ML20245D047
| ML20245D047 | |
| Person / Time | |
|---|---|
| Site: | West Valley Demonstration Project |
| Issue date: | 12/31/1987 |
| From: | Yager R INTERIOR, DEPT. OF, GEOLOGICAL SURVEY |
| To: | NRC |
| Shared Package | |
| ML20245D040 | List: |
| References | |
| 85-4308, NUDOCS 8711040327 | |
| Download: ML20245D047 (67) | |
Text
-
Simulation of Ground-Water Flow Near the Nuclear-Fuel Reprocessing Facility at the Western New York Nuclear Service Center, Cattaraugus County, New York 1
- cattaraugus county T
.f U.S. GEOLOGIC AL SURVEY D
Water-Resources investigations Report 85-4308 8711040327 971001 PDR ADOCK 0500o201 P
PDR Prepared in cooperation with the f
U.S.' NUCLE AR' REGULATORY COMMISSION
SIMULATION OF GROUND-WATER FLOW NEAR THE NUCLE AR-FUEL REPROCESSING FACILITY AT THE WESTERN NEW YORK NUCLEAR SERVICE CENTER, CATTARAUGUS COUNTY, NEW YORK by Richard M. Yager U.S. GEOLOGICAL SURVEY Wa ter-Resources Investigations Re port 85-4308
- pf}Nk 4;-@f%
gc4 e, _
f',_,,
q' *
'/
Prepared in coope ration with the
'naD U. S. NUCLEAR REGULATORY CCNMISSION It haca, New Yo rk 1987
UNITED STATES DEPAR'DiENT OF TitB INTERIOR DONALD PAUL 110 DEL, Secretary -
GEOLOGICAL SURVEY Daller, L.
Peck, Directo r I
I f
s Fo r additional informa tion Copies of this report can write to:
be purchased f rom:
Subdistrict Chief U. S. Ge ological Su rvey U. S. Ge ologic al Sur vey Books and Open-File Reports 521 W. Se neca St reet Fede ral Ce nter, Bldg. 41 It haca, New York 14850 Box 2 5425 Telephone: (607) 272-8722 Denver, Colorado A0225 Te l ep hon e: (303) 236-7476 4
11
CONTENTS Page 1
Abstract 2
Introduc tion.
5 Purpose and scope.
5 Ackn owl ed gme nt s.
5 Site-description and history.
5 Reprocessing plant facilities.
8 Migration of radioisotopes.
12 Hydrogeologic se tting.
12 Drainage.
12 Climate.
13 Geology.
14 Wa ter-bearing units.
15 Hydrology of the surficial sand and gravel.
Flow direction under undisturbed conditions.
17 Influence of plant f aellities on ground-water flow.
17 High-level liquid-was te-tank complex.
17 17 Drainage struct ures 17 Wastewater lagoons.
Other s tructures.
19 Water-transmitting properties of sand and gravel deposit 19 21 Recharge.
22 Discharge.
Seasonal fluctuation of ground-water levels and discharge.
23 24 Seasonal patterns Calculation of monthly recharge fo r transient-state model.
25 26 Simula tion of ground-wa ter flow.
Flow model.
27 27 De sig n.
29 Input data.
29 Hydraulic conductivity.
29 Re charge and evapo transpiration.
31 Steady-state simulations.
Calibrat ion.
31 31 Se nsitivity.
Results.
34 Transient-state simulations.
39 Calibration.
40 40 Sensitivity.
Re sul t s.
40 46 Model applica tion.
Ground wa ter movement.
46 Analysis of past t ritium migration.
47 Summary and conclusions.
51 Ref erences ci ted.
52 Appendix--Es timation of hyd raulic conduct ivi ty 54 PLATE (in pocke t)
Plate 1.
Map showing locations of wells and test borings on north plateau.
iii
ILLUSTRATIONS Page Figures 1-3.
Maps showing location of:
1.
Western New York Nuclear Se rvices Center in Ca ttaraugus Creek drainage basin.
3 2.
Nuclear-fuel-reprocessing plant and related waste f acilities 4
3.
Reprocessing plant facilities and repositories of radioactive liquid waste and wastewater.
6 4.
Flow diagram of low-level radioactive wastewater-7 l
t reatme nt sys tem.
1 5.
Graph showing t rit ium concent rations in ground water that discharged to the wetland and f rench drain, 1974-81.
9 l
6.
Map showing tritium concentrations of ground-water samples collected from the north plateau in:
A.
1974.
B.
1978.
10 7.
Histogram showing monthly precipitation and potential evapotranepiration on the north pla teau, October 1982 t hrough September 1983.
13 8.
Generalized hydrogeologic sections:
A-A ' f r om main reprocessing plant area to Franks Creek showing major lithologic units, and water levels measured in well 82-4E, July 1983:
B-B',
f rom shale uplands to Franks Creek showing water-table altitude on the north plateau.
14 9-11.
Maps of north plateau showing:
9.
Saturated thickness of sur ficial gravel, May 10, 1983.
16 10.
Water-level altitude and dir ection of ground-water flow, May 10, 1983.
18 11.
Hydraulic conductivity values used in model simulations of sand and gravel.
20 12.
Hydrography of wells 80-4 and 80-8 on north plateau, wa te r ye a rs 1982-83.
24 13.
Graph showing monthly 1982-83 recharge rates calculated by eq uat ions 3 and 4 and by soil-moisture model used to t rans ient-s ta te s imul at ions.
26 14.
Map showing finite-d if fe rence grid and boun da ry condi t ions used in model simulations 28 15-16.
Diagrams showing :
15.
Tbe f our t ypes of drains used in the ground-water flow model.
30 16.
Evapotranspiration rate as a function of water level in the surf icial grave 1.
30
.7-20.
Map s of north pla teau showing:
17.
Model ec11s in which recharge and evapotrans-piration were not simulated.
33 18.
Simulated steady-state and measured water levels.
35 19.
Predicted steady-state rates of recharge and discharge from constant-flux and drain boundaries 37 20.
Predicted steady-cate distribution of evapotranspiratiot:
38 s
iv
ILLUSTRATIONS (continued)
Page Figures 21.
Vertical section through north plateau showing simulated I
ground-water levels near a seepage face based on variable grid and uniform grid spacing.
39 22-23.
Graphs showing:
22.
Observed ground-water levels in four wells in relation to changes in seasonal recharge conputed by soil-moisture-deficit method and by soil-moisture model of Steenhuis and others (1983).
41 23.
Observed departur es' of ground-water levels f rom initial level at wells 80-3 and 80-4 in relation to simulated values computed f rom two magnitudes of 42 s pecific yield.
24-25.
Graphs showing:
24.
Measured ground water discharges at sites NP-1 and NP-3 in relation to discharges simulated from constant and variable drain-conduc tance value s.
43 25.
Measured and simulated water levels in eight wells.
44 26-28.
Maps of north plateau showing:
26.
Ground-water flow paths through sand and gravel as predicted by steady-state model.
48 27.
Flow paths and traveltimes of water f rom two potential sources of tritium contamination as predicted by steady-state model.
49 28.
1972 ground-water flow paths simulated by steady-state model and tritium concentrations in ground-water samples collected in 1974.
50 A-1.
Map showing till surface altitude near main plant and location of buried strewn channel.
56 A-2.
Box plot analysis illustrating dif ferences in hyd raulic conductivity at observation wells grouped according to their location relative to the buried channel.
58 TABLES Table 1.
Tritium concentration of liquid waste and wastewater in reprocessing plant facilities.
8 2.
Ground-water budget for sand and gravel deposit on north plateau, October 1982 t hrough Se ptember 19 83.
21 3.
Measured ground-water discharges f rom north plateau, March, July, and Oc tober 1983.
24 4
Optimum values obtained through steady-state simulation and result s of sensitivity analysis.
32 5.
Recharge and discharge values calculated by steady-state mode l.
36 A - 1.
Hydraulic conduc tivity values for sand and gravel on the north plateau as de termined f rom slug-test data by Cooper me thod.
54 A-2.
Relative measures of hydraulic conductivity developed f rom pinping and slug-t est data f rom 15-cm-diame ter wells.
57 v
. CONVERSION FACTORS AND ABBREVIATIONS Conversion f actors for the terms used in this report are given for readers who prefer to use inch pound units rather than In ternational System (SI) units.
Multiply SI unit By To obtain inch nound units Length l
millime ter (mm) 0.03937 inch (in)
. me ter (m) 3.281 fo ot (ft) kilometer (km) 0.6214 mile (mi) 3 cent tmeter (cm) 0.3937 inch (in)
Area 2
square meter (m )
10.76 squa re f oot (ft2) 2 1.196 square ya rd (yd )
0.0002471 acre hectare (ha) 2.471 acre 2
square kilometer (km )
0.3861 square mile (ti2)
Volume 3
cubic meter (m )
35.31 cubic f oot ( f t3) 3 1.307 cubic yard (yd )
0.0008107 acre-foot (acre-ft) 264.2 gallon (gal) liter (L) 1.0577 auart (qt)
Flow 3
liter per second (L/s) 0.0353L cubic foot per second (ft /s) 3 3
cubic meter per second (m /s) 35.31 cubic foot pe r second ( f t /s) 3 cubic meter per day (m /d) 0.0004086 cubic foot per second (ft3/s)
Hydraulle Units me ter pe r day (m/d) 3.281 feet per day (ft/d)
Temperature degree Ce lsiu s ('C)
'F => (9 / 5 *C) + 3 2 degree Fahrenheit ( *F)
Other Abbreviation picocur ie per milliliter
( pC L / mL)
./
National Geodetic Vertical Datun of 1929 (NGVD of 1929): A geodet ic datum derived f rom a general adjustment of the first-order level nets of both the United States a nd Ca nad s, f o rme r ly ca l led "Me a n Se a Le ve l. "
vi
Simulation of Ground-Water Flow Near the Nuclear-Fuel Reprocessing Facility at the Western New York Nuclear Service Center, Cattaraugus County, New York By Richard M. Yager ABSTRACT A two-dimensional finite-dif ference model was developed to simulate g round-water flow in a surficial sand and gravel de posit underlying the nuclear-fuel reprocessing facility at Western New York Nuclear Service Center near West Valley, N.Y.
The sand and gravel deposit over11es a till plateau that abuts an upland area of siltstone and shale on its west side, and is bounded on the other three sides by deeply incised stream channels that drain to Buttermilk Creek, a tributary to Cattaraugus Creek. Radioactive materials are stored within the reprocessing plant and are also buried wi thin a till deposit at the facility. Tritiated water is stored in a lagoon system near the plant and released under permit to Franks Creek, a tributary to Buttermilk Creek.
Ground-water levels predicted by steady-state simulations closely matched those measured in 23 observation wells, with an average error of 0.5 meter.
Simulated ground-water discharges to two stream channels and a subsurf ace d rain were within 5 percent of recorded values.
Steady state simulations used an average annual recharge rate of 46 centimeters per year; predicted evapo-trans piration los s f rom the ground was 20 centimeters per year. The lateral range in hyd raulic conduc tivity obtained through model calibration was 0.6 to 10 meters per day. This range compares f avorably with that calculated f rom slug tests at observation wells, although the mean value of 4.0 used in the model is considerably higher than the geometric mean value of 0.6 meter per day obtained f r om slug-test da ta.
tbdel simula tions indicated that 33 percent of the gr ound wa ter discharged from the sand and gravel unit (2.6 liters per second) is lost by e va pot rans pir a tion, 39 percent (3.0 liters per second) flows to seepage f aces at the periphery of the plateau, 20 pe rcent (1.6 liters per second) discharges t o s tream channels that d rain a large wetland area near the center of the pla-t eau, and the remaining 8 percent (0.6 liter per second) discharges to a sub-surf ace f rench drain and to a wa stewater-t reatment system.
Ground-water levels computed by a transient-state simulation of an annual climatic cycle, including seasonal variation in recharge and evapotrans pira-tion, closely ma tched water levels measured in e ight obs erva tion wells.
The dif ference between conputed and observed ground-water levels could largely be explained by uncertainty in the timing and volume of recharge.
The hyd raulic conductance of seepage f aces was varied seasonally to match measured base fl ows.
The model was used to delinea te ground-wa ter flow paths and to est ima te travel times fran potential sources of radioisotope contamination to discharge a reas on the plateau.
The model predicted t hat the subsur f ace drain and the 1
stream channel that drains the wetland would intercept most of the recharge originating near the reprocessing plant.
A slug of water introduced at the main plant building would take approximately 500 days to reach either discharge point.
The model also wa s used to simulate groaad-water flows of 19 72, when tritium was detected in ground wa ter discharging into the we Lland.
Fl ow pa ths predicted by the model do not support the assumptiton that le akage f rom waste-water lagoons 200 meters south of the we tland was the source of the tritium.
N I; TRODUCTION The We stern New Yo rk Nuc lear Se rvice Ce nter is on a 1,350-ha tract of land acq uired in 1961 by the New York State Of fice of Atomic Development near the village of West Valley in northern Cattaraugus County, about 48 km south of Bu ffalo (fig. 1).
In 19 63, t he U. S. At o mi c En ergy Co mmi s sion i s s ued a pe r-mit to a private operator authorizing development of about 100 ha of the tract for construc tion of a nucle ar-f uel reprocessing plant and its facili ties. The f acili ties included a receiving and s torage f acility for irradiated f uel rods, an underground tank complex f or storage of liq uid high-level radioact ive wastes generated by reprocessing, and a low-level radioact ive-wastewater-t reatme nt pla nt.
The site also included two areas for shallow burial of solid radioactive wastes-a 4-ha area licensed by the St ate of New York for burial of conmercial low-level radioactive wastes and a 2.9-ha area previously licensed by the U.S. Nuclear Regula tory Commission for burial of radioact ive materials f rom the reprocessing plant, called the facility's disposal area.
Locations of the f acilities are shown in figure 2.
In 1982, t he reprocessing plant wa s turned over to the U.S.
De pa r tme nt o f Energy (DOE), which contracted the operation of the f acili ty ta a private operator, We st Valley Nucle ar Se rvices, Inc.
The aim of the DOE is to decommission the reprocessing facilities and to solidify the high-level liquid radioactive waste stored at the site for future disposal at a high-level-waste repository.
The U.S. Geological Survey conducted studies f ran 1975 through 1980 to evaluate the potential for radioisotope migration fr om the St ate-licensed bur ial grouw3. The studies we re part of a national program to determine the principal f act ors that control the subsur face movement of radioisotopes.
Complementary sttdies by the New York St ate Geological Survey were done to evaluate other processes of radioisotope migration fr om the burial ground (Prud ic, 1986).
In 1980,4 the U.S. Geological Survey, under contract with s,he U. S. Nuclear Re gulato ry Oc.mni rsion, began a study to inve stigate the hvd rogeolor;y and ground-water Y ae near the repr ocessing plant and its f acilities. Th e repro-cessing pTant'is on the 42-ha north plateau, which is separated f rom t he birial ground and facili ty's dic90 sal area by a deeply incised s tream channel, Er dma n B r ook. A companion study, also begun in 1980, examined the hydro-geology and ground-water flow in the f acility's disposal arna (Bergeron and Bugliosi,.in pr es s).
2
i Tc E l
]s(
' j
,; 2 l
Li -
(
{(
q-)
i i
i
).
.4g 4,
g h
.k "
79 W OO" 78'30'00" 4
3
,- p,.
. i,r.
/.
7-
/
$ l
,t,
- 'j:::
BUFFALCi J-I f
i
[NEW YORK j
'M
/
{
f6 Cattaraugus Coi
- ares, go g
LAKf y
=
< ~.
./ q.'
1 s
y
.,t
/
}
h Y
/,'
T t
J
't r
j N'
r k
({,-;g
(
O
[~'
'/
,/ f lo
[
f
('
l
.tA
/
c
},\\
..\\
/
/g
. /
E R1FsCOUNTY Z
/
)4 l(,
,A
?
't 2
- g
(
3 f '{ /
\\!
ARCApE g
ft
\\
/\\
SFRJH3 VILLI.mg
. ~ %[ - -
?
e
- l
'f gy.4 C
b)
, gs 14i t
j k
f-l I
\\
p l
(10WAl 1,'M N
)l' j l f('
Catta *09,
1 0
t, jutermik f
r 4
>s o',
Creek y
,\\
j(,
\\
h/
e
4 WESTERN NEW YORK x-f1 l
\\
J NUCLEAN SF.RVICE CENT R
\\* WEST VALLEY N
\\,
')
t N,
I
>k y'l r.z D
/
O
],
U
,' Connoisarauley e
Creek FRANKLINVILLE S
\\. & *'.. / )
,i E
?
/
a 1'
U a
,/
, i[ T i.,
a
/',
-)
Z
,1 i
4 i ((, ' ' C A T T ih R A U G U S 4
I COUNTY O
Ul m
\\
l1
}
/
./
'd y'
i AREA DEPICllD IN FIGURE 2 x
f N..N DRAINAGE-ARN' BOUNDARY OLEAN I
/
V m.
'l
.e, f
b.m
\\
s f
42*
NEW YODt h
PE NNSYLV sNI A 0 10 Kli.OMET EMS l
s A
+
I 0
'O I
- A w
Base from U S Geologicat $dv( [ 974 15 4
3 Siate of New York. 1 500.000
.y ' '(l 1
l
\\
s Figu re 1. -L is of al iy area ui thl1 cattavatqu e Creek, drainage basin.
