University of New Mexico UNM Digital Repository

Earth and Planetary Sciences ETDs Electronic Theses and Dissertations

12-1-2014 Continental-scale Scott aJ sechko

Follow this and additional works at: https://digitalrepository.unm.edu/eps_etds

Recommended Citation Jasechko, Scott. C" ontinental-scale isotope hydrology." (2014). https://digitalrepository.unm.edu/eps_etds/40

This Dissertation is brought to you for free and open access by the Electronic Theses and Dissertations at UNM Digital Repository. It has been accepted for inclusion in Earth and Planetary Sciences ETDs by an authorized administrator of UNM Digital Repository. For more information, please contact [email protected].

Scott Jasechko Candidate

Earth and Planetary Sciences Department

This dissertation is approved, and it is acceptable in quality and form for publication:

Approved by the Dissertation Committee:

Dr. Zachary D. Sharp , Co-chairperson

Dr. Peter J. Fawcett , Co-chairperson

Dr. Joseph Galewsky

Dr. Juske Horita

CONTINENTAL-SCALE ISOTOPE HYDROLOGY

by

SCOTT ALLAN JASECHKO

B.Sc., University of Victoria, 2009 M.Sc. University of Waterloo, 2011

DISSERTATION

Submitted in Partial Fulfillment of the Requirements for the Degree of

Doctor of Philosophy Earth and Planetary Sciences

The University of New Mexico Albuquerque, New Mexico

December, 2014

DEDICATION

To Jennifer, Gordon, Glenn and Edith – for your love and your encouragement.

iii ACKNOWLEDGMENTS

I am grateful to Zachary Sharp, Peter Fawcett and Joseph Galewsky for supporting, challenging and guiding me throughout my Ph.D. education. I am thankful for my friends and mentors who continue to grant me the joy of belonging in a community.

iv CONTINENTAL-SCALE ISOTOPE HYDROLOGY

By

Scott Jasechko

B.Sc., Physical Geography and Earth and Ocean Sciences, University of Victoria, 2009

M.Sc., Earth and Environmental Sciences, University of Waterloo, 2011

Ph.D., Earth and Planetary Sciences, University of New Mexico, 2014

ABSTRACT

Providing sustainable sources of for a growing population of 7 billion people is one of the grand challenges of the 21st century. This dissertation outlines several applications of isotope hydrology to address four previously unknown questions involving surface- and ground-water resources at regional- to continental-spatial scales over contemporary- to millennial-temporal scales.

The four chapters in this dissertation investigate (1) the rate of plant transpiration, (2) the seasonality of groundwater recharge, (3) the climate of the last ice age, and (4) the chemistry of Ugandan waters.

(1) Chapter one presents a new global compilation of lake water isotopic data, river isotopic data, stand-level transpiration rates, and water use efficiency measurements, and analyzes the newly synthesized data to show that plant transpiration is the largest water flux from Earth’s continents, exceeding both physical evaporation and continental runoff. (2) Chapter two presents a new global synthesis of rain, snow and groundwater isotopic compositions, and analyzes the paired

v precipitation-groundwater dataset to show that the percentage of precipitation that recharges is at a maximum during the winter (extra-tropics) and wet (tropics) seasons. (3) Chapter three presents a new global compilation of groundwater radiocarbon, tritium, and stable O and H isotopic data, and maps the isotopic shift of meteoric waters since the last ice age. The analysis shows that the majority (~90%) of precipitation during the last ice age had lower 18O/16O and 2H/1H ratios than the modern day, except in some exclusively coastal locations. We also show that current isotope-enabled general circulation models capture some, but not all, spatial variability in ice-age-to-late-Holocene

18O/16O and 2H/1H shifts, providing a new calibration tool that can be used to improve our understanding of glacial climate dynamics. (4) Chapter four presents isotopic and chemical analyses of Ugandan lake, river, rain, and ground water collected during a field expedition led in July of 2013.

Analysis of this new dataset reveals new estimates of lake water balances across Uganda.

vi TABLE OF CONTENTS DEDICATION ...... iii ACKNOWLEDGMENTS ...... iv ABSTRACT ...... v TABLE OF CONTENTS ...... vii PREFACE...... 1 CHAPTER 1 — GLOBAL PLANT TRANSPIRATION FLUXES ...... 9 1.1 Abstract ...... 9 1.2 Introduction ...... 9 1.3 Dataset and methods ...... 32 1.4 Results ...... 54 1.5 Discussion ...... 61 1.6 References ...... 63 CHAPTER 2 — THE SEASONALITY OF GLOBAL GROUNDWATER RECHARGE ...... 83 2.1 Abstract ...... 83 2.2 Introduction ...... 83 2.3 Dataset and methods ...... 88 2.4 Results ...... 101 2.5 Discussion ...... 106 2.6 References ...... 121 CHAPTER 3 — THE ISOTOPIC COMPOSITION OF ICE AGE GROUNDWATERS ...... 142 3.1 Abstract ...... 142 3.2 Introduction ...... 143 3.3 Dataset and Methods...... 148 3.4 Results ...... 151 3.5 Discussion ...... 164 3.6 References ...... 188 CHAPTER 4 — THE ISOTOPE HYDROLOGY OF UGANDA ...... 212 4.1 Abstract ...... 212 4.2 Introduction ...... 212 4.3 Dataset and methods ...... 214 4.4 Results ...... 217 4.5 Discussion ...... 231 4.6 References ...... 241 vii

List of Figures 1-1. Schematic of fresh water fluxes ...... 10

1-2. Schematic of plant transpiration ...... 12

1-3. Water yields before and after the clearing of vegetation in experimental watersheds ...... 15

1-4. Compiled transpiration/evapotranspiration ratios ...... 17

1-5. Compiled transpiration/evapotranspiration ratios and precipitation rates ...... 28

1-6. Transpiration/evapotranspiration measurements sorted by technique...... 29

1-7. Compiled desert transpiration/evapotranspiration ratios and precipitation rates ...... 30

1-8. Locations of transpiration study watersheds ...... 34

1-9. Water use efficiency as a function of vapor pressure deficit ...... 50

1-10. Spatial distribution of water use efficiency ...... 50

1-11. The deuterium excess of 31 major rivers ...... 52

1-12. The O and H isotopic composition of large lakes ...... 54

1-13. Heterogeneity of lake water O and H ...... 55

1-14. Temperature and O isotopic composition of Baikal and Tanganyika ...... 56

1-15. The isotopic composition of the North American Great Lakes at depth ...... 56

1-16. The transpiration rate calculated for 54 lake catchments grouped by biome...... 58

1-17. The transpiration rate for 10% of Earth’s ice free land area ...... 59

1-18. The transpiration rate for 73 catchments ...... 60

1-19. Gross primary productivity for 10% of ice free land areas ...... 61

2-1. An estimate of the global annual groundwater recharge ratio ...... 86

2-2. Locations of paired precipitation-groundwater isotopic data ...... 91

2-3. The change in meteoric 3H from 1930 to 2009...... 96

2-4. The isotopic approach to quantifying recharge/precipitation seasonality ...... 100

2-5. Comparison of groundwater and precipitation O and H isotopic data ...... 101

viii 2-6. The seasonality of groundwater recharge ratios at 54 locations ...... 103

2-7. Comparison of groundwater and precipitation isotopic compositions ...... 106

2-8. The seasonality of O and H isotopic data in precipitation ...... 108

2-9. Seasonality of normalized difference vegetation indices across land surfaces ...... 110

2-10. A comparison of modelled and isotope-based recharge ratio seasonality ...... 113

2-11. Cross plot of modelled and isotope-based recharge ratio seasonalities ...... 115

3-1. Temperature changes from the last glacial maximum to the modern day ...... 145

3-2. Map of O and H isotopic change from the last ice age to the late-Holocene...... 152

3-3. Ranges of isotopic change from the last ice age to the late-Holocene observed in records ...... 162

3-4. The difference between δ18Oice age and δ18Olate-Holocene with latitude ...... 163

3-5. The modelled (CCSM) precipitation δ18Olast glacial maximum – δ18Opre-industrial ...... 167

3-6. The modelled (ECHAM) precipitation δ18Olast glacial maximum – δ18Opre-industrial ...... 168

3-7. The modelled (IsoGSM) precipitation δ18Olast glacial maximum – δ18Opre-industrial ...... 169

3-8. The modelled (LMDZ) precipitation δ18Olast glacial maximum – δ18Opre-industrial ...... 170

3-9. Locations where all models agree on the sign of δ18Olast glacial maximum – δ18Opre-industrial ...... 171

3-10. Agreement for 3 of 4 models on the sign of δ18Olast glacial maximum – δ18Opre-industrial ...... 172

4-1. The O and H isotopic composition of Ugandan waters ...... 218

4-2. Sampling locations of Ugandan waters ...... 221

4-3. A Piper diagram showing the hydrochemistry of Ugandan waters ...... 229

4-4. Stable-isotope-based evaporation/input ratios for 24 Ugandan Lakes ...... 230

4-5. O isotopic composition and conductivity of Ugandan lakes and groundwaters ...... 232

4-6. Deuterium excess of Ugandan lakes and groundwaters with electrical conductivity ...... 233

4-7. The deuterium excess and sample elevation of Ugandan rivers and groundwaters ...... 234

4-8. Stable-isotope-based evaporation/input ratios based on O and H isotopes ...... 237

ix List of Tables

1-1. Compiled transpiration/evapotranspiration ratios ...... 19–26

1-2. Compiled transpiration/evapotranspiration sorted by ecoregion ...... 27

1-3. Modelled transpiration/evapotranspiration fractions at a global scale ...... 32

1-4. Study lake information ...... 35–36

1-5. Study lake hydrography ...... 37–38

1-6. Transpiration model input parameters (isotope) ...... 44–45

1-7. Transpiration model input parameters (isotope, temperature and humidity) ...... 45–46

1-8. Compiled large lake isotopic investigations ...... 48–49

1-9. Compiled plant water use efficiencies...... 51

1-10. Deuterium excess of major rivers...... 53

1-11. The terrestrial sublimation flux estimated in previous studies ...... 62

2-1. Locations of paired groundwater and precipitation data ...... 89–90

2-2. Seasonal groundwater recharge ratio results ...... 104–105

3-1. Modern and ice age physical and isotopic data for the oceans and the cryosphere ...... 147

3-2. Differences in the δ18O value of the last ice age and the late-Holocene...... 153–154

3-3. Observed ranges of δ18Oice age and δ18Olate-Holocene values in groundwaters ...... 155–158

3-4. Speleothem δ18O from the last ice age to the late-Holocene ...... 159

3-5. Ice core δ18O from the last ice age to the late-Holocene ...... 160

4-1. The isotopic composition of Ugandan waters ...... 220

4-2. The isotopic composition of Ugandan waters sampled and measured in this study ...... 222–227

4-3. Major ion chemistry of Ugandan waters ...... 235–236

x PREFACE

This dissertation uses stable O and H isotopic compositions of meteoric waters to quantify the sources and processes that govern the storage and movement of water on continents. This preface is divided into two parts: (Part A) an outline of the publication yields from this dissertation, specifically addressing the publication authorship guidelines as stipulated by The University of New

Mexico’s Department of Earth and Planetary Sciences, and (Part B) an outline for this dissertation, with brief terminology and background information relevant to all four chapters within this dissertation.

(Part A) This dissertation is linked to six planned-or-published peer reviewed publications.

Three articles have already been published in top-tier peer-reviewed journals. One other publication is currently in review. Two manuscripts are being prepared for submission to a peer reviewed journal at the time this dissertation is submitted. Chapter 1 is linked to three publications in the journals

Nature (two publications) and Agricultural and Forest Meteorology. Chapter 2 is linked to a manuscript that is currently undergoing peer review. Publications linked to chapters 3 and 4 are being prepared for submission at the time that this dissertation is being submitted.

S. Jasechko is the lead author of three published articles, is the second author on the third publication that has only two authors, in total. S. Jasechko is the lead author on the fourth publication that is currently in review, and will also be the lead author for publication six that results from dissertation chapter 4. To follow the requirements of the Department of Earth and Planetary

Sciences, the contributions and roles of each co-author within each publication are outlined below.

Chapter 1 uses a global dataset of isotopic data compiled for lakes and rivers to quantify the rate that plants uptake water. Chapter 1 has yielded three publications in peer reviewed journals, the first of which was published in April-2013 in Nature, the second was published in February-2014 in

1 Nature, and the third was published in June-2014 in Agricultural and Forest Meteorology. The first of three Chapter one publications included five analysis components, all of which were completed by

S. Jasechko in entirety: (i) compilation and synthesis of data, (ii) geospatial analyses, (iii) development of methodology and equations, (iv) analysis of global remote sensing data, and (v) calculation of transpiration fluxes. The manuscripts published in Nature were written by S. Jasechko with comments and suggestions from Z. D. Sharp, J. J. Gibson, S. J. Birks, Y. Yi and P. J. Fawcett. The publication in Agricultural and Forest Meteorology represents a collaboration between W.

Schlesinger and S. Jasechko, with W. Schlesinger and S. Jasechko sharing in all stages of manuscript development (data compilation, figure development, statistical analyses and writing of the manuscript text.).

Chapter 2 synthesizes global isotopic datasets of modern groundwater and precipitation and analyzes the database to quantify the seasonal differences in groundwater recharge ratios at 54 globally-distributed locations, where the “groundwater recharge ratio” is defined as the fraction of precipitation that recharges groundwater aquifers. Chapter 2 has yielded one manuscript that is in press for publication in Water Resources Research at the time that this dissertation was submitted.

The Chapter 2 project included six parts: (i) compilation of a global groundwater isotopic dataset, (ii) amalgamation of three continental-scale precipitation isotopic datasets, (iii) geospatial synthesis of groundwater isotopic data with precipitation data, (iv) development of a new set of equations, (v) calculation of groundwater recharge ratios, (vi) comparison of results with a state-of-the-art global hydrological model. S. Jasechko led all components of this analysis, and worked together with collaborators on two components: part (ii): “amalgamation of three continental-scale precipitation isotopic datasets” (working with S. J. Birks and J. M. Welker), and part (vi): “comparison of results with a state-of-the-art global hydrological model” (working with T. Gleeson and Y. Wada). The manuscript was prepared by S. Jasechko (first author), and incorporates comments and suggestions

2 from all co-authors: S. J. Birks, T. Gleeson, Y. Wada, P. J. Fawcett, Z. D. Sharp, J. J. McDonnell and

J. M. Welker.

Chapter 3 uses a newly developed global groundwater dataset of radioactive carbon, radioactive hydrogen, stable isotopes and stable hydrogen isotopes to delineate ice age groundwaters and to map the distribution of 18O/16O and 2H/1H ratios of meteoric waters from the last ice age. A manuscript linked to Chapter 3 — which will be submitted for publication — has been prepared by S. Jasechko (first author) and is now circulating amongst co-authors at the time this dissertation is submitted. The project included four components: (i) groundwater isotopic data compilation, (ii) radiocarbon dating of groundwaters, (iii) statistical comparison of modern- and paleo-groundwater isotopic compositions, (iv) geospatial comparison of groundwater isotopic observations with the results of four isotope-enabled general circulation models run under last ice age conditions. S. Jasechko led each component, and worked with leaders of four general circulation models on component (iv): “geospatial comparison of groundwater isotopic observations with the results of four isotope-enabled general circulation models run under last ice age conditions.” The manuscript in preparation was written by S. Jasechko incorporates comments and suggestions from all co-authors.

Chapter 4 uses a newly developed isotopic dataset of Ugandan groundwaters, lakes, rivers, precipitation and springs to quantify hydrological processes controlling water availability throughout the country. Chapter 4 presents a newly developed dataset resulting from fieldwork planned and led by S. Jasechko, with on-the-ground collaborative support from M. Kizza (Makerere University) and

M. GebreEgziabher (Addis Ababa University). The cost of travel, lodging, transportation, sampling equipment and geochemical analysis were supported in entirety by the combined graduate student research funds awarded through four graduate student research grants to S. Jasechko: (i) the

Consortium of Universities for the Advancement of Hydrologic Science’s Pathfinder Fellowship, (ii)

3 the American Geophysical Union’s Horton Hydrology Research Grant, (iii) the Geological Society of

America’s Graduate Student Research Grant, and (iv) the Caswell-Silver Foundation’s Kelley-Silver

Graduate Fellowship research allotment. Special sampling equipment was loaned to S. Jasechko by T.

Fischer, L. Crossey and K. Karlstrom. Sample preparation and chemical analyses were completed by

S. Jasechko with guidance from A. S. Ali. Oxygen, hydrogen and carbon isotopic analyses were completed by S. Jasechko with laboratory support from V. Atudorei. The results of this project will be submitted to an appropriate peer reviewed journal, with S. Jasechko listed as the lead author.

(Part B) This dissertation is divided into four chapters (in order): (1) Global plant transpiration fluxes, (2) The seasonality of global groundwater recharge, (3) A global database of ice age groundwaters, and (4) The isotope hydrology of Uganda. The four chapters examine a variety of spatial scales, ranging from local (~101 km2) and regional scales (~104 km2; e.g., Chapter 4) to continental scales (~106 to 107 km2). The four chapters explore different time periods, ranging from the climate of the last ice age (104 years before today) to the climate of the present day. The cross- cutting theme that binds there four chapters into one dissertation is that all four chapters investigate distributions of 18O/16O and 2H/1H ratios in environmental waters. An introduction to the application of oxygen and hydrogen isotopes in hydrology is presented next before delving into each chapter.

The elements of oxygen and hydrogen were first discovered in the 18th century by Henry

Cavendish and Antoine Lavoisier. However, it was not for another 150 years that the isotopes of oxygen and hydrogen were first discovered. The first isotopes to be identified were of thorium and uranium (McCoy and Ross, 1907) and were first acknowledged by F. Soddy (Soddy, 1913), who received the Nobel Prize in 1922 for this work. Soddy used the term isotopes to describe radionuclides that had different decay rates but seemed at the time to be identical in all other manners: "Put colloquially, their atoms have identical outsides but different insides... These elements

4 which are identical in their whole chemical character and are not separable by any method of chemical analysis are now called isotopes" (F. Soddy’s Nobel Prize address in 1922; within reference:

Soddy, 1966).

Soddy first identified differences in the radioactive decay rates of isotopes (Soddy, 1913).

Since his work more than 100 years ago, we have learned that small chemical differences between stable isotopes exist. The differences arise due to differences in mass between isotopes, which is defined by the sum of protons (Z) and neutrons (N) within the atomic nucleus (minus a small amount of “missing” mass that has been converted to nuclear binding energy). The discovery of the existence of stable isotopes of oxygen can be credited to Blackett (1925) who used photography to document the production of 17O from the capture of an alpha particle (i.e., He2+, a particle comprised of two neutrons and two protons) by a common nitrogen atom (14N). Soon after this laboratory experiment, different stable isotopes of oxygen (16O, 17O, 18O) were discovered to be naturally occurring within Earth’s atmosphere (Giauque and Johnson, 1929a; 1929b). The discovery of a stable isotope of hydrogen is credited to Urey (1932) who applied electrolysis to natural waters to extract deuterium and confirm the existence of two stable hydrogen isotopes in nature (2H, 1H).

The discovery of naturally occurring stable isotopes of O and H have led to a vast array of hydrological and paleo-climate investigations. Landmark work in the 1950s and 1960s identified several features of the global isotopic data that have been reproduced many times over since their foundation (i.e., Friedman, 1953, Craig, 1961, Dansgaard, 1964): (i) the ratios of 18O/16O and 2H/1H covary in precipitation (Friedman 1953; Craig, 1961), (ii) the isotopic composition of precipitation is controlled by temperature-dependent fractionation during rainout, leading to lower 18O/16O and

2H/1H ratios farther from moisture sources (Dansgaard, 1964), (iii) the process of evaporation changes 18O/16O and 2H/1H ratios in different proportions than condensation due to additional kinetic (i.e., disequilibrium) isotope effects (Craig, 1961), (iv) plant transpiration does not modify the

5 isotopic composition of water (Wershaw et al., 1966), (v) the isotopic composition of waters has changed over Earth’s history and provides information about past climates (e.g., Urey et al., 1951).

Many other interesting discoveries have been made in the field of stable isotope hydrology over the past 60 years; however, aforementioned discoveries are of particular importance to the discoveries outlined in the forthcoming chapters.

Last, before I begin chapter 1, some isotopic terminology must be presented. Isotopic data are presented in per mille notation on a scale that ranges from −1000 to +∞. Delta notation is described mathematically as δ = (Rsample/Rstandard) × 1000 ‰, where R represented the ratio of

18O/16O or the ratio of 2H/1H, and the subscripts sample and standard refer to the ratio in the measured sample or in an international standard, respectively. The international standard most commonly applied to O and H isotopes is oceanic water: “standard mean ocean water” (or, SMOW), that has an 18O/16O ratio of 0.00200520±0.00000043 and a 2H/1H ratio of 0.00015575±0.00000008

(Baertschi, 1976; de Wit et al., 1980).

6 References (preface)

Baertschi, P. (1976), Absolute 18O content of standard mean ocean water, Earth and Planetary

Science Letters, 31, 341–344.

Blackett, P. M. S. (1927), The Ejection of Protons from Nitrogen Nuclei, Photographed by the Wilson Method, Proceedings of the Royal Society of London, 107, 49–360.

Craig, H. (1961), Isotopic variations in meteoric waters, Science, 133, 1702–1703.

Dansgaard, W. (1964), Stable isotopes in precipitation, Tellus, 16, 436–468.

de Wit J.C., van der Straaten, C. M., Mook, W. G. (1980), Determination of the absolute hydrogen isotopic ratio of V-SMOW and SLAP, Geostandards Newsletter 4, 33–36.

Friedman, I. (1953), Deuterium content of natural water and other substances. Geochimica

Cosmochimica Acta, 4, 89–103.

Giauque, W. F., and Johnson, H. L. (1929a), An isotope of oxygen of mass 17 in the earth’s atmosphere, Nature, 123, 831.

Giauque, W. F., and Johnson, H. L. (1929b), An isotope of oxygen, mass 18, Nature, 123,

318.

McCoy, H. N., and Ross, W. H. (1907), The specific radioactivity of thorium and the variation of the activity with chemical treatment and with time. Journal of the American Chemical Society,

29, 1709–1718.

Soddy, F. (1913), The Radio-elements and the Periodic Law, Chemical News, 107, 97–99.

7 Soddy, F. (1966), The Origins of the Conceptions of Isotopes, Nobel Lecture, December 12,

1922, In Nobel Lectures Including Presentation Speeches and Laureates' Biographies: Chemistry

1901–1921, New York: Elsevier, pp. 367–401.

Urey H. C., Brickwedde, F. G., and Murphy, G. M. (1932), A Hydrogen Isotope of Mass 2,

Physical Review, 39, 164–165.

Urey, H. C., Lowenstam, H. A., Epstein, S., & McKinney, C. R. (1951), Measurement of paleotemperatures and temperatures of the Upper Cretaceous of England, Denmark, and the southeastern United States, Geological Society of America Bulletin, 62, 399–416.

8 CHAPTER 1 — GLOBAL PLANT TRANSPIRATION FLUXES

1.1 Abstract Terrestrial water stores are balanced by inputs from rainfall and snowfall and losses via evaporation, transpiration, river discharges and submarine groundwater discharges. Two-thirds of all precipitation on land surfaces is vaporized by transpiration or by evaporation, but current general circulation and land surface models span a wide range of predicted transpiration/evaporation ratios. Here I analyze a global dataset of river and lake water isotopic data to show that gas exchange at plant stoma represents both (i) the largest outgoing water flux from Earth’s continents, and (ii) the greatest assimilation of CO2 in the global climate system. This result suggests that current land surface and climate models can prioritize biological, rather than physical (evaporation), water fluxes to enhance predictions of water availability under varying climate and land use futures.

1.2 Introduction

Chapter 1 describes a new approach to quantifying transpiration using isotopic data in lakes and rivers. The approach and results of Chapter 1 were published in April of 2013 (Jasechko et al.,

2013). Three subsequent works have been published (or are in review) since April of 2013 as a result of this initial publication (Schlesinger and Jasechko, 2014; Jasechko, 2014, Evaristo et al., in review).

This chapter presents a background to plant transpiration investigations in hydrology, discusses the isotopic dataset and approach taken to quantify transpiration, and concludes by discussing the ramifications of this work and presenting a vision for this field moving ahead.

Water transport on continents is replenished by precipitation on land surfaces that provides about 110,000 km3 of fresh water each year (Oki and Kanae, 2006). The path that water takes after falling on the land surface involves mixing, storage and transportation either as a liquid (i.e., advection-dispersion through porous media, or streamflow) or through vapourization via evaporation or plant transpiration. It has long been recognized that evapotranspiration outweighs 9 streamflow on continents by a factor of close to 2. Evapotranspiration consumes two-thirds of all precipitation on continents, with an annual flux close to 70,000 km3/year (Jung et al., 2011).

Streamflow, on the other hand, has an annual flux of ~36,000 to ~38,000 km3/year (Dai and

Trenberth, 2002; Syed et al., 2010; Figure 1-1).

Figure 1-1. Schematic of current knowledge of fresh water fluxes on continents using rounded numbers (note, submarine groundwater discharge not depicted, although this flux is expected to be

10 to 10,000 times less than continental runoff in rivers; Taniguchi et al., 2002).

Evapotranspiration is comprised of two components: plant transpiration (a biological process) and evaporation (a physical process). Transpiration supports multiple life-sustaining roles for plants. First, plants – like humans – require water for their cellular structures. Where humans are about 70 % water, plants are ~80 % water. Transpiration supports cellular growth through the provision of fresh water to plant cells. Second, plants move nutrients from the subsurface into photosynthetically-active regions within the plant. For tall trees in forests this is often the canopy, which can be in excess of 10s of meters above the ground surface. Third, plant transpiration requires

10 energy to convert liquid water into vapor (i.e., latent energy); plants capitalize upon this vaporization energy requirement to cool off their leaf surfaces and, thus, moderate their growing leaf temperature close to a cool, pan-biome temperature of 21°C (Helliker and Richter, 2008; Figure 1-2).

11

Figure 1-2. Schematic of gas exchange at plant stoma. Plants draw water from soil and groundwater reservoirs, moving the water and entrained nutrients up the xylem by capitalizing on capillary action.

At stoma (upper right) plants passively release H2O (liquid to vapor conversion) via evaporation at leaf surfaces (termed transpiration), which also cools leaf surfaces and maintains leaf temperatures that are optimal for growth (Helliker and Richter, 2008). 12 The separate fluxes of evaporation and transpiration have been considered as a single component in hydrological investigations, lumped into the single term: evapotranspiration. However, separating the fluxes of evaporation and transpiration is important for hydroclimatology because the processes of evaporation and transpiration are different, and will respond differently to land use and climate modifications. Where evaporation is a physical process, transpiration is a biological process that is central to primary production on continents: the largest carbon flux in the climate system

(~120 Gt C/year; Beer et al., 2010). The response of evaporation and transpiration to a changing climate will be different. Evaporation is a physical process, and the potential for evaporation can broadly be expected to increase with future warming, although some evaporation pan data suggest that other factors may exert a stronger control than temperature alone (e.g., Roderick and Farquhar,

2002). The response of plant transpiration to a warming climate is, on the other hand, complicated by several sometimes conflicting responses. For example, warmer temperatures and fertilization of the biosphere through enriched atmospheric CO2 is expected to increase plant productivity, and consequentially increase transpiration. However, the enrichment of CO2 in the atmosphere has also been predicted to increase the water use efficiency of plants (i.e., the ratio of H2O transpired to CO2 uptake), which is expected to decrease transpiration. Indeed a shift to more water-efficient ecosystems has recently been observed across a variety of biomes in North America (Keenan et al.,

2013). Examining the supplementary data within this publication (Keenan et al., 2013) shows that water use efficiency may be changing at a greater rate (i.e., as percentage of the total flux) than streamflow (e.g., Labat et al., 2002; Peterson et al., 2002; McClelland et al., 2006), precipitation

(Zhang et al., 2007) or evapotranspiration (Jung et al., 2010; Miralles et al., 2014), highlighting that knowledge of the exact flux of transpiration on continents is important to accurately predicting change in the global hydrological cycle. These qualitative predictions are further complicated by expected limitations to the “CO2 fertilization effect” imparted by fast approaching nitrogen limitations upon the extent of CO2 fertilization. The broad implication of this analysis, is that models

13 of the Earth’s critical zone and the climate that neglect any one of physics, chemistry or biology are at risk of missing important feedbacks, interactions and thresholds between the atmosphere, biosphere, hydrosphere and lithosphere.

Changing land uses via deforestation and agriculture change transpiration fluxes and influence downstream liquid water yields in rivers. Deforestation by humans has, globally, led to a decrease in natural evapotranspiration of ~3000 km3 per year (about 4% of global evapotranspiration; Gordon et al., 2005; Jung et al., 2010), and irrigation for agriculture – which reactivated groundwater that is part of long flow paths, often removed from the “active” hydrosphere, on human time scales – has increased terrestrial vapor fluxes by ~2600 km3 per year

(Gordon et al., 2005). This pumping of unsustainable groundwater sources has been investigated and mapped at a global scale (Wada et al., 2012).

Other examples of land use modifications and impacts upon downstream hydrology date back to experimental watershed work completed in the 1970s- and 1980s. Bosch and Hewlett (1983) present a series of watershed studies where river flows were measured before and after the complete deforestation of the upstream watershed. Follow deforestation, water yields downstream increased by

75% due to reduced evapotranspiration fluxes following forest clearing, highlighting the important role of transpiration in total evapotranspiration fluxes (average of n = 25 experimental watersheds that were completely deforested; 10th-90th percentile spans +10% increase to 194% increase in water yields; Figure 1-3; Bosch and Hewlett, 1983).

14

Figure 1-3. Water yields before and after the clearing of vegetation within experimental watersheds show an increase in runoff following deforestation. Each point represents a single watershed, which had its outflow monitored before and after forest clearing. Data presented in this figure from Bosch and Hewlett (1983).

Changing climate chemistry is expected to impact transpiration, and will produce an important feedback to regional warming. For example, a general circulation model (Community Land and Community Atmosphere Model) simulation using a transpiration/evapotranspiration ratio of

40% (pers. comm. L. Cao; Cao et al., 2010) showed that the physiological response to a CO2 enriched atmosphere was a decrease in transpiration that reduced latent heat fluxes on continents and ultimately accounted for ~15% of land surface warming, with the remainder largely attributed to CO2 radiative forcing and other feedback mechanisms. Similarly, more than half (8.4%) of the predicted increase in future global river discharges (predicted runoff increase of 15% of present day) in this model were predicted to be derived from reduced transpiration water fluxes. Given the potential of changes to transpiration to warm land surfaces (Cao et al., 2010) and the relatively rapid increases in water use efficiency reported by Keenan et al. (2013) there is a need to quantify the proportion of evapotranspiration completed by vegetation through transpiration.

15 An extensive review of ~100 peer-reviewed publications was completed to uncover studies that have decoupled evapotranspiration into its components: evaporation, and transpiration

(Schlesinger and Jasechko, 2014, building upon a compilation originally presented by Schlesinger and

Bernhardt, 2013). Three groups of studies exist: (i) forest- or cropland-scale field measurements using a suite of techniques (e.g., sap flow meters, radial flow meters, isotope partitioning), (ii) general circulation models, and (iii) land surface models. These three groups of approaches are reviewed in the coming sections.

More than 80 studies have quantified transpiration fluxes at forest stand scales over the past

50 years. The results of these studies were recently compiled and reviewed by Schlesinger and

Jasechko (2014). The locations and transpiration fluxes (reported as a percentage of annual evapotranspiration) of the compiled studies are presented in Figure 1-4. Existing transpiration flux measurements have been completed on all continents and span a variety of biomes with different plant life forms. The studies use on a variety of different approaches to estimate transpiration as a proportion of evapotranspiration.

16

Figure 1-4. (top). Locations of stand-level measurements transpiration (T) as a proportion of evapotranspiration (ET; i.e., T/ET). Colors mark the share of total evapotranspiration accounted for by transpiration alone (Figure reproduced from Schlesinger and Jasechko, 2014). (bottom) Ranges of transpiration/evapotranspiration ratios compiled from 81 studies sorted into major biomes. Bars mark the 25th-75th percentile range of compiled studies for each biome; whiskers mark the 10th-90th percentile range of compiled studies for each biome. Colors delineate the annual flux of evapotranspiration from each biome as a proportion of total terrestrial evapotranspiration, which is

~70,000 km3/year (Jung et al., 2010). Evapotranspiration rates across each biome were obtained from long term mean annual satellite-based evapotranspiration flux data (Mu et al., 2011).

Transpiration can be estimated using a variety of approaches. The most commonly applied approaches broadly fall under the category of hydroclimatological models that utilize meteorological measurements and are often coupled to sap flow measurements or transpiration fluxes (43 studies

17 compiled by Schlesinger and Jasechko, 2014). Other approaches that have been used to measure transpiration fluxes include radial sap flow meters (e.g., Nizinski et al., 2011), energy balance models

(e.g., Tajchman, 1972; Liu et al., 2012), stand level O and H isotope based models (e.g, Hsieh et al.,

1998; Ferretti et al., 2003; Wang et al., 2013), catchment scale O and H isotope based models (e.g.,

Telmer and Veizer, 2000; Gibson and Edwards, 2002), satellite-based estimates (e.g., Tian et al.,

2013) and a water balance approaches comparing water fluxes from control and bare-soil plots

(Schlesinger et al., 1987). A comprehensive review of available stand level measurements is presented in Table 1-1.

18 Table 1-1. Compiled transpiration/evapotranspiration studies.

Ecoregion Country Latitude Longitude Location Tropical 1 India 22.5 87.3 Arabari Range Rainforest 2 Temperate Forest Germany 52.4 13.8 Berlin 3 Temperate Forest Germany 52.4 13.8 Berlin United 4 Temperate Forest - - Pinus sylvestris plantation Kingdom Temperate 5 Deciduous Russia 50.8 42.5 Tellermanovsky Forests 6 Boreal Forest - - - - 7 Boreal Forest Germany 51.8 10.5 Harz Mountains Temperate United States Wyoming: High Plains Grasslands 8 41.1 -104.7 Grassland of America Research Station Tropical 9 Puerto Rico 18.3 -65.7 Luquillo Experimental Forest Rainforest 10 Temperate Forest Germany 48.0 11.6 Near Munich Gubantonggut Desert: Fukang 11 Desert 44.3 87.9 Station of Desert Ecology 12 Boreal Forest Canada 63.4 -114.3 Northwest Territories and Nunavut 13 Boreal Forest Canada 45.7 -76.9 Ottawa River basin 14 Tundra Canada 64.5 -112.7 Northwest Territories and Nunavut Tropical United States 15 20.1 -155.8 Kohala, Hawaii Grassland of America Tropical United States 16 20.1 -155.8 Kohala, Hawaii Grassland of America Tropical United States 17 20.1 -155.8 Kohala, Hawaii Grassland of America Tropical United States 18 20.1 -155.8 Kohala, Hawaii Grassland of America Temperate United States Colorado: Central Plains 19 40.7 -104.8 Grassland of America Experimental Range Temperate United States Oklahoma: Kessler Farm field 20 35.0 -97.5 Grassland of America laboratory Tropical 21 Brazil -3.0 -60.0 Manaus Rainforest Tropical 22 Brazil -3.0 -60.0 Manaus Rainforest Tropical 23 Brazil -3.0 -60.0 Manaus Rainforest Tropical 24 Brazil -3.1 -60.0 Ducke Forest Reserve Rainforest Tropical 25 Indonesia -6.6 106.3 Janlappa Nature Reserve Rainforest 19 Ecoregion Country Latitude Longitude Location Czech 26 Temperate Forest 49.1 13.7 National Park Sumava Republic United 27 Temperate Forest 52.4 0.7 East Anglia Kingdom United 28 Temperate Forest 52.0 -3.5 East of Aberystwyth Kingdom Temperate The 29 Deciduous 52.5 4.6 North Holland Netherlands Forests Temperate Hainich National Park, Central 30 Deciduous Germany 51.1 10.5 Germany Forests Temperate 31 Deciduous Denmark 56.4 9.3 Hald Ege Forests Mediterranean United States 32 32.8 -116.4 Echo Valley, California Shrubland of America Mediterranean United States 33 32.8 -116.4 Echo Valley, California Shrubland of America Mediterranean 34 Chile -33.1 -71.0 Fundo Santa Laura Shrubland Temperate United States Colorado: Central Plains 35 40.7 -104.8 Grassland of America Experimental Range Temperate United States Colorado: Long term Ecological 36 40.5 -104.8 Grassland of America Research Station Temperate 37 China 37.6 101.7 Shidi Grassland Temperate 38 China 37.7 101.7 Gancaitan Grassland Temperate 39 China 30.9 91.1 Dangxiong Grassland Temperate 40 China 43.6 116.7 Neimeng Grassland 41 Steppe Tunisia 35.8 9.2 Southern Tunisia Inner Mongolia Grassland 42 Steppe China 43.5 116.7 Ecosystem Research Station United States Nevada: Mojave Global Change 43 Desert 36.6 -115.7 of America Facility United States 44 Desert 36.9 -116.6 Nevada test site of America 45 Desert Israel 30.9 34.4 Negev Desert United States 46 Desert 32.0 -112.9 Arizona: Ajo Mountains of America Arizona: Walnut Gulch United States 47 Desert 31.7 -110.1 Experimental Watershed (Sonoran of America and Chihuahuan Deserts)

20 Ecoregion Country Latitude Longitude Location Arizona: Walnut Gulch United States 48 Desert 31.7 -110.1 Experimental Watershed (Sonoran of America and Chihuahuan Deserts) United States Nebraska, near Central City (Platte 49 Wetland 41.1 -97.9 of America River) 50 Agricultural Australia -34.3 142.2 Red Cliffs 51 Agricultural France 43.9 1.2 Auradé 52 Agricultural France 43.8 1.4 Lamasquère Kahoku Expt. Watershed, Kyushu 53 Temperate Forest Japan 33.1 130.7 Island Temperate United States 54 Deciduous 36.0 -84.3 Oak Ridge, Tennessee of America Forests Mediterranean 55 Israel 31.4 35.0 Yatir Forest Shrubland United States 56 Desert 31.9 -110.8 Arizona: Sonoran Desert of America Arizona: Walnut Gulch United States 57 Desert 31.7 -110.1 Experimental Watershed (Sonoran of America and Chihuahuan Deserts) Arizona: Walnut Gulch United States 58 Desert 31.7 -110.1 Experimental Watershed (Sonoran of America and Chihuahuan Deserts) United States 59 Desert 31.4 -110.4 Huachuca Mountains of America United States 60 Desert 31.4 -110.4 Huachuca Mountains of America 61 Desert Israel 32.8 35.2 Alon ha’Galil 62 Agricultural Argentina -28.6 -66.8 Northwestern Argentina Vanuatu Agricultural Research and 63 Agricultural Vanuatu -15.4 167.2 Technical Center United States 64 Temperate Forest 34.6 -111.8 Arizona, Beaver Creek of America United States 65 Temperate Forest 34.0 -85.8 Southeastern U.S.A. of America United States 66 Temperate Forest 35.1 -83.4 Cowetta (Pine) of America United States 67 Temperate Forest 44.2 -122.3 Oregon, Andrews watershed of America Temperate United States 68 Deciduous 35.1 -83.4 Cowetta (Hardwood) of America Forests 69 Steppe Argentina -45.0 -70.0 Southern Argentina United States 70 Desert 35.8 -116.1 Mojave Desert of America

21 Ecoregion Country Latitude Longitude Location Tropical 71 D.R. Congo -4.7 12.1 Pointe–Noire Rainforest Tropical 72 D.R. Congo -4.7 12.1 Pointe–Noire Grassland 73 Temperate Forest New Zealand -43.2 170.3 Okarito Forest, Westland Temperate 74 Deciduous Australia -32.3 117.9 Corrigin, Western Australia Forests Temperate 75 Deciduous Portugal 38.5 -8.0 Herdade da Alfarrobeira Forests Temperate 76 Deciduous France 48.7 7.1 Hesse Forests Temperate United States 77 Deciduous 46.2 -89.3 Ottawa National Forest of America Forests 78 Boreal Forest Sweden 60.0 17.3 Uppsala 79 Boreal Forest Sweden 60.0 17.3 Uppsala 80 Desert China 39.8 99.5 Heihe River Basin United States 81 Desert 32.5 -106.8 Jornada Experimental Range of America

22 Table 1-1. (continued)

T E Q T/ET Method P* (% (% of (% of Reference (%) of P) P) P) 1 - 1623 45 56 45 Banerjee in Galoux et al., 1981 Lutzke & Simon in Galoux et al., 2 - 626 50 49 50 1981 Lutzke & Simon in Galoux et al., 3 - 627 40 48 41 1981 4 - 710 47 39 55 Rutter cited in Galoux et al., 1981 Molchanov cited in Galoux et al., 5 - 513 49 36 5 58 1981 Ten studies by Molchanov 1963, 6 - 502 39 35 53 cited by Choudhury et al., 1998 Two studies by Delfs (1967), cited 7 - 1237 19 26 42 by Choudhury et al., 1998 Diffusion 8 365 65a Trlica and Biondini, 1990 porometer Diurnal water table 9 3725 14 9 61 Frangi and Lugo, 1985 changes Energy balance 10 725 37 22 41 63 Tajchman, 1972 model Energy balance 11 150 38 62 38 Liu et al., 2012 model Isotope-based 12 340 71 18 12 81 Gibson and Edwards, 2002 (catchment) Isotope-based 13 872 45 8 85 Telmer and Veizer, 2000 (catchment) Isotope-based 14 310 34 8 58 80 Gibson and Edwards, 2002 (catchment) Isotope-based 15 1410* 32b 68 32 Hsieh et al., 1998 (stand level) Isotope-based 16 1410* 59 41 59 Hsieh et al., 1998 (stand level) Isotope-based 17 1380 61 39 61 Hsieh et al., 1998 (stand level) Isotope-based 18 2500 72 28 72 Hsieh et al., 1998 (stand level) Isotope-based 19 329 93 Ferretti et al., 2003 (stand level) Isotope-based 20 911 65-77 Wang et al., 2013 (stand level) Model (with met. 21 2000 49 26 26 65 Salati and Vose, 1984 data) Model (with met. 22 2000 62 19 19 77 Salati and Vose, 1984 data) Model (with met. 23 2232* 40 10 50 80 Shuttleworth, 1988 data) 23 T E Q T/ET Method P* (% (% of (% of Reference (%) of P) P) P) Model (with met. 24 2209 56 11 32 84 Leopoldo et al., 1995 data) Model (with met. 25 2851 31 21 60 Calder et al., 1986 data) Model (with met. 26 366 52a 53 52 Prazak et al., 1994 data) Model (with met. 27 595 59 36 55 Gash and Stewart, 1997 data) Model (with met. 28 2620 7 23 23 Hudson, 1988 data) Model (with met. 29 234 93 17 84 Dolman, 1988 data) Model (with met. 28- 30 590 Gebauer et al., 2012 data) 47 Model (with met. 31 549 54 9 86 Ladekari, 1998 data) Model (with met. 32 475 60 40 60 Poole et al., 1981 data) Model (with met. 33 475 32 51 4 39 Poole et al., 1981 data) Model (with met. 34 590 35 55 10 39 Poole et al., 1981 data) Model (with met. 35 335 46 54 51 Lauenroth and Bradford, 2006 data) Model (with met. 36 379* 67 33 0 67 Massman, 1992 data) Model (with met. 37 350 39 73 39 Hu et al., 2009 data) Model (with met. 38 477 37 67 37 Hu et al., 2009 data) Model (with met. 39 580 56 83 56 Hu et al., 2009 data) Model (with met. 40 580 39 49 44 Hu et al., 2009 data) Model (with met. 41 144 45 55 0 45 Floret et al., 1982 data) Model (with met. 42 275 55a 34 62 Huang et al., 2010 data) Model (with met. 43 74 40a 60 40 Young et al., 2009 data) Model (with met. 44 150 35 65 35 Smith et al., 1995 data) Model (with met. 45 170 41 Littman and Veste, 2006 data) Model (with met. 46 200 80 20 80 Liu et al., 1995 data)

24 T E Q T/ET Method P* (% (% of (% of Reference (%) of P) P) P) Model (with met. 47 223 64 64 Moran et al., 2009 data) Model (with met. 48 233 79 79 Moran et al., 2009 data) Model (with met. 49 687* 63a Kabenge and Irmak, 2012 data) Model (with met. 50 476 67 63 52 Yunusa et al., 1997 data) Model (with met. 51 615 42 51 46 Beziat et al., 2013 data) Model (with met. 52 684 23 53 33 Beziat et al., 2013 data) Model (with met. 53 2128 23 20 53 Kumagai et al., (in press) data), sap flow Model (with met. 54 1333 19a 14 58 Wilson et al., 2001 data), sap flow Model (with met. 55 285 45 46 48 Raz-Yaseef et al., 2012 data), sap flow Model (with met. 56 212 21a 27 47 Cavanaugh et al., 2011 data), sap flow Model (with met. 57 260 21a 36 42 Cavanaugh et al., 2011 data), sap flow Model (with met. 58 322 37 63 58 Scott et al., 2006 data), sap flow Model (with met. 59 400 >45 Ffolliott et al., 2003 data), sap flow Model (with met. 60 477 >75 Ffolliott et al., 2003 data), sap flow Model (with met. 61 515 >40 Ffolliott et al., 2003 data), sap flow Model (with met. 62 455* 70-80 Rousseaux et al., 2009 data), sap flow Model (with met. 63 2763 68 Roupsard et al., 2006 data), sap flow 64 Modelled (no obs.) 1085 49 15 41 76 Waring et al., 1981 65 Modelled (no obs.) 1225 49 15 38 77 McNulty et al., 1996 66 Modelled (no obs.) 2175 35 15 46 70 Waring et al., 1981 67 Modelled (no obs.) 2355 16 11 72 59 Waring et al., 1981 68 Modelled (no obs.) 2175 28 12 55 70 Waring et al., 1981 69 Modelled (no obs.) 150 34 56 10 38 Paruelo and Sala, 1995 70 Modelled (no obs.) 165 27 73 27 Lane et al., 1984 71 Radial flow meter 1019 81 12 87 Nizinski et al., 2011 72 Radial flow meter 1019 58 11 84 Nizinski et al., 2011 73 Sap flow 1127 8a 12 80 39 Barbour et al., 2005 74 Sap flow 265 53 78 40 Mitchell et al., 2009 25 T E Q T/ET Method P* (% (% of (% of Reference (%) of P) P) P) 75 Sap flow 669 73 27 73 Paco et al., 2009 76 Sap flow 763 33a 15 69 Granier et al., 2000 77 Sap flow 896 65a Tang et al., 2006 78 Sap flow 250 46a 25 65 Grelle et al., 1997 79 Sap flow 271 51a Cienciala et al., 1997 38- 80 Satellite-based 285 Tian et al., 2013 73 Water-balance; 81 control and bare 210 72 28 72 Schlesinger et al., 1987 plots

Compiled transpiration/evapotranspiration ratios range from minimums of 23% (United

Kingdom, East of Aberystwyth; Hudson, 1988) to 93% (Colorado: Central Plains Experimental

Range; Ferretti et al., 2003; Table 1-1). The average T/ET ratio for compiled studies is 60%.

Compiled transpiration/evapotranspiration ratios are found to be highest in the tropics (e.g., tropical rainforest T/ET of 70%±14%, tropical grassland T/ET of 62%±19%) and lower in Mediterranean climates (47%±10%; Table 1-2; Figure 1-4). The highest evapotranspiration fluxes off of the continents are from tropical regions. Spatially-weighting transpiration/evapotranspiration to the percent of terrestrial evapotranspiration accounted for by each biome yields a transpiration/evapotranspiration ratio of ~61%. This ratio is equivalent to a transpiration/evaporation ratio of ~1.5, or a 50% greater transpiration flux than evaporation flux on continents.

26 Table 1-2. Compiled transpiration/evapotranspiration (Schlesinger and Jasechko, 2014)

Percent Percent of T/ET percent Land P ET of Ecoregion land average ±1 s.d. area (%) (mm/yr) (mm/yr) terrestrial precipitation ET Tropical 70±14 (n = 8) 16 1830 35 1076 33.1 Rainforest Tropical 62±19 (n = 5) 12 950 14 583 13.9 Grassland Temperate Deciduous 67±14 (n = 9) 9 850 10 549 10.1 Forests Boreal Forest 65±18 (n = 5) 14 500 8 356 9.5 Temperate 57±19 (n = 8) 8 470 5 332 5.4 Grassland Desert 54±18 (n = 14) 18 180 4 209 7.3 Temperate Coniferous 55±15 (n = 13) 4 880 4 458 3.4 Forest Steppe 48±12 (n = 3) 4 440 2 467 3.4 Mediterranean 47±10 (n = 4) 2 480 1 302 1.0 shrubland

The stand level measurements were scaled up in Schlesinger and Jasechko (2014) to estimate global fluxes. However, we note that transpiration/evapotranspiration ratios in some studies neglect understory transpiration fluxes, suggesting that the reported terrestrial transpiration/evapotranspiration flux is likely to be a low end member of the actual terrestrial transpiration/evapotranspiration ratio (Schlesinger and Jasechko, 2014).

The compiled data showed little spatial coherence in transpiration/evapotranspiration ratios.

First, studies completed at the same research site produced very different estimates of transpiration/evapotranspiration. For example, Cavanaugh et al. (2011) and Moran et al. (2009) both investigated a research site in Arizona (U.S.A.) and produced transpiration/evapotranspiration estimates of 42% and 79%, respectively. The difference in these two results highlights the difficulty

27 associated with measuring transpiration fluxes and the uncertainties coupled to the scaling of point

(often tree-size scale) observations up to regional scales.

No trend was observed between precipitation rates and transpiration/evapotranspiration ratios within the compiled data (Figure 1-5), highlighting that climate is not the only control upon ecosystem productivity and primary production. Indeed, satellite based investigations of climate controls upon terrestrial primary production reveals a three tier set of controls that includes temperature, sunlight and water. Water is limiting in arid and semi-arid regions, but is a less important control in other regions (e.g., the Amazon basin; Running et al., 2004). Primary production in cold regions — which cover half of Earth’s of ice-free land surfaces (Jasechko et al., in review) under the definition of Bates and Bilello (1966) — is limited by temperature and sunlight, and primary production in tropical forests is limited by sunlight (Running et al., 2004).

Figure 1-5. Transpiration/evapotranspiration ratios compared to site-specific precipitation rates.

Reproduced from Schlesinger and Jasechko (2014).

28

Figure 1-6. Compiled estimates of transpiration/evapotranspiration ratios sorted into study approach. Whiskers mark the 10th-90th percentile range of the data, the shaded rectangles mark the

25th-75th percentile range, the black line marks the median of each dataset.

