C'A /A center of excellence in earth sciences and engineering A Division of Southwest Research Institute' 6220 Culebra Road ° San Antonio, Texas, U.S.A. 78228-5166 (210) 522-5160 - Fax (210) 522-5155

April 2, 2002 Contract No. NRC-02-97-009 Account No. 20.01402.861

U.S. Nuclear Regulatory Commission ATTN: Mr. Neil M. Coleman Division of Waste Management Two White Flint North Mail Stop 7C 13 Washington, DC 20555

Subject: Transmittal 6f "Late Pleistocene and Holocene Groundwater Recharge from the Chloride Mass Balance Method and Chlorine-36 Data-Journal Paper" (IM 20.01402.861.240)

Dear Mr. Coleman:

This letter transmits Intermediate Milestone 20.01402.861.240 "Late Pleistocene and Holocene Groundwater Recharge from the Chloride Mass Balance Method and Chlorine-36 Data." This work was undertaken as an activity under the Unsaturated and Saturated Flow Under Isothermal Conditions Key Technical Issue. We propose to submit this paper for publication in the journal Water Resources Research.

Temporal variability of atmospheric chloride input, as well as annual precipitation, was considered in this study to estimate recharge rates for both Yucca Mountain, Nevada, and Black Mesa, Arizona, using the chloride mass balance method. The resulting average recharge estimates for Black Mesa are 9 mm/yr [0.35 in/yr] for Holocene and 35 mm/yr [1.38 in/yr] for late Pleistocene. Local recharge rates at Yucca Mountain were estimated from the Chlorine-36/Chlorine ratios and chloride concentrations in perched waters. Based on this approach, the estimated recharge for Yucca Mountain is 5 mm/yr (0.20 in/yr] for Holocene and 15 mm/yr [0.59 in/yr] for late Pleistocene. These recharge rates are comparable to results of independent numerical groundwater flow models and watershed-scale and infiltration models at Black Mesa and Yucca Mountain, respectively.

The results for the Yucca Mountain portion of this study provide a useful test of estimates of recharge used in U.S. Department of Energy performance assessments for present and future climate conditions at the potential Yucca Mountain repository. This information supports ongoing reviews of DOE and NRC agreement items and related issue resolution activities. Further, the results of this study demonstrate that estimates of recharge using the chloride mass balance method can be significantly biased when present-day chloride inputs are used in waters that are of Pleistocene origin or are a mixture of Holocene and Pleistocene recharge.

Washington Office ° Twinbrook Metro Plaza #210

12300 Twinbrook Parkway ° Rockville, Maryland 20852-1606

I

Neil M. Coleman April 2, 2002 Page 2

If you have any questions or comments about this deliverable, please contact me (210)522-5540 or Mr. James Winterle (210) 522-5249.

Sincerely,

CcEnglish C. Pearcy, Manager Geohydrology and Geochemistry

e Iwinterielintermediate milestones/861 240 /ph Enclosure cc. J. Linehan J Clocco

D DeMarco H Arlt E. Whitt J. Bradbury B. Meehan W. Ford J Greeves L Harmdan W. Reamer D Brooks J Schlueter B. Leslie K Stablein W. Dam

W. Patrick CNWRA Directors CNWRA Element Mgrs T. Nagy (SwRI contracts) P Maldonado

Late Pleistocene and Holocene Groundwater Recharge from the Chloride

Mass Balance Method and'Chlorine-36 Data

Chen Zhu1*, James R. Winterle2, and Erica I. Love'

'Department of Geology and Planetary Science, University of Pittsburgh, Pittsburgh, PA

15260, USA;

2Center for Nuclear Waste Regulatory Analyses,'Southwest Research Institute, 6220

Culebra Road, San Antonio, TX 78238, USA

*To whom all correspondence should be addressed. Email: czhu@pitt.edu; Phone: (412)

624-8766; Fax: (412) 624-9662.

To be submitted to

Water Resources Research

D:\CV\Papers\ChenZhu et al\cmb_28-cz-jrw.doc 12105/01

Abstract

The chloride mass balance (CMB) method for estimating groundwater recharge is

economic and effective, provided that the hydrological conditions for its applications are met and

the modeling parameters are known. However, modeling parameters such as precipitation and CU

deposition rates vary temporally, most notably as a result of the drastic climatic changes from

late Pleistocene to Holocene. A common practice in the application of the CMB method is to use

present day average precipitation and effective Cl deposition rates to calculate recharge rates

without regard to the age of the groundwater. This study shows that CI- deposition rates,

estimated from 36C1 data, were lower in late Pleistocene than Holocene at Yucca Mountain,

Nevada, but higher in late Pleistocene than Holocene at Black Mesa, Arizona.

Temporal variability of atmospheric CU input, as well as annual precipitation, was

considered in this study to estimate recharge rates using the CMB method. The resulting average

recharge estimates for Black Mesa are 9 mm/yr for Holocene and 35 mmryr for late Pleistocene.

Local recharge rates at Yucca Mountain were estimated from the 36CI/Cl ratios and CI

concentrations in perched waters. The estimated recharge for Yucca Mountain is 5 mm/yr for

Holocene and 15 mm/yr for late Pleistocene. These recharge rates are comparable to results of

independent numerical groundwater flow models and watershed-scale and infiltration models at

Black Mesa and Yucca Mountain, respectively.

D:\CV\Papers\ChenZhu et aI\cmb_28-cz-jrw doc 12/05/01

31

1. Introduction

An accurate estimate of recharge to aquifers in arid and semiarid areas remains a

challenge in hydrogeology because recharge fluxes under these climate conditions are low and

spatially and temporally variable [see e.g., Lerner et al., 1990]. However, recharge is an

important component of the water budget, and the extent of groundwater resources available'is in

part determined by the recharge and discharge into and ouf of aquifers [Domenico and Schwartz,

1998]. In particular, the predictive results from numerical groundwater flow models, now

commonly employed for evaluations of water resources and environmental sustainability, are

largely affected by the recharge values used as model input. Various field methods of recharge

estimates have been proposed and recently reviewed by Gee and Hillel [1988], Allison [1988],

Lerner et al. [1990], Allison et al. [1994], Phillips [1994], and Hendrickx and Walker [1997].

The Chloride Mass Balance (CMB) method, when applied to the saturated zone [Erikson and

Khunakasern, 1969; Allison and Hughes, 1978; Wood and Sanford, 1995; Wood, 1999, and

others cited in Herczeg and Edmunds, 2000], provides a first approximation of recharge. The

CMB method is still widely used [e.g., Dettinger, 1989; Lichty and McKinley, 1995; Bazuhair

and Wood, 1996; Fabryka-Martin et al., 2000; Meijer and Kwicklis, 2000], and the results of

these estimates can have significant socio-economic effects.

Because of the lack of data, an assumption commonly used in practical applications of

the CMB method to saturated aquifers is that atmospheric chloride inputs are constant through

time. The two exceptions known to the authors are the work of Stute et at. [1993] and Andrews et

al. [1994]. Stute et al. [1993] estimated different Cl 'deposition rates at present'and during the

last glacial maximum (LGM) in the recharge areas of Carrizo aquifer in Texas, based on the

distance to the Gulf of Mexico coastline. They also reconstructed the paleo- 36CI deposition rates

D.\CV\Papers\ChenZhu et al\cmb_28-cz-jrw.doc 12/05/01

4

on the basis of both Earth's magnetic field and 36C1/CI ratios in packrat middens in the deserts of

the western United States [Plummer et al., 1997]. Both estimates show that recharge rates

between 16 and 20 ka were up to two times higher than modem recharge rates. Andrews et al.

[1994] studied the East Midlands Triassic sandstone aquifer in UK. They used the CMB method

and 36CI data to explore different precipitation, evapotranspiration (ET), and CI deposition rates

during the Holocene and Pleistocene.

In this study, we used C1- and the 36C I/Cl ratios recorded in packrat middens [Plummer et

al., 1997] and Pleistocene groundwater to calculate the late Pleistocene CI deposition rates at

Yucca Mountain, southern Nevada and at Black Mesa, northeastern Arizona, respectively. Using

age-appropriate parameters for groundwater of late-Pleistocene and Holocene ages, we then

estimated recharge rates from groundwater Cr- concentrations for the different time periods.

In the and to semi-arid Black Mesa area (Figure 1), the N aquifer is the only local source

of drinking water. Withdrawals from the aquifer in the Black Mesa area have been increasing

dunng the last 30 years, and water levels have declined significantly [Thomas and Truini, 2000].

Additionally, Peabody Western Coal Company operates a strip coal mine and uses the N-aquifer

water to transport pulverized coals to the Mojave desert through a slurry pipe. Numerical

groundwater flow models have been used to evaluate the long-term environmental effects of

these groundwater withdrawals [Eychaner, 1983; Brown and Eychaner, 1988; GeoTrans, 1987].

The predictions from these numerical models, however, depend on the assumed recharge rates

[Lopes and Hoffman, 1997]. Recharge rates have been estimated using the Maxey-Eakin method

[Brown and Eychaner, 1988; GeoTrans, 1987] and spatial distribution of 14 C ages [Lopes and

Hoffman, 1997; Zhu, 2000]. However, recharge values are still a controversial topic in the Black

D:\CV\Papers\ChenZhu et al\cmb_28-cz-jrw doc 12/05/01

5

Mesa region, and the estimated recharge rates bear significantly on water resource management

decisions.