.m.
I s'
s
)-
u ol j
u y
s
.t Y l --
78'3915" 78"39'0(P 1
I ye'?'.
\\
-' 42' Od n
\\
NORTH PLATEAU -
2P 15" Fond T
STUDY AREA n
,.. [ ;r
.K
.(M g
/
k} Wetland t
l
- HIGH LEVt.
i HADlQ ACTl; 'E
'. 5.
t j
l WASTE + STOR NGE-3 4
/
T ANK COMPT (X
' h. /.
'iNASTE-WATER..AGOON
. LOW-LEVEL RADIOACTIVE I.'
n-i
/
.{
j QD $'
/
q>
V*.
FUEL PEPROCr SING
'l N
PL/.I
\\
4.f
\\
, g/
\\.
f s-R
'h IC
,s y
SN*s La f-ACidTY'S x:
s
\\
]
\\
(,,
OfSPOSAL
\\
i
/
\\s AREA 4
'N<.
-,f
,k
\\+' \\
l}h i
\\
\\
l
~
O p
i
/
$0x,/,'
" t'\\ \\O 1
/ \\ STATE UCENSED V A5H'-\\[
l 1.
\\ \\
l'
\\
DISPOSAL AREA
+
/
w[~l N/
\\,
/
\\
0 5,0 102 160 METERS
/
6 250 shorEFr
\\
l i
.a
\\.
e lo sv.
7 [ (g )
Figure 2.--Location of nuclear-fuel-reprocessing plant and related uaste facilities.
(General location la shown in fig.1.)
l 4
l ai P'urpose and Scope j
1This report describes (1) hydrogeologic conditions and ground-wa.ter flow neari he reprocessing plant and its f acilities on the north plateau, (2) the t
reprocessing plant facilities and the migration of radioisotopes in the area, (3) ground-water. flow patterns on the north plateau within the surficial sand and gravel deposit, (4) the development and calibration of a two-dimensional finite-dif ference model.used to. simulate steady-and transient-state flow within the surficial material, and (5) the application of the model to analyze past tritium migration and to predict flow paths and velocities of g round wa ter f rom' two ' po tential sources of tritium detected in 1972.
Acknowledgments The author thanks the staf f of the West Valley Nuclear Service Center for cooperation and assistance in recording water levels and supplying data on' the f
plant operation and construction. Thanks are also extended to T. S. Steenhuf f 1
. professor of Agricultural Engineering at Cornell University, for assistance in l
dev' eloping a model to simulate soil-moisture content that was used in calibrating values of ground-wa ter. recharge for the flow model.
.l SITE DESCRIPTION AND HISTORY
{
The Western New York Nuclear Service Center was operated as a nuclear-f uel. reprocessing f acility during 1966-72, during which time it received spent f uel-rod assemblies f rom nuclear reactors and processed the fuel elements to recover uranita and plutonium.
Until 1975 the facility continued to receive f uel-rod assemblies.
In 1985 the site contained spent fuel-rod assemblies that had not been reprocessed, high-level: radioactive liquid wastes generated by the recovery process, and a variety of low-level radioact ive solid wastes generated by the reprocessing facility and received from of fsite commercial installa tions.
Reprocessing-Plant Facilities The reprocessing plant consists of many f acilities used in the recovery process (fig. 3).
Three of these--the fuel-receiving and stcrage area, the I
high-level radioactive liquid-waste-tank complex, and the low-level radioac-tive wastewater-treatment system (fig. 3)- contain radioactive liquid waste and wastewater and are of concern as a potential source of radioisotope migra-l tion-to ground water.
j The fuel-receiving and storage area, which occupies the east part of the main plant, was the point of entry for f uel-rod assemblies received at the 1
site. The area includes a fuel-storage pool in which the fuel-rod assemblies we re submerged in demineralized wa ter that is ke pt between 27' and 32*C.
The f ael-storage pool is constructed of concrete lined with carbaloy paint and contains 3.0 x 106 liters of water.
L The high-level liquid-waste-tank complex (fig. 3) serves as a temporary repository for waste solutions from the recovery process used in the reproc-essing facility. The tank complex consists of two underground concrete vaults, l
5
78;39 15" 78 :39 'XT
,. - ~
i
{
/
A NP 2 STREAMFLOW-GAGING STATION
/
i
/'
j* a,
\\
j
's%,
'S fr" 4,
\\
f 4 25
- g 2(~
)
/ \\
I6
+
\\
J
%gt f'
\\
knuf E\\ TENT OF SAND AND g.w"*s.,%. 3
/\\
GRAVEL DEPOS:rs
\\
NP-24 i
NP )L #L*
~~p_. -*
[
r
,.+s
\\.
t ff,/ */
.k
+
xm
\\
?
,/
/ E' '
?k
\\
\\
WEllAND
' NP-3
\\c l
f'
/,/ l
,{
' Q --
l 3
/
I
/^
/
/
t h
/ M s'
I I
/
l i
7 . - - HARDST AN D j
n y
j i
/
FRENCH i
\\
l DRAIN y
\\
/
./
\\
[
p p%
. /
(
v ANJ/
1
\\
Y Ny-HIGH LEVEL
^
[
LMy4NS V [/,/
)
l I
N RADIOACTIVE s
I fl f
(
WASTE $TOHAGl -
j,,A.
l
,\\
-\\\\
\\
TANK COMPLEX f </ j /
k, Y/
k
/+
g\\
SEPTIC- (V x
FUEL-RECEIVING--
AND TANK
?
gdg /
\\g I. * 'pE AREA (NN STORA 1 LEACHFIELD' "y-wd, _
1
\\
's V
1 m,,
\\,.
K REPROCF5S:% 4 i '.N ptAg7 f
- r.
At w tA T
\\,
l PLANT I
m\\
! OUTFALL
'7 4\\
\\
f CHANNEL jj
\\
k
/
/
c w,
8
!J'
+
\\
+\\
r s
N N
/
\\
.,g\\
y
<.,,:, Q..
v,.
s' oc~
\\\\
"g 9.e j
\\\\ w,,%,....,,,,"'" * ** m.,,,#j#,,
y s,
gt.s M
i r A,1ITV 5 N
,\\
g,f g',
ut r-P:;;y. 4 L A p g A R
x s
\\\\t o
w tuo tso vm ns v, '. -
p ____u,.a
- s. s s s
NN, o
- ,o or un y,
i
'\\
\\
.. _ -. ~..,..AL,11_..
o.n,+.. - u
- s., +,
,,v.,
.i.
s l
A., %.i H,, :.#,
i u.'7 $,.1 y O t
I i
Figure 3.--Location of reprocessing-plant facilities anl repositories of radioactive liqubi txiste and vasteuater on north plateau.
6
]
l i
l each of which encases two cylindrl;:al stainless-steel tanks. The entire
. complex is ' backfilled with 8 m of silty clay till. An external hydraulic pressure is maintained by a water-injection system in a 1.2-m layer of pea gravel beneath.the tank complex to prevent leakage from the concrete vaults.
Nearly 98 percent Lof the 2.2 x 106 liters of liquid waste is stored in one t ank. Vapar ventilated from the concrete vault is regularly monitored for radioactive leakage from the storage tanks.
The low-level wastewater-treatment system, 100 m east of the main plant, was designed to' remove radioisotopes from wastewater generated by reprocessing
. operations. and to release the. treated water to surface water at a controlled rate.. Since the shutdown of the reprocessing plant in 1972, the system has treated 20 to 60 x.106 m/yr of wastewater from (1) precipitation that had infiltrated into the State-licensed burial ground, (2) condensate from the cooling system of the fuel-storage, pool, and (3) rinse water f rom decontamina-i
. t ion ac tivities. -
The f acility originally included five lagoons for storage of processed
. and unproces sed was tewa ter.
A schematic diagram of the low-level radioactive wastewater system showing the relative sizes of the lagoons and direction of flow is given in figure 4.
The wastewater entered the system through lagoon 1, passed to lagoon 2 f or temporary storage, and was periodically withd rawn f rom lagoon 2 and back through lagoon 1 f or treatment.
The treated water was pumped to lagoons 4 and 5' and then drained to lagoon 3, f rom which it was released to Erdman Brook (fig.
2).
In 1984 lagoon I was backfilled, and wastewater was discharged directly into lagoon 2.
LOW LEVEtrRADIOACUVE g
g WASTEWATER-TREATMENT PLANT H SLUDGE TO BURIAL 4- - - - - - - - -
--+
LAGOON 4 LAGOON 5 p
h h
it l
y U
9 INTERCEPTORS
' PROCESSING-PLANT WASTE 4- -
-h
-D LAGOON 1 LAGOON 2 LAGOON 3
-9
-O f
NElfiRAUZER PIT l
l 1
FROM DISCHARGE WEIR WASTE-8URIAL AREA TO ERrME4 BROOK l
Figure 4.--Nou diagmm of Lou-level radioactive vasteuater-treatment l
eyatem.
(Modified fvan U.S. Department of Energy,1979.)
1 7
1
--l
A subsurf ace (french) drain was installed on the north and west sides of lagoons 2 and 3 to intercept and reduce ground-water seepage into them.
The drain consists of a 15 cm-diameter perforated pipe buried about 3 m below land surface. The drain discharges to Erdman Brook east of lagoon 3.
(See fig. 3.)
~
Migration of Radioisotopes Migration of radioisotopes f rom the reprocessing f acility has been docu-mented by radiation surveys of the land surface and soll samples (L. Roberts, West Valley Nuclear Se rvices Co., written canmun.,19 84). Gamma-radiat ion surveys on the 42-ha north plateau detected surface radiation 10 to 100 times the background level recorded of fsite. The shape of the radiation field indi-cated that the probable source of the surface contamination was particulate fallout from the ventilation stack of the main plant building. A soil-L sampling survey showed that surf ace deposits of particulate radioactive materials were. retained in the upper 25 cm of soil (L. Roberts, written conmun., 1984). This indicates that most radioactive-decay products generated by the reprocessing f acility adsorb readily to clay surfaces in subsurf ace -
materials and do not tend to taigrate with ground water.
l l
Tritium is the most mobile radioisotope found in ground water and is the l
only one detected in ground-water samples f rom the north plateau (L. Ro be r t s,
L written commun., 1984). Concentrations of tritium at selected locations are listed in table 1.
Above-background concentrations were detected in 1972 in ground water that discharged to the wetland and the french drain.
In response q
to this discovery, several shallow wells were installed near the main plant to determine the tritium source.
Analysis of water samples f rom these wells indica ted that the low-level wastewater-treatment system was the probable source of tritium in ground water. Fallout from the ventilation stack was less likely to contribute tritium to ground water because the tritium emerged f rom the stack in the f orm of water vapor and would thus be carried away f rom the plateau by air currents.
Table 1.--Tritium concentration of liquid vaste and vaste-uater in reprocessing-plant facilities,1979.
Location Concentration (pCi/mL)
Fuel-s torage pool 600 liigh-level liquid-was te tank 22,000 La go on 1 1,500-100,000 Unaf fected areas 2
(background levels)
The concentrations of tritium in ground water that discharged to the wetland and the f rench drain during 1973-81 are plotted in figure 5; those in wells near the lagoons in 1974 and 1978 are shown in figures 6A and 6B, r es pe ct ive ly.
The abrupt rise in tritium concent ration at the french drain in 1976 is attributed to overflow from lagoon 3.
The concentrations declined as the lagoon was emptied but remained above the background concentration in l
sampics collected of f site.
The source of this continuing radioactivity was assumed to be lagoon 1, which was unif ned and hydraulically connected to g r oun d wa te r.
Lagoon I was backfilled in 1984 and is no longer part of the t reatme nt system.
8
I i
The concentrations of tritium in ground water at wells near the lagoons in June 1974 are sinwn in figure 6A.
Tritita concentrations at most wells declined af ter lagoons 4 and 5 were sealed with rubber. liners in October 1974 and reprocessing activities ceased; the concentrations at the same wells in 1978 are shown in figure 6B. Tritium concentrations were generally highest near lagoon 1, but relatively high trititu concentrations were also found in I
wells north of the reprocessing plant.
The tritium in ground water north of the plant could have originated f rom contamination that has occurred beneath the main plant building (New York State Department of Environmental Co nse rva tion, 1975, p. 12). Other possible sources could be tritium released f rom the ventila tion stack during cold weather and carried to the land surf ace by snowfall, or leaking containers of tritiated water that may have been stored on the hardstand, a small paved area north of the reprocessing plant (fig. 3).
l 10.000 i
x 1,000 j
e a.
O
"%g x
x FRENCH ORAIN x
x G
x-,
8 h
WETLAND T
f 10 x
g xN x
Normal Background Concentration Recorded Offsate i
i i
i i
j 1974 1975 1976 1977 1978 1979 1980 1981 Figure 5.--Tvitinn concentrations in ground vater that disehargod to the vetland and french drain, 1974-81.
9
78'39'15" 78"39 LW j
,/ ~-
A NP-2 STREAMFLOW-GAGING STATION Gj, '
7 40 OBSERVATION WELL-Number is j
O triflum Concentration, in J
picocuries per liter; letter is local well symbol, p,
n
- 200 POINT OF GROUND-WATER Q?fr EXTENT OF SAND AND DISCHARGE--Number is
\\ GRAVEL DEPOSITS l
tritium concentration, in X
picocuries per liter,
)
l x4,g
- g,%,
42o s
N 27;
,p y
5 lg,';5[
m.,
~~
f NP-[g '[
- g. ~.4 - *'
j
(
g/N
)
, l's
'\\
i
'S WETLANO /
\\
n i
+ g g
100 sgt% "f '
i*
' ['
HARDSTAND
)
HIGH LE7EL FRENCH x' A).]
RADIOACTIVE DRAIN WAS T6-S TORAGE-300
^
- 's '
y
/ \\
TANK COMPLEX 160 350 400 0 5
\\
l O OO K
' /
\\
J-8 J-9 F
/
g s
p7 3 V
%' 'N s
70 O M-2
$G NS 200 x
)
\\
P O
,//.,
,/'
/
j u
i LOW-LEVEL g (Y,*,
40 L-5 j
I WASTE WATER 5
e' f
i TREATMENT PLANT g/e v
+
6 i
N
' FUEt
{
! ST ATE UGYiED l
REPROCESSING i /
W AS il
\\
PLANT
/
! DISPOSAL APE A 1
f
/
k\\
l \\[
/
\\
]
/ J
,7 p
+
r
(
\\
d 00
l
,o s
\\
\\
A
\\
/
ph
'L mmm Rosa
/
\\
j/,s*',
(
- /
Q-v-
- wn,
'Q
..,-; y
\\
\\ --
f
\\
o
!.o 1p 150 MrrEns
\\
/
\\
j l
0 S,o Sbreer
\\/
iGsh, n u s ca.,,, a su~.,
- ~~'
~ ' ' - ~ ~
~
- ~
AsMyme ow 19 M t 2, cm n
Figure 6A.--Tritiin concentrations in ground-oater samples collected from r: orth plateau in 1974, before lagoons were sealed.
(Location is shoon in fig. 2.)
10 L
et s
- i J
.78"39 00".
78*3915
ANP-2 STREAMFLOW-GAGING STATION '
g e
,40 :. OBSERVATION WELL.-Number is tritfurn -
L.5 concentration, in picocuries per liter:
two numbers indicate range of
[
~
q concentration. Letter is local EXTENT OF SAND AND GRAVEL DEPOSITS
- m 10-POINT OF GROUND-WATER N
k
DISCHARGE -Number is tritium concentration in picoeuries -
g*
per liter.
4% /
15" (NP-21 n
"d
~\\
j NP-1 jk P\\
j
/./
N+ '\\
~(
e3 p
- # WETLAND 4
10 Qs g
i
,g 1-LOW-LEVEL WASTEWATER 1
1 TREATMENT PLANT k
i 1
HAROSTANO
/
s.
HIGH-LEVEL RADIOACTIVE FRENCH
~
DRAIN
- WASTE-STORAGE.
7'h iip 130 p. g fj j
, e'
\\
TANK COMPLEX j ;.
915 20 V /
/
\\
f-23 0 e eJ.9]0 O
3 70y0 6
ONO,/
3 0 0-90 t x
- 2. %L-3 N G90NS/
^N h00,
38 70 -
)
7"'
j 25pg 5/ 40 /
n
/7
/
/,
/
409 C 290415 #
j 110-17 1500 /
g E5g i
f
.p M260 AG \\
g6151050 t i
X 38 9'16 G s
g j
M Y
~'*%,,,
' FUEL l
j-STATE LICENSED i
WASTF '
REPROCESSING b j 015POSAI. ARE A
\\
PLANT l
1
+^
j.
i O
[
4
\\
42*
/-
7
'\\
j 2T i
/
i
\\
pM k
00" 4
g iAyvon Road e
+
9
,/
D er 9
}(p*,fp, --
/
j
\\
/ F AC:UTY $
'""'*% %"'"*A.,#
/) \\
i, Y
(
DISPOSAL AR[ A
' "g
\\
/
\\
p 3
0 5,0 103 150 METE R$
\\
\\v/
\\
0 260 f.;M FEET
+
~J
. tin. nom o s ce a. su,ve,
~--
As%n1 Hrdlow 1979124 0rr)
Figure 88.--Trititrt concentrations of ground-vater samples collected from north platems in 1978, after lagoons 4 and S hai been seated for 4 years.. (Location is shart in fig. 2.)
