads are higher in well 462 than in well 463. The vertical component of ground-water flow at the south cluster is, therefore, up towards the regolith hydrogeologic unit. Heads in all wells, except well 464. are above land surface at this cluster.

Despite the vertical flow components, most ground water within the regolith hydrogeologic unit flows laterally to streams. Preliminary model analysis by Tucci (1986, p. 15) indicated that most ground-water flow is within the regolith hydrogeologic unit, and that less than 3 percent of the total ground-water flow in the valley is from the regolith hydrogeologic unit to bedrock.

Quantitative analysis of vertical ground-water flow is discussed further in the "Ground-Water-Flow Models" section.

Annual water-level fluctuations range from less than 1 foot to 12 feet (Webster and Bradley, 1988,

p. 47). Water-level fluctuations are usually minimal in low-lying areas near streams, and are usually greatest in upland areas.

For example, well U19, which is close to the Clinch River and about 150 feet from Whiteoak Creek below Whiteoak Dam, has an average-annual water-level fluctuation of about 1.5 feet (fig. 20). The largest average-annual water-level fluctuation recorded during this study was about 10 feet in well U35 (fig. 20), which is located on the flanks of Haw Ridge. Annual water-level fluctuations for 15 wells completed in the regolith hydrogeologic unit for which continuous data are available average about 4 feet.

Moore (1988b, p. 60) calculated the geometric mean of water-level fluctuations for 586 wells in Melton and Bethel Valleys to be 5 feet; however, this data set included wells completed below the bottom of the regolith hydrogeologic unit.

Recharge and Discharge Recharge to the regolith hydrogeologic unit is primarily from infiltration of precipitation. A small amount of recharge also may result from infiltration of streamflow during the summer, when ground-water levels fall below stream levels that are maintained by artificial discharge of effluent from ORNL plant facil-ities. Estimates of average-annual recharge vary and are discussed further in the following paragraphs.

Estimated average-annual precipitation at ORNL for 1951 through 1988 is about 50 in/yr. This estimate is based on the ratio (0.94) of average-annual precipitation at ORNL (47.03 in/yr) to average-annual precipitation at Oak Ridge (50.0 in/yr) for the period 1977 through 1988. Average-annual precipitation at Oak Ridge (53.27 inches) for the period 1951 through 1988 is multiplied by this ratio to estimate the average-annual precipitation at ORNL for the same period.

Moore (1988b, p. 33) estimated that about 57 percent of annual precipitation is lost to evaporation and tran-spiration. This loss is 28.5 in/yr, assuming an average-annual precipitation of 50 inches. The remaining 21.5 in/yr is discharged to streams and springs, either by runoff or discharge through the ground-water system.

Moore (1988b,

p. 71) concluded that little 34 Hydrology of Melton Valley at Oak Ridge National Laboratory, Tennessee

815 810 805 -

800 -

795 o

3 fc K 775 770 I I I I I I I I I I I I I I I I I I H.l I I I I I I I I I I

I I

I I I I I I I I 1 I

I I I I I I I I I I I. I I

I I I I I

I I I I I I

I I I I I OD MMJSNJMMJSNJMMJSN JMMJSNJMMJSNJMMJS 1982 1983 1984 1985 1986 1987 1988 WATER YEAR 765 760 755 I I I

I I

I I

I I

I I

I I

l I

I I

I I

I I

I I

I I

I I

I I

I I

I I

I l

I I

i I

I I

1 I

I 1

I I

SOUTH.

I I I

I I I I I I l I

I I

I I l I l I l I I I I I

I I

I I I

I I I

I I

I I I

I I I I I I I 1

0 NDJFMAMJJASONDJFMAMJJASONDJFMAMJJASOND JFMAMJJAS 1984 1985 1986 1987 1988 WATER YEAR Figure 19. Hydrographs for north well cluster (April 1983 through September 1988) and south well cluster (October 1984 through September 1988), burial ground 5.

Hydrology of Melton Valley 35

OCT JULY Figure 20. Water levels in wells U19 and U35, January 1986 through September 1988.

precipitation discharges to streams as overland runoff, but that 90 to 95 percent of the discharge to streams is through the stormflow zone above the water table.

The remaining 1.1 to 2.2 inches of the 50in/yr average-annual precipitation is recharge to the water table, based on Moore's (1988b) estimates.

Tucci (1986, p. 11) estimated an average-annual recharge rate of 3.2 in/yr in the preliminary model analysis of Melton Valley.

Estimates of average-annual recharge can be made on the basis of base-flow analysis, assuming that base flow to streams is equal to recharge to the ground-water system. Data for stations Ml and M2 (fig. 12), which cover a wide range of precipitation, are generally less influenced by effluent discharge than stations on Whiteoak Creek and probably provide a more accurate estimate of recharge than1 data for the other stations. Average-annual base flow, assuming a linear relation between base flow and precipitation (fig. 12) and an average-annual precipitation of 50 in/yr at ORNL, is estimated to be 210 (tf/s-dymi2, or about 8 in/yr. This recharge is about 16 percent of average-annual precipitation. Recharge for the rela-tively dry years 1986 through 1988, based on this analysis, was about 9 percent of the 38.4 inches of average-annual precipitation for those years (table 6),

or about 4 in/yr.

36 Hydrology of Melton Valley at Oak Ridge National Laboratory, Tennessee

Recharge to the ground-water system primarily occurs on the upland areas (Haw and Copper Ridges) and on low hills in the middle of the valley. Preci-pitation that falls in low-lying areas, where the water table is at, or close to, land surface, tends to be rejected or transpired. Low-lying areas are primarily discharge areas, where ground water moves laterally and up to streams.

Discharge from the regolith hydrogeologic unit is primarily to streams, springs, and Whiteoak Lake.

A minor amount of ground water flows from the regolith to deeper units. Tucci (1986, p. 15) estimated that 0.02 ft3/s, or about 3 percent of flow in the regolith, is discharged to the underlying bedrock unit.

Details of the water budget are discussed further in the "Areal Model" section.

Bedrock The bedrock hydrogeologic unit consists of weathered and unweathered bedrock at depths gener-ally greater than 50 feet below land surface. Locally, this unit includes pockets of regolith where regolith is more than 50 feet deep. The bedrock unit is further subdivided, on the basis of differences in hydraulic conductivity, into shallow, intermediate, and deep zones. The bedrock unit extends to depths of at least 600 feet, at which depth saline ground water is often present (J. Switek, Martin-Marietta Energy Systems, written commun., 1988).

The base of the fresh ground-water-flow system is considered to be at about 600 feet below land surface; however, brines have been reported in wells as shallow as 365 feet (J. Switek, Martin-Marietta Energy Systems, written commun., 1988). Ground-water flow in bedrock is primarily through fractures, faults, and solution openings. The largest component of flow is along laterally continuous bedding-plane openings (Webster and Bradley, 1988, p. 79).

Hydraulic Conductivity Hydraulic conductivity of the bedrock unit varies over several orders of magnitude, but generally decreases with increasing depth (fig. 15). The shal-low zone of the bedrock unit is from 50 to 100 feet below land surface. Hydraulic-conductivity values for the shallow zone range from 1.2 x 10~3 to 2.4 ft/d, based on slug tests conducted on 45 wells completed in this zone (fig. 15).

The median hydraulic-conductivity value of the shallow zone is 0.05 ft/d, and the mean value is 0.33 ft/d.

Wells completed within the Maryville Limestone and Nolichucky Shale have the highest hydraulic conductivities in this zone; however, wells completed in these geologic units com-prise 87 percent of all the wells completed in this zone. Hydraulic-conductivity values of this zone are similar to those of the regolith, but are somewhat lower overall.

The largest hydraulic-conductivity values in this zone are associated with wells com-pleted in the uppermost part at depths less than 60 feet. The shallow bedrock zone is the most perme-able of the bedrock zones because it is the most weathered, particularly in the upper part. Open frac-tures probably are more common in this zone than in deeper zones.

The intermediate bedrock zone consists of un-weathered bedrock and extends from 100 to 250 feet below land surface. Hydraulic-conductivity values for 39 wells completed in this zone range from 1 x 10"5 to 0.19 ft/d (fig. 15).

The median hydraulic-conductivity value is 0.01 ft/d, and the mean value is 0.03 ft/d. Nearly all hydraulic-conductivity values are less than 0.1 ft/d, and about 25 percent of the values are less than 0.001 ft/d (fig. 14). Hydraulic conduc-tivity appears to be largest in wells completed in the Maryville Limestone and the Nolichucky Shale at depths greater than 100 feet.

The deep bedrock zone consists of unweathered bedrock and extends from 250 to 600 feet below land surface. In this zone, the rocks are generally much less permeable than in the overlying zones. Hydraulic-conductivity values for 12 wells completed in this zone range from slightly less than 1 x 10'5 to 2 x 10"3 ft/d (fig. 15). The median hydraulic-conductivity value is slightly more than 1 x 10'5 ft/d, and the mean value is about 3.0 x 10"4 ft/d.

Most hydraulic-conductivity values in this zone are less than 2 x 10~3 ft/d. Although all data available for Melton Val-ley indicate low hydraulic-conductivity values at depths greater than 250 feet, King and Haase (1988,

p. 34) reported values as large as 1.16 (Rome Forma-tion) and 0.41 ft/d (Maynardville Limestone) at depths of about 655 and 1,000 feet, respectively, in Bear Creek Valley. Zones of relatively large hydraulic con-ductivity within the deep bedrock zone, therefore, may be present in Melton Valley, but are as yet undetected.

Because of the dominant influence of fractures on ground-water flow in bedrock, the hydraulic Hydrology of Melton Valley 37

conductivity of the rocks is anisotropic. The maxi-mum hydraulic-conductivity vector generally is paral-lel to strike.

Reported anisotropy values (strike-normal to strike-parallel) range from 1:3 to 1:20 (Tucci, 1986, p. 5). Fractures or faults that cut across strike may locally influence the direction of the maximum hydraulic-conductivity vector and ground-water flow (Dreier and Toran, 1989, p. 63-64).

Ground-Water Flow Ground-water flow in bedrock is primarily through bedding-plane openings, fractures, faults, and solution openings. The orientation of these features, therefore, has a significant influence on the direction of ground-water flow.

Anisotropy of bedrock hydraulic conductivity resulting from these features causes ground-water flow to be skewed in the direction of the maximum hydraulic-conductivity vector, so that flow lines are not always perpendicular to equipotential lines.

For example, the potentio-metric surface of the shallow bedrock zone (fig. 21) is similar to that of the overlying regolith hydrogeologic unit (fig. 18); however, because of the influence of secondary openings in bedrock, the direction of ground-water flow may be different from that in the regolith hydrogeologic unit, particularly when the hydraulic gradient is perpendicular to strike (Webster and Bradley, 1988, p. 80).

Data for the intermediate and deep bedrock zones are insufficient to adequately describe flow within those zones. The potentiometric surfaces of those zones are believed to be similar to those of the overlying zones, but gradients are probably less.

As previously stated, differences in potentio-metric head between regolith and bedrock indicate the presence of vertical flow components. These compo-nents are down from regolith to bedrock in upland areas and up from bedrock to regolith in lowland areas (fig. 19). In the north well cluster heads are similar in the 99 to 151 foot interval, indicating a larger lateral flow component than vertical flow component. Poten-tiometric data from well clusters in the western and central parts of burial ground 5 indicate flow compo-nents down from regolith and up from the intermediate bedrock zone to the shallow bedrock zone (Webster and Bradley, 1988, p. 58).

Water-level fluctuations in bedrock are slightly less variable than those in regolith. Continuous water-level data for 30 wells completed in bedrock indicate annual fluctuations that range from about 1 to 10 feet.

The mean annual water-level fluctuation calculated for the 30 bedrock wells is 3.7 feet.

Recharge and Discharge All recharge to bedrock is believed to occur through the regolith. Data are insufficient to deter-mine any recharge to bedrock as underflow from other areas near Melton Valley. Recharge to bedrock is primarily in the upland and mid-slope areas, where a downward potentiometric gradient from the regolith exists. Tucci (1986, p. 15) estimated that recharge to bedrock from regolith was about 3 percent of the estimated recharge to the regolith (3.2 in/yr), or 0.09 in/yr. Moore (1988b, p. 84) estimated that about 1 percent of the total ground-water flow occurs at depths greater than 200 feet.

Ground-water flow in bedrock is assumed to discharge back to the regolith because the system is believed to be in equilibrium. Discharge from bed-rock occurs in low lying areas near streams, where an upward potentiometric gradient from bedrock to the regolith exists. Data are insufficient to determine the occurrence of discharge from bedrock as underflow out of the valley.

Simulation of Ground-Water Flow Numerical models were used to provide a better understanding of the ground-water-flow system of the study area. Models are useful tools for this purpose because they can incorporate all of the major compo-nents that affect ground-water flow, and they allow for the evaluation of the interactions of the various components.

A computer program written by McDonald and Harbaugh (1988) was used to simulate ground-water flow in the study area.

The program uses finite-difference techniques to solve the ground-water-flow equation for three-dimensional, steady or non-steady flow in an anisotropic, heterogeneous medium. Two models were constructed to simulate (1) flow in a two-dimensional cross section from Haw Ridge to Copper Ridge through burial ground 5, and (2) areal, three-dimensional flow in Melton Valley. The models were constructed and calibrated to simulate only steady-state ground-water flow. Transient conditions were not simulated, and the models are not intended for use 38 Hydrology of Melton Valley at Oak Ridge National Laboratory, Tennessee

-2*0 O

O_

X O

O L.

o

.a

^00 00

£2 w

<< 00

£ ^

V- <<

u O

'Z. O

" *~

E o IT) o Q.

I I

CM D>

Hydrology of Melton Valley 39

in any transient or predictive ground-water-flow problems. A basic assumption of the models is that ground-water flow in the fractured rocks of the study area can be approximated as flow through an anisotropic, porous medium. Flow through individual fractures is not directly simulated in the models.

