W

Copyright

by
IE
Bradley Donald Cey

2008
EV
PR
The Dissertation Committee for Bradley Donald Cey Certifies that this is the
approved version of the following dissertation:

Dissolved Noble Gases in Groundwater

W
Committee:
IE
Bridget R. Scanlon, Supervisor
EV

Philip C. Bennett

John M. Sharp
PR

Zong-Liang Yang

G. Bryant Hudson
 Dissolved Noble Gases in Groundwater

by

Bradley Donald Cey, B.Sc., M.Sc.

W
Dissertation
IE
Presented to the Faculty of the Graduate School of

The University of Texas at Austin

EV
in Partial Fulfillment

of the Requirements

for the Degree of

PR

Doctor of Philosophy

The University of Texas at Austin

May 2008
 3324840

W
IE
EV
PR

3324840
2008
 Acknowledgements

Financial assistance for my studies was provided by the Jackson School of

Geosciences, the Glenn T. Seaborg Institute at Lawrence Livermore National Laboratory,

W
California State Water Resources Control Board Groundwater Ambient Monitoring and

Assessment Program, and the Geological Society of America. The funding provided by
IE
these organizations was critical to the completion of this degree and is gratefully

appreciated.
EV
Thank you to the landowner of the Kings County field site for generously

allowing access for this study. I have enjoyed my studies at UT in large part because of

the people I have had the pleasure of meeting and working with. Thank you to my
PR

dissertation committee members Phil Bennett, Jack Sharp, and Liang Yang, for their

insightful assistance. I am thankful for the opportunity to have worked on an interesting

and challenging research project with scientists from Lawrence Livermore National

Laboratory. I thank Jean Moran, Bryant Hudson, Brad Esser, Mike Singleton, and Steve

Carle for allowing me to join them in their research. Special thanks to Mike, Bryant, Gail

Eaton, Wayne Culham, and Bob Reedy for assistance in the lab. Thank you to Jean for

her support and encouragement throughout the course of my research. I am especially

grateful to Bryant for his willingness to serve on my dissertation committee and for

sharing his enthusiasm for science and knowledge of noble gases with me.

iv
 Thank you to Gerald Flerchinger for allowing me to use the SHAW numerical

modeling code, and to Lindsey Gulden and Enrique Rosero for assistance with accessing

data and computer issues.

Thank you to Jack Sharp for his assistance during with my application to UT.

Without Jack’s persistence, it is unlikely that I would have come to study at UT.

I am grateful to Bridget Scanlon for supervising my dissertation. She believed in

me throughout—during the successes and during the challenges. Thank you so very

much for your enthusiastic encouragement.

Most importantly, thank you to my family. To Anna and Clare for their love and

W
their zest for life. To my wife Misty, for her trust, love, encouragement, and giving so

freely of herself to allow me to complete this work.

IE
EV
PR

v
 Dissolved Noble Gases in Groundwater

Publication No._____________

Bradley Donald Cey, Ph.D.

The University of Texas at Austin, 2008

Supervisor: Bridget R. Scanlon

W
Atmospheric noble gases (He, Ne, Ar, Kr, and Xe) dissolved in groundwater are a
IE
valuable tool in hydrology. Numerous studies have relied on groundwater recharge

temperatures calculated from dissolved noble gas data (noble gas temperatures, NGT) to
EV
infer paleoclimate conditions. This research investigated gas dissolution during

groundwater recharge and critically examined the use of dissolved noble gas data in

groundwater research. A detailed investigation of an agriculturally impacted shallow

PR

aquifer allowed comparison of measured water table temperatures (WTT) with calculated

NGT. Results suggest that NGT calculated from widely used noble gas interpretive
models do reflect measured WTT, supporting the use of dissolved noble gases to deduce

recharge temperatures. Samples having dissolved gas concentrations below the

equilibrium concentration with respect to atmospheric pressure were attributed to

denitrification induced gas stripping in the saturated zone. Modeling indicated that minor

degassing (<10% ΔNe) may cause underestimation of groundwater recharge temperature

by up to 2 °C. In another study a large dissolved noble gas data set (905 samples) from