11 1
1
-6
Ia
'b-0
+
HYDROGEOLOGIC SETTING 1
Dralnage The-study area lies along the west side of the. Buttermilk Creek valley (fig. 1). : Buttermilk Creek flows' northwestward along the. castiside of che site :to Cattaraugus Creek near Springville, about 3 km north of the site.
2 Cattaraugus Creek drains to Lake Erie.
- Franks Creek, a maj or tributary. to Buttermilk Creek, d rains' the entire E
site.
lt. has a drainage area of 6.35 km2 and borders the ' north plateau on the-
_ east (fig. 2). Two of its smaller tributaries, Quarry Cr eek and Erdman Br ook,-
' border the north plateau on the northwest and southeast, respectively.
M-The ' north ' plateau is drained. by-three small manamed' tributary streams.
K The:westernmost stream, with a gaging station designated North Plateau 1 (hereaf ter.ref erred to as NP1), d rains the ~ west side of the ^ plateau and is tributary to. Quarry Creek (fig. 3)..The. gaged drainage area is 10.4 ha.
The second stream is ' upstream' f rom gaging station NP3 (fig. 3) and is intermit-tent; it drains 9.8 ha and 19 tributary to Franks Creek. This channel
. receives flow f rom the center. of the plateau, including the we tland and most L of the' reprocessing f acilites.
The third stream drains a 1.8-ha area upstream.
L f rom the ' partial-record station designated NP2 (fig. 3); it also is intermit-tent and is tributary to Franks Creek. This channel was the outlet of the wetland, which now drains past station NP3'.as a result of topographic modifi-cations during site development. Most of the water discharged from the easternmost 7.5 ha of the plateau is f rom the main plant's steam-condensation a yatem combined with overflow f rom the plant's wa ter-supply system. Th e f rench-drain system that surrounds Lagoons 2 and 3 (fig. 6B) also discharges
- perennially to Erdman Brook.
Climate Mean annual' temperature at the site is 7.2*C.
The warmest month is July, with a mean temperature of 19.6*C; the coldest month is February, with a mean t emperatur e of -5.7'C.
Mean annual precipitation is about 100 cm; this amount
-i is distributed f airly evenly throughout the year.
ebnthly precipitation and estimated potential evapotranspiration (fig. 7) l were used to estimate total recharge to ground water on the north plateau f rom.
1 October 1982 through September 1983. Appr oxima tely 13 percent of the precipi-l tation flows over the surf ace of the north plateau as storm runof f (Kappel and Ha rding, 1987). Es timated potential evapotrans piration was calculated by a method based' on the Penman equation (Steenhuis and others, 1983). The maximum l
L and minimum ' daily temperature values used were those recorded at weather sta-tions in Gowanda, Franklinv111e, and Arcade (fig. 1); percent cloud cover and average wind speed were recorded at the Buf f alo Airport.
Annual precipitation t.
' f or the 12 month period was 92 cm, and the po tential evapotranspiration was estimated to be 90 cm.
Evapotranspiration f ran native grasses alone would be 68 cm, assuming a consumptive-use coef ficient of 0. 7 5 (Gray,1970,
- p. 354).
L 12 n
] POTENTIAL EVAPOTRANSPIRATION
~
16'O MONTHLY PRECIPITAT!ON a:
$ to.o 3
o l;.
p g
- s E
g 5.0 0
OCT NOV DEC JAN FEB MAR APR MAY JUN JUL AUG SEP 1982 1983 Figure 7.--Monthly precipitation and potential evapotranspiration on the north plateau, October 1982 through September 1983.
}
I Geology The Western New York Nuclear Service Center site is in the glaciated section of the Appalachian Plateau in western New York State, an area charac-The terized by deeply dissected glacial drif t overlying shale and sandstone.
north plateau consists of a sequence of glacial and postglacial deposits that successive periods of glaciation as described by LaFleur (1979).
ref le ct Geologic data were available f rom 11 test borings completed by the U.S.
Geological Survey and 87 test borings and pits completed by other investiga-Logs of these borings are presented in Bergeron (1985).
Locations of tors.
wells and test borings in the study area are shown on plate 1.
The plateau is covered by a layer of silty sand and gravel deposited as an alluvial f an at the edge of a postglacial lake that formed in the This Buttermilk Creek valley during the recession of the last ice margin.
deposit overlies a sequence of till deposits that probably correspond to three advances of the ice margin into the Buttermilk Creek valley. The till deposits are separated by lacustrine silt and sand, and by alluvial sand and gravel deposited during the recession of each ice margin. This sequence is depicted in section A-A' in figure 8A (lines of section are shown in pl. 1).
The surficial sand and gravel deposit and the underlying glacial deposits have been eroded by Buttermilk Creek and its tributaries to a maximum depth of The 40 m.
Today, the plateau area is bordered on three sides by gullies.
shale bedrock is at land surface at the southwes t edge of the plateau, where it is in contact with the surficial sand and gravel.
13
,]
U-t q
L3 t
y e' in l
a' f
, 5iw
.['
e MAIN BUILDING At 44o SHOWING FOUNDATION DEPTH 430
%t g
'2'.
t t,,,
i
~
3 l-
%*N'4
' 410 h
t
.h Sh81'-
\\
i N Silty Clayey Till -
- {-
g
'h.400 fg
?
%s,_
t WELL-SCREEN POSITION:
'e 390 b
380 WATER LEVEL IN CASING
\\
j s'a*nd nd grated Fine p
2 E 380 SCREENED THROUGH UNIT ~ l N
j----- 384 s
In n,eters above NGVD of 1929 l N
370 N,
. GEOLOGIC CONTACT -Dashed s'
% where inferred
\\
Silty Clayey Till g
g 360 l
\\
I 350 0
5,0 1QO 150 METERS
,: Sand and Sitt 379 (
[
N s Silty Clay Till 0-250 500 FEET N
Shale Vertical Aeration X10
'1 Figure 8A.--General hydrogeologic section A-A' shouing major Lithologic l
d units from the main reprocessing-plant area to Franks Creek and
. uater levela measured in well 82-42, July 1983. Location of vells arti line of 'section are shoun on plate 1.
(%dified from Bergeron l
ani othere,1987.)
[
Water-Bearing Units Unconfined ground water saturates the lowe r part of the surficial sand and gravel where it immediately overlies the till.
The general position of the water table and the direction of ground-wa ter flow through the sur ficial sand and gravel along section B-B' are shown in figure 88. - The saturated thickness of the sand and gravel on May 10, 1983 is depicted on the map shown in figure 9 (p. 16).
1 The sand and gravel deposit varies in composition but averages 55 percent gravel, 20 percent sand, and 2 5 percent silt.
It is mostly silty gravel with some sand and has a high hydraulic conductivity. Where the deposit is thick, it 'contains thin layers of gravelly silt of lower hyd raulic conduc tivity.
p The
(
. sand and gravel on the lower (central) part of the north plateau near well 80-6 (pl.1) is covered with 3 m of silty till that was applied as fill to cover a wetland area.
14 l
SOUTHWEST NORTH B
B' inflow from I
Upland Area Shale Bedrock kl
.... APPROXIMATE POSITION OF 445 - N Roe ngs di WATER TABLE
~
DIRECTION OF GROUND-440
$I WATER FLOW Spring Parking dl LO' 1
% BEDROCK CONTACT
~
435 SiltySa and Gravel 430 l
Evapctronspiraton
[425 2 ;..,
2 Silty. clay till
d.,
is 420 e
N O 50 100150 METERS l
SP'i"9
\\
I 8
2 415 O
250 500 FEET Silty.ciay till g
4 Vertical Exaggeration X 10 Figure 8B.--Generalized hydrogeologic section B-B' front shale uplands to Franks Creek shouing vater-table altitude on the north plateau, May 10,1983.
((Ane of section is shoun on pl.1.)
The only other major water-bearing unit on the north plateau is the upper 1 m of the shale bedrock. The upper part of the bed rock is fractured and yields approximately 0.6 L/s to borehole 83-4E (fig. 8A).
Although largcly saturated, the till deposits underlying the surf telal sand and gravel do not transmit significant quantities of ground water because of their low hydraulic conductivity. The recessional deposits of sand and silt that separate the till deposits are also partly saturated and yield some water to borehole 83-4E (fig. 8A). However, the thickness and low hydraulic conductivity of the till deposits restricts ground water flow to the lower recessional deposits and bedrock units.
Thus, the surficial sand and gravel unit is the most significant aquifer near the reprocessing plant f ac il ities.
HYDROLOGY OF THE SURFICIAL SAND AND GRAVEL The following discussion is based on data collected f rom October 1981 through September 1983 and on the results of the ground-wa ter flow model simu-lations discussed f urther on.
During t he data-colle ct ion period, ground-water levels were measured monthly in 25 observation wells finished in the surf telal sand and gravel, and slug tests were done to obtain estimates of hydraulic conductivity f or use in the model.
Streamflow in the two intermittent-stream channels that d rain the north plateau to Quarry and Fr anks Creek wa s recorded continuously at stations NP-1 and NP-3.
Altitudes of springs and seepage f aces along the periphery of the plateau were surveyed, and discharges from these areas were measured periodically.
15
i
?
E 4
'78'39'15" 78 '39 (Xr i
i k
g),'/ ~*-
ANP2 STREAMFLOW-GAGING STATION '
e
- 3 LINE OF EQUAL SATURATED THICKNESSI OF SAND AND GRAVELIN METERS. ]
I e 2.0 OBSERVATION WELL AND
.j SATURATED THICKNESS OF SAND ge,, V, AND GRAVEL. IN METERS EXTENT OF SAND AND
\\
GRAVEL DEPOSITS i
X X
42' g
A
$** /\\
5
/
fN
~
a NP
/
gk[gg34
-* " ~ h 2-
/
['
\\
NP T.
-s 1
,1 0-NP3
./
0.'Q W -* 1 WETLAND - /
N
\\Q 4
/
'2' 1
3
.f HAP AN.
a
)
l
,x
.9 / '.k' s
(
. 4.9 j
s x
0.3 3.0 e
0.5 1,5 f
N 2
5 JREEH
/-ilGH LEVEL 9
06
/
- ORAIN, e 2.8 RADIOACTIVE 12 s
TANK COMPLEX
% M3 0 RAD 40 ACTIVE
!I WASTE-STORAGE-HIGH - LEVEL v
\\
l JGP
\\.,,_
/
STE STORA0
/
^N'
/
X K CO E1
?
LAOOONS
~,
5.5 N
g i l
2.1
/
)
I h
2.2 w
y Q-q = r) (I f
b y
l
' i JEL
' 1.6
/ STAfL UCEN5f D G.3 b$
f ID ARA
'~
4 t
s
/
e
(
/N '
l Y
l f
Q 3
}
j
}/
42+
'4'
'I 27 f
\\
39 1
g '/
\\
'\\
.00" O
.c g
9 e o w.or, %,r q
x creen,.-
\\
\\.
.3
/
g%
%*****,,/'/,/,p y
\\
-s
\\
\\
p,.__.',
FAOUTY'S
\\~'g s.t gd['
DISPOSAL AHE A 7
\\
\\
/
o S,o ngo 150 MrTias s
x o
2so 560 ner
.\\
'9 '
/
\\_
ease non v s c m. ace son, y t..
Ashf mf HoHow 19791.24 OrX) t Figure 9.--Saturated thickness of surficial gravel on north plateau, May 10,1983.
(Modified front Bergeron and othere,1987.)
16
Flow Direction Under Undisturbed Conditions Ground water in the surficial sand and gravel flows radially away from the apex of the alluvial f an at the upland (southern) boundary and northeast-ward toward stream channels bordering the plateau. Some of the water from the f ractured bedrock along the upland boundary probably enters the sand and gravel deposit; the remainder moves northward within the bedrock. Ground-water flow through the sand and gravel is predominantly horizontal, and leakage into the underlying till is inconsequential. Discharge from the sand and gravel occurs by evapotranspiration and by seepage to intermittent-stream channels that drain the plateau and to springs and seepage f aces above the contact between the gravel and till along the periphery of the plateau. Th e directions of ground-water flow through the sand and gravel, shown in figure 10, are based on ground-water levels recorded in May 1983 and the altitudes of springs.
Influence of Plant Facilities on Ground-Water Flow The plant facilities indicated on figure 3 have altered the natural ground-water flow pattern by obstructing flow in some areas and providing pre-f erential discharge points in others. The ef fect s of these structures are discussed below.
High-Level Liquid-Waste-Tank Complex The high-level liquid-was te-tank complex and the f uel-storage pool f ully penetrate the surficial sand and gravel deposit and prevent the flow of ground water through these areas. The backfill that sur rounds these structures is less permeable than the original materials, and this res tricts ground-water flow.
Drainage Structures Two drainage structures--the f rench drain adiacent to lagoons 2 and 3 and the ditch connecting the wetland to the stream channel above station NP-3 (fig. 9)--were ins talled to dewater parts of the north plateau. Th e s e s t a-tions receive most ground-water discharge, and ground water flows toward these areas. Flow at both stations is continuous through the year.
Wastcoater Lagoons The low-level was te-treatme nt system has influenced the ground-water flow s ys t em in t he past by serving both as a source of recharge and a point of d isch a rg e.
Lagoons 4 and 5 were built above land surf ace in 1971 and, although their bottoms were sealed with silty clay till, wastewater leaked to t he surficial s and and gravel.
In 19 74 the lagoons were lined with a synthe-tic materlat to prevent f ur the r leakage.
Lagoons 2 and 3 were both excavated into the till underlying the sand and gravel, yet wastewater can leak into the surficial deposits whenever the water level in the lagoons rises above the contact be tween the till and the gravel.
Since such an incident in 1976, water levels have been controlled to prevent lateral leaka ge f rom the lagoons.
The f rench drain has reduced seepage to lagoons 2 and 3, but seepage still occurs along the southwest f ace of lagoon 2.
17
h f
78'39 15
78#39 00' A NP-2 STREAMFLOW GAGING STATION c
~4184 LINE OF EQUAL WATER-TABLE ALTITUDE, MAY 10,1983-Contour interval 3 meters. Datum is L3.6 EXTENT OF SAND ANO NGVD of 1929.'
GRAVEL DEPOSITS 5419.6 OBSERVATION WELL-Number is/
Nb water lovel, in meters above
- M--
.j NGVD of 1929i l
' [
p DIRECTION OF GROUND-15~
WATER FLOW 418.5. SPRING-Number is 413.(
41 3.6 omd h water level altitude'.
'I 417'4 44 8.2 NP 2 I e
in meters above 416.
El v$[on
-XU
.x ggtyg N't14,2 g NGVD of 1929,
/
!q
/h\\
I a
//0 416.3 D
?*a 'D,. I"[k- !?p4p1
/
g 3
15.4 s
419.6 417 414.6 V C
416.0 7 HARDSTAND 16.8 x
/
421.8
/
N.
424.2 422.9 e 5 k
/
\\
g 4274$M j23.1 O
423.2 a
3 hive
' OWNS
/
N 24 o
T 429 4 N ST STORAGi423.3 M
K COMPLEX g
)
b./427.3 4224
+ l 4:0.7 g
S 3
432.8
-4 k
h Q&44 L.,,
3 Wates tiev f
a J
ST 429
'FU _
/ ATE-UCENSED DiSPh5 L AREA 430.9 P
420 18.g 1
435.9
/
%4 423 18.5, l
42*
1 429.3O 426I 9f4 s,
L *2g'f~d
+
43s
%^L
\\
\\
b d
ois'POIAL A
.o Sp 100 160 METERS 0
250 soo rEET j
i i
NEEidiIw"fNYi"24Y I
Figure 10.--Water-levet altitude and direction of ground-vater l
l fice on the north plateau, May 10, 1983.
18
Be fo re lagoon 'I was removed f rom the low-level waste-t reatment system in 1984, it was hydraulically connected to ground water in the surficial sand and gravel, which allowed leakage both to and f rom the lagoon.
Lagoon 1 a ccepted wastewater fran the reprocessing f acility and the burial grounds only period-ically; mos t of the time, ground water seeped into lagoon 1 and flowed f rom there into lagoon 2.
During periods in which wastewater was transferred into lagoon 1, the temporary increase in hydraulic head probably caused leakage from the lagoon into the ground water.
The lagoon also caused a net loss of ground water through evaporation and overflow into lagoon 2 and thus represented a ground-wa ter-discharge point.
Other Structures Other structures that influence ground-water flow through the surficial sand and gravel are plant buildings and parking lots, which, together with the lagoons, drain to stream channels and have reduced by 17 percent the permeable area on the plateau through which rainfall can infiltrate and recharge the g round-water system.
Potential sources of ground-water recharge include a septic-tank leach field attached to the maintenance shop, possible leaks f rom underground water lines northeast of the plant, and infiltration f rom the plant's outf all channel that crosses the sand and gravel near lagoon 1 (fig. 3).
The outf all channel carries condensate and backwash f rom water filters and discharges to Erdman Brock east of lagoons 2 and 3.
(Se e fig. 3. )
Water-Transmitting Properties of Sand and Gravel Deposit Hydraulic conductivity of the sand and gravel was estimated from slug-test data and later modified during calibration of the flow model.
The proce-dure for estimating hydraulic conductivity from slug-test data is explained in the appendix. The assumed lateral distribution of hydraulic conductivity is s hown in figure 11.