The deep unsaturated zone beneath Yucca Mountain is presently being evaluated as the

proposed site of a geologic repository for high-level nuclear waste (Figure 2). Reliable estimates

of recharge at the local scale are important because infiltrating waters provide a potential means

for mobilization of radionuclides. The CMB method was previously used by Fabrka-Martin et

al. [2000] to estimate recharge from unsaturated zone water at Yucca Mountain. However, water

squeezed from rock cores in the unsaturated zone represents point measurements and thus may

not be amenable to the CMB method, which is more appropriate for basin-scale recharge

estimates. Additionally, matrix water from cores may not be representative in formations where

fracture flow is prevalent. Meijer and Kwicklis [2000] used the CMB method to estimate

recharge from saturated zone water beneath Yucca Mountain, but those samples may be more

representative of recharge in the higher elevationsto the north of Yucca Mountain. No one has

used this method for samples pumped or bailed from wells in perched zones at Yucca Mountain,

which are derived from local recharge and represent a larger scale than water obtained from core

samples. Flint et al. [1996] used a watershed-scale numerical model to estimate spatially

distributed mean annual infiltration at Yucca Mountain for both modem and pluvial climate

conditions. The CMB method, as'applied in this study, provides a useful tool for validating both

the modem and pluvial recharge rates estimated from the numerical approach.

2. The Chloride Mass Balance Method

Wood and Sanford [1995] and Wood [1999] have eloquently stated the necessary

conditions for and limitations of successful applications of the CMB method for estimating

D.\CV\Papers\ChenZhu et aI\cmb_28-cz-irw.doc 12/05/01

6

recharge fluxes to groundwater aquifers. Only a brief introduction of the method is presented

here, with discussion limited to the use of CU- concentrations from the saturated formations only.

If it can be assumed that atmospheric deposition is the only source for CU in infiltrating

water, that CU- behaves as a conservative tracer along its path, and that surface run-on and run-off

can be neglected, then the CI- concentration of groundwater recharge is a result of

evapotranspiration (ET) loss of water. Thus, if the amount of precipitation and the Cl

concentration in precipitation are known, recharge to the aquifer (I.T), R, can be estimated from

the mass balance equation [Wood, 1999],

R=P Cleff( ClGW

where ClGwV (M/L 3) stands for Cl- concentration in groundwater, Cleff(MIL 3) is the "effective"

chloride concentration of surface infiltration resulting from both wet and dry deposition, and P

(U/T) denotes the mean annual precipitation (MAP). The mass balance equation (1) is also

applicable to 36C1 [Stute et al., 1993; Andrews et al., 1994].

Dry fallout of CU- is retained in the soil and leached during recharge, and hence must be

accounted for in the mass balance [Dettinger, 1989; Wood and Sanford, 1995]. Dry deposition

rates are quite variable, depending on distance from the ocean, wind patterns, and local sources

of CU [Plumnnier, 1996; Davis et al., 1998; Fabryka-Martin et al., 2000; Harrison et al., 20011.

Dettinger [1989] suggested that dry deposition accounts for 33% of the total CU surface fluxes in

the Great Basins of Nevada and Utah. Climate changes over geological time may have different,

or even contrary influences on the contribution from wet and dry deposition [Fabryka-Martin et

al., 19971.

D \CVXPapers\ChenZhu et al\cmb_28-cz-jrw doc 12/05/01

I

7

In regional aquifers, the groundwater residence time may span tens of thousands of years,

during which time major climate changes have occurred. Hence, each vanable in equation (1) is

time-dependent [Phillips, 1994; Wood and Sanford, 1995; Tyler et al., 1996; Wood, 1999; Love

et al., 2000; Scanlon, 2000], and equation (1) can be rewritten as

Cl~ff (t) R(t) = P(t) CIG; W (2) ClGw (t)

where t stands for time, or

CAJ

RI = p& Cleff (3) Cl~lI•

where At is a time bracket or time period of concern. Parameters denoted by At are average

values during the time brackets.

However, in the vast majority of practical applications of the CMB method, Cl4,tris often

obtained from measurement of CI" concentrations in present-day precipitations and is assumed to

be constant through time. The value of P used is usually the yearly average, extending back to as

many years as records are available, which is usually a few decades. However, the Cr

concentrations in groundwater, ClGw. may represent recharge conditions (i.e., Cleff and P) from

thousands of years ago. Therefore, an application of equation (1) may mix parameters of

different times.

Studies of Greenland Ice Sheet Project (GISP2) ice cores have clearly demonstrated

temporal variability of atmospheric CI deposition [Figure 3a, Mayeivski et al., 1994]. In the

GISP2 ice core, CI" preserved in Pleistocene ice (- 100 ppb) was an order of magnitude higher

than in Holocene ice (- 10 ppb). The abrupt decrease of CI concentrations in the ice core

D.\CV\Papers\ChenZhu et ai\cmb_28-cz-jrw.doc 12/05/01, -

j_______

8

occurred at -11 ka, during the transition from Pleistocene to Holocene. This decrease of Cl

concentrations is accompanied by a decrease of Na' and S042 concentrations. Because the Na'

concentrations are not affected by the loss of HCI and, hence, are more indicative of sea salt

input at the time (R. Edwards, Old Dominion University, personal communication, 2000). While

direct extrapolation of the Greenland ice core data to the Southwestern United States is not

warranted, the marked changes in CI- concentrations in the ice core that accompany the known

drastic changes in paleoclimate make the assumption of a constant Cr surface flux questionable.

Chlorine in the atmosphere mostly arises from sea salts [Feth, 1981]. Variations of the prevailing

atmospheric circulation patterns accompanying a major paleoclimate shift at the transition from

late Pleistocene to Holocene on the North American continent [Thompson et al., 1993] would

have affected the Cl- deposition rate.

Additionally, the long-term ClI surface mass flux is more appropriate for use in the CMB

method because short-term measurements of wet and dry deposition rates are known to have

large spatial, temporal, and seasonal variability [Phillips, 1994; Fabryka-Martin et al., 2000;

Scanlon, 20001. The results of short-term measurements also vary with methods of sample

collection [Phillips, 1994]. As an alternative approach, the long-term average deposition rate

(Dcr) can be estimated from 36C1 data through the following equation [Phillips, 1994; Andrews et

al., 1994; Fabrvka-Martin et al., 2000; Scanlon, 2000],

DC ( 36 (4) D 1(6Cl /Cl)(4

D:\CV\Papers\ChenZhu et ai\cmb_28-cz-jrw.doc 12/05/01

9

where D3M denotes the prebomb deposition rate of 36 C1 and (36C! I CI) is the prebomb ratio. D36

varies with latitude and is also a function of precipitation. Phillips [2000] provided an empirical

relationship regressed from the available data between latitudes 46 to 340N:

'Di 6 =0.047P + 8.09, (5)

where D36 has units of atoms nm2 s-1 and, here, P is' the average annual precipitation in mm yf r.

The variable Clff can be calculated from the relation Clog = Dcl'l P. Phillips [1994] used this

method to estimate Dc, in New Mexico. Fabryka-Martin et al. [2000] used equations (4) and (5)

to calculate the long-term (Holocene) chloride deposition rate at Yucca Mountain. Scanlon

[2000] calculated the "long-term" Dci in the Chihuahua Desert, Texas.

Reconstruction of the atmospheric 36 C/C1 from packrat urine data in Nevada indicates

that the 36C1/Cl ratio was higher by a factor of 1.5 to 2 in late Pleistocene (800 to 1,000 x 10"5)

compared to Holocene (500 to 600 x 10-15) [Plunumer et at., 1997]. Similarly higher 36C1/Cl

ratios were found in soil water from the deep vadose zones in the Nevada Test Site, which are

believed to be older than 20 ka [Tyler et al., 1996]. Plummer et al. [1997] calculated that the

average 36CI atmospheric production rate from 12 to 40 ka was 33% higher than the present

value, which is consistent with calculations from geomagnetic field intensity data [Davis et al.,

1998]. If a linear correlation exists between production rate and deposition rate, as used by Stute

et al. [ 1993], a factor of 1.33 should be applied to equation (5) for calculating D36 for the late

Pleistocene. It is appropriate, when the CMB method is applied to late Pleistocene groundwater

in which the CI" concentration is measured, for the late Pleistocene CI" deposition rate to be used.

D.\CV\Papers\ChenZhu et aI\cmb_28-cz-jrw.doc 12/05/01

I________

10

3. Black Mesa, Arizona

3.1 Hydrogeologic Setting

The N aquifer in the Black Mesa basin, Arizona, is a 14,000 km 2 regional aquifer and the

most important source of potable water for the region [Cooley et al., 1969; Evychaner, 1983; Zhl,

2000]. The most productive unit in the N aquifer is the Navajo Sandstone, an Upper

Triassic/Lower Jurassic aeolian quartz sandstone well known for its large cross bedding and

spectacular landforms. Groundwater in the N aquifer moves southeast from the Shonto area

toward the center of the mesa, where the flow path diverges northeastward toward Laguna Creek

and Chinle Wash and southwestward toward Moenkopi Wash [Cooley et al., 1969; Brown and

Eychaner, 1988; Figure 11. The aquifer is recharged seasonally from precipitation in the

highlands, principally during the winter and spring. The Shonto area, in the northwestern corner

of the mesa, accounts for approximately one-third of the total recharge in the basin and most of

the water that flows into the center of the basin [Cooley et al., 1969]. The principal discharge

areas are along the Moenkopi Wash near Tuba City and into the alluvium along Dinnebito and

Oraibi Washes.

The N aquifer is overlain unconformably by the silty Carmel Formation, which

hydraulically separates the N and D aquifers in the Black Mesa area (Figure 4). Groundwater

from the D aquifer is characterized by high total dissolved solids (TDS), elevated S02-, CU, and

Na' concentrations, and lower dissolved oxygen (DO). Leaking of D water into the N aquifer is

restricted to the southeast edges of the mesa, but some wells in other areas are partially screened

in the D.

The current climate in the Black Mesa area is semiarid. The MAP is less than 300 mm in

much of the area but as much as 320 mm in areas above 1,800 m in elevation. The mean annual

D:\CV\Papers\ChenZhu et aI\cmb_28-cz-jrw doc 12/05/01

11

temperature is about 10 PC in the Shonto area. Vegetation consists of grass shrubs at altitudes

below 1,670 m, pinion-juniper between 1,670 and 2,280 m, and pine forest above 2,280 m

[Cooley et al., 1969].