Cross-Sectional Model A two-dimensional, cross-sectional model of ground-water flow was used to provide general information concerning distribution of recharge across Melton Valley and relative amounts of ground-water flow with depth. The model also was used to test two alternative distributions of hydraulic conductivity with depth. Results of simulation using the cross-sectional model provided information for the areal flow model.

The cross-sectional model simulates ground-water flow through a 1-foot wide section of the aquifer along a flow line. Hydraulic conditions along that section are assumed to be representative of conditions along similar sections across the valley. A hydrogeo-logic section was constructed, approximately perpen-dicular to strike, from Haw Ridge, through burial ground 5, to Copper Ridge (fig. 18). This section was chosen because of the availability of deep potentio-metric, subsurface geologic, and abundant water-table data. Ground-water flow along this section is from the ridges to streams within the valley.

Model Construction, Assumptions, and Boundary Conditions The model represents a line of section about 1-mile long, and consists of variably spaced grid blocks that range in area from 3,200 to 24,000 ft2 (fig. 22). The smallest grid blocks generally represent HAW RIDGE COPPER RIDGE NO-FLOW BOUNDARY Geology modified from R.B. Dr<<l<<r, C.S. Haas*. C.M. Bvaudoln. H.L. King, and J. Swltok, 1987 EXPLANATION GRID BLOCK SIMULATING:

CONSTANT-HEAD NODE D

ACTIVE NODE M

INACTIVE NODE H

MEASURED WELL Figure 22. -Cross-sectional model grid and boundaries.

40 Hydrology of Melton Valley at Oak Ridge National Laboratory, Tennessee

the regolith hydrogeologic unit, and the largest grid blocks represent deep bedrock. The maximum depth represented by the model ranges from about 750 feet below Melton Branch to about 1,200 feet below Copper Ridge.

The water table is the upper boundary of the model and is simulated as a constant-head boundary within the upper part of the regolith hydrogeologic unit (fig. 22). Water-table elevations along the sec-tion, mapped from water-level measurements made on June 28, 1988, from more than 200 wells throughout the valley, were assigned to each constant-head node.

These measurements provide excellent control on the configuration of the water table in the valley on that date; however, water levels were not available for most areas on Haw Ridge and Copper Ridge. The model-calculated ground-water flow from the constant-head nodes represents recharge to and dis-charge from the regolith hydrogeologic unit.

The sides and bottom of the model are simulated as no-flow boundaries (fig. 22). The sides of the model correspond to topographic divides and are assumed to correspond to ground-water divides below the ridges.

Ground water is generally saline at depths greater than 600 feet below Melton Branch, but the model was extended to 750 feet to assess the relative amount of ground-water flow from the surface that could circu-late to depths greater than 600 feet.

Two different hydraulic-conductivity distribu-tions were simulated in the model. The first distri-bution (simulation 1) represented the concept that hydraulic conductivity varies primarily with depth, regardless of geologic unit.

Median values of hydraulic conductivity for depth intervals of 0 to 50, 50 to 100, 100 to 250, and greater than 250 feet (fig. 15) were input to the model (fig. 23, simulation 1). Hydraulic conductivity ranged from 1.0 x 10'5 to 0.19 ft/d for these depth intervals.

The second hydraulic-conductivity distribution (simulation 2) represented the concept that hydraulic conductivity varies with both depth and geologic unit.

Data for this distribution (fig. 23, simulation 2) are based on hydraulic-conductivity data discussed previ-ously in this report. Hydraulic conductivity ranged from 1.0 x 10"5 to 0.63 ft/d for this simulation.

Hydraulic-conductivity values at depths greater than 400 feet were set equal to 1 x 10"5 ft/d, which is the median value calculated from slug tests in 12 wells open at a depth of 400 feet (fig. 15).

The hydraulic-conductivity values for both sim-ulations were multiplied by a factor of 0.33 to account for horizontal anisotropy, in which hydraulic conduc-tivity is assumed to be greater parallel to strike than normal to strike. The factor used (0.33) represents the 1:3 ratio used by Tucci (1986, p. 7) in the preliminary ground-water-flow model of Melton Valley. A ratio of 10:1 for horizontal to vertical hydraulic conduc-tivity was assumed and input to the model. This ratio is somewhat lower than ratios commonly used in hori-zontally layered aquifers (100:1 or greater) because of the dip of the rocks and the potential for downward ground-water flow along bedding planes.

Simulated water levels were compared to water levels measured on June 28, 1988, in six deep wells (fig. 22) along the line of section. Water levels at that time were about 3 feet lower, overall, than average-annual water levels (fig. 24), so that model-calculated recharge rates based on those water levels will be somewhat lower than average-annual recharge rates.

However, the areal distribution of simulated recharge and discharge should be consistent throughout the year, because the overall shape of the water table gen-erally remains constant.

Results of Simulations Results of both cross-sectional model simula-tions support the concept that recharge to the ground-water system takes place on the ridges and on the lower, secondary hills in the valley, and the concept that nearly all ground-water flow is at depths of less than 250 feet below Melton Branch (table 8).

Simulated ground-water flow was from Haw Ridge and the northern part of burial ground 5 to the area near an unnamed tributary north of the burial ground, and from the southern part of burial ground 5 and Copper Ridge to Melton Branch (fig. 25a and 25b).

Simulated water levels generally were within 10 feet of measured water levels in the six wells used for comparison of simulation results, and were generally lower than measured water levels (fig. 25a and 25b).

The differences in the two simulations were in the total amount of model-calculated recharge and in the relative amounts of ground-water flow at shallow depths.

In simulation 1, in which hydraulic conduc-tivity varies only with depth, total model-calculated recharge was about 4.7 in/yr (table 8).

Recharge Hydrology of Melton Valley 41

(inches per year) is derived by dividing the model-calculated flow from each constant-head cell (cubic feet per day) by the area of the flow face (square feet) and dividing by 4,380 (conversion of feet per day to inches per year). Of this recharge, about 48 percent (2.2 in/yr) occurred on Copper Ridge, 21 percent (l.Oin/yr) occurred on Haw Ridge, and 31 percent (1.5 in/yr) occurred on the low hills within the valley (table 8).

A small amount of recharge (0.3 in/yr) occurred in one grid block near the base of Copper Ridge, and is included in the amount calculated for Copper Ridge. However, the recharge for that grid block may result either from an error in the estimated water-table elevations in that area or from the model discretization. Model-calculated discharge at the base of the ridges represents discharge by springs and evapotranspiration at lower elevations where plant roots intercept the shallow water table.

Model-calculated ground-water flow for simula-tion 1 decreases with increasing depth. In the follow-ing discussions, references to depths are depths below Melton Branch. In order to calculate percent flow, model-calculated flow at the bottom face of each grid cell, just above the specified depth, is summed and then divided by the total flow.

Most ground-water flow (96 percent) is at depths of less than 50 feet (table 8). Ninety-seven percent of the total ground-water flow is within the upper 100 feet and nearly all SIMULATION 1

o.o 0.19 0.05 0.01 Assumes that hydraulic conductivity varies primarily with depth, regardless of geologic unit 1.0 X 10~5 SIMULATION 2 0.0 0.12 0.025 n m

' 1 0.28 1 0.035 1

0.17 0.05 1.0 X 10~3 1.0 X 10~5 Assumes that hydraulic conductivity varies with depth and geologic unit 3.0 X lO"3 1.0 X 10~5 HYDRAULIC-CONDUCTIVITY VALUES. IN FEET PER DAY 0

200 400 600 800 1,000 FEET I ' H h ' '

0 100 200 METERS Figure 23. Hydraulic-conductivity distributions, unadjusted for horizontal anistropy, for cross-sectional model simulations 1

and 2.

42 Hydrology of Melton Valley at Oak Ridge National Laboratory, Tennessee

Approximate average water level 0

D MMJSNJMMJSNJMMJSNJMMJSNJMMJSNJMMJS 1983 1984 1985 1986 T987 1988 WATER YEAR Figure 24. Water levels in well 474 showing fluctuations from April 1983 through September 1988.

(greater than 99 percent) is within the upper 250 feet (table 8). At depths greater than 550 feet, ground-water flow is negligible. This result is supported by the presence of saline water at depths of 600 feet below Melton Branch.

Simulated water levels were within about

+/-5 feet of measured water levels, except on Haw Ridge, in simulation 1. The 27-foot difference on the ridge is because of a simulated vertical head gradient that does not occur at this site. The lack of vertical gradient at this site may indicate horizontal ground-water flow beneath the ridge or a vertical hydraulic connection with anomalously large permeability from the regolith hydrogeologic unit to depths of at least 200 feet. Cavities, which could provide such a vertical connection, were encountered during the drilling of well UC2 (Tucci and Hanchar, 1989, p. 15). The dip of the rocks is steep in this area (45 to 90 degrees), so that vertical flow components may be facilitated. The vertical anisotropy of hydraulic conductivity, which must be uniformly applied by the model over the entire model area, correctly simulates vertical grad-ients that occur in all other parts of the model area, but not at the Haw Ridge site.

In simulation 2, in which hydraulic conduc-tivity varies with both depth and geologic unit, total model-calculated recharge, 2.1 in/yr (table 8), was somewhat lower than in simulation 1. The hydraulic Hydrology of Melton Valley 43

HAW RIDGE COPPER RIDGE Model boundary 200 400 600 800 1,000 FEET EXPLANATION 800 SIMULATED WATER-LEVEL CONTOUR Contour Interval 20 feet.

Datum Is sea level GENERALIZED DIRECTION OF GROUND-WATER FLOW

-3 DIFFERENCE BETWEEN MEASURED AND SIMULATED WATER LEVEL, IN FEET Note:

Trace of section Is shown on figure 18.

Figure 25. Simulated water levels and general directions of ground water flow for cross sectional model simulations (a)1 and (b)2.

conductivity of the regolith is lower, overall, for simulation 2 than for simulation 1, so that model-calculated recharge also is lower. The percentage of recharge on the ridges also differs from that in simu-lation 1.

A greater percentage of recharge (67 per-cent) occurs on Copper Ridge, and lesser per-centages of recharge occur on Haw Ridge (10 percent) and on the secondary hills (23 percent) in simulation 2 than in simulation 1.

Model-calculated ground-water flow also decreases with increasing depth in simulation 2; how-ever, the percentage of total flow is less at shallow depths in simulation 2 than in simulation 1. Ground-water flow in the upper 50 feet is 91 percent of the total flow and flow in the upper 100 feet is 97 percent of the total flow (table 8). Nearly all (greater than 99 percent) of the total ground-water flow occurs at depths less than 250 feet, and ground-water flow is 44 Hydrology of Melton Valley at Oak Ridge National Laboratory, Tennessee

Table 8.

Model-calculated water-budget components for cross-sectional model simulations Budget or flow component Simulation 1 a Simulation 2b Recharge (inches per year):

Total 4.7 Copper Ridge 2.2 Haw Ridge 1.0 Mid-valley 1.5 Percentage of total flow in indicated depth interval:

0- 50 feet 96 50-100 feet 97 100-250 feet 99.7 250-550 feet 100 2.1 1.4

.2

.5 91 97 99.7 100 aAssumes that hydraulic conductivity varies primarily with depth, regardless of geologic unit.

bAssumes hydraulic conductivity varies with both depth and geologic unit.

negligible at depths greater than 550 feet (table 8), as in simulation 1.

Simulated water levels do not match measured water levels as closely in simulation 2 as in simulation 1, except on Haw Ridge; however, simulated water levels are generally within 10 feet of measured water levels.

The closer match beneath Haw Ridge is because of the larger hydraulic-conductivity values assigned to the Rome Formation at depths of 200 feet in the simulation. These larger hydraulic-conductivity values more closely approximate the enhanced vertical hydraulic connection thought to exist at this site.

Comparison of the two model simulations indicates that model-calculated recharge is sensitive to regolith hydraulic conductivity. Recharge is greater in simulation 1 than in simulation 2 because the hydraulic conductivity of the regolith is greater, over-all, in simulation 1 than in simulation 2. The model-calculated recharge values for both simulations, however, are within the range of values indicated from base-flow analysis and values reported by previ-ous investigators (Tucci, 1986; Moore, 1988b).

Because water-table elevations for June 1988, a period of lower-than-average water levels, were used for the simulations, model-calculated recharge may be somewhat less than average-annual recharge. Simu-lated water levels beneath Haw Ridge more closely match measured water levels when an enhanced verti-cal hydraulic connection is simulated in that area.

Simulated heads also are sensitive to hydraulic-conductivity values; however, the few wells along the line of section available for comparison to model results preclude quantitative evaluation of this sensi-tivity.

Model results concerning flow at depths greater than 200 feet should be considered somewhat speculative because of a lack of water-level data below that depth for comparison to model results.

Because water-table elevations on Copper Ridge are unknown, additional simulations were made in which the water table was both raised and lowered by 25 feet near the crest of the ridge. Simulated heads were only locally affected by these changes, and model-calculated recharge was within +/-5 percent of the recharge calculated in the original simulations. Poten-tial errors in model results due to inaccurate simu-lation of the water table on Copper Ridge, therefore, are considered to be small.

Areal Model A three-dimensional model was constructed to test and refine concepts of ground-water flow for Melton Valley. The model also may provide valuable information for more site-specific studies that are ongoing or planned within the valley. Information on the distribution of recharge across the valley obtained from the cross-sectional-model analysis was incorpo-rated into the areal model.

Model Construction, Assumptions, and Boundary Conditions The model consists of four layers that corre-spond to the depth intervals and distribution of hydraulic conductivity discussed in the "Ground-Water Hydrology" section. The layers represent a con-tinous ground-water-flow system, in which hydraulic conductivity decreases with depth, rather than a series of aquifers separated by confining units typical of many flow models.

The layers and corresponding depth intervals are as follows:

Layer 1 water table to 50 feet, Layer 2 50 to 100 feet, Layer 3 100 to 250 feet, and Layer 4 250 to 600 feet.