The range of values, 0.6 to 10.0 m/d, compares f avorably with the range obtained f rom slug tests (see table 7, f ur t he r on), a lt hough the average value of 5.0 m/d is significantly higher than the geometric mean of 0.6 m/d from slug-test results.
This dif ference is probably due to errors in interpretation of the slug-test data, as discussed in the appendix.
Low hydraulic conductivity values are associated with backfilled areas near the main plant building, the high-level liquid-waste complex, and the low-level wa ste-treatment f a ci li ty.
The low hyd raulic conductivity in these areas causes a steeper hydraulic gradient west of the main plant than to the north (fig. 10). Areas of high hydraulic conductivity correspond to a buried stream channel on the surface of the till underlying the sand and gravel and produce the flatter gradient north of the plant.
The channel may be an ero-sions) f eature that marks the location of a former stream channel cut into the sur f ace of the till plateau. The resulting channel deposit would be composed of coarser material than that of the surrounding area.
The location of the channel is shown in figure A-1 in the appendix.
The specific yield of the surficial sand and gravel was assumed to range f rom 0.10 to 0.2 5 in accordance with values reported in Todd (1980,
- p. 38).
Lower values correspond to areas with a high silt content or to areas where a confining layer of silt and clay or fill overlies the de posit.
Higher values were asstned to represent areas of well-sorted sand and gravel.
19
a 78*3915" 78*39'00" i
s A NP2 STREAMFLOW-GAGING STATION f
AREA OF EQUAL HYDRAULIC CONDUCTIVITY (METERS PER DAY) l EXTENT OF SAND AND
[*
0.1-0.6 GRAVEL DEPOSITS
,e 0.6-1.8 1.8-2.4 Y-J h
5.010,0
%nd u4
, NP-2 y'
14 e
p
-s 03 e'.js i MM C
WETLAND O
f l-HA8
[
x i
i FRENCH l.
m i
ic l
fl
- h -Dh-C
/EL g
f A
's
'\\ \\
INE 2
j
_/
\\
A \\ vw
,RA d
i I
N V i.
PL 2
\\ \\
\\t N
I
\\ X
\\ -
~
N
/
\\ N s
Ju
\\
\\
"1 T"3 g
1 t
x x x r
c_-
l
]
l
\\
\\
2'
/
J at X
'JL jr
/
t y
\\
'L
' t.F.% /
\\
NI FUEL STATE-UCENSED VPROCESSING WASTE -
PLANT OtSPOSAL AREA k
jf_
h 1
$9 at l
g Logoon Road
- c l
A g[
OfSPOSA A EA O
SO 100 150 METERS Y
O 250
$boffET l
Baso from O S Geolowcat Surv7y' i
Ashfort1 Hollow. 1979 1 24 000 l
l Figure 11.--Hydraulic-conductivity values used in modet simulations of sand arri gravet on north plateau.
l 20 l
Recharge Ground water enters the sand and gravel on the north plateau as precipit-ation that percolates into the soil and as underflow f rom the f ractured bedrock along the upland (southern) boundary. The estimated magnitude of flow f rom these sources is given in the ground-water budget in table 2.
Some recharge may emanate as leakage f rom the main plant's outf all channel (fig. 2),
. underground water lines, and the leach field, but little information on these sources is available.
Annual recharge f rom precipitation was indicated to be 50 cm/yr by a mathematical model that simulates soil moisture content (Steenhuis and others,
{
1983). The model predicted direct runof f and percolation f rom the soil pro-file through a mass-balance approach to provide a daily accounting of. soil-moisture content.
Soil moisture in the model was increased by precipitation and snowmelt and decreased by evapotranspiration from the root zone.
Daily i
temperature was used to determine the timing and volume of snowmelt and evapo-transpiration. The volumetric flux of water through the soil profile was calculated as a function of the soil-moisture gradient and the saturated hydraulic conductivity of the soil.
Daily values of precipitation used in the soil moisture model were averaged from data recorded at three rain gages on the north plateau (Kappel and liarding,19 87). These values were increased' by 10 percent to account for the observed surf ace runof f.
The correction f actor is similar to the magnitude-of-measurement error reported by Winter (1981, p. 86) for rain gages installed above the land surf ace without use of wind shields.
Daily tempera-ture was averaged from records for the three weather stations mentioned earlier, and potential evapotranspiration was calculated from the modified Penman equation mentioned previously (Steenhuis and others, 1983). A vegetative cover of grass was assumed to grow f rom May to October, with a maximum root-zone depth of 90 cm in July. Af ter subtracting losses through surface runof f, the model calculated that half the 100 cm of the remaining rainf all becomes recharge, and half becomes evapotranspiration.
Table 2.--Ground-vater budget for sand and gravel de osit on 1.
nort;h plateau, October 1982 through Septem er 1983.
(Values are in centimeters per year.]
Recharge Discharge In filt ra t ion from St ream channels 13 f
precipitation 50 Springs and seepage faces 21 French drains Unde rflow from Low-level waste-(
bed rock 12 t rea ttae nt system 2
l Vertical leakage to till 1 Le akage f rom main Ground-water plant's out f all evapotranspiration 18 channel
_4, Change in storag e 4
Total 66 61 Mass balance error: 8 percent 21
Underflow to the surficial sand and gravel was estimated f rom Darcy's law.
If the saturated thickness of the deposit is assumed to be 1.0 m, the cross-sectional flow area near the upland boundary is 380 m. From figure 10, the r
2 hydraulic gradient near the boundary is 4.5 m divided by 60 m or 0.075.
From an average hydraulic conductivity of 5.0 m/d, the volumet ric flux, Q, is:
2 3
Q = (5.0 m/d)(380 m )(0.075) = 142 m /d (1)
I Expressed on an areal basis, annual recharge f rom underflow is 12 cm/yr. This 1
estimate could be in error by a factor as great as 4 as a result of uncer-tainty in the saturated' thickness and hydraulic-conductivity values in this area. For this reason, the volume of underflow used in the ground-water-flow l
model was included in a sensitivity analysis, discussed further on.
\\
Infiltration f rom the plant's outf all channel, which discharges steam condensate f rom the main plant building to Erdman Brook, was also simulated in j
the model. Although this discharge is variable, a flow of 500 m /d was 3
assumed representative of normal conditions on the basis of streamflow me as u reme nt s.
Infiltration of 10 percent of this volume would give a recharge rate of about 4 cm/yr.
Potential recharge from other buried plant facilities mentioned earlier was not considered in this model.
I Discharge Ground water discharges f rom the surficial sand and gravel through (1) drainage to s tream channels, springs, and seepage f aces; (2) leakage to the f rench drain and the low-level waste-treatment system; (3) vertical leakage into the underlying till; and (4) evapotranspriation f rom the water table.
l The estimated magnitudes of these discharges are given in table 2.
l Ground-water discharge to stream channels was estimated to be 148 m /d or 3
13 cm/yr over the entire surface area of the pla teau.
Discharges to stream channels were estimated from continuous streamfloa values recorded at stations NP-1 and NP-3 (Kappel and lia rding, 1987). Average monthly base flow at each station was determined by applying base-flow-recession techniq ues described in Todd (1980,
- p. 2 27) to streamflow hyd rographs f or Oc tobe r 1982 t hrough Se ptember 1983 (Kappel and 11a rding, 1987).
Ground-water discharges through springs and seepage f aces are dif ficult to measure because they occur over large, poorly def ined areas.
Some are i nt e rmi t tent and cease during the summer.
The estimated discharge to springs 3
and seepage f aces of 240 m /d or 2 L cm/yr is probably less than the actual volume because not all discharge could be measured.
Volume t ric measureme nta of discharge from springs and seepage faces along the northeast and northwest s(des of the plateau, which dealn to Ouarry and 3
Franks Creeks, indicated a tot al discharge of 20 m /d or 1.8 cm/yr (Kappel and lia rding, 1987). Discharges along the south side of the plateau wera estimated indirectly f rom streamflow measurements made on Franks Creek in April 1978.
These measurements indicated that flow f rom Erdman Brook on this side of the 3
plateau contributed about 810 m /d to Franks Creek.
9ubtracting from this value the measured surf ace flow f rom low-level was te-burial-g round drainage 22
and estimated flows f rom the f rench drain and the plant-outf all channel pro-vided an estimate of seepage from the north plateau to this tributary of 180 3
to 260 m /d (16 to 23 cm/yr).
Volumetric measurements of ground-water discharge f rom the f rench drain summer, and f all of 1983.
Ground-water discharge were made in the spring /d (1.7 to 2.4 cm/yr).
ranged f rom 19 to 2 7 m Ground-water discharge to the low-level waste-treatment system was estimated by subt racting additions to the lagoon from the reprocessing plant and losses through evaporation from the measured increase in storage.
The increase in storage in lagoon 2 was calcu-lated for f our periods f rom June through August 1983 during which no precipi-3 tation occur red.
Ground-water discharges ranged f rom 9 to 40 m /d and 3
averaged 26 m /d (2.3 cm/yr).
Vertical leakage from the sur ficial sand and gravel to the underlying till was estimated to be less than I cm/yr.
The calculation was made by applying 2 area on the cill surf ace.
The vertical hydraulic gra-Da rcy's law to a 1.0-m dient between the surficial deposit and the saturated lacustrine deposit below it is approximately 1.0 m/m.
(See fig. 8B.) The saturated hydraulic conduc-tivity of the intervening till deposit was given by Prudic (1981) as 2.0 x 10-8 cm /s (0. 6 cm/yr). Vertical leakage (Q) f rom the sand and gravel is thus:
2 Q = (0.6 cm/yr)(1.0 m )(1.0) = 0. 6 cm/yr (2)
Ground-water discharge through evapotranspiration includes the amount of the soil-moisture deficit that is replenished by ground water and direct evapo-ration where the water table is at land surface.
The soil moisture model pre-dicted that the root zone would provide 50 cm/yr of the total evapotranspira-tion loss of 68 cm/yr f rom native grasses, mentioned earlier.
Evapotranspira-t ion losses f rom ground water would therefore total 18 cm/yr.
This estimate would be low for areas in which the water table is at land surface and high for areas where the depth to water exceeds 1.0 m, the approximate base of the root zone.
The annual change in ground-water storage in the surficial sand and gravel was estimated f rom dif ferences between water levels recorded in obser-vation wells 80-1 through 80-8 (pl.1) in Oc tober 1982 and those recorded in September 19 83.
Two wells on the upper (south) part of the plateau (80-1 and 80-2) showed an increase in water level; 80-8 showed no change, and the rest showed declines.
The cumulative change indicated a net decline of 0. 2 m.
Assuming a specific yield of 0.20, the change in storage represents a net ground-water discharge of 4 cm/yr.
Seasonal Fluctuation of Ground-Water Levels and Discharge Ground-water levels in observation wells fluctuate as much as 2 m during the year, and ground-water discharges to some strean channels, springs, and seepage faces can change by nearly 100 percent.
Seasonal fluctuations of ground-water levels in well 80-8 south of the reprocessing plant and 80-4 near the wetland f rom October 1982 through September 1983 are illustrated in figure 12; base-flow. discharges at two t ribut aries, the french drain, and springs and seeps are given in table 3.
23
t Table 3.--Maaoured ground-vater discharges from north plateau in March, 7uly, and October 1983.
\\
(Drainage area 41.4 hectares.]
Measured discharge
'(liters per second)
Location 3/3/83 7/5/83 10/6/83 i
f N P-1 0.37 0.03 0.62 NP-2
.09
.03
.10 N P-3 1,76
.59 1.56 French drain at lagoons
.28
.22
.31 Springs and seepage f aces
.58
.27-
.62 l
TOTAL 3.08 1.14 3.21 430.5 WELL 80 8 1982 WATER YEAR *
--- 1983 WATER YEAR 430.0 i
429.5
/
rs
\\ / \\
.cf...........................
'/
\\
......~....N.....
...............M..E.A..N..
429 0 ks 6
A vs a
{ 428.5 y'
2 420.5 WER E
1 g\\
420 0
,/
J f -t
"\\
N 419 5 f
j s
a
..3...
................e..
1 419 0
/
N N'
418.5 f'
f
/
' Water year extends from Octoter through September o' following year.
(/
418.0 OCT NOV DEC JAN FEB MAR APR MAY JUN JUL AUG SEP I
Figure 12.--Hydrography of celle 80-4 and 80-8 on the north plateau, tJaber years 19 82-83.
(Vell locations are shoun on pl. 1.)
Seasonal Patterna l
Water levels were highest in the spring af ter periods of snowmelt and
{
precipitation, when large ground-water discharges also occur red.
Water levels declined during the late spring and reached the lowest level in July, when the rate of evapotranspiration was highest.
Gr ound-water discharges generally 24 l
I declined by 60 percent f ran March through July, although flow from the f rench J
d rain remained f airly constant because the saturated thickness in that part of the north plateau was sufficient to maintain discharge throughout the year.
Wa ter levels rose with increased recharge in the fall, and ground-water dis-charges returned to high levels in the spring.
Wa ter levels continued to rise through the early winter until f reezing temperatures began to limit recharge.
Ground-water levels and discharge fluctuate in response to seasonal var-intion in recharge.
Precipitation falls mainly as snow f rom December through February and remains on land surf ace until it melt s.
Some of the snowmelt infiltrates to ground water and, when the snowpack melts in March and April, provides the greatest volume of recharge.
Alt hough brief warm spells can occur during any of the winter months, the potential for infiltration during mid-winter thaws is limited by the extent to which the soil remains frozen.
As temperatures increase and vegetation growth resumes during spring and summer, evapotrans piration removes much of the moisture held in the unsatur-ated zone.
Evapotranspiration directly f rom ground water occurs in areas where the wa ter table is near land surf ace.
Only a small amount of precipita-tion recharges ground water during this period because evapotranspiration keeps the moisture content of the unsaturated zone below saturation.
Re cha rge is possible during periods of extended precipitation, however, when the infil-tration rate exceeds the evapotranspiration rate.
Evapotranspiration dimin-ishes with killing f rost in the fall and thereaf ter allows a greater propor-tion of the precipitation to recharge the ground water.
Calculation of Monthly Recharge for Transient-State Model Seasonal changes in recharge were incorporated into transient-state simu-lations discussed further on.
Two sets of monthly recharge rates for October 1982 through September 1983 were used in the simulations; these were calcu-lated from (1) the soil moisture model previously mentioned, and (2) a method based on the monthly soil-moisture deficit.
The latter method uses precipita-tion and evapotranspiration losses for each month and the annual recharge rate determined by the soil-moisture model.
Evapotranspiration rates we re calcu-lated by applying a consumptive-use coef ficient of 0.75 to the monthly poten-tial evapotranspiration values, shown in figure 7.
Monthly recharge rates, RECH, were then calculated by:
i At=
PRECIP4 (RECHavg - ET )
i PRECIP (3) avg and RECHi= At - a in (4 )
m where: PRECI Pi
= measured precipitation in month 1, PRECIPavg = mean monthly precipitation, RECH
= mean monthly recharge rate give n by soil-moisture model, avg ETi
= estimated eva po transpiration rate i n mo n t h 1, 6 min
= minimum value of 6 f or i = 1 through 12, and R ECHi
= recharge applied to model for month L.
Th e A value def ined by eq ua tion 3 is related to the soil moisture deficit and weights the recharge in each month by the measured precipitation and the 25
s es'timated evapotranspiration rate.
Equation 4 ensured that some recharge Loccurred in every month for which At> Amin because Omin was less than zero.
Recharge was assumed to be zero in July (for which At = o in) and in January m
and' Februa ry, when the soil was largely frozen.
Pot ential recharge f rom the.
latter.2 -months was assumed to occur in March.
4 4
~ Monthly recharge calculated by the soil-moisture model was greatest.in March and declined to zero by June. A lesser amount of recharge was calcu-lated for' September through December.. Recharge calculated by the soil-
' i moisture-deficit method was distributed more uniformly through the year. -
5 Monthly ' recharge rates calculated by the two methods are presented in figure 13.
i i
~
f
-X -VALUES OBTAINED THROUGH "
l.
Ig EQUATIONS 3 AND 4 l-l\\
--o-- VALUES OBTAINED FROM
\\
'll SOIL MOISTURE MODEL
~
.\\
iSteenhuis and others,1983):
a g
\\
k I
t g
V 5
I b
\\
'I
\\
\\
l l
10.0 I
s x
n/
I
\\.
1
\\
- e- - e l
\\,
1 i
s N
I X
\\
i s0 I
\\
x\\
g I
\\
l
[
s i
s s
\\
on
/.
OCT NOV DEC JAN FEB MAR APR MAY JUN JUL AUG SEP Figure 13.--Monthly recharge rates, 1982-83, calculated by quations 3 l
ani 4 and by soil-moisture modet used in transient-state eimulatione.
SIMULATION OF GROUND-WATER FLOW The ground-water flow model developed in this study was calibrated to I
ground-wa ter levels and discharges measured on the north plateau during 1982-83.
By calculating ground-water flow paths and velocities through the suf ficial Sand and gravel de posit, the model can be used to predict travel-j times of conservative solutes in the ground water.
The model was calibrated I
through steady-state simulations that represented average flow conditions and l
[
t ransient-state simulations that incorporated the ef fects of storage during a l
l year period, October 1982 through September 1983).