3.2 Groundwater Chloride Data

In the N aquifer, Cl concentrations are considerably higher in Holocene water than in

late Pleistocene water (Figure 3b). The groundwater ages on the horizontal axis in Figure 3b are

estimated from geochemical modeling and age corrections [Zhu, 2000]. In Figure 5, the Cl

concentrations are plotted together with the 8D and 818 0 data of the same samples. It is generally

known that Pleistocene water has depleted H and 0 stable isotope values with respect to recent

water because of a cooler and more humid climate in late Pleistocene [Merlivat and Jouzel,

1979; Clark and Fritz, 1997]. Figure 5 shows that the lower CI" concentrations are associated

with _8D- and 8180-depleted water. It should be noted that processes that possibly affect Cr

concentrations in the N aquifer at Black Mesa all tend to increase CI concentration in the

downgradient and older groundwater; these include (1) diffusion or leakage from the overlying D

aquifer, (2) dissolution of CF bearing minerals (e.g., biotite) or fluid inclusions, and (3) sampling

artifacts such as partial screening in the D or Carmel Formation.

Different Ci" concentrations in groundwater recharged during the LGM (compared to

Holocene groundwater) were also found in other aquifers. For example, studies of the Great

Artesian Basin in Australia found a similar decrease of Cr with age along the hydraulic gradient.

Love et al. [2000] attributed this to higher past recharge rates. Stute et al. [1993] obtained 14C age

estimates ranging from modem to about 35 ka for waters from the Carrizo aquifer. of Texas, and

found minimum CI" concentrations around 12-17 ka. Purdy et al. [1996] found a C1'

D.\CV\Papers\ChenZhu et al\cmb_28-cz-jrw doc 12/05/01

12

concentration minimum in the Aquia aquifer in the Atlantic Coastal Plain in an area that also has

depleted 8D and 8180, which is consistent with being recharged during the LGM.

3.3 36C!1Ci data and Chloride Deposition Rates

In this study, CI- deposition rates for late Pleistocene and Holocene in the Black Mesa

area were estimated from the 36Cl/Cl data, following the method described in Section 2. 36 C1/C1

ratios in the N aquifer were measured to be 840 and 882 x 10-15 for groundwater samples from

two wells [Lopes and Hoffinan, 1997], which have corrected 14C ages of 25 and 27 ka,

respectively [Zhu, 2000). Lopes and Hoffinan [1997] measured 36C1/CI in four additional

groundwater samples from the N aquifer, but these samples appear to have mixed with

groundwater from the overlying D aquifer [Davis et al., 2000]. The D-aquifer groundwater has

36Cl/Cl ratios from 10 to 127 x 1015 [Lopes and Hoffinan, 1997] and CI concentrations up to 100

mg/L. The Cl/Br ratios in many D-aquifer samples is greater than 1000, indicating dissolution of

evaporites. In contrast, N-aquifer water has Cl/Br ratios generally less than 200. Mixing of a

small portion of the D-aquifer water in the N-aquifer, which can result from both inter-aquifer

leakage or partial screening of the sampling well in the D-aquifer, will skew the 36C1/Cl ratio of

the N-aquifer water. Therefore, the N-aquifer samples that are suspected to be mixed with D

aquifer water are excluded from further discussion.

Additional 36C1/C1 ratio measurements for groundwater are available in the nearby San

Juan basin, New Mexico [Plummer, 1996; Phillips, 2000]. The Black Mesa and San Juan basin

are only 150 km apart. The 36C1/Cl ratio in Holocene San Juan basin groundwater is 765 x 1015,

but late Pleistocene age groundwater (based on corrected 14C dates) shows 36C /C ratios around

1000 x 10.15 (Figure 3b). The temporal variation patterns of 36C1/C1 ratios in these radiocarbon-

D:\CV\Papers\ChenZhu et aI\cmb_28-cz-jrw doc 12/05/01

13

dated groundwater samples are similar to that documented for the fossil rat urine sequence

[Plunmer et al., 1997; Phillips, 2000]. Thus, the 36CI/Cl data for Black Mesa and San Juan basin

together provide the basis for the following calculations.

For calculating the Holocene Cleff, the prebomb 36C1/Cl ratio of 765 x 1015 from Plu miner

[1996] and Phillips [2000] and the average MAP of 305 mm/yr for the last 60 years (1948-1998)

were used (Western Regional Climate Center, www.wrcc.dri.edu). Black Mesa is located within

the latitudes where the parameters in equation (5) are regressed, so we assume equation (5) is

applicable, which yields a calculated D36 of 22 atoms m-2 s-1. Using equation (4), the calculated

Dct is 54 mg m-2 yr-'. That is equivalent to a Cleff of 0.18 mg[L.

This Holocene value of Cleffis about half the estimate of 0.35 mgJL from the estimates of

present day wet and dry deposition [Zhu, 2000]. Phillips [1994] estimated from 36C1 data a Dc, of

75 to 125 mg m-2 yfI for sites in New Mexico. Scanlon [2000] used the same method and

estimated Dc, of 87 mg m2 yfr1 for the Chihuahua Desert, Texas. This rate is equivalent to a Clef

of 0.27 mgL for a MAP of 320 mm/yr, which is about twice as large as the weighted average of

0.14 ± 0.03 mg/L for Cliff obtained from monthly measurements of precipitation and dry fallout

for two years. These differences between Clff 'estimated from ?6C1 data and MAP estimates and

Cliff estimated from average values in precipitation illustrate some of the uncertainty in

obtaining recharge estimates using the CMB method. Scanlon [2000] considered the Cl input

estimated from 36 C1 data more valid, and it is likewise favored in this study, because it represents

a longer-term period. While uncertainty in the Cleff input results in proportional uncertainty in the

absolute value of the recharge estimated using the CMB method, greater confidence can be

placed in the relative difference between recharge values estimated for the late Pleistocene and

the Holocene.

D.\CV\Papers\ChenZhu et al\cmb_28-cz-jrw.doc 12/05/01

I_________

14

For calculating the late Pleistocene Clff, a value of 1000 x 10-15, the average of 36C!/Cl

ratios in Pleistocene groundwater at Black Mesa and San Juan basin, was used as the background

ratio at that time. Implicitly assumed is that the atmospheric signature is preserved and there was

no mixing of groundwater or significant cosmogenic production in the aquifer. Radioactive

decay of C36C would decrease the initial ratio by only about 8% for our oldest sample (35 ka),

which is approximately equal to the analytical error of measurement [Plummer, 1996].

Therefore, decay is neglected in our discussion of both the Black Mesa and Yucca Mountain

36Cl/CI data. Similarly, because of short groundwater residence time (<< 301 ka. the half life of 36D

36Cl), production of 36Cl in groundwater is assumed to be negligible [Stute et al., 1993; Plummer,

1996]. The 36CI deposition rate was calculated from equation (5), but, as previously discussed,

33% more deposition was added (factor of 1.33) assuming a linear correlation to atmospheric

production in late Pleistocene, as discussed in Section 2. The precipitation rate in the late

Pleistocene is unknown, but studies of paleorecharge show that the recharge rates were two to

three times higher in late Pleistocene than today [Zhu et al., 1998]. The main recharge area at

Black Mesa (Shonto area) is above 7,500 feet, and most likely precipitation dunng the last

glacial period was > 500 mm/yr (J. Betancourt, U.S. Geological Survey, personal

communication, 2001). In the absence of independent estimates, we assume that the same factor

for precipitation increase that was estimated for the Pleistocene at Yucca Mountain (see Section

4), also applies to Black Mesa. This leads to an estimate of 538 mm/yr MAP for the glacial

period. Note that P appears twice in calculations, once in equation (5) and again in converting

Dc, to Clet. Using these parameters, the calculated D 36 in late Pleistocene is 44 atoms m2 s-l, the

calculated Dct is 82 mg m2 yrl1, and the calculated Cleffis 0.15 mg/L.

D.\CV\Pavers\ChenZhu et al\cmb 28-cz-irw doc 12/05/01

15

Compared to the Holocene 36C1 deposition rate, D 3 6 calculated for the late Pleistocene

period was twice as high because of both higher precipitation and higher atmospheric production

of 36CI. The estimated Dci in the late Pleistocene is 1.5 times higher than that in the Holocene.

However, because of higher precipitation, the Cleffis lower than that in the Holocene, which is

consistent with the general observation that CI" concentrations decrease with increasing

precipitation [Phillips, 1994].

The wet Cl deposition rate is a function of distance to the oceanic sources of sea salt

[Junge and Werby, 1957]. Stute et al. [1993] believed that the distance from the recharge areas of

the Carrizo aquifer to the Gulf of Mexico coastline during the LGM was 325 km versus about

200 km today because, during the LGM, sea level was -120 m lower. Based on that, Stutte et al.

[1993] postulated a 23% reduction in CI" deposition during the Pleistocene compared to today.

Purdy et al. [1996] also used the distances from the recharge area of the Aquia aquifer to the

Atlantic coast as the basis for different Cl" deposition rates in late Pleistocene and Holocene.

However, the Black Mesa area is located in the continental interior. If CU concentration in the

wet deposition is-an exponential function of the distance to the sea, then this effect is diminished

inland. Furthermore, Stute et al. [1993] considered only wet deposition, not dry deposition. This

is also true in the Purdy et al. [1996] study. Strong winds in late Pleistocene would have had an

effect on dry CI" deposition [Harrison et al., 2001].