Layer 1 consists of saturated regolith and the uppermost, partly weathered bedrock, and corre-sponds to the regolith hydrogeologic unit. Layer 2 consists of shallow, weathered and unweathered bed-rock. Layers 3 and 4 consist of deep unweathered Hydrology of Melton Valley 45

bedrock. The bottom of layer 4 (600 feet) is assumed to be the bottom of the freshwater flow system.

The total model area is about 6.1 mi2; however, only about 4.9 mi2 of this area is active within the model (fig. 26). The remaining 1.2 mi2 of the total model area is assumed to be outside the area of contri-bution of ground-water flow for the valley.

The model area includes all of the Whiteoak Creek drain-age within Melton Valley, as well as drainage towards Melton Hill Lake at the northeastern end of the valley, and a small tributary drainage to the Clinch River at the southwestern end of the valley (fig. 26).

Lateral boundaries are primarily no-flow bound-aries located along surface-water divides that are assumed to coincide with ground-water divides.

Because the water table (fig. 18) is similar in shape to the topography, this assumption is probably valid.

Boundaries along the southwestern and northeastern ends of the model that correspond to the Clinch River and Melton Hill Lake, respectively, are simulated as constant-head boundaries within layer 1. Water levels in these surface-water bodies do fluctuate during the year; however, the fluctuations are generally only a few feet above or below their average stage because of regulation of flow at dams along the Clinch River.

Average stages, 741 feet for the Clinch River and 795 feet for Melton Hill Lake, were assigned as the constant-head values in layer 1 in the areas underlying these water bodies. Lateral boundaries in all other layers are assumed to be no-flow, so that all discharge from and recharge to these layers is assumed to be to or from layer 1.

Data are insufficient to determine whether deep ground water flows into or out of the valley or beyond the Clinch River.

Whiteoak Lake was simulated by river nodes (fig. 26), which allow leakage to or from the aquifer through the lake or river bottom. An average lake-stage elevation of 745 feet and a uniform lake-bottom-sediment thickness of 2 feet was assumed. Whiteoak Creek and the lower half of Melton Branch also were simulated as river nodes (fig. 26).

These streams receive effluent discharged from various ORNL plant facilities, so that base flow is artificially maintained and may provide recharge to the ground-water system during summer months when ground-water levels fall below stream stages.

All other streams were simulated as drain nodes which only allow leakage from the aquifer to the stream. Stream widths were either measured during seepage investigations or estimated from topographic maps; a minimum width of 1 foot was assumed.

Streambed thickness for all streams was assumed to be 1 foot. A uniform vertical hydraulic-conductivity value of 0.1 ft/d was assumed for all streambeds. This value was chosen on the basis of preliminary model analysis (Tucci, 1986, p. 11) and was tested during model-sensitivity analysis. All river and drain nodes are within layer 1.

Recharge to the water table by precipitation was simulated in layer 1 only on the ridges and on the low hills within the valley (fig. 27). This distribution is based on the results of the cross-sectional model analysis discussed in the previous section. Modeling of ground-water flow in Bear Creek Valley, which has a similar hydrogeologic setting, indicated that recharge only occurs on the bordering ridges (Bailey and Lee, 1991). An estimate of recharge on the ridges and hills of about 5.0 in/yr, resulting in a total simulated recharge to the valley of about 2.0 in/yr, was initially used in the model. This estimate was tested during model calibration, and the effects of using different recharge rates on model results was examined as part of the sensitivity analysis. The total recharge rate for the valley (2 in/yr) is slightly less than that used (3.2 in/yr) in the preliminary model analysis (Tucci, 1986, p. 11), but is within the range of recharge rates calculated on the basis of analyses by Moore (1988b, p. 71). The rates used on the ridges are similar to the rates used in an area! ground-water-flow model of Bear Creek Valley, 4.0 to 5.0 in/yr (Bailey and Lee, 1991).

The model code requires the use of a vertical leakance value between model layers. The leakance value usually represents the vertical hydraulic con-ductivity divided by the thickness of a confining bed separating two aquifers; however, in this analysis the leakance value represents the vertical connection between two adjacent model layers. Vertical leakance for two adjacent model layers is calculated by the equation (McDonald and Harbaugh, 1988, p. 5-13) 1

+ 02/2) 46 Hydrology of Melton Valley at Oak Ridge National Laboratory, Tennessee

r r O.S 1

KILOMETER EXPLANATION BURIAL GROUND AND NUMBER GRID BLOCK SIMULATING:

CONSTANT HEAD RIVER DRAIN o

INACTIVE AREA Figure 26. Areal model grid and boundary conditions.

0.5 1 KILOMETER EXPLANATION BURIAL GROUND AND NUMBER GRID BLOCK IN WHICH RECHARGE IS SIMULATED Figure 27. Areal distribution of simulated recharge.

Kz\\

KZ2 where Vc is vertical leakance, in feet per day per foot; bi is thickness of the upper model layer, in feet;

£>2 is thickness of the lower model layer, in feet; is vertical hydraulic conductivity of the upper layer, in feet per day; and is vertical hydraulic conductivity of the lower layer, in feet per day.

Vertical leakance values were assigned between each model layer using equation 3. Vertical hydraulic conductivity was assumed to be equal to one-tenth the median horizontal hydraulic-conductivity value for each layer because of the relatively shallow dip of the beds (10 to 20 degrees) and the abundance of low permeability shale layers within most rock units.

Calculated vertical leakance between layers 1 and 2, assuming an average thickness of 30 feet for layer 1, is 1.8 x 10'4 (ft/d)/ft.

Calculated vertical leakance between layers 2 and 3 and layers 3 and 4 are 1.3 x 10'5 and 1.3 x 10"7 (ft/d)/ft, respectively.

Transmissivity of layer 1 was initially input to the model by multiplying the median hydraulic conductivity of that layer (0.19 ft/d) by the measured or estimated saturated thickness of layer 1 for hydrologic conditions on June 28, 1988 (fig. 28).

Transmissivities of layers 2 and 3 were initially input by multiplying the median hydraulic conductivity of each geologic unit in those layers, where known, by the layer thickness 50 feet for layer 2 and 150 feet for layer 3. A uniform transmissivity value of 7.9 x 10-2 ft2/d was assumed for layer 4 based on an

,-4 assumed hydraulic conductivity of 2.25 x 10 ft/d and Table 9. Adjustments to June 28, 1988, water levels used to estimate average-annual water levels

[All adjustments are in feet]

Adjustment, in feet, for indicated topographic setting Model layer 1

2 3

Valley bottom

+1.0

+1.0

+.5 Mid-slope

+2.0

+2.0

+1.0 Hilltop

+5.0

+5.0

+1.0 a layer thickness of 350 feet. Additional detail in the transmissivity distribution of layer 4 is not justified because of the lack of hydrologic data for that layer.

Transmissivity values of layers 1, 2, and 3 were tested during model calibration and sensitivity analysis and those adjustments are discussed in subsequent sections.

An areal anisotropy ratio of 1:3 (strike-normal to strike-parallel) for transmissivity of all layers was initially chosen, based on the preliminary model analysis of Tucci (1986, p. 7).

Simulated water levels were compared to estimated average water levels for wells within 238 model grid blocks (164 in layer 1, 48 in layer 2, and 26 in layer 3). Water-level data are not available for layer 4.

Average water levels were estimated by adjusting the water levels measured on June 28, 1988.

The amount of adjustment was based on comparison of June 28, 1988, water levels to available long-term water-level records for each well.

Adjustments ranged from 0.5 to 5.0 feet according to model layer and topographic setting (table 9). The largest adjust-ments were required for wells completed in layers 1 and 2 that are located on hilltops, and the smallest adjust-ments were required for wells completed in layer 3 and located in valley bottoms.

Results of Simulations Comparison of simulated water levels to esti-mated average-annual water levels was the primary calibration criteria for the areal flow model. Rigorous comparison of model-calculated to measured ground-water seepage to streams was not a part of the calibration process because average-annual seepage to streams that are unaffected by effluent discharge is not well defined.

Overall, simulated water levels were in reason-ably good agreement with estimated water levels, in that a little more than one-half of the simulated water levels were within +/-5 feet of estimated average-annual water levels (table 10, figs. 29-31).

About 97 percent of the simulated water levels in layer 1 and about 96 percent of the water levels in all layers were within 20 feet of the estimated levels. The maximum difference between simulated and estimated water levels was 41 feet in layer 1 near the entrance to burial ground 5.

Simulated water levels gener-ally were higher than estimated levels in layer 1, lower than estimated levels in layer 2, and higher than estimated levels in layer 3.

Hydrology of Melton Valley 49

a 3s

<fc.$<<.

S o I

<Sx..

^

OAK RIDGE BETHEL NATIONAL LABORATORY VALLEY

~

0.5 1 MILE 35-0.5 1 KILOMETER EXPLANATION BURIAL GROUND AND NUMBER LINE OF EQUAL THICKNESS OF SATURATED MATERIAL Interval 5 feet i

i iQ o/o

?4 Figure 28.--Saturated thickness of layer 1, June 28, 1988.

s.

0.5 1 MILE 0.5 1

KILOMETER DATUM IS SEA LEVEL 900-EXPLANATION BURIAL GROUND AND NUMBER WATER-LEVEL CONTOUR Shows simulated water-level altitude. Contour interval 25 feet. Datum is sea level Figure 29.--Simulated water levels for layer 1

II o

S*.

& o at ft I

OAK RIDGE NATIONAL LABORATORY I ETHEL V A 0.5 1 MILE 0.5 1 KILOMETER 900-EXPLANATION BURIAL GROUND AND NUMBER WATER-LEVEL CONTOUR Shows simulated water-level altitude. Contour interval 25 feet. Datum Is sea level Figure 30.--Simulated water levels for layer 2.

OAK RIDGE BETHEL 0.5 1 MILE 0.5 1 KILOMETER 5 I s

EXPLANATION BURIAL GROUND AND NUMBER 900 WATER-LEVEL CONTOUR Shows simulated water-level altitude. Contour interval 25 feet. Datum is sea level Figure 31.--Simulated water levels for layer 3.

The shape of the simulated water table in layer 1 and potentiometric surfaces of layers 2 and 3 (figs. 29-31) are similar to the shape of the estimated surfaces. The shape of all simulated potentiometric surfaces reflects the shape of the land surface, although the potentiometric contours become more generalized and less convoluted with increasing depth, particularly in layer 4 (fig. 32).

A ground-water divide, which separates flow to the Clinch River from flow to Melton Hill Lake, is present in all layers below the surface-water divide in the northeastern end of the model area.

The root mean square error (RMSE) was calculated in order to compare simulated and estimated water levels for each simulation, and to provide a comparison between simulations. RMSE, in feet, is calculated by RMSE =

N E

i= 1 (4)

N where N is the number of observations; hi" is the estimated water levels, in feet; and hf is the simulated water level, in feet.

The calibrated values of RMSE were 7.8 feet for layer 1, 9.6 feet for layer 2, 8.4 feet for layer 3, and 8.3 feet for all layers. Lower RMSE values may indicate more accurate model results, in that simulated water levels are more similar to estimated water levels. Conversely, higher RMSE values may indicate less accurate model results.

Table 10. Percentage of simulated water levels within 20, 10, and 5 feet of estimated average-annual water level Percentage of simulated water levels within indicated distance of estimated average-annual water level Model layer Layer 1 Layer 2 Layer 3 All layers 20 feet 97 92 96 96 10 feet 85 75 77 83 5 feet 58 56 50 57 Results of simulations were considered to be acceptable using most of the input values discussed in the previous section.

Simulation of layer 1 as an isotropic system, in which hydraulic conductivity is equal in all horizontal directions, produced the best simulation results. Similar results were obtained in a flow model of Bear Creek Valley, in which the regolith was simulated as isotropic (Bailey and Lee, 1991).

Simulation results produced lower RMSE values using the minimum interquartile values for hydraulic conductivity of layers 2 and 3 (0.02 and 1.6 x 10"3 ftVd, respectively) than when using the median hydraulic-conductivity values.

Several other conceptual models were tested during model calibration, but the simulations pro-duced greater RMSE values. Some of the concepts tested included:

Variable hydraulic conductivity with geologic unit as well as depth (RMSE = 13.7 feet for all layers);

Uniform recharge of 2 in/yr over the entire model area (RMSE = 14.5 feet for all layers);

and Variable recharge rates (3 in/yr on Haw Ridge and mid-valley, and 10 in/yr on Copper Ridge) totaling about 2 in/yr, and variable hydraulic conductivity with depth and geologic unit (RMSE = 11.5 feet for all layers).

Simulations produced lower RMSE values, when uniform, generalized hydraulic-conductivity values were used for each layer than when hydraulic-conductivity values were varied by geologic unit within layers. These results indicate that within the study area, depth is a more important control on hydraulic conductivity than geologic unit.

The calibrated model is non-unique in that other combinations of model-input parameters pro-duced similar, or in some cases lower, RMSE values.

The values chosen for calibration, however, are believed to be the most representative of actual field conditions given the uncertainties of the data, partic-ularly recharge. The non-uniqueness of the model is discussed further in the following section, "Sensitivity Analysis."

All recharge to the ground-water system is by infiltration of precipitation to the water table. A total recharge rate of 2.0 in/yr was used in model 54 Hydrology of Melton Valley at Oak Ridge National Laboratory, Tennessee

OAK RIDGE NATIONAL LABORATORY BETHEL MODEL BdtlNDARY

^

0.5 1 MILE 0.5 1 KILOMETER I

s *I EXPLANATION BURIAL GROUND AND NUMBER 900 WATER-LEVEL CONTOUR Shows simulated water-level altitude. Contour interval 25 feet. Datum Is sea level Figure 32.- Simulated water levels for layer 4.

calibration. This value is less than the average-annual recharge rate of 6.7 in/yr estimated by base-flow analysis, and also is less than the rates calculated in the cross-sectional-model simulations. However, the calibrated recharge rate is within the range of values (1.1 to 2.2 in/yr) estimated on the basis of analysis by Moore (1988b, p. 71).