)
l 1
26 l
l
4
(;',
I-Flow Model.
Ground water 'in the surficial sand 'and gravel on the north plateau is unconfined and flows laterally, parallel to the surface of the ' underlying till.
Ground-wa ter flow in this system can be described. by the following partial differential. equation governing two-dimensional flow through a porous medium:
i Bh B_
T 33 +
8_,
T B) -W=Sy xx yy 0x.
Bx By By at (5)
H
-1 where: - x and '-y ' =~ cartesian coordinates align ed ' on t he major axe s of transmissivity, Txx, Tyy, h = hydraulic head (L),
W = volumetric flux per unit area representing sources -
and/or sinks. of water (L/t),
S
= specific yield (dimensionless), and yt = t ime ( t ).
The model;used in this study solves a set of finite-difference approximations -
l to this eq ua tion. ' It uses a grid to divide the aq uifer system into an array',
i of rows and columns.- A complete discussion of ' the model derivation-is given in Mcdonald and Harbaugh (1984).
Design The north plateau area was simulated on a finite-dif ference grid that j
represents an area of 41.4 ha.
The model grid (fig. 14) contains 644 cells (23 rows by 24 columns), of which 202 are inactive. Each grid cell represents
- 30. 5 x 30. 5 m, r oughly 1/10 h a.
Model boundary conditions were selected to approximate the effect of real i.
hydrologic bondaries.. The types and locations of boundaries used in the model are indicated in figure 14 No-flow boundaries were specified to correspond to' the high-level liquid-waste-tank complex and the main plant building. ' The
-l surf ace of the underlying till was also treated as a no-flow boundary because j
the quantity of flow through this unit represents'less than 2 percent of. the estimated ground-wa ter discharge f rom the no rth plateau (table 2).
q The surficial sand and gravel is recharged by underflow f rom bedrock west
. of the main p1 ant building along Rock Springs -Road (fig. 14). This boundary was initially represented in the model by constant-head cells in which the water table was maintained at a constant level, and later represented by.
constant-flux cells. 'Ihe volume of underflow entering the surficial gravel at each constant-flux cell was determined by the model with the cons tant-head boun dary. A range of constant fluxe s was then used with the steady-state l
model to test the effect of underflow on the predicted hydraulic-head distri-l-
bu t ion. Constant-flux cells were also used to simulate leakage f rom the ou t-
' f all channel that d rains the main plant area.
l Ground water discharges f rom the plateau to stream channels upstream f rom l
stations NP1, NP2, and NP3, the f rench d rain, and the low-level waste-treatment j.
system (fig. 3) and also through seepage faces along the edges of the plateau.
These discharge points were simulated in the model by drains.
Each drain assigns seepage frm a grid cell at a rate proportional-to the difference in I
s
~
i 27
~
78'30'15" 78*39'00" i
ANP2 STREAMFLOW-GAGING STATION DRAIN CELL-3 CONSTANT-FLUX CELL
. h NO FLOW CELL EXTENT OF SAND AND 0
e,%
r 0 0 GRAVEL DEPOSITS o
D 1
42' h
O 2r 15*
h O
1
(
O N o
3 yo p
O
["O O
ql.
c j
O g
s O
'8
,o N
Q
\\
r O
O g
)
- b do a
g A,. !!i. ~
4
< o R.
to Tiv G
V 2T G
j!!,)
O L
t O.
O >
\\/
,4 o ~~
l
/ /
FL.
O STATE-UCENSED
/
4 7
,e y
O~
WASTE -
O.
DISPOSAL AREA p
P
\\
/
/
42"
/
2T 9
l
/
f n
/
$c"e
$g p
O
/
O Erdf N
DISPOS L A EA s
0 50 100 150 METERS O
250 Eb1 FEET
\\,
Base froin U S. Geosegita: Survey Ash'<v d Hnllew 19 791 ^4 000 Figure 14.--Finite-difference grid and bourdary conditione usai in model simulations.
28
j
. altitude between the water table and the drain.
The model calculates the rate of seepage thr ough the equation:
Q = C(h-d)
(6) l 3
seepage rate (m /d),
where: Q
=
conductance of the interface between the aq uifer and C
=
the drain (m /d),
j 2
hyd raulic head in the aquifer (m), and h
=
elevation of t he d rain (m).
d
=
Drain conductance was defined as:
C = AK (7) 1 1
2 y
average cros s-sectional flow area (m ),
where: A
=
hyd raulic conductivity of the interface (m/d), and K
=
1.= flow path length (m).
j Drain altitudes used in the model were assumed to be 0.3 m above the surf ace of the till to provide suf ficient saturated thickness for seepage to occur.
Estimation of drain conductance is described in the discussion of hydraulic conductivity f urther on.
J l
Input Data Hydraulic Conductivity.--Hyd raulic conduct ivity of grid cells ranged f rge
?
0.6 to 10.0 m/d, in accordance with the distribution shown in figure 4.
Th e model computed saturated thickness of each grid cell by subtracting the alti-tude of the bottom of the gravel deposit f rom the altitude of the simulated 1
water-table surf ace. The bottom altitude of the surficial gravel was deter-mined f rom a contour map of the surface of the upper till unit ( Albanese and others, 1983). Altitudes of the till surf ace at the edge of the plateau were f:
interpolated f rom the altitudes of springs shown in figure 10.
6 i
Hydraulic conductivity of the surficial sand and gravel was used to esti-J mate the conductance term, C, in eq uation 6.
Mcdonald and Harbaugh (1984) discuss a variety of f actors that may af fect drain conduct ance, including the thickness of the interf ace between the aquifer and the drain, and the differ-ence in permeability between the aquifer material and drain material.
No inf ormation on the hyd raulic properties of the interf ace was available.
Drain conductance was estimated on the assumption that the length of the flow path thr ough t he in te rf ac e wa s 0. 3 m.
The hydraulic conductivity of the interf ace was assumed to be 0.03 m/d, which is within the range of values for fine sand and s il t (To dd, 19 80, p. 71). Diagrams of the cros s-sectional areas used to estimate conduct ance of the four types of d rains simula ted by the model are shown in f igur e 15.
Initial estimates of drain conductance were modified during' calibration to improve model predictions of ground-water discharge and
' hyd raulic head.
i Recharge and Evapotranspiration.--Ground-water recharge was assumed to
{
occur at a unif orm rate across the north plateau. The amoun t of recharge l
8 applied to each gr id cell totaled 50 cm/yr, the value obtained f rom the soil-moisture model mentioned earlier.
This rate was later adjusted during cali-l bration to determine model sensit ivity to recharge.
j 29 j
I
. ! P.
- M;t3
~
'1 pu
$?.
k
. g;cf.
O E'
,j
' lf
'[
/;j#
)
g r;+ f h
t
'gf
-f r
's,
is
.!q
/
}
g
-f,
/
s 2'
)
!/
4 <'.
s
'b
, J. - - / * ! Tp l
0
,.l s
8 l) l
/
8 ag
/
l.,/f'S' l
/
l v
(
'i T
STREAM CHANNEL AREA = 2xBxS FRENCH DRAIN AREA = wCS SEEFAGE-f ACE AREA = 8xS
-l ((
r l
. MN AO
_w<,,
~g
/
[
t'.
[ /
WATER TABLE
's t
- {_
.p, SEEPAGE FACE
/ I j,
't 4
i_' 'f q.
-~_
, i '.
g}_yry q
/
I
- r
- Y.tY
, f liUHHilU r Figure 15
.r.
. t-4,f
\\
r-
[f q'
The four types of draine u' sed in the ground-vater f2ou mclet, b
l a
uith terme for calculation of
~
yl croea-sectional area. ^ (Locatione l
are indicated in fig. ^ 14.)
f LAGOON 1 AREA = LxW
- r.
.,(
e g
,f t
- e. f i
^
Evapotransp15ation from ground water onithe north plateau is not uniform n
but depends upon the depth of the wateirtabid.
Evepotranspiration from each
~
' grid cell in the model was determihed)f rom the f unction -illustrated in figure r
- 16. - No evapotranspiration occurred 1p/Mid cells where the water table was below the root zone, which was es>hmac[d to be 1 m thick. In grid cells where
.I the water table was at or'above the land surface, evapotranspiration was I
fi assumed to occur at the actual evapot'ranspiy' ation rate. For water levels q
- r -
,within the root zoie, f.he evapot transpiration rate was propo rtional to the,
j dept,h to water.
3[
yt '('
1 l
j,)
/I' Land surface p'
7.
i2 E
a t
I,
\\
d d.
f c
3v lp Il Figure 16 1
l~
-9 o
i, e
lo I;f Relationch,ip betueen..
/
aase of root ron.
, Y'
~~fL ~~ ~ ~~
vater tevit in model cell J
J and evapt; inspiration rate.,
.i i l
(Modified f ecn Mcdonald I
arti Harbaugh,1984, p. 317. )
f,0 inc,,,,4 f,,
y,,l mum
-l EvAPOTRANSPIRATION PATE fo
/
f f.
77 g r; r
h
- h a
_ \\.l *
,} < '
sk V
4 V %g1 y's s ) ' y'
]
l A
\\
y I '=
h Qt
,j
,y Y
3
',i s
,\\
The O phc f acilities and associated paved areas prevent recharge and H
ra tion Wa r a s ig nif ' ean t, area of tha plateau. Acti $.Vrid cell /,, T g'
evapot ransF {harge and evapotrans piration were not in which ru.
alloma to ocgr are indi-
l 17.
.t cate(dinfigur:
. ' ~
i, \\
p s
q 1
At-Steady-State Simulations i
, \\{ s
/
a i y
T
-)
]j Steady r(cate simulations were uased to calibrate the ground-water-flow
\\
l
(
~) m> del,to >ty('/(cWthe mean annual water-table altitude.
For the purpose of t't }# study, v.L Mmulated head distribution was compared with ground-wa ter l
1evels measured in 23 observation wells on May 10, 1983, when water levels l
were close to the mean of levels recorded f rom Octobe r 1981: through September j
19 83 (f ig, /12). Durig this period, the total precipitation 1 (92 cm) was slightly lower than the r9ean annual rate of 100 cm/yr.
Sidlated ground-wa ter.
discharges were compared with mean daily discharges derivcd f rom the annual j
discharges of cercam channels above stations NP1, NP2, and NP3, and the french drain.
(
3 n"N f
Calibration The model wh calibt\\ted by comparing simulated ground-wa ter levels and I
ground-water didharges with those observed in the field.
The calibration i
f procedure, entailed trial-and-error adjustments of boundary conditions and j
l hvdraulic conductivity, drain conductance, r e ch a rge, and po tential evapotrane-gtration. Values were selected to cover the range of uncertainty associated with each term.
Progress,in model calibration was measured by the mean abso-lute dif fejence of crtimate between the simulated and observed hydraulic head in the 23 observation mila, and the root-mean-square dif ference of estimate between simulated and oberved diset:arges into the NPI and NP3 channels and the f'rench drain !
j S.Inc itivity Optimum coef ficients obtaine:d with the steady-state model are presented 7
ir table 4, which also su:rmarize s the results of sensitivity testing to eval-unte encla terp/ n relative ef fect on simulated ground-water levels and discharges.\\ Reruits of 10 sviady-stste simulations are listed, in which the calibrated val'ue of a single variable was changed while the values of the,
- l n
1 W ln atisolute dif ference = E (O P ),
i i t
n
[
. b, 1
2~ 1/2 root-meanjquare di'forence i E (O _Piy i
i U1 i
n l
rhere: Ot = observed value at point i, Pi= predicted value at point i, n = number of poihts.
)
31 J
7 2
7 7
9 w
2 3
o l
0 0
f r
e 5
3 2
d 5
5 7
1 n
0 2
U 0
0 6
1 6
6 y
0 6
1 1
t si 0
0 n
i v d
o si i
yt t
l c 6
8 0
5 l
nd 0
6 a
au 1
1 u
an O
0 m
o C
i yc s
t i n 5
e vi 1
9 4
2 t
i a 0
2 5
3 a
t r t
iD O
0 s
s C
n y
e d
s 0
a.
y es nt 0
ti ii 2
i.
ss v
6 0
1 y
di 2
hl et 6
3 2
e ga t c 1
un su 0
oa ed r
t n 0
h y o
9 5
3 2
tt sc 5
2 8
i e
dv uc 3
0 ei l i nt al O
ii V u 0
7 8
as a
3 7
1 t n r
0 b e d
0 os y
5 H
a sf eo u
4 l e e
7 8
4 8
6 1
at g
2 vl r
0 u
a 0,
ns h
we c
2 4
5 1
7 8
mr e
4 1
i R
ti 0
0 pn Oa d
m 0
e yeue 3
v d t mg 6
0 0
0 1
0 r
1 1
e 4
0 2
5 4
aain 4
6 1
s et t a b
e t s pr 0
6 S
o os
)
M
)
e t
a r
nu n
T a
y el e
)
rm/
ea c
)
r oo m wv r
r y
pr(
t e
y on b
/
af ed p
/
p o y
y m
v r
b e r
r(
m ai) t t
(
es e t
e e
(
vt r i
i et ea t) t s eay cv v
w d s a cl a m ae e
e r/
i i i
o esw nu w(
wg l
g mim l t t
l t o em
- r b
r u p(
uc c
f al d ri d s d a a
a ms au) nu) r l
n es nl nh i
h i ne rdd i dd e
unu f
ue uc r
c xat d n/
an/
d moo f d ov os a
e a ra yom rom n
i i r i n r e ri V
R Mt r c(
D c(
U S t g D a Gl Gd l
I U
,1 1
78'39'15" 78*39'00" I
i
- A NP2, STREAMFLOW-GAGING STATION
- MODEL CELLS IN WHICH RECHARQd 9:y. '., AND EVAPOTRANSPIRATION WERE NOT SIMULATED,
/
ss EXTENT OF SAND AND x
xh GRAVEL DEPOSITS 2T
~
~
%nd e
t.
" 7Pd 7s p
l WETLAND
-3 g 9
h
./
E
/*
(
HARDSTAND FRENCHf DRAIN
<k
\\x 6
s 4
/
N
): <
~
H H WEL
,/
4 f)
STE -STO AG e.
.T L
f
~ "l
+
e STATE-LICENSED F L P
/
OISPOS AL AREA
/
4 42a 2r~
(
ef'/1.
\\
x g
j' uagn,noas s
,s OfS50IA?A EA l
0 50 100 150 METERS Y
b 250 550 FEET
=
l Figure 17.--Model cells in which recharge awi evapotranapiration vere not simulated.
33 i
l
other four variables were held cons tant.
The table includes the simulated ground-water evapotranspiration value calculated for each simulation and the resulting dif ference between observed.and simulated values of water levels and di sc ha rges.
Ground-water evapotranspiration served to reduce the ef fect of changing the optimum value s.
In simulations that produced higher water levels than were obtained with the optimum values,- ground-water evapotrans pirat.on increased, which limited the resulting rise in water levels.
Similarly, in simulations in which water levels were lower, ground-water evapotranspiration decreased.
The ef fect was that water levels were restricted to a fairly narrow range in all simulations used in the sensitivity analysis.
Recharge was the term to which the model was most sensitive in steady-state simulations. Table 4 indicates that reducing ground-water recharge by 10 percent f rom 46 to 42 cm/yr increased the dif ference between simulated and I
observed water levels f rom 0.51 to 0. 75 m and the dif ference between simulated' and observed grour.d-water discharges f rom 4 to 18 percent.
Decreasing recharge also caused parts of the simulated area to go dry. Changes in recharge also af fected the volume of ground-water discharges.
In general, the lower values of recharge could not produce the discharges observed in the stream channels at KP-1 and NP-3 nor those in the f rench drain.
Average hydraulic-conductivity values of the surficial gravel used during steady-state calibration ranged f rom 0. 6 m/d to 10.0 m/d.
Decreasing the optimum values of hydraulic conductivity by 50 percent increased predicted i
ground-water levels by 1 m and produced much smaller ground-water discharges
{
than those observed.
Doubling the optimum values lowered ground-water levels and dewatered parts of the plateau. The spatial distribution of hydraulic conductivity used in steady-state simulations significantly improved model predictions.
This can be seen by comparing errors associated with the optimum model run with that based on a uniform hydraulic-conduc tivity value of _ 5.0 m/d (table 4).
Jhanges in drain conductance af fected the ground-water discharges and some water levels near the seepage-f ace boundary. Discharge into the low-level waste-treatment system was extremely sensitive to changes in water level. A water-level decline of about 0.3 m caused a nearly fivefold decrease in discharge. This predicted decrease in discharge agrees with the estimated decrease to eepage losses to the lagoon during a period when the water table declined by a similar amount.
The steady-state model was also sensitive to changes in underflow through the upland boundary. Doubling the underflow caused water levels to rise I to 2 m in the upper (south) part of the plateau and greatly increased the rate of i
evapo t rans pi ra t ion.
Decreasing underflow by 50 percent lowered water levels by a similar amount and dewatered a large section of the plateau.
l Re sults Water-level altitudes computed by the steady-state model and those measured on May 10, 1983 a re plotted in figure 18.
The mean absolute dif fer-ence between computei water levels and those measured in the field is 0. 5 m, 34
i 78'39'15" 78'39 00~-
.j i
i i
-)
SIMUL ATED BY GROUND WATER FLOG. g)[
]
LINE OF EQUAL HYDRAULIC HEAD e
- 418 ---
/.