Reducing the dataset in Figure 1-5 to include only studies within desert and steppe biomes

— which are expected to broadly be water-limited ecosystems — improves the trend between precipitation and transpiration/evapotranspiration ratios slightly (R2 of 0.07; Figure 1-7) over the entire, global compilation (R2 of 0.01). This suggests that further site-specific studies in deserts could help to enhance our understanding of how water-limited ecosystems might respond to changes in precipitation amounts.

29

Figure 1-7. Arid region transpiration/evapotranspiration ratios compared to site-specific precipitation rates. A regression through the data reveals a significant (p < 0.05) trend towards higher transpiration/evapotranspiration ratios with increasing precipitation amount.

Different methodologies used to calculate transpiration fluxes are found to produce slightly different transpiration/evapotranspiration ratios. Isotope-based studies have a higher average transpiration/evapotranspiration ratio of ~70%, whereas sap flow and meteorological models suggest an average transpiration/evapotranspiration ratio of ~55% (Figure 1-6).

Several general circulation model based estimates of transpiration fluxes have been reported over the past decade. The general circulation model estimates of transpiration/evapotranspiration ratios are shown in Figure 1-1All alongside the compiled stand level data.

Generally, the GCMs have lower transpiration/evapotranspiration ratios than those suggested by stand level measurements. GCM transpiration/evapotranspiration ratios range from

25% to 65%, whereas stand level transpiration measurements indicate a global transpiration flux of closer to 60%, although this is likely to be a low end-member because many transpiration studies do not include understory transpiration fluxes. Lawrence et al. (2007) first pointed out that the

Community Land Model (version 3) was underpredicting transpiration fluxes. General circulation 30 model estimates of transpiration/evapotranspiration range from 13 % (Community Land Model 3, without improvements made by Lawrence et al., 2007), to 65 % (Lund–Potsdam–Jena model; Gerten et al., 2005). A global land surface model that integrates satellite data (Miralles et al., 2011) has a transpiration/evapotranspiration ratio of 80 %. A biophysical model developed by Choudhury et al.

(1998) proposes a transpiration/evapotranspiration ratio of 52 %. The broad range of transpiration/evapotranspiration ratios estimated by earlier works (Table 1-3) highlights the immense challenge of estimating this ratio. Upscaling a compilation of stand level measurements

(transpiration/evapotranspiration ratio of 61 %; Schlesinger and Jasechko, 2014) and a continental- scale isotope-based approach (transpiration/evapotranspiration of 80 to 90 %; Jasechko et al., 2013) suggest that the majority of general circulation models underestimate the role of transpiration in the global , and that transpiration is the largest water flux from Earth’s continents.

31 Table 1-3. Global transpiration estimates. Climate model Transpiration / Evapotranspiration Reference Community Land Model 13 % (prior to improvements by Lawrence et al., 2007 (version 3) Lawrence et al., 2007) Community Land Model 41 % (with improvements by Lawrence et al., 2007 (version 3) Lawrence et al., 2007) Community Land Model 3.5 coupled to Community 40 % Cao et al., 2010 Atmosphere Model 3.5 Joint UK Land Environment 38 % to 48 % Alton et al., 2009 Simulator Lund–Potsdam–Jena model 65 % Gerten et al., 2005 Global Soil Wetness Project 48 % Dirmeyer et al., 2006 n/a 52 % Choudhury et al., 1998 Global Land-surface Evaporation: the Amsterdam 80 % Miralles et al., 2011 Methodology Vegetation Integrative Simulator 24 % Ito and Inatomi, 2012 for Trace Gases

1.3 Dataset and methods

The development of isotope-based transpiration calculations is divided into three sections:

(i) development of a global lake water O and H isotope database and geospatial analysis of lake catchments, (iii) calculation setup, geospatial data extraction, and analysis.

1.3.1 Development of a global lake water O and H isotope database

To develop a continental scale estimate of transpiration fluxes we required a continental scale isotopic dataset. A lake-by-lake compilation of isotopic data was completed over four months

(September 2011 to December 2011) and the resulting compilation was presented at the American

Geophysical Union Fall Meeting in December of 2011 (Jasechko et al., 2011). 32 The dataset spans lakes from all continents with the exception of Antarctica. The dataset contains 2129 measurements of δ18O and 2098 measurements of lake water δ2H from 73 unique lakes compiled from 61 published datasets. Only lakes with surface areas on the order of 102-104 km2 were included in the large lake isotopic database.

The location of each of the compiled lakes is presented in Tables 1-4 and 1-5 and Figure 1-8.

Large lakes are concentrated geographically into scoured basins at the margins of the Laurentide and

Fennoscandanavian ice sheets that resided over North America and Eurasia during the last ice age.

These lakes include Great Bear, Great Slave, Lake Winnipeg, Lake Superior, Lake Huron, Lake

Michigan, Lake Erie and Lake Ontario in North America, and Lake Ladoga and Lake Onega in

Eurasia. Other large lakes are concentrated in geological rift valleys and include Lake Baikal and Lake

Tanganyika, which combine to a total volume that comprises more than one-third of all fresh water at Earth’s surface. The majority of compiled lakes are exorheic (i.e., externally drained), with a minority of endorheic (i.e., closed basin) lakes that include the Aral and Caspian Seas, Great Salt

Lake, and Lake .

Before analyzing hydroclimate and hydrological data for each lake, the potential contributing area to each lake was quantified by delineating watersheds for each of the 73 lakes in our database.

Lake catchment areas were delineated using the Shuttle Radar Topography Mission

(www2.jpl.nasa.gov/srtm) Digital Elevation Model and global river spatial data (waterbase.org).

Catchments were delineated by hand in a geographic information system on the basis of topographic highs from the Shuttle Radar Topography Mission data and drainage basin data.

33

Figure 1-8. Locations of lake catchments studied in chapter one. The entire set of catchments covers

~10% of Earth’s surface. Small catchments are delineated with diamonds for clarity. Insets are shown for the western region of North America, eastern Africa and the Tibetan plateau.

34 Table 1-4. Lake information Elevation Lake Basin type Outflow Lat. Lon. (m.a.s.l.) Abhe Endorheic - 11.1 41.8 240 Abiyata Endorheic - 7.8 38.7 1573 Afdera Endorheic - 13.3 40.9 -100 Albert Chain White Nile 1.7 30.9 615 Aral Sea Endorheic - 45.1 58.3 53 Athabasca Headwater Slave River 59.0 -110.0 213 Awasa Endorheic - 7.1 38.5 1708 Baikal Headwater Angara River 53.1 107.7 450 Baringo Endorheic - 0.6 36.1 970 Beysehir Endorheic - 37.7 31.5 1116 Biwa Headwater Seta River 35.3 136.2 86 Caspian Endorheic - 42.0 51.0 -28 Chad Endorheic - 13.0 14.2 244 Chamo Endorheic - 5.9 37.6 1110 Dagze Co Endorheic - 31.9 87.6 4478 Dead Sea Endorheic - 31.3 35.5 -420 Edward Headwater Semliki River -0.4 29.6 912 Egridir Endorheic - 38 30.9 924 Elephant Butte Headwater Rio Grande 33.4 107.2 1312 Erie Chain Niagara River 42.5 -79.6 173 Garda Headwater Mincio 45.6 10.7 65 Geneva Headwater Rhone River 46.4 6.6 372 Great Bear Headwater Great Bear R. 66.0 -120.0 156 Great Salt Endorheic - 41.2 -112.6 1270 Great Slave Chain Mackenzie R. 61.8 -114 176 Huron Chain St. Clair River 43.5 -82 176 Issyk-Kul Endorheic - 42.5 77.3 1600 Jackson Headwater Snake River 43.9 -110.6 2067 Kainji Headwater Niger River 10.4 4.6 139 Kivu Headwater Ruzizi River -2.0 29.0 1460 Kluane Headwater Kluane River 61.1 -138.5 781 Ladoga Chain Neva River 60.8 31.4 11 Lucern Headwater Reuss River 47.0 8.4 433 Malawi Headwater Shire River -12.0 34.5 471 Manasarovar Endorheic - 30.7 81.5 4584 Mar Chiquita Endorheic - -30.5 -62.7 67

35 Table 1-4. Lake information (continued)

Elevation Lake Basin type Outflow Lat. Lon. (m.a.s.l.)

Mead Chain Colorado River 36.1 -114.7 367 Michigan Chain Mackinac 42.4 -87.0 176 Naivasha Headwater - -0.8 36.4 1884 Namco Endorheic - 30.7 90.6 4718 Nasser Chain Nile 22.3 31.7 179 Ngangla Ringco Endorheic - 31.4 83.4 4724 Nicaragua Headwater San Juan River 11.2 -85.5 31 Oahe Chain Missouri River 44.4 -100.4 490 Okanagan Headwater Okanagan River 50.2 -119.4 345 Onega Headwater Svir River 61.9 35.4 56 Ontario Chain St. Lawrence R. 43.5 -79.4 86 Powell Headwater Colorado River 36.9 -111.5 1130 Poyang Headwater Changjiang 29.1 116.3 10 Qarhan Endorheic - 37.0 95.1 2685 Qinghai Hu Endorheic - 36.9 100.1 3200 Rukwa Endorheic - -8.4 32.7 800 Sakakawea Chain Missouri River 47.5 -101.4 561 Salton Sea Endorheic - 33.2 -115.7 -71 Sambhar Salt Endorheic - 27.0 75.1 360 Shala Endorheic - 7.4 38.6 1559 Superior Headwater St. Marys River 47.0 -85.2 183 Tahoe Headwater Truckee River 39.1 -120.1 1900 Tana Headwater Blue Nile 11.6 37.4 1790 Tanganyika Chain Rukuga River -4.9 29.5 773 Taro Co Endorheic - 31.1 84.3 4579 Taupo Headwater Waikato River -38.8 175.9 395 Titicaca Endorheic - -15.5 -69.4 3827 Tonlé Sap Chain Tonlé Sap River 11.6 104.9 14 Turkana Endorheic - 4.0 36.0 360 Valencia Endorheic - 10.2 -68.1 410 Van Endorheic - -38.7 -43.4 1646 Victoria Headwater White Nile -1.0 33.0 1133 Winnipeg Headwater Nelson River 52.1 -97.8 217 Yamdruk-tso Endorheic - 28.8 90.6 4458 Yellowstone Headwater Yellowstone R. 44.5 -110.4 2357 Endorheic - 31.1 85.4 4624 Zige Tangco Endorheic - 32.0 90.8 4575

36 Table 1-5. Physical hydrology of lakes Catchment area Open water Volume τ * Lake (km2) (km2) (km3) (years) Abhe 94200 1600 6 50 Abiyata 10400 830 1.4 5 Afdera 7100 110 6 90 Albert 58800 5900 132 3 Aral Sea 949500 77900 193 6 Athabasca 271100 26900 110 1.8 Awasa 1500 100 1 10 Baikal 583200 37900 23600 280 Baringo 6600 150 0.2 2.3 Beysehir 15400 880 2 2 Biwa 3700 680 28 6.5 Caspian 3024400 428800 78000 260 Chad 976300 26200 72 4 Chamo 1900 320 4 17 Dagze Co 12800 640 3 6 Dead Sea 43200 1200 136 160 Edward 26800 2800 77 6 Egridir 3300 480 10 25 Elephant Butte 89900 820 2.5 2 Erie 103700 27300 484 2.1 Garda 2200 370 50 28 Geneva 7900 700 90 14 Great Bear 148500 41100 2300 55 Great Salt 81900 7700 20 6 Great Slave 702200 73400 2090 11 Huron 192100 66000 3540 15 Issyk-Kul 22000 6300 1740 170 Jackson 2000 150 6 7 Kainji 1565300 12800 15 1 Kivu 7500 2400 350 60 Kluane 5500 400 12 9 Ladoga 225800 34400 850 9 Lucern 2200 120 12 3 Malawi 124900 29100 7775 250 Manasarovar 5100 490 20 20 Mar Chiquita 129700 3100 6 3 *τ: approximate residence time of each lake

37 Table 1-5. Physical hydrology of lakes (continued) Catchment area Open water Volume τ * Lake (km2) (km2) (km3) (years) Mead 147200 1100 25 2 Michigan 174400 60100 4920 64 Naivasha 3200 110 1 2 Namco 10700 1900 63 25 Nasser 2330500 16000 132 2 Ngangla Ringco 12500 610 5 6 Nicaragua 27900 8900 108 6 Oahe 150400 2400 25 1 Okanagan 6000 390 25 23 Onega 54500 12200 280 8 Ontario 82000 21000 1640 7 Powell 278600 3300 33 1.5 Poyang 161500 4100 3 1 Qarhan Salt 109700 820 50 8 Qinghai Hu 29600 4700 70 20 Rukwa 79300 3000 40 12 Sakakawea 461900 6400 29 1.4 Salton Sea 20000 930 9 3 Sambhar Salt 5900 20 0.2 1 Shala 4100 310 40 43 Superior 226200 92700 12100 88 Tahoe 1300 500 160 130 Tana 15000 3100 28 4 Tanganyika 230800 34000 19000 400 Taro Co 16800 830 5 6 Taupo 3500 630 60 12 Titicaca 56900 8600 900 160 Tonlé Sap 58800 ~3200 ~160 ~1 Turkana 180400 8700 200 40 Valencia 3000 360 6 10 Van 17100 3700 607 62 Victoria 264100 68400 2750 26 Winnipeg 1048200 88100 284 3 Yamdruk-tso 10000 1100 20 41 Yellowstone 2700 370 15 8 Zhari Namco 20100 1400 30 40 Zige Tangco 3300 190 3 19 *τ: approximate residence time of each lake

38 1.3.2 Calculation approach

To calculate transpiration rates we first developed a set of equations that can be applied to estimate transpiration/evaporation ratios. A hydrological catchment’s water balance can be described by water fluxes and changes to water storages (Equation 1.1): dV  I  xP  E T  Q dt Equation 1.1 where dV/dt is the rate of change in water storage in the catchment, I represents the flux of precipitation entering the catchment plus any upstream liquid inflows from chain lake systems, E represents physical evaporation losses from a catchment, T represents transpiration water losses from a catchment, Q represents liquid losses via stream discharges and via groundwater recharge and advection out of the basin, x represents the fraction of precipitation (P) that is intercepted by vegetation and returned to the atmosphere through evaporation. At steady state Equation 1.1 reduces to (Equation 1.2):

I  xP  E  T  Q Equation 1.2

In addition to the physical water balance, a steady state stable isotope mass balance can be described as (Equation 1.3):

 I I   P xP  E E TT QQ Equation 1.3

where δI is the flux-weighted isotopic composition of inputs (precipitation and chain lake inflows), δP is the isotopic composition of precipitation, δE is the isotopic composition of evaporating moisture

(isotope fractionation labelled), δT is the isotopic composition of water used by plants in transpiration

(not isotope fractionation labelled) and δi is the isotopic composition of intercepted rain and snow.

39 Combining equations 1 and 2 yields a single equation representing the transpiration flux exiting a hydrological catchment under steady state conditions (Equation 1.4):

I    Q    xP    T  I E Q E P E    T E Equation 1.4

1.3.3 Geospatial analysis

Next, each of the inputs into equation 1.4 were quantified using multiple global geospatial datasets. The following paragraphs examine each of the input parameters in equation 1.4 one by one.

The precipitation input to each hydrological catchment (P) was calculated using global high resolution precipitation data spanning the continents (New et al., 2002). The catchment area of each lake was calculated, as was the mean annual precipitation rate for each catchment. The two components were multiplied together to estimate the annual flux of precipitation inputs for each basin. Annual water inputs to each catchment (I) were calculated as the sum of precipitation inputs

(P) plus contributions from upstream chain lake systems.

The liquid fluxes out of each catchment (Q) were compiled on a river-by-river basis using data within the primary literature. These water fluxes were also used to quantify chain lake inflows into downstream lake basins where appropriate. Groundwater fluxes out of lakes were also collected on a lake-by-lake basin for endorheic basins with known connections with regional systems (e.g., Isiorho et al., 1996; Ojiambo et al., 2003).

The proportion of precipitation intercepted and returned to the atmosphere (x) was calculated using satellite-based grids developed by Miralles et al. (2010) coupled to annual precipitation rates (New et al., 2002).

40 δP, the flux-weighted isotopic composition of precipitation inputs entering each hydrological catchment. δP was estimated for each hydrological catchment considering seasonal and spatial variability in precipitation amounts. Geospatial grids of monthly precipitation isotope compositions are available for download from waterisotopes.org following methods of Bowen and Revenaugh

(2003), Bowen and Wilkinson (2002) and Bowen (2010). The seasonality of precipitation amount was quantified by flux weighting each grid call at a monthly time step (where δP(j) is the isotopic composition of precipitation for month j, and Pj represents the monthly precipitation rate (mm per month) at each grid cell). Our calculation also accounts for the spatial distribution of precipitation was included by weighting the individual grid cells to their respective precipitation amounts (i.e., grid cell i) following equation 1.5:

 12 P   n  j1 j Pj    P  i1 12 P  i   j1 j    i Equation 1.5 P n P  i1 i

The isotopic composition of water inputs (δI) to each catchment was calculated by flux weighting the isotopic compositions of precipitation (δP) against contributions from upstream lakes

(i.e., river inflows from upstream lakes).

The isotopic composition of evaporate from each catchment (δE) was calculated using an evaporation model (Craig and Gordon, 1965; Equation 1.6):

   */  * h     Lake A K Equation 1.6 E 1 h   K where δLake is the isotopic composition of lake water (compiled from primary literature), α* is an isotopic equilibrium fractionation factor (temperature dependent; temperature data from New et al.,

2002)), ε* is an equilibrium isotopic separation factor (approximated as: α* − 1), h is the relative humidity of the catchment (calculated for each catchment using geospatial data from New et al., 41 2002), δA is the isotopic composition of atmospheric vapor (estimated in two ways: once by using precipitation as a liquid signature of atmospheric vapor, and back calculating the vapor isotope composition using temperature data (New et al., 2002), and second by compiling δA values from an isotope-enabled general circulation model developed by Yoshimura et al., 2008) and εK is a kinetic isotopic separation factor calculated by CK·[1 – h] (Gonfiantini, 1986).

The isotopic composition of transpired moisture (δT) was estimated across the continents using isotopic data for precipitation and seasonality in primary productivity. We weighted the isotopic composition of precipitation15 spatially (i) to long-term monthly mean normalized difference vegetation indices (NDVI; proxy for chlorophyll abundance), with NDVI values below zero set to a value of zero. A range of two temporal (j) weighting approaches is used for δT, one weighted to growing season (representing shallow rooted end-member; Equation 1.7) and another to monthly precipitation (representing a deep rooted end-member, i.e., a phreatophyte; Equation 1.8):

 12 NDVI    n  j1 j Pj    NDVI  i1 12  i j1NDVIj   i TSHALLOW  n Equation 1.7  i1NDVIi  12 P   n  j1 j Pj   NDVI  i1 12  i j1Pj   i TDEEP  n Equation 1.8  i1NDVIi

Water use efficiency functions were compiled from the primary literature to develop catchment wide estimates of water use efficiency. A review of water use efficiency data as a function of humidity reveals differences between C3 and C4 plants (Table 1-9; reproduced from Jasechko et al.,

2013), such that a global grid of C3/C4 photosynthesis types was also sought after. We assessed spatial variability in C3/C4 species abundances using grids developed by Still et al. (2003), downloaded from http://webmap.ornl.gov/wcsdown/dataset.jsp?ds_id=932. Power regressions of

42 C3 and C4 datasets were applied to develop water use efficiency/climate relationships for each photosynthetic pathway: C3: Water use efficiency = 4.21×(Vapor pressure deficit)-0.67 and C4: Water use efficiency = 6.91×(Vapor pressure deficit)-0.40. Daytime vapor pressure deficit grids were then estimated at a monthly time step by averaging the maximum and average monthly mean temperatures at each grid cell (data from Hijmans et al., 2005) and catchment-wide water use efficiencies for each basin were calculated (Figure 1-9, 1-10). Resulting transpiration fluxes were then converted into gross primary productivity using the catchment water use efficiency data. The inputs for each calculation are presented in Tables 1-6 to 1-8.

43 Table 1-6. Model input parameters (± 1 s.d. uncertainty shown)

δ18OL δ2HL δ18OT δ2HT δ18OI δ2HI Lake (‰) (‰) (‰) (‰) (‰) (‰) Abhe 3.7±0.5 -4±4 0.2±1.4 12±11 -0.3±1 8±9 Abiyata 8.5±1.3 59±11 -1.8±1.3 -3±10 -2.0±1 -4±9 Afdera 6.4±0.5 28±2 0.9±1.5 14±11 0.5±1 13±9 Albert 5.2±0.5 37±4 -2.8±1.3 -10±10 -0.8±0.7 -1±7 Aral Sea 2.8±1.8 0±11 -5.8±2.7 -32±22 -9.9±1 -65±9 Athabasca -16.8±1.3 -131±4 -16.6±1.7 -127±14 -17.9±1 -137±9 Awasa 7.5±0.7 51±4 -1.7±1.4 -3±11 -2.3±1 -6±9 Baikal -15.8±0.3 -123±1 -12.2±1.2 -91±10 -12.3±1 -92±9 Baringo 7.5±1.3 42±8 -3.8±1.5 -17±12 -4.4±1 -21±9 Beysehir -1.4±0.9 -18±4 -6.8±1.9 -40±14 -8.2±1 -50±9 Biwa -7.0±0.5 -44±5 -8.1±1.2 -53±10 -8.0±1 -52±9 Caspian -1.7±0.2 -20±3 -10.5±2 -76±16 -11.6±1 -85±9 Chad 8.2±3.6 45±19 -1.8±1.9 -8±14 -3.2±1 -17±9 Chamo 7.7±0.8 50±3 -1.4±1.2 0±9 -1.4±1 0±9 Dagze Co -6.4±0.5 -69±4 -16.1±1.4 -115±10 -16.5±1 -117±9 Dead Sea 1.4±2.3 4±2 -5.1±1.6 -25±11 -6.1±1 -28±9 Edward 4.2±0.2 30±1 -3.5±1.3 -15±11 -3.7±1 -17±9 Egridir -2.4±0.6 -21±2 -7.4±1.7 -48±13 -8.5±1 -53±9 Elephant Butte -7.8±1.2 -68±6 -12.9±1.2 -91±10 -12.8±1 -93±9 Erie -6.6±0.3 -48±9 -7.1±1.9 -46±15 -7.6±0.5 -55±9 Garda -7.3±0.2 -55±1 -8.0±1.2 -52±9 -7.8±1 -51±9 Geneva -12.3±0.1 -88±2 -9.1±1.9 -61±14 -10.9±1 -74±9 Great Bear -18.7±0.5 -155±4 -17.2±3.7 -130±30 -22.3±1 -171±9 Great Salt -4.8±1.0 -67±9 -12.5±2.4 -93±18 -14.8±1 -110±9 Great Slave -17.8±0.3 -141±3 -16.5±2.3 -127±18 -18.6±0.9 -143±9 Huron -7.1±0.1 -54±2 -8.1±2.2 -54±17 -9.1±0.6 -64±6 Issyk-Kul -0.7±0.1 -9±2 -10.2±1.5 -62±15 -10.6±1 -72±9 Jackson -17.9±0.1 -141±4 -13.3±3.2 -96±26 -17.4±1 -129±9 Kainji -17±11 -2.2±2.2 -12±16 -4.5±1 -27±9 Kivu 1.5±1.4 18±6 -4.3±1.4 -20±12 -4.7±1 -25±9 Kluane -22.6±0.5 -177±3 -18.5±2.3 -149±16 -21.8±1 -169±9 Ladoga -9.5±0.5 -10.8±1.8 -78±15 -11.9±0.9 -76±8 Lucern -12.7±0.5 -9.9±1.7 -67±13 -11.0±1 -75±9 Malawi 2.0±0.1 12±1 -3.3±2 -14±16 -4.6±1 -24±9 Manasarovar -5.5±3.5 -58±16 -17.1±1.5 -112±12 -16.4±1 -117±9 Mar Chiquita 3.1±0.2 18±1 -5.4±1.4 -32±10 -5.0±1 -31±9

44 Table 1-6. Model input parameters (± 1 s.d. uncertainty shown; continued)

δ18OL δ2HL δ18OT δ2HT δ18OI δ2HI Lake (‰) (‰) (‰) (‰) (‰) (‰) Mead -12.9±0.7 -103±5 -11.4±1.5 -85±12 -12.9±0.9 -97±8 Michigan -5.8±0.1 -44±1 -7.9±1.8 -54±14 -8.6±0.8 -61±8 Naivasha 4.6±1.1 26±7 -5.2±1.3 -28±10 -5.5±1 -29±9 Namco -7.3±0.4 -70±3 -17.8±1.3 -130±11 -17.5±1 -126±9 Nasser 0.0±1.4 8±8 -0.9±1.5 1±12 -1.5±1 -3±9 Ngangla Ringco -4.2±0.5 -57±4 -16.6±1.2 -120±10 -16.6±1 -118±9 Nicaragua -2.0±0.5 -9±4 -4.6±1.8 -27±15 -5.7±1 -37±9 Oahe -14.2±0.2 -116±2 -11.8±1.4 -88±11 -13.0±0.8 -98±8 Okanagan -11.4±0.5 -103±3 -12.8±1.9 -99±15 -14.5±1 -111±9 Onega -10.4±0.7 -10.8±2.2 -78±18 -12.8±1 -95±9 Ontario -6.6±0.1 -49±1 -8.2±2.1 -51±17 -7.4±0.3 -54±3 Powell -15.0±0.2 -115±2 -12.6±2.1 -93±16 -14.6±1 -107±9 Poyang -6.8±1.1 -38±9 -7.0±1.4 -46±11 -6.5±1 -42±9 Qarhan Salt 6.6±0.5 -16±4 -12.6±1.3 -92±10 -13.3±1 -95±9 Qinghai Hu 2.4±0.7 12±5 -12.5±1.4 -86±10 -12.0±1 -86±9 Rukwa 4.3±0.2 26±2 -3.6±1.7 -17±14 -4.7±1 -26±9 Sakakawea -15.5±0.2 -124±1 -13.3±1.8 -99±14 -14.5±1 -109±9 Salton Sea -3.6±2.4 -52±12 -6.5±1.8 -52±14 -8.4±1 -66±9 Sambhar Salt 11.5±11.2 -4.9±1.2 -31±10 -5.0±1 -30±9 Shala 7.5±0.7 52±3 -1.3±1.2 1±10 -1.4±1 0±9 Superior -8.6±0.1 -66±1 -9.6±2.1 -67±17 -11.4±1 -81±9 Tahoe -5.5±0.3 -59±16 -11.2±2.5 -87±18 -13.8±1 -103±9 Tana 4.5±0.9 35±6 -2.4±1.4 -9±11 -2.7±1 -11±9 Tanganyika 3.8±0.4 26±2 -3.3±1.6 -15±13 -4.0±1 -20±9 Taro Co -5.6±0.5 -68±4 -16.3±1.3 -118±10 -16.6±1 -119±9 Taupo -5.3±0.4 -33±3 -7.1±1.3 -44±10 -6.9±1 -43±9 Titicaca -3.8±0.7 -50±3 -12.7±1.7 -86±14 -13.6±1 -94±9 Tonlé Sap -5.2±1.0 -5.5±1.5 -34±12 -7.0±1 -25±6 Turkana 5.6±0.4 38±4 -1.6±1.2 -1±10 -1.7±1 -2±9 Valencia 22±4 -4.1±1.4 -29±11 -4.6±1 -32±9 Van 1.0±0.1 -7±0 -7.4±2.9 -45±22 -10.7±1 -70±9 Victoria 3.5±0.5 -3.5±1.3 -15±11 -3.6±1 -16±9 Winnipeg -10.4±0.5 -79±8 -11.6±2.3 -92±15 -14.3±1 -107±9 Yamdruk-tso -5.5±0.5 -68±4 -18.0±1.8 -136±17 -16.7±1 -121±9 Yellowstone -16.5±0.2 -135±4 -15.1±2.6 -118±17 -17.8±1 -133±9 Zhari Namco -6.7±0.5 -75±4 -17.3±1.3 -122±10 -17.0±1 -122±9 Zige Tangco -6.1±0.5 -68±4 -16.1±1.6 -115±12 -17.0±1 -122±9

45 Table 1-7. Model input parameters (± 1 s.d. uncertainty shown)

δ18OP δ2HP δ18OA δ2HA TL hA Lake (‰) (‰) (‰) (‰) (°C) (%) Abhe -0.3±1 8±9 -7.7±2 -51±33 26.1±1 67±3 Abiyata -2.0±1 -4±9 -10.5±1 -73±12 20.0±1 59±3 Afdera 0.5±1 13±9 -7.1±3 -46±37 27.1±1 68±3 Albert -2.9±1 -11±9 -11.3±1 -77±9 24.4±1 69±3 Aral Sea -9.9±1 -65±9 -18.4±1 -142±9 14.2±1 55±3 Athabasca -17.9±1 -137±9 -35.5±4 -288±85 -9.3±1 74±3 Awasa -2.3±1 -6±9 -11.5±1 -81±9 18.0±1 59±3 Baikal -12.3±1 -92±9 -30.6±2 -251±53 -8.4±1 79±3 Baringo -4.4±1 -21±9 -12.4±1 -85±9 24.2±1 54±3 Beysehir -8.2±1 -50±9 -16.7±1 -123±21 14.0±1 56±3 Biwa -8.0±1 -52±9 -18.0±2 -134±30 14.9±1 75±3 Caspian -11.6±1 -85±9 -17.7±1 -136±27 16.0±1 68±3 Chad -3.2±1 -17±9 -9.3±3 -69±39 27.4±1 36±3 Chamo -1.4±1 0±9 -10.4±1 -72±18 21.9±1 57±3 Dagze Co -16.5±1 -117±9 -28.9±4 -225±66 0.1±1 58±3 Dead Sea -6.1±1 -28±9 -13.1±1 -91±9 23.3±1 56±3 Edward -3.7±1 -17±9 -12.2±1 -85±9 23.6±1 70±3 Egridir -8.5±1 -53±9 -17.1±2 -126±26 14.7±1 55±3 Elephant Butte -12.8±1 -93±9 -25.6±2 -200±47 5.4±1 49±3 Erie -8.7±1 -58±9 -19.4±2 -146±30 11.0±1 74±3 Garda -7.8±1 -51±9 -16.0±1 -118±15 18.0±1 75±3 Geneva -10.9±1 -74±9 -17.2±2 -129±28 13.8±1 71±3 Great Bear -22.3±1 -171±9 -39.3±5 -318±96 -14.2±1 73±3 Great Salt -14.8±1 -110±9 -21.7±1 -171±24 16.2±1 46±3 Great Slave -18.9±1 -145±9 -36.9±4 -300±87 -11.6±1 74±3 Huron -10.1±1 -69±9 -23.7±2 -183±38 1.6±1 79±3 Issyk-Kul -10.6±1 -72±9 -20.7±1 -158±31 9.7±1 52±3 Jackson -17.4±1 -129±9 -24.9±3 -197±51 7.0±1 51±3 Kainji -4.5±1 -27±9 -9.2±3 -68±42 28.2±1 41±3 Kivu -4.7±1 -25±9 -13.8±1 -98±10 19.5±1 74±3 Kluane -21.8±1 -169±9 -31.0±4 -251±77 3.8±1 75±3 Ladoga -12.1±1 -90±9 -23.7±3 -188±52 4.9±1 85±3 Lucern -11.0±1 -75±9 -18.9±2 -143±38 10.9±1 76±3 Malawi -4.6±1 -24±9 -12.1±1 -84±9 24.1±1 71±3 Manasarovar -16.4±1 -117±9 -27.1±6 -209±98 0.5±1 60±3 Mar Chiquita -5.0±1 -31±9 -14.6±1 -107±12 21.2±1 71±3

46 Table 1-7. Model input parameters (± 1 s.d. uncertainty shown; continued)

δ18OP δ2HP δ18OA δ2HA TL hA Lake (‰) (‰) (‰) (‰) (°C) (%) Mead -12.0±1 -89±9 -19.5±1 -152±12 19.9±1 34±3 Michigan -9.0±1 -62±9 -22.6±2 -175±38 4.3±1 76±3 Naivasha -5.5±1 -29±9 -15.4±1 -111±17 14.0±1 68±3 Namco -17.5±1 -126±9 -28.6±4 -223±77 1.3±1 57±3 Nasser -1.5±1 -3±9 -10.1±1 -68±20 27.7±1 44±3 Ngangla Ringco -16.6±1 -118±9 -28.0±4 -217±79 -0.4±1 58±3 Nicaragua -5.7±1 -37±9 -13.7±1 -99±9 26.6±1 80±3 Oahe -12.3±1 -91±9 -23.4±2 -183±43 11.1±1 62±3 Okanagan -14.5±1 -111±9 -22.8±2 -184±42 10.0±1 60±3 Onega -12.8±1 -95±9 -24.4±3 -194±53 4.1±1 86±3 Ontario -9.9±1 -66±9 -22.8±2 -174±37 4.4±1 76±3 Powell -14.6±1 -107±9 -23.2±1 -181±29 11.9±1 46±3 Poyang -6.5±1 -42±9 -16.4±1 -121±20 21.8±1 76±3 Qarhan Salt -13.3±1 -95±9 -25.9±2 -201±44 2.4±1 48±3 Qinghai Hu -12.0±1 -86±9 -23.9±2 -185±35 4.2±1 49±3 Rukwa -4.7±1 -26±9 -12.6±1 -88±9 23.9±1 67±3 Sakakawea -14.5±1 -108±9 -25.4±3 -200±50 9.1±1 59±3 Salton Sea -8.4±1 -66±9 -15.5±1 -126±9 25.1±1 51±3 Sambhar Salt -5.0±1 -30±9 -12.7±1 -92±9 27.2±1 43±3 Shala -1.4±1 0±9 -10.5±1 -73±11 19.8±1 59±3 Superior -11.4±1 -81±9 -27.3±3 -215±54 -3.1±1 75±3 Tahoe -13.8±1 -103±9 -22.5±2 -181±46 7.9±1 53±3 Tana -2.7±1 -11±9 -11.1±1 -78±9 20.1±1 53±3 Tanganyika -4.1±1 -21±9 -12.2±1 -85±9 23.8±1 71±3 Taro Co -16.6±1 -119±9 -27.0±4 -209±76 1.9±1 59±3 Taupo -6.9±1 -43±9 -16.4±2 -123±26 13.4±1 81±3 Titicaca -13.6±1 -94±9 -21.9±3 -165±57 8.4±1 57±3 Tonlé Sap -6.3±1 -40±9 -14.5±1 -104±9 27.6±1 80±3 Turkana -1.7±1 -2±9 -9.4±2 -62±26 29.0±1 52±3 Valencia -4.6±1 -32±9 -12.6±1 -98±14 23.4±1 74±3 Van -10.7±1 -70±9 -18.2±2 -136±34 12.0±1 57±3 Victoria -3.6±1 -16±9 -12.3±1 -86±9 22.2±1 72±3 Winnipeg -14.3±1 -107±9 -22.2±2 -174±36 12.9±1 69±3 Yamdruk-tso -16.7±1 -121±9 -26.9±4 -210±71 3.1±1 55±3 Yellowstone -17.8±1 -133±9 -25.3±3 -200±54 6.6±1 51±3 Zhari Namco -17±1 -122±9 -27.8±5 -216±91 1.7±1 59±3 Zige Tangco -17±1 -122±9 -28.9±2 -225±40 0.7±1 59±3

47 Table 1-8. Lake isotope investigations Lake n Reference Abhe 1 Kebede et al., 2009 Abiyata 7 Kebede et al., 2009; Craig et al., 1977 Afdera 11 Gonfiantini et al., 1973 Albert 1 Bahati et al., 2005 Aral Sea 36 Oberhansli et al., 2009 Athabasca 4 Hitchon and Karouse, 1972; Wolfe et al., 2007 Awasa 33 Kebede et al., 2009; Craig et al., 1977; Darling et al., 1996 Baikal 32 Seal and Shanks, 1998 Baringo 3 Cerling et al., 1988; Becht et al., 2005 Beysehir 11 Dincer et al., 1968 Biwa 15 Taniguichi et al., 2001 Caspian 25 Froehlich et al., 2000 Chad 95 Fontes et al., 1970 Chamo 13 Kebede et al., 2009 Dagze Co 1 Yuan et al., 2011 Dead Sea 27 Gat et al., 1984 Edward 4 Rossel et al., 2006 Egridir 10 Dincer et al., 1968 Elephant Butte 12 Phillips et al., 2003; This work Erie 151 Yang et al., 1996; Karim et al., 2008; Jasechko et al., 2014 Garda 176 Longinelli et al., 2008 Geneva 9 Fontes et al., 1970 Great Bear 32 Hitchon et Krouse, 1972; This work Great Salt 32 Nielson and Bowen, 2010 Great Slave 7 Hitchon and Krouse, 1972; Brock et al., 2009 Huron 142 Yang et al., 1996; Karim et al., 2008; Jasechko et al., 2014 Issyk-Kul 7 Ricketts et al., 2001 Jackson 2 This work Kainji 18 Zimmerman et al., 1976 Kivu 19 Cohen et al., 1997 Kluane 15 Brahney, 2007 Ladoga 3 Luz and Barkan et al., 2010 Lucern 1 Luz and Barkan et al., 2010 Malawi 21 Gonfiantini et al., 1979 Manasarovar 7 Yao et al., 2009 Mar Chiquita 31 Dapena et al., 1997

48 Table 1-8. Lake isotope investigations Lake n Reference Mead 12 Craig, 1966; This work Michigan 80 Jasechko et al., 2014 Naivasha 9 Darling et al., 1997; Cerling et al., 1988; Becht et al., 2005 Namco 2 Liu et al., 2009 Nasser 41 Aly et al., 1993 Ngangla Ringco 1 Yuan et al., 2011 Nicaragua 1 Lachniet et al., 2002 Oahe 11 Kendall and Coplen, 2001 Okanagan 36 Wassenaar et al., 2011 Onega 2 Luz and Barkan, 2010 Ontario 68 Yang et al., 1996; Karim et al., 2008; Jasechko et al., 2014 Powell 118 Kendall and Coplen, 2001; This work Poyang 35 Wenbin et al., 2007 Qarhan Salt 1 Yuan et al., 2011 Qinghai Hu 10 Henderson et al., 2010 Rukwa 4 Bergonzini et al., 2001 Sakakawea 15 Kendall and Coplen, 2001 Salton Sea 3 Mazzini et al., 2011 Sambhar Salt 8 Yadav et al., 1997 Shala 19 Craig et al., 1977 Superior 161 Karim et al., 2008; Jasechko et al., 2014 Tahoe 4 McKenna, 1992 Tana 52 Gonfiantini et al., 1973 Tanganyika 48 Craig et al., 1975 Taro Co 1 Yuan et al., 2011 Taupo 1 Stewart et al., 1981 Titicaca 12 Fontes et al., 1979 Tonlé Sap 14 Kabeya et al., 2008 Turkana 9 Cerling et al., 1988 Valencia 1 Friedman et al., 1964 Van 2 Kwiecien, 2011 Victoria 1 Beuing et al., 1997 Winnipeg 3 Buhay et al., 1998; This work Yamdruk-tso 1 Yuan et al., 2011 Yellowstone 1 This work Zhari Namco 1 Yuan et al., 2011

49

Figure 1-9. Compiled water use efficiency and vapor pressure deficit relationshiops (compiled data and original references presented in Jasechko et al., 2013). Black lines mark C4 pathways and grey lines mark C3 pathways. Thick black and red lines mark the regressions through the entire C4 (red) and C3 (black) datasets, respectively.

Figure 1-10. Estimated spatial distributions of water use efficiency, accounting for climate and photosynthesis types (C3 and C4) Figure based upon work by Jasechko et al. (2013).

50 Table 1-9. Plant water use efficiency (WUE) as a function of vapor pressure deficit (VPD) WUE Plant C3/C4 VPD range (kPa) (mmol CO2/mol H2O) Acer saccharum, Betula C3 2.8×(VPD)−0.37 <0.1 to 1.7 alleghaniesis, Tsuga canadensis

Arachis C3 1.4 to 2.3 2.2

Deciduous C3 5.9×(VPD)−0.98 0.3 to 0.9

Encilia fariosa C3 3.5×(VPD)−1.02 1.1 to 3.8

Evergreen C3 7.1×(VPD)−0.93 0.3 to 0.9

Hordeum vulgare C3 3.0×(VPD)−0.39 0.6 to 1.9

Ipomoea vagans C3 8.0×(VPD)−0.77 0.6 to 3.0

Larrea tridentata C3 3.6×(VPD)−0.38 0.5 to 2.0

Nicotiana glauca C3 5.0×(VPD)−0.80 0.5 to 2.0

Olea europaea L. C3 5.2×(VPD)−0.94 0.2 to 6.8

Oryza sativa C3 4.8×(VPD)−0.48 0.5 to 2.0

Oryza sativa L. C3 1.8×(VPD)−0.28 0.3 to 1.6

Phalaris aquatica C3 3.2×(VPD)−0.44 0.5 to 2.0

Phaseolus vulgaris C3 4.8×(VPD)−0.76 0.5 to 2.0

Pinus sylvestris, Picea abies C3 5.8×(VPD)−0.39 0.1 to 1.3

Populus tremuloides C3 5.5×(VPD)−0.82 0.5 to 4.5

Prosopis juliflora C3 10.6×(VPD)−1.38 1.0 to 9.6

Pseudotsuga C3 8.9×(VPD)−0.49 0.3 to 3.2

Quercus C3 6.6×(VPD)−0.44 0.2 to 1.8

Salix viminalis C3 7.8×(VPD)−0.58 0.2 to 2.1

Triticum C3 3.0×(VPD)−0.67 0.5 to 1.5

Triticum C3 5.9×(VPD)−0.50 <0.1 to 2.8

Dactyloctenium aegyptium C4 14.3×(VPD)−0.73 1.2 to 3.8

Eragrostis tremula C4 17.9×(VPD)−1.12 2.1 to 5.0 Miscanthus giganteus, Spartina C4 5.3×(VPD)−1.18 1.0 to 1.2 cynosuroides

Paspalum plicatulum C4 5.7×(VPD)−0.35 0.5 to 2.0

Pleuraphis rigida C4 7.0×(VPD)−0.78 1.3 to 3.8

Schoenefeldia gracilis C4 10.1×(VPD)−1.05 0.9 to 3.0

Zea mays C4 8.9×(VPD)−0.35 <0.1 to 2.9

Zea mays C4 7.7×(VPD)−0.47 0.5 to 2.0

References for each of the above studies compiled within Jasechko et al. (2013)

51 A global in scale calculation was developed using a new global compilation of river isotopic data (Table 1-10). A deuterium excess mass balance of the continents was used to estimate the global transpiration/evapotranspiration ratio on land surfaces:

Id I  d E   Qd Q  d E  xPd P  d E  T  Equation 1.9 d T  d E where d represents the deuterium excess of each flux, I represents the flux of precipitation entering the catchment, E represents physical evaporation losses from a catchment, T represents transpiration water losses from a catchment, Q represents liquid losses via runoff and groundwater discharge out of the basin, x represents the fraction of precipitation (P) that is intercepted by vegetation and returned to the atmosphere through evaporation. The global water use efficiency was estimated by spatially weighting our grid cell estimates of water use efficiency to mean annual mean normalized difference vegetation indices (values less than zero assigned a value of zero), and was found to be close to 3.2±0.9 mmol CO2 per mol H2O.

Figure 1-11. The deuterium excess of 31 major rivers (a) and associated annual streamflow (b). Discharge data from Dai and Trenberth (2002). 52 Table 1-10. Deuterium excess of major rivers ranked by discharge

River Q (km3/y)** Rank** Deuterium excess* (‰) Amazon 6642 1 9.8±1.8 Changjiang 944 4 8.8±2.5 Mississippi 610 6 8.5±1.8 Yenisey 599 7 8.2±2.6 Paraná 568 8 7.7±4.0 Lena 531 9 7.2±1.3 Mekong 525 10 5.9±4.5 Ob 412 13 5.8±2.1 St Lawrence 363 16 3.1±1.3 Amur 354 17 6.3±1.3 Mackenzie 290 19 -1.0±1.2 Columbia 252 21 6.6±1.6 Yukon 212 24 3.5±1.6 Danube 202 26 9.0±1.2 Fraser 144 30 2.9±3.8 Kolyma 118 35 5.7±1.9 Indus 104 38 13.5±5.9 Neva 79 45 6.4±1.0 Sacramento 69 50 8.6±0.9 Kuskokwim 57 54 5.7±1.2 Alabama 51 68 10.4±2.3 Stikine 51 69 8.3±1.1 Susquehanna 46 75 12.9±2.3 Susitna 45 78 4.7±1.1 Volta 37 86 1.3±6.9 Copper 34 96 6.5±1.4 Nushagak 31 109 5.7±1.3 Tombigbee 27 124 11.6±2.9 Colorado R. 12 165 -1.8±1.2 Brazos 7 180 2.9±5.4 Colorado (TX) 3 195 4.6±8.3 Rio Grande 2 196 -1.5±1.2 * Deuterium excess (d) defined as (Dansgaard, 1964): d = δ2H − 8∙δ18O (±1 s.d. shown), references to original data sources presented in Jasechko et al. (2013). ** Runoff data from Dai and Trenberth (2002)

53 1.4 Results

Figure 1-12 presents the range of δ18O and δ2H values observed in Earth’s large lakes. The lowest δ18O and δ2H values in large lakes are generally found at high altitudes and latitudes, whereas the highest δ18O and δ2H values are found in lakes that are located at low latitudes and altitudes. The entire dataset spans the range of −23‰ to +15‰ in δ18O and −180‰ to +80‰ in δ2H values. The majority of lakes are found to plot below a regression of meteoric waters (i.e., the “global meteoric water line;” Craig, 1961) because of kinetic isotope effect occurring during the process of evaporation (Craig, 1961).

Figure 1-12. The stable O and H isotopic composition of Earth’s large lakes and inland/semi- enclosed seas. The lowest δ18O and δ2H values are observed at high altitudes and latitudes (e.g., Kluane Lake) and the highest δ18O and δ2H values are observed at low latitudes and altitudes (e.g., Lake Turkana). Dots mark individual water samples, with shaded areas enclosing all points for a single lake (“convex hull” – from Jasechko et al., 2013). References presented in Table 1-8.

54 Some lakes have greater internal variability in δ18O and δ2H values than others. Generally, stratified and shallow lakes (e.g., Lake Chad, Great Salt Lake and the Aral Sea) have larger variations in δ18O and δ2H values than well mixed, deep lakes (e.g., the North American Laurentian Great

Lakes or Lake Baikal; Figure 1-13). Stratified lakes (e.g., Tanganyika) have different δ18O values at the lake surface compared to values at depth (Figure 1-14); whereas, well-mixed lakes such as Lake Baikal

(Figure 1-14) and each of the North American Great Lakes (Figure 1-15) have relatively homogenous

δ18O and δ2H values.

Figure 1-13. The internal variability in a selection of large lake δ18O and δ2H values. The upper pane highlights two well-mixed, deep (>400 m maximum depth) lakes that have homogenous isotopic compositions. The lower pane delineates heterogeneous lake δ18O and δ2H values found in shallow large lakes and inland seas.

55

Figure 1-14. Temperature (top row) and δ18O (bottom row) profiles for the two largest (volumetric) lakes on Earth: Tanganyika (left) and Baikal (right; data from Craig, 1975 and Seal and Shanks, 1998). Baikal has a heterogenous temperature profile, but a homogenous isotopic composition. Tanganyika has a heterogeneous isotopic profile and a heterogenous temperature profile.

Figure 1-15. The isotopic composition and temperature (April) of the North American Great Lakes (profile shown here spans Superior, Huron, Erie and Ontario, but skips Michigan, which has the highest δ18O and δ2H values of the Laurentian Great Lakes; from Jasechko et al., 2014).

56 The rate of transpiration spanning 10% of Earth’s ice free area based on stable isotopic data is presented in Figure 1-17, both as a percentage of total evapotranspiration (Figure 1-17 a) and as a transpiration rate (Figure 1-17 b). The rate of transpiration is greatest in the humid tropics where primary production is rarely limited by water and temperature (e.g., east African Great Lakes; Figure

1-16; Running et al., 2004). The rate of transpiration – in our dataset – is smallest in the boreal forest, where growth is limited to a short growing season due to the pronounced seasonality of the high latitudes. Transpiration rates are also discovered to be small in arid climates, where growth is expected to be limited by water supplies.

Global transpiration fluxes calculated by a deuterium excess mass balance of continental waters support conclusions reached using our global lake dataset: transpiration is the largest H2O flux from the continents. Our global analysis suggests that 80 to 90% of vapor flows from the continents are funneled through transpiration, with smaller percentages left to evaporation or sublimation.

Volumetrically, isotopic data support a transpiration flux of ~60,000 km3 of water per year, which has a corresponding latent energy requirement of 33 W/m2, suggesting that a substantial percentage of radiation absorbed by Earth’s surfaces is appropriated to the vaporization of water at plant leaf surfaces (“solar absorbed at the surface” reported as ranging between 147 W/m2 and 174 W/m2

(Trenberth et al., 2009).

57

Figure 1-16. The transpiration rate for seven ecozones (each bar is one lake catchment). The total evapotranspiration flux for each catchment is marked by a black line, and the median result of Monte Carlo calculation realizations is represented by a square. Bars extend to the 25th-75th percentiles of Monte Carlo realizations.

58

Figure 1-17. Transpiration rates for 10% of ice free land areas. (a) Transpiration is shown both as a proportion of total evapotranspiration (%; pane a) and as a rate (mm H2O year-1; b).

59

Figure 1-18. Catchment-by-catchment transpiration rates for 73 lakes. Colors mark ecoregions, with grey bars marking lake catchments that fall into more than one ecoregion. Total evapotranspiration rates are marked as a dash, whereas the median of calculation results is marked by a square.

Global primary production assimilates ~123±8 Gt of carbon each year (Beer et al., 2010).

The transpiration fluxes reported in this study can be used to calculate gross primary production by applying the water use efficiency data calculated within each catchment to transpiration fluxes (i.e., converting mm H2O transpired per year into g C assimilated per year). Gross primary production rate calculated using transpiration fluxes are presented in Figure 1-19.

60

Figure 1-19. Gross primary productivity (catchment averages) for 10% of Earth’s ice free land area calculated by coupling isotope-based transpiration fluxes to water use efficiency data.