3.4 Estimate of Recharge Rates

Using the parameters estimated in the above section and listed in Table 1, recharge rates

were calculated using equation (3) (Figure 3c). For the Black Mesa groundwater chloride data in

Figure 3b, recharge rates estimated for late Holocene (6 ka to present) range from 3 to 20 mm/yr

D:\CV\Papers\ChenZhu et al\cmb_28-cz-jrw doc 12/05101

16

and average 9±5 mm/yr (±_Io, and thereafter). Only one sample has a radiocarbon age between

11 ka and 6 ka, and the estimated recharge rate for that sample is 19 mm/yr. Recharge rates

estimated for late Pleistocene were higher, ranging from 7 to 95 mm/yr and averaging 35_22

mm/yr. There was a period of peak recharge of about 60 mm/yr, occurring between 14 and 17 ka.

Note that the percentage of precipitation that becomes recharge, or the ratio RIP, is

independent from the assumption of P. As illustrated in Figure 3d, we calculated the percentage

of recharge to be 4±2% for the Holocene and 8±1_% for the late Pleistocene.

3.5. Discussion of the Black Mesa Recharge Estimates

Recharge rates estimated from the CMB method indicate pronounced variation through

time. As shown in Figure 3c, the timing and magnitude of variation in recharge estimates using

the CMB method are consistent with recharge estimates from numerical models that are

calibrated to radiocarbon ages [Zhu et al., 1998; Zhu, 2000], which are entirely independent from

the C- data. Note that the trend of the recharge variation from 14C ages and numerical modeling

is more reliable than the estimated recharge rate values because the recharge estimates depend on

the effective porosity values in the transport model. What is important to note is that both the

CMB method and numerical modeling indicate that recharge in late Pleistocene was 2 to 3 times

higher than today. These recharge changes correlate well with known paleoclimate changes

during late Pleistocene and Holocene documented by a variety of climate proxies [see Zhu et al.,

1998 for references]. The Mean Annual Temperature (MAT) during the LGM was 5-6 'C cooler

than today, but during the drier Holocene maximum, the summer temperature was 2-4 TC warmer

than today [Thompson et al., 1993].

D \CV\Papers\ChenZhu et aI\cmb_28-cz-jrw doc 12/05/01

17

Higher percentages of precipitation that becomes recharge in late Pleistocene (Figure 3d;

-8%) appear to be consistent with the known cooler and more humid climate in late Pleistocene.

Additionally, precipitation during late Pleistocene was shifted to winter from summer, which

may also result in a higher percentage of MAP that becomes recharge. The estimated 4%

recharge in Holocene is generally consistent with other estimates for arid and semiarid areas

[Wood and Sanford, 1995].

4. Yucca Mountain, Southern Nevada

4.1 Hydrogeologic Setting

The unsaturated zone at Yucca Mountain comprises gently dipping layers of Tertiary

volcanic tuffs of differing hydrologic properties. At the surface, the amount of precipitation that

becomes recharge varies with changes in slope,'orientation, vegetation, depth of alluvial cover,

and bedrock properties [Flint et al., 1996]. The episodic nature of recharge at Yucca Mountain is

greatly attenuated by the presence of non-welded tuff layers with relatively few fractures and

high hydraulic diffusivity that are situated above the geologic horizon proposed for construction

of the nuclear waste iepository. The proposed 'r6pository horizon is situated in densely welded

tuffs with very low matrix permeability (<1018 m2 ) but considerable fracture permeability (>10

12 m2). Layers of sparsely fractured non-welded tuffs lie between the water table and the

proposed repository-horizons; perched water has been observed where the permeability of these

layers has been significantly reduced by zeolitic alteration.

4.2 Chloride Deposition Rates

D.\CV\Papers\ChenZhu et aI\cmb_28-cz-jrw.doc 12/05101

I _________

18

From a prebomb background 36C1/C1 ratio of 500 x 1015, Fabka-Martin et al. [2000]

calculated a Holocene C1- deposition rate of 60 mg m- yr- and an average Clff value of 0.35

mgIL (see Table 1 for the parameters they used). This long-term deposition rate compares

favorably to an average of 0.5 mg/L for present day precipitation at Yucca Mountain measured in

111 samples. Dettinger [1989] studied 15 groundwater basins in Nevada and used an average

Cleff of 0.4 mgfL to estimate recharge. Measurements of CI- concentration in precipitation at Red

Rock Canyon outside of Las Vegas, as a part of the National Atmospheric Deposition Program

(NADP). resulted in an average of 0.16 mgfL for an average MAP of 162 mm/yr. Measurements

at Springs Basin, north of Yucca Mountain, believed to be an analog for pluvial climate

conditions, resulted in an average Cleff of 0.51 mgfL with MAP of 335 mm/yr [Fabryka-Martin

et al., 2000].

In this study, we calculated the CI deposition rate for late Pleistocene while we adopted

the Holocene deposition rate from Fabryka-Martin et al. [2000]. A study of the timing of climate

change based on 180 data from calcite in Devil's Hole, Southern Nevada, indicates that climate

in the Yucca Mountain region during the late Pleistocene was indicative of a transitional or

intermediate state between full glacial and interglacial states [Winograd et al., 1997; Landwehr

et al., 1997]. Estimates for a long-term average P for the glacial transition climate at Yucca

Mountain were obtained from analog sites with MAP rates from about 200 to 430 mm/yr [U.S.

Department of Energy, 2000a]. For this study, an intermediate value of 300 mm/yr is used. The

36 C1 deposition rate was calculated from equation (5), but assuming 33% more atmospheric 36)

production of 36Cl and a linear correlation to atmospheric production in late Pleistocene (see

Section 2). A higher local meteoric 36C1/C1 at Yucca Mountain (1000 x 10-15) was assumed for

late Pleistocene [Plunmer et al., 1997, Fabrvka-Martin et al., 2000]. The calculated D36 is 22

D:\CVXPapers\ChenZhu et aI\cmb_28-cz-jrw~doc 12/05/01

19

-2-I m-2

atoms m-2 s-1 and a Dct of 41 mg m yfl that is equivalent to a Cleff Of 0.18 mgJL for MAP of

300 mm/yr.

The calculated Pleistocene Cleff is about half that of the Holocene. Fabryka-Martin et al.

[1997] pointed out that the Cleff could be lower in the Pleistocene because dry lakebeds today

were water-filled during the last glacial period, lowering the wind blown Cr near Yucca

Mountain. Meanwhile, inspection of the deposition rates during El Nihio episodes showed that

wet CI deposition rates were similar in El Nifio years to normal years [Fabryka-Martin et al.,

1998]. Paleoclimatic conditions during the glacial times were believed to be similar to those

prevailing during an extreme El Nifio event [Fabryka-Martin et al., 1998].

4.3. Recharge Estimates for Saturated Groundwater

Recharge rates were calculated from the CI" concentrations in groundwater beneath

Yucca Mountain from the CMB method (Table 2). In calculating recharge rates, we assumed that

groundwater from the saturated zone is of late Pleistocene origin and that the late Pleistocene

modeling parameters (300 mm/yr for P and 0.18 mg/L for Cleff) should be used. The calculated

recharge rates range from 5 to 9 mm/yr and average 8 t 1 mm/yr. It is thus estimated that about

3% of precipitation became recharge during the late Pleistocene.

The recharge rates estimated in this study are close to the lower end of the range

estimated by Meijer and Kwicklis [2000] (see Figure 6), who applied the CMB method using

parameters for the Holocene period, including two C/ff values, 0.3 mg/L and 0.6 mg/L, to

evaluate uncertainty in Cl deposition. In addition to a Holocene Cleff, Meijer and Kwicklis

[2000] also used a modem MAP rate of 170 mm/yr. Radiocarbon, 8D, and 8180 data; however,

D:\CV\Papers\ChenZhu et al\cmb_28-cz-jrw.doc 12/05/01

20

show that saturated zone groundwater beneath Yucca Mountain has a late Pleistocene recharge

age [Meijer and Kwicklis, 2000].

Although our calculation may be better suited for late Pleistocene recharge, it is still of

limited use for Yucca Mountain site investigations because saturated zone water beneath Yucca

Mountain may be more indicative of recharge from areas that are hydraulically upstream. Hence,

modeling parameters that better represent upstream geographic locations would be more

appropriate. To obtain a more representative estimate of local recharge at Yucca Mountain, we

estimate recharge from samples of perched water at Yucca Mountain in the following section.

4.4. Recharge Estimates for Perched Water

Fabryka-Martin et al. [2000], among others, summarized the perched water bodies found

in volcanic tuffs beneath Yucca Mountain. Meijer and Kivicklis [2000] discussed the quality and

representativeness of samples, including whether the samples were obtained by bailing or

pumping after purges for a few pore volumes. They concluded that the most representative

samples are from SD-7 on and after March 16, 1995; from UZ-14 PT-4 on August 27, 1993; and

from WT-24 on October 22, 1997. Yang et al. [1996] show that UZ-14 PT-I and PT-2 were also

sampled during the pumping test, and UZ-14 D was bailed after the pumping test. These samples

are also included in the analysis of this study. Chloride concentrations in the six samples from

these three locations range from 4.1 to 9 mg/L.

The ongin and timing of the perched water bodies at Yucca Mountain remains uncertain.

Fabryka-Martin et al. [2000] summanzed the following hypotheses: (1) they represent a

transient rise in water level in the saturated zone; (2) they represent a remnant from a time during

which infiltration rates were higher; (3) they reflect long-term steady-state conditions.

D.\CV\Papers\ChenZhu et aI\cmb_28-cz-jrw.doc 12/05/01

21

Major ion chemistry and various isotope ratios have been used to compare the perched

water with unsaturated zone matrix water and saturated zone water at Yucca Mountain [e.g.,

Fabryka-Martin et al., 2000; Meijer and Kwicklis, 2000]. In general, macro-ion chemistry of

perched water shows a similarity with saturated zone water beneath Yucca Mountain, but is in

sharp contrast with the unsaturated zone rock-matrix pore water, which shows much higher

concentrations of dissolved solids [Meijer and Kwicklis, 2000; Fabryka-Martin et al., 2000]. On

the other hand, 87Sr/ 86Sr data show that perched water is isotopically distinct from saturated zone

water [Fabrka-Martin et al., 2000].