All ground-water discharge is assumed to be to the surface-water system. Model-calculated ground-water discharge to the Whiteoak Creek drainage system, including Melton Branch and its tributaries, is 0.39 ft3/s (table 11).

Although the average-annual ground-water seepage to streams is not known, the model-calculated rate is within the range believed to be realistic for the study area. Seepage analyses conducted during this study consistently showed total ground-water seepage to the Whiteoak Creek drainage system that was less than 1 ft3/s. Because the seepage was usually within the range of measurement error, the exact value of seepage is not known, but can be assumed to be less than 1 ftVs.

Model-calculated ground-water discharge to other streams within the valley totaled 0.23 ft3/s, and discharge to Whiteoak Lake was 0.03 ft3/s (table 11).

Boundary discharge to the Clinch River and Melton Hill Lake was 0.01 and 0.06 ftVs, respectively (table 11). These rates, although unsubstantiated by field data, are believed to be of the same order of magnitude as the actual rates.

Table 11.

Model-calculated water-budget components for area! model simulations Water-budget component Flow, in cubic feet per second INFLOW Recharge Total inflow 10.72 0.72 OUTFLOW Seepage to:

Whiteoak Creek drainage Whiteoak Lake Other streams Clinch River Melton Hill Lake Total outflow 0.39

.03

.23

.01

.06 0.72 Equivalent to 2.0 inches per year.

The model-calculated distribution of flow with depth is similar to the distributions obtained in the cross-sectional-model analysis.

About 9 percent of simulated recharge to the ground-water system circulates below layer 1 to depths greater than 50 feet.

About 98 percent of the total model-calculated ground-water flow in the valley is within layers 1 and 2, at depths less than 100 feet, and only 0.2 percent of the flow circulates below layer 3 to depths greater than 250 feet. All of this deep flow, however, is assumed to circulate back up to the surface-water system.

Sensitivity Analysis The response of the model to changes in various model-input parameters was evaluated by sensitivity analysis.

The relative sensitivity of the model to these changes indicates the degree of importance of individual parameters to the simulation of ground-water flow and can provide an indication of the uniqueness of the model calibration. For example, if the simulations produce similar results when a model-input parameter is varied over a large range of values from the calibrated value, then the model is "insensitive" to that parameter and the model solution can be considered as non-unique. Additionally, if the model is insensitive to variations of a particular parameter, then obtaining additional field information to refine knowledge of that parameter would do little to improve the results.

The parameters varied in this sensitivity analysis were recharge, hydraulic conductivity of layer 1, transmissivity of layer 2, row-to-column anisotropy, vertical leakance between layers 1 and 2, and river and drain conductance values.

Each parameter was adjusted uniformly over the entire model area, where applicable, and RMSE was calculated and compared to the calibrated RMSE values.

Except where noted, each parameter was evaluated independently, in that all other parameters were held constant while the tested parameter was varied. Model sensitivity to spatial discretization was not tested because the grid spacing, 150 to 500 feet, was small enough to adequately address the modeling objectives.

The model is sensitive to changes in recharge and layer 1

hydraulic conductivity when these parameters are evaluated independently (fig. 33).

Relatively small decreases in layer 1

hydraulic 56 Hydrology of Melton Valley at Oak Ridge National Laboratory, Tennessee

16 14 12 13o 10 Layer 1 RECHARGE, IN INCHES PER YEAR 16 14 12 10 Layer 1 0.2 0.4 0.6 0.8 HYDRAULIC CONDUCTIVITY, IN FEET PER DAY 1.0 Figure 33. Changes in root mean square error of simulated water levels with respect to changes in (a) recharge rate and (b) layer 1 hydraulic conductivity.

Hydrology of Melton Valley 57

conductivity and small increases in recharge produced large increases in RMSE.

The model was less sensitive to increases in hydraulic conductivity and decreases in recharge. Although use of a hydraulic-conductivity value of 0.38 ft/d for layer 1 and a recharge rate of 1.5 in/yr produced RMSE values for this layer that were less than the calibrated value, RMSE values of other layers increased when these parameters were used (fig. 33).

When hydraulic conductivity of layer 1 and recharge were varied together, the model was less sensitive to the changes (fig. 34), and the non-uniqueness of the model was shown. If RMSE values of 8 or less for layer 1, and 10 or less for all layers, are considered as indicative of acceptable simulations, then any combination of recharge and hydraulic conductivity of layer 1 that resulted in RMSE values within these limits would be acceptable. Recharge rates from 1.0 to 4.0 in/yr and layer 1 hydraulic-conductivity values from 0.1 to 0.6 ft/d, in the proper combination, could produce acceptable simulations.

A ratio of recharge to hydraulic conductivity of about 6.7 (in/yr)/(ft/d) produces optimal simulation results.

The best results, in terms of lowest RMSE values, were obtained using a recharge rate of 1.0 in/yr and a hydraulic-conductivity value of 0.1 ft/d (fig. 34).

These values may indeed be representative of real-world conditions, but are close to the lower limits of estimates for these parameters.

Because of the relatively large number (228) of slug tests used in the statistical analysis of the hydraulic-conductivity data (fig. 15), the use of the median value of layer 1 hydraulic conductivity, 0.19 ft/d, is believed to be more representative of regolith hydraulic conductivity throughout Melton Valley than a lower value.

Overall, simulations produced the smallest RMSE values when the hydraulic conductivity of layer 1 (regolith) was assumed to be isotropic (row-to-column anisotropy = 1.0). When layer 1 hydraulic-conductivity values were simulated as greater parallel to strike than perpendicular to strike (anisotropy less than 1.0), RMSE values increased (fig.

35a).

Increasing anisotropy values, in which hydraulic conductivity would be greater perpendicular to strike than parallel to strike, was not tested in the sensitivity analysis because this scenario is hydrogeologically unreasonable given the geologic structure of the valley. Varying the anisotropy values of the bedrock units (layers 2, 3, and 4) had no effect on simulation results.

Increasing the transmissivity of layer 2 improved simulation results (lower RMSE values) in layer 1, but produced less accurate results (higher RMSE values) in layer 2 (fig. 35b). Reducing the transmissivity of layer 2 by several orders of magnitude had little effect on results (fig. 35b).

The model was sensitive to reductions in river and drain conductance values (fig. 36a) because the streams are the primary outlets for ground-water flow.

Reducing the connection between the streams and the ground-water system by reducing the conductance increases water levels and gradients to the streams to enable ground water to discharge to the streams.

Increasing the conductance by about twice the initial value produced lower RMSE values by increasing the connection to the streams and ground-water system.

Consequently, simulated water levels in layer 1, which were greater overall than estimated water levels, were lowered, resulting in a closer match and lower RMSE (fig. 36a). Use of a vertical hydraulic-conductivity value for streambeds that is about equal to the median hydraulic conductivity of the regolith may be more appropriate in future modeling in Melton Valley.

Heads in layers 2, 3, and 4 were sensitive to reductions of more than an order of magnitude from calibrated vertical-leakance values between layers 1 and 2, but were insensitive to increases in leakance (fig. 36b). Because the only source of water to layers 2, 3, and 4 is by leakage from layer 1, reducing the connection between layers 1 and 2 greatly reduces heads in lower layers. Most of the discharge from layer 1 is to streams (91 percent) rather than to lower layers, therefore, reduction of vertical leakage has little effect on layer 1 (fig. 36b).

Model Limitations Modeling of ground-water flow in Melton Valley provided some valuable insights into the operation of the flow system. However, results of the simulations must be evaluated within the limitations inherent in the model assumptions and the data used to construct and calibrate the model.

The models assume porous-media flow and, on a valley-wide scale, this assumption seems to be valid. The models do not simulate flow in individual fractures.

Areas in which simulated and estimated water levels differ significantly may indicate areas in 58 Hydrology of Melton Valley at Oak Ridge National Laboratory, Tennessee

10 11.4 6.3

' 7.8 0

0.2 0.4 0.6 0.8 1.0 HYDRAULIC CONDUCTIVITY, IN FEET PER DAY All layers EXPLANATION

-14 LINE OF EQUAL ROOT MEAN SQUARE ERROR Dashed where approx-imately located.

Interval 2 feet 15.0 ROOT MEAN SQUARE ERROR Values In feet 0.2 0.4 0.6 0.8 HYDRAULIC CONDUCTIVITY, IN FEET PER DAY 1.0 Figure 34.--Changes in root mean square error of simulated water levels with respect to simultaneous changes in recharge and hydraulic conductivity.

Hydrology of Melton Valley 59

14 12 10 ALL LAYERS LAYER 1

22

~ 18 cr O

Of 10 0

0.5 1.0 1.5 ROW-TO-COLUMN ANISOTROPY VALUE 0.0001 0.001 0.01 0.1 1.0 10 100 1.000 MULTIPLE OF LAYER 2 TRANSMISSIVITY Figure 35. Changes in root mean square error of simulated water levels with respect to changes in (a) layer 1 row to column anisotropy and (b) layer 2 transmissivity.

which fractures or other structural controls on ground-water flow, which are not simulated in the models, predominate.

Although the model grids are fairly detailed, the scales of the models are too large to adequately address flow or transport problems in individual burial grounds or trenches. The models could be used, however, to provide initial estimates of boundary conditions for more detailed models.

The numerical models support the conceptual model of the ground-water system in Melton Valley in that most flow is at relatively shallow depths.

Because of the paucity of data on hydrologic con-ditions at depths greater than 100 feet, however, quan-titative simulation results for deep flow should be considered as tentative. Ground water may flow through individual fractures or fracture zones at depths of at least 600 feet in Melton Valley, but this deep flow probably represents a small percentage of the total ground-water flow.

Simulated water levels were compared to estimated average-annual water levels for the areal model, so that the average-annual recharge rate simulated in the model (2.0in/yr) also should be considered as an estimate. Calibration of simulated water levels to those measured on June 28, 1988, would require a simulated recharge lower than average. Because the average water-level fluctuation in Melton Valley is relatively small (less than 5 feet),

use of estimated average-annual water levels based on the measured levels of June 28, 1988, probably introduces only a relatively small error in the simulations.

Areal model simulation results are non-unique in terms of total recharge rates and average hydraulic conductivity of the regolith hydrogeologic unit.

A ratio of recharge to hydraulic conductivity of about 6.7 (in/yr)/(ft/d) produces optimal simulation results for recharge values ranging from 1.0 to 4.0 in/yr and average hydraulic conductivity values of the regolith hydrogeologic unit ranging from 0.1 to 0.6 ft/d.

Simulated recharge rates to the ground-water system are consistently lower than rates estimated on the 60 Hydrology of Melton Valley at Oak Ridge National Laboratory, Tennessee

20 15 13 a:

tn 10 ALL LAYERS LAYER 1

234567 89 10 11 MULTIPLE OF RIVER AND DRAIN CONDUCTANCE VALUES 12 GC.o on on o

tn 22 18 U

10 ALL LAYERS 6

0.0001 0.001 6

i 0.01 0.1 1.0 10 100 MULTIPLE OF VERTICAL LEAKANCE VALUE 1.000 10,000 Figure 36. Changes in root mean square error of simulated water levels with respect to changes in (a) river and drain conductance values and (b) vertical leakance between layers 1 and 2.

Hydrology of Melton Valley 61

basis of analysis of streamflow data. Ground-water flow within the stormflow zone (Moore, 1988b,

p. 35), which may discharge fairly slowly to streams, could produce misleading recharge estimates in hydrograph-separation analysis of streamflow data.

Because of this relatively slow discharge from the stormflow zone and the relatively high frequency of rainfall in the Oak Ridge area, the true contribution of ground-water flow from the regolith hydrogeologic unit may not be apparent in the hydrographs. The lack of long-term hydrographs for streams not affected by effluent discharge in Melton Valley adds to the difficulty of estimating recharge from stream-flow records.

Additional long-term streamflow records could greatly aid in estimating true average-annual recharge rates, thereby restricting the range of hydraulic-conductivity values for the regolith hydro-geologic unit that produce optimal simulation results.

Use of uniform hydraulic-conductivity values for a layer, correlated to depth rather than geologic unit, resulted in the best match between simulated and estimated water levels. These results seem to support the conclusion by Moore (1988b, p. 50) that there is no significant difference between hydraulic-conductivity values of major geologic units in the area. Moore's analysis, however, was between the Rome Formation, Conasauga, Knox, and Chicka-mauga Groups, and was not a comparison of hydraulic-conductivity values within units of the Conasauga Group.

For smaller scale, site-specific problems, a more detailed distribution of hydraulic conductivity, and perhaps recharge, would be required.

Limited hydrogeologic information is available on Copper and Haw Ridges. Because these areas are the primary recharge areas, additional water-level and hydraulic data could provide valuable information with which to refine knowledge of hydrologic conditions in the study area. All simulated recharge is assumed to be from precipitation, and no ground water is assumed to enter or leave the valley as underflow. Additional water-level data near boundary areas, particularly at depths greater than 100 feet, could help to validate these assumptions.

The models were calibrated only to estimated average-annual hydrologic conditions.

No attempt was made to evaluate the storage properties of the ground-water system by means of transient calibration.

Application of the present models to problems of transient flow, therefore, would be inappropriate.

Water Chemistry and Radionuclides Samples of water from streams and wells in the study area (fig. 37) were collected for chemical analy-ses to aid in describing the ground-water-flow system and to determine possible effects on local water resources of leachate migrating from radioactive waste disposal trenches. Waste materials in the burial grounds have been described as mostly solid waste products having low levels of radioactivity.

These materials have been dumped in unlined trenches that have been capped by spoils from the excavations.

Ground water in burial ground 6 was reported to have activities of as much as 97,200 pCi/L strontium-90 (90Sr), 506,000 pCi/L gross beta, 9,100,000 pCi/L tritium, plus that of other radionuclides (Solomon and

others, 1988).

Similar levels of activity for radionuclides in leachates and ground water near ILW Trench 7, one of the trenches in Melton Valley in which intermediate-level liquid wastes formerly were disposed, also have been described (Olsen and others, 1983).