MODEL-- Contour interval 3 meters /
Datum.is NGVD of 1929, 0419.6 WATER-TABLE ALTITUDE MEAS-y URED IN OBSERVATION WELL v9
/
EXTENT OF SAND AND d~NP-2 STREAMFLOW-GAGING STATION GRAVEL DEPOSITS.
o
.e, 42' k
2T
~
15' 437 417 4
+
O 'cond u
p 34-0 1Y.9
. \\
- )
s 4
f'[\\
I 41 g
g
/
/
1 WETLAND /
\\
\\ -
e' 419.6 0
(
426 RDSTAND 420
%- g FRENCH q
If8
]29 '
/
421.8 ORAIN
,\\
~J
, [#
N 427,2 423.2 0
4 % 427.9 O
O O
O o
/
N
> h 424.2 423.1 422.9 f
3
,- x
^
{
Lp (As
,/.
\\
l 4
432 42b B OC E
)
(
M4PSTE S]MPtpRAG423.
/
h T ANK CO X
p
/l 423 423 r
4 P
O e
432.9 y
1,j f
/
/<
' F UE STATE LICENSFO 435 D'5P StL ARE A 4309 P
C' 426
/
[/
\\
~
y*
i 432 00" j
42'9.30
,,,nf '
J
/ ogoon Road
" /
{\\
/, /,. '
x Q:-
-A
\\
<^
FACILITY'S
)
a j, god,,-
DISPOSAL ARE A 0
5,0 100 150 METERS O
250 500 F EET
\\/
l t
Base from U b Geolowcal Survey
_.... ~ _. _
.-.L__._
1 Ashfved Hollow. 1979124.0fK)
Figure 18.--Simulated stealy-state and neasured water levels on the north plateau.
35 1
fl
which is less than.5 percent of the 16-m difference between hydraulic head measured at' the highest and lowest points on the plateau. Ground-water levels computed by the model are within 1.5 m of water levels measured in the 23 observation wells.. The principal inflows and discharges of ground watet-pr e-dicted by the model are listed in table 5.
The root mean-square dif ference between measured and simulated ground-wa ter discharges to stream channels upstream f rom s tations NPI and NP3 and the f rench drain was 4 percent.
I Discharges to the-low-level waste-treatment system and the Franks Creek tribu-tary compared favorably with spot measurements made at these points.
The ground-water budge t computed by the steady-state simulations closely
. parallels the-budget given in table 3 (p. 24). The areal recharge rate of 46 cm/yr used was nearly 50 percent of the measured precipitation and slightly lower than the 50 cm/yr predicted by the soil-moisture model.
Underflow from upland sources was estimated to be 17 percent of the total recharge.
Combined discharge through seepage f aces was.nearly twice the discharge to the channel above station NP-3.
Evapotranspiration f rom the water table was estimated to be 20 cm/yr... Including evapot transpiration from the root zone would give a total annual rate of 70 cm/yr, or 77 percent of the calculated potential rate.
This agrees closely with the evapotranspiration estimate obtained f rom the Penman equation, as described earlier.
' Table 5.--Recharge and discharge values calculated by steady-etate model.
(All values are in centimeters per yearl Recharge sources Discharge sites Infilt ration f rom Eva po tr an s pir atio n 20.0 precipitation 46.0 NP-1 channel
- 2. 2 Underflow f rom bed rock 10.4 NP-3 channel 10.0 Le ak ag e f r om French d rain 2.1 outfall channel 3.7 Low-level wa st e t rea tme nt system
- 2. 2 Franks Creek tributary 14.5 Ot her seepage faces 9.0 TOTAL 60.1 60.0 distribution of discharge rates simula ted by the model is plot ted in f
a 19; the distribution of evaporation is shown in figure 20.
The main ocharge areas were on the southeast boundary of the plateau along Erdman Brook and on the northern part of the plateau near the NP-1 and NP-3 channels.
Eva pot rans piration fran the water table, which is restricted to areas where l
the depth to water is less than 1 m, occured mainly in the we tland area or the north plateau and along the north boundary near the seepage faces.
36
i G.
f.1.'
'l 78'391S*
78"39'0(F
' (). ; DISCHARGE-AREA ROONDARY.
W'
- Ql r RECHARGE-AREA BOUNDARY.-
i
-0.3 '-
. RECHARGE (a) OR DISCHARGE (-)
-04,5%. -
RATES,lN LITER $ PER SECOND',
s
.f
(,
4% '
PERCENTAGE OF TOTAL' LRECHARGE OR DISCHARGE /
EXTENT OF SAND AND fg 4
A N P-2 STREAMFLOW GAGING '
"f N.
GRAVEL DEPOSITS STATION
- ,l
'g 4,, g s --
,YI
's a
~
42' l'
\\
'+,'y--
-',/
27'
~
[ 'j
.lll g
- 15
y N
- g
- < "m y
NP-2
/\\
p__4 ~t[., ) !.Y N-K p"* s i
g.
A
-T NP.i
(-
^*
. Ob,4',/
-s s
1 g,7 q
tj 1
' }, N 4 )WULAND f(
I r w\\
s P.3 '
g g [' '
f
[~.-
f 3,4%$
[315- -e' b;
f'- 0.5.6%,e
,,/
A
-1.3.17%
't-1 i
A}
[
FRENCH
- g HARDSTAND g
s ORAIN
-ff g
[
N
-(\\..
l N
>J e*y *'.m x
4 i
l
\\p RADIOACTIVE LAC /*f+A
%,,,,N f
HiGH LEVEL s
i.
. }
f-l
(
WASTE STORAGE -
p,
/ 'A
}
)
TANK COMPLEX f,
f
/ \\\\
f g
-0.3,4%
,j
,)
+0.5.6% '
' pun SI A f f lKt NSED l
//[
\\.
PLANT WO6 At AL'f A WAW REPROCESSING '
j
/\\ )!
l,'/ 'li
./
/
L
\\
/
\\
QN
/)
(~/f
/
g/ *1.8.23% (
a
\\
/
27 00"
+1.3,17%
j
/
j \\
s f
i 's._ /
mes.w/.'
g s
y 3
.r
/'
,/
\\
f
. = =
,/
A
^
s-.*
\\
r ACluty L
, k.'< g, p
(/re osm a u f, \\
g.
s+
g O
5,0 ICO 160 METE R$
)
\\
\\
b 20 (2)Q FEET
\\
/
7 B'aw from U S Geologral Survey
~
~~
Ashfort! Hollow 1979124 (m Figure 19.--Prsdicted steady-atate rates of' recharge and discharge from contant-f7ux and drain boundwies.
37
. - ~.
18'3915-78"39 007 I
. 't ' *)
BOUNDARY OF AREA IN WHICH GROUND ep WATER EVAPOTRANSPIRATION WAS,)
SIMULATED
. BOUNDARY OF MODELED AREA
- 0,2 GROUND-WATER EVAPOTRANS-PIRATION RATE,IN LITERS -
Th 4'}
. PER SECOND EXTENT OF SAND A' D N
3%
PERCENTAGE OF TOTAL m
GRAVEL DEPOSITS 4 2',
./
4-kh O!SCHARGE 3
27'
~
A NP-2 16" STREAMFLOW-GAGING STATION f
~,/
r YPond
-0.6. 8%
e a
p
'/
NP 1 4 %
o$/y i
r
,,\\
l
'a V \\
\\
i
"",f' I
(@. @ J WEThD /
i O
/
42;3%
rJL.
/'
% ~ % W )./
s )~ \\
r
\\%
.)
//\\
\\
N
% 434%
/ l
/
/
N/
y
.(
\\
AROSTAND %
p
'g
- = /
FRENCH-N
\\
AREAL RECHARGE:
DRAIN
'7%.,
N
-0 2,2%
(
+0.6,77%
g s, ) (^ % -
9 's Y
,, j,
/
]
s
\\\\-
/
\\
3 g
HIGH ' LEVEL '
f5
' /
's's.
RADIOACTIVE LAf
)
)
WAS TE -5TOH AGE - f/
f j
/-
ac g
TANK COMPLEX e/ 2 f
lt l
/j
/
I
(
t
./
n /
P
^l 1
s y
/
/
%s 0%
.+
g
(
_ g
--y)"
jM - -
l/
~1 n
f 3, srm uctssm
' rut i
REPH SING D$P A\\ AREA g
-9.3;16%
/
/
I
\\
l/
f-f' 42*
- g
[< !'
'\\
2T
~
g N
5
\\
00" 5
- 89
,d,
\\
u l
j iagoon Road
~
Crech /
\\
-y #
1
\\
/
i p
F AOUTT5 gp DISPOSAL ARLA,/ }
\\
,s
\\\\\\
0 60 100 150 METE RS
}
HL A
1 0
250 LOO F EET
/
\\
[
, Haso from U $ GevkignM Survey
_ _ _..... ~. -. _..... -
L --- -
Ashfortf Hodow. l$'/9124 (W Figure 20.--Prediatel steady-state distribution of evapotranspiration.
38
The most significant discrepancies between the. hydraulic-head distribu-tion generated by the model and the measured water-table altitude were near.
seepage f aces along the plateau's southeastern ' perimeter. Water levels gener-ated by the model were 2 to 3 m above land surf ace in several grid cells along this discharge boundary. This dif ference is attributed to the uniform grid spacing selected for the model, which resulted. In an abrupt change in average saturated thickness along a' flow path approaching the' boundary.
Decreasing the grid: spacing by 50 percent along this boundary lowered the simulated water
' levels by 1 to. 2 m, ' as shown in figure 21.-
.1 C
C' l
o GROUND-WATER LEVEL PREDICTED
- l WITH UNIFORM GRID SPACING a
1 435 o\\
'X GROUND-WATER LEVEL PREDICTED [
~
\\
WITH VARIABLE GRID SPACING
\\
Land Surface i o i MODEL CELL l
- s \\
~
s
,430
% 'Os,
[
k Surface of Lavery Till l
W Surficial Sand
' o %;-
425 l
and Gravel
'x *g ;
, O 60 10 ) METERS 4
420 0
250 FEET Vertical Exa90eration X 10 Grid j Uniforms o e o e o e o e o e o e o e o
e o e o-t o a spacing l Variable s o e o
s o e o
I o
e o e o e o e o eofo1 oeoa Figure 21.--Simulated ground-vater levels near a seepage face based on
{
variable-grid and unifom-grid spacing. (Location of section ahoun ' on fig. 18. )
Transient-State Simulations Transient-state simulations of the ground-wa ter-level response to varia-tions in monthly recharge and evapo transpiration we re used to estimate the specific yield of the surficial sand and gravel and to verif y the hydrologic values obtained f rom steady-state simulations.
Transient-state simulations represented an annual climatic cycle beginning in October 1982, a period when water levels were close to the mean levels used to calibrate steady-state s imulat ions. The hydraulic-head distribution computed by the calibrated steady-state model was the initial condition used in the transient-state simu-lation. The transient simulation was divided into 12 one-month periods for which average monthly value s of evapot transpiration, recharge, and underflow we re specified. Each monthly period was simulated as four time steps to I
dampen the e f fect of abrupt changes in recharge f rom month to month.
A 2 year period was simulated by repeating the annual cycle to minimize the ef feet of error in the initial water levels specified in the transient-state model.
l 39 i
Calibration Ground-water levels ' computed by transient-state simulations were compared' with average monthly levels recorded in wells 80-1 through 80-8 f rom October 1982 through Septenber 1983, and : computed ground-water discharges to streams were compared with base-flow hydrography recorded at stations NP1 and NP3.
The sensitivity of predicted water levels to changes in storage was tested
- through a range of specific yield values, and the sensitivity of water levels to monthlyJrecharge was tested through the two sets of monthly recharge rates give n in figur e ' 13.
Underflow; through the upland boundary' of.the modeled area was adjusted by a factor proportional to the recharge assumed to occur in each month.
I Sensitivity Wa ter levels conputed by the transient-state model were sensitive to monthly recharge rates and less sensitive to s pecific yield.
Comparison of computed with observed water levels at wells 80-3 through 80-6 (fig. 22) shows that the computed water levels follow the observed seasonal fluctuations but are of le sse r magnitude.
The disc repancy be twe en canputed and observed water levels can largely be attributed to error in the timing and snount of recharge specified in the transient-state model; water levels were also af fected by the value of specific yield (fig. 23). ' Die lower value of specific yield (0.10) produced a close approximation of the large water-level fluctuations recorded in wells 80-4, 80-5, and 80-6, whereas a specific yield of 0. 20 produced a 4
better match ' for the other wells.
Computed base flow at stations NP1 and NP3 was sensitive to drain con-j, ductance (fig. 24). Ground-wat er discharges pr edicted f rom average d rain conductances obtained through steady-state simulations did not correspond closely to the estimated base flow at these two stations.
The discrepancies I
between computed and observed discharges can largely be attributed to seasonal changes in drain conductance. As shown in equation 6 and figure 15, drain conductance is directly proportional to the average cross-sectional flow area entering the drain and therefore to the height of the seepage face on the char nel bank.
Because the saturated thickness of the surficial gravel d ratning to the stream. channel varies throughout the year, the conductance of the channels can be assumed to vary as well.
1 In a separate transient-state simulation, drain-conductance values for each month were adjusted by a factor proportional to the saturated thickness of adjacent grid cella computed in a previous simulation that assumed constant drain conductance.
The discharges at stations NP-1 and NP-3 predicted from f
variable drain conductances were closer to the observed base-flow values and did not substantially alter the resulting water-level hyd rographs (fig. 24).
l Resulte Hydrography of measured and simulated ground-wa ter levels at wells 80-1 l
through 80-8 are presented in figure 25; measured and simulated ground-water discharges to channels above stations NP-1 and NP-3 are plotted in figure 24.
Model predictions were obtained from the set of optimum hydrologic values listed in table 4, the set of monthly recharge rates calculated f rom equations 3 and 4, and values of specific yield ranging f r om 0.10 t o 0. 20.
40
i i
i i
e i
i i
i WELL 80-3 0.5 4
+
X 5
+
X x
0.0 - X
+
X
+
h X
+
X
+
X
-0.5 X
i I
i i
g g
g g
g g
i i
i i
e i
i i
i i
i WELL 804 05 4
X m
+
00 - X X
2 X
+
X 2
X X
j h
-0.5 OBSERVED WATER LEVEL X
X LEVEL CALCULATED BY SOIL-g MOISTURE DEFICIT METHOD r<
-1.0
+
LEVEL CALCULATED BY SOIL-3: -
MOISTURE MODEL (Steenhuis,1983)
_J y
i i
i i
i i
i i
e E_
e i
i e
i i
e i
i i
WE180-5 2
0.5 b
+
+
e x
2 x
+
+
0.0 - x
+
x X
6 x
+
+
o x
+
4 x
g X
-0,5 i
i i
. m i
i i
i i
i WELL 80-8 0.5
+
+
X 4
+
X 0.0 - X X
X
+
X X
X
+
x
-05 AbG SEP OCT NOV DEC JAN FB MAR APR MAY JUN JUL Figure 22.--Observed ground-vater levels in four cells in relation to seasonal recharge computed by soil-moisture deficit method and by soil-moisture model of Steenhuis and others (1983).
41
t r>
q.
)
1 g
aj
'd 7]
t q
to a
i i
e i
i.
~ WELL 80 0.5
.o O'
X L X,
h
.y60.0
'X.
.O y
Z O
X
)
O M-O g
o
-0.5 O
i e
i s-i i
e i
i
,g g 1.0 i.
j WELL 80-4
)
0.5 l
O g
P o
-X L5 ' 00 x
g-x.
6 C'
o X
- (
X 0
0 X
0 0.5 OBSERVED WATER LEVEL 0
0:
- SPECIFIC YlELD =0.10
-o' X:
SPECIFIC YlELD =0.20 i
H
-1.0 l
i i
e i
i i
.OCT-NOV DEC-JAN FEB.
^ Figure 23.--Observed departures of ground-water levels from initial' Levet at uette 80-3 and 80-4 in relation to simulated values computed fran tuo magnitudes of specific yield.
l The well hydrograpv generated by the transient-state model (fig. 25) closely correspond to recorded water levels.
Simulated ground-water discharges follow the annual pattern of high flow in the winter and spring and j
declining flow through the summer.
Flue tuations of both simulated ground-l
' water levels and discharges are less extreme than those observed, mainly l
' because of the timing, volume, and distribution of recharge.
Error in esti-l mating the lateral distribution of hydraulic conductivity and specific yield
- H
_ would also cause simulated values to deviate f rom observed values.
Discrepancies between computed and observed discharge can be partly explained;by error in the recharge specified in the model; other sources of error could be seasonal f actors that af fect drain conductance, such as (1) the lengthening of drainage channels during we t pe riods, and (2) reduc ed perme-ability of the interf ace between aquif er and d rain surf ace during dry periods.
42
)
i
50 SfTE NP 1 O
O g
O o
M O
O 0
0 0
x x
20 x
x x
x x
x x
x x
E O 10 f5 a.
y 0.0
.e u
srrE NP 3
$ 200 u
E 5
cc 0
150
\\0 s
0 b
0 O
x X
x X
0 x
x x
O X
100 OBSERVED DISCHARGE O
DISCHARGE CALCULATED WITH O
g CONSTANT DRAIN CONDUCTANCE DISCHARGt CALCULATED WITH O
O VARIABLE D9AIN CONDUCTANCE O
50 - i i
i e
i i
i i
i OCT NOV DEC JAN FEB MAR APR MAY JUN JUL AUG SEP aites NP-1 and NP-3 in Figure 24.--Measured ground-vater discharges at relation to discharges simulated from constant and variable drain-conductance values.