Three environmental isotopes potentially proyide the timing information for recharge of

perched water: (1) 0 and H stable isotopes, (2) 14C, and (3) 36C1. The 0 and H isotopes are

advantageous because they are a part of the water molecules. The 0 and H stable isotope values

of perched waters fall within the range of those found in saturated zone water beneath Yucca

Mountain and those found in groundwater beneath northern Fortymile Wash (Figure 7).

Groundwater beneath northern Fortymile.Wash, sampled from well UE-29 a#2, represents recent

Holocene recharge [Meijer and Kwicklis, 2000; Fabryka-Martin et al., 2000]. Note in Figure 7

that groundwater sampled from well UE-25 p#1 is from a deep, confined carbonate aquifer and is

likely far removed from any recent source of recharge. All other data shown in Figure7 are from

wells sampled in the upper volcanic tuff aquifer beneath the Yucca Mountain area.

The same perched water data for Yucca Mountain have been interpreted differently by

different authors: they are taken to indicate Holocene recharge by Fabryka-Martin et al. [2000],

and late Pleistocene recharge by Meijer and Kwicklis [2000]. As elaborated below, the isotopic

data are consistent with the hypothesis that most perched waters at Yucca Mountain are a

mixture of Pleistocene and Holocene water [e.g., Sonnethal and Bodi'arsson, 1999]. The fact that

D:\CV\Papers\ChenZhu et al\cmb_28-cz-irw.doc 12/05/01

I _________

perched water has much lower C1- concentrations than is found in pore waters of rock matrix

core samples in the unsaturated zone indicates that fracture flow responding to extreme weather

events rather than pervasive flow through the matrix is the most likely cause for Holocene

replenishment [Fabryka-Martin et al., 1997].

Because the recharge estimate from the saturated zone C1- concentration data may not

indicate local recharge at Yucca Mountain, the perched water bodies provide an opportunity to

decipher local recharge history near the proposed repository. However, in using the CMB

method to estimate recharge rates, the modeling parameters for Holocene and Pleistocene ought

to be different (see Table 1). It is necessary then to estimate the mixing proportions of the

Pleistocene and Holocene inputs that compose perched water.

Because the 36C1/Cl ratio would not change with ET [Clark and Fritz, 1997], the

observed 36Cl/Cl is only a result of mixing of Holocene and Pleistocene water. Therefore, the

36CI/Cl ratios can be used to estimate the proportions of Holocene and Pleistocene recharged

water (Figure 8). The relationship between 36C1/Cl ratio and Cf concentration in a two

component system is governed by the mixing equation (Faure, 1996, p142)

C l I/ ci . 3 i j I _ r P NJ H

t-c-7-) Clnl"(ClIb -ClP,) + .C/,, Cl•,) (6)

where the superscripts ti, P, and H denote mixture, Pleistocene, and Holocene water,

respectively. The 36C1/CI ratios in the H and P perched water should be the same as their

respective meteoric input because ET would not change them. The measured Clg,, of perched

D \CV\Papers\ChenZhu et aI\cmb_28-cz-jrw.doc 12/05/01

23

water is'a mixture of H and P recharged water. C14, and CiP, are unknown but can be estirmated

from a mixing analysis [see Faure, 1996, p. 144].

A number of 36CI/Cl and CI" concentration measurements are available for perched waters

at Yucca Mountain (Table 2). As mentioned earlier, the samples from the three pumped wells,

WT-24, UZ-14, and SD-7, were considered more reliable and hence were favored for this mixing

analysis. However, as shown in Figure 8, data points from these three wells cannot be fitted

along a single mixing line. Note that fitting a mixing'line between data'points from different

wells implies an assumption that the percentage of precipitation that becomes recharge is the

same at each well location. In an area of such varied topography, soil cover, and vegetation as

one finds at Yucca Mountain, this assumption must be carefully evaluated. Wells UZ-14 and

WT-24 were chosen because both intersect perched water bodies located above the basal

vitrophyre of Topopah Spring welded (TSw) unit, while SD-7 perched water is above a zeolitic

portion of the upper Calico Hills non-welded (CHn) unit [Yang et al., 1996]. Additionally, WT

24 and UZ-14 are both located north of the proposed repository and both tap into what is

conceptualized as a relatively extensive perched water system covering an area of several km 2

[U.S. Dept. of Energy, .2000b]. Conversely, the perched water body sampled by SD-7 is located

to the south of the proposed repository area and is thought to be very small 'and isolated, as

inferred from pumping tests [U.S. Dept. of Energy, 2000b]. SD-7 is characterized by CI"

concentrations (4.1 mg/L) that are much lower than other perched water or saturated zone water.

Numerical modeling of CI" distribution in perched waters predicted higher CU" concentrations for

SD-7 under both modem and glacial climate scenarios [Sonnethal and Bodvarsson, 1999].

234U/238U ratios in wells WT-24 and UZ-14 are similar and are much higher than those in SD-7

[Meijer and Kwicklis, 2000]. Sr concentrations in WT-24 and UZ-14 are also much higher than

D:\CV\Papers\ChenZhu et a1\cmb_28-cz-jrw.doc 12/05/01

24

in SD-7 [Sonnethal and Bodvarsson, 1999]. Based on these considerations, the mixing line for

this analysis was fitted between UZ-14 and WT-24 data points (Figure 8).

Note that the three pumped UZ-14 samples (PT1, PT2, and PT3) are all aligned along the

mixing line. Arguably, water was drawn from different parts of the perched water body as

pumping continued. Based on the line in Figure 8, the UZ-14 PT1 sample contains about 63%

Pleistocene water; PT2 shows slightly more Pleistocene water as pumping continued. By the

time PT4 was sampled, well UZ-14 was drawing about 68% Pleistocene water. Therefore,

although there are only two well locations, several samples from well UZ-14 that represent

different volumes of the perched water body surrounding the well are all consistent with the

mixing line. Note also that sample UZ-14 D was bailed after pumping ceased. This point also

falls close to the mixing line in Figure 8 and shows a relatively high proportion of Pleistocene

water.

From the fitted mixing line in Figure 8 and the assumed 36CI/Cl of the H and P end

members (see Table 1), C1 H., and Cl , are calculated to be 13.5 and 3.5 mg/L, respectively.

Mixing calculations show that Holocene precipitation-denved water composes 56% of WT-24

perched water and -32% of UZ-14 perched water. Then, using equation (3) and the P and Cle1o

values in Table 1, we calculated the Holocene and Pleistocene recharge rates as 5 mm/yr and 15

mm/yr, respectively.

The estimated Cr concentration in Holocene-recharged perched water is much lower than

unsaturated zone rock-matrix pore water, which varies from 32 to 7,400 mg/L [Fabryka-Martin

et al., 20001. This is consistent with a fracture flow ongin of perched water. The estimated CI

concentrations in Pleistocene-denved perched water is lower than the saturated zone water

beneath Yucca Mountain. Compared to the saturated zone water beneath Yucca Mountain, the

D:\CV\Papers\ChenZhu et al\cmb_28-cz-jrw doc 12/05/01

25

perched water has slightly higher radiocarbon activities and more negative 813C values. This is

consistent with a mixture of Pleistocene and Holocene water.

4.6. Discussion of the Yucca Mountain Recharge Estimates

The results of the mixing analysis indicate that perched waters at Yucca Mountain are a

mixture of Pleistocene and Holocene recharge, and that recharge rates in the northern portion of

Yucca Mountain were approximately three times as great during the late-Pleistocene. This

interpretation is consistent with paleoclimate reconstruction studies that suggest the Yucca

Mountain region was cooler and wetter during the late-Pleistocene period of transition from

glacial to interglacial climate conditions. .

As previously mentioned, there is considerable uncertainty in the values of estimated

recharge obtained from this analysis, owing to the inherent uncertainties in the necessary

assumptions regarding long-term precipitation and chloride deposition rates, and the timing and

duration of changes in climatic conditions. Interestingly, the estimated recharge rates of 5 mm/yr

and 15 mm/yr for Holocene and late-Pleistocene periods are consistent with recharge rates

estimated from numerical watershed modeling conducted for the Yucca Mountain Project. For

example, the Flint et al. [1996] infiltration model yielded a spatially averaged net infiltration

estimate of 4.5 mm/yr over the 124,km 2 model domain using the same MAP of 170 mm/yr used

in this analysis for the Holocene period. The U.S. Dept. of Energy [2000c] further refined the

Flint et al. [1996] model and obtained a best-estimate infiltration rate of 4.7 mm/yrfor the 4.7

kin2 area above the proposed repository location using a MAP of 197 mm/yr to represent modem

climate. Note that the perched-water tapped by wells UZ-14 and WT-24 underlies approximately

the northern half of the proposed repository location. The U.S. Dept. of Energy [2000c] also

obtained a best-estimate infiltration rate of 19.8 mm/yr over the proposed repository location

D.\CV\Papers\ChenZhu et al\cmb_28-cz-jrw.doc 12/05/01

26

using a MAP of 323 mm/yr to represent a glacial-transition climate, which is comparable to the

MAP of 300 mm/yr used to represent late Pleistocene in this analysis.

Infiltration estimates obtained from numerical watershed modeling are subject to the

same uncertainties regarding long-term average precipitation rates. Uncertainty in the

meteorological boundary conditions (i.e., magnitude and timing of temperature and precipitation)

has been evaluated by the U.S. Dept. of Energy [2000c] and upper and lower bounds for net

infiltration have been estimated for present and potential future climates. As expected, the

infiltration estimates are sensitive to the assumed meteorological input, as are recharge estimates

obtained in this study using the CMB method. This uncertainty notwithstanding, the CMB and

mixing analysis used in this study provide useful validation of the numerical estimates of the

percentage of precipitation that becomes recharge.