Samples collected from streams and ground water in 1985 and resampled in 1987 were processed according to standard methods of the U. S. Geological Survey (Brown and others, 1970; Claassen, 1982; Skougstad and others, 1979).

Historical major-ion data collected by USGS personnel for water from wells near trenches in 1983 are included in the data tables and subsequent computation of means of the data. Analyses were performed at USGS laboratories for major ions in 1983 and radiochemical constituents in

1985, and at Tennessee Valley Authority laboratories for major ions in 1985 and all constituents in 1987.

Radiochemical analyses were made for dissolved gross alpha, gross beta, 90Sr, and tritium.

Differences in gross alpha activity levels between the 1985 and 1987 samples are primarily a consequence of the use of different laboratory methods. The standard used for the 1985 analyses was natural uranium, whereas the standard used for the 1987 analyses was americium-241.

The gross alpha activities were so small in most samples that the possible errors from counting uncertainty in the analyses equaled or exceeded most gross alpha values.

The gross alpha data, therefore, are not considered in the analyses discussed here.

The on standard for all analyses of gross beta was Sr. The gross-beta analyses do not include the low-energy beta particles emanating from tritium in the samples.

62 Hydrology of Melton Valley at Oak Ridge National Laboratory, Tennessee

Jfc.s*.

fc.

OAK RIDGE

\\BETHEL NATIONAL LABORATORY V

• i i £ Y UC1 rG^A-X^UDI UC2 n BG2 fh*t*oftfcx

^x"~"

^

""O¥GI RiDG£ Creefc M £ L r o N "^

Iffviteoo-fe COPPER

(

V A L L RIDGE 0.5 1 MILE 0.5 1

KILOMETER EXPLANATION

'M2 BURIAL GROUND AND NUMBER INTERMEDIATE-LEVEL LIQUID WASTE TRENCH AND NUMBER STREAM SAMPLING SITE M

designation is for Melton Branch and W designation Is for White-oak Creek UAI WELL SAMPLING SITE 1 to 3 wells WELL SAMPLING SITE 4 to 5 wells; A, wells 439, 440, 458, 459, and 460; B, wells 461, 462, 463, and 464; C, wells 469A, 470, 471, and 472; D, wells 473, 474, 475, and 476 Figure 37.--Location of water-quality sampling sites and burial grounds.

Surface-Water Quality Samples were collected from five surface-water sites in Melton Valley for analyses of principal ions and radionuclides (fig. 37). Samples were collected at the notches in weir plates or flumes when all flow was through the notches.

Results of the analyses are presented in table 12.

Major-ion chemistry in surface water Surface water in Melton Valley is principally a calcium bicarbonate type at station Ml, and becomes a calcium bicarbonate sulfate type at station M2 (table 12).

Concentrations of all of the major ions, and values for properties except alkalinity, increased between the two stations in the samples of 1985.

Dissolved solids increased from 200 mg/L to 450 mg/L. No sample was collected at Ml in August 1987 because the streambed was dry at that location.

The increases in concentrations of major ions between stations Ml and M2 during the April 1985 sampling are probably due to effluent from the ORNL facility located northwest of station Ml.

The major-ion concentrations downstream at stations W3, W4, and W5, on Whiteoak Creek, are influenced by effluent from other ORNL facilities, as is indicated by the present radiochemical data and documented in reports of earlier studies.

The water-quality data from streams are not considered in the remainder of this section because of their limited use in describing the chemistry of natural streamflow and of natural discharge from the ground-water system.

Radionuclides in surface water Station Ml, located at the upstream end of Melton Branch, is the only surface-water sampling site that does not receive effluent from ORNL facilities. In the April 1985 sampling, 90Sr activity at Ml was less than the detection limit of 0.4 pCi/L, gross beta activity was 2.2 pCi/L, and tritium was 3,900 pCi/L. At M2, 90Sr activity had increased to 420 pCi/L; gross beta, to 880 pCi/L; and tritium, to 2,900,000 pCi/L. Activities smaller than those at M2 were measured in samples from the downstream stations on Whiteoak Creek. In the 1987 sampling, the activity level of each constituent was substantially less than the corresponding value in the 1985 sampling.

The 2.2 picocuries per liter (pCi/L) gross beta at station Ml (table 13) is probably the natural background level of activity for Melton Branch. The 3,900 pCi/L of tritium at Ml probably is caused by tritium released into the air and transported to the Table 12. Concentrations of major ions in water from streamflow stations in the vicinity of burial grounds 4, 5, and 6

[Analyses by Tennessee Valley Authority Laboratory]

Station desig-nation 1 W3 W3 W4 W4 W5 W5 Date sam-pled 4/85 8/87 4/85 8/87 4/85 8/87 Cal-ium (mg/L) 45.5 37.0 45.9 36.1 53.4 40.1 Mag-nesium (mg/L) 11.2 9.5 11.0 9.5 12.1 9.9 Potas-sium (mg/L) 1.9 1.6 1.8 1.7 2.3 1.8 Sod-ium (mg/L) 12.0 11.0 13.0 12.0 16.0 13.0 Alkali-nity as CaCO3 (mg/L) 116 108 118 112 121 114 Chlo-ride (mg/L) 9.0 10.0 9.0 10.0 11.0 12.0 Sul-fate (mg/L) 31.0 23.0 32.0 29.0 58.0 27.0 Dis-solved solids (mg/L) 190 130 200 220 240 230 PH (stan-dard units) 7.7 8.0 7.7 8.2 7.7 8.9 M1 M2 M2 4/85 4/85 8/87 65.8 100 43.2 6.5 22.1 9.9 1.2 3.6 1.8 2.6 16.0 9.2 172 108 124 3.0 16.0 12.0 10.0 250 28.0 200 450 240 7.7 8.4 8.8 locations shown on figure 37.

64 Hydrology of Melton Valley at Oak Ridge National Laboratory, Tennessee

Table 13. Activities of radiochemical constituents in water from streamflow stations and wells distant from trenches

[Values in parentheses are possible error due to accounting uncertainties and are + the measured value; activities without values in parentheses indicate that possibleerror was not reported by lab. Analyses in 1985 by U.S. Geological Survey Laboratoryand analyses in 1987 by Tennessee Valley Authority Laboratory. pCi/L, picocuries perliter; <, less than]

Station desig-nation 1 W3 W3 W4 W4 W5 W5 Date sam-pled 4/85 8/87 4/85 8/87 4/85 8/87 Gross alpha2 (pCi/L) 20 4.4 45 4.1 37 4.1 (4.6)

(4.2)

(4.3)

Gross beta3 (pCi/L) 320 179 330 152 390 204 (14.1)

(12.7)

(15.3)

Strontium-90 (pCi/L) 130 53.0 130 63.6 200 104 (8.5)

(4.1)

(12.5)

Tritium4 (PCi/L) 0.34E+05 0.57E+04 0.45E+05 (0.30E+03) 0.58E+04(0.31E+03) 0.41 E+06 0.61 E+05 (0.30E+04)

M1 M2 M2 4/85 4/85 8/87 3.3 49 4.3 (4.5) 2.2 880 289 (19.6)

<0.4 420 135 (15.3) 0.39E+04 0.29E+07 0.89E+06 (0.45E+05) 1 Locations shown on figure 37.

2Standard for analyses of samples collected in 4/85 was natural uranium, and standard for analyses of samples collected in 4/87 was americium-241.

3Standard for all gross beta analyses was strontium-90.

4E is exponent to base 10.

stream by rainfall and runoff. Atmospheric tritium releases at ORNL have been documented annually in environmental monitoring reports, and a survey was made in 1986 of tritium activities in soil moisture and atmospheric moisture above ground level (Amano and others, 1987).

Although tritium activity in atmos-pheric precipitation was not reported, tritium in excess of 400 pCi/L was measured in atmospheric moisture (Amano and others, 1987).

The radiochemical data from Melton Branch and Whiteoak Creek also are of limited use in describ-ing natural streamflow chemistry because of effluent discharge from ORNL facilities. The effluent from the facility northwest of station Ml (fig. 37) probably caused the increased gross beta, ^Sr, and tritium activities at station M2 on Melton Branch (table 13).

Waste radionuclides are dissolved in ground water discharged from the burial ground 5 area (Steuber and others, 1978; Cerling and Spalding, 1981; Webster and Bradley, 1988, p. 102), and also contribute to the radioactivity in water from Melton Branch at station M2. Stations W3, W4, and W5 on Whiteoak Creek (fig. 37) receive effluent from other ORNL facilities, and this is reflected in above-background activities of all radiochemical constituents (table 13) for these stations.

Ground-Water Quality Water samples were collected from 31 wells in Melton Valley in April 1985, and August-September 1987, for analyses of major ions and radionuclides (fig. 37). The samples were withdrawn from the wells using a submersible pump.

Nineteen of the wells were located near radioactive-waste disposal trenches and were evacuated at least three times in the years between well installation and sampling in 1985 or 1987. Twelve wells located away from the trenches had been evacuated at least twice in the period between well completion and the sampling period.

During installation of the 12 distant wells, water for drilling and water for cementing were spiked with 13 mg/L of the tracer sodium ben/oate (Malcolm and others, 1980). Samples collected in September 1987, after several evacuations of these wells, contained less than 0.7 mg/L sodium benzoate. The samples in 1987 Hydrology of Melton Valley 65

from most of the wells are assumed, therefore, to represent natural ground water. Concentrations of some of the constituents in at least three wells (UB2, UC2, and UG3), however, indicate possible lingering effects of water used for well construction.

Nine of the 12 wells distant from trenches were completed in regolith and have open intervals at depths of 86 feet or less (table 14). Three of the wells distant from trenches were completed in bedrock.

Other distant wells completed in bedrock were not sampled because their water levels recovered slowly and useful potentiometric head data would have been forfeited as a result of evacuation and sampling during the time span of the project.

Of the 19 wells near trenches, 5 were com-pleted in regolith and have open intervals at depths of 36 feet or less (table 14). Twelve were completed in bedrock and have open intervals at depths of 88 feet or more. Wells 108 and 109, which have open intervals from 40 to 126 feet and 38 to 126 feet, respectively, are completed in both regolith and bedrock.

Most wells near trenches were completely evacuated before sampling, but discharge volume was limited in some of the wells completed in bedrock.

About 5 gallons of water were pumped from wells 461, 463, 469A, 470, and 473 before sampling in 1985 and 1987 because water levels in these wells recovered slowly from discharges, and like the wells distant from trenches, useful potentiometric head data would have been forfeited at these sites had the wells been evacuated. About 3 gallons of water were pumped from well 476 at the time of sampling because the water was expected to contain several thousand picocuries per liter 90Sr, which posed a potential health hazard to field personnel.

The assumption was made, for purposes of data comparison, that water chemistry in samples collected from wells near trenches in late August 1987 would not have been significantly different from that 3 weeks later in mid-September 1987, when samples were collected from wells distant from trenches.

PH The pH of ground-water samples ranged from 6.4 to 11.8. Of the 59 analyses included in table 14, 35 had a pH of less that 8.9. The pH was greater than 9.0 in at least one of the samples collected from each of the following wells near trenches: 458, 459, 461, 462, 463, 469A, 470, 473, 474, and 475 (table 14).

Samples having the highest pH values were from wells 461, 463, 469A, 470, and 473. These five wells had the least number of evacuations and smallest volumes of water withdrawn for sampling. They also are among those having the smallest hydraulic-conductivity values (Webster and Bradley, 1988,

p. 71), which reflects the slow rate of water movement in the rock surrounding the open interval of each well. These two factors may indicate that water used in cementing the casings remains at or near the wells.

There also is the possibility that the high pH results from a reaction between ground water and the cement seal. The bottom of the seal in each of these wells is at the bottom of the casing where the seal is exposed to water, and the leaching potential of water, all of the time. A pH greater than 9.0 also could indicate that water (leachate) flows from the waste burial trenches to the wells.

In samples from wells 459, 469A, and 475, the pH increased by more than 0.5 unit to 9.2 or greater during the period 1983 to 1987 (table 14).

The reasons for the increase have not been determined, but contributing factors could have been contact of ground water with the cement seals and the migration of leachate from trenches to the wells. However, if contact of ground water with the seals was solely responsible for the increases in pH, the pH probably should have decreased, rather than increased, with time, by dilution from recharge after each evacuation and sampling. If leachates are solely responsible, the pH of the trench fluid would have to be strongly alkaline (pH 10 to 11) to cause the observed increases in the pH of the well water. Such high pH values are suspect in view of values ranging from 5.8 to 8.0 for 21 samples of trench fluid collected at 13 trenches in burial ground 6 (Solomon and others, 1988, p. 19-20),

which has wastes generally similar to those in burial ground 5.

The pH was greater than 9.0 in water from wells UB2, UC2, and UG3, all distant from trenches, and probably is not representative of the pH of natural ground water. Some of the water from drilling and cementing likely remained near the wells at the time of sampling and was included in the samples, even though the wells were evacuated twice before sampling and little of the benzoate tracer was found in the samples.