California was analyzed. Noble gas modeling using the same interpretive models

vi
indicates that multiple models may fit measured data within measurement uncertainty,

suggesting that goodness-of-fit is not a robust indicator of model appropriateness. A

unique aspect of this study was the high Ne and excess air concentrations associated with

surficial artificial recharge facilities. A final study examined whether climatic/hydrologic

changes occurring over glacial-interglacial time periods could impact the accuracy of

NGT used in paleoclimate studies. Numerical modeling experiments estimated WTT

sensitivity to changes in: 1) precipitation amount, 2) water table depth, and 3) air

temperature. Precipitation and water table depth had a minor impact on WTT (~0.2 °C).

In contrast, the impact of air temperature changes on WTT was more pronounced.

W
Results suggest that air temperatures inferred from NGT data may underestimate actual

air temperature change since the last glacial maximum by ~1 °C at sites having seasonal
IE
snowcover. These results suggest despite uncertainty in the exact physical processes

controlling gas dissolution during groundwater recharge, NGT do reflect WTT.

EV
However, inferring paleo-air temperatures from NGT are subject to error, especially

locations with seasonal snowcover.

PR

vii
 Table of Contents

List of Tables ...........................................................................................................x

List of Figures ........................................................................................................ xi

Chapter 1 : Introduction .........................................................................................1

Overview.........................................................................................................1
Research Questions.........................................................................................2
Hypotheses and Research Approach...............................................................3

Chapter 2 : Noble Gas Recharge Temperature Studies at an Agricultural Site in

California ........................................................................................................5

W
Introduction.....................................................................................................6
Materials and Methods....................................................................................8
IE
Study Site ...............................................................................................8
Noble Gas Modeling ............................................................................11
Results and Discussion .................................................................................16
EV
Soil Gas Pressure .................................................................................16
Temperature .........................................................................................16
Oxygen Isotope Data (δ18O) ................................................................17
PR

Dissolved Noble Gases ........................................................................18

Undersaturation...........................................................................18
Noble Gas Temperatures.............................................................20
Modeling Degassing ............................................................................23
Conclusions...................................................................................................25

Chapter 3 : Impact of Artificial Recharge on Dissolved Noble Gases in Groundwater

in California ..................................................................................................45
Introduction...................................................................................................46
Materials and Methods..................................................................................48
Overview of Study ...............................................................................48
Sampling and Analysis ........................................................................49
Modeling ..............................................................................................50

viii
 Results and Discussion .................................................................................51
Spatial Characterization .......................................................................51
Noble Gas Models................................................................................53
Excess Air ............................................................................................54
3
H–3He Age-dating ..............................................................................56
Recharge Temperature .........................................................................57
Artificial Recharge...............................................................................58
Conclusions...................................................................................................60

Chapter 4 : On the Accuracy of Noble Gas Paleotemperatures over Glacial-

Interglacial Periods .......................................................................................76
Introduction...................................................................................................77

W
Methods.........................................................................................................80
Model Description ...............................................................................80
Numerical Experiments .......................................................................82
IE
Results...........................................................................................................83
Base Case .............................................................................................83
EV
Precipitation .........................................................................................84
Water Table Depth...............................................................................85
Temperature .........................................................................................85
Discussion .....................................................................................................87
PR

Conclusions...................................................................................................89

Chapter 5 : Conclusions and Outlook ................................................................105

Appendix A : Laboratory Soil Testing Results from Kings County Field Site .108

Appendix B : Field Measurements from Kings County Field Site....................118

Appendix C : Dissolved Noble Gas Data and NOBLE90 Modeling Results ....125

Appendix D : SHAW Model Input Files ...........................................................158

Bibliography ........................................................................................................166

Vita ....................................................................................................................179

ix
 List of Tables
Table 2.1. Well depths, sample collection dates, and selected measured data (originally
published in McNab et al., 2007; Singleton et al., 2007).................................................. 27