5. Concluding Remarks

It appears that lower ET as a fraction of total precipitation or lower Cl- deposition during

the LGM when compared to today caused lower Cr concentrations in groundwater in the N

aquifer of Black Mesa, Arizona, as reported in this study. Similar conclusions have been reached

for the Aquia aquifer of Atlantic Coastal Plain, Maryland [Purdy et al., 1996], the Carrizo

aquifer, Texas [Stute et al., 19931, and the Great Artesian Basin, Australia [Love et al., 20001.

These records show the usefulness of groundwater as an archive of paleoclimate changes in a

wide range of geographic and climatic regions. However, water-rock interactions, diffusion from

less permeable units containing higher Cl- pore waters, interaquifer leakage, and the dissolution

of evapontes all tend to smear this record. Thus, the paleoclimate CI- signature is only preserved

D.\CV\Papers\ChenZhu et al\cmb_28-cz-jrw doc 12/05/01

27

in aquifers under favorable tornditions, and the failure of many hquifers to show this temporal

trend may simply be due to additional Ci inputs.

The presence of lower CI" concentrations in groundwater recharged during the LGM

highlights the need to consider temporal variability when using the CMB method to support

water resource management decisions. Chloride concentrations in groundwater are a function of

both ET and CI atmospheric input, and quantifying the two effects can be challenging.

Atmospheric inputs of chloride most likely varied temporally amid the profound climate changes

from Pleistocene to Holocene (distances from the recharge area to the sea coast, atmospheric

circulations, position of the jet stream). Because data are often lacking, the common practice is to

use present-day measurements of precipitation and CI deposition rates to estimate recharge rates

from the CI concentrations in groundwater of late Pleistocene age. This study shows that using

the modem precipitation and effective CI concentration can result in an erroneous estimate of

paleorecharge from the CMB method.

Significantly higher meteoric 36 C1/Cl ratios in the late Pleistocene than in the Holocene

are recorded in packrat middens [Plummer et al., 1997]; from the deep vadoze zone in Nevada

[Tyler et al., 1996]; and I saturated zone water at Black Mesa, Arizona [Lopes and Hoffinan,

1997] and San Juan basin, New Mexico [Plummer, 1996]. This temporal variability is combined

with the temporal variations of 36C, production rates [Davis et al., 1998] to estimate different CU"

deposition rates for the late Pleistocene and the Holocene. For Yucca Mountain, the calculated

late Pleistocene CU deposition rate was 8% less than the Holocene rate. This greater modem

deposition rate is probably attributable to dust from desiccated lakes in the region during the

Holocene whereas these lakebeds were covered with water during LGM [Plummer, 1996].

Conversely, the CU deposition rate for Black Mesa was 50% higher in late Pleistocene than

D:\CV\Papers\ChenZhu et aI\cmb_28-cz-jrw.doc 12/05/01

28

Holocene. A likely cause for this higher Ce deposition during late Pleistocene is from the shifted

southern branch of the jet stream Because of the presence of the ice sheet, the jet stream was

split in two and the southern branch was positioned in the southwestern United States

[COHMAP, 1988].

The calculated CI deposition rates were used to estimate recharge rates from the CMB

method for Black Mesa and Yucca Mountain. For Black Mesa, the average recharge rate in the

late Pleistocene is 35 mm/yr and in the Holocene it is 9 mm/yr. At Yucca Mountain, the best

record for local recharge may come from perched water, because the saturated zone waters

beneath Yucca Mountain are of late Pleistocene ages and may represent the conditions in

recharge areas upstream of Yucca Mountain. However, perched water at Yucca Mountain is

most likely a mixture of late Pleistocene and Holocene water [Sonnethal and Bodvarsson, 1999.

Fabrvka-Martin et al., 2000). From 3 6 C1 Isotopic mixing calculations, it is estimated that the

perched water bodies tapped by wells WT-24 and UZ-14 are comprised of 56% and 32%

Holocene water, respectively. Recharge rates were then calculated to be 5 mm/yr in Holocene

and 15 mm/yr in late Pleistocene.

The mixing analysis for perched waters at Yucca Mountain highlights the difficulty often

encountered in obtaining water samples of sufficient quality for conducting such analyses.

Measurements of Cl- concentrations and 36Cl/Cl ratios from bailed water samples were quite

variable within the same well, whereas measurements from pumped samples were more

consistent. It is also necessary to have confidence that perched-water samples from different

wells to be used in mixing analyses are derived from either the same perched water body or from

a similar source of recharge.

D.\CV\Papers\ChenZhu et al\cmb_28-cz-jrw doc 12/05/01

29

The estimated recharge rates at both Black Mesa and Yucca Mountain are consistent with

independent recharge estimates. At Black Mesa, paleorecharge rates were estimated from

simultaneous calibration of numerical groundwater flow models to radiocarbon ages and water

level data [Zhiu et al., 1998; Zhu, 2000]. At Yucca Mountain, infiltration rates estimated from

numerical watershed modeling for modem climate conditions and for a cooler, wetter glacial

transition type climate yielded results similar to those of the CMB in terms of both the absolute

value of the recharge estimate and the percentage of precipitation that becomes recharge.

Acknowledgement: This paper documents work performed in part for the U.S. Nuclear

Regulatory Commission (NRC) by the Center for Nuclear Waste Regulatory Analyses under

contract NRC-02-97-009. This work was conducted on behalf of the NRC Office of Nuclear

Material Safety and Safeguards, Division of Waste Management. The paper is an independent

product and does not necessarily reflect the views or regulatory position of the NRC. We thank

Charles Jones for his insights regarding the isotopic mixing calculations. We also thank Cynthia

Dinwiddie for a thorough technical review and many excellent suggestions.

D.\CV\Papers\ChenZhu et aI\cmb_28-cz-jrw.doc 12/05101

30

References

Allison, G. B., and M. W. Hughes, The use of environmental chloride and tritium to estimate total recharge to an unconfined aquifer, Aust. J. Soil Res., 16, 181-195, 1978

Allison, G. B., A review of some physical, chemical, and isotopic techniques available for estimating groundwater recharge, in Proceedings of NATO ASI Series, Series C, 222. 4972, 1988.

Allison, G. B., G. W. Gee, and S. W. Tyler, Vadoze-zone techniques for estimating groundwater recharge in and regions, Soil Sci. Soc. Ant. J., 58, 6-14, 1994.

Andrews, J. N., W. M. Edmunds, P. L. Smedley, J. C. Fontes, L. K. Fifield, and G. L. Allen, Chlorine-36 in groundwater as a palaeoclimatic indicator: the East Midlands Triassic Sandstone aquifer (UK), Earth and Plan. Sci. Let., 122, 159-171. 1994

Bazuhair, A. S., and W. W. Wood, Chloride mass-balance method for estimating ground water recharge in and areas: examples from western Saudi Arabia, J. Hydrol., 186, 153-159, 1996.

Benson, L., and H. Klieforth, Stable isotopes in precipitation and ground water in the Yucca Mountain region, southern Nevada: paleoclimatic implications, in Geophysical Monograph 55, pp. 41-59, American Geophysical Union, 1989.

Brown, J. G., and J. H. Eychaner, Simulation of five ground-water withdrawal projections for the Black Mesa area, Navajo and Hopi Indian Reservations, Arizona, Water-Revour. hives. Rep. 88-4000, U. S. Geol. Surv., 1988.

Clark, I., and P. Fritz, Environmental Isotopes in Hvdrogeology, Lewis Publishers, New York, 1997.

COHMAP Members, Climate changes of the last 18,000 years: observations and model simulations, Science, 241, 1043-1052, 1988.

Cooley, M. E., J. W. Harshbarger, J. P. Akers, and W. F Hardt, Regional hydrogeology of the Navajo and Hopi Indian Reservations. Arizona, New Mexico, and Utah, Profersional Paper 521-A U. S. Geol. Surv., 1969.

Craig, H., Isotopic variations in meteoric water, Science, 133, 1702-1703, 1961.

Davis, S. N., D. Cecil, M. Zreda, and P. Sharma, Chlonne-36 and the initial value problem, Hydrogeo. J., 6, 104-114, 1998.

Davis, S. N., J. Fabryka-Martin, L. Wolfsberg, S. Moysey, R. Shaver, C. Alexander Jr., and N. Kroth, Chlorine-36 in ground water containing low chloride concentrations, Ground Water, 6, 912-921, 2000.

D \CV\Papers\ChenZhu et aI\cmb_28-cz-jrw doc 12/05/01

31

Dettinger, M. D., Reconnaissance estimates of natural rechargeto desert basins in Nevada, U.S.A., by using chloride-balance calculations,J. Hydrol., 106, 55-78, 1989.

Domenico, P. A., and R. W. Schwartz, Physical and Chemical Hydrogeology, John Wiley & Sons, Inc.' New York, 1998. 1

Erikson, E., and V. Khunakasem, Chloride concentration in groundwater, recharge rate and rate of deposition of chloride in Israel Coastal Plain, J. Hydrol., 7, 178-197, 1969.

Eychaner, J. H., Geohydrology and effects of water use in the Black Mesa area, Navajo and Hopi Indian Reservations, Arizona, Water-Supply Pap. 2201, U. S. Geol. Surv., 1983.

Fabryka-Martin, J., A. V. Wolfsberg, P. R. Dixon, S. Levy, J. Musgrave, and H. J. Turin, Summary report of chlorine-36 studies: sampling, analysis and simulation of chlorine-36 in the Exploratory Studies Facility, Milestone Report 3783M Los Alamos Natl. Lab., Los Alamos, NM, 1997.

Fabryka-Martin, J. T., H. J. Turin, A. V. ,Woisberg, D. L. Brener, P. R. Dixon, and'J. A. Musgrave, Summary report of chlorine-36 studies as of August 1996, LA-13458-MS, Los Alamos Natl. Labs., Los Alamos, NM, 1998.