66 Hydrology of Melton Valley at Oak Ridge National Laboratory, Tennessee

Table 14. Concentrations of major ions in water from wells in the vicinity of burial grounds 5 and 6

[Analyses in 1983 by U.S. Geological Survey Laboratory and analyses in 1985 and 1987 by Tennessee Valley Authority Laboratory. mg/L, milligrams per liter; CKMG, Chickamauga Limestone; NCCK, Nolichucky Shale; MRVL, Maryville Limestone; PPKV, Pumpkin Valley Shale; RGVL, Rogersville Shale; ROME, Rome Sandstone. Formation not determined for wells 108 and 109]

Well 1 Open inter-val Forma-(feet) tion2 Date sam-pled Cal-cium (mg/L)

Mag-nesium (mg/L)

Potas-sium (mg/L)

Sod-ium (mg/L)

Alkalin-ity as CaCO3 (mg/L)

Chlo-ride (mg/L)

Sul-fate (mg/L)

Dis-solved solids (mg/L) pH (stan-dard units)

Wells near trenches 439 439 439 440 440 440 458 458 458 459 459 459 460 460 460 461 461 462 462 463 463 464 464 469A 25-35 25-35 25-35 26-36 26-36 26-36 188-203 188-203 188-203 128-140 128-140 128-140 90-100 90-100 90-100 191-201 191-201 141-151 141-151 88-100 88-100 6-11 6-11 191-201 469A191-201 469A 470 470 471 471 472 472 473 473 473 474 474 474 475 475 475 476 476 108 108 109 109 191-201 141-151 141-151 89-99 89-99 15-20 15-20 190-200 190-200 1 90-200 142-151 142-151 142-151 91-100 91-100 91-100 25-30 25-30 40-126 40-126 38-126 38-126 MRVL MRVL MRVL MRVL MRVL MRVL MRVL MRVL MRVL MRVL MRVL MRVL MRVL MRVL MRVL MRVL MRVL MRVL MRVL MRVL MRVL NCCK NCCK RGVL RGVL RGVL MRVL MRVL MRVL MRVL MRVL MRVL MRVL MRVL MRVL MRVL MRVL MRVL MRVL MRVL MRVL MRVL MRVL 8/83 4/85 8/87 8/83 4/85 8/87 8/83 4/85 8/87 8/83 4/85 8/87 8/83 4/85 8/87 4/85 8/87 4/85 8/87 4/85 8/87 4/85 8/87 8/83 4/85 8/87 4/85 8/87 4/85 8/87 4/85 8/87 8/83 4/85 8/87 8/83 4/85 8/87 8/83 4/85 8/87 4/85 8/87 4/85 8/87 4/85 8/87 150 146 144 150 148 141 8.8 6.3 6.4 15.0 13.7 10.2 92.0 105 125 1.3

.9 1.1 1.5 1.4 1.1 105 77.0 1.1

.6

.6 3.5 1.8 36.0 25.0 107 150 110

.7

.6

.8 1.6 1.4 12.0 12.7 4.2 147 147 54.6 57.4 66.0 60.7 15.0 14.5 15.1 17.0 19.1 23.5 7.7 7.2 7.1 13.0 6.3 3.9 40.0 54.6 53.9

.4

.4

.4

.4

.3

.1 16.1 11.5

.0

.1

.0

.1

.0 19.0 19.1 21.9 21.8

<.1

.0

.0

.0 1.9

.6 20.0 24.0 19.4 20.4 19.5 6.8 8.5 13.0 31.0 0.1 1.2 1.2 2.1 2.0 2.3 4.4 5.0 5.3 11.0 9.2 8.3 6.0 11.0 7.0 17.0 9.1 3.9 2.3 3.9 1.6 1.6 1.4 7.1 5.6 3.0 21.0 9.8 8.2 7.8 2.3 2.1 44.0 8.1 4.6 13.0 6.1 4.5 4.7 7.5 7.5 2.0 1.7

.8

.9 1.2 3.6 26.0 28.0 25.0 18.0 20.0 23.0 65.0 54.0 57.0 14.0 14.0 12.0 12.0 14.0 13.0 350 360 220 240 140 140 14.0 11.0 190 190 210 170 200 18.0 15.0 44.0 43.0 300 230 250 110 110 140 20.0 26.0 23.0 24.0 24.0 4.1 4.5 5.1 12.0 250 400 442 420 366 430 200 160 172 82.0 84.0 68.0 220 182 200 708 714 454 466 346 274 312 250 417 383 388 398 390 184 164 500 556 1,010 465 470 248 247 250 138 212 174 389 438 146 164 192 264 22.0 17.0 21.0 36.0 30.0 32.0 2.2 2.0 2.0 32.0 25.0 24.0 170 225 260 5.0 4.0 2.0 2.0 2.0 1.0 26.0 26.0 2.1 2.0 2.0 2.0 2.0 5.0 4.0 28.0 23.0 3.4 2.0 2.0 1.5 2.0 2.0 1.8 1.0 2.0 23.0 25.0 2.0 3.0 4.0 14.0 24.0 21.0 24.0 33.0 28.0 26.0 12.0 5.0 2.0 23.0 8.0 6.0 7.6 16.0

<.1 15.0 29.0 13.0 11.0 8.0 11.0

<.1

<.1 30.0 29.0 24.0 16.0 17.0 14.0 12.0 3.0 1.0 46.0 43.0 45.0 15.0 14.0 8.0 22.0 14.0 4.0 6.0 19.0 8.0 13.0 14.0 16.0 573 350 310 540 350 550 227 160 190 172 90 120 525 610 980 610 800 500 560 300 340 220 380 499 450 470 440 470 190 200 300 600 1,040 570 560 296 290 300 164 160 150 310 520 180 230 170 520 6.7 6.6 6.6 6.4 6.5 6.9 7.9 9.2 8.8 9.2 9.6 10.7 6.8 7.4 7.5 10.2 10.0 9.6 9.9 10.1 10.6 7.4 7.2 10.2 10.4 11.0 11.3 11.8 8.1 8.4 6.9 6.7 11.8 11.2 11.3 10.6 9.5 9.9 8.6 8.7 9.2 6.5 6.7 7.7 7.5 7.3 7.5 Hydrology of Melton Valley 67

Table 14. Concentrations of major ions in water from wells in the vicinity of burial grounds 5 and 6-Continued Open inter-val Forma-Well1 (feet) tion2 Date sam-pled Cal-cium (mg/L)

Mag-nesium (mg/L)

Potas-sium (mg/L)

Sod-ium (mg/L)

Alkalin-ity as CaCO3 (mg/L)

Chlo-ride (mg/L)

Sul-fate (mg/L)

Dis-solved solids (mg/L)

PH (stan-dard units)

Wells distant from trenches UA1 42-51 UB1 26-36 UB2 101-126 UC1 77-86 UC2 188-206 UD1 17-29 UE1 69-76 UF1 17-24 UG1 25-32 UG3 180-200 UH1 19-26 UI1 18-25 NCCK MRVL MRVL ROME CKMG PPKV MRVL MRVL NCCK MRVL NCCK NCCK 9/87 9/87 9/87 9/87 9/87 9/87 9/87 9/87 9/87 9/87 9/87 9/87 29.2 27.5

.9 20.6 6.1 38.7 83.7 59.7 33.9

.9 47.8 21.0 12.6 7.1

.1 21.2 6.8 13.2 14.5 10.6 9.8 3

7.9 11.9 3.5 1.5 4.4 3.8 16.0 2.4 2.2 3.6 2.8 2.2 1.4 4.2 19.0 8.0 110 3.6 40.0 14.0 6.8 10.0 11.0 220 11.0 70.0 130 90.0 220 152 55.0 170 364 208 150 684 166 254 24.0 6.0 20.0 2.0 55.0 4.0 8.0 4.0 27.0 6.0 6.0 3.0 16.0 10.0 12.0 12.0 12.0 11.0 13.0 9.0 16.0 19.0 16.0 14.0 190 130 290 220 180 200 300 240 180 530 200 300 7.6 7.6 10.2 8.2 9.5 7.9 7.3 7.7 8.6 9.6 7.5 7.9 locations shown on figure 37.

formation at open interval of well.

3Laboratory value.

Calculated value.

Major-Ion Chemistry of Ground Water The ground waters sampled contained con-centrations of dissolved solids ranging from 90 to 1,040 mg/L. Most samples contained concentrations of dissolved solids below 500 mg/L.

Concentrations of major ions are significantly different for water samples collected from depths greater than 100 feet, compared to samples from shal-lower depths. Major-ion chemistry of most samples from wells of less than 100-foot depth indicates a calcium bicarbonate and calcium magnesium bicar-bonate type water, whereas the chemistry of most samples from wells deeper than 100 feet reflect sodium bicarbonate type water (fig. 38). There are exceptions, however. Samples from well 463, open from 88 to 100 feet, are more characteristic of water from depths greater than 100 feet, and may reflect the upwelling of water from deeper rock to places of discharge into Melton Branch nearby, as indicated by potentiometric-head data.

Samples from well UI1, open from 18 to 25 feet, and well 459, open from 128 to 140 feet, appear to represent a mixture of water from both the shallow and deeper zones. Water from wells UC2 and 460 contains significant percentages of dissolved chloride. The chloride at well UC2 may be an artifact of well construction.

At well 460, the increasing chloride concentrations over time probably indicate the presence of leachate from trenches in the ground water of that area.

The difference in major-ion chemistry of water at about 100 feet depth indicates that two flow pat-terns are present in the subsurface. Shallow ground-water circulation (less than 100 feet below land surface) is in regolith and upper bedrock.

Deeper ground-water circulation (greater than 100 feet below land surface) is in bedrock only.

This concept of differences in water chemistry between shallow and deeper zones has been previously described in Melton Valley (Webster and Bradley, 1988) and in Bear Creek Valley (Bailey and Lee, 1991).

Chemical differences are consistent with the conceptual model of ground-water circulation where shallow ground water has less "time-in-contact" with the aquifer minerals and thus, less chemical evolu-tion. Deeper ground water has more time-in-contact with aquifer minerals and thus, more chemical evolution. A plausible pathway of chemical evolution in bedrock is from a calcium bicarbonate water chemistry to a sodium bicarbonate water chemistry.

68 Hydrology of Melton Valley at Oak Ridge National Laboratory, Tennessee

459 EXPLANATION WELL WITH BOTTOM OF OPEN INTERVAL AT DEPTH OF 100 FEET OR LESS WELL WITH TOP OF OPEN INTERVAL AT DEPTH GREATER THAN 100 FEET SELECTED WELLS IDENTIFIED BY NUMBER PERCENTAGE REACTING VALUES Figure 38. Trilinear diagram of principal ion data of water from 31 wells in Melton Valley.

Means of data are plotted for those wells with multiple analyses.

Insufficient chemical data are available to adequately assess the effects of the different geologic units containing the water.

Radionuclides in Ground Water Samples for radionuclide analyses were col-lected from wells at the same times as the samples for major-ion determinations. Gross alpha activity in the samples of 1985 ranged from 2.6 to 1,000 pCi/L; gross beta, from 2.8 to 20,000 pCi/L; 90Sr, from 0.4 to 11,100 pCi/L; and

tritium, from 1,100 to 60,000,000 pCi/L (table 15). In the samples of 1987, gross alpha activity ranged from 0 to 9.3 pCi/L; gross beta, from 1.6 to 10,000 pCi/L; ^Sr, from 0 to 10,000 pCi/L; and

tritium, from 170 to 360,000,000 pCi/L. The largest values are associated with water from regolith wells near trenches in burial ground 5; the smallest values are associated mainly with wells distant from trenches and wells 108 and 109, lower regolith-bedrock wells in burial ground 6.

The change in gross beta activity between sampling periods in 1985 and 1987 was not more than a factor of two for all wells near trenches, except for wells 440, 464, and 474. The change in tritium activity was not more than a factor of five in all wells near trenches, except for wells 463, 464, 469A, 471, 472, 108, and 109.

To help determine background levels of gross beta and tritium in ground water, results of the analyses for these constituents in water from all wells sampled in 1987 are plotted as bar graphs (fig. 39a and b). The 1987 period is used because more wells were sampled then than in 1985.

For plotting purposes, tritium activities of less than the minimum detectable level of 150 to 200 pCi/L are plotted as 200 pCi/L.

Natural background activities of gross beta in ground water probably are less than about 5 pCi/L, based on the lowest activity in water from wells sampled (fig. 39a). The lowest gross beta activities were measured in samples from wells 462, 463, 108, and wells distant from trenches (except UC2). The somewhat greater activity of gross beta at well UC2, Hydrology of Melton Valley 69

Table 15. Activities of radiochemical constituents in water from wells in the vicinity of burial grounds 5 and 6

[Values in parentheses are possible error due to counting uncertainties and are +/- the measured value; blanks indicate not reported by lab.

Analyses in 1985 by U.S. Geological Survey Laboratory and analyses in 1987 by Tennessee Valley Authority Laboratory. -, no value; pCi/L, picocuries per liter]

Well1 439 439 440 440 458 458 459 459 460 460 461 461 462 462 463 463 464 464 469A 469A 470 470 471 471 472 472 473 473 474 474 475 475 476 476 108 108 109 109 UA1 UB1 UB2 UC1 UC2 UD1 UE1 UF1 UG1 UG3 UH1 UI1 Date sam-pled 4/85 8/87 4/85 8/87 4/85 8/87 4/85 8/87 4/85 8/87 4/85 8/87 4/85 8/87 4/85 8/87 4/85 8/87 4/85 8/87 4/85 8/87 4/85 8/87 4/85 8/87 4/85 8/87 4/85 8/87 4/85 8/87 4/85 8/87 4/85 8/87 4/85 8/87 9/87 9/87 9/87 9/87 9/87 9/87 9/87 9/87 9/87 9/87 9/87 9/87 Open in-terval (ft) 25-35 25-35 26-36 26-36 188-203 188-203 128-140 128-140 90-100 90-100 191-201 191-201 141-151 141-151 88-100 88-100 6-11 6-11 191-201 191-201 141-151 141-151 89-99 89-99 15-20 15-20 190-200 190-200 142-151 142-151 91-100 91-100 25-30 25-30 40-126 40-126 38-126 38-126 42-51 26-36 101-126 77-86 188-206 17-29 69-76 17-24 25-32 180-200 19-26 18-25 Gross alpha2 (PCi/L) 9.5 0.0 7.5 0.0 3.5 0.0 2.6 0.4 9.5 0.0 14 1.1 9.5 0.9 0.2 32 2.1 8.8 0.0 9.5 0.2 3.9 0.6 100 0.0 12 0.6 61 1.9 4.8 0.7 1,000 9.3 3.7 0.6 4.8 0.0 1.7 0.9 0.7 0.5 0.7 0.0 0.2 1.1 0.0 1.5 0.2 0.4 (3.9)

(3.9)

(0.4)

(0.5)

(0.8)

(1.2)

(1.0)

(0-6)

(1.2)

(0.7)

(0.8)

(0.6)