Table 2.2. Dissolved noble gas data (originally published in Singleton et al., 2007)...... 29

Table 2.3. Results of noble gas modeling. Modeling was done using He, Ne, Ar, Kr, and
Xe for the UA, PR, and CE models. Additional modeling was done using only Ne, Ar,
Kr, and Xe with the CE model. Results rejected because of poor fitting (i.e. p < 0.05) are
not shown. ......................................................................................................................... 31

Table 3.1. Number of samples fit by each model for a given region according to criteria
discussed in the Supporting Information. AR refers to artificial recharge impacted areas
(i.e. SFBA, LAB, and Bakersfield)................................................................................... 62

W
Table 3.2. Statistical summary of modeled excess air concentrations (% ΔNe). AR refers
to artificial recharge impacted areas (i.e. SFBA, LAB, and Bakersfield). ....................... 63
IE
Table 3.3. Statistical summary of calculated recharge temperatures (°C). AR refers to
artificial recharge impacted areas (i.e. SFBA, LAB, and Bakersfield)............................. 64

Table 4.1. Soil properties used in SHAW model (from Rawls and Brakensiek, 1989)... 90
EV

Table 4.2. Soil thermal properties used in SHAW model................................................ 91

Table 4.3. Meteorological forcing data used in the Base Case scenario.......................... 92
PR

x
 List of Figures
Figure 2.1. Map of study site. Only the sampled irrigation wells are uniquely labeled.
Irrigation wells owned by other landowners are not shown. ............................................ 35

Figure 2.2. Monitor well depths. Ground surface at each monitor well location shown as
horizontal black line, screened intervals shown in gray. Elevation in meters above mean
sea level............................................................................................................................. 36

Figure 2.3. Atmospheric pressure data from nearby National Climatic Data Center
(NCDC) station showing a) seasonal fluctuations and (b) diurnal fluctuations. The
dashed line in (b) is 3 day moving average. Panel (c) shows the impact of the May 22,
2005 irrigation event at 2S on soil gas pressure (site 2S sensors – gray, sites 3S and 5S –
black)................................................................................................................................. 37

Figure 2.4. Subsurface temperature data from 3S location. Local water table fluctuated

W
between 3.8 and 5.5 m below ground surface (BGS) during the study. Subsurface
temperature data from the other instrumentation locations (2S and 5S) are similar. Air
temperature data are from the nearby National Climatic Data Center (NCDC) station. .. 38
IE
Figure 2.5. Water table temperatures for the three instrumented locations. Curves for 3S
and 5S are data from the deepest heat dissipation sensor. The lower 2006 maximum
temperature at 5S is attributed to increased vegetation cover in 2006. The curve for 2S is
EV
extrapolated from measured data assuming exponential decay of the seasonal temperature
signal with depth (water table was ~1.1 m below the deepest 2S sensor). ....................... 39

Figure 2.6. Oxygen isotope (δ18O) data reported as parts per thousand (‰) relative to
Vienna Standard Mean Ocean Water (VSMOW)............................................................. 40
PR

Figure 2.7. Helium versus Ne concentrations of undersaturated samples. Equilibrium

concentration given for T = 19 °C, S = 0, p = 0.991 atm. Lines show impact of degassing
for each of the three degassing models. ............................................................................ 41

Figure 2.8. Ne and Xe concentrations of samples from the shallowest wells. Analytical
uncertainties shown (Ne 2%, Xe 3%). Equilibrium solubilities from T = 17–20 °C (S = 0,
p = 0.991 atm) are shown. Unfractionated excess air concentrations are also shown..... 42

Figure 2.9. Noble Gas Temperatures (NGTs) calculated for samples from the shallowest
wells for four different models: Unfractionated Air (UA), Closed system Equilibrium
(CE), Partial Re-equilibration (PR), and Diffusive Degassing (DD). The shaded region
indicates the water table temperature (WTT) range for that location. For wells with
multiple samples, the sample number is noted. ................................................................ 43