1.5 Discussion

Transpiration is found to account for more than two-thirds of evapotranspiration in more than 80% of catchments studied. We find that even though potential evaporation rates likely exceed transpiration on land surfaces, the rate of evaporation is limited by the small area of open water on

“land” surfaces (~3%, globally; Downing et al., 2006). Therefore our results suggest that biological fluxes of water into the atmosphere is the greatest vapor flow from the continents, rather than evaporation. Plant roots tap into ground- and soil-water reservoirs and effectively move these underground water sources upward to Earth’s boundary layer for evaporation at leaf surfaces, whereas evaporation is limited in its water supply to water that is at or near the surface.

Our analysis neglects snow sublimation, which has been proposed as a non-fractionating process (although more recent work suggests that sublimation is indeed a fractionation labelled process; Koeniger et al., 2006), is thought to be very small at pan-continental scales. A compilation of 61 three global climate and land surface model estimates places sublimation at less than 2% of terrestrial evapotranspiration (not considered in one, ~1% in another, and ~2% in the third). Although locally sublimation may be an important H2O vaporization process, current land surface and general circulation models suggest that >98% of continental vapor flows are represented by transpiration, interception and evaporation (i.e., the vapor flows considered in this study).

Table 1-11. Published estimates of sublimation relative to terrestrial evapotranspiration Sublimation / total Study Model or methodology evapotranspiration Calculation: Sublimation of 0.35 mm/mo (i.e., 4.2 mm/year) is shown. Even including seasonally snow Dirmeyer et al., <1% covered regions and ice-caps that cover 46,000,000 2006 km2 (see *), this calculation yields a sublimation flux of ~200 km3/yr, or less than 0.5 % of terrestrial ET. Lawrence et al., Not considered Community Land Model Version 3 2007 Miralles et al., Global Land-surface Evaporation: the Amsterdam 2% 2011 Methodology * National Snow and Ice Data Center: Snow and Climate. nsidc.org/cryosphere/snow/climate.html

The connections between carbon and water fluxes made in this study highlight a novel approach for quantifying water and carbon fluxes on continents. Continental water and carbon cycles are connected by water use efficiency ratios – compiled in this study, for the first time – highlighting an opportunity for atmospheric models to take advantage of this natural H2O-CO2 accounting system on continents. Because of the higher proportion of transpiration/evapotranspiration discovered here, this analysis suggests that biological changes due to climate and land use modifications will exert a dominant impact upon fresh water fluxes, and associated transports of nutrients, contaminants, sediment and other solutes. The results of this study also highlight that changes to vegetation in the past, such as the evolution and spread of the C4 photosynthetic pathway or the emergence of vascular plants onto continents, are likely to have profoundly modified the global water and carbon cycles.

62 1.6 References

Alton, P., Fisher, R., Los, S., and Williams, M. (2009), Simulations of global evapotranspiration using semiempirical and mechanistic schemes of plant hydrology, Global

Biogeochemical Cycles, 23, GB4032.

Aly, A. I. M., Froehlich, K., Nada, A., Hamza, M. and Salem, W. M. (1993), Study of environmental isotope distribution in the Aswan High Dam Lake (Egypt) for estimating the evaporation of lake water and its recharge to adjacent groundwater. Environmental Geochemistry and

Health, 15, 37–49.

Bahati, G., Pang, Z., Armannsson, H., Isabirye, E.M. and Kato, V. (2005), Hydrology and reservoir characteristics of three geothermal systems in Western Uganda, Geothermics, 34, 568–591.

Barbour, M. M., Hunt, J. E., Walcroft, A. S., Rogers, G. N. D., McSeveny, T. M. and

Whitehead, D. (2005), Components of ecosystem evaporation in a temperate coniferous rainforest, with canopy transpiration scaled using sap–wood density, New Phytologist, 165, 549–558.

Bates, R. E. and Bilello, M. (1966), Defining the cold regions of the northern hemisphere,

U.S.A/ Cold Regions Research and Engineering Laboratory, Technical Report 178.

Becht, R., Mwango, F. and Muno, F. A. (2005), Groundwater links between Kenyan Rift

Valley lakes (eds. Odada et al.) 7 – 14 (Proceedings of 11th World Lakes Conference, Nairobi,

Kenya, 2005).

Beer, C. et al. (2010), Terrestrial gross carbon dioxide uptake: global distribution and covariation with climate, Science, 329, 834–838.

63 Bergonzini, L., Gibert, E., Winckel, A. and Merdaci, O. (2001), Bilans hydrologique et isotopique (18O et 2H) du Lac Massoko, Tanzanie. Quantification des échanges lac–eaux souterraines.

C. R. Acad. Sci. Paris, Sciences de la Terre et des Planètes 333, 617–623.

Beuning K. R. M., Kelts, K., Ito, E. and Johnson, T. C. (1997), Paleohydrology of Lake

Victoria, East Africa, inferred from 18O/16O ratios in sediment cellulose. Geology, 25, 1083–1086.

Beziat, P., Rivalland, V., Tallec, T., Jarosz, N., Boulet, G., Gentine, P. and Ceschia, E. (2013),

Evaluation of a simple approach for crop evapotranspiration partitioning and analysis of the water budget distribution for several crop species, Agricultural and Forest Meteorology, 177, 46–56.

Bosch, J. M., and J. D. Hewlett (1983), A review of catchment experiments to determine the effect of vegetation changes on water yield and evapotranspiration, Journal of Hydrology, 55, 3–23.

Bowen, G. J. (2010), Statistical and geostatistical mapping of precipitation water isotope ratios. In Isoscapes, Springer, Netherlands, pp. 139–160.

Bowen, G. J., and Revenaugh, J. (2003), Interpolating the isotopic composition of modern meteoric precipitation. Water Resources Research, 39, 1299.

Bowen, G. J., and Wilkinson, B. (2002), Spatial distribution of δ18O in meteoric precipitation,

Geology, 30, 315–318.

Brahney, J. Paleolimnology of Kluane Lake. (2007), 116 pp. (M.Sc. thesis, Simon Fraser

University, Vancouver, British Columbia, Canada)

Brock, B. E., Yi, Y., Clogg–Wright, K. P., Edwards, T. W. D. and Wolfe, B. B. (2009),

Multi–year landscape–scale assessment of lakewater balances in the Slave River Delta, NWT, using water isotope tracers, Journal of Hydrology, 379, 81–91.

64 Buhay, W. M. and Betcher, R. N. (1998), Paleohydrologic implications of 18O enriched Lake

Agassiz water. Journal of Paleolimnology, 19, 285–296.

Kendall, C. and Coplen, T. B. (2001), Distribution of oxygen–18 and deuterium in river waters across the United States. Hydrological Processes, 15, 1363–1393.

Calder, I. R., Wright, I. R. and Murdiyarso, D. (1986), A study of evaporation from tropical rainforest—West Java, Journal of Hydrology, 89, 13–31.

Cao, L., Bala, G., Caldeira, K., Nemani, R., and Ban–Weiss, G. (2010), Importance of carbon dioxide physiological forcing to future climate change, Proceedings of the National Academy of Sciences,

107, 9513–9518.

Cavanaugh, M. L., Kurc, S. A. and Scott, R. L. (2011), Evapotranspiration partitioning in semiarid shrubland ecosystems, a two–site evaluation of soil moisture content on transpiration,

Ecohydrology, 4, 671–681.

Cerling, T. E., Bowman, J. R. and O’Neil J. R. (1988), An isotopic study of a fluvial lacustrine sequence: the Plio–Pleistocene Koobi Fora sequence, East Africa, Palaeogeography,

Palaeoclimatology, Palaeoecology, 63, 335–356.

Choudhury, B. J. and DiGirolamo, N. E. (1998), A biophysical process–based estimate of global land surface evaporation using satellite and ancillary data. I. Model description and comparison with observations, Journal of Hydrology, 205, 164–185.

Choudhury, B. J., DiGirolamo, N. E., Susskind, J., Darnell, W. L., Gupta, S. K., and Asrar,

G. (1998), A biophysical process–based estimate of global land surface evaporation using satellite and ancillary data, II. Regional and global patterns of seasonal and annual variations, Journal of Hydrology,

205, 186–204.

65 Cienciala, E., Kucera, J., Lindroth, A., Cermak, J., Grelle, A. and Halldin, S. (1997), Canopy transpiration from a boreal forest in Sweden during a dry year, Agricultural and Forest Meteorology, 86,

157–167.

Cohen, A. S., Talbot, M. R., Awramik, S. M., Dettman, D. L. and Abell, P. (1997), Lake level and paleoenvironmental history of Lake Tanganyika, Africa, as inferred from late-Holocene and modern stromatolites, Geological Society of America Bulletin, 109, 444–460.

Craig, H. (1975), Lake Tanganyika geochemical and hydrographic study: 1973 expedition.

Scripps Institute of Oceanography, Reference 75–5, La Jolla, California, 83 pp.

Craig, H. (1975), Lake Tanganyika geochemical and hydrographic study: 1973 expedition.

Report 75–5, 83 pp. (Scripps Institute of Oceanography La Jolla, California, 1975).

Craig, H. (1996), Isotopic composition of the Red Sea and Salton Sea geothermal brines.

Science, 154, 1544–1548.

Craig, H. and Gordon, L. I. (1965), in Stable Isotopes in Oceanographic Studies and

Paleotemperatures (ed. Tongiorgi, E.) Lab. Geol. Nucl., pp. 9–130.

Craig, H., Lupton, J. E. and Horowiff, R. M. (1977), Isotope Geochemistry and Hydrology of geothermal waters in the Ethiopian rift valley, Report 77–14, 140 pp. (Scripps Institute of

Oceanography, La Jolla, California).

Dai, A., and Trenberth, K. E. (2002), Estimates of freshwater discharge from continents:

Latitudinal and seasonal variations, Journal of Hydrometeorology, 3, 660–687.

Dapena, C. and Panarello, H. O. (1997), Isotopic study of the “Laguna Mar Chiquita,”

Córdoba, Argentina and its hydrogeological and paleoclimatological implications. 7–15 (Proc. Int.

Symp. International Atomic Energy Agency, Vienna). 66 Darling, W. G., Berhanu, G. and Arusei, M. K. (1996), Lake–groundwater relationships and fluid–rock interaction in the East African Rift Valley: isotopic evidence. Journal of African Earth

Sciences, 22, 423–431.

Dinçer, T. (1968), The use of oxygen–18 and deuterium concentrations in the water balance of lakes, Water Resources Research, 4, 1289–1305.

Dirmeyer, P. A., Gao, X., Zhao, M., Guo, Z., Oki, T., and Hanasaki, N. (2006), GSWP–2: multimodel analysis and implications for our perception of the land surface, Bulletin of the American

Meteorological Society, 87, 1381–1397.

Dolman, A. J (1988), Transpiration of an oak forest as predicted from porometer and weather data, Journal of Hydrology, 97, 225–234.

Downing, J. A. et al. The global abundance and size distribution of lakes, ponds, and impoundments. Limnol. Oceanogr. 51, 2388–2397 (2006).

Evaristo, J., Jasechko, J., McDonnell, J. J. (in review), Global separation of plant transpiration from groundwater recharge and streamflow.

Ferretti, D.F., Pendall, E., Morgan, J. A., Nelson, J. A., LeCain, D. and Mosier, A. R. (2003),

Partitioning evapotranspiration fluxes from a Colorado grassland using stable isotopes: Seasonal variations and ecosystem implications of elevated atmospheric CO2, Plant and Soil, 254, 291–303.

Ffolliott, P.F., Gottfried, G. J., Cohen, Y. and Schiller, G. (2003), Transpiration by dryland oaks, studies in the south–western United States and northern Israel, Journal of Arid Environments, 55,

595–605.

Floret, C., Pontanier, R., and Rambal, S. (1982), Measurement and modeling of primary production and water use in a South Tunisian steppe, Journal of Arid Environments, 5, 77–90. 67 Fontes, J.–Ch. and Gonfiantini, R. (1970), Composition isotopique et origine de la vapeur d’eau atmosphérique dans la région du Lac Leman, Earth and Planetary Science Letters, 7, 325–329.

Fontes, J–Ch., Boulange, B., Carmouze, J.P. and Florkowski, T. Preliminary oxygen–18 and deuterium study of the dynamics of Lake Titicaca. 145–150 (Proc. Int. Symp. International Atomic

Energy Agency, Vienna.

Fontes, J–Ch., Gonfiantini, R. and Roche, M. A. (1970), Deutérium et oxygène–18 dans les eaux du lac Tchad. 387–404 (Proc. Int. Symp. International Atomic Energy Agency, Vienna).

Frangi, J. L. and Lugo, A. E. (1985), Ecosystem dynamics of a subtropical floodplain forest,

Ecological Monographs, 55, 351–369.

Friedman, I., Redfeld, A. C., Scohoen, B., Harris, J. (1964), The variation of the deuterium content of natural waters in the hydrologic cycle. Reviews of Geophyics, 2, 177–224.

Froehlich, K. (2000), Evaluating the water balance of inland seas using isotopic tracers: the

Caspian Sea experience, Hydrological Processes, 14, 1371–1383.

Galoux, A., Benecke, P., Gietl, G., Hager, H., Kayser, C., Kiese, O., Knoerr, K. R., Murphy,

C. E., Schnock, G. and Sinclair, T. R. (1981), Radiation, heat, water and carbon dioxide balances, In

D.E. Reichle (ed). Dynamic Properties of Forest Ecosystems, Cambridge University Press, pp. 87–

204.

Gash, J. H. C. and Stewart, J. B. (1997), The evaporation from Thetford Forest during 1975,

Journal of Hydrology, 35, 385–396.

Gat, J. R. (1984), The stable isotope composition of Dead Sea waters, Earth and Planetary

Science Letters, 71, 361–376.

68 Gebauer, T., Homa, V. and Leuschner, C. (2012), Canopy transpiration of pure and mixed forest stands with variable abundance of European beech, Journal of Hydrology, 442/443, 2–14.

Gerten, D., Hoff, H., Bondeau, A., Lucht, W., Smith, P., and Zaehle, S. (2005),

Contemporary “green” water flows: simulations with a dynamic global vegetation and water balance model, Physics and Chemistry of the Earth, 30, 334–338.

Gibson, J.J. and Edwards, T. W. D. (2002), Regional water balance trends and evaporation– transpiration partitioning from a stable isotope survey of lakes in southern Canada, Global

Biogeochemical Cycles 16.

Gonfiantini, R. (1986), Environmental isotopes in lake studies, In Handbook of

Environmental Isotope Geochemistry (Eds. P. Fritz, J.C. Fontes), Elsevier, pp. 113–168.

Gonfiantini, R., Borsi, S., Ferrara, G. and Panichi, C. (1973), Isotopic composition of waters from the Danakil Depression (), Earth and Planetary Science Letters, 18, 13–21.

Gonfiantini, R., Zuppi, G. M., Eccles, D. H. and Ferro, W. (1979), Isotope investigation of

Lake Malawi. (Proc. Int. Symp. International Atomic Energy Agency, Vienna).

Gordon, L. J., Steffen, W., Jönsson, B. F., Folke, C., Falkenmark, M. and Johannessen, Å.

(2005), Human modification of global water vapor flows from the land surface, Proceedings of the

National Academy of Sciences of the United States of America, 102, 7612–7617.

Granier, A., Biron, P. and Lemoine, D. (2000), Water balance, transpiration and canopy conductance in two beech stands. Agricultural and Forest Meteorology, 100, 291–308.

Grelle, A., Lundberg, A., Lindroth, A., Moren, A. S. and Cienciala, E. (1997), Evaporation components of a boreal forest, variations during the growing season, Journal of Hydrology, 197, 70–87.

69 Helliker, B. R., and Richter, S. L. (2008), Subtropical to boreal convergence of tree–leaf temperatures. Nature, 454, 511–514.

Henderson A. C. G., Holmes, J. A. and Leng, M. J. (2010), late-Holocene isotope hydrology of Lake Qinghai, NE Tibetan Plateau: effective moisture variability and atmospheric circulation changes. Quaternary Science Reviews, 29, 2215–2223.

Hijmans, R. J., Cameron, S. E., Parra, J. L., Jones, P. G., and Jarvis, A. (2005). Very high resolution interpolated climate surfaces for global land areas. International journal of climatology,

25(15), 1965–1978.

Hitchon, B. and Krouse, H. R. (1972), Hydrogeochemistry of surface waters of the

Mackenzie River drainage basin, Canada—III. Stable isotopes of oxygen, carbon and sulphur,

Geochimica et Cosmochimica Acta, 36, 1337–1357.

Hsieh, J. C. C., Chadwick, O. A., Kelly. E. F. and Savin, S. M. (1998), Oxygen isotopic composition of soil water, Quantifying evaporation and transpiration, Geoderma, 82, 269–293.

Hu, Z. et al. (2009), Partitioning of evapotranspiration and its controls in four grassland ecosystems, Application of a two–source model, Agricultural and Forest Meteorology, 149, 1410–1420.

Huang, X., Hao, Y., Wang, Y., Cui, X., Mo, X. and X. Zhou (2010), Partitioning of evapotranspiration and its relation to carbon dioxide fluxes in Inner Mongolia steppe, Journal of Arid

Environments, 74, 1616–1623.

Hudson, J. A (1988), The contribution of soil moisture storage to the water balances of upland forested and grassland catchment, Hydrologic Science Journal, 33, 289–308.

70 Isiorho, A. A., Matisoff and Wehn, G. K. S. (1996), Seepage relationships between Lake

Chad and the Chad aquifers, Ground Water, 34, 819–826. relationships between Lake Chad and the

Chad aquifers. Ground Water 34, 819–826 (1996).

Ito, A., and Inatomi, M. (2012). Water–use efficiency of the terrestrial biosphere: a model analysis focusing on interactions between the global carbon and water cycles. Journal of

Hydrometeorology, 13, 681–694.

Jasechko, S. (2011), Stable isotope mass balance of the Laurentian Great Lakes, (M.Sc. thesis,

University of Waterloo, Waterloo, Ontario, Canada).

Jasechko, S. (2014), Global plant breathing revealed by stable isotopes, American

Geophysical Union Hydrology Section July 2014 Newsletter.

Jasechko, S., Birks, S. J., Gleeson, T., Wada, Y., Fawcett, P. J., Sharp, Z. D., McDonnell, J. J. and Welker, J. M. (in review), The pronounced seasonality of global groundwater recharge.

Jasechko, S., Gibson, J. J. and Edwards, T. W. D. (2014), Stable isotope mass balance of the

Laurentian Great Lakes. Journal of Great Lakes Research, 40, 336–346.

Jasechko, S., Gibson, J. J., YI, Y., Birks, S. J., and Sharp, Z. D. (2011). Stable isotope composition of Earth's large lakes. In AGU Fall Meeting Abstracts (0358).

Jasechko, S., Sharp, Z. D., Gibson, J. J., Birks, S. J., Yi, Y., and Fawcett, P. J. (2013),

Terrestrial water fluxes dominated by transpiration, Nature, 496, 347–350.

Jung, M. et al. (2010), Recent decline in the global land evapotranspiration trend due to limited moisture supply, Nature, 467, 951–954.

71 Kabenge, I. and Irmak, S. (2012), Evaporative losses from a common reed–dominated peachleaf willow and cottonwood riparian community, Water Resources Research, 48, doi,10.1029/2012WR011902.

Karim, A., Veizer, J. and Barth, J. (2008), Net ecosystem production in the great lakes basin and its implications for the North American missing carbon sink: A hydrologic and stable isotope approach, Global Planetary Change, 61, 15–27.

Kebede, S., Travi, Y. and Rozanski, K. (2009), The δ18O and δ2H enrichment of Ethiopian lakes, Journal of Hydrology, 365, 173–182 (2009).

Keenan, T. F., Hollinger, D. Y., Bohrer, G., Dragoni, D., Munger, J. W., Schmid, H. P., and

Richardson, A. D. (2013), Increase in forest water–use efficiency as atmospheric carbon dioxide concentrations rise, Nature, 499, 324–327.

Koeniger, P., Hubbart, J. A., Link, T., Marshall, J. D. Isotopic variation of snow cover and streamflow in response to changes in canopy structure in a snow–dominated mountain catchment.

Hydrol. Processes 22, 557–566 (2006).

Kumagai, T., Takiko Tateishi, M., Miyazawa, Y., Kobayashi, M., Yoshifugi, N., Komatsu, H. and T. Shimizu (2014), Estimation of annual forest evapotranspiration from a coniferous plantation in Japan (1), Water use components in Japanese cedar stands, Journal of Hydrology, 508, 66–76.

Kwiecien, O. et al. (2011), Hydrological Conditions in Eastern Anatolia Over the Last CA.

500 KA – First Insights from Lake Van. Presentation at the American Geophysical Union Fall

Meeting 2011, December 5–9, San Francisco, U. S. A.

Labat, D., Goddéris, Y., Probst, J. L., and Guyot, J. L. (2004), Evidence for global runoff increase related to climate warming. Advances in Water Resources, 27, 631–642.

72 Lachniet, M. S. and Patterson, W. P. (2002), Stable isotope values of Costa Rican surface waters. Journal of Hydrology, 260, 135–150.

Ladekarl, U. L (1998), Estimation of the components of soil–water balance in a Danish oak stand from measurements of soil moisture using TDR, Forest Ecology and Management, 104, 227–238.

Lane, L.J., Romney, E. M. and Hakonson, T. E. (1984), Water–balance calculations and net production of perennial vegetation in the northern Mojave desert. Journal of Range Management, 37, 12–

18.

Lauenroth, W. K. and Bradford, J. B. (2006), Ecohydrology and the partitioning AET between transpiration and evaporation in a semiarid steppe, Ecosystems, 9, 756–767.

Lawrence, D. M., Thornton, P. E., Oleson, K. W., and Bonan, G. B. (2007), The partitioning of evapotranspiration into transpiration, soil evaporation, and canopy evaporation in a GCM: impacts on land–atmosphere interaction, Journal of Hydrometeorology, 8, 862–880.

Leopoldo, P. R., Franken, W. K. and Nova, N. A. V. (1995), Real evapotranspiration and transpiration through a tropical rain forest in central Amazonia as estimated by the water–balance method, Forest Ecology and Management, 73, 185–195.

Littmann, T. and M. Veste (2006), Determination of actual evapotranspiration and transpiration in desert sand dunes (Negev Desert) using different approaches, Forestry Studies in China,

8, 1–9.

Liu, B.L., Phillips, F., Hoines, S., Campbell, A. R. and Sharma, P. (1995), Water movement in desert soil traced by hydrogen and oxygen isotopes, chloride, and 36Cl, southern Arizona, Journal of

Hydrology, 168, 91–110.

73 Liu, R., Li, Y. and Wang, Q–X. (2012), Variations in water and CO2 fluxes over a saline desert in western China, Hydrological Processes, 26, 513–522.

Liu, X. and Chen, J. (2009), Studying of Model of Stable Isotope Fractionation in Lake –

Taking the Nam Co Lake as an example. Proceedings of the International Forum on Porous Flow and Applications, Wuhan City, China.

Longinelli, A. et al. (2008), A stable isotope study of the Garda lake, Northern Italy: its hydrological balance, Journal of Hydrology, 360, 103–116.

Luz B. and Barkan E. (2010), Variations of 17O/16O and 18O/16O in meteoric waters,

Geochimica et Cosmochimica Acta, 74, 6276–6286.

Massman, W. J (1992), A surface–energy balance method for partitioning evapotranspiration data into plant and soil components for a surface with partial canopy cover, Water Resources Research,

28, 1723–1732.

Mazzini, A., Svensen, H., Etiope, G., Onderdonk, N. and Banks, D. (2011), Fluid origin, gas fluxes and plumbing system in the sediment–hosted Salton Sea Geothermal System (California,

USA). Journal of Volcanology and Geothermal Research, 205, 67–83.

McClelland, J. W., S. J. Déry, B. J. Peterson, R. M. Holmes, and E. F. Wood (2006), A pan– arctic evaluation of changes in river discharge during the latter half of the 20th century, Geophys.

Res. Lett., 33, L06715.

McKenna, S. A., Ingraham, N., Jacobson, R. L. and Cochran, G. F. (1992), A stable isotopic study of Bank storage mechanisms in the Truckee River Basin, Journal of Hydrology, 134, 203–219.

McNulty, S. G., Vose, J. M. and Swank, W. T. (1996), Loblolly pine hydrology and productivity across the southern United States, Forest Ecology and Management, 86, 241–251. 74 Miralles, D. G. et al. (2014), El Niño–La Niña cycle and recent trends in continental evaporation, Nature Climate Change, 4, 122–126.

Miralles, D. G., De Jeu, R. A., Gash, J. H., Holmes, T. R., and Dolman, A. J. (2011),

Magnitude and variability of land evaporation and its components at the global scale. Hydrology and

Earth System Sciences, 15, 967–981.

Miralles, D. G., Gash, J. H., Holmes, T. R., de Jeu, R. A. and Dolman, A. J. (2010), Global canopy interception from satellite observations, Journal of Geophysical Research: Atmospheres, 115,

D16122.

Mitchell, P. J., Veneklaas, E., Lambers, H. and Burgess, S. S. O. (2009), Partitioning of evapotranspiration in a semi–arid eucalypt woodland in south–western Australia, Agricultural and

Forest Meteorology, 149, 25–37.

Moran, M. S., Scott, R. L., Keefer, T. O., Emmerich, W. E., Hernandez, M., Nearing, G. S.,

Paige, G. B., Cosh, M. H. and O’Neill, P.E. (2009), Partitioning evapotranspiration in semiarid grassland and shrubland ecosystems using time series of soil surface temperature. Agricultural and

Forest Meteorology, 149, 59–72.

Mu, Q., M. Zhao, S. W. Running (2011), Improvements to a MODIS Global Terrestrial

Evapotranspiration Algorithm, Remote Sensing of Environment, 115, 1781–1800

New, M., Lister, D., Hulme, M. and Makin, I. (2002), A high–resolution data set of surface climate over global land areas, Climate Research, 21, 1–25.

Nielson, K. E. and Bowen, G. J. (2010), Hydrogen and oxygen in brine shrimp chitin reflect environmental water and dietary isotopic composition, Geochimica et Cosmochimica Acta, 74, 1812–1822.

75 Nizinski, J. J., Galat, G. and Galat–Luong, A. (2011), Water balance and sustainability of

Eucalyptus plantations in the Kouilou Basin (Congo–Brazzaville), Russian Journal of Ecology, 42, 305–

314.

Oberhänsli, H., Weise, S. M. and Stanichny, S. (2009), Oxygen and hydrogen isotopic water characteristics of the Aral Sea, Central Asia, Journal of Marine Systems, 76, 310–321.

Ojiambo, S. B., Lyons, W. B., Welch, K. A., Poreda, R. J. and Johannesson, K. H. (2003),

Strontium isotopes and rare earth elements as tracers of groundwater–lake water interactions, Lake

Naivasha, Kenya, Applied Geochemistry, 18, 1789–1805.

Oki, T., and Kanae, S. (2006), Global hydrological cycles and world water resources, Science,

313, 1068–1072.

Paco, T.A., David, T. S., Henriques, M. O., Pereira, J. S., Valente, F., Banza, J., Pereira, F. L.,

Pinto, C. and David, J. S. (2009), Evapotranspiration from a Mediterranean evergreen oak savannah.

The role of trees and pasture, Journal of Hydrology, 369, 98–106.

Paruelo, J. M. and Sala, O. E. (1995), Water losses in the Patagonian steppe—a modeling approach, Ecology, 76, 510–520.

Peterson, B. J., Holmes, R. M., McClelland, J. W., Vörösmarty, C. J., Lammers, R. B.,

Shiklomanov, A. I., Shiklamanov, I. A., and Rahmstorf, S. (2002), Increasing river discharge to the

Arctic Ocean, Science, 298, 2171–2173.

Phillips, F. M., Mills, S., Hendrickx, J. M. H. and Hogan, J. (2003), Environmental tracers applied to quantifying causes of salinity in arid–region rivers; results from the Rio Grande Basin, southwestern USA, Developments in Water Science, 50, 327– 334.

76 Poole, D.K., Roberts, S. W. and Miller, P.C. (1981), Water utilization, in P.C. Miller (ed).

Resource Use by Chaparral and Matorral, Springer, pp. 123–149.

Prazak, J., Sir, M. and Tesar, M. (1994), Estimation of plant transpiration from meteorological data under conditions of sufficient soil moisture, Journal of Hydrology, 162, 409–427.

Raz–Yaseef, N., Yakir, D., Schiller, G. and Cohen, S. (2012), Dynamics of evapotranspiration partitioning in a semi–arid forest as affected by temporal rainfall patterns,

Agricultural and Forest Meteorology, 157, 77–85.

Ricketts, R. D. Johnson, T. C., Brown, E. T., Rasmussen, K. A. and Romanovsky, V. V.

(2001), Trace element and stable isotope study of the Holocene paleoclimate of Lake Issyk–Kul,

Palaeogeography, Palaeoclimatology, Palaeoecology, 176, 207–227.

Roderick, M. L., and Farquhar, G. D. (2002), The cause of decreased pan evaporation over the past 50 years, Science, 298, 1410–1411.

Roupsard, O. et al. (2006), Partitioning energy and evapotranspiration above and below a tropical palm canopy, Agricultural and Forest Meteorology, 139, 252–268.

Rousseaux, M. C., Figuerola,, P. I., Corfrea–Tedesco, G. and Searles, P. S. (2009), Seasonal variation in sap flow and soil evapotranspiration in an olive (Olea europaea L.) grove under two irrigation regimes in an arid region of Argentina, Agricultural Water Management, 96, 1037–1044.

Running, S. W., Nemani, R. R., Heinsch, F. A., Zhao, M., Reeves, M., and Hashimoto, H.

(2004), A continuous satellite–derived measure of global terrestrial primary production. Bioscience, 54,

547–560.

Russell, J. M. and Johnson, T. C. (2006), The water balance and stable isotope hydrology of

Lake Edward, Uganda–Congo, Journal of Great Lakes Research, 32, 77–90 (2006). 77 Salati, E. and Vose, P. B. (1984), Amazon basin—a system in equilibrium, Science, 225, 129–

138.

Schlesinger, W. H., and Bernhardt, E. S. (2013), Biogeochemistry: an analysis of global change, Academic press.

Schlesinger, W. H., and Jasechko, S. (2014), Transpiration in the global water cycle,

Agricultural and Forest Meteorology, 189, 115–117.

Schlesinger, W. H., Fonteyn, P. J. and Marion, G. M. (1987), Soil moisture content and plant transpiration in the Chihuahuan desert of New Mexico, Journal of Arid Environments, 12, 119–126.

Scott, R. L., Huxman, T. E., Travis, E. J., Cable, W. L. and Emmerich, W. E. (2006),

Partitioning of evapotranspiration and its relation to carbon dioxide exchange in a Chihuahuan

Desert grassland, Hydrological Processes, 20, 3227–3243.

Seal, R. R. and Shanks, W. C. (1998), Oxygen and hydrogen isotope systematics of Lake

Baikal, Siberia: implications for paleoclimate studies, Limnology and Oceanography, 43, 1251–1261.

Shuttleworth, W. J (1988), Evaporation from Amazonian rainforest, Proceedings of the Royal

Society of London, 233B, 321–346.

Smith, S. D., Herr, C. A., Leary, K. L. and Piorkowski, J. M. (1995), Soil–plant water relations in a Mojave desert mixed shrub community, A comparison of three geomorphic surfaces,

Journal of Arid Environments, 29, 339–351.

Stewart, M. K. and Taylor, C. B. Environmental isotopes in New Zealand hydrology: 1

Introduction: The role of oxygen–18, deuterium, and tritium in hydrology. New. Zeal. J. Sci. 24, 295–

311 (1981).

78 Still, C. J., Berry, J. A., Collatz, G. J., and DeFries, R. S. (2003), Global distribution of C3 and

C4 vegetation: carbon cycle implications. Global Biogeochemical Cycles, 17(1), 6–1.

Syed, T. H., Famiglietti, J. S., Chambers, D. P., Willis, J. K., and Hilburn, K. (2010), Satellite– based global–ocean mass balance estimates of interannual variability and emerging trends in continental freshwater discharge. Proceedings of the National Academy of Sciences, 107, 17916–17921.

Tajchman, S. L (1972), The radiative and energy balance of coniferous and deciduous forests, Journal of Applied Ecology, 9, 359–375.

Tang, J., Bolstad, P. V., Ewers, B. E., Desai, A. R., Davis, K. J. and Carey, E. V. (2006), Sap flux–upscaled canopy transpiration, stomatal conductance, and water use efficiency in an old growth forest in the Great Lakes region of the United States, Journal of Geophysical Research, 111, doi,

10.1029/2005JG000083

Taniguchi, M., Burnett, W. C., Cable, J. E., and Turner, J. V. (2002), Investigation of submarine groundwater discharge, Hydrological Processes, 16, 2115–2129.

Taniguchi, M., Nakayama, T., Tase, N. and Shimada, J. (2001), Stable isotope studies of precipitation and river water in the Lake Biwa basin, Japan, Hydrological Processes, 14, 539–556.

Telmer, K. and Veizer, J. (2000), Isotopic constraints on the transpiration, evaporation, energy, and gross primary production budgets of a large boreal watershed, Ottawa River basin,

Canada, Global Biogeochemical Cycles, 14, 149–165.

Tian, F., Qiu, G. Y., Yang, Y., Lu, Y. and Xiong, Y. (2013), Estimation of evapotranspiration and its partition based on an extended three–temperature model and MODIS products, Journal of

Hydrology, 498, 210–220.

79 Trenberth, K. E., Fasullo, J. T., and Kiehl, J. (2009), Earth's global energy budget, Bulletin of the American Meteorological Society, 90, 311–323.

Trlica, M. J. and Biondini, M. E. (1990), Soil–water dynamics, transpiration, and water losses in a crested wheatgrass and native shortgrass ecosystem, Plant and Soil, 126, 187–201.

Wada, Y., L. P. H. van Beek, and M. F. P. Bierkens (2012), Nonsustainable groundwater sustaining irrigation: A global assessment, Water Resources Research, 48, W00L06.

Wang, L. X., Niu, S. L., Good, S. P., Soderberg, K., McCabe, M. F., Sherry, R. A., Luo, Y.

Q., Zhou, X. H., Xia, J. Y. and Caylor, K. K. (2013), The effect of warming on grassland evapotranspiration partitioning using laser–based isotope monitoring techniques, Geochimica et

Cosmochimica Acta, 111, 28–38.

Waring, R. H., Rogers, J. J. and Swank, W.T. (1981), Water relations and hydrologic cycles,

In Dynamic Properties of Forest Ecosystems (ed. D.E. Reichle) Cambridge University Press, pp.

205–264.

Wassenaar, L. I., Athanasopoulos, P. and Hendry, M. J. (2011), Isotope hydrology of precipitation, surface and ground waters in the Okanagan Valley, British Columbia, Canada. Journal of

Hydrology, 411, 37–48.

Wenbin, Z., Wang, M., Chunhua, H. and Huayin, X. X. (2007), Preliminary isotope studies in

Poyang Lake, 427–435 (Proc. Int. Symp. International Atomic Energy Agency, Vienna).

Wilson, K. B., Hanson, P. J., Mulholland, P. J., Baldocchi, D. D. and Wullschleger, S. D.

(2001), A comparison of methods for determining forest evapotranspiration and its components, sap–flow, soil water budget, eddy covariance, and catchment water balance, Agricultural and Forest

Meteorology, 106, 153–168.

80 Wolfe, B.B. et al. (2007), Classification of hydrological regimes of northern floodplain basins

(Peace–Athabasca Delta, Canada) from analysis of stable isotopes (δ18O, δ2H) and water chemistry.

Hydrological Processes, 21, 151–168.

Yadav, D. N. (1997), Oxygen isotope study of evaporating brines in Sambhar Lake,

Rajasthan, India, Chemical Geology, 138, 109–118.

Yang, C., Telmer, K. and Veizer, J. (1996), Chemical dynamics of the St. Lawrence riverine system: δDH2O, δ13CDIC, δ34Ssulphate, and dissolved 87Sr/86Sr, Geochimica et Cosmochimica Acta, 60, 851–

866.

Yao, Z. J. et al. (2009), Characteristics of isotope in precipitation, river water and lake water in the Manasarovar Basin of Qinghai: Tibet Plateau, Environmental Geology, 57, 551–556.

Yoshimura, K., Kanamitsu, M., Noone, D. & Oki, T. Historical Isotope Simulation using

Reanalysis Atmospheric Data, J. Geophys. Res. 113, D19108 (2008).

Young, M. H., Caldwell, T. G., Meadows, D. G. and Fenstermaker, L.F (2009), Variability of soil physical and hydraulic properties at the Mojave Global Change Facility, Nevada, Implications for water budget and evapotranspiration, Journal of Arid Environments, 73, 733–744.

Yuan, F. et al. (2011), Evaporative enrichment of oxygen–18 and deuterium in lake waters on the Tibetan Plateau, Journal of Paleolimnology, 46, 291–307.

Yunusa, I. A. M., Walker, R. R. and Guy, J. R. (1997), Partitioning of seasonal evapotranspiration from a commercial furrow–irrigated Sultana vineyard, Irrigation Science, 18, 45–54.

Zhang, X., Zwiers, F. W., Hegerl, G. C., Lambert, F. H., Gillett, N. P., Solomon, S., Stott, P.

A., and Nozawa, T. (2007), Detection of human influence on twentieth–century precipitation trends,

Nature, 448, 461–465. 81 Zimmermann, U., Baumann, U., Imevbore, A., Henderson, F. and Adeniji, H. A. (1976),

Study of the mixing patterns of Lake Kainji using stable isotopes, Catena, 3, 63–76.

82 CHAPTER 2 — THE SEASONALITY OF GLOBAL GROUNDWATER RECHARGE

2.1 Abstract

Groundwater is recharged as rain and snowmelt infiltrate underground into aquifers.

Groundwater is a vital resource that sustains 40% of crop irrigation. Many studies report annual groundwater recharge rates, yet few studies report seasonal differences in groundwater recharge rates, particularly in light of differing precipitation fluxes between seasons. In this chapter I define the

“groundwater recharge ratio” as the proportion of rain and snow that recharges groundwater aquifers. On the basis of a newly compiled set of 54 paired precipitation-groundwater isotopic data, I show that groundwater recharge ratios are highest during the winter in most arid and temperate climates, and are at a maximum during the wet season in the tropics. The isotope-based seasonal assessment of groundwater recharge ratios are compared with the outputs of a global hydrological model (PCR-GLOBWB), and the model is found to compare closely with the isotope observations in most, but not all locations. The seasonal difference in the efficiency of groundwater recharge suggests that changes to winter (temperate and arid regions) and wet season (tropics) hydrological processes will be the most important to future changes in groundwater recharge fluxes.

2.2 Introduction Groundwater supplies one third of modern-day human water uses (Wada et al., 2014) and represents the lion’s share (~99%) of unfrozen terrestrial water (Aeschbach-Hertig and Gleeson,

2012). Groundwater is replenished by rain and snowmelt that infiltrates through the critical zone near to Earth’s surface and into aquifers. Groundwater is depleted by natural discharges of groundwater flow paths into the water at the surface – such as streams, lakes and seas – and also is depleted by human extractions via wells. Humans need groundwater to sustain modern livelihoods. Groundwater supplied drinking water for two billion people, and sustains about 40% of global cropland irrigation

83 (Siebert et al., 2010; Foley et al., 2011). Although groundwater is a pivotal component of modern human livelihoods, the extractions of groundwater by humans are unsustainable and are draining aquifers at the global- (Konikow and Kendy, 2005; Wada et al., 2010; Konikow, 2011; Gleeson et al.,

2012) and regional scales (Rodell et al., 2009; Famiglietti et al., 2011; Scanlon et al., 2012; Feng et al.,

2013; Steward et al., 2013; Voss et al., 2013; Joodaki et al., 2014). Unsustainable groundwater extractions have been spotlighted in multiple regional scale studies including the northern Gangetic

Plain (India, Rodell et al., 2009), the North China Plain (Feng et al., 2013), the Middle East (Voss et al., 2013; Joodaki et al., 2014), the High Plains of the central United States of America (Scanlon et al.,

2012; Steward et al., 2013) and the Californian Central Valley (western U.S.A., Famiglietti et al., 2011;

Scanlon et al., 2012) and the Colorado River basin (Castle et al., 2014). To reverse these examples of regional-scale, non-sustainable pumping, groundwater managers will need to set and achieve long term pumping rate goals that will realize sustainable withdrawals (Gleeson et al., 2012; Aeschbach-

Hertig and Gleeson, 2012). However, these pumping rate goals must be established in the face of a changing climate, and, therefore, a moving target. In order to predict future groundwater replenishment rates, it is important that the best information possible be made available regarding natural groundwater recharge fluxes and their controlling processes which include: the physical state, amount and intensity of precipitation; topography; water table characteristics; geology; soil type; vegetation characteristics; boundary layer climatology; irrigation return flows).

Some previous work has evaluated controls upon groundwater recharge fluxes. A compilation of chloride mass balance recharge estimates suggests that plant life form distributions are a leading determinant for groundwater recharge, falling second only to precipitation amounts in terms of importance (Kim and Jackson, 2012). Work presented in Chapter one has indeed shown that transpiration is a dominant process in the global hydrological cycle (Jasechko et al., 2013).

84 Knowledge of annual groundwater recharge fluxes are a common research target in regional and continental scale scientific investigations (e.g., Scanlon et al., 2006; Döll and Fielder, 2007; Wada et al., 2010). Fewer studies have explored seasonal differences in recharge fluxes as a proportion of precipitation. Understanding the seasonal distribution of groundwater recharge is important because climate change will impact the hydrology of each season in different ways.

Here we define the groundwater recharge ratio: “the groundwater recharge (R) flux as a proportion of precipitation (P): R/P.” Previous modelling studies have estimated that the annual groundwater recharge ratio is close to ~10% (Figure 2-1). The groundwater recharge is estimated to be lowest in arid climates (average of 4%) and higher in boreal, temperate, and moist tropical forests

(averages recharge ratios of ~14%, ~15%, and ~16%, respectively; recharge estimates from Döll and

Fielder, 2007 and precipitation data from the Global Precipitation Climatology Project, accessed at www.gewex.org). However, these estimates are highly uncertain as a result of land use and irrigation return flows not embedded within most hydrological models, in addition to the immense challenges associated with accurately representing complex interactions of plants, rocks, and climate at Earth’s critical zone (where these interactions are at a maximum).

85

Figure 2-1. The global groundwater recharge ratio (recharge data from Döll and Fielder, 2007; precipitation data from the Global Precipitation Climatology Project: www.gewex.org) in map form (a) and presented as ecozone statistics (b; colored bars mark 25th-75th percentiles, lines mark 10th-90th percentile distribution).

Previous fieldwork has revealed that winter groundwater recharge ratios are higher than summer groundwater recharge ratios (Heppner et al., 2007; Jukić and Denić-Jukić, 2009; Yeh and

Famiglietti, 2009; Dripps and Bradbury, 2010; Dripps, 2012; Leterme et al., 2012). Seasonality of groundwater recharge ratios have been assessed in Belgium (Leterme et al., 2012), Greenland

(Leterme et al., 2012), the northeastern U.S.A. (Heppner et al., 2007; Yeh and Famiglietti, 2009;

Dripps and Bradbury, 2010; Dripps, 2012) and Croatia (Jukić and Denić-Jukić, 2009). In some cases, summer groundwater recharge has been shown to be restricted solely to high intensity thundershowers (Wisconsin, U.S.A.; Dripps, 2012). Field based monitoring of groundwater recharge in Tanzania has shown that groundwater recharge ratios are at their highest when rainfall is most intense (Taylor et al., 2013), suggesting that an intensifying hydrosphere (Durack et al., 2012) could, in fact, be beneficial from the sole standpoint of groundwater recharge fluxes. Temperate climate groundwater recharge has been found to be extremely and rapid process during snowmelt (Gleeson 86 et al., 2009), with the ice content of the shallow subsurface being a controlling factor upon on the proportion of snowmelt that recharges the subsurface aquifers (Granger et al., 1984). Groundwater recharge investigations in the mid-western United States of America has found that snowmelt can comprise two thirds of annual groundwater recharge (Delin et al., 2007; Dripps, 2012). Yet, in spite of these examples of seasonal biases in the efficiency of groundwater recharge, different recharge ratios between different seasons have not been observed in in all cases (e.g., Spain, Leterme et al.,

2012), opening an opportunity to calculate and assess the potential for seasonality in groundwater recharge ratios across different biomes with different lithologies, plant life forms and hydroclimates.

In this chapter, I hypothesize that by coupling groundwater and precipitation isotopic data, one may calculate seasonal differences in the groundwater recharge ratio. Several studies have compared precipitation and groundwater isotopic compositions. These studies have found differences in some cases, and no differences in other cases, between precipitation and groundwater

δ18O and δ2H values.

Studies finding similarities in the isotopic compositions of flux-weighted annual precipitation and modern groundwater include locations such as China (Li et al., 2000), Finland (Kortelainen,

2004), France (Genty et al., 2014), Israel (Even et al., 1986), Italy (Madonia et al., 2013), Korea (Lee et al., 1999; Lee and Kim, 2007), New Zealand (Williams and Fowler, 2002), Tasmania (Goede et al.,

1982), the United Kingdom (Darling and Bath, 1988; Darling et al., 2003) and the United States of

America (Yonge et al., 1985; van Beynen and Febbroriello, 2006). These finding suggest at first glance – without statistical analysis, per se – that groundwater recharge ratios in these sites are similar year round.

Studies finding differences in the isotopic compositions of flux-weighted annual precipitation and modern groundwater include field sites in South Africa (Vogel et al., 1963), the south-western United States (Arizona, Simpson et al., 1972, Kalin, 1994; Nevada, Winograd et al., 87 1998), the north-eastern United States (Pennsylvania, O’driscoll et al., 2005; Vermont: Abbott et al.,

2000), central Canada (Alberta, Maulé et al., 1994; Grasby et al., 2010), southern Canada (Ontario;

Huddart et al., 1999), French Guyana (Negrel et al., 2010), St. Croix (Gill, 1994), Spain (Julian et al.,

1992), Barbados, Puerto Rico and Guam (Jones et al., 2000; 2003). These differences have been interpreted as a reflection of higher groundwater recharge ratios during winter (Vogel et al., 1963;

Simpson et al., 1972; Maulé et al., 1994; Kalin, 1994; Winograd et al., 1998; Abbott et al., 2000;

O’driscoll et al., 2005) and wet seasons (Jones et al., 2000; 2003; Negrel; et al., 2010). These isotope- based results have never been synthesized at a global scale, nor have all studies quantitatively assessed the seasonal difference in groundwater recharge ratios, expressing these observations qualitatively instead.

The objective of this chapter of my dissertation is to test for, and quantify, seasonal differences in groundwater recharge ratios across a variety of field sites by analyzing a newly compiled global dataset of precipitation and groundwater isotopic data.

2.3 Dataset and methods

Here I calculate the seasonality of groundwater recharge ratios (two seasons) for 54 globally- distributed locations (Figure 2-2). I analyze global isotopic data for precipitation from regional and global monitoring networks (Araguás-Araguás et al., 2000; Welker, 2000; Birks and Edwards, 2009;

Welker, 2012) and compare precipitation isotopic data to nearby groundwater isotopic data that have been compiled from previous field reports. Precipitation data are available through the International

Atomic Energy Agency (e.g., Araguás-Araguás et al., 2000), the United States Network for Isotopes in Precipitation (Welker, 2000; Welker, 2012) and the Canadian Network for Isotopes in

Precipitation (Birks and Edwards, 2009). Groundwater isotopic data were compiled from >40 published datasets within the primary literature. Original field studies have been properly credited and are referenced within Table 2-1.