Fabryka-Martin, J., A. Meijer, B. Marshall, L. Neymark, J. Paces, J. Whelan, and A. Yang, Analysis of geochemical data for the unsaturated zone, ANL-NBS-HS-000017, Office of Civilian Radio. Waste Manag., 2000.

,d Faure, G., Principles of Isotope Geology, 2n edition, John Wiley and Sons, Inc., New York, 1996.

Feth, J. H., Chloride in natural continental water-A review, Water-Supply'Pap. 2176, U. S. Geol. Surv., 1981.

Flint, A. L., J. A. Hevesi, and L. E. Flint, Conceptual and Numerical Model of Infiltration for the Yucca Mountain Area, Nevada, Water Resour. Inves. Rep. Milestone 3GU1623M, U. S. Geol. Surv., Denver, Co, 1996.

Gee, G. W., and D. Hillel, Groundwater recharge in arid regions: review and critique of estimate methods, Hydrol. Processes, 2, 255-266, 1988.

GeoTrans, Inc., A two-dimensi6nal, finite-difference flow model simulIating the effects of withdrawal to the N aquifer, Black Mesa area, Arizona, Submitted to the Office of Sufface 'Mining on behalf of Peabody Western Coal Comjpany, 1987.

Harrison, S. P., K.E. Kohfeld, C. Roelandtl, T. Claquin, The role of dust in chmatechariges today, at the last glacial maximum and in the future, Earth-Science Review, 54, 43-80, '2001. -

D'\CV\Papers\ChenZhu et al\cmb_28-cz-jrw doc 12/05/01

32

Hendrickx, J., and G. Walker, Recharge from precipitation, in Recharge of Phreatic Aquifers in (senn) Arid Areas, edited by I. Simmers, pp. 19-98, A.A. Balkema, 1997.

Herczeg, A. L., and W. M. Edmunds, Chapter 2 Inorganic Tracers, in Environmental Tracers in Subsurface Hydrology, edited by P. G. Cook and A. L. Herczeg, pp. 31-78,Kluwer Academic Publishers, Norwell, Mass, 2000.

Junge, C. E., and R. T. Werby, The concentration of chloride sodium, potassium, calcium, and sulfate in rain water over the United States, J. Meteor., 15, 417, 1957.

Landwehr, J. M., T. B. Coplen, K. R. Ludwig, I. J. Winograd, and A. C. Rigs, Data from Devils Hole core DH- 11, Open-File Report 97-792 U. S. Geol. Surv., Reston, VA, 1997

Lerner, D. N., A. S. Issar, and I. Simmers, Groundwater Recharge -- A guide to understanding and estimating natural recharge, Internat. Contrib. Hydrol., 8, 345, 1990.

Lichty, R. W., and P. W. McKinley, Estimates of Ground-Water Recharge Rates for Two Small Basins in Central Nevada, Water-Res. Inves. Report 94-4104, U. S. Geol. Surv., 1995.

Lopes, T. J., and J. P. Hoffman, Geochemical analysis of ground-water ages, recharge rates, and hydraulic conductivity of the N aquifer, Black Mesa area, Arizona, Water-Resour. Inves. Rep. 96-4190, U.S. Geol. Surv., 1997.

Love, A. J., A. Herczeg, L. Sampson, R. Cresswell, and L. Fifield, Sources of chloride and implications for 36C1 dating of old groundwater, southwestern Great Artesian Basin, Australia, Water Resour. Res., 36, 1561-1574, 2000.

Mayewski, P A., L. Meeker, S. Whitlow, M. Twickler, M. Morrson, P. Bloomfield, G. Bond, R. Alley, A. Gow, P. Grootes, D. Meese, M. Ram, K. Taylor, and W. Wumkes, Changes in atmospheric circulation and ocean ice cover over the north Atlantic during the last 41,000 years, Science, 263, 1747-175 1, 1994.

Meijer, A., and E. Kwicklis, Geochemical and isotopic constraints on groundwater flow directions, mixing, and recharge at Yucca Mountain, Nevada, ANL-NBS-HS-000021, Office of Civilian Radi. Waste Manag., 2000.

Merlivat, L., and J. Jouzel, Global climatic interpretation of the deuterium-oxygen-18 relationship for precipitation, J. Geophys. Res., 84, 5029-5033, 1979.

Oliver, T., and T. Root, Hydrochemical database from the Yucca Mountain area, Nye County, Nevada, Memorandum to the Technical Project Officer, 9 pp. with figures and Excel database, U.S. Geol. Surv., ACL: MOL., 19980302.0367, 1997.

Patterson, G., Chemical, and isotopic data from precipitation sample collected behind service station in Area 25 and groundwater samples collected at boreholes UE-25 c#2, UE-25

D:\CV\Papers\ChenZhu et al\cmb_28-cz-irw doc 12/05/01

33

c#3, USW UZ-14, UE-25 WT #3, UE-25 WT#17 and USW WT-24, 10/06/97 to

07/01/98, U.S. Geological Survey, 1998.

Phillips, F. M., Environmental tracers for water movement in desert soils of the American

southwest, Soil Sci. Soc. Am. J., 58, 15-24, 1994.

Phillips, F. M., Chapter 10 Chlorine-36, in Environmental Tracers in Subsurface Hydrology, edited P. G. Cook and A. L. Herczeg, pp. 299-348, Kluwer Academic Publisher, Norwell, Mass., 2000.

Plummer, M. A., Secular variation of cosmogenic nuclide production from chlorine-36 in fossil pack rat middens, Masters Thesis, New Mexico Institute of Mining and Technology, Socorro, NM, 1996.

Plummer, M. A., F. M. Phillips, J. Fabryka-Martin, H. J. Turin, P. E. Wigand, and P. Sharma, Chlorine-36 in fossil rat urine: An archive of cosmogenic nuclide deposition during the past 40,000 years, Science, 277, 538-541, 1997.

Purdy, C. B., G. R. Helz, A. C. Mignerey, P. W. Kubik, d. Elmore, P. Sharma, and T. Hemmick, Aquia aquifer dissolved ClI and 36C1/Cl: Implications for flow velocities, Water Resour. Res., 32(5), 1163-1171, 1996.

Scanlon, B. R., Uncertainties in estimating water fluxes and residence times using environmental tracers in an arid unsaturated zone, WaterResour. Res., 36(2), 395-409, 2000.

Sonnethal, E. L., and G. S. Bodvarsson, Constraints on the hydrology of the unsaturated zone at Yucca Mountain, NV from three-dimensional models of the chloride and strontium geochemistry, J. Cont. Hydr., 38, 107-156, 1999.

Steinkampf, W. C., Field, chemical, and isotopic data describing water samples'collected in Death Valley National Monument and at various boreholes in and around Yucca Mountain, Nevada, between 1992 and 1995, 1995.

Stute, M., J. F. Clark, R. M. Phillips, and D. Elmore, Reconstruction of late glacial climate from

the groundwater archive: CI" and 36C1 in the Carrizo Aquifer, .Texas, in-Isotope ' Techniques in the Study of Past and Current Environmental Changes in the Hfrdrosphere

and the Atmosphere, Proceedings of An International Symposium on Applications of Isotope Techniques in Studying Past and Current Environmental Changes in the Hydrosphere and the Atmosphere, pp.'259-270, IAEA, Vienna; 1993.

Thomas, B. E., and M. Truini, Ground-water, surface-water, and water-chemistry data, Black

Mesa area, Northern Arizona-1999, Open-File Report 00-453, U. S. Geol. Surv., 2000.

Thompson, R. S., C. Whitlock, P. J. Bartlein, S. P. Harrison, and W. G. Spaulding, Climatic changes in the western United States since 18,000 years ago, in Global climate since the

D.\CV\Papers\ChenZhu et al\cmb_28-cz-jrw.doc 12/05/01

34

last glacial maxinium, edited by H. E. Wright and K. Kutzbach, University of Minnesota Press, Minneapolis, Minn., 1993.

Tyler, S. W., J. B. Chapman, S. H. Conrad, D. P. Hamermeister, D. 0. Blout, J. J. Miller, M. J Sully, and J. M. Ginanni, Soil-water flux in the southern Great Basin, United States: temporal and spatial variations over the last 120,000 years, Water Resour. Res., 32, 148 11499, 1996.

U.S. Dept. of Energy, Future Climate Analysis, Civilian Radioactive Waste Management, Management and Operating Contractor. Analysis/Model Report ANL-NBS-GS-000008. 2000a

U.S. Dept. of Energy, Unsaturated Zone Flow and Transport Model Process Model Report, Civilian Radioactive Waste Management, Management and Operating Contractor. Process Model Report TDR-NBS-HS-000002, REVO0, ICN02. 2000b.

U.S. Dept. of Energy, Simulation of Net Infiltration for Modern and Potential Future Climates, Civilian Radioactive Waste Management, Management and Operating Contractor. Analysis/Model Report ANL-NBS-HS-000032, REVO0. 2000c.

Winograd, I. J., J. M. Landwehr, K. R. Ludwig, T. B. Coplen, and A. C. Riggs, Duration and structure of the past four interglaciations, Quat. Res., 48, 141-154, 1997.

Wood, W. W., and W. E. Sanford, Chemical and isotopic methods for quantifying groundwater recharge in a regional, semiarid environment, Ground Water, 33(3), 456-468, 1995

Wood, W. W., Use and misuse of the chloride mass balance method in estimating ground water recharge, Ground Water, 37(1), 2-3, 1999.

Yang, I. C., G. W. Rattray, and P. Yu, Interpretations of chemical and isotopic data from boreholes in the unsaturated-zone at Yucca Mountain, Nevada, Water-Resources Investigation Report 96-4058, U.S. Geol. Surv., Denver, CO, 1996.

Zhu, C., R. K. Waddell, I. Star, and M. Ostrander, The responses of groundwater in the Black Mesa basin, northeastern Arizona to paleoclimatic changes during late Pleistocene and Holocene, Geol., 26, 127-130, 1998.