Figure 2.10. Impact of degassing on calculated NGTs. These synthetic samples had
excess air added (ΔNe = 30%), some unfractionated and others fractionated according to
the CE model (both F = 0.65 and 0.75). Samples were then degassed by various amounts

xi
according to the DS1 and DD models. The resultant gas concentrations were then
modeled by the CE and UA models. Panels (a) and (b) show the difference between
original recharge temperature (19 °C) and modeled recharge temperature after degassing.
Panels (c) and (d) show model goodness-of-fit (probability of χ2 being greater than a
given value obtained from the χ2 distribution for the appropriate number of degrees of
freedom). Panel (e) shows the modeled CE fractionation parameter, F (note that the UA
model is a limiting case of the CE model in which F = 0). .............................................. 44

Figure 3.1. Map of California showing sample locations in each of the six regions: Los
Angeles Basin (LAB), Mojave Desert Basin (MDB), northern California (NC), northern
portion of the Central Valley (NCV), southern portion of the Central Valley (SCV), and
San Francisco Bay Area (SFBA). ..................................................................................... 65

Figure 3.2. Histogram of Ne concentrations for: literature data (Aeschbach-Hertig et al.,

2002c; Andrews et al., 1991; Andrews et al., 1994; Beyerle et al., 1998; Beyerle et al.,
2003; Clark et al., 1998; Clark et al., 1997; Dennis et al., 1997; Fontes et al., 1991; Hall

W
et al., 2005; Kulongoski et al., 2004; Ma et al., 2004; Saar et al., 2005; Stute et al., 1995a;
Stute and Deak, 1989; Stute et al., 1995b; Stute and Sonntag, 1992; Thomas et al., 2003;
Weyhenmeyer et al., 2000; Wilson et al., 1990; Wilson et al., 1994; Zuber et al., 2004;
IE
Zuber et al., 2000), artificial recharge impacted areas (SFBA, LAB, and Bakersfield), and
non-artificial recharge impacted areas. ............................................................................. 66

Figure 3.3. CE modeled versus measured gas concentrations. Note that sample 100706
EV
plots off scale for Kr. ........................................................................................................ 67

Figure 3.4. PR modeled versus measured gas concentrations. Note that sample 100706
plots off scale for Kr. ........................................................................................................ 68
PR

Figure 3.5. UA modeled versus measured gas concentrations. Note that sample 100706
plots off scale for Kr. ........................................................................................................ 69

Figure 3.6. Scatter plots of model error (relative difference between modeled and
measured gas concentrations) for samples fit by the CE model. Results from the other
models are comparable. Note that analytical uncertainties are approximately 2% for Ne
and Ar, and 3% for Kr and Xe. ......................................................................................... 70

Figure 3.7. Element ratios for: 1) measured gas concentrations for artificial recharge
impacted areas (AR) and non-AR areas, and 2) total model predicted gas concentrations
(UA, CE, and PR). Measured ratios below the cluster of CE points can not be explained
by the UA or CE models and are consistent with diffusive degassing or other
fractionating processes. The UA model predicted ratios lie on the line shown. Not all
data are shown at this scale............................................................................................... 71

Figure 3.8. Relationship between CE model parameters and excess air. As excess air
increases, the degree of fractionation decreases. Note that the CE model reduces to the

xii
UA model when F = 0. One data point is not shown at the given scale (ΔNe = 737%,
F = 0.031, q = 5.8). ........................................................................................................... 72

Figure 3.9. Histogram of calculated 3H–3He ages for each of three models. .................. 73

Figure 3.10. Modeled recharge temperatures for samples fit by all three excess air
models. In general, the PR recharge temperature > CE recharge temperature > UA
recharge temperature......................................................................................................... 74

Figure 3.11. Median calculated recharge temperatures from each of the three models.
The measured data are mean annual air temperatures taken from the U.S. Historical
Climatology Network (http://www.ncdc.noaa.gov/ol/climate/research/ushcn/ushcn.html)
and are the medians of stations nearest each sample location. ......................................... 75

Figure 4.1. Schematic diagram showing processes modeled, input forcings, and
boundary conditions.......................................................................................................... 93