88 Table 2-1. Locations of paired precipitation and groundwater isotopic data

Station Data Lon. Lat. Reference Negrel and Petelet-Giraud, Cayenne IAEA -52.4 4.8 Guyana Shield 2010 Taguac IAEA 144.8 13.6 Guam caves Jones and Banner, 2003 Seawell IAEA -59.5 13.1 Barbados aqfr. Jones et al., 2000 Jakarta IAEA 106.8 -6.2 Jakarta aqfr. Kagabu et al., 2011 Das et al., 1988; Lorenzen et New Dehli IAEA 77.2 28.6 Gangetic Plain al., 2012 Dar es Salaam IAEA 39.2 -6.9 Coastal aqfr. Bakari et al., 2012 Demlie et al., 2007; Kebede Addis Ababa IAEA 38.7 9.0 Akaki Volcanics et al., 2007; Rango et al., 2010; Bretzler et al., 2011 Santa Maria IAEA -120.5 34.9 CA Coast www.waterqualitydata.us Beit Dagan IAEA 34.8 32.0 Israel coast aqfr. Yechieli et al., 2008 Pisa IAEA 10.4 43.7 Pisa Plain Grassi and Cortecci, 2005 Trout Lake USNIP -89.7 46.1 Surficial aqfr. www.waterqualitydata.us Yellowstone USNIP -110.4 44.9 Alluvial aqfr. www.waterqualitydata.us Smith's Ferry USNIP -116.1 44.3 Idaho Batholith Schlegel et al., 2009 Lake Geneva USNIP -88.5 42.6 Surficial aqfr. www.waterqualitydata.us East MA USNIP -71.2 42.4 Surficial aqfr. www.waterqualitydata.us Niwot Saddle USNIP -105.6 40.1 Surficial aqfr. www.waterqualitydata.us Wye USNIP -76.2 38.9 Aquia aqfr. Aeschbach-Hertig et al. 2002 Purdue Agr. USNIP -87.5 38.7 Surficial aqfr. www.waterqualitydata.us Clinton Stn. USNIP -78.3 35.0 Atlantic Plain www.waterqualitydata.us Caddo Valley USNIP -93.1 34.2 MI River Valley www.waterqualitydata.us Coffeeville USNIP -89.8 34.0 MI Embayment www.waterqualitydata.us Saturna CNIP -123.2 48.8 Surficial aqfr. Allan, 2003 Ottawa CNIP -75.7 45.3 Surficial aqfr. Praamsma et al., 2009 Elliot et al., 1999; Darling et Wallingford IAEA -1.1 51.6 London Chalk al., 1997 P. Douradas IAEA -7.6 40.4 Serra da Estrela Carreira et al., 2011 Malm Zuber et al. 2004; Krakow IAEA 19.9 50.1 Limestones Samborska et al., 2012 Cuxhave IAEA 8.7 53.9 N. German Bsn. Kloppman et al., 1998 Orleans IAEA 1.9 47.9 Paris Bsn. Kloppman et al., 1998 Melbourne IAEA 145.0 -37.8 Yarra Bsn. Tweed et al., 2004 Newcastle IAEA -104.2 43.9 Surficial aqfr. www.waterqualitydata.us Little Bighorn USNIP -107.4 45.6 Surficial aqfr. www.waterqualitydata.us Berg and Person, 2012; Lamberton USNIP -95.3 44.2 Mt. Simon aqfr. www.waterqualitydata.us N. Platte Agr. USNIP -100.8 41.1 N. High Plains McMahon et al., 2006 Mon Mouth USNIP -90.7 40.9 Surficial aqfr. www.waterqualitydata.us Great Plains USNIP -97.5 35.0 Arbuckle aqfr. www.waterqualitydata.us Edmonton CNIP -113.5 53.6 Surficial aqfr. Maule et al., 1994 Saskatoon CNIP -106.6 52.1 Dalmeny aqfrs. Fortin et al., 1991 Wynyard CNIP -104.2 51.8 Surficial aqfr. unpublished data 89 Station Data Lon. Lat. Aquifer Reference Esther CNIP -110.2 51.7 Surficial aqfr. Wallick, 1981 Lanza, 2009; Cheung and Calgary CNIP -114.0 51.0 Surficial aqfr. Mayer, 2009; Rock and Mayer, 2009 Icelandic Park USNIP -97.8 48.8 Winnipeg fm. Ferguson et al., 2007 / Gimli / CNIP Craters of the USNIP -113.6 43.5 Surficial aqfr. www.waterqualitydata.us Moon Pinedale USNIP -109.8 42.9 Colorado Plat. www.waterqualitydata.us Sand Spring USNIP -107.7 40.5 Surficial aqfr. www.waterqualitydata.us Smith Valley USNIP -119.3 38.8 Basin & Range www.waterqualitydata.us Tuscon ** -110.8 32.2 Tucson Basin Cunningham et al., 1998 Chihuahua IAEA -106.1 28.6 Chihuahua Plain Wassenaar et al., 2009 Alice Springs IAEA 133.9 -23.8 Amadeus Bsn. Wischusen et al., 2000 Zhangye IAEA 100.4 38.9 Hexi Corridor Qin et al., 2011 Yinchuan IAEA 106.2 38.5 Yinchuan Plain Wang, L. et al., 2012 Yellowknife CNIP -114.3 62.3 Con Mine Douglas et al., 2000 Whitehorse CNIP -135.1 60.7 Surficial aqfr. Carey and Quinton, 2005 Chapais CNIP -75.0 49.8 Surficial aqfr. Boutin, 2009

90

Figure 2-2. Locations where precipitation and groundwater isotopic data are available. Circles mark

54 study sites where sufficient groundwater and precipitation were available to assess groundwater recharge ratio seasonality. Diamonds mark locations where only a comparison of groundwater and precipitation isotopic data could be made (i.e., no seasonal recharge ratio

Most studies have only grab samples of groundwater, and do not report long term monitoring isotopic data. However, the multi-year isotopic monitoring studies of groundwater δ18O and δ2H values show little seasonal variability, suggesting that grab samples are suitable archives of multi-year recharge fluxes. Examples of long term groundwater monitoring for isotopic data include records analyzed in Finland (Kortelainen et al., 2004), Italy (Iacumin et al., 2009), the United

Kingdom (Darling et al., 2003), New Zealand (Williams and Fowler, 2002), eastern Canada (Savard et al., 2007) and France (Genty et al., 2014. The temporal homogeneity of groundwater δ18O and δ2H values is interpreted to be the result of hydrodynamic dispersion and multi-year groundwater residence times.

91 Precipitation isotopic data has been collected for more than 50 years by the International Atomic

Energy Agency (Araguás-Araguás et al., 2000) and other country-wide precipitation networks

(Welker, 2000; Kurita et al., 2004; Birks and Edwards, 2009; Welker 2012, Liu et al., 2013). The analysis of precipitation data for this study involved to steps: (i) calculation of amount-weighted annual precipitation isotopic compositions, and (ii) calculation of amount-weighted precipitation isotopic compositions for two six-month seasons for each meteorology station. Seasons are defined as winter and summer in the extra-tropics, and the wettest and driest consecutive six-month interval in the tropics.

First, the amount-weighted isotopic composition of precipitation (δP(annual)) was calculated following

(Equation 2.1):

12 ∑i=1 δP(i)Pi δP(annual) = 12 Equation 2.1 ∑i=1 Pi

where δP(i) is the monthly isotopic composition of precipitation during month i, and Pi is the amount of precipitation (i.e., the rate) during month i.

Second, the amount-weighted isotopic composition of season 1 (δP(season 1); defined as October-March in the northern hemisphere extra-tropics, and the wettest consecutive six month interval in the tropics) and season 2 (δP(season 2); defined as April-September in the extra-tropics, and the driest consecutive six month interval in the tropics) precipitation was calculated following (Equations 2.2 and 2.3):

δP(10)P10+δP(11)P11+δP(12)P12+δP(1)P1+δP(2)P2+δP(3)P3 δP(season 1) = Equation 2.2 P10+P11+P12+P1+P2+P3

δP(4)P4+δP(5)P5+δP(6)P6+δP(7)P7+δP(8)P8+δP(9)P9 δP(season 2) = Equation 2.3 P4+P5+P6+P7+P8+P9

92 Following the same symbology as outlined for Equation 2.1. Southern hemisphere (e.g., Melbourne,

Australia) sites had winter and summer months inverted.

Before completing an analysis of groundwater and precipitation data at similar locations, paleo-groundwater were required to be delineated and removed from the analysis because these groundwaters are not reflective of the modern climate where precipitation measurements have been made. Indeed fossil groundwaters that recharged during the Pleistocene (i.e., fossil groundwaters) have been shown to be different than modern groundwaters by −8 to +2 ‰ in δ18O due to different

Pleistocene hydroclimatology (e.g., Plummer, 1993; Edmunds, 2009) or due to subglacial recharge of groundwaters beneath the Laurentide and Fennoscandanavian ice sheets that resided over the northern portions of Eurasia and North America 20,000 years ago (e.g., Estonia, Karro et al., 2004; central Canada, Grasby and Chen, 2005; reviews by Jiráková et al., 2011 and McIntosh et al., 2012).

Groundwater δ18O and δ2H values, well depths, and 3H and 14C radioactivity levels were compiled from earlier works (Table 2-1). Considerations of (i) possible effects of evaporation during groundwater recharge, and (ii) possible shifts in δ18O and δ2H values related to paleoclimates recorded in fossil groundwaters were made before comparing precipitation and groundwater stable isotopic data.

First, partially evaporated groundwater samples were removed from our analysis using the deuterium excess parameter (Dansgaard, 1964). Partial evaporation leads to changes in isotopic compositions along δ2H/δ18O slopes of less than eight because of differences in the vapor pressures of the 1H1H16O, 1H1H18O and 2H1H16O isotopologues. The deuterium excess parameter (d = δ2H –

8×δ18O; Dansgaard, 1964) integrates information within both δ18O and δ2H values and is used here to test for modifications to the isotopic composition of groundwaters due to partial evaporation.

Samples bearing an evaporative signature will have a lower deuterium excess than that of meteoric waters, which have a global mean deuterium excess of close to +10 ‰. All groundwater samples with

93 a deuterium excess value of less than zero were removed from this analysis to ensure that the calculation of seasonal groundwater recharge ratios was not biased to evaporative influences.

Second, groundwater ages in excess of ~10,000 years were removed from this analysis on the basis of 3H, 14C and well depths because fossil groundwaters have different δ18O and δ2H values from modern groundwaters (Plummer, 1993; Edmunds and Milne, 2001; Grasby and Chen, 2005;

Karro et al., 2004; Edmunds, 2009; Jiráková et al., 2011; McIntosh et al., 2012). Samples with a 14C concentration of greater than 60 p.m.C. were included in this study, as the maximum groundwater age of samples with 60 p.m.C. can be no more than 5,000 years. Paleo-groundwater shifts in δ18O and δ2H values do not become apparent until ~12 ka (Edmunds and Milne, 2001; Edmunds, 2009;

Darling, 2011), such that groundwaters with 14C activities of exceeding 60 p.m.C. should be suitable for comparison with modern precipitation.

A similar delineation of “modern” groundwater can be derived from tritium groundwater data, a commonly applied age tracer in groundwater investigations. Here I a mixing model that accounts for groundwater residence time and mixing of different groundwaters with different ages

(i.e., time elapsed since recharge). I apply a mixing model that defines modern and old groundwater using the year 1950 as a threshold, where “post-1950 groundwater” is defined as groundwater having recharged after the year 1950 and “pre-1950 groundwater” is defined as groundwater that recharged prior to 1950.

To use the compiled 3H groundwater data to quantify the mixing components of post-1950 and pre-1950 groundwater an estimate of the activity of 3H in meteoric waters was required. I downloaded and analyzed a global dataset of 3H in precipitation measurements made since ~1960 at various locations by the International Atomic Energy Agency. Records of pre-1950 3H activities in meteoric waters are available from wine and ice cores compiled by Kotzer et al. (2000).

94 The 3H mixing model used to calculate pre- and post-1950 age components of each groundwater sample is presented next. The mass (m) fraction of groundwater having recharged after the year 1950 for a water sample (mpost-1950/msample) is calculated as:

3 3 mpost−1950 Hsample− Hpre−1950 = 3 3 Equation 2.4. msample Hpost−1950− Hpre−1950

where 3Hsample represents the 3H concentration of a given sample, 3Hpre-1950 represents the range of possible 3H values for groundwater that recharged prior to 1950, and or 3Hpost-1950 represents the range of possible 3H values for groundwater that recharged after 1950.

The calculation includes spatio-temporal variations in meteoric tritium activities in addition to the radioactive decay of 3H. Possible 3Hpre-1950 and 3Hpost-1950 concentrations were determined using linear regressions of latitude against the annual average 3H concentration in precipitation at various locations (Figure 2-3).

95

Figure 2-3. The variations of tritium in meteoric water over time. The left pane shows linear regressions through International Atomic Energy Agency precipitation stations (mean annual 3H values computed for every site, for every year). The color of each line marks the corresponding year that the regression was completed. The right pane shows an example of changes to 3H in precipitation over time for 45°N calculated using the regressions presented in the left pane. The squares and diamonds mark wine and ice core data (Kotzer et al., 2000). The darker lines and points show the 2009-equivilent 3H activity after considering radioactive decay, whereas the lighter (grey) points and line mark the uncorrected (i.e., “real time”) 3H activity of precipitation.

Regressions of 3H and latitude were developed using all International Atomic Energy

Stations with data for any given year (the number of stations available ranged from 12 to 89 sites, annually). Regressions for the southern and northern hemisphere were completed separately because of known inter-hemispheric differences in the 3H activity of precipitation, imparted because the atmosphere is not completely mixed (Rozanski et al., 1991). The latitude and sample date for every groundwater well location was entered into the latitude-time regressions of 3H in precipitation to develop a range of possible 3Hpost-1950 and 3Hpre-1950 activities, with considerations for radioactive decay made by calculating an equivalent 3H concentration for the date that each groundwater sample was collected (i.e., meteoric tritium decay corrected up to the date that the compiled groundwater sample was collected; 3H half-life of 12.3 years).

96 3Hsample values (i.e., measured groundwater tritium activity) and corresponding ranges for

3Hpost-1950 and 3Hpre-1950 were input into equation 2.4 to quantify the mixing proportion of “modern”

(i.e., post-1950) groundwater within each groundwater sample. All samples being comprised of

>80% “post-1950 groundwater” (median value from calculation used) were included in this study, as these were presumed to have recharged during the contemporary climate, where precipitation data is also available. All samples that did not meet this threshold were removed from our calculation, as mixing with paleo-waters could not be precluded. This dataset reduction step is likely to be conservative as many “3H-dead” (i.e., below detection tritium activities) groundwater samples may have recharged more recently than the mid-Holocene, and, therefore, could have been compared to modern precipitation in principle.

Finally, 90 percent of samples obtained from depths shallower than 40 meters underground were found to meet the aforementioned “modern groundwater” criteria set for 14C and 3H data.

Therefore, a depth threshold of 40 meters below ground level was set as a threshold for modern groundwater for compiled datasets that present groundwater δ18O or δ2H measurements but do not present 14C and 3H data. Additional care was taken on an aquifer-by-aquifer basis where paleo-waters are known to occur (e.g., Ferguson et al., 2007) in order to ensure that paleo-water isotopic data did not propagate into the calculation of groundwater recharge

Now that modern groundwaters have been delineated using the above 3H, 14C and well depth based methods, a stable isotope based calculation of groundwater recharge ratio seasonality can proceed. To calculate the seasonal difference in groundwater recharge ratios, we compare modern groundwaters (delineated sing 14C, 3H and well depths as per the preceding paragraphs) with modern precipitation data by combining a water budget (equation 2.5) and an isotopic (equation 2.6) mass balance:

Pannual = Pseason 1 + Pseason 2 Equation 2.5

97 PannualδP(annual) = Pseason 1δP(season 1) + Pseason 2δP(season 2) Equation 2.6

where Pannual, Pseason 1 and Pseason 2 are the precipitation rates for the year (i.e., annual), for season 1 (i.e., winter in the extra-tropics, and the wet season in the tropics), and for season 2 (summer in the extra- tropics, and the dry season in the tropics). Similarly, δP(annual), δP(season 1) and δP(season 2) are the amount- weighted isotopic compositions for annual, season 1 or season 2 time intervals. Combining equations

2.5 and 2.6 yields an isotope-based solution for the contribution of season 2 (i.e., summer or dry season) rainfall to total annual precipitation:

P δ −δ season 2 = P(annual) P(season 1) Equation 2.7 Pannual δP(season 2)−δP(season 1)

A similar set of equations can be derived for groundwater recharge rates (R) rather than precipitation rates.

Rannual = Rseason 1 + Rseason 2 Equation 2.8

Rannualδgroundwater = Rseason 1δP(season 1) + Rseason 1δP(season 2) Equation 2.9

where Rannual, Rseason 1 and Rseason 2 are annual, season 1 and season 2 recharge rates, and δgroundwater is the isotopic composition of recently recharged groundwater. Combining equations 2.8 and 2.9 yields the an equation representing theproportion of season 2 recharge as a ratio of annual recharge f

(equation 2.10).

R δ −δ season 2 = groundwater P(season 1) Equation 2.10 Rannual δP(season 2)−δP(season 1)

Combining equations 2.7 and 2.10 yields the isotope-based equation for the recharge ratio during the summertime (extra-tropics) or during the dry season (tropics; Rseason 2/Pseason 2; equation 2.11).

98 R δ −δ R season 2 = groundwater P(season 1) ( annual) Equation 2.11 Pseason 2 δP(annual)−δP(season 1) Pannual

A similar derivation (i.e., equations 4 – 10) can be made to calculate the recharge ratio during season

1 (Rseason 1/Pseason 1; equation 2.12):

R δ −δ R season 1 = groundwater P(season 2) ( annual) Equation 2.12 Pseason 1 δP(annual)−δP(season 2) Pannual

Finally, the isotope-based equation representing the seasonal difference in the groundwater recharge ratio (R/P) between season 1 and season 2 can be made – without knowledge of annual precipitation and recharge fluxes – by combining equations 2.11 and 2.12, yielding (Equation 2.13):

(R/P) δ −δ δ −δ season 1 = ( groundwater P(season 2)) / ( groundwater P(season 1)) Equation 2.13 (R/P)season 2 δP(annual)−δP(season 2) δP(annual)−δP(season 1)

This isotopic derivation of seasonal differences in groundwater recharge ratios is presented schematically in Figure 2-4 (lower axis).

99

Figure 2-4. A schematic representation of the isotope-based approach to estimating seasonal differences in the groundwater recharge ratio, defined as the proportion of rain and snow that infiltrates into groundwater aquifers. The four isotopic data shown are: the flux weighted isotopic composition of season one precipitation (δP(season 1)), the flux weighted isotopic composition of season two precipitation (δP(season 2)), the flux weighted isotopic composition of annual precipitation (δP(annual)), and the isotopic composition of groundwater (δgroundwater).

Uncertainties were estimated by completing the calculation using every combination of input data and subsequently computing percentile ranges from the various calculation results on a site-by- site basis. The calculation of seasonality in groundwater recharge ratios was only made for locations that had at least three groundwater δ18O or δ2H values and three annual amount-weighted δ18O and

δ2H values for precipitation. 16 stations were excluded in the analysis because no precipitation data were available for the summer or the winter season (e.g., Damascus, Syria) or because the δ18O and

δ2H values of winter and summer precipitation were not consistently higher or lower than the opposing season (e.g., Quincy and Kennedy Space Center in Florida, U.S.A.). Locations that not

100 included in this study of seasonal differences in the groundwater recharge ratio are marked as diamonds in Figure 2-2.

2.4 Results

Paired measurements of the isotopic composition of precipitation and groundwater at 54 globally-distributed locations are shown in Figure 2-5. Results show that, for the majority of samples, precipitation and groundwater isotopic compositions are similar, or that groundwater δ18O or δ2H values are lower than annual precipitation δ18O or δ2H values.

Figure 2-5. Comparison of groundwater and mean annual precipitation isotopic data at 54 globally- distributed locations. Vertical error bars mark one standard deviation of inter-annual variability in amount-weighted isotopic compositions of precipitation. Horizontal error bars bracket one standard deviation of groundwater isotopic data.

101

Isotope-based calculation results of seasonal differences of groundwater recharge ratios (i.e.,

(R/P)winter/(R/P)summer) are presented in Figure 2-6. Similarly, Table 2-2 presents 25th-75th percentile ranges of our isotope-based calculations of (i) the ratio of the summer groundwater recharge fluxes relative to winter groundwater recharge fluxes (Rsummer/Rwinter), (ii) summer recharge efficiencies

(Rsummer/Psummer), and (iii) winter recharge efficiencies (Rwinter/Pwinter) for each study site.

Winter groundwater recharge ratios are higher than summer groundwater recharge ratios for

93% of desert (7 of a total of 9), temperate grassland (11 of a total of 13) or temperate forest locations (16 of a total of 18; median of δ18O-based results of Monte-Carlo realizations). Winter recharge is at least twice as effective (i.e., higher recharge/precipitation ratio) as summer recharge for half of all temperate grasslands and temperate forests (15 of 31 locations) and for three-quarters of deserts and xeric shrublands (7 of 9 locations). Also, one quarter of temperate or arid locations have a winter groundwater recharge efficiency that is more than five times that of the summer.

Seasonal changes in the groundwater recharge ratios for tropical climates (n = 7) show that all of the tropical sites tested here have higher groundwater recharge ratios during the wet season relative to the dry season (i.e., (R/P)wet >> (R/P)dry; Figure 5). Only a few locations were available for

Mediterranean climates (n = 3) and boreal forests (n = 3). Mediterranean climates examined here showed very little variability between summer and winter precipitation δ18O and δ2H values, resulting in highly uncertain isotope-based calculations of groundwater recharge ratios for these coastal locations (i.e., small change between δP(summer) and δP(winter); Figure 6). The few boreal sites (n = 3) have a similar groundwater recharge ratio during the summer and winter seasons. Analysis of groundwater recharge ratios in boreal forests are limited by the lack of groundwater isotopic data, likely associated with the low boreal population density (~2.5 persons/km2) relative to the global average (~50 persons/km2; population dataset from www.sedac.ciesin.columbia.edu/data/collection/gpw-v3). 102

Figure 2-6. Seasonal differences in groundwater recharge ratios (recharge/precipitation: R/P) between the (a) summer and winter seasons (extra-tropics), or between the (b) wet and dry seasons

(tropics). Colored bars mark the 25th-75th percentile ranges of calculation results and lines mark the

10th-90th percentile range of calculation results. 103

Table 2-2. Seasonal groundwater recharge ratios (isotope-based)

-

-

winter

season1 season2

/R

P(winter)

P(annual)

P(winter)

P(annual)

P(summer) P(summer)

(%) (%)

P(summer) P(summer)

O

O

H

H

Station

O 2

2

18 H

18

2

δ

δ

summer

18

δ

δ

(R/P) (R/P)

δ

δ

R

Cayenne −2.2 −10 1.4 4 0 0 - 39 0 Taguac −5.3 −33 2.2 19 0 - 0.3 65 - 100 0 - 48 Seawell −1.9 −6 1.8 13 0 - 0.1 17 - 35 0 - 6 Jakarta −5.6 −35 1.1 9 0 - 0.2 48 - 100 0 - 26 New Dehli −5.8 −38 5.0 41 0 - 0.3 11 - 23 0 - 21 Dar es Salaam −2.6 −12 1.7 15 0 - 0.3 3.9 - 16 0 Addis Ababa −1.3 +3 0.9 9 0 29 - 96 0 Santa Maria −5.0 −35 2.0 12 0 - 0.1 0 - 30 0 - 100 Beit Dagan −5.1 −22 1.7 6 0 - 0.4 0 - 15 0 - 34 Pisa −5.5 −33 0.7 n/a 0 - 0.2 34 - 75 0 - 10 Trout Lake −11.1 −77 6.1 49 0 - 0.6 85 - 100 0 - 24 Yellowstone −16.2 −122 9.4 69 0.2 - 0.6 12 - 26 3.2 - 6 Smith's Ferry −15.6 −118 4.9 36 0 - 0.3 11 - 16 0 - 5 Lake Geneva −7.6 −53 4.5 32 0.4 - 1.1 33 - 39 21 - 24 East MA −7.5 −51 2.2 25 0 - 1.4 17 - 56 0 - 30 Niwot Saddle −17.6 −130 8.5 65 0.3 - 0.7 3.9 - 5 2.2 - 4.4 Wye −7.3 −44 2.8 17 0 - 1.7 1.2 - 16 16 - 26 Purdue Agr. −5.7 −33 3.4 24 0 - 1.5 22 - 40 0 - 18 Clinton Stn. −5.0 −29 1.8 16 0 - 0.8 0 - 27 3.6 - 18 Caddo Valley −4.9 −27 2.1 15 0 - 0.6 16 - 21 0.8 - 9 Coffeeville −5.0 −32 1.5 12 0 - 1.9 0 - 24 21 - 72 Saturna −10.9 −79 2.2 14 0.2 - 2.4 6 - 57 64 - 100 Ottawa −11.0 −75 5.5 38 0.6 - 1.7 46 - 85 40 - 71 Wallingford −7.2 −49 1.5 10 0 - 0.6 15 - 54 0 - 25 P. Douradas −7.6 −45 0.9 6 0 - 0.5 12 - 42 0 - 39 Krakow −9.1 −65 3.8 29 0.2 – 1.0 22 - 38 4.4 - 13 Cuxhave −7.0 −49 1.4 9 0 - 0.5 22 - 51 0 - 17 Orleans −6.9 −46 1.8 13 0.1 - 1.4 0 - 16 0 - 6 Melbourne −5.0 −28 1.4 15 0 - 0.1 8 - 25 0 - 1.4 Newcastle −11.2 −89 4.3 47 0 - 2.0 1.3 - 8 0 - 1.2 Little Bighorn −15.1 −115 5.9 44 0 - 0.5 1.9 - 3.1 0 - 0.4 Lamberton −7.6 −51 6.7 37 0.6 - 2.0 31 - 47 7 - 11 N. Platte Agr. −9.0 −61 5.6 53 1.3 - 4.6 13 - 36 5 - 8

104

-

-

winter

season1 season2

/R

P(winter)

P(annual)

P(winter)

P(annual)

P(summer) P(summer)

(%) (%)

P(summer) P(summer)

O

O

H

H

Station

O 2

2

18 H

18

2

δ

δ

summer

18

δ

δ

(R/P) (R/P)

δ

δ

R

Mon Mouth −6.7 −41 3.5 24 0.5 - 2.1 14 - 25 8 - 15 Great Plains −5.8 −35 2.4 17 0.7 - 4.7 0 - 22 22 - 50 Edmonton −17.6 −131 10.6 84 1.1 - 2.7 68 - 100 43 - 56 Saskatoon −14.3 −111 9.0 76 0.1 - 0.5 up to 100 10 - 27 Wynyard −16.0 −124 7.8 62 0.6 - 1.4 78 - 100 33 - 52 Esther −15.7 −124 10.0 72 0.3 - 1.0 up to 100 17 - 37 Calgary −17.8 −138 8.6 49 0.2 - 3.7 48 - 100 36 - 63 Gimli −14.0 −102 11.3 72 1.8 - 2.8 4.1 - 6 4.4 - 5 Craters of Moon −16.9 −128 3.3 52 0.1 - 0.5 0 - 0.1 0 - 0.1 Pinedale −14.8 −110 9.9 38 0.1 - 0.7 6 - 14 2.1 - 6 Sand Spring −12.8 −96 6.5 61 0 - 0.1 21 - 31 0 - 1.6 Smith Valley −12.4 −94 3.8 24 0 - 0.4 33 - 100 0 - 55 Tuscon −7.1 −53 2.5 8 0 13 - 25 0 Chihuahua −4.1 −26 6.1 42 0 - 1.9 0 - 18 0 - 2.4 Alice Springs −5.2 −22 1.6 20 0 - 0.6 0 - 15 0 - 0 Zhangye −6.7 −46 9.4 61 0.6 - 2.9 1.2 - 2.2 0.3 - 0.5 Yinchuan −7.4 −48 8.9 50 0.8 - 1.9 4.6 - 14 0 - 0.7 Yellowknife −20.7 −158 2.5 21 0 - 1.9 0 - 64 56 - 100 Whitehorse −21.3 −164 4.9 31 0.2 - 2.9 48 - 100 25 - 66 Chapais −13.5 −97 5.2 45 1.2 - 1.5 54 - 100 44 - 64 ** Seawell (Barbados) recharge data from Jones and Banner, 2000 * Taguac (Guam) recharge data from Jocson et al., 2002

105 2.5 Discussion

Recharge ratios were calculated for 54 aquifer-precipitation pairings. Further, an additional

16 sites were available for a comparison of precipitation and groundwater δ18O and δ2H values, but were not suited for quantifying groundwater recharge ratios due to the lack of summer or winter precipitation end-members. A comparison of δ18O and δ2H values for the amount-weighted isotopic composition of precipitation and groundwater is shown in Figure 2-7 for these 70 locations (average

±1 s.d. uncertainty). Groundwater matched the amount-weighted precipitation from nearby monitoring stations within 1 ‰ for δ18O and within 9 ‰ for δ2H for half of the locations in this study or 2 ‰ for δ18O and within 16 ‰ for δ2H for four-fifths of study locations.

Figure 2-7. Differences in the stable oxygen and hydrogen isotopic compositions of amount- weighted precipitation (δP(annual)) and local groundwaters (δGroundwater). Error bars mark one standard deviation from the mean.

106 The difference between precipitation and groundwater isotopic compositions ranges from

+1.8 ‰ to −5.6 ‰ for δ18O and from +9 ‰ to −45 ‰ for δ2H. The closest match between the isotopic composition of groundwater and precipitation were found in the tropics. All locations having an average groundwater δ18O value of greater than −5 ‰ have an amount-weighted precipitation value that matches groundwater is within 1.5 ‰. In contrast, regions with lower δ18O groundwater values have a broader range of differences between groundwater and precipitation. At locations where groundwater δ18O values are less than −10 ‰ (n=24) the range in δ18OGroundwater −

δ18OP(annual) was between −5.6 ‰ and +1.0 ‰.

More tightly constrained groundwater-precipitation isotopic data in regions with higher

δ18O values is reconciled by an examination of seasonal fluctuations in δ18O and δ2H values.

Regions having higher δ18O and δ2H values also have more subdued seasonal fluctuations in the isotopic composition of precipitation. Conversely, regions with lower δ18Oprecipitation and δ2Hprecipitation values tend to exhibit greater seasonal changes in the isotopic composition of precipitation (Figure 2-

8).

107

Figure 2-8. The absolute value of the difference between summer and winter δ18O values for 333 globally-distributed locations. The top pane shows the seasonality of δ18O on a global map, and this spatial presentation reveals that locations having the greatest seasonality in precipitation δ18O are located at high latitudes or far from continents. The bottom pane presents a cross plot of the seasonality of δ18O in precipitation, with each point representing on precipitation monitoring station.

The most subdue seasonal fluctuations in δ18O are also locations that have high overall δ18O values, and tend to be located in the humid tropics.

The difference between summer (April to September) and winter (October to March) δ18O values is less than 2 ‰ for the overwhelming majority (95 %) of stations that have an amount- weighted δ18Oprecipitation value greater than −3 ‰ (i.e., 18 of 19 stations). Conversely, the difference between summer and winter δ18O values is greater than 5‰ for the majority (87 %) of stations having an amount-weighted precipitation δ18O value below −15 ‰ (i.e., 27 of 31 stations). 108 Geographically, stations located within the tropics have an average difference between winter and summer δ18O values of 2.3 ‰ (s.d. of 1.6 ‰, n = 46), whereas locations in the extra-tropics have an average difference between winter and summer δ18O values of 5.0 ‰ (s.d. of 4.0 ‰, n = 176).

Offsets have also been reported for North America, where surface water is isotopically light compared to rainfall, due in large part to snow fall and snow melt water inputs in the western North

American watersheds compared to the central and eastern regions of North America (Dutton et al.,

2005).

Overall, it appears that groundwater values may be of use as a proxy for the long-term annual amount-weighted isotopic composition of precipitation in some cases, but that the application of an offset may be appropriate because the majority of groundwaters have lower δ18O and δ2H values than amount-weighted annual precipitation. There may be the potential to develop predictive models of the isotopic composition of groundwater that can complement existing global maps of the isotopic composition of precipitation (Bowen and Wilkinson, 2002; Bowen and Revenaugh, 2003).

Now this discussion will turn attention from raw isotopic datasets to groundwater recharge ratios. As a reminder, this chapter shows that arid and temperate climates have higher winter recharge ratios than summer recharge ratios. This suggests that a given unit change in winter precipitation will be more important than the same unit change in summer precipitation, from a groundwater recharge perspective.

The high groundwater recharge ratios found during the winter in arid and temperate climates may be due to seasonal changes in evapotranspiration potential. Many arid and temperate climates examined here have pronounced seasonal differences in surface temperatures and plant productivity.

Lower recharge ratios during summertime are explained in part by the higher potential for evapotranspiration. Higher winter recharge ratios are explained in part by lower potentials for evapotranspiration because of reduced atmospheric temperatures and dormant vegetation (Welker et

109 al., 1991, Blumenthal et al., 2008, Chimner and Welker, 2005; Anderson-Smith et al., 2014). A global map of the seasonality in chlorophyll abundance, calculated using long-term monthly mean values of the normalized difference vegetation index (NDVI), highlights the pronounced seasonality of plant growth (Figure 2-9).

Figure 2-9. The ratio of summer (six-month) and winter (six-month) normalized difference vegetation index spanning Earth’s continents (stippled areas highlight areas having <10% difference between summer and winter normalized difference vegetation indices; southern hemisphere locations have had summer/winter months reversed relative to the northern hemisphere).

One quarter of continental areas – mostly located in the tropics – have less than a 10% difference between summer and winter plant productivity (stippled regions in Figure 2-9), suggesting no a dominant growing season in these regions. The greatest intra-annual changes in plant activity is found in cold regions (defined as having at least one month with a mean temperature <0°C, Bates and Bilello, 1966), which cover one half of the continents (New et al., 2002). Cold regions have normalized difference vegetation indices that are 14 times higher during the summer relative to during the winter (global average), whereas non-cold regions have an average normalized difference

110 vegetation indices that have more subdued intra-annual variations (non-cold-region

NDVIsummer/NDVIwinter have a global average value of 1; Figure 2-9).

Some cold regions have seasonally frozen ground during the winter that inhibits winter groundwater recharge (Hayashi et al., 2003; Cable et al., 2013). This seasonal blocking of recharge may have an effect, but overall it does not appear to override the seasonality of groundwater recharge ratios in temperate regions. The effect of a temporally variable “frozen ground aquitard” may be offset by elevated groundwater recharge during the rapid melt of seasonal snowpack. Groundwater recharge in the United States is more than double monthly precipitation during snowmelt, implying that snowmelt constitutes an important component of annual recharge (Dripps and Bradbury, 2010;

Dripps, 2012; Allan et al., 2014).

There are four temperate locations that show summer groundwater recharge ratios that are higher than winter recharge ratios: Coffeeville (Mississippi, in the southern U.S.A.), Great Plains

Apiaries (Oklahoma, in the south-central U.S.A.) Saturna Island (British Columbia, off the coast of western Canada) and Wye (Maryland, in eastern U.S.A.). It is noteworthy that all of these locations do not have a large winter snowpack (i.e., less than 5 mm of snow-water equivalent in February, as obtained from long-term monthly mean snow water equivalent data analyzed from passive microwave satellite products: www.globsnow.info). Large-scale mapping can provide better knowledge of the importance of snowmelt to annual groundwater recharge. For some locations that have a Mediterranean-type precipitation seasonal pattern (e.g., Saturna Island), wintertime storage may fill and inhibit recharge, generating runoff instead (Sayama et al., 2011). This could in part help to explain the isotope-based observation of higher summer recharge ratios, although more detailed research in these locations is supported by the global perspective presented here.

Groundwater recharge ratios in the tropics are higher during the wet season than during the dry season in all cases examined. This finding suggests that more intense rainfall leads to higher

111 recharge as a proportion of precipitation (i.e., more intense rain leads to higher groundwater recharge ratios). This finding is consistent with previous isotope and water-level monitoring based work in

Namibia, Uganda, Ethiopia and Tanzania (Wanke et al., 2008; Owor et al., 2009; Walraevens et al.,

2009; Taylor et al., 2013). Each of these studies found that groundwater recharge is most efficient during high intensity rainfall events in the tropics. This finding implies that possible increases in the frequency of high-intensity rainfall events under in intensifying water cycle (Durack et al., 2012) may enhance groundwater availability in some tropical locations. However, a more intense water cycle may elevate geohazard risks to local communities (Belle et al., 2013).

Uncertainty in isotope-based calculations of the groundwater recharge ratio in tropical settings are greater than uncertainties than regions with more pronounced seasonality because of the small differences between summer and winter precipitation isotopic compositions (average of 1.9‰).

The intra-annual variability in δ18O values of precipitation is presented in Figure 2-8. Inland and high-latitude locations are characterized by a greater intra-annual range in δ18O and δ2H values than coastal sites and the tropics. Subdued seasonality of δ18O in the tropics results in higher uncertainties in the seasonality of the groundwater recharge ratio, suggesting that isotope-based approaches to calculating seasonal differences in groundwater recharge ratios will be better constrained outputs in hydro-climates characterized by pronounced seasonality. In spite of the high uncertainties, there exist more than 60 tropical locations with long-term precipitation isotopic data (International Atomic

Energy Agency database: www-naweb.iaea.org/napc/ih/IHS_resources_gnip.html), highlighting an unfilled opportunity to calculate groundwater recharge ratios should groundwater investigations be completed at these locations.

Next, I compare the isotope based observations of groundwater recharge ratio seasonality with outputs from a global hydrological model (pers. comm. Y. Wada; e.g., Gleeson et al., 2012). The

112 spatial comparison of the isotope-based groundwater recharge ratios with a global hydrological model is shown in Figure 2-10 (PCR-GLOBWB; Wada et al., 2010).

Figure 2-10. The seasonality of groundwater recharge ratios from isotope-based calculations (this study; points) and a global hydrological model (PCR-GLOBWB; Wada et al., 2010).

113 PCR-GLOB (Wada et al., 2010) is a global hydrological model that simulates water balances at a 0.5o×0.5o spatial resolution and a daily temporal resolution. The model is setup with two soil layers and an underlying groundwater aquifer. The model simulates exchanges such as infiltration, capillary action between the layers, and also simulates exchanges between the top soil horizon and the atmosphere via rainfall, snowmelt, evaporation and transpiration, canopy interception and snowpack storages. The groundwater aquifer in the model is representative of the deeper subsurface, such that vegetation is not considered to play a critical role in these exchanges via hydraulic lift or plant transpiration. Groundwater storage is parameterized using geospatial lithology and topography datasets. A detailed description of the model is presented within works by Wada et al. (2012a; b).

A cross plot comparison of median groundwater recharge ratios from the isotope- and model-based approaches show that PCR-GLOBWB outputs match isotopic outputs (within error) in most, but not all locations in the extra-tropics (a range of PCR-GLOB modelled recharge ratios falling within 100 km of each study location were used to bound possible model recharge ratio values). The 10th-90th percentile range of isotope-based recharge ratios overlaps with modeled PCR-

GLOBWB recharge ratios in 85% of extra-tropical locations analyzed in this chapter (Figure 2-11).

114

Figure 2-11. Comparison of recharge ratios calculated using isotope-based and hydrological modelling based approaches for (a) summer (April-September) and (b) winter (October-March).

Error bars (x-axis) mark the 10th-90th percentile ranges of isotope-based calculations (squares mark the median). PCR-GLOB error bars mark the range of model results within 100 km of each study location. Colors for each square correspond to ecoregions as shown in previous figures in this chapter.

115 The extra-tropical locations where PCR-GLOBWB recharge ratios do not overlap within the

10th-90th percentiles of isotope-based groundwater recharge ratios are all located in regions that have a February snowpack of between 18 and 81 mm (now water equivalent; Trout Lake and Craters of the Moon in the U.S.A., and Edmonton, Saskatoon, Wynyard and Esther in Canada, long-term monthly means of snow storages from www.globsnow.info). Part of the model-isotope differences observed may be the result of the fact that the isotope-based calculation assesses the seasonal distribution of recharge relative to the timing of precipitation, not necessarily the timing of recharge.

For example, recharge of snow falling during the winter (October to March) but not melting until the spring season (e.g., April to June) is – from the isotopic standpoint – winter recharge. Whereas – from the model standpoint – this snowpack-delayed recharge is considered as summer recharge.

Other sources of discrepancy between the isotope and model groundwater recharge ratio estimates include cropland irrigation return flow that are not incorporated into PCR-GLOBWB, nor the isotope based model, per se (as this flow would be evident in the groundwater isotopic data, but will not be included in precipitation fluxes). Irrigation return flows can constitute an overwhelming component of groundwater recharge in some regions, such as the California Central Valley, for example (Faunt, 2009). Irrigation can also aid recharge indirectly by enhancing the proportion of rainfall that infiltrates (e.g., Chiew and McMahon, 1991). Further, PCR-GLOBWB does not include groundwater recharge from lakes, wetlands and rivers that may account for some component of recharge in arid and semi-arid regions.

The broader implications of this study are three fold: (i) implications for climate change, (ii) impliactions for paleoclimatology, and (iii) implications for ecosystem ecology.

Current climate model projections of groundwater recharge are highly uncertain because of large differences between different general circulation models, different downscaling methods and variable coupling with hydrological models (Crosbie et al., 2011). Previous works that have assessed

116 the potential for change to groundwater recharge have found that different climate models range both in the direction and magnitude of predicted changes to groundwater recharge. The differences range from ±20 to ±50% changes to future groundwater recharge (Allan et al., 2010; Stoll et al.,

2011; Dams et al., 2012). Very few models have quantified changes to the intra-annual distribution of groundwater recharge (Dams et al., 2012). Therefore, current models are likely to be overlooking potentially important changes to individual seasons and their associated impacts upon the annual groundwater recharge flux. The isotope based recharge ratios presented here may be used to assess the most important seasonal hydrological processes governing groundwater recharge under a future, warmer and more energetic water cycle.

In temperate regions, our results suggest that a higher percentage of winter precipitation is able to recharge aquifers. This finding suggests that a unit change to winter precipitation will be more important, from a groundwater recharge perspective, than the same unit change to summer precipitation. The bias towards winter recharge could also be altered if the factors limiting summer recharge occur, such as summer evapotranspiration and storm intensities. The observed bias towards winter precipitation recharge in the extra-tropics can been attributed to the seasonal filtering of precipitation, whereby greater proportions of winter precipitation reach the water table relative to the total summer precipitation amount. This result is interpreted to be due to the high evapotranspiration rates that limit the amount of summer precipitation that recharges.

In tropical settings, we found that recharge ratios are highest during the rainy season. This finding supports the integration of rainfall intensity and intra-annual distributions of rainfall amounts as a central component of future forecasts of groundwater recharge in a warming climate. Existing studies of site-specific modeling in Uganda have found that by including intra-annual variability in precipitation amounts, and variable rainfall intensities, into projections of future groundwater recharge fluxes substantially modifies the projection of future groundwater availability, changing the

117 prediction from a 55% decrease to, instead, a 53% increase in annual groundwater recharge (Mileham et al., 2009). Given the large number of precipitation monitoring stations (e.g., 330 locations in

Figure 2-8) and the equations and approach derived in this chapter, a new opportunity now exists to quantify groundwater recharge ratios across the continents using isotopic data for groundwaters. The incorporation of these data as calibration and validation toolsets in groundwater-equipped general circulation models may help to confirm the validity of projections from these models. Similarly, the paired investigation of precipitation and groundwater isotopic compositions at the same field site can be used to test isotope-enabled general circulation models’ conceptualization of groundwater/surface-water interactions (ECHAM: Hoffman et al., 1998; CCSM: Noone et al., 2002;

IsoGSM: Yoshimura et al., 2003; GISS: Schmidt et al., 2007; LMDZ4: Risi et al., 2010; iLOVECLIM:

Roche et al., 2013).

Our finding that groundwater recharge fluxes do not match precipitation fluxes one-to-one

(Figure 2-7) has three partially-overlapping implications for the interpretation of isotope-based paleoclimate proxies such as fossil groundwaters and speleothems.

First, changes to the seasonality of precipitation fluxes may not be recorded – isotopically – on a one-to-one basis in paleoclimate records involving subsurface waters such as paleo- groundwaters, smectite, tree rings, speleothems and vein calcite (e.g., Winograd et al., 1992; Plummer,

1993; McCarroll and Loader, 2004; Asmerom et al., 2010; Stevenson et al., 2010, Winnick et al., 2013;

Mix and Chamberlain, in press). Groundwater recharge is generally a more efficient process during the winter relative to the summer as shown in this study. Paleoclimate records based on groundwaters may be more tightly linked (i.e., biased) to changes in winter (or, wet season) climate, relative to summer (or, dry season). This realization map help to explain some of the discrepancies that have been observed in fossil groundwaters and lake sediment records located near to one another. For example, paleo-limnologic records of Owens Lake, California show δ18O shifts of up to

118 10 ‰ during the past 500,000 years (Smith and Bischoff, 1997; Menking et al., 1997). Conversely, groundwater-precipitated calcite from nearby Devils Hole, Nevada shows a much smaller range of

δ18O shifts (less than 3 ‰) over the past 500,000 years (Winograd et al., 1992).

Second, the dramatic shifts in climate and biomes from the last ice age to today — such as the shift from deserts to forested climates in parts of in Europe and Alaska (Williams, 2003) — may have modified the recharge ratios in these settings, and thereby created changes to groundwater δ18O values. Our limited set of boreal observations in this study are of particular interest for further word because of the apparent similarity between precipitation and groundwater isotopic compositions in this zone. The boreal biome shifted to lower latitudes during the last glacial maximum (Williams,

2003), and could have modified the seasonality of groundwater recharge ratios, too. This research calls for more work in boreal sites that have long-term precipitation δ18O and δ2H data in order to further test this observation.

Third, seasonal changes in the isotopic composition of precipitation, shown in Figure 2-8, provide some information for the possible range of δ18O shifts in paleoclimate records that can be attributed to changes in the seasonality of precipitation. Seasonality is commonly discussed as a potential source isotopic change amongst other factors such as differences in paleo-ocean δ18O, atmospheric and sea surface temperatures and air mass trajectories. For example, a complete shutdown of precipitation from a single six month (summer or winter) interval can account for a shift no greater than ~9 ‰ in δ18O (much less in most regions), if the seasonality of precipitation

δ18O were similar in the past to today. Some lacustrine paleoclimatic records show more than 9 ‰ variation during the Pleistocene (e.g., Owens Lake, California; Smith and Bischoff, 1997), and this analysis may help to put quantitative bounds on the magnitudes of δ18O and δ2H shifts that may be attributed to seasonality when interpreting paleoclimate records.

119 Finally, ecosystem ecology is linked to groundwater recharge fluxes. The groundwater recharge ratio patterns assessed here span a variety of biomes with different plant life forms, with unique temporal and spatial partitioning of water sources by vegetation with different rooting and growth patterns (Ehleringer and Cooper 1992; Dodd et al., 1998; Alstad et al., 1999; Welker 2000;

Dawson et al., 2002; Kulmatiski et al., 2010; Leffler and Welker, 2013). In deserts, temperate grasslands and temperate forests, seasonal hydrological processes support the growth of a diversity of life forms (grasses and shrubs, trees and understory plants) that utilize soil- and ground-water resources from different depths and are thereby linked to water movements close to Earth’s surface.

Developing an improved understanding of the seasonal changes in vegetation and coupled feedbacks to groundwater recharge – such as interception, transpiration and hydraulic redistribution – will help to better predict how large-scale biome shifts may impact groundwater. For example, ongoing tree death due to mountain pine beetle infestation has recently been shown to reduce transpiration fluxes, resulting in a one-third increase in groundwater fluxes that becomes particularly apparent in late summer (Bearup et al., 2014). Changing seasonality in groundwater recharge fluxes due to vegetation shifts have important implications for aquatic species that depend upon groundwater refugia for habitat (e.g., Power et al., 1999). Plant life forms are expected to shift in a warming climate, and yet these shifts will likely contain surprises such as recent work that has shown that some plant species move downhill (toward warmer temperatures) as climate warms, an unexpected response likely related to plant’s selection of optimal water requirements over temperature trends (Crimmins et al.,

2011; Harsch and Janneke, 2014). Continuing to monitor groundwater and precipitation isotopic compositions can help to quantify vegetation water sources and to assess eco-hydrological feedbacks as transpiration fluxes are modified by changing human land uses (Gordon et al., 2005) and plant water use efficiencies (Keenan et al., 2013).

120 2.6 References

Abbott, M. D., Lini, A., and Bierman, P. R. (2000), δ18O, δD and 3H measurements constrain groundwater recharge patterns in an upland fractured bedrock aquifer, Vermont, USA.,

Journal of Hydrology, 228, 101–112.

Aeschbach-Hertig, W., and Gleeson, T. (2012), Regional strategies for the accelerating global problem of groundwater depletion, Nature Geoscience, 5, 853–861, doi:10.1038/ngeo1617.

Aeschbach-Hertig, W., Stute, M., Clark, J. F., Reuter, R. F., and Schlosser, P. (2002), A paleotemperature record derived from dissolved noble gases in groundwater of the Aquia Aquifer

(Maryland, USA), Geochimica et Cosmochimica Acta, 66, 797–817.

Allen, D. M. (2004), Sources of ground water salinity on islands using 18O, 2H, and 34S.

Ground Water, 42, 17–31.

Allen, D. M., Cannon, A. J., Toews, M. W., and Scibek, J. (2010), Variability in simulated recharge using different GCMs, Water Resources Research, 46, W00F03. doi: 10.1029/2009WR008932

Allen, D. M., Stahl, K., Whitfield, P. H., and Moore, R. D. (2014)1 Trends in groundwater levels in British Columbia. Canadian Water Resources Journal/Revue canadienne des ressources hydriques, (Ahead of print).

Alstad, K. P., Welker, J. M., Williams, S. and Trilica, M. J. (1999), Carbon and water relations of Salix monticola in response to winter browsing and changes in surface water hydrology: an isotopic study using δ13C and δ18O, Oecologia, 120, 375–385.

Anderson-Smith, A., Sullivan, P. F. and Welker, J. M. (2014), Increases in tundra shrub density is manifested in vegetation spectral properties and corresponds to greater CO2 sequestration,

Global Change Biology (in press).

121 Araguás-Araguás, L., Froehlich, K., and Rozanski, K. (2000), Deuterium and oxygen-18 isotope composition of precipitation and atmospheric moisture, Hydrological Processes 14, 1341–1355.

Asmerom, Y., Polyak, V. J., and Burns, S. J. (2010), Variable winter moisture in the southwestern United States linked to rapid glacial climate shifts, Nature Geoscience, 3, 114–117.

Bakari, S. S., Aagaard, P., Vogt, R. D., Ruden, F., Brennwald, M. S., Johansen, I., and

Gulliksen, S. (2012), Groundwater residence time and paleorecharge conditions in the deep confined aquifers of the coastal watershed, South-East Tanzania, Journal of Hydrology, 466, 127–140.

Bates, R. E. and Bilello, M. A. (1966), Defining the cold regions of the northern hemisphere,

Hanover, NH: CRREL, Technical Report 178.

Bearup, L., Maxwell, R. M., Clow, D. W., and McCray, J. E. (2014), Hydrological effects of forest transpiration loss in bark beetle-impacted watersheds, Nature Climate Change, Advance online publication (April 20), doi:10.1038/nclimate2198

Belle, P., Aunay, B., Bernardie, S., Grandjean, G., Ladouche, B., Mazué, R. and Join, J.-L.

(2013), The application of an innovative inverse model for understanding and predicting landslide movements (Salazie cirque landslides, Reunion Island), Landslides, 1–13.

Berg, J. A., and Pearson, S. R. (2011), South-central Minnesota groundwater monitoring of the Mt. Simon aquifer: Minnesota, Department of Natural Resources, Ecological and Water

Resources Division Report, 92 pp.

Birks, S. J., and Edwards, T. W. D. (2009), Atmospheric circulation controls on precipitation isotope–climate relations in western Canada, Tellus B, 61, 566–576.

122 Blumenthal, D., Chimner, R., Welker, J. M., and Morgan, J. (2008), Increased snow facilitates plant invasion in mixedgrass prairie, New Phytologist, 179, 440-448. doi:10.1111/j.1469-

8137.2008.02475.x.

Boutin, P. (2009), Etude geochmique des eaux souterraines à la mine joe mann,

Chibougamau,Québec, M.Sc. Thesis at Université du Québec, 141 pp.

Bowen, G. J., and Revenaugh, J. (2003), Interpolating the isotopic composition of modern meteoric precipitation, Water Resources Research, 39, 1299. doi:10.1029/2003WR002086

Bowen, G. J., and Wilkinson, B. (2002), Spatial distribution of δ18O in meteoric precipitation, Geology 30, 315–318.

Bretzler, A., Osenbrück, K., Gloaguen, R., Ruprecht, J. S., Kebede, S., and Stadler, S. (2011),

Groundwater origin and flow dynamics in active rift systems–A multi-isotope approach in the Main

Ethiopian Rift, Journal of Hydrology, 402, 274–289.

Burchuladze, A. A., Chudy, M., Eristavi, I. V., Pagava, S. V., Povinec, P., Sivo, A., and

Togonidze, G. I. (1989), Anthropogenic 14C variations in atmospheric CO2 and wines, Radiocarbon,

31, 771–776.

Cable, J., Olge, K., Bolton, D., Bentley, L., Romonvsky, V., Iwata, H., Harazona, Y., and

Welker, J. M. (2013), Permafrost thaw effects boreal deciduous plant transpiration through increased soil water, increased thaw and warmer temperatures, Ecohydrology DOI: 10.1002/eco.1423.