Zhu, C., Estimate of recharge from radiocarbon dating of groundwater and numerical flow and transport modeling, Water Resour. Res., 36, 2607-2620, 2000.

D \CV\Papers\ChenZhu et al\cmb_28-cz-jrw.doc 12/05/01

35

Table 1. Calculations of C6" deposition rates

Parameters 3Cl/Cl P D 36 DcO Cleff

(x 1015) (mmlyr) (atoms m-2 s"1) (mg m-2 yr-1) (mg/l)

Yucca Mountain

Holocene* 500 172 16 60 0.35

Pleistocene 1,000 300 30 55 0.18

Black Mesa

Holocene 765 305 22 54 0.18

Pleistocene 1000 538 44 82 0.15

*From Fabryka-Martin et al. [2000]

D:\CV\Papers\ChenZhu et al\cmb_28-cz-jrw.doc 12/05/01

36

Table 2 Chemical and isotopic data for saturated zone and perched water at Yucca Mountain VSMOW refers to Vienna Standard Mean Ocean Water

WellW ClBr 3-6Cl/CI 8D 8150 Recharge Veil Sampling CI" BrMIass .x 10s SOW VSMOW mm/yr Ref's

Identifier Abbrediation Date mg/L mg/L Ms 05VIIwvI~W m~r Rf Ratio Ratio per mil per mil Holocene Pleistocene

UE-25 WT#3 WT#3 Average 6 0 05 123 - -102 1* -13 6* - 9 I

UE-25 WT #12 WT#12 19880412 78 008 102 571" -102.5 -13.75 - 7 1 2 UE-25 WT-17 WT-17 Average 64 078 8 - -1019* -13.7* S 14

USW G-2 G-2 19960423 65 0 04 155 525* -98 8 -13 33 - 8 1.2 USW G-4 G-4 19821209 5.9 - - - -103 -13 8 -9

USW H-3 H-3 19840314 55 0.18 53 - -101 -139 - 6 3

USW H-4 H-4 19820517 69 - - - -104 -14 - 8 ! Saturated

USW H-5 c H-5 Average 61 - - - -102 -136 -9 1 Zone USW H-I (Tcp) H-I (Tcp) 19801020 57 - - - -103 -134 - 9 Water USW H-I (Tcb) H-I (Tcb) 19801208 5 8 - - - -101 -135 -9 1 UE-25 b#1 (bh) b#1 (bh) 19810901 8 5 - - - -101 -134 - 5 3

UE-25 b#1 (Tcb) b#1 (Tcb) 19820720 75 - - - -99.5 -135 - 7 3

UE-25 c#1 c#1 19830930 74 003 228 - -102 -13.5 - 7 1,3 UE-25 c#2 c#2 Average 7 1 016 44 - -100 -134 - 8 1.34

UE-25 c#3 c#3 Average 7.2 0 16 45 500* -103 -135 - 8 1,35 c#3('95) c#3('95) 19950523 65 - - - 997 -1338 - 8 1

ower Carbonate UE-25 p#1 p#1 19900606 253 0 09 284 132 -106 -138 - - I Aquifer

SD-7 SD-7 19950316 4 1 005 87 - -997 -134 -

SD-7 SD-7 19950317 4.1 005 87 657 -996 -134 -

SD-7 SD-7 19950320 4 1 005 87 - -996 -134 -

SD-7 SD-7 19950321 4 1 005 87 635 -996 -13.3 -"Perched UZ-14 PT-4 UZ-14 19930827 67 008 87 675 -973 -134 5 15 1.2 Vater UZ-14 PT-1 19930817 72 644 -978 -133

UZ-14 PT-2 19930819 67 675 -97.3 -134

UZ-14 D 70 19930831 70 690 -976 -13 1

USW WT-24 WT-24 19971022 9 0 586 -994 -135 5 15

(1) Meijer and Kwtcklis [2000]; (2) Fabr~ka-Martn et al. [1998], (5) Stenkampf[ 1995].

[ 1998]; (3) Oliver and Root [ 1997]; (4) Patterson

D \CV\Papers\ChenZhu et al\cmb_28-cz-jrw.doc 12/05/01

37

Figure Captions

Figure 1. Location and hydrogeological settings of the Black Mesa basin. Contours of present

day hydraulic heads (in feet). Arrows indicate the directions of groundwater flow.

Figure 2. Location of Yucca Mountain, Nevada. Yucca Mountain is the only site currently under

investigation as a potential permanent geological repository for high-level nuclear waste

in the USA.

Figure 3. (a) Variations of CI" in the GISP2 ice cores in late Pleistocene and Holocene

[Mayewski et al., 1994]. (b) Variation of ClI concentrations in groundwater at Black Mesa

(circles), and 36C1/C1 ratios at Black Mesa (solid triangles) and San Juan basin (open

triangles) through late Pleistocene and Holocene. (c) Recharge rates estimated for Black

Mesa (circles) using equation (3) and parameters in Table 1 (note that four points with

values > 50 mm/yr are plotted off the chart, highest value is 95 mm/yr); solid line

indicates normalized recharge estimated from numerical modeling calibration to 14C age

distribution [Zhit et al., 1998]. (d) Estimated percentages of precipitation that become

recharge (circles).

Figure 4. Cross-section of the Black Mesa basin. N and D stand for the N and D aquifer. The

silty Carmel Formation hydraulically separates the N and D aquifers. The D aquifer is

overlain by the Mancos Shale. Modified from Cooley et al. [1969].

Figure 5. Cl concentrations in groundwater from Black Mesa compared to (a) 8D, and (b) •I80

values of the same samples.

Figure 6. Recharge rates estimated from the saturated zone water ClI concentrations. MKI and

MK2 stand for rates estimated by Meijer and Kwicklis [2000] using values of 0.3 and 0.6

mg/L, respectively, for effective Holocene Cli concentrations.

DACV\Papers\ChenZhu et aI\cmb_28-cz-jrw.doc 12/05/01

38

Figure 7. H and 0 isotope data for perched water, saturated zone water beneath Yucca

Mountain, and Fortymile Canyon. The global meteoric water line is from Craig [1961],

and the local meteoric water line is from Benson and Klieforth [1989].

Figure 8 The relationship between 36C1/CI and reciprocal of Cl- concentrations in perched

bodies. Open symbol show bailed samples.

D \CV\Papers\ChenZhu et aI\cmb_28-cz-jrw.doc 12/05/01

Figure 1

Study A . Area

0 25 50

Kilometers

Figure 2

Arizoni

Figure 3

1000

"100 04

S10

35,000 30,000 25,000 20,000 15,000 10,000 5,000 0

Years (B.P.)

000 35,000 30,000 25,000 20,000 15,000 10,000 5,000

35,000 30,000 25,000 20,000 15,000 10,000 5,000 0

0 d0

00

0

rech2000-b.xlsfigl

0001'

40,C

0.60

0.50

_ 0.40 0 E 0.30 E L-- 0.20

"0.10

0.00 40,(

b AL• /A •A A

SdD o • _ee o o

0 e° A

1400

1200

1000!-O

800 x

600

400 g 200

0

100

80

'60

40

E.

E

4)

0

8 7

6

5 4

3 2

0

.0 � 4- 02

� �a)

x2400 0 40,000

0 •E15% 0

S10%

22 5% f.0 0

Sno,.

40,000

6

5

4

3

2

1

0

Cc

x

35,000 30,000 25,000 20,000 15,000 10,000 5,000

Corrected C-14 ages of groundwater (years B.P.)

Figure 312/l/20014:41 PM

I

Figure 4

A A Northwest Southeast

U) Meters 0

0 0 2,400 "

1,0 00

2,100 - Carmel o

1,,800--- -

1,200 - . ... -..... .. N .. ..............n

Vertical spate is exaggeratea 4 10

Scale

lometers -

20 miles

Figure 5

-60 _______________

a. open symbols > 11 ka

70 aSolid symbols < 11 ka

> @8

E, 0 .B ; -90 ES

L.D

00

-110 • 0 0 000

-120

000 0.10 020, 0.30 040 050 Cl (mmol/L)

-8

b. Open symbols > 11 ka

9Solid symbols < 11 ka

AA A -.10

O0 A A - A AAA

>.1

A A

a-x 0 ~-13

-14

-15

000 010 0.20 0.30 040 050

CI (mmol/L)

- Figure 5rech2000-b.xlsH&OACL10/812001

1C1

LL 0

0 :E 0

c#3'

-CflcA wc#2

c#3

b#1 (Tcb)

b#1 (bh)

H-1(Tcp)

HA5

H-3

....... G-2

........ W -17

--- IW#12

" ý ........ WT#

0C) CO (D C,4 C) Co co -t J 0

C\I - ,- - -~ q~

c'i u,

-o

Figure 7

-85

-90 O UZ-14

CO) > -95 b#(Tcb WT-24

L- -100 SD-7

% -0 -b- 'A Perched 0 0 # UE-25 p#1

-o 105 C#3 H-1 (Tcp) UW# . P-29a#2

ill2ý WT'1 7 WT#3 & H5 -"•"•,traig MWL

-110 A . I . -LoMWL .

-14.5 -14 -13.5 -13 -12.5, -12 5180 (per mil) VSMOW

TAble2-cz.xls Figure 7 110/812001

Figure 8

1000

Pleistocene900

800

rvv w WT-24 156% Holocene

600

500 - u z. SHolocene

4 400 r UZ-14 C

UZ-14 D UZ-14 PT-4 32% Holocene

SUZ-14PT-2 UZ-14 PT1

"•r• uz-14 A

-"4A2 ONRG-7A "+ NRG-7a )KSD-9

3000 05 01 015 02

(I I [Cl]) m L/mg

CImwxng_zhu.xls

0 1E

i SD-7 SD-7

ASD-7

Abailed

Apumped ,

0 25 03

Figure 8 12/4/2001