W
Figure 4.2. Input forcing data from year 2004 at 38.5° N, 97° W of the North American
Land Data Assimilation System (NLDAS) (Cosgrove et al., 2003). (a) Hourly
temperature data, and (b) daily precipitation totals. ......................................................... 94
IE
Figure 4.3. Profiles of mean annual soil temperature and saturation for the Base Case
scenario (mean annual air temperature, MAAT = 13.68 °C; mean annual water table
temperature, WTTloam = 14.66 °C; WTTsand = 15.40 °C). ................................................ 95
EV

Figure 4.4. Water table fluctuations of the Base Case scenarios (mean annual water table
depth of 3 m)..................................................................................................................... 96

Figure 4.5. Modeled time series of (a) temperature and (b) saturation at various depths
PR

for the loam soil Base Case scenario (mean annual air temperature, MAAT = 13.68 °C;
mean annual water table temperature, WTTloam = 14.66 °C)............................................ 97

Figure 4.6. Response of water table temperature (WTT) relative to mean annual air
temperature (MAAT) to changes in precipitation amount................................................ 98

Figure 4.7. Amount of seasonal decoupling of ground surface temperature (GST) from
mean annual air temperature (MAAT) in response to changes in precipitation amount. . 99

Figure 4.8. Response of water table temperature (WTT) relative to mean annual air
temperature (MAAT) to changes in mean annual depth of water table below ground
surface. Mean annual water table depth was 3 m in the Base Case scenario. ............... 100

Figure 4.9. Mean annual soil temperature profiles for simulations in which the water
table depth was varied from 2–5 m. Mean annual water table depth was 3 m in the Base
Case scenario. ................................................................................................................. 101

xiii
Figure 4.10. Response of water table temperature (WTT) relative to mean annual air
temperature (MAAT) to changes in MAAT. MAAT was 13.68 °C in the Base Case
scenario. .......................................................................................................................... 102

Figure 4.11. Seasonal response of ground surface temperature (GST) relative to mean
annual air temperature (MAAT) for loam. ..................................................................... 103

Figure 4.12. Response of water table temperature (WTT) relative to mean annual air
temperature (MAAT) to changes in the annual amplitude of MAAT. ........................... 104

W
IE
EV
PR

xiv
 Chapter 1:

Introduction

OVERVIEW

Atmospheric noble gases (He, Ne, Ar, Kr, and Xe) dissolved in groundwater are a

valuable tool in hydrology. They have been used in hydrologic studies for over half a

century (Oana, 1957 as referenced in Kipfer et al., 2002). An early application was to

W
determine groundwater flow velocity based on seasonal variations in dissolved gas

concentration (Sugisaki, 1961). Mazor (1972) first recognized the potential to derive
IE
paleotemperature records from dissolved noble gases in groundwater. Since then,

numerous studies have relied on groundwater recharge temperatures calculated from

EV
dissolved noble gas data (noble gas temperatures, NGT) to infer paleoclimate conditions

(Kipfer et al., 2002; Stute and Schlosser, 1993). Dissolved noble gases in groundwater

are also important for determining the age of young groundwater (<50 yr) using the 3H–
PR

3
He technique (Solomon and Cook, 2000).

The sources of dissolved noble gases in groundwater can be either atmospheric or

non-atmospheric. Atmospheric sources include equilibrium concentration with respect to

atmospheric pressure (governed by Henry’s Law) and “excess air” (Heaton and Vogel,

1981). The excess air component is the result of dissolution of air bubbles trapped in the

saturated zone (Kipfer et al., 2002). Non-atmospheric sources can be either radiogenic

(e.g. 3He from 3H decay), terrigenic (e.g. primordial 3He from the mantle), or both (e.g.
4
He from U and Th decay in aquifer rocks). For many groundwater systems, the non-

atmospheric sources are much smaller than atmospheric sources (Lehmann et al., 1993).