Carey, S. K., and Quinton, W. L. (2005), Evaluating runoff generation during summer using hydrometric, stable isotope and hydrochemical methods in a discontinuous permafrost alpine catchment, Hydrological Processes, 19, 95–114.

123 Carreira, P. M., Marques, J. M., Marques, J. E., Chaminé, H. I., Fonseca, P. E., Santos, F. M.,

Moura, R. M., and Carvalho, J. M. (2011), Defining the dynamics of groundwater in Serra da Estrela

Mountain area, central Portugal: an isotopic and hydrogeochemical approach, Journal, 19,

117–131.

Castle, S. L., Thomas, B. F., Reager, J. T., Rodell, M., Swenson, S. C. and Famiglietti, J. S.

(2014), Groundwater Depletion During Drought Threatens Future Water Security of the Colorado

River Basin. Geophysical Research Letters, doi: 10.1002/2014GL061055.

Cheung, K. and Mayer, B. (2007), Chemical and isotopic characterization of shallow groundwater from selected monitoring wells in Alberta: Part I: 2006-2007, Alberta Environment

Report, 88 pp.

Chiew, F. H. S. and McMahon, T. A. (1991), Groundwater recharge from rainfall and irrigation in the Campaspe River Basin, Soil Research, 29, 651–670.

Chimner, R. A., and Welker, J. M. (2005), Ecosystem respiration responses to experimental manipulations of winter and summer precipitation in a Mixedgrass Prairie, WY, USA., Biogeochemistry,

73, 257–270. doi:10.1007/s10533-004-1989-6.

Crimmins, S. M., Dobrowski, S. Z., Greenberg, J. A., Abatzoglou, J. T., and Mynsberge, A.

R. (2011), Changes in climatic water balance drive downhill shifts in plant species’ optimum elevations. Science 331, 324-327.

Crosbie, R. S., Dawes, W. R., Charles, S. P., Mpelasoka, F. S., Aryal, S., Barron, O., and

Summerell, G. K. (2011), Differences in future recharge estimates due to GCMs, downscaling methods and hydrological models, Geophysical Research Letters, 38, L11406, doi:10.1029/2011GL047657

124 Cunningham, E. E., Long, A., Eastoe, C., and Bassett, R. L. (1998), Migration of recharge waters downgradient from the Santa Catalina Mountains into the Tucson basin aquifer, Arizona,

USA., Hydrogeology Journal, 6, 94–103.

Dams, J., Salvadore, E., Van Daele, T., Ntegeka, V., Willems, P., Batelaan, O., and Hendricks

Franssen, H. J. (2012), Spatio-temporal impact of climate change on the groundwater system,

Hydrology and Earth System Sciences 16, 1517–1531.

Dansgaard, W. (1964), Stable isotopes in precipitation, Tellus, 16, 436–468.

Darling, W. G., and Bath, A. H. (1988), A stable isotope study of recharge processes in the

English Chalk, Journal of Hydrology, 101, 31–46.

Darling, W. G., Edmunds, W. M., and Smedley, P. L. (1997), Isotopic evidence for palaeowaters in the British Isles, Applied Geochemistry, 12, 813–829.

Darling, W. G., Bath, A. H., and Talbot, J. C. (2003), The O and H stable isotope composition of freshwaters in the British Isles. 2, surface waters and groundwater, Hydrology and Earth

System Sciences, 7, 183–195.

Das, B. K., Kakar, Y. P., Moser, H., and Stichler, W. (1988), Deuterium and oxygen-18 studies in groundwater of the Delhi area, India, Journal of Hydrology, 98, 133-146.

Dawson, T. E., Mambelli, S., Plamboeck, A. H., Templer, P. H., and Tu, K. P. (2002), Stable isotopes in plant ecology. Annual review of ecology and systematics, 33, 507–559.

Delin, G. N., Healy, R. W., Lorenz, D. L., and Nimmo, J. R. (2007), Comparison of local-to regional-scale estimates of ground-water recharge in Minnesota, USA., Journal of Hydrology, 334, 231–

249.

125 Demlie, M., Wohnlich, S., Gizaw, B., and Stichler, W. (2007), Groundwater recharge in the

Akaki catchment, central Ethiopia: evidence from environmental isotopes (δ18O, δ2H and 3H) and chloride mass balance, Hydrological Processes, 21, 807-818.

Dodd, M. B., Lauenroth, W. K., and Welker, J. M. (1998), Differential water resource use by herbaceous and woody plants in a shortgrass steppe community. Oecologia, 117, 504–512.

Döll, P., and Fiedler, K. (2007), Global-scale modeling of groundwater recharge, Hydrology and Earth System Sciences Discussions, 4, 4069–4124.

Douglas, M., Clark, I. D., Raven, K., and Bottomley, D. (2000), Groundwater mixing dynamics at a Canadian Shield mine, Journal of Hydrology, 235, 88–103.

Dripps, W. R., and Bradbury, K. R. (2010), The spatial and temporal variability of groundwater recharge in a forested basin in northern Wisconsin, 24, 383–392.

Dripps, W. R. (2012), An Integrated Field Assessment of Groundwater Recharge, Open

Hydrology Journal, 6, 15–22.

Durack, P. J., Wijffels, S. E., and Matear, R. J. (2012), Ocean salinities reveal strong global water cycle intensification during 1950 to 2000, Science, 336, 455–458.

Dutton, A. L., Wilkinson, B. H., Bowen, G., Welker, J. M., and Lohmann, K. C. (2005),

Comparison of river water and precipitation δ18O across the 48 contiguous United States. Hydrological

Processes, 19, 3551–3572.

Edmunds, W. M., and Milne, C. J. (Eds.) (2001), Palaeowaters in coastal Europe: evolution of groundwater since the late Pleistocene. Geological Society of London.

126 Edmunds, W. M. (2009), Palaeoclimate and groundwater evolution in Africa—implications for adaptation and management. Hydrological Sciences Journal, 54, 781–792.

Ehleringer, J. R. and Cooper, T. A. (1992), On the role of orientation in reducing photoinhibitory damage in photosynthetic-twig desert shrubs, Plant, Cell and Environment, 15, 301–306. doi: 10.1111/j.1365-3040.1992.tb00977.x

Elliot, T., Andrews, J. N., and Edmunds, W. M. (1999), Hydrochemical trends, palaeorecharge and groundwater ages in the fissured Chalk aquifer of the London and Berkshire

Basins, UK, Applied Geochemistry, 14, 333–363.

Even, H., Carmi, I., Magaritz, M., and Gerson, R. (1986), Timing the transport of water through the upper vadose zone in a karstic system above a cave in Israel, Earth Surface Processes and

Landforms, 11, 181–191.

Famiglietti, J. S., Lo, M., Ho, S. L., Bethune, J., Anderson, K. J., Syed, T. H., Swenson, S. C., de Linage, C. R., and Rodell, M. (2011), Satellites measure recent rates of groundwater depletion in

California's Central Valley, Geophysical Research Letters, 38, L03403, doi:10.1029/2010GL046442.

Faunt, C. C. (Ed.), (2009), Groundwater availability of the central valley aquifer, California,

United States Geological Survey Professional Paper 1766, pp. 225.

Feng, W., Zhong, M., Lemoine, J. M., Biancale, R., Hsu, H. T., and Xia, J. (2013), Evaluation of groundwater depletion in North China using the Gravity Recovery and Climate Experiment

(GRACE) data and ground‐based measurements, Water Resources Research, 49, 2110–2118.

Ferguson, G. A., Betcher, R. N., and Grasby, S. E. (2007), Hydrogeology of the Winnipeg formation in Manitoba, Canada, Hydrogeology Journal, 15, 573–587.

127 Foley, J. A., Ramankutty, N., Brauman, K. A., Cassidy, E. S., Gerber, J. S., Johnston, M.,

Mueller, N. D., O’Connell, C., Ray, D. K., West, P. C., Balzer, C., Bennett, E. M., Carpenter, S. R.,

Hill, J., Monfreda, C., Polasky, S., Rockström, J., Sheehan, J., Siebert, S., Tilamn, D., and Zaks, D. P.

(2011). Solutions for a cultivated planet, Nature, 478, 337–342.

Fortin, G., van der Kamp, G., Cherry, J. A. (1991), Hydrogeology and hydrochemistry of an aquifer-aquitard system within glacial deposits, Saskatchewan, Canada, Journal of Hydrology, 126, 265–

292.

Friedman, I. (1953), Deuterium content of natural waters and other substances, Geochimica et

Cosmochimica Acta, 4, 89–103.

Genty, D., Labuhn, I., Hoffmann, G., Danis, P. A., Mestre, O., Bourges, F., Wainer, K.,

Massault, M., Van Exter, S., Régnier, E., Orengo, P., Falourd, S., and Minster, B. (2014), Rainfall and cave water isotopic relationships in two South-France sites, Geochimica et Cosmochimica Acta, 131, 323–

343.

Gibson, J. J., Birks, S. J., and Edwards, T. W. D. (2008), Global prediction of δA and δ2H-

δ18O evaporation slopes for lakes and soil water accounting for seasonality, Global Biogeochemical Cycles,

22, GB2031. doi:10.1029/2007GB002997

Gill, I. (1994), Groundwater geochemistry of the Kingshill aquifer system, St. Croix,

Environmental Geosciences, 1, 40–49.

Gleeson, T., Novakowski, K., and Kyser, T. K. (2009), Extremely rapid and localized recharge to a fractured rock aquifer, Journal of Hydrology 376, 496–509.

Gleeson, T., Wada, Y., Bierkens, M. F., and van Beek, L. P. (2012), Water balance of global aquifers revealed by groundwater footprint, Nature, 488, 197–200.

128 Goede, A., Green, D. C., and Harmon R. S. (1982), Isotopic composition of precipitation, cave drips and actively forming speleothems at three Tasmanian cave sites, Helictite, 20, 17–29.

Gordon, L. J., Steffen, W., Jönsson, B. F., Folke, C., Falkenmark, M. and Johannessen, Å.

(2005), Human modification of global water vapor flows from the land surface, Proceedings of the

National Academy of Sciences of the United States of America, 102, 7612–7617.

Granger, R. J., Gray, D. M., and Dyck, G. E. (1984), Snowmelt infiltration to frozen Prairie soils, Canadian Journal of Earth Sciences, 21, 669–677

Grasby, S. E., and Chen, Z. (2005), Subglacial recharge into the Western Canada

Sedimentary Basin—Impact of Pleistocene glaciation on basin hydrodynamics, Geological Society of

America Bulletin, 117, 500–514.

Grasby, S. E., Osborn, J., Chen, Z., and Wozniak, P. R. (2010), Influence of till provenance on regional groundwater geochemistry. Chemical Geology, 273, 225–237.

Grassi, S., and Cortecci, G. (2005), Hydrogeology and geochemistry of the multilayered confined aquifer of the Pisa plain (Tuscany–central Italy), Applied Geochemistry, 20, 41–54.

Harsch, M. A. and HilleRisLambers, J. (2014), Species distributions shift downward across western North America, Global Change Biology, doi: 10.1111/gcb.12697

Hayashi, M., van der Kamp, G., and Schmidt, R. (2003), Focused infiltration of snowmelt water in partially frozen soil under small depressions, Journal of Hydrology, 270, 214–229.

Heppner, C. S., Nimmo, J. R., Folmar, G. J., Gburek, W. J., and Risser, D. W. (2007),

Multiple-methods investigation of recharge at a humid-region fractured rock site, Pennsylvania, USA.

Hydrogeology Journal, 15, 915–927.

129 Hoffmann, G., Werner, M., and Heimann, M. (1998), Water isotope module of the ECHAM atmospheric general circulation model: A study on timescales from days to several years, Journal of

Geophysical Research: Atmospheres, 103, 16871–16896.

Huddart, P. A., Longstaffe, F. J., and Crowe, A. S. (1999), δD and δ18O evidence for inputs to groundwater at a wetland coastal boundary in the southern Great Lakes region of Canada, Journal of Hydrology, 214, 18–31.

Iacumin, P., Venturelli, G., and Selmo, E. (2009), Isotopic features of rivers and groundwater of the Parma Province (Northern Italy) and their relationships with precipitation. Journal of Geochemical Exploration, 102, 56–62.

Jasechko, S., Sharp, Z. D., Gibson, J. J., Birks, S. J., Yi, Y., and Fawcett, P. J. (2013),

Terrestrial water fluxes dominated by transpiration, Nature, 496, 347–350.

Jiráková, H., Huneau, F., Celle-Jeanton, H., Hrkal, Z., and La Coustumer, P. L. (2011),

Insights into palaeorecharge conditions for European deep aquifers, Hydrogeology Journal, 19, 1545–

1562.

Jocson, J.M.U., Jenson, J.W., and Contractor, D.N. (2002), Recharge and Aquifer response:

Northern Guam Lens Aquifer, Guam, Mariana Islands, Journal of Hydrology, 260, 231–254.

Jones, I. C., Banner, J. L., and Humphrey, J. D. (2000), Estimating recharge in a tropical karst aquifer, Water Resources Research, 36, 1289–1299.

Jones, I. C., and Banner, J. L. (2003), Estimating recharge thresholds in tropical karst island aquifers: Barbados, Puerto Rico and Guam, Journal of Hydrology, 278, 131–143.

130 Joodaki, G., Wahr, J. and Swenson, S. (2014), Estimating the human contribution to groundwater depletion in the Middle East, from GRACE data, land surface models, and well observations, Water Resources Research, 50, doi:10.1002/2013WR014633.

Jukić, D., and Denić-Jukić V. (2009), Groundwater balance estimation in karst by using a conceptual rainfall–runoff model, Journal of Hydrology, 373, 302–315.

Julian, J., Araguás-Araguás, L., Rozanski, K., Benavente, J., Cardenal, J., Hidalgo, M. C.,

Garcia-Lopez, S., Martinez-Garrido, J. C. Moral, F., and Olias. M. (1992), Sources of precipitation over South‐Eastern Spain and groundwater recharge. An isotopic study, Tellus B, 44, 226–236.

Kagabu, M., Shimada, J., Delinom, R., Tsujimura, M., and Taniguchi, M. (2011),

Groundwater flow system under a rapidly urbanizing coastal city as determined by hydrogeochemistry, Journal of Asian Earth Sciences, 40, 226–239.

Karro, E., Marandi, A., and Vaikmäe, R. (2004), The origin of increased salinity in the

Cambrian-Vendian aquifer system on the Kopli Peninsula, northern Estonia. Hydrogeology Journal, 12,

424–435.

Kalin, R. M. (1994), The hydrogeochemical evolution of the groundwater of the Tucson

Basin with application to 3-dimensional groundwater flow modelling, Ph.D. thesis at the University of Arizona, 510 pp.

Kebede, S., Travi, Y., Asrat, A., Alemayehu, T., Ayenew, T., and Tessema, Z. (2008),

Groundwater origin and flow along selected transects in Ethiopian rift volcanic aquifers, Hydrogeology

Journal, 16, 55–73.

131 Keenan, T. F., Hollinger, D. Y., Bohrer, G., Dragoni, D., Munger, J. W., Schmid, H. P., and

Richardson, A. D. (2013), Increase in forest water-use efficiency as atmospheric carbon dioxide concentrations rise, Nature, 499, 324–327.

Kim, J. H., and Jackson, R. B. (2012), A global analysis of groundwater recharge for vegetation, climate, and soils. Vadose Zone Journal, 11, doi:10.2136/vzj2011.0021RA

Kloppmann, W., Dever, L., and Edmunds, W. M. (1998), Residence time of Chalk groundwaters in the Paris Basin and the North German Basin: a geochemical approach, Applied

Geochemistry, 13, 593–606.

Konikow, L. F. (2011), Contribution of global groundwater depletion since 1900 to sea‐level rise, Geophysical Research Letters, 38, L17401.

Konikow, L. F., and Kendy, E. (2005), Groundwater depletion: A global problem,

Hydrogeology Journal, 13, 317–320, doi:10.1007/s10040-004-0411-8

Kortelainen, N. M., and Karhu, J. A. (2004), Regional and seasonal trends in the oxygen and hydrogen isotope ratios of Finnish groundwaters: a key for mean annual precipitation, Journal of

Hydrology, 285, 143–157.

Kotzer, T. G., Kudo, A., Zheng, J., and Workman, W. (2000), Natural and anthropog en ic levels of tritium in a Canadian Arctic ice core, Agassiz Ice Cap, Ellesmere Island, and comparison with other radionuclides. Journal of Glaciology 46, 35–40.

Kulmatiski, A., Beard, K. H., Verweij, R. J., and February, E. C. (2010), A depth‐controlled tracer technique measures vertical, horizontal and temporal patterns of water use by trees and grasses in a subtropical savanna. New Phytologist, 188, 199–209.

132 Kurita, N., Yoshida, N., Inoue, G., and Chayanova, E. A. (2004), Modern isotope climatology of Russia: A first assessment, Journal of Geophysical Research: Atmospheres, 109, D03102, doi:10.1029/2003JD003404.

Lanza, S. (2009), Groundwater anammox at an industrial site in Calgary, M.Sc. Thesis at the

University of Calgary, 96 pp.

Lee, K. S., Wenner, D. B., and Lee, I. (1999), Using H-and O-isotopic data for estimating the relative contributions of rainy and dry season precipitation to groundwater: example from Cheju

Island, Korea, Journal of Hydrology, 222, 65–74.

Lee, K. S., and Kim, Y. (2007), Determining the seasonality of groundwater recharge using water isotopes: a case study from the upper North Han River basin, Korea. Environmental Geology, 52,

853–859.

Leffler, J., and Welker, J. M. (2013), Long-term increases in snow elevate leaf N and photosynthesis in Salix arctica: response to a snow fence experiment in NW Greenland, Environmental

Research Letters, 8, 025023.

Leterme, B., Mallants, D., and Jacques, D. (2012), Sensitivity of groundwater recharge using climatic analogues and HYDRUS-1D, Hydrology and Earth System Sciences, 16, 2485–2497.

Li, B., Yuan, D., Qin, J., Lin, Y., and Zhang, M. (2000), Oxygen and carbon isotopic characteristics of rainwater, drip water and present speleothems in a cave in Guilin area, and their environmental meanings. Science in China Series D: Earth Sciences, 43, 277–285.

Liu, Z., Bowen, G., Yoshimura, K., and Welker, J. M. (2013), Pacific North American teleconnection controls on precipitation isotopes (δ18O) across the contiguous USA and adjacent regions: A GCM-Based Analysis, Journal of Climate, 27, 1046-1061, doi:10.1175/JCLI-D-13-00334.1.

133 Liu, Z., Yoshimura, K., Bowen, G. J., Buenning, N. H., Risi, C., Welker, J. M. and Yuan, F.

(2014), Paired oxygen isotope records reveal modern North American atmospheric dynamics during the Holocene, Nature Communications, 5, doi:10.1038/ncomms4701

Lorenzen, G., Sprenger, C., Baudron, P., Gupta, D., and Pekdeger, A. (2012), Origin and dynamics of groundwater salinity in the alluvial plains of western Delhi and adjacent territories of

Haryana State, India, Hydrological Processes, 26, 2333–2345.

Madonia, P., D’Aleo, R., Di Maggio, C., Favara, R., and Hartwig, A. (2013), The use of shallow dripwater as an isotopic marker of seepage in karst areas: A comparison between Western

Sicily (Italy) and the Harz Mountains (Germany), Applied Geochemistry, 34, 231–239.

Maulé, C. P., Chanasyk, D. S., and Muehlenbachs, K. (1994), Isotopic determination of snow-water contribution to soil water and groundwater, Journal of Hydrology, 155, 73–91.

McCarroll, D., and Loader, N. J. (2004), Stable isotopes in tree rings, Quaternary Science

Reviews, 23, 771–801.

McIntosh, J. C., Schlegel, M. E., and Person, M. (2012), Glacial impacts on hydrologic processes in sedimentary basins: evidence from natural tracer studies, Geofluids, 12, 7–12. doi:10.1111/j.1468-8123.2011.00344.x

McMahon, P. B., Bohlke, J. K., and Carney, C. P. (2007), Vertical gradients in water chemistry and age in the northern High Plains aquifer, Nebraska, 2003, US Department of the

Interior, US Geological Survey Scientific Investigations Report 2006–5294, 66 pp.

Menking, K. M., Bischoff, J. L., Fitzpatrick, J. A., Burdette, J. W., and Rye, R. O. (1997),

Climatic/hydrologic oscillations since 155,000 yr BP at Owens Lake, California, reflected in abundance and stable isotope composition of sediment carbonate, Quaternary Research, 48, 58–68.

134 Mileham, L., Taylor, R. G., Todd, M., Tindimugaya, C., and Thompson, J. (2009), The impact of climate change on groundwater recharge and runoff in a humid, equatorial catchment: sensitivity of projections to rainfall intensity, Hydrological Sciences Journal, 54, 727–738.

Mix, H. T., and Chamberlain, C. P. (2014), Stable isotope records of hydrologic change and paleotemperature from smectite in Cenozoic western North America. Geochimica et Cosmochimica Acta, doi: 10.1016/j.gca.2014.07.008

Négrel, P., and Giraud, E. P. (2010), Geochemistry, isotopic composition (δ18O, δ2H,

87Sr/86Sr, 143Nd/144Nd) in the groundwater of French Guiana as indicators of their origin, interrelations, Comptes Rendus Géosciences, 342, 786–795.

New, M., Lister, D., Hulme, M., and Makin, I. (2002), A high-resolution data set of surface climate over global land areas, Climate Research, 21, 1–25.

Noone, D., and Simmonds, I. (2002), Associations between δ18O of water and climate parameters in a simulation of atmospheric circulation for 1979–95, Journal of Climate, 15, 3150–3169.

O'driscoll, M. A., DeWalle, D. R., McGuire, K. J., and Gburek, W. J. (2005), Seasonal 18O variations and groundwater recharge for three landscape types in central Pennsylvania, USA., Journal of Hydrology, 303, 108–124.

Owor, M., Taylor, R. G., Tindimugaya, C., and Mwesigwa, D. (2009), Rainfall intensity and groundwater recharge: empirical evidence from the Upper Nile Basin. Environmental Research Letters, 4,

035009.

Plummer, L. N. (1993), Stable isotope enrichment in paleowaters of the southeast Atlantic

Coastal Plain, United States, Science, 262, 2016–2020.

135 Power, G., Brown, R. S. and Imhof, J. G. (1999), Groundwater and fish—insights from northern North America, Hydrological Processes, 13, 401–422

Praamsma, T., Novakowski, K., Kyser, K., and Hall, K. (2009), Using stable isotopes and hydraulic head data to investigate groundwater recharge and discharge in a fractured rock aquifer,

Journal of Hydrology, 366, 35–45.

Qin, D., Qian, Y., Han, L., Wang, Z., Li, C., and Zhao, Z. (2011). Assessing impact of irrigation water on groundwater recharge and quality in arid environment using CFCs, tritium and stable isotopes, in the Zhangye Basin, Northwest China, Journal of Hydrology, 405, 194–208.

Rango, T., Petrini, R., Stenni, B., Bianchini, G., Slejko, F., Beccaluva, L., and Ayenew, T.

(2010), The dynamics of central Main Ethiopian Rift waters: Evidence from δD, δ18O and 87Sr/86Sr ratios, Applied Geochemistry, 25, 1860–1871.

Risi, C., Bony, S., Vimeux, F., and Jouzel, J. (2010), Water‐stable isotopes in the LMDZ4 general circulation model: Model evaluation for present‐day and past climates and applications to climatic interpretations of tropical isotopic records, Journal of Geophysical Research: Atmospheres, 115,

D12118. doi:10.1029/2009JD013255

Roche, D. M. (2013), δ18O water isotope in the iLOVECLIM model (version 1.0)-Part 1:

Implementation and verification, Geoscientific Model Development Discussions, 6, 1481–1491.

Rock, L., and Mayer, B. (2009), Identifying the influence of geology, land use, and anthropogenic activities on riverine sulfate on a watershed scale by combining hydrometric, chemical and isotopic approaches, Chemical Geology, 262, 121–130.

Rodell, M., Velicogna, I., and Famiglietti, J. S. (2009), Satellite-based estimates of groundwater depletion in India, Nature, 460, 999–1002.

136 Rozanski, K., Gonfiantini, R., and Araguás-Araguás, L. (1991), Tritium in the global atmosphere: distribution patterns and recent trends, Journal of Physics G: Nuclear and Particle Physics, 17,

S523.

Samborska, K., Różkowski, A., and Małoszewski, P. (2013), Estimation of groundwater residence time using environmental radioisotopes (14C, T) in carbonate aquifers, southern Poland,

Isotopes in Environmental and Health Studies, 49, 73–97.

Savard, M. M., Paradis, D., Somers, G., Liao, S., and van Bochove, E. (2007), Winter nitrification contributes to excess NO3− in groundwater of an agricultural region: A dual‐isotope study, Water Resources Research, 43, W06422, doi:10.1029/2006WR005469.

Sayama, T., McDonnell, J. J., Dhakal, A. and Sullivan, K. (2011), How much water can a watershed store? Hydrological Processes, 25, 3899–3908. doi:10.1002/hyp.8288

Siebert, S, Burke, J., Faures, J. M., Frenken, K., Hoogeveen, J., Döll, P., and Portmann, F. T.

(2010), Groundwater use for irrigation – a global inventory, Hydrology and Earth System Sciences, 14,

1863–1880, doi:10.5194/hess-14-1863-2010

Scanlon, B. R., Keese, K. E., Flint, A. L., Flint, L. E., Gaye, C. B., Edmunds, W. M., and

Simmers, I. (2006), Global synthesis of groundwater recharge in semiarid and arid regions,

Hydrological Processes, 20, 3335–3370.

Scanlon, B. R., Faunt, C. C., Longuevergne, L., Reedy, R. C., Alley, W. M., McGuire, V. L., and McMahon, P. B. (2012), Groundwater depletion and sustainability of irrigation in the US High

Plains and Central Valley, Proceedings of the National Academy of Sciences, 109, 9320–9325.

137 Schmidt, G. A., LeGrande, A. N., and Hoffmann, G. (2007), Water isotope expressions of intrinsic and forced variability in a coupled ocean‐atmosphere model, Journal of Geophysical Research:

Atmospheres, 112, D10103, doi:10.1029/2006JD007781.

Simpson, E. S., Thorud, D. B., and Friedman, I. (1972), Distinguishing seasonal recharge to groundwater by deuterium analysis in southern Arizona, Proceedings of the Reeding Symposium,

International Association of Scientific Hydrology, 113–121.

Smith, G.I., and Bischoff, J. L., Eds. (1997), An 800,000-year Paleoclimatic Record from

Core OL-92, Owens Lake, Southeast California, Geological Society of America Special Paper 317, 37–47.

Stevenson, B. A., Kelly, E., McDonald, E., Busacca, A., and Welker, J. M. (2010), Oxygen isotope ratios in Holocene carbonates across a climatic gradient, eastern Washington State, USA:

Evidence for seasonal effects on pedogenic mineral isotopic composition, The Holocene. 20, 575–583, doi:10.1177/0959683609356588.

Steward, D. R., Bruss, P. J., Yang, X., Staggenborg, S. A., Welch, S. M., and Apley, M. D.

(2013), Tapping unsustainable groundwater stores for agricultural production in the High Plains

Aquifer of Kansas, projections to 2110, Proceedings of the National Academy of Sciences, 110, E3477–

E3486.

Stoll, S., Franssen, H. J., Butts, M., and Kinzelbach, W. (2011), Analysis of the impact of climate change on groundwater related hydrological fluxes: a multi-model approach including different downscaling methods, Hydrology and Earth System Sciences, 15, 21–38. doi:10.5194/hess-15-21-

2011

138 Taylor, R. G., Todd, M. C., Kongola, L., Maurice, L., Nahozya, E., Sanga, H., and

MacDonald, A. M. (2013), Evidence of the dependence of groundwater resources on extreme rainfall in East Africa, Nature Climate Change, 3, 374–378.

Tweed, S. O., Weaver, T. R., and Cartwright, I. (2005), Distinguishing groundwater flow paths in different fractured-rock aquifers using groundwater chemistry: Dandenong Ranges, southeast Australia, Hydrogeology Journal, 13, 771–786.

Vachon, R. W., Welker, J. M., White, J. W. C., and Vaughn, B. H. (2010), Monthly precipitation isoscapes (δ18O) of the United States: Connections with surface temperatures, moisture source conditions, and air mass trajectories, Journal of Geophysical Research: Atmospheres, 115, D21126. doi:10.1029/2010JD014105.

Van Beynen, P., and Febbroriello, P. (2006), Seasonal isotopic variability of precipitation and cave drip water at Indian Oven Cave, New York. Hydrological Processes, 20, 1793–1803.

Vogel, J. C., Ehhalt, D., and Roether, W. (1963), A survey of the natural isotopes of water in

South Africa, Proceedings of Tokyo Conference on Radioisotopes in Hydrology, 407–416.

Voss, K. A., Famiglietti, J. S., Lo, M., Linage, C., Rodell, M., and Swenson, S. C. (2013),

Groundwater depletion in the Middle East from GRACE with implications for transboundary water management in the Tigris‐Euphrates‐Western Iran region, Water Resources Research, 49, 904–914.

Wada, Y., van Beek, L. P. H., van Kempen, C. M., Reckman, J. W. T. M., Vasek, S.,

Bierkens, M. F. P. (2010), Global depletion of groundwater resources. Geophysical Research. Letters, 38,

L20402. doi:10.1029/2010GL044571.

139 Wada, Y., L. P. H. van Beek, and M. F. P. Bierkens (2012a), Nonsustainable groundwater sustaining irrigation: A global assessment, Water Resources Research, 48, W00L06, doi:10.1029/2011WR010562.

Wada, Y., L. P. H. van Beek, F. C. S. Weiland, B. F. Chao, Y.-H. Wu, and M. F. P. Bierkens

(2012b), Past and future contribution of global groundwater depletion to sea-level rise, Geophysical

Research Letters, 39, L09402, doi:10.1029/2012GL051230.

Wada, Y., Wisser, D., and Bierkens, M. F. P. (2014), Global modeling of withdrawal, allocation and consumptive use of surface water and groundwater resources, Earth System Dynamics

Discussions, 4, 355–392. doi:10.5194/esd-5-15-2014

Wallick, E. I. (1981), Chemical evolution of groundwater in a drainage basin of Holocene age, east-central Alberta, Canada. Journal of Hydrology, 54, 245–283.

Walraevens, K., Vandecasteele, I., Martens, K., Nyssen, J., Moeyersons, J., Gebreyohannes,

T., de Smedt, F., Poesen, J., Deckers, J. and Van Camp, M. (2009), Groundwater recharge and flow in a small mountain catchment in northern Ethiopia. Hydrological Sciences Journal, 54, 739–753.

Wanke, H., Dünkeloh, A., and Udluft, P. (2008), Groundwater recharge assessment for the

Kalahari catchment of north-eastern Namibia and north-western Botswana with a regional-scale water balance model, Water Resources Management, 22, 1143–1158.

Wang, L., Hu, F., Yin, L., Wan, L., and Yu, Q. (2013), Hydrochemical and isotopic study of groundwater in the Yinchuan plain, China, Environmental Earth Sciences, 69, 2037–2057.

Wassenaar, L. I., Van Wilgenburg, S. L., Larson, K., and Hobson, K. A. (2009), A groundwater isoscape (δD, δ18O) for Mexico, Journal of Geochemical Exploration, 102, 123–136.

140 Welker, J. M., McClelland, S., and Weaver, T. W. (1991), Soil water retention after natural and simulated rainfall on a temperate grassland. Theoretical and Applied Climatology, 44, 447-453.

Welker, J. M. (2000), Isotopic (δ18O) characteristics of weekly precipitation collected across the USA: an initial analysis with application to water source studies. Hydrological Processes, 14, 1449–

1464.

Welker, J. M. (2012), ENSO effects on the isotopic (δ18O, δ2H and d-excess) of precipitation across the US using a long-term network (USNIP), Rapid Communication in Mass Spectrometery, 17,

1655–1660.

West, J. B., Bowen, G. J., Cerling, T. E., and Ehleringer, J. R. (2006), Stable isotopes as one of nature's ecological recorders. Trends in Ecology and Evolution, 21, 408–414.

Williams, P. W., and Fowler, A. (2002), Relationship between oxygen isotopes in rainfall, cave percolation waters at Waitomo, New Zealand, Journal of Hydrology, 41, 53–70.

Williams, J. W. (2003), Variations in tree cover in North America since the last glacial maximum, Global and Planetary Change, 35, 1–23.

141 CHAPTER 3 — THE ISOTOPIC COMPOSITION OF ICE AGE GROUNDWATERS

3.1 Abstract

In Chapter 3 I present a global compilation of the isotopic composition of groundwater recharge from the late-Holocene (δ18Olate-Holocene) and the last ice age (δ18Oice age). Changes to meteoric water δ18O values from the last ice age to the late-Holocene are described herein as Δδ18Oice age

(where, Δδ18Oice age = δ18Olast ice age − δ18Olate-Holocene). Groundwater Δδ18Oice age values range from −3.6

‰ (i.e., δ18Olast ice age < δ18Olate-Holocene) to +2.0 ‰ (i.e., δ18Olast ice age > δ18Olate-Holocene). More than 90% of study aquifers have δ18Olast ice age < δ18Olate-Holocene, in spite of higher-than-modern seawater δ18O values during the last ice age. The few study aquifers where δ18Olast ice age > δ18Olate-Holocene are found exclusively within 300 km of coasts and generally confined to the subtropics. Groundwater Δδ18Oice age values are closer to zero (average groundwater Δδ18Oice age of −0.6 ‰) than Greenland and

Antarctic ice cores (average polar ice core Δδ18Oice age of −5.5 ‰). Δδ18Oice age values from four different isotope-enabled general circulation models are able to reproduce some but not all ice-age- to-late-Holocene δ18O shifts (pre-industrial and last glacial maximum climate simulations). Each of the four models do not reproduce the negative Δδ18Oice age values over tropical Africa and South

America, potentially reflecting imperfect parameterization of convective precipitation. The four isotope enabled general circulation models have a similar sign of Δδ18Oice age for about half of Earth’s areas, generally showing multi-model agreement upon positive Δδ18Oice age over the tropical oceans, and negative Δδ18Oice age over extra-tropical land areas. However, simulated Δδ18Oice age values do not reproduce the observed negative Δδ18Oice age values over Africa and Brazil, potentially due to different or incomplete model parameterization of convective rainfall during the last glacial maximum and present day.

142 3.2 Introduction

The isotopic composition of groundwater recharge from the last ice age has been used to improve the scientific community’s understanding of water availability under climates of the past for more than a half century (e.g., Thatcher et al., 1961; Tamers, 1967; Gat et al., 1969; Salati et al., 1974).

The use of paleo-groundwaters to reconstruct past climates has both advantages and disadvantages in comparison to other types of paleoclimate records. For example, shifts in climate on time scales of

100 to 103 years that can be distinguished in lacustrine (e.g., von Grafenstein et al., 1999) and speleothem (e.g., Denniston et al., 2007) records often cannot be identified in groundwater records because of hydrodynamic dispersion during multi-millennia groundwater residence times. However, groundwaters have an advantage over other paleoclimate records because paleo-groundwaters are a direct sample of paleo-precipitation and because paleo-groundwaters are found in many regions. As such, the chemistry of groundwaters provide insights into hydrologic and climate changes since the last ice age such as changes to air mass trajectories (Rozanski et al., 1985; Legrande and Schmidt,

2009) and land surface temperatures (Stute et al., 1989; 1995a; 1995b; Clark et al., 1998; Aeschbach-

Hertig et al., 2002). Groundwater paleoclimate archives are not subject to complicating effects that must be accounted for in other paleoclimate archives before interpreting the isotopic composition of paleowaters. For example, extracting quantitative paleoclimate information from the isotopic compositions of lake sediments, tree rings or speleothems is challenging due to several factors, including: (1) uncertainty in paleo-temperatures, which directly impact water-proxy isotopic fractionation factors (e.g., lake sediment calcite, diatom silica and sediment cellulose; tree ring cellulose; speleothem calcite), (2) uncertainty in the magnitude of evaporation-induced 18O- enrichment of surface waters prior to preservation in proxy records (e.g., lake sediment calcite, diatom silica and sediment cellulose; Leng and Marshall, 2004), (3) uncertainty and variability in the timing and seasonality of mineral precipitation and bioform growth, which impacts both the isotopic composition of the water being preserved in the proxy record, and the water-proxy fractionation

143 factor due to seasonality in temperatures (e.g., lake sediment calcite, diatom silica and sediment cellulose), (4) uncertainty in the impact of diagenesis (e.g., travertine; O’Brien et al., 2006) and (5) uncertainty in paleo-atmospheric humidity (e.g., tree ring cellulose; Roden and Ehleringer, 1999), each of which impacts the isotopic offset between paleo-waters and the preserved proxy record. In contrast, the relationship between paleo-precipitation groundwater isotope compositions is more direct and reliably identifiable. A recent study of 70 paired precipitation-groundwater isotopic datasets found systematic relationships between the isotopic composition of annual precipitation and groundwater. Differences between modern annual precipitation and groundwater isotopic compositions are related to the ratio of recharge as a proportion of precipitation (i.e., recharge/precipitation: Jasechko et al., 2014).

Ice-age-to-late-Holocene changes to isotopic compositions measured in proxy records have been ascribed in earlier works to several partially overlapping influences. Perhaps the two best- constrained and global-in-scale changes from the last ice age to the late-Holocene are (i) the change to atmospheric and surface ocean temperatures (MARGO Members, 2009; Annan and Hargreaves,

2013), and (ii) the change to the isotopic composition of the ocean (Emiliani, 1955; Dansgaard and

Tauber, 1969; Schrag et al., 1996). Atmospheric temperatures have increased by a global average of about 4°C since the last glacial maximum, as constrained by compilations of proxy-based reconstructions (Shakun and Carlson, 2010, Annan and Hargreaves, 2013). Climate warming over the past 20,000 years is thought to have been greatest in the extra-tropics (average increase of 6.3°C for latitudes of greater than 25°; Annan and Hargreaves, 2013) and more subdued in the tropics (average increase of 1.7°C for latitudes of less than 25°; Annan and Hargreaves, 2013; Figure 3-1), although some terrestrial noble gas based temperature reconstructions suggest much greater tropical cooling

(e.g., eastern Brazil 5.4°C cooler than today during the last glacial maximum; Stute et al., 1995b).

144

Figure 3-1. The change in surface air temperatures from the last glacial maximum to the preindustrial era (gridded data from Annan and Hargreaves, 2013). (a) Percentile ranges of temperature changes since the last glacial maximum for 10 degree latitudinal bands. Blue shading mark 25th-75th percentile ranges and the thin horizontal lines mark 10th-90th percentile ranges. The grey band shows the globally-averaged estimate of temperature change since the last glacial maximum of −4.0±0.8 °C

(Annan and Hargreaves, 2013). (b) Gridded surface air temperature anomaly from the last glacial maximum to the preindustrial era (Annan and Hargreaves, 2013).

The δ18O value of the last glacial period ocean was 1.0±0.1 ‰ higher than the modern ocean, as constrained by paleo-ocean water samples collected from pore waters trapped within sea floor sediments (Schrag et al., 2002; where δ18O = (18O/16Osample) / (18O/16Ostandard – 1)×1000). In addition to differences in temperatures and ocean water isotope compositions, a variety of additional explanations for observed changes to δ18O values found in paleoclimate records have been proposed, including variations in hurricane and storm intensity (e.g., Plummer et al., 1993), changes to large- scale atmospheric circulation patterns (e.g., Rozanski et al., 1985; Weyhenmeyer et al., 2000;

McDermott et al., 2001; Asmerom et al., 2010), shifts in monsoon strength (e.g., Denniston et al.,

2000; 2013; Lachniet et al., 2004; Liu et al., 2007; Pausata et al., 2011), fluctuations in the seasonality of precipitation (e.g., Cruz et al., 2005), and modifications to El Niño-Southern Oscillation patterns

(e.g., Vuille and Werner, 2005). 145 In this study I present a global compilation of proxy isotope records of groundwaters (n =

65), speleothems (n = 15) and ice cores (n = 11) spanning both the last ice age and the late-

Holocene. Global compilations already exist for speleothems (Shah et al., 2013) and polar ice cores

(Pedro et al., 2011; Stenni et al., 2011; Clark et al., 2012); this compilation is the first global compilation of isotopic data for groundwaters from the last ice age, building upon existing reviews of paleowaters across Europe (Edmunds and Milne, 2001; Jiráková et al., 2011) and Africa (Edmunds,

2009).

The objective of this study is to develop and analyze a global database of ice age groundwater chemistry and constrain the processes and mechanisms controlling meteoric water isotopic changes since the last ice age. This new compilation provides a global scale perspective that can be used to quantitatively interpret the magnitudes of δ18O and δ2H anomalies observed within various Quaternary climate records and to validate outputs from isotope-enabled general circulation models.

146 Table 3-1. Modern and ice age physical and isotopic data for the oceans and the cryosphere

Ice: Ice: Ice: Laurentide and Present day Ocean Antarctica Greenland Fennoscandinavian Volume 13.7 × 109 km3 d 28 × 106 km3 d 3.1 × 106 km3 0 km3 Depth 3800 m - - - f −30 to δ18O value 0 ‰ e −20 to −60 ‰ - −45 ‰

Last glacial Ice: Ice: Ice: Laurentide and Ocean maximum Antarctica Greenland Fennoscandinavian Volume a 13.2 × 109 km3 d 38 × 106 km3 d 4.4 × 106 km3 g 20 to 60 × 106 km3 Depth b 3665 to 3680 m - - - f −30 to δ18O value c +1.0±0.1 ‰ e −20 to −60 ‰ h −22 to −25 ‰ −45 ‰ Sea level change b change of −120 d −19.2 m d −3.1 m g −40 to −115 m * m to −135 m

a Lambeck et al., 2000 b Clark and Mix, 2002 c Schrag et al., 2002 d Huybrechts., 2002 (maximum values shown) e Range of Antarctic ice cores: Byrd Glacier, (Blunier and Brook, 2001) Dome Fuji, (Kawamura et al., 2007) Dronning Maud, (EPICA Community, 2006) Law Dome, (Pedro et al., 2011) Siple Dome (Pedro et al., 2011) and TALD Ice (Buiron et al., 2011)

f Range from NGRIP1 core g Range of model predictions from Beghin et al., 2013 h Subglacial recharge from the Laurentide and Fennoscandinavian ice sheets (Karro et al., 2004; Grasby and Chen, 2005; Ferguson et al., 2007; Stotler et al., 2010) * Presented as relative to the modern ocean level

147 3.3 Dataset and Methods

I compiled 18O/16O ratios, 2H/1H ratios, 13C/12C ratios, 3H activities, 14C activities for 1640 groundwater samples collected from 65 aquifers as reported in 68 publications. Each compiled aquifer dataset has between 4 and 182 groundwater samples (average of 27) that had previously been analyzed for both stable oxygen and hydrogen isotopic compositions and for radioactive carbon activities (14C). 14C data were required to ensure that compiled samples identifiably recharged during the last ice age. A mean 14C-based groundwater age (t, the time elapsed since recharge) was calculated for each sample by accounting for the radioactive decay of 14C and for the dissolution of 14C-dead aquifer materials (Clark and Fritz, 1997):

퐴 푡 = −8267 × 푙푛 ( ) Equation 3.1 푞×퐴표 where t is the time elapsed since the groundwater sample recharged (i.e., groundwater age), A is the measured 14C activity in a groundwater sample, Ao is the initial 14C activity (100 pmC) and q is a correction factor applied to account for the dissolution of aquifer material with zero 14C (i.e., 14C- dead). In cases where δ13C data was available q was calculated as:

13 13 훿 퐶푚푒푎푠푢푟푒푑 − 훿 퐶푎푞푢𝑖푓푒푟 푞 = 13 13 Equation 3.2 훿 퐶푟푒푐ℎ푎푟푔푒 − 훿 퐶푎푞푢𝑖푓푒푟

where δ13Cmeasured, δ13Caquifer and δ13Crecharge represent the carbon isotope composition of a groundwater sample, the aquifer and recharging groundwater. δ13Caquifer was set to 1.1±1.6 ‰ PDB as determined by the 25th to 75th percentile range of δ13C values for 16359 rock and sediment samples compiled and presented by Veizer et al. (1999). δ13Crecharge was set to −12.8±3.1 ‰ PDB as determined by the 25th to

75th percentiles of compiled δ13C values for 261 groundwater samples having a 14C activity of greater than 90 p.m.C. (i.e., recently recharged water bearing near-atmospheric radioactive activities;

Burchuladze et al., 1989). q was set to 0.86±0.14 in cases where δ13Cmeasured data were unavailable, as

148 determined by the most common δ13C-based q values (q = 0.86±0.14 represents the average and one standard deviation of δ13C-based q values).

Equivalent calendar year ages were estimated from 14C-ages by applying a polynomial fit of compiled 14C-to-calender age corrections (Fairbanks et al., 2005). Samples were then divided into two age categories: (i) the late-Holocene (14C-based age of less than 5,000 calendar years before present, or a 3H activity of greater than 5 T.U.), or (ii) the last ice age (14C-based age of 19,500 to ~50,000 calendar years before present (Clark et al., 2009) and samples with 14C activities below analytical detection). An upper ice age limit of ~50,000 years before present was set because of limitations with

14C age calculations, even though the most recent ice age extends to ~110,000 years before present

(Lisiecki and Raymo, 2005).

Groundwater δ18O and δ2H values for the late-Holocene and the last ice age were analyzed on an aquifer-by-aquifer basis. Only aquifers with a minimum of two samples dated to both the late-

Holocene and the last glacial time periods were included in this analysis. Comparisons of isotopic data for the last ice age and the late-Holocene were made by subtracting median δ18O and δ2H values from each age group, with errors calculated by maximizing the 25th to 75th percentile distributions for the two data groups (i.e., late-Holocene and last glacial period age groups). Samples were omitted from our analysis if they exhibited an evaporative signature (δ2H – 8×δ18O of less than 0), contained a mixture of modern and ice age groundwater (3H activity of greater than 1 tritium unit and a 14C-age of more than 19,500 calendar years before present), were suspected to have mixed with intruding seawater (e.g., Bouchaou et al., 2008; Morrissey et al., 2010) or were presumed to have been recharged by subglacial meltwaters from the Fennoscandinavian (e.g., Karro et al., 2004) or the

Laurentide (e.g., Grasby and Chen, 2005; Ferguson et al., 2007; Stotler et al., 2010) ice sheets (review by McIntosh et al., 2012).

149 A correction to speleothem δ18O values was applied because of the different H2O-CaCO3 isotopic fractionation factor for the last ice age and the late-Holocene imparted by the different atmospheric temperatures during each time period (Shakun and Carlson, 2010). Temperature-based

H2O-CaCO3 fractionation factors were obtained from O’Neil et al. (1969) with temperatures calculated under the assumption that atmospheric temperatures are indicative of temperatures in the shallow subsurface. Temperatures for the late-Holocene were assumed to be equivalent to modern mean annual near surface temperatures (New et al., 2002), potentially introducing <1°C of error because of temperature change throughout the last 5,000 years (Marcott et al., 2013). Adding 1°C of added uncertainty into late-Holocene temperature equates to an added ±0.4 ‰ (δ18O) of uncertainty in the temperature-corrected difference between ice age and late-Holocene δ18O values (O’Neil et al.,

1969). Last glacial period temperatures were calculated by applying the temperature offset of the last glacial maximum (Figure 3-1; Annan and Hargreaves, 2013) to gridded values of modern mean annual air temperatures (New et al., 2002).

With the help of F. Pausata, M. Werner, C. Risi, K. Yoshimura I assembled modelled precipitation Δδ18Oice age values from four isotope-enabled general circulation models: (i) CCSM3

(e.g., Pausata et al., 2011), (ii) ECHAM (e.g., Hoffman et al., 1998; Werner et al., 2011), (iii) IsoGSM

(e.g., Yoshimura et al., 2003) and (iv) LMDZ4 (e.g., Risi et al., 2010a). The models span a range of spatio-temporal resolutions and isotopic/atmospheric parameterizations that are explained in detail in the above references. Simulations run for the last glacial maximum and pre-industrial time periods were assembled to analyze global grids of the amount-weighted isotopic composition of precipitation for each time period. General circulation model outputs were compared to ice age groundwater data by extracting model estimates of the annual isotopic composition of precipitation at the locations of each of the 65 aquifers analyzed in this study. We acknowledge that the general circulation models explicitly analyze the last glacial maximum and the pre-industrial climate scenarios, whereas the

150 groundwater aquifers integrate hydroclimatology over longer (103 year) time scales that will damp abrupt δ18O changes because of storage and mixing.

3.4 Results

Groundwater, speleothem and ice core data sources, locations and isotopic changes since the last glacial maximum are presented in Figure 3-2 and in Tables 3-2 and Table 3-4. Fossil groundwater

δ18O records are unevenly distributed amongst Europe (n = 13), Africa (n = 19), Asia (n = 13),

Oceania and the Malay Archipelago (n = 4), North America (n = 13) and South America (n = 3;

Table 3-2), with 30% of records located in the tropics and 70% of records located in the extra-tropics

(tropics defined as spanning 0° to 25° latitude). The magnitude of change in meteoric δ18O from the last ice age is described herein as Δδ18Oice age (where Δδ18Oice age = δ18Olast ice age − δ18Olate-Holocene).

151

Figure 3-2. Meteoric water δ18O changes from the last ice age (19,500 to ~50,000 years ago) to the late-Holocene (within past ~5,000 years): Δδ18Oice age = δ18Olast ice age − δ18Olate-Holocene.