43
.u R
i e
n 0.5 Ma pi X
0.0 X
X X
-06 i
X 5
e i
i 0.5 X
Ma B2 X
h 0.0 X
X X
t N
2 X
E d
{ -0 5 x-x x
l J
I I
i i
t I
t t
t I
I t
ccw r-i i
i
(
WELL 80-3 y _ 0.6 b
X l
Z 2
X X
X X
0.0 X
X 8-
/
x x
p X
< 0.5 x
g-e a
i t
a 1
i t
i I
I I
gjo WELL 80 4 06 X
X 0.0 - x
~
x X
X X
X X
X X
~
OBSERVED VALUE X
S!MULATED VALUE
-1.0 l
t t
I i
1 a
i 1
a i
e f
[
OCT NOV OEC JAN FEB MAR APR MAY JUN JUL AUG SEP t
Figure 25.--Measural and sirrutated oater Levels 44
WELL 80-5,
0.5 X
x
~
0.0 - x x
x x
X x
x X
X
~
-0.5 X
i i
i i
WELL 80-6,
0.5 w
x x
E x
x x
~
0.0 - X z
x cf x
3 x
x y.05
- / -
<D g
(
ELL 80-7 _
- 05 x
x t
z__
x
.s x
x
~
0.0 zo x
/
P
<t x
5 b -05 4
~
WELL 80-8 05 x
x x
x x
~
0.0 - X X
X x
X OBSERVED VALUE
~
-0.5 x
SIMULATED VALUE
-LO OCT NV DEC JAN FEB MAR APR MAY JUN JUL AUG SEP in eight vella.
(Well locatione are chasn on pl.1.)
45
MODEL APPLICATION Tritiated water could infilt rate into the surficial sand and gravel on the north pla teau fem leaking storage facilities or accidental spills during trans port of liq uid materials. The ground-wa ter flow model can be used to predict the flow path and velocity of this wa ter, although it neglects the
{
ef fects of radioactive decay and of mixing during transport.
St eady-s ta te j
simulations calcula ted fim paths and traveltimes of water f rom two potential sources of tritium in ' the reprocessing f acility--the main plant building, including the f uel-storage pool, and the high-level liquid-waste-tank cmplex.
l Past migration of tritiated water f rom the low-level waste-treatment system j
was also simulated, and the results were empared with tritita concentrations l
measured in 1974.
l Ground-Water Movement Ground-water flow paths and velocities were calculated from the volu-me tric flux across cell boundaries computed by the steady-state model f or each grid. cell. The volumetric flux is related to the average linear velocity, V
'(m/d), by:
_v = g_
nA (8) 3 where:
Q = volumetric flux, m /d, n = rorosity of the saturated material, dimensionless, and i
2 A = average cross-sectional flow area, m,
Y is de fined as the ratio between the traveled distance of a ground-water tracer and its time of travel (Freeze and Cherry, 1979,
- p. 7 9). The average cross-sectional flow area of a grid cell in the model was assumed to be the product of the saturated thickness of the cell and the grid spacing.
The t
l porosity of the sur ficial gravel wa s assumed eq ual to values of specific yield obtained in transient-state simulations. The average linear velocity calcu-lated for each grid cell was plotted as a vector to indicate the rate and direction of ground-water flow.
t Ground-water flow paths through the north plateau as predicted by the steady-state model are plotted in figure 26.
Mo st of the ground water entering the north plateau through the upland (southwestern) boundary is diverted by building foundations and backfill associated with the main plant f
f acility to seepage f aces above the tributary to Franks Creek.
Ground water flowing thro 6gh the northwestern part of the plateau discharges to the NP-1 channel and the wetland above the NP-3 channel.
j h
Predic ted gr ound-water flow paths f rom the main plant to discharge points j
along the east boundary of the plateau are slown in figure 27, which also j
l; indicates t ravelt imes f rom the f uel-storage pool and the high-level liquid-1 waste-tank emple x t o di scharge points. The final destination of a slug of
)
water t raveling t hrough the surficial sand and gr avel will depe nd upo n where l
it is introduced. The model predicted that the NP-3 channel and the f rench drain will intercept most of the flow downgradient f rom these two potential sources and that the remainder will be discharged at seepage faces along the j
cast boundary of the plateau. Water f rom the main plant area would reach the i
i NP-3 channel and f rench d rain within 500 days and would arrive at the seepage 46
f aces within 800 days. The area of high permeability northeast of the main plant building (fig. 11) significantly decreased the traveltime of water flowing to the NP-3 channel.
Doubling the model hydraulic conductivity did not significantly alter these traveltime predictions, mainly because it created a lower hydraulic gra-dient in the area of higher permeability, which leaves the ground-water veloc-ity unchanged.
Decreasing hyd raulic conductivity by 50 percent increased trav-eltimes f rom the main plant area to the NP-3 channel f rom 500 days to 800 days.
Analysis of Past Tritium Migration Before the detection of tritium in ground water in 1972, lagoons 4 and 5 were leaking, and the channel above station NP-3 had not been construct ed.
These conditions were incorporated into the model by draining the model wet-land through the former channel above station NP-2 and simulating infiltration f rom the lagoons as leakage through a confining laye r.
The amount of leakage predicted by the model was 0.6 L/s, or nearly 50 percent of the total esti-mated volume of wastewater processed (P. Burn, West Valley Nuclear Service Ce nt e r, oral c ommun., 1984). Although this leakage value is unrealistically l a rge, it represents a " worst case" f rom which to interpret model results.
Steady-state ground-water flow paths in 1972 were predicted f rom the opti-mum hydraulic values shown in table 4.
Because annual precipitation in 1972 (117 cm) was greater than the annual rate of 92 cm/yr used to estimate recharge for the steady-state calibration, it is likely that the recharge rate in 1972 was also greater than ansumed. The increased recharge in steady-state simula-tions did not have a significant ef fect on ground-water flow paths through the north plateau, howeve r.
The 1972 ground-water flow paths predicted by steady-state simulation are plotted in figure 28 with tritium concentrations in ground-water samples collected in 1974. The flow lines indicate that leakage from lagoons 1, 4,
and 5 moves to the f rench drain and seepage f aces along the eastern boundary of the plateau and do not indicate migration of tritium f rom lagoons 4 and 5 to the wetland, even at the higher simulated rate of leakage.
Tritium data f rom 1978 (fig. 6) indicate that tritium levels in ground water west of lagoons 4 and 5 and in the wetland declined af ter lagoons 4 a nd 5 were sealed in 1972.
If this decline was caused by the sealing of the lagoons, some mechanism such as dispersion or flow through a horied conduit is needed to explain the lateral migration across the prevailing hydraulic gradient. However, tritium concentrations between the hardstand and the main pl ant buildLng remained elevated in 1978, which suggests other sources. The measured water leve ls (f ig. 10) and estimated distribution of hydraulic con-duct ivity ( fig. 11) indicate that most of the ground water discharging to the wetland would follow the buried channel on the till surf ace northwa rd f rom the reprocessing plant.
Reprocessing act ivities we re halted in 19 72 ; t hus, the tritium sources responsible for contamination of the wetland could have been along this flow path and removed since then, which would explain the general decline in t ritium concentrations.
The pot ential s our ces me nt ioned earli er--
storage of tritiated material on the hardstand, leakage beneath the reproc-essing plant, and fa11mit f rom the ventilation stack, are likely sources of the t ritium in the we tland.
47
iff H
'l 78'39 15'-
78'39 00" N
.J i
i
-- + GROUND-WATER FLOW PATH.
gf PREDICTED BY STEADY-STATE
-Cr SIMULATION 3,.
' A NP-2 STREAMFLOW GAGING STATIONf
/"
EXTENT OF SAND AND
\\
GRAVEL DEPOSITS -
.,c fx 1
1 q.1 s
4"%-a/,s' e
x h
i oy-
[l'; \\
,///
( 4 4. L - - - %
g J WETLAND Q-y#
\\
/
i l
I
- f. /
/.
l l ~.
.j NO
/
l
,$l, # HARDS i.. q
. /.
l
/
/
FRENCH l
A
/
DRAIN "~ R i
\\
,A
,Y R
l
\\_,
/
p
/'^h4
'{
/
/
/
p
/
v /
[
\\'
- /
/ h-HIGHAEVEL -
^
^cv" '
'N
/ '
j
/.
WA STORE %E -
/
\\ p/
.[i
./
K COMP) TEX
/
/
\\1,.,.
/
y l
4 yl
/,
- ~ ~
/
a, kN /
~~'%'
._m/
',,.A ITE7,
w STATE UCENSED l'
N
__+
WASTE '
g = =RE.PROCES$l DISPOSAL AREA gr I
- c p/~.~% %,~~_
s
~~, % j p,J.
p 426 1
g:
M' x
q
"~
W $e"ek
?
~ ~ - -
C m
-r
\\
f,.
FACluTY'S
}
h
"(rg/
OfSPOSAL AREA l
l 0
60 100 150 METE RS M
-j 1
D.
250 f,3) F EET l
i i
y df [d. k w h
[
4 f-l Figure 26.--Ground-vater flou pathe through sand and gravel on
(
north plateau ao predicted by steady-state modet.
I 48
78539'15" 78 39 Orr
-- + GROUND-WATER FLOW PATH
- 400 - EXTENT OF GROUND-WATER TRAVEL ~ -
f AFTER INDICATED NUMBER OF DAYS
'& NP 2 - STREAMFLOW-GAGING STATION f*%,
.\\ s Dfr EXTENT OF SAND AND
.'\\
. GRAVEL DEPOSITS X
Ms
\\
x x
.K 4p
~-
g i.
N%
2r
/
15" k
f/
N'd N P.1 p,-.gi - - P N
_j[
q s
/gf
^
N
, }
(M kf WET ANO f/
n l
L l
400
/
/ /
/
?
[
p!
4
/ /
/ [
~k HARDS 7A) /
'%.)
/ /, /,?
/ /
pggggy y
/
DRAIN - -t M
/
300
- oo y
,/ f,f'
/
^
y
,,,,},, p,/ '
2 f
/
.p
,\\
'{
K/
/
f
,. 3 g
Tih / 'l \\GO, 5
\\.
/'
'j
.-(2/ /f I l')
,e/\\\\
'q
/
/
' F UEL g [ STATE UCENM D e
W AUTE '
REPROCES$tNG.
'/
J DISPOSAL. APF A -
1 y
Pt. ANT
/. I O/
k y/
/
y l
/
i 00" A.
s
/ Lagoon Road 9
/
cre
.e L
yg i
\\
FACllITY S
,,,,,,,,,g d.
oispoSAL AREA y
\\
\\
',r/g
/
\\
\\
/
\\
o sp 100 iso verERS 6
2so s6c rrET
\\/
s
\\
i
_..-.__._.._._..__._._t..__
Base from u s Geaksical Surve)
AshfLvd Hodow 1979 1.24 000 7
1 Figure 27.--Flou paths and travettimes of oater from too potential sources of tritiwt and contamination as predicted by steady-state model.
I 1
1 49 i
1
78 3915' 78"39 00" N'8 '
e400 OBSERVATION WELL-Numt r is tritium
/
concentration of ground water in l
1972 samples, in pCi/L j/
/
l
{
% 200 GROUNDWATER DISCHARGE POINT--
l
\\
Number indicates tritium concentration,in pCi/L N,Ye
- - -oi.-
GROUNDWATER FLOW PATH
,,S
'Wf EXTENT OF SAND AND g
\\
STREAMFLOW. GAGING STATION l M-Mb'*!
A NP-2 GRAVEL DEPOSITS l
N 4p
's.
)
p
,s ty
,/,
n w#
ovna N P-2
, #f,JA j
/
\\
+g{_Y,j__-A A
NP1 N /
I
/
- f,~
$l
/r l
e hf' \\
/',f/.
(h k JWETLfNil l
/ x
/
b l l /
NP.3 h
/
/
\\
\\
- Q
,/
/,
[
/
HAROSff[NL/
///
-7 HitH LEVEL RADIOACTIVE
/
- //
,e"
/ j WASTE, STORAGE- (
j// II /
FRENCH
~ ' '
\\(
^ TANK COMPLEX 4
j
/J-9 / 300l l,
,. DRAIN j/
\\( g',Q','E $if.'.&'\\/ /g^'
?
}
/
/!
- GC I4b
/
00
- 160
~?f'/j'fl_.
l '>
\\
e, ','
,/
\\'
/ //
/
6 s
u
/
G h
\\
l f'
k\\
/
/'
~D i
' i "/,/,7ml,,,'t u
[*
')STArf UcFNsto j
- 4. @ E L 1
^^h il REPROCESSING 1
t Pu NT
/
l'3bPOS^l AN U
(
,0
/
Ny
\\
\\
,/
/
/,-
\\
l
~
\\
/ g x
\\
\\
,f' ~5 l
L# ~L_y
\\A)\\/\\
\\
\\ s, * ~,,,',,,,,,,,,,'"+i,,#',e**p/,/
.,' Disl'05 AL ARE A XN
.A
\\
/
\\
\\
t Acitars f 4 'f.
6
\\
(
f-y g
..g,
\\
,/
~
g 0
59 100 150 ME TE R5
\\\\
l
/
\\
H N
\\\\ \\
s
/. -
s o
250 woer
\\\\'
s j
L..
,_\\1 i
n ~ e.e. m 3 o.m,
... s... w,
khhi@ d %>lhm i 4 /1 1 24 (Yu )
Figure 23.--1972 ground-uater flou pathe simulatal ly steady-state model and tritium concentrations in ground-water samples collected in 1974.
50
SUMMARY
AND CONCLUSIONS A 3 year study was conducted f rom 1980 through 1983 at the Western New York Nuclear Se rvice Ce nter near West Valley, N.Y., to investigate the hydro-geology and ground water flow near the former nuclear-f uel-reprocessing plant and its f acilities. Radioactive materials are stored within the reprocessing plant and in burial grounds at the site.
Tritiated water is stored in a lagoon system near the plant and released under pennit to a nearby stream channel.
A two-dimensional finite-dif ference model wa s developed to simulate ground-water flow in the surficial sand and gravel deposit underlying the reprocessing plant.
The 41.4-ha deposit overlies a till plateau that abuts an upland area of silt stone and shale on its southwest side and is bounded on the other three sides by deeply incised stream channels.
The channels d rain to Buttermilk Creek, a tributary of Ca ttaraugus Creek, which drains to Lake Erie.
The ground-wa ter flow model provides a reasonably accurate simulation of lateral ground-wa ter flow through the surficial sand and gravel on the north plateau.
Steady-state simulations closely matched the average of ground-water levels recorded in 23 observation wells and ground-wa ter discharges measured at gaging stations NP-1 and NP-3 and the french d rain.
Transient-state simu-la t ions indicated that the model could reproduce seasonal changes in ground-water levels by having the recharge and evapotranspiration rates varied monthly. Transient-state simulation of ground-water discharges matched observed base-flow hydrography most closely when the values of drain conductance were varied monthly.
Model-generated ground-water levels and discharges were found to be sen-s it ive to the values of recharge and drain conductance.
Er ror in the cali-brated values of these terms would affect the ground-water budget predicted by the model or the simulated volume of ground water flowing through the system.
However, model pr edictions of g round-water flow paths and velocities were relatively insensitive to these terms.
Co nc lu sions f rom the model analysis can be sanmarized as follows:
1.
Model simulat ions indicated that most g round wa ter flowing from po tential sources of tritium near the main plant building would discharge into the s tream channel above station NP-3 or the f rench d rain.
Some ground water originating near the main plant would discharge through seepage faces along the southeast border of the plateau.
2.
The estima ted t raveltime of ground wa ter f ran the main plant building to the closest discharge point, calculated f rom hydraulic conduc tivity and porosity (assuned equal to specific yield) values obtained through model calibration, was 500 days.
Doubling the hydraulic conductivity did not significantly alter this estimate, but decreasing hyd raulic conductivity by 50 percent increased the traveltime to 800 days.
3.
The model was used to investigate possible sources of the tritiun that was f ound in the we tland on the north plateau in 1972.
Although the tritium levels declined af ter wastewater lagoons 4 and 5 were sealed in 1974, ground-wa ter flow paths predicted by steady-state simulations did not 51
^
j indicate 1the lagoons to-be a probable source. of the tritium.
Possible reasons f or the migration 'of tritium from the Llagoons to the wetland. could j
include.other transport processes, dispersion, or. buried conduits. : Al ter-1
- natively,' the1 tritium ;could have infiltrated to the wetland f rom other -
. sources.
- 14. ~ Calibrated model values of hydraulic conductivity,of the surficial sand' and' gravel ranged from 0.6 ' to 10.0 m/d.
The highest values were applied;tosan
~
' area northeast of the main plant building that overlies n' buried stream
~
channel' incised =in the surface of the till. This high permeability sig :
nificantly decreased traveltime.of ground. water from the main. plant-building to the NP-3 channel.
5.
More than 75 percent of' the ground water on the north plateau (46 cm/yr) is.
l
- derived from precipitation.. Underflow from the fractured bedrock along the.
upland (south) boundary of the plateau and. leakage f rom the outf all channel
- f rom the main plant building into the sand and gravel account for the 3
remainder. Evapotranspiration from the North Plateau totals about 20' cm/yr.