1
 If non-atmospheric components are insignificant, equilibrium concentration with

respect to atmospheric pressure and excess air components can be distinguished by

measuring multiple gases because of the differences in gas solubility. According to

Henry’s Law, equilibrium concentration is a function of temperature, gas partial pressure,

and water salinity. In most noble gas studies, the recharge elevation is assumed which

allows the calculation of gas partial pressure (Kipfer et al., 2002). Also, the recharging

water is commonly assumed to be fresh. Therefore, the recharge temperature (i.e. NGT)

can be calculated from the dissolved gas concentration.

NGT are commonly used in paleoclimate studies of the last glacial maximum

W
(LGM). NGT are important because they are few other quantitative temperature proxies

for low-elevation terrestrial sites (Farrera et al., 1999).

IE
Despite the importance of dissolved noble gas data, questions about the

interpretation of these data remain (Hall et al., 2005). Most interpretations rely on simple
EV
models of gas dissolution, with little experimental confirmation of these models.

Furthermore, the use of NGT in paleoclimatology requires multiple assumptions which

have received minimal attention to date (Castro et al., 2007).

PR

RESEARCH QUESTIONS

The research questions addressed in this work are in the following areas:

1. Validity of NGT as a measure of water table temperature: Do

measured NGT accurately reflect water table temperature? What, if any,

processes affect dissolved gas concentrations after groundwater recharge?

If such processes occur, what is the impact on calculated NGT?

2. Noble gas interpretive models: How do noble gas interpretive models

differ? Which model best fits measured data? What are the implications

of using an “incorrect” model?

2
 3. Noble gas paleothermometry: Do climatic/hydrological changes that

occur over glacial-interglacial periods affect paleo-air temperatures

inferred from NGT?

HYPOTHESES AND RESEARCH APPROACH

The aim of this dissertation is to improve our understanding of gas dissolution

during groundwater recharge and to critically examine the utility of dissolved noble gas

data in groundwater research. Multiple methods were used to achieve this goal.

Chapter 2. I hypothesized that calculated NGT reflect measured water table

temperatures. Additionally, I hypothesized that differing recharge regimes would affect

W
the concentrations of noble gases dissolved during recharge. To test these hypotheses a

shallow aquifer in Kings County, California was instrumented and monitored. This field
IE
study also examined the impact of saturated zone degassing on interpreted NGT. This

work was an extension of a research program on denitrification conducted by Lawrence

EV

Livermore National Laboratory (LLNL). This program involved multiple research sites

and focused on denitrification of nitrate discharged from agricultural operations in

California’s Central Valley. LLNL generated the geochemical and dissolved noble gas
PR

data reported in Chapter 2.

Chapter 3. Multiple models exist to interpret dissolved noble gas data. I

hypothesized that differences in model formulation would lead to differences in model

output. This hypothesis was tested by comparing model results from a large dissolved

noble gas data set (905 samples). The data set included samples from across California—

including many samples impacted by artificial recharge. The impact of artificial recharge

on gas dissolution was examined. This study was an extension of work performed at

LLNL as part of the California Aquifer Susceptibility (CAS) Assessment of the

Groundwater Ambient Monitoring and Assessment (GAMA) program run by California’s

3
State Water Resources Control Board. CAS examined water quality and relative

susceptibility to contamination of groundwater that serves as a source for public drinking

water. The data set analyzed in Chapter 3 was generated by LLNL as part of CAS and

GAMA.

Chapter 4. Studies investigating paleoclimate are primarily concerned with air

temperatures rather than subsurface temperatures. The value of NGT data to

paleoclimate studies relies on the temporal constancy of the coupling between water table

temperature and air temperature. I hypothesized that water table temperatures and air

temperatures do not remain coupled over glacial-interglacial time scales. To test this

W
hypothesis, a series of numerical modeling experiments were completed.
Chapter 5. This chapter summarizes the key findings uncovered during the
IE
course of this research. It also suggests areas of future work for hydrologic studies using

dissolved noble gas data.