152 Table 3-2. Groundwater datasets compiled in this study

Country Aquifer Citation(s) Algeria Great Oriental Erg: CI Edmunds et al., 2003 Botswana Kalahari: Ntane Kulongoski et al., 2004 Botswana Lokalane-Nakojane Rahube, 2003 Burkina Faso Taoudeni basin Huneau et al., 2011 Chad Chad aquifer Edmunds, 2009 Egypt Nubian aquifer Patterson et al., 2005 Mali Mali aquifer Edmunds, 2009 Morocco N. Morocco aquifer Winckel et al., 2002 Morocco Tadla basin Bouchaou et al., 2009 Morocco Nappe des sables Castany et al., 1974 Namibia Omatako basin Külls, 2000 Niger Djardo-Bilma Dodo and Zuppi, 1997; 1999 Niger Irhazer: CI Andrews et al., 1994 Niger Illumeden: CT Beyerle et al. 2003 Nigeria Chad basin Maduabuchi et al., 2006 Senegal Senegalese CT Castany et al., 1974 South Africa Uitenhage aquifer Heaton et al., 1986 Tunisia Kairouan Plain Derwich et al., 2012 Zimbabwe Zimbabwe aquifer Larsen et al., 2002 Australia Canning basin Harrington et al., 2011 Australia Ngalia and Amadeus Leaney and Allison, 1986 Australia Murray aquifer Cresswell et al., 1999 Bangladesh Bengal basin Aggarwal et al., 2000 China Songnen plain Chen et al. 2011 China Hexi Corridor: east Gates et al., 2008 China North China Plain Kreuzer et al. 2009 China Yuncheng basin Currell et al., 2010 India Cuddalore sandstone Sukhija et al., 1998 India Tiruvadanai aquifer Kumar et al., 2009 Indonesia Jakarta basin Geyh and Sofner, 1989 Israel Israel coastal aquifer Yechieli et al., 2008 Israel Dead Sea rift valley Gat and Galai, 1982 Robinson and Gunatilaka, 1991; Kuwait Kuwait aquifer Al-Ruwaih et al., 2004 Oman Batinah coastal plain Weyhenmeyer et al., 2002 Oman Najd aquifer Al-Mashaikhi et al., 2012 Syria Aleppo basin Stadler et al., 2012 Belgium Ledo-Paniselian Blaser et al., 2010 Czech Rep. Sokolov aquifer Noseck et al., 2009 France Bathonian coast Barbecot et al., 2000 France Lorraine sandstone Celle-Jeanton et al., 2009 France Aquitaine basin Jiráková et al., 2009 Hungary Great Hungarian Plain Stute and Deak, 1989 153 Country Aquifer Citation(s) Hungary Pannonian basin Varsanyi et al., 2011 Poland Mazovian basin Zuber et al., 2000 Poland S. Poland carbonates Samborska et al., 2012 Poland Malm limestone Zuber et al. 2004 Portugal Sado basin Fernandes and Carreira, 2008 United Kingdom Lincolnshire limestone Darling et al., 1997 Darling and Bath, 1988; Dennis et al., 1997; United Kingdom Chalk aquifer Elliot et al., 1999; Gooddy et al., 2006 U.S.A. Columbia Flood Bslts. Douglas et al., 2007 U.S.A. Black Hills: Pahasapa Back et al., 1983 U.S.A. Idaho Batholith Schlegel et al., 2009 U.S.A. Cambrian-Ordovician Siegel, 1991 U.S.A. High Plains: North Gosselin et al., 2001 U.S.A. Mahomet aquifer Hackley et al., 2010 U.S.A. Aquia aquifer Aeschbach-Hertig et al. 2002 U.S.A. High Plains: Central Clark et al. 1998 U.S.A. San Juan Basin Stute et al., 1995a U.S.A. Middle Rio Grande Plummer et al., 2011 U.S.A. Los Angeles Basin Swarzenski et al., 2013 U.S.A. Floridan aquifer Clark et al., 1997 U.S.A. Floridan surficial aqfr. Morrissey et al., 2010 Brazil Portigar basin: Acu Salati et al., 1974 Brazil Botacatu: central Gouvea da Silva, 1983 Brazil Botucatu: south Roboucas and Santiago, 1989

154 Table 3-2 (continued). Observed Δδ18Oice age values in groundwaters

Δδ18Oice age Country Aquifer Lon. Lat. (‰ V−SMOW) Africa Algeria Great Oriental Erg: CI 5.9 32.4 −0.5 (−1.6 to −0.3) Botswana Kalahari: Ntane 25.2 −24.0 −0.5 (−0.7 to −0.3) Botswana Lokalane-Nakojane 22.0 −22.3 −1.1 (−1.2 to −0.9) Burkina Faso Taoudeni basin −4.7 12.8 −0.5 (−1.0 to −0.3) Chad Chad aquifer 18.3 11.2 −0.9 (−2.4 to −0.5) Egypt Nubian aquifer 28.9 25.7 −1.6 (−3.6 to −0.3) Mali Mali aquifer −7.2 15.2 −0.5 (−1.3 to +0.1) Morocco N. Morocco aquifer −4.9 34.0 −0.6 (−2.2 to 1.4) Morocco Tadla basin −6.7 32.6 −0.9 (−1.5 to −0.1) Morocco Nappe des sables −14.5 15.4 +0.2 (−0.5 to +0.8) Namibia Omatako basin 17.9 −20.1 −0.9 (−1.3 to −0.1) Niger Djardo-Bilma 12.9 18.9 +2.0 (−0.4 to +2.5) Niger Irhazer: CI 7.5 17.3 −0.9 (−2.3 to +0.2) Niger Illumeden: CT 2.8 13.6 −3.0 (−3.7 to −2.1) Nigeria Chad basin 13.0 12.0 −0.3 (−2.2 to +0.5) Senegal Senegalese CT −16.4 15.2 +0.3 (−0.6 to +1.0) South Africa Uitenhage aquifer 25.5 −33.7 −0.5 (−1.0 to −0.4) Tunisia Kairouan Plain 10.0 35.5 +0.2 (−0.9 to +0.4) Zimbabwe Zimbabwe aquifer 28.1 −19.5 −0.9 (−1.3 to −0.5) Asia and western Pacific Australia Canning basin 125.1 −17.5 −1.0 (−2.3 to +0.9) Australia Ngalia and Amadeus 131.9 −23.4 −0.3 (−0.6 to +1.1) Australia Murray aquifer 140.2 −34.2 −0.3 (−1.1 to +0.1) Bangladesh Bengal basin 90.0 23.6 +1.6 (+0.9 to +2.3) China Songnen plain 124.5 45.9 −0.2 (−0.8 to 0.3) China Hexi Corridor: east 102.1 38.7 −1.4 (−2.4 to −0.3) China North China Plain 114.9 38.0 −2.3 (−2.6 to −1.7) China Yuncheng basin 110.6 35.0 −1.1 (−2.1 to −0.1) India Cuddalore sandstone 79.5 11.4 +0.8 (0.1 to +1.5) India Tiruvadanai aquifer 78.7 10.0 −0.9 (−1.5 to −0.5) Indonesia Jakarta basin 106.8 −6.3 +0.1 (−0.1 to +0.5) Israel Israel coastal aquifer 34.8 32.0 +0.2 (−0.1 to +0.5) Israel Dead Sea rift valley 35.2 30.7 −1.4 (−2.0 to −0.8) Kuwait Kuwait aquifer 47.7 29.8 −1.6 (−2.1 to −1.3) Oman Batinah coastal plain 57.7 23.6 +1.1 (−0.2 to +2.0) Oman Najd aquifer 53.9 18.1 −0.6 (−3.4 to +2.3) Syria Aleppo basin 37.3 35.9 −1.4 (−2.5 to −0.1) Europe Belgium Ledo-Paniselian 3.5 51.2 −0.5 (−1.0 to +0.1) Czech Rep. Sokolov aquifer 12.7 50.2 −0.8 (−0.8 to −0.3) France Bathonian coast −0.2 49.2 −0.4 (−0.6 to −0.3) France Lorraine sandstone 6.6 48.9 −1.0 (−1.5 to −0.8)

155 Δδ18Oice age Country Aquifer Lon. Lat. (‰ V−SMOW) France Aquitaine basin −0.4 45.9 −1.5 (−2.5 to −0.1) Hungary Great Hungarian Plain 20.8 47.6 −1.8 (−2.3 to −1.3) Hungary Pannonian basin 20.1 46.3 −3.6 (−4.1 to −2.7) Poland Mazovian basin 21.0 52.2 0.0 (−0.4 to +0.2) Poland S. Poland carbonates 19.2 50.6 −2.0 (−2.5 to −1.1) Poland Malm limestone 19.8 50.0 −1.0 (−1.9 to +0.4) Portugal Sado basin −8.5 38.3 +0.1 (−0.3 to +0.2) United Kingdom Lincolnshire limestone −0.4 52.7 −0.4 (−0.4 to −0.2) United Kingdom Chalk aquifer −1.4 51.5 −0.4 (−0.4 to −0.3) The Americas U.S.A. Columbia Flood Basalts −119.0 46.6 −2.8 (−3.6 to −1.2) U.S.A. Black Hills: Pahasapa −103.5 44.3 −0.4 (−0.9 to +1.6) U.S.A. Idaho Batholith −116.1 43.7 −0.7 (−1.1 to +0.1) U.S.A. Cambrian-Ordovician −93.2 42.9 0.0 (−0.6 to +0.5) U.S.A. High Plains: North −101.3 40.9 +0.3 (−1.3 to +2.2) U.S.A. Mahomet aquifer −88.8 39.9 −0.2 (−0.5 to +0.1) U.S.A. Aquia aquifer −76.6 38.7 0.0 (−0.4 to +0.3) U.S.A. High Plains: Central −101.0 37.5 −2.6 (−4.3 to −1.4) U.S.A. San Juan Basin −107.8 36.5 −3.1 (−3.4 to −2.5) U.S.A. Middle Rio Grande −106.4 35.1 −0.6 (−2.8 to +0.8) U.S.A. Los Angeles Basin −118.2 33.8 −1.4 (−2.0 to −0.8) U.S.A. Floridan aquifer −82.1 32.0 +1.0 (+0.6 to +1.3) U.S.A. Floridan surficial aqfr. −81.0 26.7 +1.8 (+0.6 to +2.1) Brazil Portigar basin: Acu −38.0 −5.6 −2.2 (−3.3 to −1.9) Brazil Botacatu: central −48.7 −22.3 0.0 (−1.4 to +1.8) Brazil Botucatu: south −52.9 −28.1 −1.5 (−1.9 to −1.0)

156 Table 3-3. Observed ranges of groundwater δ18Oice age and δ18Olate-Holocene (Shown as δ18OHolo.) values

18 18 Aquifer δ OHolo. HighHolo. LowHolo. δ Oice age Highice age Lowice age Great Oriental Erg −7.9 −7.0 −8.0 −8.4 −8.3 −8.6 Kalahari: Ntane −5.6 −5.2 −5.8 −6.1 −5.9 −6.1 Lokalane-Nakojane −6.1 −6.1 −6.3 −7.2 −7.2 −7.3 Taoudeni basin −5.2 −5.0 −5.4 −5.8 −5.7 −5.9 Chad aquifer −4.3 −3.5 −4.6 −5.3 −5.1 −5.9 Nubian aquifer −8.9 −7.0 −9.9 −10.5 −10.2 −10.6 Mali aquifer −6.0 −5.4 −6.5 −6.5 −6.4 −6.7 N. Morocco aquifer −6.4 −6.0 −6.6 −7.0 −5.2 −8.2 Tadla basin −5.6 −5.1 −6.4 −6.6 −6.5 −6.6 Nappe des sables −6.1 −5.6 −6.6 −5.9 −5.8 −6.1 Omatako basin −8.4 −8.1 −9.1 −9.3 −9.2 −9.3 Djardo-Bilma −10.1 −7.9 −10.4 −8.1 −7.9 −8.2 Irhazer: CI −6.4 −5.5 −7.0 −7.3 −6.8 −7.7 Illumeden: CT −4.4 −3.8 −5.2 −7.4 −7.3 −7.6 Chad basin −5.9 −4.4 −6.4 −6.2 −5.9 −6.6 Senegalese CT −6.2 −5.6 −6.5 −5.9 −5.5 −6.2 Uitenhage aquifer −4.9 −4.5 −5.0 −5.4 −5.4 −5.5 Kairouan Plain −5.8 −5.4 −5.9 −5.6 −5.5 −6.2 Zimbabwe aquifer −6.0 −5.7 −6.3 −6.9 −6.8 −7.0 Canning basin −6.6 −5.5 −8.4 −7.6 −7.5 −7.8 Ngalia/Amadeus −4.3 −4.0 −4.5 −4.5 −4.4 −5.1 Murray aquifer −7.0 −6.8 −8.4 −7.3 −7.2 −7.4 Bengal basin −4.7 −4.2 −5.3 −3.1 −3.0 −3.3 Songnen plain −10.0 −9.5 −10.3 −10.2 −10.0 −10.3 Hexi Corridor: east −9.1 −8.4 −9.8 −10.5 −10.1 −10.8 North China Plain −8.6 −8.3 −8.9 −10.8 −10.6 −10.9 Yuncheng basin −9.2 −8.4 −9.4 −10.3 −9.5 −10.5 Cuddalore sst. −5.6 −5.5 −5.8 −4.9 −4.4 −5.4 Tiruvadanai aquifer −4.1 −3.7 −4.4 −5.0 −4.9 −5.2 Jakarta basin −6.1 −5.7 −6.2 −6.0 −5.6 −6.2 Israel coastal −4.7 −4.5 −5.0 −4.5 −4.5 −4.6 Dead Sea rift valley −4.8 −4.8 −5.4 −6.2 −6.2 −6.8 Kuwait aquifer −2.9 −2.6 −3.0 −4.5 −4.4 −4.7 Batinah coast −2.7 −1.6 −3.4 −1.6 −1.4 −1.8 Najd aquifer −3.1 −0.7 −5.4 −3.6 −3.2 −4.1 Aleppo basin −6.0 −5.4 −6.9 −7.4 −6.9 −7.9 Ledo-Paniselian −6.5 −6.1 −6.9 −7.0 −6.8 −7.1

157 18 18 Aquifer δ OHolo. HighHolo. LowHolo. δ Oice age Highice age Lowice age Sokolov aquifer −9.0 −8.9 −9.1 −9.8 −9.4 −9.8 Bathonian coast −6.6 −6.5 −6.7 −7.0 −7.0 −7.1 Lorraine sandstone −8.9 −8.7 −9.0 −10.0 −9.8 −10.2 Aquitaine basin −5.8 −5.6 −6.2 −7.3 −6.3 −8.1 Grt Hungarian Plain −9.5 −9.3 −9.6 −11.3 −10.9 −11.6 Pannonian basin −9.4 −9.1 −9.6 −13.0 −12.3 −13.2 Mazovian basin −10.2 −10.0 −10.2 −10.1 −10.0 −10.3 S. Poland −9.9 −9.6 −10.1 −11.8 −11.2 −12.1 Malm limestone −10.1 −10.0 −11.1 −11.1 −10.7 −11.9 Sado basin −4.8 −4.4 −4.8 −4.7 −4.7 −4.7 Lincolnshire limest. −7.8 −7.8 −7.9 −8.2 −8.1 −8.2 Chalk aquifer −7.4 −7.4 −7.4 −7.8 −7.7 −7.8 Columbia Floods −15.3 −14.8 −16.4 −18.1 −17.6 −18.5 Black Hills −17.1 −16.7 −17.4 −17.5 −15.8 −17.6 Idaho Batholith −16.8 −16.5 −17.4 −17.5 −17.2 −17.6 Cambrian-Ordo. −8.8 −8.2 −9.2 −8.8 −8.7 −8.8 High Plains: North −9.9 −9.3 −10.8 −9.6 −8.5 −10.6 Mahomet aquifer −6.8 −6.7 −7.0 −7.0 −6.9 −7.2 Aquia aquifer −7.1 −7.0 −7.1 −7.1 −6.8 −7.4 High Plains: Cent. −9.5 −8.1 −10.5 −12.1 −11.9 −12.4 San Juan Basin −11.4 −11.2 −11.7 −14.5 −14.2 −14.6 Middle Rio Grande −11.8 −10.2 −12.9 −12.5 −12.1 −13.0 Los Angeles Basin −7.3 −7.2 −7.3 −8.6 −8.1 −9.2 Floridan aquifer −4.7 −4.6 −4.7 −3.7 −3.4 −4.0 Floridan surf. aqfr. −3.4 −2.4 −3.5 −1.6 −1.4 −1.8 Portigar basin: Acu −2.3 −1.4 −2.5 −4.5 −4.4 −4.6 Botacatu: central −8.6 −7.4 −9.3 −8.6 −7.5 −8.8 Botucatu: south −6.2 −5.9 −6.4 −7.6 −7.4 −7.8 * High and Low refer to 25th and 75th percentile ranges of modern (mod. i.e., late-Holocene) and glacial (i.e., last ice age) data groups.

158 Table 3-4. Speleothem δ18O from the last ice age to the late-Holocene

Δδ18Oice age Cave Country Reference Lon. Lat. (‰ Corr.x V−SMOW) Speleothems +1.9 (+1.8 Gunung Buda Borneo Partin et al., 2007 114.8 4.0 0.9±0.2 to +2.4) Cruz et al., 2005; −0.7 (−1.1 Botuverá Cave Brazil −49.2 −27.2 0.9±0.2 Wang et al., 2007 to −0.3) Dykoski et al., 2005; +2.5 (+2.1 Dongge t China 108.1 25.3 0.9±0.2 Yuan et al., 2004 to +3.2) +1.6 (+1.4 Hulu * China Wang et al., 2001 119.2 32.5 1.0±0.2 to +1.9) +1.0 (−0.8 Jiuxian t China Cai et al., 2010 109.1 33.6 1.0±0.2 to +3.7) +2.9 (+2.3 Yaman Cave t China Yang et al., 2010 107.9 24.5 0.9±0.2 to +3.3) Bar-Matthews et al., +2.3 (+2.1 Soreq Israel 35.0 31.5 1.0±0.1 2003 to +2.5) Bar-Matthews et al., +2.1 (+1.8 Peqin* Israel 35.2 32.6 0.9±0.2 2003 to +2.3) +1.6 (+1.4 Jerusalem W. Israel Frumkin et al., 1999 35.2 31.7 1.0±0.2 to +2.3) NW South New +0.4 (+0.2 Williams et al., 2010 172.0 −42.0 1.0±0.3 Island Zealand to +0.6) South +1.2 (+0.7 Cold Air Cave Holmgren et al., 2003 29.1 −24.0 1.0±0.1 Africa to +1.7) −4.7 (−4.8 Sofular Turkey Fleitmann et al., 2009 31.9 41.4 1.0±0.2 to −4.5) −2.0 (−2.8 Fort Stanton* U.S.A. Asmerom et al., 2010 −105.3 33.3 1.0±0.2 to −1.2) Cave of the −2.4 (−2.6 U.S.A. Wagner et al., 2010 −110.8 31.8 1.0±0.2 bells* to −2.4) +2.3 (+1.7 Moomi* Yemen Shakun et al., 2007 54.0 12.5 0.9±0.2 to +2.9) * early Holocene value used (i.e., shift likely larger than shown) t values from 15.0 ka used x Calcite-water fractionation correction subtracted from raw observed Δδ18Oice age to correct for the

4.0±0.8 °C colder climate (Annan and Hargreaves, 2013) at the last glacial stage (from O’Neil et al.,

1969; modern temperatures from New et al., 2002; Δδ18Oice age values in preceding column are shown in raw (i.e, uncorrected) form.

159 Table 3-5. Ice core δ18O from the last ice age to the late-Holocene

Δδ18Oice age Ice core Country Reference Lon. Lat. (‰ V−SMOW) Ice cores Sajama Bolivia Thompson et al., 1998 −63.9 −18.1 −4.9 (−3.7 to −5.6) Huascaran Peru Thompson et al., 1995 −77.6 −9.1 −6.5 (−5.9 to −7.4) Qinghai- Tibet Thompson et al., 1997 81.5 35.3 −0.9 (+2.0 to −3.2) Tibetan TALD Ice Antarctica Buiron et al., 2011 159.2 −72.8 −3.9 (−3.3 to −4.3) Blunier and Brook, Byrd Glacier Antarctica −119.5 −80.0 −6.0 (−5.0 to −6.9) 2001 Dome Fuji Antarctica Kawamura et al. 2007 39.7 −77.3 −3.6 (−2.7 to −4.5) Dronning EPICA Community Antarctica 2.0 −75.0 −5.2 (−4.5 to −6.0) Maud Members, 2006 Law Dome Antarctica Pedro et al. 2011 112.8 −66.8 −6.9 (−6.0 to −7.5) Siple Dome Antarctica Pedro et al. 2011 −148.8 −81.7 −7.1 (−6.2 to −7.9) Renland ice Greenland Vinther et al., 2008 −27.0 71.0 −4.1 (−2.8 to −4.9) core NGRIP1 Greenland Vinther et al., 2006 −42.3 75.1 −7.3 (−5.5 to −8.7)

160 The magnitude of change in δ18O values from the last ice age (19,500 to 50,000 years before present) to the late-Holocene (<5,000 years before present; i.e., Δδ18Oice age) is shown in Figures 3-3 and 3-5. The Δδ18Oice age value of groundwater ranges from −3.6 ‰ (i.e., δ18Olast ice age < δ18Olate-

Holocene) to +2.0 (i.e., δ18Olast ice age > δ18Olate-Holocene), with more than 90 percent of aquifers having negative Δδ18Oice age values (Figure 3-4). No systematic latitudinal trend in Δδ18Oice age values can be observed for either the fossil groundwater or speleothem records (Figure 3-4), unlike temperature

(Figure 3-1). However, cases where δ18Oice age values exceed δ18Olate-Holocene values are constrained to coastal aquifers in the subtropics (e.g., Bangladesh: +1.6 ‰, less than 300 km from the coast; Florida:

+1.0 and +1.8 ‰, less than 100 km from the coast; southern India: +0.9 ‰, less than 100 km from the coast). In comparison, aquifers characterized by lower δ18Oice age values than δ18Olate-Holocene values are found in both coastal regions and farther inland. Aquifers located farthest from coastlines exhibit the lowest Δδ18Oice age values (e.g., Hungary: −3.6 ‰, ~500 km inland; New Mexico: −3.1 ‰, ~1000 km inland; Niger: −3.0 ‰, ~800 km inland). Greenland and Antarctic ice cores have consistently negative Δδ18Oice age values that are of a greater magnitude (average of −5.5 ‰, range from −3.6 ‰ to −7.3 ‰) than groundwater Δδ18Oice age values (average of −0.6 ‰, range from −3.6 ‰ to +2.0 ‰;

Figure 3-3).

161

Figure 3-3. The Δδ18Oice age value of

groundwaters, speleothems and ice

cores. Colored bars mark the 25th-

75th percentile ranges of late-

Holocene and last glacial stage

datasets for each aquifer (blue

shades, unique to geographic

regions), speleothem (red), and ice

core (light brown marks non-polar,

dark brown marks ice cores for

Antarctica and Greenland).

Speleothem data are corrected for

the different isotope effects during

precipitation due to different ice age

and modem temperatures. An early

Holocene δ18O range was used for

the Byrd, Dronning Maud, Law

Dome ice cores and Dongge, Fort

Stanton, Hulu, Jiuxian and Moomi

speleothems due to lacking late-

Holocene data in these records. See

Tables 3-2 through 3-5 for

descriptions.

162

Figure 3-4. (top pane) Latitudinal variations of Δδ18Oice age values of groundwater (circles, each circle is one aquifer), ice cores (squares) and caves (i.e., speleothems; triangles). Dashed lines mark 10° zonal means of terrestrial precipitation δ18O values predicted by four different general circulation models (CCSM, ECHAM, LMDZ and IsoGSM). (bottom pane) Histogram of observed Δδ18Oice age values for in speleothems, ice cores and groundwaters (n = 92 records, in total). Red bars mark records where δ18Oice age > δ18Olate-Holocene, blue bars mark records where δ18Oice age < δ18Olate-Holocene.

163 3.5 Discussion

3.5.1 Ice age groundwaters as a paleoclimate proxy

The isotopic composition of groundwater from the last ice age provides a valuable tracer of the isotopic composition of past meteoric waters. The meteoric nature of ice-age-to-late-Holocene

δ18O and δ2H shifts found in the compiled groundwater data demonstrates that paleo-groundwaters with minimal evaporative influence can be readily identified, making them valuable archives of paleoclimate information. Such identification of evaporative influence is often difficult to decouple for other records based solely upon δ18O or δ2H values (e.g., lake sediments), highlighting the value of groundwater archives for paleoclimate investigations. The ability to measure both δ18O and δ2H values of groundwater and ice core paleoclimate records is another primary advantage of these “dual isotope” (i.e., both δ18O and δ2H) paleoclimate records over “single isotope” paleoclimate records that can only provide either δ18O (e.g., speleothem, lake sediment carbonate, diatom or cellulose records) or δ2H (e.g., lake sediment leaf wax records) values, but not both. Because of the ability to examine deuterium excess values, ice age groundwater records may be better suited than speleothems for determining changes to moisture sources as recently evidenced by isotope enabled general circulation model reanalysis of ice core isotopic data (Lewis et al., 2013).

Where groundwater data may be advantageous over other records in its availability of “dual isotope” data, these records suffer in temporal resolution. Lake sediment, ice core and speleothem records can often be resolved at time scales of 100 to 103 years, whereas groundwater paleoclimate records can be resolved at time scales of >103 years because of uncertainties in corrected 14C-based groundwater ages and because of hydrodynamic dispersion that “smears” the groundwater isotopic record. The impact of hydrodynamic dispersion and long groundwater residence times may help to explain a portion of the discrepancies in the magnitude of δ18O shifts observed in lake sediment and groundwater δ18O records at similar locations, just as comparing the standard deviation of daily precipitation will differ from the standard deviation of monthly precipitation at the same location. 164 Each type of paleoclimate record has advantages and disadvantages, and all records are useful to advancing our understanding of the climate during the last ice age. Groundwater records of last glacial climate are globally-distributed and are able to be analyzed for “dual isotopes” to confirm the meteoric nature of the paleoclimate record as completed in this study. However, paleo- groundwater records of past climates have a poor temporal resolution (>103 years) that negates the detection of rapid and dramatic shifts in climate. Speleothem isotopic records of last glacial climate have high temporal resolution (100 to 102 years) but only have a single isotope available (δ18O in carbonate) and are not as common (n = 15) as groundwater records (n = 65). Ice core records of last glacial climate can be analyzed for both oxygen and hydrogen isotopic data and have a high temporal resolution, but are very uncommon on land masses other than Antarctica and Greenland. Lake sediment isotopic records of last glacial climate can have a high temporal resolution and are available for a multitude of globally-distributed locations, however, lake sediment records have large uncertainties in reconstructing past changes to meteoric δ18O because of the need to (i) quantify the temperature of the water that the paleoclimate archive (e.g., lake sediment diatom, cellulose and carbonate; speleothem carbonate) precipitated from in the past, and (ii) know the impact of evaporation upon isotopic composition of water in the past, both of which are highly difficult to reconstruct considering that most lake sediment archives are “single isotope” records (i.e., one of

δ18O or δ2H analyzed).

3.5.2 Isotope-enabled general circulation models

The δ18O value of annual precipitation from four isotope-enabled general circulation models analyzed for the pre-industrial era and the last glacial maximum scenarios are shown in Figures 3-5,

3-6, 3-7 and 3-8. Points in each of the figures mark the observed ice-age-to-modern differences in

δ18O observed in groundwater, speleothems and ice cores. Locations where three or four models agree on the sign of Δδ18Oice age are shown in Figures 3-9 and 3-10.

165 Generally, modelled Δδ18Oice age values are lowest over the Fennoscandanavian and

Laurentide ice sheets (less than 3 per mille), and highest over the tropical oceans. Positive modelled

Δδ18Oice age values occur near to coasts in the tropics and subtropics. Extra-tropical land surfaces generally have negative Δδ18Oice age values, whereas tropical land surface Δδ18Oice age values are more variable in their sign amongst the models (Figure 3-9 and 3-10). The disagreement amongst the four models on the sign of Δδ18Oice age values is potentially related to different parameterizations of convective rainfall along air mass trajectories, which is a leading control upon precipitation δ18O values in tropical regions (Risi et al., 2008; Risi et al., 2010b; Lee et al., 2009; 2012; Lekshmy et al.,

2014; Samuels-Crow et al., 2014).

Simulated Δδ18Oice age values reproduce the sign of observed Δδ18Oice age values across North

America and Europe (extratropics) better than over Africa and South America (Figure 3-9 and 3-10).

Simulated isotopic compositions of rain over tropical Africa and South America have both disagreement amongst different models on Δδ18Oice age values, and also Δδ18Oice age disagreement between simulated (generally positive) and observed (generally negative) values. The negative

Δδ18Oice age values across Africa are consistent with enhanced air mass distillation during transport due to higher-than-modern upstream rainout during the last ice age. Assuming that convection is a leading control upon the isotopic composition of tropical precipitation, the models perhaps overestimate the change in convection from the last glacial maximum to the present day. The stronger agreement between simulated and observed Δδ18Oice age values in the extratropics relative to the tropics suggests that models may simulate isotopic distillation via frontal advective hydroclimates better than via convective rainout. The poorer simulation of Δδ18Oice age values in the tropics than in the extratropics is consistent with other works that show that simulating the isotopic composition of convective rains is highly sensitive to model parameterization (Lee et al., 2009).

166

Figure 3-5. The modelled difference in the δ18O value of precipitation from the last glacial maximum to the pre-industrial time period (CCSM, pers. comm. F. Pausata): δ18Olast glacial maximum – δ18Olate-Holocene.

Circles show the observed Δδ18Oice age values (median) from groundwater, speleothem and ice core records compiled and analyzed in this study.

167

Figure 3-6. The modelled difference in the δ18O value of precipitation from the last glacial maximum to the pre-industrial time period (ECHAM, pers. comm. M. Werner): δ18Olast glacial maximum – δ18Olate-

Holocene. Circles show the observed Δδ18Oice age values (median) from groundwater, speleothem and ice core records compiled and analyzed in this study.

168

Figure 3-7. The modelled difference in the δ18O value of precipitation from the last glacial maximum to the pre-industrial time period (IsoGSM, pers. comm. K. Yoshimura): δ18Olast glacial maximum – δ18Olate-

Holocene. Circles show the observed Δδ18Oice age values (median) from groundwater, speleothem and ice core records compiled and analyzed in this study.

169

Figure 3-8. The modelled difference in the δ18O value of precipitation from the last glacial maximum to the pre-industrial time period (LMDZ, pers. comm. C. Risi): δ18Olast glacial maximum – δ18Olate-Holocene.

Circles show the observed Δδ18Oice age values (median) from groundwater, speleothem and ice core records compiled and analyzed in this study.

170

Figure 3-9. Locations where all four models agree on the sign of Δδ18Oice age values (i.e., positive or negative). Red colors mark regions where there is unanimous prediction of higher-than-modern δ18O values at the last ice age amongst the four models, whereas blues colors mark regions where there is unanimous prediction of lower-than-modern δ18O values at the last ice age amongst the four models.

The shades of red and blue are the multi-model average of modelled ice-age-to-late-Holocene changes in the δ18O value of meteoric water. White regions show areas where at least one of the four models predicts a different sign of Δδ18Oice age values (i.e., some models predict negative glacial to modern shifts, other models predict positive glacial to modern shifts).

171

Figure 3-10. Locations where at least three of four models agree on the sign of Δδ18Oice age values (i.e., positive or negative). Red colors mark regions where there is higher-than-modern δ18O values at the last ice age amongst the models, whereas blues colors mark regions where there is lower-than- modern δ18O values at the last ice age amongst the models. The shades of red and blue are the multi- model average of modelled ice-age-to-late-Holocene changes in the δ18O value of meteoric water.

White regions show areas where at least one of the four models predicts a different sign of Δδ18Oice age values (i.e., some models predict negative glacial to modern shifts, other models predict positive glacial to modern shifts).

172 Models predict similar Δδ18Oice age values in some regions (e.g., precipitation over the tropical oceans) and different Δδ18Oice age values in other regions (e.g., Africa). Figure 3-9 delineates locations where all four models agree on the sign of Δδ18Oice age values (i.e., all models are positive, or all models are negative), and shows that all models agree on the sign of Δδ18Oice age over ~40% of

Earth’s surface. All four models agree on the sign of Δδ18Oice age values for half of continental areas and for one-third ocean areas. For continental precipitation, 80% of locations where all models agree on the sign of Δδ18Oice age values have a unanimously negative simulated Δδ18Oice age value. Conversely,

75% of cases where all models agree on the sign of over-ocean precipitation Δδ18Oice age values have a unanimously positive simulated Δδ18Oice age value.

All four models have positive Δδ18Oice age values over the western and southern Pacific

Ocean, the tropical and mid-latitude Atlantic Ocean, the southeastern United States of America, northeast Brazil, western Africa, eastern China and southwestern Australia. All four models predict negative Δδ18Oice age values over the western United States of America, and northern and western

Canada, the southern margins of Argentina, northern Europe, the Norwegian Sea, throughout

Russia, Kazakhstan, Uzbekistan, and Turkmenistan, and over northern Mongolia and the Tibetan plateau.

The general circulation model ice-age-to-modern δ18O changes agree with the observations for some, but not all, locations. In general, simulated Δδ18Oice age values match observed Δδ18Oice age values more closely in the extratropics than in the tropics. For example, general observed Δδ18Oice age patterns over North America and Europe are reproduced by most general circulation models. In contract, observed Δδ18Oice age values over Africa and South America are not reproduced by most models. The modelled Δδ18Oice age reproduces some of the observed positive and observed negative

Δδ18Oice age values in groundwaters, speleothems and ice cores (Figure 3-9 and 3-10), highlighting the

173 much greater potential of these models for reconstruction of Δδ18Oice age values than temperature-

δ18O regressions, alone (e.g., Dansgaard, 1964).

3.5.3 Regional Δδ18Oice age values

3.5.3.1 Australia and Oceania

Australian records of Δδ18Oice age (n=3) range from -1.0‰ (Canning Basin) to -0.3‰ (Ngalia and Amadeus, Murray aquifers; Table 3-2). General circulation models predict positive Δδ18Oice age values across Australia; whereas the three observed groundwater records have negative Δδ18Oice age values. Observed Δδ18Oice age values are similar for all three Australian records despite different climates amongst the records that range from humid northern regions (Canning Basin) to more arid interior settings (Ngalia and Amadeus basins).

Spatial differences in climate change across the Australia continent are evidenced by higher- than-modern lake levels during the last glacial maximum in southeastern Australia (Galloway, 1965;

Williams, 2001), but lower-than-modern ice age lake levels in central Australia (Hope, 2005). The climate at the last ice age in parts of Australia was more arid (Nanson et al., 1992), dustier (Chen et al., 1993) and ~10°C cooler (Miller et al., 1997) than present day. Observations of higher-than- modern ice age lake levels are attributed to lower evaporative potential at the last glacial maximum

(Hope, 2005). Observed negative Δδ18Oice age values are consistent with cooler-than-modern condensation temperatures (i.e., enhanced air mass distillation) supported by 10°C cooler land surface temperatures (Miller et al., 1997). Alternatively, atmospheric models suggest that precipitation was more seasonal during the last glacial maximum than today due to cooler-than-modern sea surface temperatures (Hope, 2005) able to produce negative Δδ18Oice age values.

Oceania records of paleoclimate include groundwater data for the Jakarta basin (Geyh and

Sofner, 1989), and speleothem data in Borneo (Partin et al., 2007) and New Zealand (Williams et al.,

174 2010). Oceania isotopic records of speleothems (Partin et al., 2007) and groundwaters (Aggarwal et al., 2004) from Vietnam, Thailand, The Philippines and Borneo each have near-zero or positive

Δδ18Oice age values that have been attributed to ice-age-to-modern changes in monsoonal strength and atmospheric convection (Aggarwal et al., 2004; Partin et al., 2007). Simulated Δδ18Oice age values are generally positive or near-zero over Bangladesh, Vietnam, Thailand, The Philippines and Borneo

(Figure 3-9 and 3-10) consistent with the sign of observed Δδ18Oice age values. Simulated precipitation

δ18O values overlying Borneo are controlled by changes to precipitation amount caused by spatial shifts in the position of the intertropical convergence zone (Lewis et al., 2010; 2011), suggesting that the +1‰ higher-than-modern ice age seawater value was offset in part by drier-than-modern climate during the last glacial maximum(applying interpretation of Lewis et al., 2011).

3.5.3.2 Southeast Asia

Southeast Asian Δδ18Oice age values range from −2.3 ‰ to +1.9 ‰ (n = 13). The highest regional Δδ18Oice age values are found in Bangladesh (Δδ18Oice age of +1.6 ‰; Aggarwal et al., 2000) and in central and south-eastern China (Δδ18Oice age of 0.0 ‰ to +1.9 ‰; Wang et al., 2001; Yuan et al.,

2004; Dykoski et al., 2005; Cai et al., 2010; Yang et al., 2010). The high Bangladeshi Δδ18Oice age value of +1.6 ‰ cannot be explained solely by ice-age-to-modern changes in seawater δ18O (δ18Oice age seawater > δ18Omodern seawater by +1 ‰), suggesting that changes to temperature and humidity of the over- ocean moisture sources, air mass rainout history, precipitation seasonality, or seasonal filtering of groundwater recharge must have occurred in Bangladesh between the last ice age and the late-

Holocene.

Chinese speleothem records located between latitudes 25°N to 35°N have near-zero or positive Δδ18Oice age values. The Chinese speleothem records have been interpreted to reflect the strength of the East Asian (Wang et al., 2001; Dykoski et al., 2005; Cosford et al., 2008) or Indian monsoons (Pausata et al., 2011). Recent work proposes that interpreting Chinese speleothem isotopic

175 data as records of the summer monsoon strength is incorrect (Caley et al., 2014). Proxy evidence suggest a weaker-than-modern summer monsoon at the last glacial maximum (Wang et al., 2001) and stronger-than-modern winter precipitation (Sagawa et al., 2011; Clark et al., 2012). The North China

Plain (northeastern China; Zongyu et al., 2005) and the eastern Hexi Corridor (northern China; Gates et al., 2008) aquifers have the lowest Δδ18Oice age values observed across east Asia (Δδ18Oice age of −1.4

‰ and −2.3 ‰). Combining northern Chinese groundwater Δδ18Oice age observations (Zongyu et al.,

2005; Gates et al., 2008) with the observed positive Δδ18Oice age values of across central China (Wang et al., 2001; Yuan et al., 2004; Dykoski et al., 2005; Cai et al., 2010; Yang et al., 2010) reveals a south- to-north decrease in Δδ18Oice age (Figure 3-2).

The observed north-to-south Δδ18Oice age decrease is spatially consistent with intra-annual precipitation δ18O seasonality across southeastern Asia. Previous studies have identified increasing precipitation δ18O values from the coast (i.e., Hong Kong) to inland China (e.g., Zhangye) during the wet season, sharply contrasting spatial patterns expected from Rayleigh distillation (Aragúas-Aragúas et al., 1998). This pattern has been interpreted as the maximum northward extent of the intertropical convergence zone (i.e., broad scale Hadley circulation; Aragúas-Aragúas et al., 1998). However, more recent work suggests that low wet-season precipitation δ18O values over southern Chinese are controlled by the deflection of westerlies from the Tibetan Plateau, whereas precipitation δ18O over northern China is controlled by local-scale precipitation fluxes and subsequent evaporation of falling raindrops (Lee et al., 2012). Therefore observed Δδ18Oice age values in southern China may be reflective of broader scale atmospheric circulation pattern changes, whereas Δδ18Oice age over northern

China could reflect ice-age-to-modern changes to local meteorology. The source of precipitation over

China varies on intra-annual time scales, and about half of all rainfall is sourced from continental moisture recycling (Lewis et al., 2013). Generally, Chinese atmospheric vapor sourced from the

Indian Ocean is at a maximum during the summer, whereas Pacific-sourced moisture is greatest during the winter (Lewis et al., 2013). The strong modelled intra-annual variation in moisture sources

176 over China (Lewis et al., 2013) suggests that observed Δδ18Oice age values may represent broad-scale changes to moisture sources and associated air mass trajectories that have resulted in a strengthening of monsoon rains from the last ice age to the present day.

General circulation models predict positive Δδ18Oice age values near to the Chinese coastlines, and negative Δδ18Oice age values in western and northern China (Figures 3-9 and 3-10), consistent with observed south-to-north decrease in Δδ18Oice age values. Generally, spatial patterns of the sign of the multi-model average Δδ18Oice age agrees with the sign of the observed Δδ18Oice age (Figures 3-9 and 3-

10) in these monsoonal regions. A single hydrological process that explains all observed Δδ18Oice age values is not identifiable nor expected given the variety of processes controlling modern precipitation in southeastern Asia (Aragúas-Aragúas et al., 1998; Lee et al., 2012). However, the strong inter-model agreement on Δδ18Oice age values and model capture of the south-to-north decrease in Δδ18Oice age implies that general circulation models reproduce the broad atmospheric boundary defining the different hydrological processes governing southern vs. northern Chinese precipitation regimes.

3.5.3.3 The Middle East

The Middle East has four Δδ18Oice age records ranging from −1.6 ‰ (Kuwait) to +1.4 ‰

(Yemen); all four sites in the Middle East are located within 100 km of a coast. Records collected in

Kuwait and Yemen have both been interpreted to reflect an ice age climate that was wetter than today’s (Al-Ruwaih and Shehata, 2004; Shakun et al., 2007), although the Δδ18Oice age value is of a different sign (i.e., Kuwait being negative, and Yemen positive). Groundwater noble gas records in

Oman reveal a 6.5°C temperature increase from the last ice age to the late-Holocene (Weyhenmeyer et al., 2000). Groundwaters have an ice-age-to-late-Holocene deuterium excess increase (i.e., dice age < dmodern) interpreted to be the result of a switch in the moisture source to Oman: from the Indian

Ocean during the last ice age, to the Mediterranean Ocean today (Weyhenmeyer et al., 2000) that may be a leading control upon other records in the region.

177 3.5.3.4 Africa

African Δδ18Oice age values range from −3.0 ‰ to near-zero (Figure 3-2). 80% of African

Δδ18Oice age values are negative. Near-zero Δδ18Oice age values are generally found near to coasts (e.g.,

Senegal Δδ18Oice age of +0.3 ‰) whereas the lowest African Δδ18Oice age value is located in Niger

(Δδ18Oice age of −3.0 ‰) 800 kilometers from the coast. Records located north and south of the equator have negative Δδ18Oice age values.

Northern African changes to hydrological processes are complicated by multiple interlinked controls such as the strength of Atlantic meridional overturning circulation (Jullien et al., 2007) and meridional shifts in the position of the intertropical convergence zone (Arbuszewski et al., 2013).

Paleowater isotopic records indicate that northern Africa was 2-3°C cooler than today (Guendouz et al., 1998) and that westerly winds transporting moisture to northern Africa were stronger than present day (Sultan et al., 1997; Abouelmagd et al., 2014). North African Δδ18Oice age observations were also likely impacted by higher-than-modern sea surface humidity as evidenced by lower ice age deuterium excess values in paleowaters (Rozanski, 1985). Potentially cooler-than-modern final air mass condensation temperatures during the last ice age coupled to changes in moisture source and sea surface temperature and humidity have each been suggested and result in unanimously negative

Δδ18Oice age values across northern Africa.

Southern Africa lacustrine sediment records recovered at Lake Tanganyika and Malawi show that the eastern Africa was 2°C to 4°C cooler than modern, and that the isotopic composition of leaf waxes was highly variable between the early and late-Holocene (Powers et al., 2005; Tiereny, 2008;

2013). These records are interpreted as indicative of precipitation variations imparted by changes to

Indian Ocean temperatures (Tiereny, 2008; 2013); although this interpretation is not supported by all

(Schefuß et al., 2011). Pollen records suggest that the African tropics were both cooler and more arid at the last glacial maximum (Gasse, 2000). General circulation model simulate lower-than-modern ice

178 age precipitation fluxes over tropical Africa (Otto-Bliesner et al., 2006). Pollen records and climate model simulation models both suggest a more arid region, indicative of lower-than-modern moisture recycling during the last ice age. Rainfall originates from both Indian and Atlantic Oceanic sources, with Atlantic-sourced moisture travelling across the Congo rainforest (Levin et al., 2009). A reduction in continental moisture recycling is consistent with the observed negative Δδ18Oice age values across southern tropical Africa. Model simulations of Heinrich event precipitation δ18O confirm changes to moisture recycling and transport distance can change Δδ18Oice age values over southern

Africa. Higher-than-modern upwind convection during the last ice age may have produced negative

Δδ18Oice age values (e.g., Lekshmy et al., 2014) consistent with observed negative Δδ18Oice age values.

However, stronger-than-modern convective rainout at the last ice age is contrary to the cooler-than- modern land surface temperatures, suggesting that increases to transport distance and vapor origin changes are more likely sources of the observed negative Δδ18Oice age values (Lewis et al., 2010).

Isotope enabled general circulation model Δδ18Oice age values and observed groundwater

Δδ18Oice age values are shown in Figures 3-9 and 3-10. The four general circulation models do not agree with each other nor with multiple compiled Δδ18Oice age values over the majority of Africa

(Figure 3-9 and Figure 3-10). While the source of this discrepancy remains unclear, different parameterizations of convective rainfall amongst the models may help to explain disagreements between models (e.g., inter-model differences in the timescale for consumption of convective available potential energy; Lee et al., 2009). Indeed, recent work has shown that convection, not precipitation amount (the “amount effect”), drives tropical variations in meteoric water δ18O values

(Lekshmy et al., 2014). However, the observed negative Δδ18Oice age values are consistent with higher- than-modern upwind convection during the last ice age. Higher-than-modern convection during the last ice age is difficult to reconcile given the cooler-than-modern land surface temperatures at the last ice age (Figure 3-1). However, the observed negative Δδ18Oice age values must reflect rainout

179 processes, since ice-age-to-modern changes to seawater δ18O induce an opposing effect (i.e., positive

Δδ18Oice age) to observations.

3.5.3.5 Europe and Mediterranean nations

Europe and nations bordering the eastern Mediterranean Sea have Δδ18Oice age values ranging from −5.7 ‰ to near-zero (n = 20; Figure 3-2). 80% of European Δδ18Oice age values are negative.

Δδ18Oice age values are generally higher in western Europe (+0.1 ‰ to −1.5 ‰ in Portugal and the

United Kingdom and France) than in eastern Europe (−1.8 ‰ to −3.6 ‰ in Poland and Hungary).

The lowest ice Δδ18Oice age value is observed in a speleothem in Turkey near to the Black Sea (−5.7

‰), which is interpreted to be the dominant source of moisture over the region (Fleitmann et al.,

2009). The interpretation of a change in moisture source is consistent with recent reporting of lower- than-modern deuterium excess values during the last ice age (Arslan et al., 2013) and pollen records indicative of a drier ice age climate in eastern Turkey (Kaplan, 2013). Indeed there is large potential for an ice-age-to-modern change in the moisture sources and air mass trajectories over Turkey given the large number of potential moisture sources (e.g., Black Sea, Mediterranean Sea, Atlantic Ocean) and the cave’s location near to the margin of the Fennoscandanavian ice sheet at the last glacial maximum. The highest European Δδ18Oice age value (near-zero) is found along the Portugal coast

(Galego Fernandes and Carreira, 2008). The near-zero Δδ18Oice age value in the Portugal aquifer suggests that the effect of the higher ice age δ18Oseawater value is cancelled out by a combination of ice- age-to-modern changes in sea surface temperature and humidity, cooler condensation temperatures at the last ice age and/or greater fluxes of winter precipitation entering aquifers at the last glacial maximum.

Positive Δδ18Oice age values are observed in the eastern Mediterranean speleothems found in

Israel (Frumkin et al., 1999; Bar-Matthews et al., 2003; Ayalon et al., 2013), although groundwater aquifers in eastern Israel and in Syria have negative Δδ18Oice age values. These records are near to one

180 another, such that the opposing sign of Δδ18Oice age observed in each record is surprising. The groundwater record may record information from higher recharge-zone elevations located ~102 km from the measurement location due to advection along regional-scale flowpaths. This compilation and spatial analysis advocated for further comparative research into speleothem and groundwater isotopic compositions to ensure each record indeed reflects paleo-meteoric water δ18O values unaltered by subsequent effects (e.g., partial evaporation, etc.).

Some European aquifers have a prolonged gap in 14C-based groundwater ages interpreted to be the result of the inhibition of recharge due to permafrost aggradation (Darling, 2004). Changes to freeze-thaw conditions of the ground surface between the last ice age the modern climate may have also impacted the seasonality of groundwater recharge ratios (Darling, 2011; Jasechko et al., 2014), suggesting that recharge dynamics may represent a process that has not yet been applied to reconcile observed Δδ18Oice age values. Indeed, pollen records indicate that northern Europe was tundra-like at the last glacial maximum and that southern Europe was semi-arid, receiving ~300 mm less precipitation than modern day (Clark et al., 2012 and references therein). The glacial-to-modern transition from semi-arid deserts to temperate forests may have modified the seasonality of the groundwater recharge ratio as evidenced by modern day recharge being much more efficient during the winter (Jasechko et al., 2014). This potential ice-age-to-modern change in recharge/precipitation ratios would have likely enhanced winter recharge fluxes resulting in negative shifts in Δδ18Oice age values consistent with observations.

General circulation model outputs of Δδ18Oice age values over Europe are unanimously negative (i.e., all four models agree on the sign of Δδ18Oice age Figure 3-9; 3-10), with the exception of southern Portugal and Spain. The model predictions across Europe closely match the observations of

Δδ18Oice age values in groundwaters and speleothems. Earlier works have suggested the European moisture sources and air mass trajectories have not changed considerably since the last ice age

181 (Rozanski et al., 1985; Loosli et al., 2001). The match between simulated and modelled negative

Δδ18Oice age values implies that the last ice age had cooler final air mass condensation temperatures, higher winter groundwater recharge ratios or higher proportions of winter precipitation as a proportion of annual totals. However, modelled Δδ18Oice age values agree less frequently over Israel and Syria, where aquifer and speleothem observations also show both positive and negative Δδ18Oice age values. The conflicting model outputs and observations of Δδ18Oice age values in the eastern

Mediterranean suggest that moisture sources, air mass trajectories and meteorology is highly sensitive to change in the eastern Mediterranean, and that this sensitivity varies over distances of ~102 kilometers.

3.5.3.6 South America

South American Δδ18Oice age values range from −6.5 ‰ to 0.0 ‰ (Figure 3-2), with the lowest values found in ice cores in the Andes (Bolivia: δ18O anomaly of −4.9 ‰; Thompson et al.,

1998; Bolivia: δ18O anomaly of −6.5 ‰; Thompson et al., 1998). The Δδ18Oice age values found in the ice cores have been interpreted to have been coupled to substantial cooling of the tropics (quoted as

8°C to 12°C; Thompson et al., 1995), and may also be related to changes in moisture recycling over the Amazon, which is the dominant moisture source to the Andes (Thompson et al., 1998).