Ground-water ' discharge through seepage faces along the periphery :
of: the plateau totals 3.0 L/s, and discharge to the NP-3 channel totals 1. 3.
L/s.. Discharges to the NP-l~ channel,ithe french drain, and the low-level
- radioactive wastewater-treatment. system account for the' remaining 0.8 L/s.
l i
REFERENCES CITED '
l Alba nes e, 'J. R., Anderson, S. L., Dunne, L. A.,- and Weir, B.'A.,'1983, Geologic and hydrologic research at the Western New York Nuclear Service Center, West Valley, New York, til U. S. ' Nuclear Regulatory Commission,
. Annuali Report,' August 1981-July 1982:
NUREG/CR-3 207, 397 p.
Bergeron, M. P.,1985, Records of wells, test borings, and geologic sections near West Valley, New York:
U.S. Geological Survey Open-File Report 83-682, 95 p.
Bergeron, M. P..and Bugliosi, E. F., Ground-water flow near two radioactive-waste disposal areas at the Western New York Nuclear Service Center, Cattaraugus County, New York--results of flow simulation:
U.S. Geological Survey. Water-Resources Investigations Report 86-4351 (in press).
Be rge ron, M. P., Ka ppel, W. M., and Yager, R. M.,1987, Geohydrologic condi-tions at the nuclear-f uels reprocessing plant and waste-management if acilities at the Western New York Nuclear Service Center, Cattaraugus County, New York:
U.S. Geological Survey Water-Re sources Investigations.
' Re po r t 8 5-414 5, 4 9 p.
Co oper, H. H.
Jr., Br edehoef t, J. D., and Pa padopulos, S. S., 19 67, Re spons e of a finite-diameter well to an instantaneous charge of water: Water
- Re sources Resea rch, v. 3, p. 262-269.
Draper, N. R., and Smith, H. H., 19 81, Applied regression analysis, 2nd ed. :
New York, John Wiley, 709 p.
52 l
3
REFERENCES CITED (Continued)
Fr eeze, R. A., and Ch er ry, J. A., 19 79, Gr oundwa te r: Englewood Clif f s, N.J.,
Prentice-Hall, Inc., 604 p.
Gray, D. M.,1970, Principles of hydrology:
Canadian National Committee for the International Hydrological Decade, Ottawa, Canada, 6 35 p.
Kappel, W. M. and liarding, W. E.,19 87, Sur face-water hyd rology of the Western New York Nuclear Service Ce nter, Cattaraugus County, New York:
U. S.
Geological Survey Wa ter-Re sources Inve stigations Re port 8 5-4 309, 36 p.
LaFle ur, R. G.,1979, Glacial geology and stratigraphy of western New Yo rk Nuclear Service Center and vicinity, Cattaraugus and Erie Counties, New Yo rk:
U.S. Geological Survey Open-File Report 7 9-989, 17 p.
Mcdonald, M. G., and Ib rbaugh, A. W.,19 84, A modular three-dimensional finite-dif ference gr ound-water flow model:
U. S. Geological Survey Open-File Repor t 8 3-875, 528 p.
New York St ate De partment of Environmental Conservation, 1975, Annual report of environmental radiation in New York:
Al bany, N.Y., 51 p.
Prudic, D. E.,1981, Computer simulation of ground-wa ter flow at a commercial radioactive-wa ste landfill near West Valley, Ca ttaraugu s Coun ty, N.Y., in Li ttle, C. A., and St ratton, L. E., (eds. ), Modeling and low-level wa ste manageme nt -an interage ncy workshop:
Oak Rid ge, Te nn., Oak Rid ge Na tional Laborato ry, OR0821, p. 215-2 48.
1986, Gr ound-water hyd rologic and subs ur f ace mig ration of radionuclides at a commercial radioactive-waste burial site, West Valley, Ca ttaraugu s County, New York:
U.S. Geological Survey Pr of es sional Paper 1325, 83 p.
St eenhuis, T. S., Muck, R. E., and Walt er, M. F., 1983, Prediction of water budgets with or without a hardpan, jy1 Pr oceedings fr an Conference on advances in infiltration:
American So ciety of Agricult ural Engineers,
monog raph series, December 1983, p.
1-13.
Todd, D. K.,1980, Gr ound wa ter hyd rology:
New Yo rk, Jo hn Wiley, 53 5 p.
U.S. Department of Energy,1979, Western New York Nuclear Se rvice Ce nter Study:
Companion Repo rt TID-28905-2, 500 p.
Ve11eman, P. F., a nd Hoaglin, D.C., 1981, Appli ca tions, ba sics, a nd computing of exploratory data analysis:
Boston, Ma ss., Dux bury Pr ess, 3 54 p.
Winter, T. C.,1981, Uncertainties in estimating the wa ter balance of lakes:
Wa ter Resour ces Bulletin, v. 17, no.
1, p. 8 2-115.
t 53
t i
A.
APPENDIX ESTIMATION OF HYDRAULIC CONDUCTIVITY
, Estimates olf. hhd raulic conductivity of the. surficial san'd and gravel' were
~ derived by. applying the method of Cooperzand others;(1967) to -the"results ofi 1
slugLtests"done :at ~ eight observation wells. - The ' application of the Cooper:
3 method is ' described:in Berge ron and others (1987), and values of hydraulic'
'conduc tivity 'obtained f rom the slug-test da ta are summarized in table A-1 r
.The Cooper. method assumes that an instantaneous slug.of waterLis added to a well t hat fully. penetrates a ' confined, wa ter-bearing layer and that the flow f rom 'the well is therefore horizontal with~ no vertical component.
The geome tric'mean hydraulic-conductivity value of 0.6 m/d given in table A-1 served only as a starting point;in the' development of the model because 1
these assumptions were not fully met. on the north plateau. First, ground water in the sand-and gravel is unconfined; t hus the saturated thickness near '
the well changed during slug tests, which gave rise' to vertical. flow compo-
. nents. The assunptionsLwere 'more valid for wells in which the saturated' thickness was large relative ~ to the initial increase in water level whereby the variation-in saturated' thickness was relatively smaller.
The water level in the well dropped 'substantially before the J first water-
- level measurement 'could be made because fof the rapid movement of wa ter. from,
the well. to the' g ravel, lThis problem was. partly mitigated by the small diameter. (5 Jcm)-of the observation wells, but initial wa ter levels were extrapolated f rom the data for use by"theLCooper method.
An' error of L25 per-cent.in the initial water-level estimate resulted in an error of more than 40 percent in the calculated hydraulic conductivity.
Table A-1.--Hydraulic-corduc tivity values for sand and'
' gmoel on the north plateau as determined fvan elug-test data by the method of Cooper and othere (1967).
(Well locations are shown on pl. 1.]
Initial Hyd raulic Sa tur ated increase Well c onduc tivi ty thickness in water number (m/d)
(m) level (m) 80-1 2.5 4.5
- 1. 0 80-2
.2
- 2. 5 1.1 80-3 7.9
- 1. 0
- 0. 7 80-4
.2
- 1. 6
- 1. 2 80-5
.2
- 3. 3 1.1 80-6
.1
.8
- 1. 7
)
80-7
.4
.6 1.2 l
80-8 1.5
- 2. 8 2.0 1
Geometric mean 0.6 54
I l
The lateral variation in hyd raulic conductivity of the surficial sand and gravel was investigated at 24 observation wells that had been previously l
installed on the north plateau for ground-water sampling.
These wells were constructed with 15-cmedf ame ter perf orated pipe,.so rapid losses of water to the unsaturated zone during a slug test precluded t he application of the method of Cooper and others (1967) to determine hydraulic conductivity.
Therefore, relative estimates of hydraulic conductivity were calculated from pumping and slug-test data from these wells.
Data on well yields were obtained from records of the previous site oper-ator.
The yields were calculated f rom the maximum discharge that could be derived fran the wells af ter 30 minutes of pumping. These data (table A-2) were normalized by dividing the yield by the saturated thickness recorded at e ach well.
During slug tests performed in this study, 80 liters of water were introduced into each well, and the rise in water level was measured af ter 30 seconds.
(These data are also listed in table 3. ) Regression analysis of the normalized yield in relation to water-level rise indicated an inverse linear rela tionship between the two variables (F = 17.6 at the 90 percent confidence i nte rva l; the slope of the line was significantly dif fe rent f rom zero).
Howeve r, the regression eq untion was a poor predictor of the variance of nor-malized yield (r2 <0. 5; less than 50 percent of variance was explained by regression) (Draper and Smith, 19 81, p. 32).
i Examination of the surface elevation of the upper till unit near the main plant building (fig. A-1) indicates that a buried stream channel underlies l
part of the sand and gravel where many of the sanpling wells are located. The channel may be an erosional feature that marks the location of a former stream channel cut into the surf ace of the till pla teau.
To test the hypothesis that i
the buried channel may af fect the hyd raulic conductivity at the wells, the data in table-3 we re subjected to a box plot analysis.
Box plot analysis is a nonparametric statistical techniq ue used to determine whether groups of data a re significantly dif ferent f rom each other (Velleman and Hoaglin, 1981,
- p. 65).
Results of a box plot analysis are given in figure A-2, which uses three hyd raulic-conductivity categories f rom table A-2 to group the data.
The plots indicate that the distribution of data from wells over the buried channel (group 2) is significantly dif ferent from that for wella south of the channel (group 1), as evidenced by the lack of overlap in the data.
This indicates that the hyd raulic conduct ivity of the sand and gravel overlying the buried channel is higher than that of the gravel east of the channel at a signifi-cance level of 5 percent.
The sand and gravel west of the channel (group 3) is also significantly dif fe rent fran the other two areas and has an inter-mediate hydraulic conductivity value.
These findings were incorpo rated into the model and are discussed in the section on the results of steady-state s imula tio ns.
55
4-l 78'39'15'
- 7843poo-
- 1380- ALTITUDE'OF TILL SURFACE--Contour A'
'y intervnl 2 feet. Datum is NGVD of 1929.9-
/'
L
~
OBSERVATION WELLS: GROUPED ACCORDING TO
. LOCATION RELATIVE TO,/
f 1
, GROUP 1 - e' CHANNEL ON TILL SURFACE- /"a==.
\\
' GROUP 2 0.
f Number is local well number./l
\\
yD6
)
Group locations are given J GROUP 3 e in table A 2 footnote.
EXTENT OF SAND AND
.l ANP-2 STREAMFLOW-GAGING STATION, I r
\\
GRAVEL DEPOSITS:
s.
f%'
420
/
Ns.
I 27'
/
15"
,M
'~, f o,u,s.
N P-2
-P I
N P-1
\\
_.g K
_.4
,/
j h
j j
N lp1 J
/(p.s pf s
/
/.K \\
- &.. NP-3 i
\\
\\
1 v-
/
s'E '.' ; WETLAND
'N i
Q I
/,/
-K
's;-et t i-q
/
s,
/
-o_.-
) i p
j 1390
(
/ \\
{
1370.
6 i
F.
ni' $6?
'l
{ '
l, '
d 1
1 FRENCH !
~A
^-
n
.i s
e b
DRAIN 'i /
"A
- )
Of[
i ir' Jh
/
f J-1 0 '-
J-2 J-6 0
/ e
%[/
1 W
% N-
\\
g ?*
^9 l
3
?/'
~
J-4
.\\
)
' N,2 & Ji 't' 4 e#s v,9
- )</' #;Y (,
N_._j
,-'p
%dQ
(
M3 pB E'v
)
g.
p
RI b
/
j F
122
{
\\
a k
HIGH-LEVEL RADIOACTlyE
+
I
{
\\ WAST 0 STORAGE-TANK COMPLEX
~
,7~~**41d-~
\\
M 390UEL STAlf UCENSF 0
[
WASTC REPROCES$1NG
,'D&OSA',AMA PLANT
/
j T\\
/
/
l
/\\
/
/
/
/
.g w
/
i
/
\\
~
yl s
4 s
A
\\
- i m,,:ady s
c, x
n
~ g#
/
\\
j I
%'N M" -,,%,j3.s,/,f oiscosAL AR[ A
- ')
i p
- Aciurrs y,p..
(/
\\ \\(+
j
\\
-\\e x
/
s
)
\\
/
\\
o so un teo urras
,i
\\
g N
f' o
250 wo r eer sy,/
\\
l
_.___._L..._
eem,. o s ce%u,i swv.w Ashi.,J Hollow 19 79 f.74 0r.X)
Figure A-1.--Till-surface altitude near main p'lant and location of buried strean channel.
56
h ;
'r g
' f. :
" Table A-2.--Relative measures of h draulic conductivity developM from.
}
b pwtping ad slug-test data fran-15-cn-diameter cette, f
[ Locations of wells: and buried channel are shown in fig. ' A-1. )
1 1
E
. Hyd raulic g
Water-level rise. conductivity -
H 30' seconds af ter '
'g rou p '
Well yield ' normalized.
. addition of'30 used inL by' saturated" thickness' liters of' water statistical.
- j Well
-(L/s)/m-(m)
- analysis 1
'.002 1.32 1
B-E
.0 04-'
2.01~
1 F
.010 1.50 1
l J-l '
.186
.66:
31
- 4-1 JJ-2
.0 64
.65'
.3' J 4~
.031'
.76 3
k LJ-6
.079 1.15 3
l J-7'
.052
.73' 2.
J-8
.110
.44 2
I L
J-9
.10? -
.81 1
l l
K
. 010 1.43
'l L-2
. 010' 1.40 1
.L-3
.010 2.~ 11 1
- 1 L-4'
>.103
.83 2
- M L-5;
. 058
.29 2.g jf I
<a
'M-1
>.145
.79 2
M-2
. ~>. 0 7 4
.16 2
M-3
>. 0 93
.45' 2
M-5:
.116
.34 2
1 N-2
.006
.87
'3 f
N-3
.037 1.64
'3 P
.103
.21 2'
Q
.103
.73 1,
R
.010 1.35 1
1 Hyd raulic-Re lat tve conductivity hyd raulic.
Number of group Location
_ conductivity observations
/
1 East of buried channel low 9
(
2 Over buried channel Hig h 9
3 We st of buried channel Intermediate 6
t'l 57
.t
.}'
- $ r?
>\\
HYDRAUUC
' l
.(
,.)
's,
r-4 ) ', -
CONDUCTIVITY i'
AM (4
RANGE
- .n g ',
<J i
t('
,,/
). - -
{[
f HIGH
+
\\
l
')
-(
)------
INTERMEDIATE
+
./-}
s 09 1.0 1.5 2.0 WATEAUAEL RISE,lN METERS r
i i
j e
-(+ )
M,2
/f l 1
LOW y.
//
o
.. j ;
I
'e
\\
.t r
HIGH
(
+
)- +
d--
4'
/
INTERMEDIATE *---(
)
+
(
i
\\
\\
(,,,
,9 0.0 0 005 0.010 0.016 WEU.YlELD,IN LITERS PER SECOND PER METER Q
EXPLAN ATIO N
+
MEDIAN VALUE
--- RANGE OF OBSERVAT ONS OUTSID v 95%
(
,)
APPROXIMATE 19% i CONFl'pNCE INTERVAL CONFIDENCE INTEHVAL ABOUT MEDIAN N
OUTLIER OU'lSiDF 39%
CONFIDENCE INTErWAL Figure A-2.--Recutte of box-plot anatyais ittuetrating d1lferencea in hydraulic conductivity at observation vette grouped as hi.gh, mediu, or too acconiing to the!.r location in relation to the buried channel, f
c l'.S. GOVEftNMENT Ph1NTING OTTICE: 1987-712-715 t7 58
- g
[I;'
^ '!
y ' {}.
> $ llf*[te ' _
-- y x
4,
- l.
4.j,
- t
-y OVERSIZE
%^
c o
HDOClqMENTs@4 G R_AGEPULL_ED af nM e*_
. Q!
./.6i 1
AFIDs}4
.-v
, t s q) -
.c
.j SEE AP$RTUREC'r-WL
- h VMT'"-
,m Yg jl;;k zi
- s
.gr
.[-
/.(
.. NUMBER OF OVERCt2ELPAGES FIL.MED ON,,APEF(((9,RSCARDS l
- e
]
4. A f.
Jl L
_ i fL y
.?;
q" M) e '!,
[i.
]
' (
[jy it 4 r-s.
_4 h,,
5M p., y
'x g
3
' ' cw 7
/
.c
< APERTURE CARD /HARD COPY AVAILABLE [ ROM RECORD $1RVICES. BRANCH A IFTS 492 = 8989
,e e,
\\
(
7,.,
.h
'T
~\\
,k[.1i '
jb>
.g. ;.
i Q
. s;,,-
t.
j (,e
',)
s
[
,\\
.s
,,'3
,c y
'E
)
l f l
./
('
i
- ' s 1
f
- qij'
/[i I
j
/>
i 1
2 z
g M
S.3GU kro Y
weN a
r t
t E
W e
h t
t k
e r
y o t Y i
ilc w a e F N g
n y ist s n e u c o eC rp su g
R u c
ter e e t
F t
eC n
e,r lc e ut n N
e eC h h t
r ic a
v e
r N eS wora lF e l
r c euN tr W
-dnuo rG fa no ita lu eu tS RE G
A Y
,