EV
PR

4
 Chapter 2:

Noble Gas Recharge Temperature Studies

at an Agricultural Site in California

Abstract
Recharge temperature inferred from dissolved noble gas concentrations (noble gas

temperatures, NGT) are useful as a quantitative proxy for air temperature change since

the last glacial maximum. Despite their importance in paleoclimate research, few studies

W
have investigated the relationship between NGT and actual recharge temperatures in field

settings. This study presents dissolved noble gas data from a series of monitor wells in a
IE
shallow unconfined aquifer in an area heavily impacted by agriculture. Multiple samples

had dissolved gas concentrations below the equilibrium concentration with respect to
EV
atmospheric pressure, indicating degassing. The NGT calculated from common

physically based interpretive gas dissolution models on samples unaffected by degassing

do reflect the measured water table temperatures (WTT). The ability to fit the data to
PR

multiple interpretive models indicates that model goodness-of-fit does not necessarily

mean that the model reflects the actual gas dissolution process. Although NGT are useful

in that they reflect WTT, caution is recommended when using these interpretive models.

There was no measurable difference in excess air characteristics (amount and degree of

fractionation) between the two recharge regimes studied. Geochemical and dissolved gas

data indicate that saturated zone denitrification caused degassing by gas stripping.

Modeling indicates that minor degassing (<10% ΔNe) may cause underestimation of

groundwater recharge temperature by 2 °C. Such errors are problematic because

degassing may not be apparent and degassed samples may be fit by a model with a high

5
degree of certainty. This source of error has implications for application of NGT data in

paleoclimate studies.

INTRODUCTION

Dissolved noble gases (He, Ne, Ar, Kr, and Xe) impart unique and valuable

information in hydrologic studies. The conservative behavior of noble gases allows

estimation of groundwater recharge temperatures (noble gas temperatures, NGT) as well

as groundwater ages. NGT are particularly important in paleoclimate research for

quantifying the temperature difference from the last glacial maximum (LGM) to present

(e.g. Farrera et al., 1999).

W
It is common for groundwater to contain dissolved gas concentrations greater than

equilibrium concentration with respect to atmospheric pressure. The additional dissolved

IE
gas is termed excess air because of its compositional similarity to air (Heaton and Vogel,

1981). Accurate determination of excess air is necessary for groundwater age-dating

EV

using the 3H–3He technique (Solomon and Cook, 2000). Some suggest that excess air

may itself be a valuable paleoclimate proxy (Aeschbach-Hertig et al., 2002b; Castro et

al., 2007).
PR

NGT are much more sensitive to concentrations of heavier gases (e.g. Xe, Kr)

because the solubility of these gases have much greater temperature dependency. In

contrast, excess air is much more sensitive to concentrations of lighter gases (e.g. He,

Ne). It is common to measure multiple gases to calculate NGT and excess air

simultaneously using an error weighted inverse modeling procedure (Aeschbach-Hertig

et al., 1999; Aeschbach-Hertig et al., 2000).

Despite the importance of NGT in paleoclimate research, few studies have

attempted to experimentally confirm that NGT accurately reflect water table temperatures

(WTT). This deficiency is critical given recent work examining assumptions in NGT

6
calculations (Castro et al., 2007; Hall et al., 2005). Most noble gas studies report the

sampled water temperature and mean annual air temperature (MAAT); however, it is

extremely rare for researchers to directly measure WTT.

Holocher et al. (2002) completed a series of laboratory column experiments in

which excess air was generated. The NGT matched the column temperature within

measurement uncertainty for all samples. Stute and Sonntag (1992) investigated the

relationship between NGT and subsurface temperature. NGT at a site near Bocholt,

Germany showed evidence of recharge from two areas (i.e. forest and field/meadow)

having different soil thermal regimes. Subsurface temperature data from ~1 m above the

W
water table were available from a nearby meteorological station having field/meadow

vegetation. No temperature measurements for the forested area were reported. The NGT
IE
of groundwater recharged in the field/meadow was the same as the measured soil

temperature.
EV
In a regional study Castro et al. (2007) compared calculated NGT to recharge

zone WTT. Details of the WTT data (e.g. precise location, measurement date(s), etc.)

used in the study were not reported. While unable to connect the WTT to NGT using
PR

common gas dissolution models, the NGT results matched WTT if subsurface noble gas

partial pressures were assumed to be greater than their respective atmospheric partial
pressures. Subsurface noble gas partial pressures could be elevated relative to

atmospheric conditions from O2 consumption by biological processes and subsequent

dissolution of the produced CO2 (Stute and Schlosser, 2000).