Paleowater Δδ18Oice age data is available in the semi-arid eastern portion of Brazil. The interpretation of this record was that rainfall was higher-than-modern during the Pleistocene (Salati et al., 1974), consistent with greater upwind convective rainfall leading to negative Δδ18Oice age values. However, recent work proposes that precipitation was lower-than-modern in eastern Brazil at the last glacial maximum (Cruz et al., 2009; Clark et al., 2012). Eastern Brazilian precipitation is anti-phased with precipitation fluxes in the South American monsoon region (Cruz et al., 2009) where ice age precipitation fluxes are thought to be higher-than-modern (Wang et al., 2007). This aquifer is located

~100 km from the Atlantic Ocean at a latitude of 5°S, and the region was 5.4°C cooler than today during the last glacial maximum (Stute et al., 1995b). The intra-annual variability in the isotopic

182 composition of precipitation is subdued, with the summer/wet-season (April to September) having a nearly identical isotopic composition (−2.2‰) to that of the winter/dry-season (−2.0‰; data from

Ceara Mirim, Brazil; data accessed from www.iaea.org/water), suggesting that changes to the seasonality of precipitation amounts or the seasonality of the groundwater recharge ratio are not the source of observed negative Δδ18Oice age value in eastern Brazil. Possible processes that may explain the negative Δδ18Oice age values in eastern Brazil include higher-than-modern upwind convection during the last ice age (Salati et al., 1974), supported as a leading control on Δδ18Oice age by general circulation model simulations that suggest local rainfall amounts govern precipitation δ18O (Lewis et al., 2010). Observed Δδ18Oice age values in eastern Brazil support the interpretation of Salati et al.

(1974) that eastern Brazil was wetter than today during the last glacial maximum.

General circulation models have unanimously positive Δδ18Oice age values over semi-arid eastern Brazil (Figures 3-9 and 3-10). Interestingly, the Δδ18Oice age value observed in this region is negative (Salati et al., 1974). It is clear that the models have not captured all processes in this region, as the predicted Δδ18Oice age values are of a different sign than the observed Δδ18Oice age value. I have ruled out seasonality of precipitation fluxes as the sole process controlling Δδ18Oice age values in this region. However, changes to moisture sources, air mass recycling, rainout history and moisture recycling (i.e., processes ii through v) may each be an important control upon Δδ18Oice age values in eastern Brazil. Similarly, upstream convective rainstorms (potentially not accurately parameterized within all general circulation models) stronger-than-modern during the last ice age could have contributed to observed negative Δδ18Oice age values.

3.5.3.7 North America

North American Δδ18Oice age records are all located in the United States of America and range from −3.1 ‰ to +1.8 ‰ (n = 14). Easternmost USA has Δδ18Oice age values that are positive or near- zero. The positive Δδ18Oice age values are highest in Florida (latitude: 27°N; Δδ18Oice age of +1.8 ‰)

183 and decrease northward through Georgia (latitude: 32°N; Δδ18Oice age of +1.0 ‰) to coastal Maryland

(latitude 39°N; Δδ18Oice age of 0.0 ‰). The decreasing Δδ18Oice age values observed with increasing latitude along the USA eastern seaboard that may be partially explained by the isotopic distillation of air masses advecting northward from the tropics. The effect of the higher ice age δ18Oseawater values has been shown to be offset by the lowering of the sea level during the last ice age which increases the distance-to-the-coast because of sea level regression (Clark et al., 1997; Aeschbach-Hertig et al.,

2002). Other potentially important processes that may reconcile the observed Δδ18Oice age values include changes to seasonal precipitation rates, changes to moisture recycling due to differing

Pleistocene vegetation in the region (Harrison et al., 2003) or changes to hurricane frequency and intensity (i.e., precipitation seasonality; Plummer, 1993) could have impacted Δδ18Oice age values.

Further, recent research show that seawater δ18O values changed over time in the Gulf of Mexico

(i.e., one of the moisture sources to central and southeastern USA; Feng et al., 2014). δ18Oseawater changes from the last ice age to the present day due to fluctuations in Mississippi River discharge may have impacted terrestrial Δδ18Oice age values, with higher Mississippi discharges leading to lower seawater δ18O and lower terrestrial precipitation δ18O values.

Westernmost USA has negative Δδ18Oice age values (e.g., Los Angeles basin Δδ18Oice age of

−1.4 ‰), contrasting Δδ18Oice age values observed along the eastern coast at similar latitudes.

Although the reason for this east-coast/west-coast difference may have multiple explanations, higher than modern winter precipitation fluxes during the last ice age could invoke a negative Δδ18Oice age value consistent with observations.

Central USA has the lowest Δδ18Oice age values that range from −0.6‰ to −3.1‰. The southwestern USA was 5°C cooler than today during the last glacial maximum (Stute et al., 1995a).

The low inland Δδ18Oice age values observed in central North America are consistent with the enhanced isotopic distillation of moisture advecting overland due to cooler final condensation

184 temperatures, or higher proportions of annual precipitation falling during the winter season. The observed Δδ18Oice age values in these records have been attributed to lower-than-modern summer precipitation fluxes during the late Pleistocene (New Mexico, Phillips et al., 1986), latitudinal shifts in the positions of the polar jet stream and the intertropical convergence zone (New Mexico, Asmerom et al., 2010) and changes to over-ocean humidity, temperature or moisture sources (Idaho, Schlegel et al., 2009). Pollen records indicate widespread forests throughout the present day deserts of the

American southwest, indicative of wetter-than-modern conditions at the last glacial maximum

(Williams, 2003). An isotopic record at Cave of the Bells is interpreted to reflect southwestern aridity, with reductions in paleo-δ18O interpreted to reflect a cooler and a wetter climate (Arizona; Wagner et al., 2010). Extending this interpretation to the observed negative Δδ18Oice age values throughout the southwest USA, the groundwater Δδ18Oice age values suggest that the American southwest was both cooler and more humid during the last ice age compared to present day. The source of higher-than- modern ice age humidity may be linked to changes in air mass trajectories and moisture sources to the southwestern USA (Asmerom et al., 2010; Wagner et al., 2010). Further, the strong intra-annual variability in the isotopic composition of modern day precipitation (more than a 7 ‰ difference between the summer and winter δ18O values) suggests that increases to winter precipitation or higher recharge/precipitation ratios could also contribute to the observed negative Δδ18Oice age values in southwestern USA groundwaters.

General circulation model results Δδ18Oice age values are generally positive along the eastern seaboard, the Gulf States and the central plains of the USA (Figures 3-9; 3-10). The modelled

Δδ18Oice age results are consistent with the sign of Δδ18Oice age values observed in aquifers across

Florida, Georgia and Maryland (Plummer, 1993; Clark et al., 1997; Aeschbach-Hertig et al., 2002;

Morrissey et al., 2010; Figures 3-9 and 3-10), although no isotopic record of the last glacial maximum is available for aquifers across the Gulf States (e.g., Edwards Aquifer, Texas). The general circulation models Δδ18Oice age values are generally negative west of the Rocky Mountains, consistent with the

185 sign of observed Δδ18Oice age values in Colorado, New Mexico and Idaho (Clark et al., 1998; Stute et al., 1995a; Schlegel et al., 2009; Asmerom et al., 2010).

Conclusions

The Δδ18Oice age of groundwater aquifers compiled in this study ranges from −3.6 ‰ (i.e.,

δ18Oice age < δ18Olate-Holocene) to +2.0 ‰ (i.e., δ18Oice age > δ18Olate-Holocene). ~90% of aquifers have negative Δδ18Oice age values. Aquifers with positive Δδ18Oice age values are found exclusively near to coasts. Future research may capitalize upon the broad availability of groundwater isotopic records of last ice age and late-Holocene climate in order to isolate and constrain hydrological processes responsible for observed Δδ18Oice age values using general circulation models (e.g., Lewis et al., 2010).

Further, observed Δδ18Oice age values are compared to general circulation model outputs of Δδ18Oice age values. The general circulation models agree in the sign and magnitude of Δδ18Oice age values for some, but not all locations. This synthesis and sensitivity analysis advocates for the use of quantitative models when interpreting precipitation δ18O paleoclimate records.

Regional paleoclimate signals show that during the last ice age:

- Australia was more arid, ~10°C cooler, and had either greater contributions of winter

precipitation or increased rainout and isotopic fractionation of air masses potentially due to

cooler-than-modern atmospheric condensation temperatures.

- Southern Chinese summer monsoons were weaker-than-modern, winter rainfall was higher-

than-modern. Northern China was 5°C cooler and more humid than present climate

conditions.

- The Middle East was 6.5°C cooler than today and received greater vapour fluxes from the

Indian Ocean than present day, producing a wetter overall climate during the last ice age.

- Northern Africa was 2°C to 3°C cooler than modern climate and had greater vapor influxes

from westerly moisture sources, creating a more humid climate than present day.

186 - Southern African was 2°C to 4°C cooler than modern, had higher-than-modern vapor inputs

from westerly moisture sources, potentially lower-than-modern moisture recycling over the

Congo rainforest and potentially greater-than-modern upwind convective rainfall.

- European climate was 3°C to 9°C cooler than present day, had broadly similar-to-present

vapor inflows from westerly moisture sources, and may have had substantially higher-than-

modern winter groundwater recharge ratios or cooler final condensation temperatures (i.e.,

greater upwind air mass distillation) leading to observed unanimously negative Δδ18Oice age

values.

- Eastern Brazil was 5°C cooler and was more humid than present day climate due to greater-

than-modern monsoonal rainstorms.

- The southwestern USA was 5°C cooler than today (Stute et al., 1995a), was more humid

than modern climate, and potentially received greater-than-modern vapor fluxes from

westerly moisture sources or higher-than-modern winter precipitation fluxes.

Acknowledgements

I thank C. Risi, F. Pausata, K. Yoshimura and M. Werner for their time and help compiling results from the isotope enabled general circulation models. I also thank A. Lechler, F. Pausata and

T. Gleeson for the valuable insights that have improved this chapter.

187 3.6 References

Abouelmagd, A. et al. (2012), Toward a better understanding of palaeoclimatic regimes that recharged the fossil aquifers in North Africa: Inferences from stable isotope and remote sensing data,

Palaeogeography, Palaeoclimatology, Palaeoecology, 329, 137-149.

Aeschbach-Hertig, W., Stute, M., Clark, J. F., Reuter, R. F., and Schlosser, P. (2002), A paleotemperature record derived from dissolved noble gases in groundwater of the Aquia Aquifer

(Maryland, USA), Geochimica et Cosmochimica Acta, 66, 797–817.

Aggarwal, P. K. et al. (2000), A report on isotope hydrology of groundwater in Bangladesh: implications for characterization and mitigation of arsenic in groundwater, International Atomic

Energy Agency, Department of Technical Co-operation, Vienna (Austria).

Aggarwal, P. K., K. Fröhlich, K. M. Kulkarni, and L. L. Gourcy (2004), Stable isotope evidence for moisture sources in the Asian summer monsoon under present and past climate regimes, Geophysical Research Letters, 31, L08203.

Al-Mashaikhi, K., Oswald, S., Attinger, S., Büchel, G., Knöller, K., and Strauch, G. (2012),

Evaluation of groundwater dynamics and quality in the Najd aquifers located in the Sultanate of

Oman, Environmental Earth Sciences, 66, 1195–1211.

Al-Ruwaih, F. M., and Shehata, M. (2004), Hydrochemical processes and environmental isotopic study of groundwater in Kuwait, Water International, 29, 158–166.

Andrews, J. N., Fontes, J. C., Aranyossy, J. F., Dodo, A., Edmunds, W. M., Joseph, A., and

Travi, Y. (1994), The evolution of alkaline groundwaters in the continental intercalaire aquifer of the

Irhazer Plain, Niger, Water Resources Research, 30, 45–61.

188 Annan, J. D., and Hargreaves, J. C. (2013), A new global reconstruction of temperature changes at the Last Glacial Maximum. Climate of the Past, 9, 367–376.

Aragúas-Aragúas, L., K. Froehlich, and Rozanski, K. (1998), Stable isotope composition of precipitation over Southeast Asia, Journal of Geophysical Research, 103, 721–742.

Arbuszewski, J. A., Cléroux, C., Bradtmiller, L., and Mix, A. (2013), Meridional shifts of the

Atlantic intertropical convergence zone since the Last Glacial Maximum, Nature Geoscience, 6, 959–

962.

Arslan, S., Yazicigil, H., Stute, M., and Schlosser, P. (2013), Environmental isotopes and noble gases in the deep aquifer system of Kazan Trona Ore Field, Ankara, central Turkey and links to paleoclimate, Quaternary Research, 79, 292–303.

Asmerom, Y., Polyak, V. J., and Burns, S. J. (2010), Variable winter moisture in the southwestern United States linked to rapid glacial climate shifts, Nature Geoscience, 3, 114–117.

Ayalon, A., Bar-Matthews, M., Frumkin, A., and Matthews, A. (2013), Last Glacial warm events on Mount Hermon: the southern extension of the Alpine karst range of the east

Mediterranean, Quaternary Science Reviews, 59, 43–56.

Back, W., Hanshaw, B. B., Plummer, L. N., Rahn, P. H., Rightmire, C. T., and Rubin, M.

(1983), Process and rate of dedolomitization: mass transfer and 14C dating in a regional carbonate aquifer, Geological Society of America Bulletin, 94, 1415–1429.

Bar-Matthews, M., Ayalon, A., Gilmour, M., Matthews, A., and Hawkesworth, C. (2003),

Sea-land oxygen isotopic relationships from planktonic foraminifera and speleothems in the Eastern

Mediterranean region and their implication for paleorainfall during interglacial intervals, Geochimica et

Cosmochimica Acta, 67, 3181–3199.

189 Barbecot, F., Marlin, C., Gibert, E., and Dever, L. (2000), Hydrochemical and isotopic characterisation of the Bathonian and Bajocian coastal aquifer of the Caen area (northern France),

Applied Geochemistry, 15, 791–805.

Beyerle, U., Rueedi, J., Leuenberger, M., Aeschbach‐Hertig, W., Peeters, F., Kipfer, R., and

Dodo, A. (2003), Evidence for periods of wetter and cooler climate in the Sahel between 6 and 40 kyr BP derived from groundwater, Geophysical Research Letters, 30, 1173.

Blaser, P. C., Kipfer, R., Loosli, H. H., Walraevens, K., Van Camp, M., and Aeschbach‐

Hertig, W. (2010), A 40 ka record of temperature and permafrost conditions in northwestern Europe from noble gases in the Ledo‐Paniselian Aquifer (Belgium), Journal of Quaternary Science, 25, 1038–

1044.

Blunier, T., and Brook, E. J. (2001), Timing of millennial-scale climate change in Antarctica and Greenland during the last glacial period. Science, 291, 109–112.

Bouchaou, L., Michelot, J. L., Vengosh, A., Hsissou, Y., Qurtobi, M., Gaye, C. B., Bullen, T.

D., and Zuppi, G. M. (2008), Application of multiple isotopic and geochemical tracers for investigation of recharge, salinization, and residence time of water in the Souss–Massa aquifer, southwest of Morocco, Journal of Hydrology, 352, 267–287.

Bouchaou, L. et al. (2009), Origin and residence time of groundwater in the Tadla basin

(Morocco) using multiple isotopic and geochemical tools, Journal of Hydrology, 379, 323–338.

Buiron, D. et al. (2011), TALDICE-1 age scale of the Talos Dome deep ice core, East

Antarctica, Climate of the Past, 7, 1–16.

190 Burchuladze, A. A., Chudy, M., Eristavi, I. V., Pagava, S. V., Povinec, P., Sivo, A., and

Togonidze, G. I. (1989), Anthropogenic 14C variations in atmospheric CO2 and wines, Radiocarbon,

31, 771–776.

Cai, Y., Tan, L., Cheng, H., An, Z., Edwards, R. L., Kelly, M. J., Kong, X., and Wang, X.

(2010), The variation of summer monsoon precipitation in central China since the last deglaciation,

Earth and Planetary Science Letters, 291, 21–31.

Caley, T., Roche, D. M., Renssen, H. (2014), Orbital Asian summer monsoon dynamics revealed using an isotope-enabled global climate model, Nature Communications, 10, 105–148.

Castany, G., Marce, A., Margat, J., Moussu, H., Vuillaume, Y., and Evin, J. (1974), An environmental isotope study of the groundwater regime in large aquifers. In: Isotope techniques in groundwater hydrology 1974, Vol. I.

Celle-Jeanton, H., Huneau, F., Travi, Y., and Edmunds, W. M. (2009), Twenty years of groundwater evolution in the Triassic sandstone aquifer of Lorraine: impacts on baseline water quality. Applied Geochemistry, 24, 1198–1213.

Chen, X. Y., Bowler, J. M. and Magee, J. W. (1993). Late Cenozoic stratigraphy and hydrologic history of Lake Amadeus, a central Australian playa, Australian Journal of Earth Sciences,

40, 1–14.

Chen, Z., Wei, W., Liu, J., Wang, Y., and Chen, J. (2011), Identifying the recharge sources and age of groundwater in the Songnen Plain (Northeast China) using environmental isotopes.

Hydrogeology Journal, 19, 163–176.

Clark, I. D., and Fritz, P. (1997). Environmental isotopes in hydrogeology. CRC press.

191 Clark, J. F., Stute, M., Schlosser, P., Drenkard, S., and Bonani, G. (1997), A tracer study of the Floridan aquifer in southeastern Georgia: Implications for groundwater flow and paleoclimate,

Water Resources Research, 33, 281-289.

Clark, J. F., Davisson, M. L., Hudson, G. B., and Macfarlane, P. A. (1998), Noble gases, stable isotopes, and radiocarbon as tracers of flow in the Dakota aquifer, Colorado and Kansas,

Journal of Hydrology, 211, 151–167.

Clark, P. U., Dyke, A. S., Shakun, J. D., Carlson, A. E., Clark, J., Wohlfarth, B., Mitrovica, J.

X., Hostetler, S. W., and McCabe, A. M. (2009), The last glacial maximum, Science, 325, 710–714.

Clark, P. U. et al. (2012), Global climate evolution during the last deglaciation, Proceedings of the National Academy of Sciences, 109, E1134–E1142.

Cosford, J., Qing, H., Yuan, D., Zhang, M., Holmden, C., Patterson, W., and Hai, C. (2008),

Millennial-scale variability in the Asian monsoon: Evidence from oxygen isotope records from stalagmites in southeastern China, Palaeogeography, Palaeoclimatology, Palaeoecology, 266, 3–12.

Cresswell, R., Wischusen, J., Jacobson, G., and Fifield, K. (1999), Assessment of recharge to groundwater systems in the arid southwestern part of Northern Territory, Australia, using chlorine–

36, Hydrogeology Journal, 7, 393–404.

Cruz, F. W., Burns, S. J., Karmann, I., Sharp, W. D., Vuille, M., Cardoso, A. O., Ferrari, J.

A., Dias, P. L. S., and Viana, O. (2005), Insolation-driven changes in atmospheric circulation over the past 116,000 years in subtropical Brazil. Nature, 434, 63–66.

Cruz, F. W. et al. (2009), Orbitally driven east–west antiphasing of South American precipitation. Nature Geoscience, 2, 210–214.

192 Currell, M. J., Cartwright, I., Bradley, D. C., and Han, D. (2010). Recharge history and controls on groundwater quality in the Yuncheng Basin, north China, Journal of Hydrology, 385, 216–

229.

Dansgaard, W. (1964), Stable isotopes in precipitation, Tellus, 16, 436–468.

Dansgaard, W., and Tauber, H. (1969), Glacier oxygen-18 content and Pleistocene ocean temperatures, Science, 166, 499–502.

Darling, W. G. (2004), Hydrological factors in the interpretation of stable isotopic proxy data present and past: a European perspective, Quaternary Science Reviews, 23, 743–770.

Darling, W. G. (2011), The isotope hydrology of quaternary climate change, Journal of Human

Evolution, 60, 417–427.

Darling, W. G., and Bath, A. H. (1988), A stable isotope study of recharge processes in the

English Chalk, Journal of Hydrology, 101, 31–46.

Darling, W. G., Edmunds, W. M., and Smedley, P. L. (1997), Isotopic evidence for palaeowaters in the British Isles, Applied Geochemistry, 12, 813–829.

Dennis, F., Andrews, J. N., Parker, A., Poole, J. and Wolf, M. (1997), Isotopic and noble gas study of Chalk groundwater in the London Basin, England. Applied Geochemistry, 12, 763–773.

Denniston, R. F., González, L. A., Asmerom, Y., Reagan, M. K., and Recelli-Snyder, H.

(2000), Speleothem carbon isotopic records of Holocene environments in the Ozark Highlands,

USA. Quaternary International, 67, 21–27.

193 Denniston, R. F., González, L. A., Asmerom, Y., Sharma, R. H., and Reagan, M. K. (2000),

Speleothem evidence for changes in Indian summer monsoon precipitation over the last ∼2300 years, Quaternary Research, 53, 196–202.

Derwich, L. J., Zouar, K., and Michelot, J. L. (2012), Recharge and paleorecharge of the deep groundwater aquifer system in the Zeroud Basin (Kairouan plain, Central Tunisia), Quaternary

International, 257, 56–63.

Dodo, A., and Zuppi, G. M. (1997), Groundwater flow study in the Bilma–Djado Basin

(Niger) by means of environmental isotopes, Comptes Rendus de l'Academie des Sciences. Serie 2, Sciences de la Terre et des Planetes, 30, 845–852.

Dodo, A., and Zuppi, G. M. (1999), Variabilité climatique durant le Quaternaire dans la nappe du Tarat (Arlit, Niger), Comptes Rendus de l'Académie des Sciences–Series IIA–Earth and Planetary

Science, 328, 371–379.

Douglas, A. A., Osiensky, J. L., & Keller, C. K. (2007). Carbon–14 dating of ground water in the Palouse Basin of the Columbia River basalts, Journal of Hydrology, 334, 502–512.

Dykoski, C. A., Edwards, R. L., Cheng, H., Yuan, D., Cai, Y., Zhang, M., Lin, Y., Qing, J.,

An, Z., and Revenaugh, J. (2005), A high-resolution, absolute-dated Holocene and deglacial Asian monsoon record from Dongge Cave, China. Earth and Planetary Science Letters, 233, 71–86.

Edmunds, W. M. (2009), Palaeoclimate and groundwater evolution in Africa—implications for adaptation and management. Hydrological Sciences Journal, 54, 781–792.

Edmunds, W. M. et al. (2003), Groundwater evolution in the Continental Intercalaire aquifer of southern Algeria and Tunisia: trace element and isotopic indicators, Applied Geochemistry, 18, 805–

822.

194 Edmunds, W., and C. Milne (Eds.) (2001), Palaeowaters in coastal Europe: Evolution of

Groundwater since the late Pleistocene, Geol. Society Special. Publication, 189, Geological Society of

London, London.

Elliot, T., Andrews, J. N., and Edmunds, W. M. (1999), Hydrochemical trends, palaeorecharge and groundwater ages in the fissured Chalk aquifer of the London and Berkshire

Basins, UK, Applied Geochemistry, 14, 333–363.

Emiliani, C. (1955), Pleistocene temperatures, The Journal of Geology, 538–578.

EPICA community members (2006), One-to-one coupling of glacial climate variability in

Greenland and Antarctica. Nature, 444, 195–198.

Fairbanks, R. G., Mortlock, R. A., Chiu, T. C., Cao, L., Kaplan, A., Guilderson, T. P.

Fairbanks, T. W., Bloom, A. L., Grootes, P. M., and Nadeau, M.-J. (2005), Radiocarbon calibration curve spanning 0 to 50,000 years BP based on paired 230Th/234U/238U and 14C dates on pristine corals, Quaternary Science Reviews, 24, 1781-1796.

Feng, W., Casteel, R. C., Banner, J. L., and Heinze-Fry, A. (2014), Oxygen isotope variations in rainfall, drip-water and speleothem calcite from a well-ventilated cave in Texas, USA: Assessing a new speleothem temperature proxy, Geochimica et Cosmochimica Acta, 127, 233–250.

Ferguson, G. A., Betcher, R. N., and Grasby, S. E. (2007), Hydrogeology of the Winnipeg formation in Manitoba, Canada, Hydrogeology Journal, 15, 573–587.

Fleitmann, D. et al. (2009), Timing and climatic impact of Greenland interstadials recorded in stalagmites from northern Turkey, Geophysical Research Letters, 36, L19707.

Frumkin, A., Ford, D. C., and Schwarcz, H. P. (1999), Continental oxygen isotopic record of the last 170,000 years in Jerusalem, Quaternary Research, 51, 317–327. 195 Galego Fernandes, P., and Carreira, P. M. (2008), Isotopic evidence of aquifer recharge during the last ice age in Portugal, Journal of Hydrology, 361, 291–308.

Gasse F. (2000), Hydrological changes in the African tropics since the Last Glacial

Maximum, Quaternary Science Reviews, 19, 189–211.

Gat, J.R., Mazor, E. and Tzur, Y., (1969), The stable isotope composition of mineral waters in the Jordan Rift Valley, Journal of Hydrology, 7, 334–352.

Gates, J. B., Edmunds, W. M., Darling, W. G., Ma, J., Pang, Z., and Young, A. A. (2008),

Conceptual model of recharge to southeastern Badain Jaran Desert groundwater and lakes from environmental tracers, Applied Geochemistry, 23, 3519–3534.

Geyh, M. A., and Sofner, B. (1989), Groundwater analysis of environmental carbon and other isotopes from the Jakarta Basin Aquifer, Indonesia, Radiocarbon, 31, 919–925.

Gooddy, D. C., Darling, W. G., Abesser, C., and Lapworth, D. J. (2006), Using chlorofluorocarbons (CFCs) and sulphur hexafluoride SF6 to characterise groundwater movement and residence time in a lowland Chalk catchment, Journal of Hydrology, 330, 44–52.

Gosselin, D. C., Harvey, F. E., and Frost, C. D. (2001), Geochemical evolution of ground water in the Great Plains (Dakota) Aquifer of Nebraska: implications for the management of a regional aquifer system, Ground Water, 39, 98–108.

Gouvea da Silva, R. B. (1983), Estudo hidroquımico e isotópico das águas subterrâneas do aquıfero Botucatu no estado de Sao Paulo, Ph. D. thesis, University of Sao Paulo, Sao Paulo, Brazil.

Grasby, S. E., and Chen, Z. (2005), Subglacial recharge into the Western Canada

Sedimentary Basin—Impact of Pleistocene glaciation on basin hydrodynamics, Geological Society of

America Bulletin, 117, 500–514. 196 Guendouz, A., Moulla, A. S., Edmunds, W. M., Shand, P., Poole, J., Zouari, K., and Mamou,

A. (1998), Palaeoclimatic information contained in groundwaters of the Grnad Erg Oriental, northern Africa. In: Isotope Techniques in the study of Environmental Change, International Atomic

Energy Agency.

Hackley, K. C., Panno, S. V., and Anderson, T. F. (2010), Chemical and isotopic indicators of groundwater evolution in the basal sands of a buried bedrock valley in the midwestern United

States: Implications for recharge, rock–water interactions, and mixing, Geological Society of America

Bulletin, 122, 1047–1066.

Harrington, G., Stelfox, L., Gardner, W. P., Davies, P., Doble, R., and Cook, P. G. (2011),

Surface water–groundwater interactions in the lower Fitzroy River, Western Australia, CSIRO

Publications.

Harrison, S. P., and Prentice, I. C. (2003), Climate and CO2 controls on global vegetation distribution at the last glacial maximum: analysis based on palaeovegetation data, biome modelling and palaeoclimate simulations, Global Change Biology, 9, 983–1004.

Heaton, T. H. E., Talma, A. S., and Vogel, J. C. (1986), Dissolved gas paleotemperatures and

18O variations derived from groundwater near Uitenhage, South Africa, Quaternary Research, 25, 79–

88.

Hoffmann, G., Werner, M., and Heimann, M. (1998), Water isotope module of the ECHAM atmospheric general circulation model: A study on timescales from days to several years. Journal of

Geophysical Research: Atmospheres, 103, 16871–16896.

Holmgren, K., Lee-Thorp, J. A., Cooper, G. R., Lundblad, K., Partridge, T. C., Scott, L.,

Sithaldeen, R., Talma, A. S., and Tyson, P. D. (2003), Persistent millennial-scale climatic variability over the past 25,000 years in Southern Africa, Quaternary Science Reviews 22, 2311–2326. 197 Hope, P. (2005), The Weather and Climate of Australia at the Last Glacial Maximum, Ph.D.

Dissertation, The University of Melbourne, 292 pp.

Huneau, F. et al. (2011), Flow pattern and residence time of groundwater within the south– eastern Taoudeni sedimentary basin (Burkina Faso, Mali), Journal of Hydrology, 409, 423–439.

Jasechko, S., Birks, S. J., Gleeson, T., Wada, Y., Fawcett, P. J., Sharp, Z. D., McDonnell, J. J. and Welker, J. M. (2014), The pronounced seasonality of global groundwater recharge. Water

Resources Research, doi: 10.1002/2014WR015809.

Jiráková, H., Huneau, F., Celle-Jeanton, H., Hrkal, Z., and Le Coustumer, P. (2009),

Palaeorecharge conditions of the deep aquifers of the Northern Aquitaine region (France), Journal of

Hydrology, 368, 1-16.

Jiráková, H., Huneau, F., Celle-Jeanton, H., Hrkal, Z., and La Coustumer, P. L. (2011),

Insights into palaeorecharge conditions for European deep aquifers, Hydrogeology Journal, 19, 1545–

1562.

Jones, I. C., Banner, J. L., and Humphrey, J. D. (2000), Estimating recharge in a tropical karst aquifer, Water Resources Research, 36, 1289–1299.

Kaplan, G. (2013), Palynological analysis of the Late Pleistocene terrace deposits of Lake

Van, eastern Turkey: Reconstruction of paleovegetation and paleoclimate, Quaternary International,

292, 168–175.

Karro, E., Marandi, A., and Vaikmäe, R. (2004), The origin of increased salinity in the

Cambrian-Vendian aquifer system on the Kopli Peninsula, northern Estonia, Hydrogeology Journal, 12,

424–435.

198 Kawamura, K., et al. (2007), Northern Hemisphere forcing of climatic cycles in Antarctica over the past 360,000 years. Nature, 448, 912–916.

Kreuzer, A. M., von Rohden, C., Friedrich, R., Chen, Z., Shi, J., Hajdas, I., Kipfer, R., and

Aeschbach–Hertig, W. (2009), A record of temperature and monsoon intensity over the past 40 kyr from groundwater in the North China Plain, Chemical Geology, 259, 168–180.

Külls, C. (2000), Groundwater of the North–Western Kalahari, Namibia. Ph.D. Thesis,

Julius–Maximilian University of Würzburg.

Kulongoski, J. T., Hilton, D. R., and Selaolo, E. T. (2004), Climate variability in the

Botswana Kalahari from the late Pleistocene to the present day, Geophysical Research Letters, 31,

L10204.

Kumar, S. U., Sharma, S., Navada, S. V., and Deodhar, A. S. (2009), Environmental isotopes investigation on recharge processes and hydrodynamics of the coastal sedimentary aquifers of

Tiruvadanai, Tamilnadu State, India, Journal of Hydrology, 364, 23–39.

Lachniet, M. S., Asmerom, Y., Burns, S. J., Patterson, W. P., Polyak, V. J., and Seltzer, G. O.

(2004), Tropical response to the 8200 yr BP cold event? Speleothem isotopes indicate a weakened early Holocene monsoon in Costa Rica, Geology, 32, 957–960.

Lambeck, K., Yokoyama, Y., Johnston, P., and Purcell, A. (2000), Global ice volumes at the

Last Glacial Maximum and early Lateglacial, Earth and Planetary Science Letters, 181, 513–527.

Larsen, F., Owen, R., Dahlin, T., Mangeya, P., & Barmen, G. (2002). A preliminary analysis of the groundwater recharge to the Karoo formations, mid–Zambezi basin, Zimbabwe, Physics and

Chemistry of the Earth, 27, 765–772.

199 Leaney, F. W., and Allison, G. B. (1986), Carbon–14 and stable isotope data for an area in the Murray Basin: its use in estimating recharge. Journal of Hydrology, 88, 129–145.

Lee, J.-E., R. Pierrehumbert, A. Swann, and B. R. Lintner (2009), Sensitivity of stable water isotopic values to convective parameterization schemes, Geophys. Res. Lett., 36, L23801,

Lee, J.-E., Risi, C., Fung, I., Worden, J., Scheepmaker, R. A., Lintner, B., and Frankenberg,

C. (2012), Asian monsoon hydrometeorology from TES and SCIAMACHY water vapor isotope measurements and LMDZ simulations: Implications for speleothem climate record interpretation,

Journal of Geophysical Research, 117, D15112.

LeGrande, A. N., and Schmidt, G. A. (2009), Sources of Holocene variability of oxygen isotopes in paleoclimate archives, Climate of the Past, 5, 1133–1162.

Lekshmy, P. R., Midhun, M., Ramesh, R., and Jani, R. A. (2014), 18O depletion in monsoon rain relates to large scale organized convection rather than the amount of rainfall. Scientific Reports, 4.

Leng, M. J., and Marshall, J. D. (2004), Palaeoclimate interpretation of stable isotope data from lake sediment archives, Quaternary Science Reviews, 23, 811–831.

Levin, N. E., Zipser, E. J. and Cerling, T. E. (2009), Isotopic composition of waters from

Ethiopia and Kenya: Insights into moisture sources for eastern Africa, Journal of Geophysical Research,

114, D23306.

Lewis, S. C., LeGrande, A. N., Kelley, M., and Schmidt, G. A. (2010), Water vapour source impacts on oxygen isotope variability in tropical precipitation during Heinrich events, Climate of the

Past, 6, 325–343.

200 Lewis, S. C. et al. (2011). High-resolution stalagmite reconstructions of Australian–

Indonesian monsoon rainfall variability during Heinrich stadial 3 and Greenland interstadial 4, Earth and Planetary Science Letters, 303, 133–142.

Lewis, S. C., LeGrande, A. N., Kelley, M. and Schmidt, G. A. (2013), Modeling insights into deuterium excess as an indicator of water vapor source conditions, Journal of Geophysical Research, 118,

243–262.

Lisiecki, L. E., and Raymo, M. E. (2005), A Pliocene-Pleistocene stack of 57 globally distributed benthic δ18O records, Paleoceanography, 20, PA1003.

Liu, X., Shen, J., Wang, S., Wang, Y., and Liu, W. (2007), Southwest monsoon changes indicated by oxygen isotope of ostracode shells from sediments in since the late

Glacial, Chinese Science Bulletin, 52, 539–544.

Loosli, H. H. et al. (2001), Isotopic methods and their hydrogeochemical context in the investigation of palaeowaters, Geological Society, London, Special Publications, 189, 193–212.

Maduabuchi, C., Faye, S., and Maloszewski, P. (2006), Isotope evidence of palaeorecharge and palaeoclimate in the deep confined aquifers of the Chad Basin, NE Nigeria, Science of the total environment, 370, 467–479.

Marcott, S. A., Shakun, J. D., Clark, P. U., and Mix, A. C. (2013), A reconstruction of regional and global temperature for the past 11,300 years, Science 339, 1198–1201.

MARGO Members (2009), Constraints on the magnitude and patterns of ocean cooling at the Last Glacial Maximum, Nature Geoscience, 2, 127–132.

201 McDermott, F., Mattey, D. P., and Hawkesworth, C. (2001), Centennial-scale Holocene climate variability revealed by a high-resolution speleothem δ18O record from SW Ireland, Science,

294, 1328–1331.

McIntosh, J. C., Schlegel, M. E., and Person, M. (2012), Glacial impacts on hydrologic processes in sedimentary basins: evidence from natural tracer studies, Geofluids, 12, 7–12.

Miller, G. H., Magee, J. W. and Jull, A. J. T. (1997). Low-latitude glacial cooling in the

Southern Hemisphere from amino-acid racemization in emu eggshells, Nature, 385, 241–244.

Morrissey, S. K., Clark, J. F., Bennett, M., Richardson, E., and Stute, M. (2010),

Groundwater reorganization in the Floridan aquifer following Holocene sea-level rise. Nature

Geoscience, 3, 683–687.

Nanson, G., Price, D. and Short, S. (1992). Wetting and drying of Australia over the past 300 ka, Geology, 20, 791–794.

New, M., Lister, D., Hulme, M., and Makin, I. (2002), A high-resolution data set of surface climate over global land areas, Climate Research, 21, 1–25.

Noseck, U. et al. (2009), Carbon chemistry and groundwater dynamics at natural analogue site Ruprechtov, Czech Republic: insights from environmental isotopes, Applied Geochemistry, 24,

1765–1776.

O'Brien, G. R., Kaufman, D. S., Sharp, W. D., Atudorei, V., Parnell, R. A., and Crossey, L. J.

(2006), Oxygen isotope composition of annually banded modern and mid-Holocene travertine and evidence of paleomonsoon floods, Grand Canyon, Arizona, USA, Quaternary Research, 65, 366–379.

O'Neil, J. R., Clayton, R. N., and Mayeda, T. K. (1969), Oxygen isotope fractionation in divalent metal carbonates, The Journal of Chemical Physics, 51, 5547–5558. 202 Otto-Bliesner, B. L., Brady, E. C., Clauzet, G., Tomas, R., Levis, S., and Kothavala, Z.

(2006), Last glacial maximum and Holocene climate in CCSM3, Journal of Climate, 19, 2526–2544.

Partin, J. W., Cobb, K. M., Adkins, J. F., Clark, B., and D. P. Fernandez (2007), Millennial- scale trends in west Pacific warm pool hydrology since the Last Glacial Maximum, Nature, 449, 452–

455.

Patterson, L. J. et al. (2005), Cosmogenic, radiogenic, and stable isotopic constraints on groundwater residence time in the Nubian Aquifer, Western Desert of Egypt, Geochemistry, Geophysics,

Geosystems, 6.

Pausata, F. S., Battisti, D. S., Nisancioglu, K. H., and Bitz, C. M. (2011), Chinese stalagmite

δ18O controlled by changes in the Indian monsoon during a simulated Heinrich event. Nature

Geoscience, 4, 474–480.

Pedro, J. B., Van Ommen, T. D., Rasmussen, S. O., Morgan, V. I., Chappellaz, J., Moy, A.

D., Masson-Delmotte, V. and Delmotte, M. (2011), The last deglaciation: timing the bipolar seesaw.

Climate of the Past, 7, 671–683.

Phillips, F. M., Peeters, L. A., Tansey, M. K., and Davis, S. N. (1986), Paleoclimatic inferences from an isotopic investigation of groundwater in the central San Juan Basin, New Mexico,

Quaternary Research, 26, 179–193.

Plummer, L. N. (1993). Stable isotope enrichment in paleowaters of the southeast Atlantic

Coastal Plain, United States, Science, 262, 2016–2020.

Plummer, L. N., Bexfield, L. M., Anderholm, S. K., Sanford, W. E., and Busenberg, E.

(2004), Hydrochemical tracers in the middle Rio Grande Basin, USA: 1. Conceptualization of groundwater flow, Hydrogeology Journal, 12, 359–388.

203 Powers, L. A., Johnson, T. C., Werne, J. P., Castañeda, I. S., Hopmans, E. C., Sinninghe

Damsté, J. S., and Schouten, S. (2005), Large temperature variability in the southern African tropics since the Last Glacial Maximum, Geophysical Research Letters, 32, L08706.

Rahube, T. B. (2003), Recharge and groundwater resources evaluation of the Lokalane–

Ncojane Basin (Botswana) using numerical modelling. MSc Thesis, International Institute for

Geoinformation Science and Earth Observation, Enschede, Netherlands, 104.

Risi, C., S. Bony, F. Vimeux, L. Descroix, B. Ibrahim, E. Lebreton, I. Mamadou, and B.

Sultan (2008), What controls the isotopic composition of the African monsoon precipitation?

Insights from event-based precipitation collected during the 2006 AMMA field campaign, Geophysical

Research Letters, 35, L24808.

Risi, C., Bony, S., Vimeux, F., and Jouzel, J. (2010a), Water‐stable isotopes in the LMDZ4 general circulation model: Model evaluation for present‐day and past climates and applications to climatic interpretations of tropical isotopic records, Journal of Geophysical Research: Atmospheres, 115,

D12118.

Risi, C., Bony, S., Vimeux, F., Frankenberg, C., Noone, D., and Worden, J. (2010b),

Understanding the Sahelian water budget through the isotopic composition of water vapor and precipitation, Journal of Geophysical Research: Atmospheres, 115, D24110.

Robinson, B. W., and Gunatilaka, A. (1991), Stable isotope studies and the hydrological regime of sabkhas in southern Kuwait, Arabian Gulf, Sedimentary Geology, 73, 141–159.

Roboucas, A. C., and Santiago, M. F. (1989), 14C analyses of groundwater from the Botucatu aquifer system in Brazil, Radiocarbon, 31, 926–935.

204 Roden, J. S., and Ehleringer, J. R. (1999), Hydrogen and oxygen isotope ratios of tree-ring cellulose for riparian trees grown long-term under hydroponically controlled environments, Oecologia,

121, 467–477.

Rozanski, K. (1985). Deuterium and oxygen-18 in European groundwaters—links to atmospheric circulation in the past, Chemical Geology 52, 349–363.

Sagawa, T., Yokoyama, Y., Ikehara, M., and Kuwae, M. (2011), Vertical thermal structure history in the western subtropical North Pacific since the Last Glacial Maximum, Geophysical Research

Letters, 38, L00F02.

Salati, E., Menezes Leal, J., and Mendes Campos, M. (1974), Environmental isotopes used in a hydrogeological study of northeastern Brazil, In: Isotope techniques in groundwater hydrology

1974, Vol. I. 379–398.

Salati, E., Dall'Olio, A., Matsui, E., and Gat, J. R. (1979), Recycling of water in the Amazon basin: an isotopic study, Water Resources Research, 15, 1250–1258.

Samborska, K., Różkowski, A., and Małoszewski, P. (2013), Estimation of groundwater residence time using environmental radioisotopes (14C, T) in carbonate aquifers, southern Poland,

Isotopes in Environmental and Health Studies, 49, 73-97.

Samuels-Crow, K. E., Galewsky, J., Hardy, D. R., Sharp, Z. D., Worden, J., and Braun, C.

(2014), Upwind convective influences on the isotopic composition of atmospheric water vapor over the tropical Andes, Journal of Geophysical Research – Atmospheres, 119, 7051–7063,

Schefuß, E., Kuhlmann, H., Mollenhauer, G., Prange, M., and Pätzold, J. (2011), Forcing of wet phases in southeast Africa over the past 17,000 years, Nature, 480, 509–512.

205 Schlegel, M. E., Mayo, A. L., Nelson, S., Tingey, D., Henderson, R., and Eggett, D. (2009),

Paleo-climate of the Boise area, Idaho from the last glacial maximum to the present based on groundwater δ2H and δ18O compositions, Quaternary Research, 71, 172–180.

Schrag, D. P., Hampt, G.,,and Murray, D. W. (1996), Pore fluid constraints on the temperature and oxygen isotopic composition of the glacial ocean, Science, 272, 1930–1932.

Schrag, D. P., Adkins, J. F., McIntyre, K., Alexander, J. L., Hodell, D. A., Charles, C. D., and

McManus, J. F. (2002), The oxygen isotopic composition of seawater during the Last Glacial

Maximum, Quaternary Science Reviews, 21, 331–342.

Shah, A. M., Morrill, C., Gille, E. P., Gross, W. S., Anderson, D. M., Bauer, B. A. Buckner,

R. and Hartman, M. (2013), Global speleothem oxygen isotope measurements since the Last Glacial

Maximum, Dataset papers in Geosciences, 9 pp.

Shakun, J. D., and Carlson, A. E. (2010), A global perspective on Last Glacial Maximum to

Holocene climate change, Quaternary Science Reviews, 29, 1801–1816.

Shakun, J. D., Burns, S. J., Fleitmann, D., Kramers, J., Matter, A., and Al-Subary, A. (2007),

A high-resolution, absolute-dated deglacial speleothem record of Indian Ocean climate from Socotra

Island, Yemen, Earth and Planetary Science Letters, 259, 442–456.

Siegel, D. I. (1991), Evidence for dilution of deep, confined ground water by vertical recharge of isotopically heavy Pleistocene water, Geology, 19, 433–436.

Stadler, S., Geyh, M. A., Ploethner, D., and Koeniger, P. (2012), The deep Cretaceous aquifer in the Aleppo and Steppe basins of Syria: assessment of the meteoric origin and geographic source of the groundwater, Hydrogeology Journal, 20, 1007–1026.

206 Stotler, R. L., Frape, S. K., Ruskeeniemi, T., Pitkänen, P., and Blowes, D. W. (2012), The interglacial–glacial cycle and geochemical evolution of Canadian and Fennoscandian Shield groundwaters, Geochimica et Cosmochimica Acta, 76, 45–67.

Stute, M., and Deak, J. (1989), Environmental isotope study 14C, 13C, 18O, D, noble gases on deep groundwater circulation systems in Hungary with reference to paleoclimate, Radiocarbon, 31,

902–918.

Stute, M., Clark, J. F., Schlosser, P., Broecker, W. S., and Bonani, G. (1995a), A 30,000 yr continental paleotemperature record derived from noble gases dissolved in groundwater from the

San Juan Basin, New Mexico, Quaternary Research, 43, 209–220.

Stute, M., Forster, M., Frischkorn, H., Serejo, A., Clark, J. F., Schlosser, P., Broecker, W. S., and Bonani, G. (1995b), Cooling of tropical Brazil (5°C) during the Last Glacial Maximum, Science,

269, 379–379.

Sukhija, B. S., Reddy, D. V., and Nagabhushanam, P. (1998), Isotopic fingerprints of paleoclimates during the last 30,000 years in deep confined groundwaters of Southern India,

Quaternary Research, 50, 252–260.

Sultan, M., Sturchio, N., Hassan, F. A., Hamdan, M. A. R., Mahmood, A. M., Alfy, Z. E., and Stein, T. (1997), Precipitation source inferred from stable isotopic composition of Pleistocene groundwater and carbonate deposits in the Western desert of Egypt, Quaternary Research, 48, 29–37.

Swarzenski, P. W., Baskaran, M., Rosenbauer, R. J., Edwards, B. D., and Land, M. (2013), A combined radio–and stable–isotopic study of a California coastal aquifer system. Water, 5, 480–504.

207 Tamers M. A. (1967), Radiocarbon ages of groundwater in an arid zone unconfined aquifer,

In: Stout G. E. (ed.), Isotope techniques in the hydrologic cycle. American Geophysical Union

Monograph 11, 143–152.

Thatcher, L., Rubin, M., and Brown, G. F. (1961), Dating desert groundwater, Science, 134,

105–106.

Thompson, L. G., Mosley-Thompson, E., Davis, M. E., Lin, P. N., Henderson, K. A., Cole-

Dai, J., Bolzan, J. F., and Liu, K. B. (1995), Late glacial stage and Holocene tropical ice core records from Huascaran, Peru. Science, 269, 46–50.

Thompson, L. O., Yao, T., Davis, M. E., Henderson, K. A., Mosley-Thompson, E., Lin, P.

N., Beer, J., Synal, H.-A., Cole-Dai, J., and Bolzan, J. F. (1997), Tropical climate instability: The last glacial cycle from a Qinghai-Tibetan ice core. Science, 276, 1821–1825.

Thompson, L. G. et al. (1998), A 25,000-year tropical climate history from Bolivian ice cores.

Science, 282, 1858–1864.

Tierney, J. E., Russell, J. M., Huang, Y., Damsté, J. S. S., Hopmans, E. C., and Cohen, A. S.

(2008), Northern hemisphere controls on tropical southeast African climate during the past 60,000 years, Science, 322, 252–255.

Tierney, J. E., Smerdon, J. E., Anchukaitis, K. J., and Seager, R. (2013), Multidecadal variability in East African hydroclimate controlled by the Indian Ocean, Nature, 493, 389-392.

Varsányi, I., Palcsu, L., and Kovács, L. Ó. (2011), Groundwater flow system as an archive of palaeotemperature: Noble gas, radiocarbon, stable isotope and geochemical study in the Pannonian

Basin, Hungary, Applied Geochemistry, 26, 91–104.

208 Veizer, J., et al. (1999). 87Sr/86Sr, δ13C and δ18O evolution of Phanerozoic seawater. Chemical

Geology, 161, 59–88.

Vinther, B. M. et al. (2006), A synchronized dating of three Greenland ice cores throughout the Holocene, Journal of Geophysical Research, 111, D13102.

Vinther, B. M., Clausen, H. B., Fisher, D. A., Koerner, R. M., Johnsen, S. J., Andersen, K.

K., Dahl-Jensen, D., Rasmussen, S. O., Steffensen, J. P., and Svensson, A. M. (2008), Synchronizing ice cores from the Renland and Agassiz ice caps to the Greenland Ice Core Chronology, Journal of

Geophysical Research: Atmospheres, 113, D08115.

von Grafenstein, U., Erlenkeuser, H., Brauer, A., Jouzel, J., and Johnsen, S. J. (1999), A Mid-

European Decadal Isotope-Climate Record from 15,500 to 5000 Years B.P., Science, 284, 1654–1657.

Vuille, M., and Werner, M. (2005), Stable isotopes in precipitation recording South American summer monsoon and ENSO variability: observations and model results, Climate Dynamics, 25, 401–

413.

Wagner, J. D. M., Cole, J. E., Beck, J. W., Patchett, P. J., Henderson, G. M., and Barnett, H.

R. (2010), Moisture variability in the southwestern United States linked to abrupt glacial climate change, Nature Geoscience 3, 110–113.

Wang, Y. J., Cheng, H., Edwards, R. L., An, Z. S., Wu, J. Y., Shen, C-C., and Dorale, J. A.

(2001), A high-resolution absolute-dated late pleistocene monsoon record from Hulu Cave, China,

Science, 294, 2345–2348.

Wang, X. F. et al. (2007), Millennial-scale precipitation changes in southern Brazil over the past 90,000 years, Geophysical Research Letters, 34, L23701.

209 Werner, M., Langebroek, P. M., Carlsen, T., Herold, M., and Lohmann, G. (2011), Stable water isotopes in the ECHAM5 general circulation model: Toward high‐resolution isotope modeling on a global scale, Journal of Geophysical Research: Atmospheres, 116, D15109.

Weyhenmeyer, C. E., Burns, S. J., Waber, H. N., Aeschbach-Hertig, W., Kipfer, R., Loosli,

H. H., and Matter, A. (2000), Cool glacial temperatures and changes in moisture source recorded in

Oman groundwaters, Science, 287, 842–845.

Weyhenmeyer, C. E., Burns, S. J., Waber, H. N., Macumber, P. G., and Matter, A. (2002),

Isotope study of moisture sources, recharge areas, and groundwater flow paths within the eastern

Batinah coastal plain, Sultanate of Oman, Water Resources Research, 38, 1184.

Williams J.W. (2003), Variations in tree cover in North America since the last glacial maximum, Global Planetary Change, 35, 1–23.

Williams, M., Prescott, J. R., Chappell, J., Adamson, D., Cock, B., Walker, K. and Gell, P.