(4.0)

(0.9)

(0.9)

(0.6)

(0.6)

(0.6)

(0.9)

(0.9)

(0.9)

(0.7)

(0.7)

(0.6)

(0.7)

(0-9)

(0.8)

(1.3)

(0.7)

(0.8)

Gross beta3 (PCi/L) 31.0 53.5 13.0 55.6 6.8 5.6 22.0 11.5 18.0 15.0 12.0 8.9 6.7 3.4 2.8 460 146 6.4 4.9 17.0 10.0 10.0 8.9 1,200 955 10.0 5.3 450 5.9 3.9 7.5 20,000 10,000 2.8 1.6 3.5 5.2 4.9 3.2 4.7 4.5 17.0 4.1 3.2 4.2 3.9 3.9 2.6 5.0 (7.9)

(8.0)

(0.8)

(1.1)

(1.4)

(1.2)

(0.8)

(0.7)

(8.4)

(0.9)

(1.2)

(1.0)

(53)

(0.9)

(0-9)

(0.9)

(1,000)

(0.6)

(0.9)

(0.8)

(0.7)

(0.8)

(0.8)

(1.4)

(0.8)

(0.7)

(0.8)

(0.8)

(0.9)

(0.7)

(0.8)

Strontium-90 (pCi/L) 1.9 0.0 1.8 11.2 1.8 0.5 9.1 1.1 3.5 5.6 1.6 0.0 1.2 0.3 0.1 270 61.8 0.8 0.2 0.6 0.0 2.6 0.0 610 488 0.6 0.4 0.4 0.0 0.7 0.4 11,100 6,200 0.6 0.0 0.4 0.0 0.6 0.2 1.1 0.0 0.3 0.0 0.0 0.3 0.8 0.3 0.0 0.0 (4.9)

(5.1)

(0.4)

(0.4)

(0.8)

(0.4)

(0.4)

(0.4)

(6.0)

(0.6)

(0-4)

(0.4)

(44)

(0.5)

(0.5)

(0.5)

(890)

(0.4)

(0.5)

(0-3)

(0.4)

(0.4)

(0.3)

(0.4)

(0.4)

(0.4)

(0.4)

(0.5)

(0.4)

(0.5)

(0.4)

Trttlunrvt (pCi/L) 0.60E+08 0.31E+08(0.16E+07) 0.20E+08 0.14E+08 (0.75E+06) 0.17E+04 0.56E+03 (0.61E+02) 0.84E+04 0.59E+04(0.31E+03) 0.21 E+06 0.75E+05 (0.38E+05) 0.57E+04 0.13E+04 (0.92E+02) 0.24E+04 0.14E+04 (0.96E+02) 0.10E+06 0.46E+04 (0.24E+03) 0.50E+08 0.36E+09 (0.18E+08) 0.26E+04 0.15E+06 (0.79E+05) 0.48E+04 0.13E+04 (0.88E+02) 0.48E+06 0.35E+05 (0.18E+05) 0.92E+07 0.31 E+06 (0.15E+05) 0.12E+04 0.86E+04 (0.44E+03) 0.11E+04 0.57E+03 (0.62E+02) 0.91 E+04 0.90E+04 (0.46E+03) 0.20E+08 0.66E+07 (0.33E+06) 0.10E+06 0.11 E+04 (0.81E+02) 0.10E+06 0.20E+04 (0.12E+03) 0.17E+03 (0.49E+02) 0.41 E+03 (0.54E+02)

<0.20E+03 (0.44E+02) 0.19E+04 (0.11E+03) 0.18E+03 (0.49E+02) 0.77E+03 (0.67E+02) 0.10E+04 (0.78E+02) 0.15E+05 (0.77E+03) 0.32E+03 (0.52E+02)

<0.20E+03 (0.44E+02)

<0.20E+03 (0.46E+02)

<0.20E+03(0.44E+02) locations shown on figure 37. Wells 439 through 476 are grouped by location and designated A, B, C, and D on figure 37.

2Standard for analyses of samples collected in 4/85 was natural uranium, and standard for analyses of samples collected in 8/87 was americium-241.

3Standard for all gross beta analyses was strontium-90.

4E is exponent to base 10.

70 Hydrology of Melton Valley at Oak Ridge National Laboratory, Tennessee

160 140

°- 120 UJ oc

§ 100 o

a.

- 80 M

60 v)o OCo Q

40 UJ

_i O

l 20 955 10,000

a. Dissolved gross beta Wells near trenches Wells distant from trenches Concentration 5 picocuries per liter sn K^

10 10 8 oc LU 10 ID-10 2 Wells near trenches

b. Tritium Wells distant from trenches Concentration 500 picocuries per liter iA**>\\*i,*VV^

WELL Natural background concentration less than 5 pCi/L gross beta and 500 pCi/L tritium.

Figure 39. -Concentrations of (a) dissolved gross beta and (b) tritium in ground-water samples collected at Oak Ridge National Laboratory in August and September 1987.

Hydrology of Melton Valley 71

which is distant from trenches and open at the depth interval 188 to 206 feet, may result from incomplete flushing of water used in well construction. The pH of water from this well was 9.5, which indicates possible contamination of the ground water by water used in cementing the well casing.

The background level of tritium activity in ground water at ORNL is difficult to determine. Deep zones having little or no hydraulic communication with the regolith may contain water without atmospheric tritium resulting from nuclear weapons testing.

Shallow wells, and deep wells having hydraulic communication with the regolith, are more likely to produce water with tritium from atmospheric sources, and local sources such as airborne effluent from ORNL facilities, water effluent discharged to streams, and leachates from trenches.

Wells UA1, UB1, and UB2 are located near the Clinch River and, of all wells sampled, are probably least affected by the local sources of radionuclides.

Based on data from these wells, the background level of tritium activity in ground water appears to be less than about 500 pCi/L in the vicinity of ORNL.

Tritium activities are greater than the estimated 500 pCi/L background level in water from four wells distant from trenches: UF1, UC1, UD1, and UE1 (fig. 39). Well UF1 is open to regolith at a depth of 17 to 24 feet, and is located about 75 feet from Melton Branch. The 15,OOOpCi/L of tritium in water from this well is significantly greater than the 3,900 pCi/L in Melton Branch at stream-sampling site Ml.

Wastewater effluent containing large tritium activities has been discharged into Melton Branch between stations Ml and M2, and probably recharges the regolith in the area of wen UF1 during periods of high flow.

The reason that tritium activities are slightly greater than background in wells UC1, UD1, and UE1 is not known.

These wells, however, are open to water-producing zones that are less than 90 feet deep.

It is possible that ground water at that depth may contain elevated levels of tritium from recent recharge from atmospheric precipitation.

Above-background gross beta and tritium activities in water from wells near trenches are probably the result of fluids from disposal trenches mixing with ground water. All samples from wells near trenches exceeded the estimated natural background levels of 500 pCi/L tritium, and most samples from wells near trenches exceeded the estimated natural background level of 5 pCi/L gross beta (fig. 39a and b, and table 15). Analyses for wells 439, 440, 464, 472, and 476 (all shallow wells constructed near trenches) had the greatest gross beta and tritium activities and indicate probable migration of fluids from trenches to the wells. Tritium activities at least 10 times greater than background in at least one of the samples from wells 461, 469A, and 473, all deep wells in bedrock, indicate that leachates from trenches upgradient of these wells have migrated to depths of at least 190 feet, corresponding to the top of the open intervals in these wells. Samples from wells located distant from trenches did not exceed the background levels, except for tritium in wells UC1, UD1, UE1, and UF1, and gross beta in well UC2.

Although data for wells completed at depths of 40 to 90 feet are limited, the gross beta and tritium activities in ground water generally decrease below a depth of 40 feet.

The radiochemical analyses thus indicate that most radionuclide transport is in regolith above the 40-foot depth.

SUMMARY

AND CONCLUSIONS The Oak Ridge National Laboratory (ORNL) began nuclear weapons development in 1943. As a result of that development and later research programs, radioactive waste was produced and continues to be generated. Most of the solid waste is buried in trenches at three radioactive-waste burial grounds, located in Melton Valley south of the ORNL plant. Water flow through these trenches has resulted in the transport of radionuclides from the burial grounds to local streams in Melton Valley.

Most of the flow in the middle and lower reaches of Whiteoak Creek and Melton Branch, which drain Melton Valley, is wastewater from the ORNL water supply system. The water for this system is obtained from a source outside the Whiteoak Creek drainage basin. Most of the wastewater discharges from a sewage-treatment plant at ORNL.

The discharge per unit drainage area along Melton Branch is about three times greater than that along Whiteoak Creek.

Although several possible geologic and hydrologic reasons may account for the difference, the reason for the differences could not be determined with certainty on the basis of data collected during this study. Based on the relation of precipitation to effluent-adjusted base flow, the annual 72 Hydrology of Melton Valley at Oak Ridge National Laboratory, Tennessee

base-flow discharge per square mile of drainage area for a year of normal precipitation (53 inches) is about 60 ft3/s-d on Whiteoak Creek and about 200 fi^/s-d on Melton Branch. The base flow per square mile of drainage area seems to be almost four times greater in the Melton Branch basin than in the Bethel Valley part of the Whiteoak Creek drainage basin.

Melton Valley is underlain by three major stratigraphic units the Rome Formation, the Conasauga Group, and the Knox Group. The valley floor is underlain by the Conasauga Group, which is further subdivided into six formations of alternating shale and limestone lithologies. Regolith, consisting mainly of clay, silt, and rock fragments, has developed over bedrock, and locally is as much as 86 feet thick.

Alluvium, deposited by the Clinch River and averaging 26.5 feet thick, underlies the southwestern part of the study area.

The Copper Creek thrust fault, underlying Haw Ridge, is the major structural feature.

Other structural features include tear faults that cross the valley, bedding-plane faults, folds, and fractures. The complex structure probably influences ground-water flow locally.

The ground-water-flow system is divided into two hydrogeologic units regolith and bedrock. The regolith hydrogeologic unit includes

regolith, alluvium, and partly weathered bedrock at depths of 50 feet or less below land surface. Ground-water flow within the regolith hydrogeologic unit is primarily through porous media, and is unconfined. The shape of the water table is similar to the shape of the land surface, and ground water generally flows from ridges and hills to streams. Vertical components of flow exist, but are small compared to horizontal com-ponents. Recharge to the water table occurs on Haw and Copper Ridges and on low hills in the middle of the valley.

Average-annual recharge rates are not known with certainty, but have been estimated to range from about 1 to 8 in/yr.

The bedrock hydrogeologic unit consists of weathered and unweathered bedrock at depths greater than 50 feet below land surface, and locally may include pockets of regolith.

The unit is further subdivided, on the basis of differences in hydraulic conductivity, into shallow, intermediate, and deep zones. The unit extends to depths of at least 600 feet, where saline water is commonly present. Ground-water flow in bedrock is primarily through bedding-plane openings, open fractures, faults, and solution openings. The shapes of the potentiometric surfaces of the bedrock zones are similar to the shape of the water table; however, directions of ground-water flow in bedrock may be different from flow in the regolith because of the influence of secondary openings.

Recharge to the bedrock is by vertical flow from the regolith, but the amount of flow to the bedrock is small. Discharge from the bedrock unit is upward to the regolith and streams.

Hydraulic conductivity of the rocks was obtained by single-well "slug" tests because the poorly transmissive character of the rocks at ORNL precluded tests by pumping, even at low pumping rates. Aquifer-storage coefficients were not obtained because of the insensitivity of the slug test method for determining this characteristic.

Hydraulic-conductivity values are highly variable, ranging over several orders of magnitude, but generally decrease with increasing depth.

Hydraulic-conductivity values of the regolith hydrogeologic unit range from 6.6 x 10"4 to 6.9 ft/d, and the median value is 0.19 ft/d.

Hydraulic-conductivity values of shallow bedrock (50 to 100 feet deep) range from 1.2 x 10"3 to 2.4 ft/d, and the median value is 0.05 ft/d. The median value of hydraulic conductivity of rock between 100 and 250 feet deep is 0.01 ft/d, and ranges from 1 x 10"5 to 0.19 ft/d.

Hydraulic conductivity at depths greater than 250 feet generally is verylow, and the median value is slightly more than 1

x 10"5 ft/d; however, hydraulic-conductivity values as large as 0.41 ft/d have been reported for depths as great as 1,000 feet in other parts of the Oak Ridge Reservation that are underlain by the same geologic units. Values of hydraulic conductivity are believed to be greater in a direction parallel to strike than perpendicular to strike; however, this concept was not supported by ground-water-flow modeling analysis.

Both cross-sectional and areal ground-water-flow models were used to provide a better understanding of the flow system of Melton Valley.

The results of simulation made using the models indicated that from 91 to 96 percent of the recharge to the water table flows within the upper 50 feet of the ground-water system, and 97 percent of the flow is within the upper 100 feet. Less than 1 percent of the total ground-water flow circulates to depths greater than 250 feet. Recharge rates are not known with certainty, because the simulation results are non-unique with respect to recharge and hydraulic conductivity. However, simulated recharge rates are Summary and Conclusions 73

low, and range between 1.0 and 4.7 in/yr. The models simulated estimated average-annual, steady-state ground-water-flow conditions, so that application of the model to problems of transient ground-water flow may not be appropriate.

Water in Melton Branch is of the calcium bicarbonate and calcium bicarbonate sulfate types with concentrations of dissolved solids ranging from 200 to 450 mg/L. Between stations Ml and M2 representing headwaters and downstream reaches, the stream receives effluent discharged from an ORNL facility, and the concentrations of most principal ions increase.

Major-ion concentrations of water in Whiteoak Creek, downstream of Melton Branch, also are influenced by effluent discharges from other ORNL facilities.

Analyses of surface water, therefore, were of limited use in describing the chemistry of natural streamflow.

Shallow ground water (less than 100 feet deep) was dominated by

calcium, magnesium, and bicarbonate ions. Deeper ground water (greater than 100 feet deep) was dominated by sodium and bicarbonate ions.