Klump et al. (2007) reported on field scale noble gas dissolution experiments

from two sites. In situ temperatures were not taken at the two study sites; however,

subsurface temperatures were inferred from either measuring samples of recently

7
recharged water or using data from a nearby (~20 km) meteorological station. They

concluded that calculated NGT accurately reflected in situ soil temperatures.

In each of these three field studies, subsurface temperature data were considered

in an attempt to compare NGT to WTT. However, none of these studies incorporated

direct measurements of subsurface temperature to examine the relationship between NGT

of very young (weeks to years) groundwater to WTT.

The objectives of the study were to: 1) compare the modeled NGT to the

measured WTT to evaluate potential bias in NGT, 2) compare differences in gas

dissolution occurring under different recharge regimes, and 3) examine the potential

W
impact of degassing on NGT. Improved understanding of gas dissolution processes

occurring during groundwater recharge will benefit groundwater age-dating and

IE
paleoclimate studies. This study offers the following improvements over the noble gas

studies discussed above: 1) high frequency measurements of subsurface temperature

EV
throughout the unsaturated zone at multiple locations, 2) noble gas concentrations

measured at multiple locations across the site, and 3) two different recharge regimes.

This study complements recent work from the same site by Singleton et al. (2007) that
PR

focused on evidence for saturated zone denitrification, and by McNab et al. (2007) that

focused on the geochemistry of lagoon water. Singleton et al. (2007) present average

NGT and excess air for the site. All dissolved noble gas analyses presented here are

original—revised and expanded from the previous analyses by Singleton et al. (2007).

MATERIALS AND METHODS

Study Site

The study site includes a dairy farm and surrounding fields in Kings County, CA.

The climate is Mediterranean type with hot summers and mild winters

(MAAT = 16.6 °C). Mean annual precipitation is 170 mm, with 80% falling during the

8
coolest five months (November through March). Local meteorological data were

obtained from a nearby (~10 km) National Climatic Data Center (NCDC) station. The

site has minimal topographic relief and an elevation ~70 m above mean sea level. The

local geology consists of unconsolidated sediments (primarily sands and silts) that were

deposited in a series of alluvial fan systems originating where rivers exit the Sierra

Nevada mountains (Weissmann et al., 1999).

Cropland surrounding the dairy operation is flood irrigated with a combination of

groundwater and dairy wastewater (i.e. liquid manure). Occasionally water from the

Kings River is transported through unlined canals for irrigation. Groundwater used for

W
irrigation is drawn from both a shallow perched aquifer (<25 m below ground surface,

BGS) and a deeper aquifer (>40 m BGS). Water for domestic use is drawn from the deep
IE
aquifer. The water table is ~5 m BGS across the site. Groundwater flow direction within

the perched aquifer is difficult to characterize because of the many irrigation wells that
EV
pump intermittently and the seasonally filled irrigation canals. Deeper regional

groundwater flow is generally westward toward the center of the valley (Williamson et

al., 1989). Additional details of the study site are given elsewhere (McNab et al., 2007;
PR

Singleton et al., 2007).

Five sets (locations 1S, 2S, 3S, 4S, and 6S) of small diameter multi-level wells
were previously installed at the site (locations given in Figure 2.1, depths given in Table
2.1 and Figure 2.2) as part of related studies (McNab et al., 2007; Singleton et al., 2007).

These multi-level well sites are all located on the edges of fields alternatingly planted

with corn and wheat, except 2S which is beside an alfalfa field and 6S which is between

cattle pens and manure lagoons. A sixth single completion well site (well 5S1) is located

in a field ~11 m from the study area’s main irrigation canal. This 14 m wide canal is

commonly full only during spring and summer months.