This indicates that two flow patterns are present in the subsurface.

Shallow ground-water circulation, at depths of less than 100 feet, is in regolith and upper bedrock. Deeper ground-water circulation, at depths of greater than 100 feet, is in bedrock only and has undergone greater geochemical evolution.

Radiochemical constituents from water in wells included 90Sr (ranging from 0 to 11,100 pCi/L), gross beta (ranging from 1.6 to 20,000 pCi/L), and tritium (ranging from less than 170 to 360,000,000 pCi/L).

Background activities of gross beta and tritium for ground water were 5 and 500 PCi/L, respectively.

Radiochemical constituents having activity levels much greater than background indicate mixing of ground water with leachates from disposal trenches. The radiochemical data from burial grounds 5 and 6, although limited in the depth interval of about 40 to 90 feet, indicate that most radionuclide transport is in the regolith overlying bedrock above a depth of about 40 feet. Radionuclides from trenches in burial ground 5, however, have been transported to depths of at least 190 feet at wells 461,469A, and 473.

74 Hydrology of Melton Valley at Oak Ridge National Laboratory, Tennessee

SELECTED REFERENCES Amano, H., Garten, C.T.Jr., and Lomax, R.D., 1987, A field survey of environmental tritium in areas adjacent to ORNL solid-waste storage areas: U.S. Department of Energy, Oak Ridge National Laboratory (Report) ORNL/TM-10438, Environmental Sciences Division Publication no. 2914, 31 p.

Bailey, Z.C. and Lee R.W., 1991, Hydrogeology and geo-chemistry in Bear Creek and Union Valleys, near Oak Ridge, Tennessee: U.S. Geological Survey Water-Resources Investigations Report 90-4008, 72 p.

Brown, Eugene, Skougstad, M.W., and Fishman, M.J., 1970, Methods for collection and analysis of water samples for dissolved mineral and gases:

U.S. Geological Survey Techniques of Water Resources Investigations, Book 5, Chapter Al, 160 p.

Cerling, T.E., and Spalding, B.P., 1981, Area distribution of 60Co, 137Cs, and 90Sr in streambed gravels of White Oak Creek watershed, Oak Ridge, Tennessee: U.S. Department of Energy, Oak Ridge National Laboratory (Report) ORNL/

TM-7318,67p.

Claassen, H.L., 1982, Guideline and techniques for obtaining water samples that accurately represent the water chemistry of an aquifer:

U.S. Geological Survey Open-File Report 82-1024,49 p.

Connell, J.F., and Bailey, Z.C., 1989, Statistical and simulation analysis of hydraulic-conductivity data for Bear Creek and Melton Valleys, Oak Ridge Reservation, Tennessee: U.S.

Geological Survey Water-Resources Investigations Report 89-4062,49 p.

Cooper, H.H., Jr., Bredehoeft, J.D., and Papadopulos, I.S., 1967, Response of a finite-diameter well to an instantaneous charge of water: Water Resources Research, v. 3, no. 1,

p. 263-269.

Davis, B.C., Boegly, J.W., Jr., Rothschild, E.R., Spaulding, B.P.,

Vaughan, N.D., Haase, C.S., Huff, D.D., Lee, S.Y., Walls, B.C., Newbold, J.D.,

and Smith, E.D.,

1984, Site characterization techniques used at a low-level waste shallow land burial field demonstration facility: U.S.

Department of Energy, Oak Ridge National Laboratory (Report) ORNL/TM-9146, Environmental Sciences Division Publication no. 2306, 173 p.

Dreier, R.B., Haase, C.S., Beaudoin, C.M., King, H.L., and Switek, J., 1987, Summary of geologic data in the vicinity of the hydrofracture facilities: U.S. Department of Energy, Oak Ridge National Laboratory (Report) ORNL/RAP/LTR-87/26, 15 p.

Dreier, R.B., and Toran, L.E., 1989, Hydrogeology of Melton Valley determined from hydraulic head measuring station data:

U.S. Department of Energy, Oak Ridge National Laboratory (Report)

ORNL/TM-11216, Environmental Sciences Division Publication no. 3338, 205 p.

Fenneman, N.M., 1938, Physiography of eastern United States:

New York,McGraw-Hill, 714 p.

Francis, C.W., and Stansfield, R.G., 1986, Characterization plan for the old hydrofracture facility: U.S. Department of Energy, Oak Ridge National Laboratory (Report) ORNL/

TM-9991, Environmental Sciences Division Publication no.

2701, 75 p.

Glover, R.E., 1964, Ground-water movement: U.S. Bureau of Reclamation, Engineering Monograph 31, 76 p.

Haase, C.S., Switek, J., and Stow, S H., 1987, Geochemistry of formation waters in the lower Conasauga Group at the new hydrofracture facility; preliminary data from the deep monitoring (DM) wells: U.S. Department of Energy, Oak Ridge National Laboratory (Report)

ORNL/RAP-6, Environmental Sciences Division Publication no. 2887, 50 p.

Haase, C.S., Walls, B.C., and Farmer, C.D., 1985, Stratigraphic and structural data for the Conasauga Group and the Rome Formation on the Copper Creek Fault Block near Oak Ridge, Tennessee: Preliminary results from test borehole ORNL-JOY NO. 2: U.S. Department of Energy, Oak Ridge National Laboratory (Report) ORNL/TM-9159, Environ-mental Sciences Division Publication no. 2392, 88 p.

Hoos, A.B., and Bailey, Z.C., 1986, Reconnaissance of surficial geology, regolith thickness, and configuration of the bedrock surface in Bear Creek and Union Valleys, near Oak Ridge, Tennessee:

U.S. Geological Survey Water-Resources Investigations Report 86-4165, 9 p.

Hvorslev, M.J., 1951, Time lag and soil permeability in ground-water observations: Vicksburg, Mississippi, U.S. Army Corps of Engineers Waterways Experiment Station Bulletin 36, 50 p.

King, H.L., and Haase, C.S., 1987, Subsurface controlled geologic maps for the Y-12 Plant and adjacent areas of Bear Creek Valley:

U.S. Department of Energy, Oak Ridge National Laboratory (Report)

ORNL/TM-10112, Environmental Sciences Division Publication no. 2892, 17 p.

____1988, Summary of results and preliminary interpretation of hydrogeologic packer testing in core holes GW-131 through GW-135 and CH-157 Oak Ridge Y-12 Plant:

U.S.

Department of

Energy, Oak

Ridge, Tennessee, Environmental Management Department Health, Y/TS-495, 49 p.

Malcolm, R.L, Aiken, G.R, Thurman, E.M., and Avery, P.A.,

1980, Hydrophilic organic solutes as tracers in groundwater recharge studies: in Contaminants and Sediments, v. 2, edited by Robert A. Baker, Ann Arbor Science Publishers, Inc., Ann Arbor, Michigan, p. 71-88.

McDonald, M.G., and Harbaugh, A.W., 1988, A modular three-dimensional finite-difference ground-water flow model:

U.S. Geological Survey Techniques of Water-Resources Investigations, Book 6, Chapter Al, 576 p.

McMaster, W.M., 1962, Geologic map of Oak Ridge area, Tennessee: Oak Ridge, Tennessee, U.S. Atomic Energy Commission, 1 sheet, scale 1:31,680.

____1967, Hydrologic data for the Oak Ridge area Tennessee:

U.S. Geological Survey Water-Supply Paper 1839-N, 60 p.

Moore, G.K., 1988a, Analysis of hydraulic-conductivity values in fractured rocks near Oak Ridge National Laboratory, Tennessee: U.S. Department of Energy, Oak Ridge National Laboratory, Environmental Sciences Division Publication no. 3090, 13 p.

____1988b, Concepts of groundwater occurrence and flow near Oak Ridge National Laboratory, Tennessee:

U.S.

Department of Energy, Oak Ridge National Laboratory (Report)

ORNL/TM-10969, Environmental Sciences Division Publication no. 3218,96 p.

Selected References 75

____1989, Groundwater parameters and flow systems near Oak Ridge National Laboratory: Oak Ridge National Laboratory (Report)

ORNL/TM-11368, Environmental Sciences Division Publication, no. 3403, 83 p.

Olsen, C.R., Lowry, P.O., Lee, S.Y., Larsen, I.L., and Cutshall, N.H., 1983, Chemical, geological, and hydrological factors governing radionuclide migration from a formerly used seepage trench: A field study: U.S. Department of Energy, Oak Ridge National Laboratory (Report) ORNL/TM-8839, 106 p.

Papadopulos, I.S., Bredehoeft, J.D., and Cooper, H.H., Jr., 1973, On the analysis of 'slug-test' data: Water Resources Research, v. 9, no. 4, p. 1087-1089.

Rorabaugh, M.I., 1964, Estimating changes in bank storage and ground-water contribution to streamflow: International Asso-ciation of Science and Hydrology Publication 63, p. 432-441.

Rothschild, E.R., Huff, D.D., Haase, C.S., Clapp, R.B., Spalding, B.P., Farmer, C.D., and Farrow, N.D., 1984, Geohydrologic characterization of proposed solid waste storage area (SWSA) 7: Oak Ridge National Laboratory (Report) ORNL/

TM-9314, Environmental Sciences Division Publication no.

2380,262 p.

Searcy, J.K., 1959, Flow-duration curves in Manual of hydrology:

Part 2, Low-flow techniques: U.S. Geological Survey Water-Supply Paper 1542-A, 33 p.

Sledz, J.J. and Huff, D.D, 1981, Computer model for determining fracture porosity and permeability in the Conasauga Group, Oak Ridge National Laboratory, Tennessee: U.S. Depart-ment of Energy, Oak Ridge National Laboratory (Report)

ORNL/TM-7695, Environmental Sciences Division Publica-tion no. 1677, 138 p.

Smith, E.D., and Vaughan, N.D., 1985, Aquifer test analysis in nonradial flow regimes: A case study: Groundwater, v. 23, no. 2, March-April, 1985, p. 167-175.

Solomon, D.K., Haese, R.C., Dinsmore, T.V., and Kelmers, A.D.,

1988, Sampling and analysis of SWSA 6 trench leachates and groundwaters: U.S. Department of Energy, Oak Ridge National Laboratory (Report) ORNL/TM-10813, Environ-mental Sciences Division Publication no. 3117, 81 p.

Skougstad, M.W., Fishman, M.J., Friedman, L.C., Erdmann, D.E.,

and Duncan, S.S. (eds.), 1979, Methods for determination of inorganic substances in water and fluvial sediments: U.S.

Geological Survey Techniques of Water-Resources Investigations, Book 5, Chapter Al, p. 432-439.

Steuber, A.M., Edgar, D.E., McFadden, A.F., and Scott, T.G.,

1978, Preliminary investigation of ^Sr in White Oak Creek between monitoring stations 2 and 3, Oak Ridge National Laboratory:

[U.S.] Department of Energy, Oak Ridge National Laboratory (Report) ORNL/TM-6510, 80 p.

Steuber, A.M., Huff, D.D., Farrow, N.D., Jones, J.R., and Munro, I.L., 1981, An evaluation of some ^Sr sources in the White Oak Creek drainage basin:

[U.S] Department of Energy, Oak Ridge National Laboratory (Report) ORNL/TM-7290, 38 p.

Tucci, Patrick, 1986, Ground-water flow in Melton Valley, Oak Ridge Reservation, Roane County, Tennessee Prelim-inary model analysis:

U.S. Geological Survey Water-Resources Investigations Report 85-4221, 20 p.

____1987, Surface-geophysical investigations in Melton Valley, Oak Ridge Reservation, Tennessee: U.S. Geological Survey Water-Resources Investigations Report 87-4184, 29 p.

Tucci, Patrick, and Hanchar, D.W., 1989, Lithologic, geophysical, and well-construction data for observation wells in the Melton Valley area, Oak Ridge Reservation, Tennessee:

U.S. Geological Survey Water-Resources Investigations Report 88-4197, 22 p.

U.S. Department of Commerce, 1972, Daily, monthly, and annual climatological data for Oak Ridge, Tennessee, townsite and area stations January 1951 through December 1971: Oak Ridge, Tennessee, National Oceanic and Atmospheric Administration, Oak Ridge, Tennessee, 557 p.

____Climatologic

data, Annual

summary, Tennessee:

Asheville, North

Carolina, National Oceanic and Atmospheric Administration, annually 1972 through 1987,

v. 77 through 92, no. 13 (each volume).

____Climatologic data, Tennessee: Asheville, North Carolina, National Oceanic and Atmospheric Administration, monthly January through September 1988, v.93, numbers 1 through 9.

Webster, D.A., 1976, A review of hydrologic and geologic conditions related to the radioactive solid-waste burial grounds at Oak Ridge National Laboratory, Tennessee: U.S.

Geological Survey Open-File Report 76-727, 85 p.

Webster, D.A., Beatty, J.S., Benjamin, P.M., and Tranum, M.W.,

1980, Water-level data for wells in burial ground 6, Oak Ridge National Laboratory, Tennessee, 1975-1979: U.S.

Geological Survey Open-File Report 81-57, 100 p.

____1981, Water-level data for wells in and near burial ground 4, Oak Ridge National Laboratory, Tennessee, 1975-1979:

U.S. Geological Survey Open-File Report 81-339,52 p.

____1982, Water-level data for wells in burial ground 5, Oak Ridge National Laboratory, Tennessee, 1975-1979: U.S.

Geological Survey Open-File Report 82-372, 135 p.

Webster, D.A., and Bradley, M.W., 1988, Hydrology of the Melton Valley radioactive-waste burial grounds at Oak Ridge National Laboratory, Tennessee: U.S. Geological Survey Open-File Report 87-686, 115 p.

Zehner, H.H., 1989, Construction data and retrieval procedures for selected wells drilled from 1985 through 1987 at Oak Ridge National Laboratory, Tennessee: U.S. Geological Survey Open-File Report 89-61,96 p.

76 Hydrology of Melton Valley at Oak Ridge National Laboratory, Tennessee

'USGPO: 1992-750-232/80005