(2001). The enigma of a late Pleistocene wetland in the Flinders Ranges, South Australia, Quaternary

International, 83-85, 129–144.

Williams, P. W., Neil, H. L., and Zhao, J. X. (2010), Age frequency distribution and revised stable isotope curves for New Zealand speleothems: palaeoclimatic implications. International Journal of

Speleology, 39, 99–112.

Winckel, A., Marlin, C., Dever, L., Morel, J., Morabiti, K., Makhlouf, M. B., and Chalouan,

A. (2002), Recharge altitude estimation of thermal springs using stable isotopes in Morocco, Comptes

Rendus Geoscience, 334, 469–474.

210 Yang, Y., Yuan, D., Cheng, H., Zhang, M., Qin, J., Lin, Y., XiaoYan, Z. and Edwards, R. L.

(2010), Precise dating of abrupt shifts in the Asian Monsoon during the last deglaciation based on stalagmite data from Yamen Cave, Guizhou Province, China, Science China Earth Sciences, 53, 633–641.

Yechieli, Y., Kafri, U., and Sivan, O. (2009), The inter–relationship between coastal sub– aquifers and the Mediterranean Sea, deduced from radioactive isotopes analysis. Hydrogeology Journal,

17, 265–274.

Yoshimura, K., Oki, T., Ohte, N., and Kanae, S. (2003), A quantitative analysis of short‐term

18O variability with a Rayleigh‐type isotope circulation model, Journal of Geophysical Research:

Atmospheres, 108, 4647.

Yuan, D. et al. (2004), Timing, Duration, and Transitions of the Last Interglacial Asian

Monsoon, Science, 23, 575–578.

Zongyu, C., Jixiang, Q., Jianming, X., Jiaming, X., Hao, Y., and Yunju, N. (2003),

Paleoclimatic interpretation of the past 30 ka from isotopic studies of the deep confined aquifer of the North China plain, Applied Geochemistry, 18, 997–1009.

Zuber, A., Weise, S. M., Osenbrück, K., Pajnowska, H., and Grabczak, J. (2000), Age and recharge pattern of water in the Oligocene of the Mazovian basin (Poland) as indicated by environmental tracers, Journal of Hydrology, 233, 174–188.

Zuber, A., Weise, S. M., Motyka, J., Osenbrück, K., and Różański, K. (2004), Age and flow pattern of groundwater in a Jurassic limestone aquifer and related Tertiary sands derived from combined isotope, noble gas and chemical data, Journal of Hydrology, 286, 87–112.

211 CHAPTER 4 — THE ISOTOPE HYDROLOGY OF UGANDA

4.1 Abstract

In the final chapter of this dissertation I present the results of a comprehensive sampling campaign and stable O and H isotopic investigation of Ugandan rivers, lakes, wetlands and groundwaters. Surface- and ground-waters in Uganda have δ18O values ranging between −4.0 ‰ and

+8.7 ‰, δ2H values ranging between −13.2 ‰ and +55.5 ‰, and deuterium excess values ranging between −17 ‰ and +22‰. The highest δ18O and δ2H values and lowest deuterium excess values are found in Ugandan lakes; whereas, the lowest δ18O and δ2H values and highest deuterium excess values are found in springs and river waters sourced from the Rwenzori mountains of southwest

Uganda. I analyze the isotopic composition of lake waters using a stable-isotope-mass-balance to calculate the fraction of evaporation as a proportion of water inputs to 24 lakes. I show that a sample of lake water analyzed for O and H isotopes, coupled to the application of a stable-isotope-mass- balance, can rapidly delineate well flushed (low evaporation/input ratio) and terminal

(evaporation/input ratio of close to 1) lake systems.

4.2 Introduction

In Uganda, 70% of the 35 million people living there have access to an improved water source, ranking Uganda 148 out of 179 nations reporting in 2010 (Millennium Development Goals

Indicators). Groundwater is the primary drinking water source for 80% of Ugandans and cultivated lands cover one third of the country, highlighting the importance of agriculture (usually rain-fed) to the Ugandan economy. Lake Victoria, located in southeastern Uganda, sustains a commercial fishery, supports over-lake transportation, feeds municipal water supplies and generates hydroelectric power at the Kiira and Nalubaale power stations near Jinja.

Uganda is a landlocked nation. It is situated in the humid tropics of east Africa, bordered by

Sudan (north), Kenya (east), Tanzania (south), Rwanda (southwest) and the Democratic Republic of

212 the Congo (west). The country is constrained between 1°S and 4°N latitude and 29°E to 35°E longitude and covers a total of 240,000 km2. Land surface elevations range between 600 to 5,200 meters above sea level, with 80% of the country resting between 900 and 1,500 meters above sea level. The majority of land surfaces have been converted to rain-fed croplands. Natural Ugandan vegetation that remains intact includes closed canopy forests of the relatively-wet southwest and open shrublands of the relatively-arid northwest. Regolith depths can extend to depths of 30m below groundwater surfaces, and the bedrock geology includes endogenous granulites in central Uganda and metasedimentary rocks along the western arm of the east African rift system, located in western

Uganda (Taylor and Howard, 1998a; 1998b; 2000).

The hydrography of Uganda ranges from flashy systems in the steep mountainous systems of western Uganda, to slow-flowing and vast wetlands in the subdued relief in central Uganda, to semi-arid ephemeral flow systems on the westward slopes of Mt. Kenya and other mountains in northeastern Uganda. Annual rainfall ranges from minimums of ~700 mm per year in the northeast to maximums of ~1,500 mm per year in the southwest. Lake Victoria – the second largest area of fresh surface water on Earth – borders Uganda’s southeastern margins and its outflow generates the headwaters of the White Nile. Uganda’s drainage system is dominated by two flow systems: (i) drainage into Lake Victoria and Lake Kyoga, and (ii) drainage into the rift lakes of western Uganda

(i.e., Lakes George, Edward, and Albert). The two drainage systems converge at the northern margin of Lake Albert before flowing northward into .

Stable isotope investigations of Ugandan waters have targeted improved knowledge of groundwaters (Taylor and Howard, 1996; 1998a; Tindimugaya et al., 2007), surface waters (Russell and Johnson, 2006) and geothermal systems (Kato, 2000; Bahati et al., 2005). Groundwater stable isotopes have revealed that recharge occurs almost exclusively during the rainy season (April to

October) and that recharge is at a maximum during high intensity rainfall events (Taylor and Howard

213 1996; 1998)). Coupled hydrology-geomorphology investigations have shown that the regolith and subsurface bedrock are hydraulically-linked, that the regolith is more permeable than bedrock aquifers, and that this weathered mantle hosts a more active hydrosphere than underlying fractured bedrock systems (Taylor and Howard, 1996; 1998a; 2000). Annual groundwater recharge rates are spatially variable. Paired watershed studies show that recharge at each catchment varies ten-fold with recharge rates of 20 to 200 mm/year in each respective catchment (Taylor and Howard, 1998a).

Groundwater ages calculated using chlorofluorocarbons and tritium show that modern groundwater with a mean age of less than 50 years exists at depths of more than 60 meters below ground level, suggesting that certain shallow aquifers are well-flushed (Tindimugaya et al., 2007). Other isotope based investigations have quantified over-lake evaporation from Lake Edward using a stable isotope based approach (Russell and Johnson, 2006) and assessed geothermal activity (Kato, 2000; Bahati et al., 2005).

The primary objective of this study is to quantify water fluxes into and out of Ugandan surface waters via a stable isotope mass balance.

4.3 Dataset and methods

4.3.1 Sample collection and analysis Samples of water were collected in high-density polyethylene bottles over a three week sampling campaign in July of 2013. Water samples were collected from rivers, lakes, wetlands, springs and groundwater wells throughout Uganda. Two tiers of samples were collected: tier 1, where waters were sampled for analysis of 18O/16O and 2H/1H ratios, and tier 2, where waters were sampled for

18O/16O and 2H/1H ratios, concentrations of Ca2+, Mg2+, K+, Na+, Cl-, SO42- and 43 other solutes.

Tier 2 water samples (n = 45) were filtered through a 0.45 micron filter in the field. One sample was preserved with ultrapure nitric acid (major cations, trace metals and uranium series elements) and another was left without acidification (major anions). Chemical analyses were completed at the

214 University of New Mexico’s Analytical Chemistry Laboratory. Stable oxygen and hydrogen isotopic compositions of water samples were analyzed at the University of New Mexico using a Picarro

L1102-i liquid water analyzer.

The three week field campaign led to the collection of 225 water samples. Because of the short duration of our field trip we specifically targeted water bodies expected to integrate prolonged residence times such as lakes (n = 36), groundwater wells (n = 75) and springs (n = 13). We also sampled streams (n = 67), tap water (n = 22) and swamps (n = 12) throughout Uganda where sampling opportunities arose.

4.3.2 Surface waters – evaporation modelling We use the stable O and H isotopic data of lake waters to calculate evaporation losses from each system. The approach taken has been described in numerous studies and readers are directed to these earlier works for additional descriptions (Zuber, 1983; Gonfiantini, 1986; Gat et al., 1996;

Gibson, 2002; Gibson and Edwards, 2002; Gibson et al., 1996; 1998; 2002; Froehlich, 2000; Russell and Johnson, 2006; Horita et al., 2008; Yi et al., 2008; Brock et al., 2009; Turner et al., 2010; 2014).

The calculation of lake water balances via a stable-isotope-based approach couples a hydrologic

(Equation 4.1) and isotopic (Equation 4.2) mass balance under an assumption of steady state:

퐼 = 퐸 + 푄 Equation 4.1

퐼훿퐼 = 퐸훿퐸 + 푄훿푄 Equation 4.2 where I, E and Q are the fluxes of water entering the lake (I), evaporation from the lake (E) and liquid outflow from the lake via surface or groundwater discharges (Q), and δ denotes the isotopic composition of each flux. Combining equations 4.1 and 4.2 yields an estimate of the evaporation flux as a proportion of water inputs to each lake (i.e., evaporation/input ratio: E/I; Equation 4.3)

215 퐸 훿 −훿 = 퐼 퐿푎푘푒 Equation 4.3 퐼 훿퐸−훿퐿푎푘푒 assuming liquid outflows from the lake have the same isotopic composition as the bulk lake (i.e., well mixed assumption: δQ = δLake). The isotopic composition of water inputs to lakes is estimated as the intercept of a regression of lake O and H isotopic compositions (“local evaporation line: LEL;”

Figure 4-1) and a regression of Ugandan rainfall (“meteoric water line: MWL;” Figure 4-1). We find that the isotopic composition of input waters to each lake is between δ18O values of −2.1 ‰ to +0.0

‰ and δ2H input values between −3.5 ‰ and +9.8 ‰ (maximum and minimum intercepts of 95th percent confidence interval regressions of meteoric waters (MWL) and lake water (LEL)). Annual evaporation rates (i.e., mm/year, rather than E/I ratios) can be calculated by rearranging equations

4.1 and 4.2 in cases where the liquid outflow from the lake is gauged (Jasechko et al., 2014):

훿 −훿 퐸 = 푄 × 퐼 퐿푎푘푒 Equation 4.4 훿퐸−훿퐼

The isotopic composition of evaporate is estimated applying an evaporation model (Craig and Gordon, 1965):

∗ ∗ (훿퐿푎푘푒−[훼푙∙푣 −1])/훼푙∙푣 −ℎ훿퐴−(퐶푘[1−ℎ]) 훿퐸 = Equation 4.5 1−ℎ+(퐶푘[1−ℎ])

∗ where 훼푙∙푣 represents a temperature-dependent equilibrium isotope fractionation factor (Horita and

Wesolowski, 1994), h represents the relative humidity near to the lake surface (derived from New et

∗ al., 2002), δA represents the isotopic composition of the atmosphere (calculated as δA = δP – (훼푙∙푣 −

1; Gibson et al., 2002) and CK is a constant that describes kinetic isotope effect during evaporation

(CK is 13.7 to 20.7 for δ18O-based calculations and 7.5 to 16.1 for δ2H based calculations; Jasechko et al., 2014).

216 4.4 Results

4.4.1 Stable O and H isotopic composition of Ugandan waters

The isotopic composition of Ugandan surface- and ground-waters sampled in this study range from −4.0 ‰ to +8.7 ‰ in δ18O, −13.2 ‰ to +55.5 ‰ in δ2H, and −16.7 ‰ to +21.9 ‰ in deuterium excess (Figure 4-1; Table 4-1; Table 4-2). Monthly precipitation samples (n = 267) collected at Entebbe (Uganda, n = 182), Soroti (Uganda, n = 11), Jinja, (Uganda, n = 27), Masaka

(Uganda, n = 20), Wobulenzi (Uganda, n = 8) and Kericho (western Kenya, n = 19) between 1960 and 2010 by the International Atomic Energy Agency (e.g., Araguás-Araguás et al., 2002) range from

−11.6 ‰ to +11.4 ‰ in δ18O, −81.2 ‰ to +69.0 ‰ in δ2H, and −22.1 ‰ to +27.1 ‰ in deuterium excess, and plot along a regression – herein the “Ugandan meteoric water line” – of δ2H =

7.21(±0.13)×δ18O + 10.76(±0.40) (uncertainties are standard error of regression; Figure 4-1; Table

1). Water sampling locations are shown in Figure 4-2.

217

Figure 4-1. The O and H isotopic composition of Ugandan waters. Different symbols mark unique water types, including groundwaters (squares), rivers (circles) and lakes (triangles). Black lines mark linear regressions of meteoric waters (MWL) and lakes (LEL; grey funnel plot marks the 95th percent confidence interval of regressions).

Lakes have the highest δ18O and δ2H values and the lowest deuterium excess values of each of the sample groups. Lakes plot near to groundwater in some cases, but also plot along a trajectory

“beneath” meteoric waters in δ18O-δ2H space (Figure 4-1). A regression of the lake data gives a

δ2H/δ18O slope of 5.13±0.13, significantly shallower than a regression of meteoric waters (δ2H/δ18O slope of 7.21±0.13).

River and wetlands have oxygen and hydrogen isotopic compositions that are similar to

Ugandan rainfall in most cases. A subset of river samples have deuterium excess values of less than zero and plot close to lakes sampled in this study (10 of 67 river water samples, 15%). Some river

218 samples having deuterium excess values of less than zero were sampled downstream of large lakes

(e.g., outflows of Lakes Victoria and Lake Kachera). Perennial wetlands are common in central

Uganda near to Lake Kyoga. Isotopic analysis of wetland samples shows that most samples plot along the Ugandan meteoric water line near to groundwater samples.

Groundwater samples collected from wells (n = 75) plot near to the Ugandan rainfall in most cases in δ2H-δ18O space. Groundwater samples have deuterium excess values that are similar to

Ugandan rainfall (average deuterium excess values of 10.2 ‰ for groundwater and 12.3 ‰ for rainfall). Six of our 75 groundwater samples (8 % of all groundwater samples) have a deuterium excess value of less than zero, similar to Ugandan lakes (average deuterium excess of −1.3 ‰; Table

4-1).

Springs and tap waters have δ18O and δ2H values that fall within the range of precipitation in

Uganda. Springs have a deuterium excess value that is similar to rainfall (average of 15.5 ‰). The lowest δ18O value observed in our dataset (−4.0 ‰) is a hot spring sample from the foothills of the

Rwenzori Mountains collected at an elevation of 1600 meters above sea level. The sources of tap water were unknown in most cases. Tap waters are found to have stable O and H isotopic compositions that fall within the range of groundwaters and surface waters collected in this study.

219 Table 4-1. The isotopic composition of Ugandan waters

Sample type n δ18O (‰) δ2H (‰) d-excess (‰) Avg. s.d. Avg. s.d. Avg. s.d. Groundwater 75 −0.5 1.9 +6.1 10.2 +10.2 6.0 Lakes 36 +3.4 2.6 +25.9 13.3 −1.3 7.6 Rivers 67 −0.5 2.4 +5.8 12.7 +10.0 7.6 Springs 13 −2.0 1.7 −0.3 9.3 +15.5 4.7 Swamp water 12 −0.3 2.0 +8.2 12.6 +10.5 6.5 Tap water 22 +0.1 2.5 +8.8 12.8 +7.9 7.5 Rainfall * 267 −2.0 2.4 −3.6 17.7 +12.3 5.3

* rainfall statistics from the combined precipitation datasets collected at Entebbe, Soroti, Jinja,

Masaka, Wobulenzi and Kericho; data obtained from the International Atomic Energy Agency’s

Water Resources Programme: www.iaea.org/water.

220

Figure 4-2. Locations of water samples collected in Uganda: groundwater (yellow squares), crater lakes (red circles), other lakes (blue circles), rivers (blue triangles), springs (black triangles), wetlands (green circles), tap waters (small dots).

221 Table 4-2. The isotopic composition and electrical conductivity of Ugandan water samples

Lat. Lon. Alt. EC T δ18O δ2H ID Type d−excess [°] [°] [m] [μs/cm] [⁰C] [‰ SMOW] [‰ SMOW] T01 Tap 0.1 32.5 399 105 25.6 −0.24 5.9 7.9 T02 Tap 0.3 32.6 1187 132 26.3 1.88 16.6 1.6 T03 Tap 0.0 32.0 1166 −3.15 −9.9 15.4 T04 Tap −0.3 31.8 1250 −2.93 −10.2 13.2 T05 Tap 0.1 32.5 1188 3.53 26.7 −1.6 T06 Tap −0.6 31.0 1279 0.96 10.1 2.5 T07 Tap −0.5 30.7 1466 144 26.4 −1.64 0.6 13.7 T08 Tap −1.3 30.0 1888 265 17.5 3.94 27.9 −3.7 T09 Tap −0.2 30.0 980 −2.64 −3.5 17.6 T10 Tap 0.6 31.4 1286 173 29.4 0.05 6.9 6.5 T11 Tap 0.4 32.7 1196 129 24.9 3.16 25.3 0.1 T12 Tap 0.5 33.3 1195 103 26.3 3.54 26.9 −1.4 T13 Tap 0.8 33.7 1122 132 27.1 −1.83 −3.7 10.9 T14 Tap 1.1 34.2 1126 133 29.4 −1.57 2.5 15.1 T15 Tap 1.1 34.2 1126 130 26.3 −1.62 2.6 15.5 T16 Tap 0.4 32.7 1196 129 24.9 3.38 25.2 −1.9 T17 Tap 0.4 33.1 1245 106 25.0 4.06 28.8 −3.7 T18 Tap 2.3 31.6 636 943 32.1 −0.63 9.6 14.6 T19 Tap 1.4 32.3 1088 −2.17 −0.4 17.0 T20 Tap 1.6 31.7 1191 −1.42 2.1 13.5 T21 Tap 1.2 32.4 1084 53 26.3 −1.04 2.1 10.5 T22 Tap 0.2 30.1 964 −1.17 1.3 10.7 G01 Grdwtr. −0.4 31.5 1263 350 25.0 −0.64 4.7 9.8 G02 Grdwtr. −0.5 31.0 1285 258 25.1 3.44 28.1 0.6 G03 Grdwtr. −0.7 30.2 1504 71 18.0 −2.00 0.5 16.5 G04 Grdwtr. −1.0 30.2 1433 126 21.8 −1.64 −0.8 12.4 G05 Grdwtr. −0.6 29.8 1031 102 29.0 −0.98 3.6 11.4 G06 Grdwtr. −0.4 29.9 995 88 22.8 0.35 11.6 8.9 G07 Grdwtr. 0.5 30.1 1650 −2.50 −3.3 16.7 G08 Grdwtr. 1.2 34.2 1166 255 27.2 −1.32 2.0 12.5 G09 Grdwtr. 1.9 34.6 1146 726 28.4 0.21 13.8 12.1 G10 Grdwtr. 2.5 34.6 1259 652 25.5 0.08 9.5 8.8 G11 Grdwtr. 0.5 33.4 1148 239 26.4 −2.35 −5.1 13.6 G12 Grdwtr. 0.6 33.5 1144 116 27.6 3.95 29.1 −2.5 G13 Grdwtr. 0.8 33.6 1085 171 29.0 −1.79 −2.8 11.5 G14 Grdwtr. 1.2 34.3 1117 198 22.7 −1.45 2.1 13.8

222 Lat. Lon. Alt. EC T δ18O δ2H ID Type d−excess [°] [°] [m] [μs/cm] [⁰C] [‰ SMOW] [‰ SMOW] G15 Grdwtr. 1.7 34.6 1144 723 28.1 −1.44 −0.4 11.1 G16 Grdwtr. 2.5 34.7 1407 718 23.6 −1.08 4.7 13.3 G17 Grdwtr. 2.4 34.5 1205 491 27.1 0.27 5.5 3.4 G18 Grdwtr. 2.1 34.2 1180 1060 26.4 −0.40 5.4 8.5 G19 Grdwtr. 2.0 34.1 1085 394 27.5 −0.23 7.0 8.9 G20 Grdwtr. 1.9 34.0 1060 619 27.5 −0.83 4.0 10.6 G21 Grdwtr. 1.9 34.0 1099 624 26.2 −0.68 6.3 11.7 G22 Grdwtr. 1.9 33.8 1054 187 27.9 −1.79 −2.3 11.9 G23 Grdwtr. 1.9 33.8 1078 129 27.3 3.87 23.0 −8.0 G24 Grdwtr. 1.8 33.5 1067 179 27.9 −1.15 4.0 13.2 G25 Grdwtr. 1.9 33.2 1098 233 26.8 −1.29 3.6 13.9 G26 Grdwtr. 2.0 33.1 1047 371 26.8 −0.11 7.6 8.4 G27 Grdwtr. 2.1 32.9 1066 147 27.3 0.12 10.1 9.2 G28 Grdwtr. 2.2 32.3 1041 100 27.2 1.02 24.1 15.9 G29 Grdwtr. 2.2 32.3 1045 189 28.3 −1.41 2.6 13.9 G30 Grdwtr. 2.3 34.3 1175 1390 26.1 −1.24 2.8 12.7 G31 Grdwtr. 1.9 34.0 1099 178 25.4 −1.47 −3.2 8.6 G32 Grdwtr. 1.8 33.6 1075 184 28.1 −1.51 −0.9 11.2 G33 Grdwtr. 1.9 33.3 1129 152 27.6 −1.71 −4.2 9.5 G34 Grdwtr. 1.9 33.1 1055 146 27.9 −3.09 −5.8 18.9 G35 Grdwtr. 2.0 33.0 1065 460 22.9 −0.94 5.0 12.5 G36 Grdwtr. 2.3 32.4 1055 101 27.2 −0.74 5.6 11.5 G37 Grdwtr. 2.5 32.4 1067 133 27.2 −0.96 3.0 10.6 G38 Grdwtr. 2.6 32.4 1073 188 26.5 −0.68 7.3 12.7 G39 Grdwtr. 2.7 32.3 1090 161 25.8 −0.75 5.3 11.4 G40 Grdwtr. 2.8 32.2 1080 293 25.1 1.12 2.5 −6.5 G41 Grdwtr. 2.6 31.8 959 160 27.2 −0.77 3.3 9.4 G42 Grdwtr. 2.8 32.2 1096 153 25.0 0.11 5.1 4.2 G43 Grdwtr. 2.7 32.2 1084 119 26.4 −1.26 −0.2 9.9 G44 Grdwtr. 2.6 32.0 985 162 26.7 −1.25 0.5 10.6 G45 Grdwtr. 2.6 31.9 962 144 24.0 −0.96 3.5 11.2 G46 Grdwtr. 2.6 31.6 854 232 28.5 −0.74 5.5 11.4 G47 Grdwtr. 1.7 31.3 1056 118 26.1 7.80 50.4 −12.1 G48 Grdwtr. 1.5 31.3 1128 62 24.4 5.03 32.7 −7.5 G49 Grdwtr. 2.2 31.5 719 372 27.2 3.35 25.8 −1.0 G50 Grdwtr. 1.6 31.3 1079 217 27.6 0.48 11.1 7.3 G51 Grdwtr. 1.6 31.3 1090 235 25.6 −0.27 8.9 11.1 G52 Grdwtr. 1.5 31.3 1231 34 24.7 −1.39 2.0 13.1 G53 Grdwtr. 1.6 31.8 1191 194 23.0 −1.42 2.1 13.5 223 Lat. Lon. Alt. EC T δ18O δ2H ID Type d−excess [°] [°] [m] [μs/cm] [⁰C] [‰ SMOW] [‰ SMOW] G54 Grdwtr. 0.8 32.5 1117 277 26.2 2.67 25.9 4.6 G55 Grdwtr. 1.6 32.0 1053 327 25.3 −1.72 −1.9 11.8 G56 Grdwtr. 1.4 32.3 1088 82 26.2 −1.17 1.1 10.5 G57 Grdwtr. 0.1 32.5 1149 104 25.6 3.98 32.6 0.8 G58 Grdwtr. −0.6 30.6 1417 128 28.1 −1.42 3.2 14.5 G59 Grdwtr. −0.6 30.4 1469 454 25.1 −1.77 1.5 15.7 G60 Grdwtr. −1.1 29.9 1960 212 18.5 −2.17 0.7 18.0 G61 Grdwtr. −1.0 29.9 1715 305 25.7 −2.62 −2.7 18.3 G62 Grdwtr. 0.3 30.1 1202 1530 28.8 −2.03 −0.5 15.8 G63 Grdwtr. 0.5 31.1 1369 75 24.0 −1.07 4.9 13.4 G64 Grdwtr. 1.0 33.8 1079 911 28.5 −1.97 −0.7 15.0 G65 Grdwtr. 0.8 33.7 1132 −0.69 8.1 13.6 G66 Grdwtr. 1.4 34.3 1109 922 27.9 −0.78 4.8 11.0 G67 Grdwtr. 1.6 34.5 1100 1100 28.5 −2.15 −6.6 10.6 G68 Grdwtr. 2.4 34.4 1194 1000 26.3 −1.76 −1.0 13.1 G69 Grdwtr. 2.8 32.1 1068 587 25.0 −0.68 6.0 11.4 G70 Grdwtr. 2.1 31.5 664 485 31.2 −1.71 −2.9 10.8 G71 Grdwtr. 1.5 31.5 1146 74 25.7 −1.11 4.7 13.5 G72 Grdwtr. 1.4 32.3 1095 408 25.6 −1.45 0.7 12.3 G73 Grdwtr. 0.8 32.5 1113 206 25.7 −1.43 1.9 13.4 G74 Grdwtr. 0.2 30.1 1417 −1.27 4.4 14.5 G75 Grdwtr. 0.2 30.1 1417 −1.27 4.4 14.5 S01 Spring 0.2 32.5 1196 3.27 26.9 0.8 S02 Spring 0.5 30.1 1650 93 18.8 −2.11 −0.3 16.6 S03 Spring 0.5 30.1 1740 59 20.0 −2.82 −4.0 18.6 S04 Spring 0.4 30.2 1126 4100 24.8 −2.41 −1.7 17.6 S05 Spring 0.4 30.2 1126 4750 24.9 −2.41 −1.7 17.6 S06 Spring 0.4 30.2 1126 4750 24.9 −2.19 0.3 17.9 S07 Hot Spring −0.7 30.2 1632 760 54.0 −2.13 0.6 17.6 S08 Hot Spring −0.7 30.2 1632 761 54.0 −1.95 0.2 15.8 S09 Hot Spring −0.9 30.0 1382 1710 54.0 −3.15 −8.1 17.1 S10 Hot Spring 0.5 30.1 1650 1710 66.8 −4.03 −13.2 19.1 S11 Hot spring 2.8 31.9 1019 −2.23 −3.5 14.4 S12 Hot spring 2.8 31.9 1019 −2.23 −3.5 14.4 S13 Hot spring 2.8 31.9 1019 −2.12 −1.1 15.9 L01 Crater Lk. 0.4 30.2 1166 6750 28.1 7.51 46.6 −13.5 L02 Crater Lk. −0.4 30.3 1549 947 28.0 6.64 39.7 −13.4 L03 Crater Lk. 0.4 30.2 1169 1110 29.5 8.44 50.8 −16.7 L04 Crater Lk. −0.4 30.3 1533 376 26.8 −0.29 7.4 9.8 224 Lat. Lon. Alt. EC T δ18O δ2H ID Type d−excess [°] [°] [m] [μs/cm] [⁰C] [‰ SMOW] [‰ SMOW] L05 Crater Lk. −0.4 30.3 1491 413 27.5 5.05 34.1 −6.3 L06 Crater Lk. −0.4 30.3 1574 451 23.6 1.17 15.0 5.7 L07 Crater Lk. −0.5 30.3 1645 320 29.6 4.68 30.3 −7.1 L08 Crater Lk. 0.5 30.3 1471 338 28.1 4.20 30.6 −3.0 L09 Crater Lk. −0.5 30.3 1471 4.20 30.6 −3.0 L10 Crater Lk. 0.5 30.3 1346 508 3.55 27.9 −0.4 L11 Crater Lk. 0.5 30.3 1281 434 3.81 29.4 −1.1 L12 Crater Lk. −0.3 30.1 1337 119 26.2 3.18 25.3 −0.1 L13 Crater Lk. −0.3 30.1 1335 152 26.4 2.74 22.4 0.5 L14 Crater Lk. −0.3 30.1 1344 119 24.9 3.59 26.8 −2.0 L15 Crater Lk. −0.3 30.1 1304 −1.57 −0.9 11.7 L16 Crater Lk. −0.3 30.1 1304 167 28.1 2.45 18.3 −1.2 L17 Crater Lk. −0.3 30.8 1317 343 28.2 5.61 36.1 −8.8 L18 Crater Lk. −0.3 30.1 1268 469 28.2 2.17 20.1 2.7 L19 Crater Lk. −0.3 30.1 1318 745 29.7 7.17 44.7 −12.7 L20 Crater Lk. −0.3 30.1 1318 7.17 44.7 −12.7 L21 Crater Lk. −0.2 30.1 1038 728 25.1 −1.08 4.8 13.4 L22 Crater Lk. −0.3 30.1 1387 298 4.56 28.8 −7.7 L23 Crater Lk. −0.3 30.1 1387 315 27.0 5.48 34.7 −9.1 L24 Lake −0.7 30.9 1259 0.08 8.8 8.2 L25 Lake −0.7 30.9 1259 0.08 8.8 8.2 L26 Lake −0.7 30.9 1259 0.08 8.8 8.2 L27 Lake −0.6 31.0 1281 126 0.57 9.2 4.6 L28 Lake 0.1 32.5 1149 122 25.7 3.62 31.4 2.4 L29 Lake −0.5 31.2 1268 374 26.2 2.91 22.4 −0.9 L30 Lake −1.3 29.9 2122 264 20.0 4.45 32.6 −2.9 L31 Lake −0.3 29.9 914 857 29.0 4.22 34.7 0.9 L32 Lake 0.4 32.0 1168 341 28.5 5.40 36.8 −6.3 L33 Lake −1.3 29.8 1905 2.11 17.4 0.5 L34 Lake −1.2 29.7 1907 1.23 13.8 3.9 L35 Lake −1.3 29.7 1815 1.29 15.0 4.7 L36 Lake 1.8 31.3 618 591 30.0 6.22 45.8 −4.0 Q01 Swamp 0.2 32.3 1178 105 19.7 −1.44 −1.6 9.9 Q02 Swamp 0.9 33.7 1056 139 23.0 −1.41 −1.2 10.1 Q03 Swamp 1.9 33.8 1067 352 25.7 −1.39 −0.3 10.8 Q04 Swamp 1.9 33.2 1090 379 27.7 0.46 11.1 7.4 Q05 Swamp 2.0 34.1 1066 139 21.9 −2.44 −3.5 16.0 Q06 Swamp 2.0 33.0 1036 160 27.9 1.51 31.2 19.1 Q07 Swamp 2.8 31.9 1019 −2.32 −5.1 13.5 225 Lat. Lon. Alt. EC T δ18O δ2H ID Type d−excess [°] [°] [m] [μs/cm] [⁰C] [‰ SMOW] [‰ SMOW] Q08 Pond −0.6 30.3 1539 32 25.7 1.39 13.3 2.1 Q09 Pond −0.6 29.7 953 590 20.0 −1.16 0.4 9.7 Q10 Pond 2.2 34.3 1185 75 21.3 −0.24 11.7 13.6 Q11 Pond 2.3 34.4 1168 375 19.1 −1.89 1.0 16.1 Q12 Pond 0.2 32.4 1205 109 18.3 −1.13 1.2 10.2 R01 Major river 2.3 31.6 634 4.64 32.7 −4.5 R02 Major river 0.0 32.0 1197 117 21.1 −0.91 4.5 11.8 R03 Major river −1.4 30.0 1958 142 −1.41 −2.5 8.7 R04 Major river 2.5 31.5 624 2410 −2.56 −9.2 11.3 R05 Major river 1.7 32.1 1044 127 3.36 26.0 −0.9 R06 River −0.5 31.2 1265 3.00 14.7 −9.3 R07 River −0.5 30.9 1338 101 24.1 1.37 5.3 −5.7 R08 River −0.6 30.6 1417 620 26.1 −0.59 −0.6 4.1 R09 River −0.6 30.3 1524 690 24.3 −1.79 −0.4 13.9 R10 River −0.6 30.2 1515 540 25.2 −1.45 1.8 13.3 R11 River −1.3 30.1 1858 101 20.7 0.26 8.0 5.9 R12 River −1.1 29.9 1969 161 15.0 −2.56 −3.9 16.7 R13 River −1.0 29.9 1932 87 16.3 −2.75 −4.7 17.3 R14 River −0.9 30.0 1444 104 23.4 −1.21 2.1 11.8 R15 River −0.8 29.8 1372 185 24.5 −1.82 −0.7 13.8 R16 River −0.8 29.8 1449 91 24.4 −1.51 1.3 13.4 R17 River −0.7 29.9 1239 92 25.7 −1.61 0.2 13.1 R18 River −1.3 30.0 1907 168 16.2 −2.47 −5.7 14.1 R19 River −1.0 30.0 1722 385 20.9 −2.94 −7.4 16.1 R20 River −0.8 29.8 1372 286 22.5 −0.93 3.9 11.4 R21 River −0.8 29.8 1449 154 24.9 −1.49 1.8 13.7 R22 River −0.7 29.9 1170 243 25.1 −0.58 5.7 10.4 R23 River 0.0 29.8 1042 335 25.7 −1.31 −0.8 9.7 R24 River 0.0 29.9 1088 84 23.5 0.86 10.1 3.2 R25 River −0.6 29.7 969 216 23.6 0.92 14.0 6.6 R26 River −0.5 29.7 918 59 23.4 −2.04 −2.0 14.3 R27 River −0.3 29.9 986 151 23.4 −1.19 0.9 10.4 R28 River −0.2 30.0 963 334 19.1 −2.34 −2.5 16.2 R29 River −0.3 30.1 1304 85 20.9 −3.18 −6.8 18.6 R30 River 0.3 30.1 1107 86 20.0 −2.94 −4.7 18.8 R31 River 0.4 30.2 1067 −2.85 −4.9 17.9 R32 River 0.7 30.3 1520 761 20.6 −1.24 2.6 12.5 R33 River 0.6 30.4 1462 426 20.7 −1.70 0.6 14.2 R34 River 0.6 30.6 1364 −0.55 5.1 9.5 226 Lat. Lon. Alt. EC T δ18O δ2H ID Type d−excess [°] [°] [m] [μs/cm] [⁰C] [‰ SMOW] [‰ SMOW] R35 River 0.4 32.0 1178 41 21.0 −1.64 −1.1 12.0 R36 River 2.4 34.6 1257 316 30.5 4.49 30.3 −5.6 R37 River 0.4 32.8 1120 42 20.0 −1.49 −1.6 10.3 R38 River 1.1 34.2 1126 322 24.9 1.35 16.4 5.6 R39 River 1.8 34.6 1228 294 33.6 −1.35 1.4 12.2 R40 River 2.4 34.6 1268 106 25.6 1.46 32.4 20.7 R41 River 1.9 33.3 1093 132 29.5 −0.90 6.7 13.9 R42 River 2.0 33.0 1065 233 29.1 −2.81 −3.3 19.2 R43 River 2.4 32.4 1063 170 26.8 −1.12 1.4 10.3 R44 River 2.5 34.7 1407 272 18.3 −1.27 1.3 11.5 R45 River 2.4 34.5 1198 180 18.3 3.69 28.4 −1.1 R46 River 1.8 33.5 1058 208 30.1 −2.33 −4.5 14.1 R47 River 1.9 33.4 1043 195 26.4 −1.61 1.7 14.5 R48 River 1.9 33.3 1118 172 27.7 3.23 26.2 0.4 R49 River 1.9 33.2 1084 173 28.3 3.56 25.1 −3.4 R50 River 2.2 32.9 1136 149 24.2 −0.91 4.3 11.6 R51 River 2.8 32.0 1043 63 21.0 −2.30 −6.0 12.4 R52 River 2.6 32.1 990 134 25.1 8.62 55.5 −13.5 R53 River 2.8 32.0 1042 64 20.7 −0.80 5.2 11.6 R54 River 2.3 31.7 716 4.63 32.8 −4.2 R55 River 1.9 31.4 634 108 25.7 −0.55 6.7 11.1 R56 River 1.9 31.5 634 151 25.9 −0.75 5.0 11.0 R57 River 1.7 31.4 991 324 25.5 −0.86 3.3 10.1 R58 River 1.6 31.6 1111 32.3 −0.60 7.0 11.9 R59 River 1.5 32.0 1045 −1.14 2.3 11.4 R60 River 1.6 31.8 1123 151 22.5 −0.39 4.8 7.9 R61 River 1.7 32.1 1044 127 25.1 −1.18 1.9 11.3 R62 River 1.0 32.5 1076 121 22.4 −3.20 −12.2 13.4 R63 River −0.6 30.9 1280 754 21.8 −2.32 −5.2 13.4 R64 River 0.2 30.0 1420 59 17.2 −2.94 −1.7 21.9 R65 River 0.5 30.1 1650 121 15.7 −3.04 −2.7 21.6 R66 River 2.2 32.2 1047 110 27.0 4.87 36.4 −2.6 R67 River −1.2 29.7 1949 −1.05 5.3 13.7

227 4.4.2 Ugandan hydrochemistry

The salinity of Ugandan waters sampled in this study ranges from total dissolved solids values of 25 ppm to 3900 ppm. The highest salinities are observed in hot springs that have total dissolved solids ranging between 248 mg/L and 3851 mg/L. The lowest salinities are observed in rivers (n = 9) that have total dissolved solids of less than 1,000 mg/L. Crater lakes have the largest variability in total dissolved solids, with a minimum of 80 mg/L and a maximum of 2100 mg/L.

Most samples are usually Ca-HCO3 water types (Figure 4-3), with the exception of hot springs that have salinities dominated by Na+-K+ and Cl- and SO42- (Table 4-3).

Contaminants measured in this study include arsenic, fluoride and nitrate. The maximum contaminant levels set by the Environmental Protection Agency are 10 ppb (arsenic), 4 mg/L

(fluoride) and 10 mg/L (nitrate as NO3-N). Most groundwater and river water samples meet the drinking water standards. However, nitrate concentrations exceeding the maximum contaminant level concentration were found in a subset of groundwater samples (e.g., 44 mg/L NO3-N). The origin of the observed high nitrate concentrations is unknown.

228

Figure 4-3. A Piper diagram showing the major cation (lower left ternary) and major anion (lower right ternary) projected onto a combined cation-anion diamond (top-most). Bicarbonate concentrations were calculated using a charge balance because measurements in the field were not possible.

229 4.4.3 Stable isotope-based evaporation fluxes

Our stable isotope based evaporation/inflow ratios range from near-zero (i.e., well-flushed lakes) to near 100% (terminal lakes, where evaporation is the only water loss). The Bunyaruguru and

Kasenda crater lake systems (i.e., lakes with “B” and “K” in title, respectively) have a mixture of both well-flushed lakes and lakes that are terminal (Figure 4-4). This approach, in spite of large uncertainties, shows that terminal and well-flushed lakes can be distinguished on the basis of δ18O and δ2H values.

Figure 4-4. Stable-isotope-based evaporation/input ratios for 24 Ugandan Lakes. Grey bars mark the degree of flushing, with light grey being well-flushed lakes, and dark grey representing lakes having the majority of water losses via evaporation. Lakes entitled with “B” are the Bunyaruguru crater lakes system (south of Lake George) and lakes entitled with “K” are the Kasenda crater lakes system

(north of Lake George). The 18O/16O-based model results are shown here. 230 4.5 Discussion

Stable oxygen and hydrogen isotopic data reveal distinct hydrogeochemical processes across

Uganda. For example, stable oxygen isotopic data and electrical conductivity (a proxy for salinity) show that the processes of mineral dissolution and evapo-concentration can be distinguished.

Mineral dissolution does not alter the oxygen isotopic composition of water and produces an increase in the electrical conductivity of the water. Evapo-concentration, on the other hand, increases both the electrical conductivity and the δ18O value of water (Figure 4-5). The process of evapoconcentration also emerges when examining deuterium excess values and electrical conductivity. The lake having the highest electrical conductivity also has the lowest deuterium excess value, consistent with evapoconcentration. Groundwaters generally have a near-meteoric isotopic composition and a large range of electrical conductivity values, implying that the source of salinity in nearly all groundwaters is likely to be low-temperature mineral dissolution. Lakes, on the other hand, have electrical conductivities that rise with increasing δ18O, highlighting that evapoconcentration of input waters is an important control upon the salinity of certain lakes in Uganda. The process of evapoconcentration is also evidenced by the relationship between deuterium excess and electrical conductivity (Figure 4-6); the lake with the lowest deuterium excess (greatest evaporation/input ratio) also has the highest electrical conductivity (Figure 4-6). The highest

231

Figure 4-5. Oxygen isotopic composition and electrical conductivity of Ugandan lakes (triangles) and groundwaters (squares). Dashed lines mark schematic trajectories for mineral dissolution under low temperatures and evapoconcentration of waters.

The deuterium excess of Ugandan river- and ground-waters is elevated for high latitude samples

(Figure 4-7), primarily collected in southwestern Uganda near to the Rwenzori Mountains. Although the sampling site is different from source water elevations (due to streamflow and groundwater advection), this finding suggests that precipitation in the Rwenzori mountains has a higher deuterium excess than other Ugandan waters, perhaps due to moisture recycling in the Congo basin to the west

(Ndembo et al., 2007), or implicating that snowmelt comprises a portion of these samples because of the known build-up of deuterium excess in snow (Gat et al., 1994).

232

Figure 4-6. Deuterium excess and electrical conductivity of Ugandan lakes (triangles) and groundwaters (squares).

233

Figure 4-7. The deuterium excess and sample elevation of Ugandan Rivers (circles) and groundwaters. High altitude samples (i.e., altitudes above 1,600 meters above sea level) have high deuterium excess values, potentially related to kinetic isotope effects during snow formation or moisture recycling from the Congo basin to the west.

234 Table 4-3. Major ion chemistry of Ugandan waters (units of ppm)

Sample Type Ca2+ K+ Mg2+ Na+ Si Cl- SO42- L-19 Crater lake 11.0 71.6 21.4 83.1 18.2 14.6 0.7 R-63 River 58.5 4.5 23.7 43.8 21.0 19.8 321.1 GW-62 Groundwater 125.2 30.9 41.4 108.3 32.1 74.3 208.8 L-03 Crater lake 9.6 81.5 58.2 95.4 4.7 51.1 0.6 L-23 Crater lake 20.4 11.9 18.2 8.5 15.2 5.7 0.6 L-08 Crater lake 19.6 5.7 25.0 8.7 14.1 3.6 1.3 L-10 Crater lake 34.3 9.7 30.8 11.7 14.8 3.7 0.5 L-11 Crater lake 31.4 6.8 25.9 9.3 12.0 3.4 0.7 GW-64 Groundwater 49.9 10.1 12.7 97.5 29.0 199.3 22.0 GW-65 Groundwater 100.8 2.3 20.5 23.9 17.9 37.8 2.2 GW-66 Groundwater 104.8 4.6 18.7 54.3 17.0 3.9 1.1 GW-67 Groundwater bdl 13.1 bdl 244.0 8.4 11.7 5.4 GW-68 Groundwater 81.6 4.8 19.8 100.1 30.4 16.5 71.6 R-53 River 40.5 5.4 18.5 26.1 28.5 136.4 2.3 GW-46 Groundwater 8.3 34.3 8.4 35.2 31.2 2.5 1.1 R-66 River 6.0 3.0 3.3 9.8 2.4 4.3 0.9 GW-70 Groundwater 26.8 8.1 7.5 57.9 38.3 17.4 9.0 L-36 Lake 9.7 37.3 23.1 62.0 0.5 22.7 19.5 GW-71 Groundwater 3.1 0.6 1.8 6.1 6.3 4.9 1.0 GW-72 Groundwater 27.2 4.0 10.5 30.4 37.7 39.7 33.0 GW-73 Groundwater 13.0 3.0 3.7 18.7 34.9 2.2 1.8 R-04 Major river 32.7 12.4 9.8 340.6 16.5 585.0 n.a. R-05 Major river 5.8 33.6 3.4 20.0 3.1 4.9 0.7 R-03 Major river 7.1 3.5 5.2 11.6 6.4 6.3 4.0 SP-10 Hot Spring 26.7 63.3 4.5 1439.0 24.5 786.8 1434.0 L-32 Lake 17.6 11.3 6.7 32.3 8.3 24.9 0.7 L-28 Lake 8.3 3.3 2.6 10.3 0.8 6.1 3.9 GW-57 Groundwater 5.0 2.6 2.4 9.1 0.8 10.6 11.1 SW-12 Pond 6.7 1.2 3.0 4.3 12.8 1.0 0.5 R-02 Major river 4.9 2.2 2.8 12.4 7.5 6.4 n.a. L-29 Lake 22.1 9.9 11.3 33.5 10.2 49.3 12.7 GW-58 Groundwater 6.7 1.0 3.3 9.8 6.8 4.8 37.2 GW-59 Groundwater 17.6 12.4 9.2 39.1 16.5 69.2 15.9 SP-07 Hot Spring 33.1 11.0 0.3 185.1 32.2 83.8 344.7 SP-08 Hot Spring 33.2 11.1 0.2 185.7 32.4 115.9 342.4 L-30 Lake 18.2 4.3 8.6 16.3 0.4 26.2 1.6

235 Sample Type Ca2+ K+ Mg2+ Na+ Si Cl- SO42- GW-60 Groundwater 15.7 2.0 10.6 6.6 14.2 6.4 12.6 GW-61 Groundwater 26.8 3.3 15.6 5.8 7.5 12.0 15.1 SP-09 Hot Spring 66.9 26.3 18.5 434.1 34.0 218.5 469.0 L-31 Lake 13.9 57.0 33.6 77.9 5.1 22.5 25.6 R-64 River 4.6 1.3 1.6 3.5 7.2 1.1 5.9 SP-04 Spring 323.4 70.5 86.7 410.3 37.9 227.8 634.4 R-65 River 9.1 1.5 2.1 13.2 8.1 6.0 12.9 GW-74 Groundwater 42.1 14.1 19.1 9.7 11.9 4.4 13.7 L-01 Crater lake 2.8 161.1 42.8 1229.0 3.4 483.3 178.6 SP-11 Hot spring 9.1 3.9 1.6 89.0 28.4 43.9 64.6 GW-63 Groundwater 2.4 2.1 1.3 5.5 15.6 3.0 2.6

Lake evaporation to input ratios can be derived from both isotopic tracers (i.e., 18O/16O and

2H/1H). Both isotopic tracers are conservative and should yield the same evaporation/inflow ratio if all model parameters adequately represent reality. However, our results show that (i) the 2H/1H- based model is more sensitive than the 18O/16O-based model, and (ii) that the 2H/1H-based model yields higher evaporation/inflow ratios compared to results from the 18O/16O-based model (Figure

4-8).

236

Figure 4-8. Stable-isotope-based evaporation/input ratios computed using a 2H/1H-based model (y- axis) and an 18O/16O-based model (x-axis).

Previous studies recognizing a mismatch between 2H/1H- and 18O/16O-based evaporation/input ratios have multiplied α*l-v values by a constant (e.g. Bennett et al., 2008; Gibson and Reid, 2014) or have only reported 18O-based results (e.g., Zuber, 1983) as 18O/16O-based evaporation/input ratios are generally more reasonable (i.e., between 0 and 100 percent) than 2H/1H- based evaporation/input ratios. In this study I report output from each tracer and acknowledge that

2H/1H- and 18O/16O-based results do not match. Next, I reanalyze modelled 2H/1H- and 18O/16O- based evaporation/input ratios to test for the reasoning behind this discrepancy by modifying multiple model input parameters: (i) relative humidity, (ii) kinetic fractionation coefficient, (iii) atmospheric isotopic composition under changing deuterium excess, (iv) atmospheric isotopic composition under constant deuterium excess.

First, the model input relative humidity of the atmosphere near to the lake surface was modified to test if relative humidity could lead to convergence of 2H/1H- and 18O/16O-based evaporation/input ratios. This analysis showed that modifying modelled relative humidity cannot

237 explain the observed mismatch of 2H/1H- and 18O/16O-based evaporation/input ratios. This finding is consistent with expectations because changes to model input relative humidity simultaneously impacts both 2H/1H- and 18O/16O-based evaporation/input ratios. Changing relative humidity does not allow 2H/1H- and 18O/16O-based evaporation/input ratios to converge because the ratio of

CK(δ18O model) / CK(δ2H model) is close to one (see Equation 4.5).

Second, modifying the constant describing the kinetic evaporative isotope effect of an open water body (CK) was able to converge 18O and 2H-based evaporation/input ratios when the ratio of

CK(δ18O model) / CK(δ2H model) was set to 0.20 to 0.45. To test to see if these CK(δ18O model) /

CK(δ2H model) ratios are reasonable I examined a compilation of empirically-based CK values for

δ18O and δ2H (Jasechko et al., 2014) that show CK(δ18O model) / CK(δ2H model) ratios to be between 0.8 to 2.8. The CK(δ18O model) / CK(δ2H model) ratio required for convergence of 2H/1H- and 18O/16O-based evaporation/input ratios (0.20 to 0.45) is lower than empirical CK(δ18O model) /

CK(δ2H model) ratios (0.8 to 2.8), suggesting that the constant describing the kinetic evaporative isotope effect of an open water body (CK) is not the primary source of the observed difference between 2H/1H- and 18O/16O-based evaporation/input ratios.

Third, modifying the modelled deuterium excess of the atmosphere converged the 2H/1H- and 18O/16O-based evaporation/input ratios if the deuterium excess of δA is increased. However, the modelled deuterium excess of atmospheric vapor must be increased to between +15 ‰ and +60 ‰ before the two evaporation/input ratios match. Although not conclusive, I propose that a build-up of evaporated moisture over the lake surfaces may impact the isotopic composition of the atmosphere during evaporation, thereby increasing the deuterium excess of δA and providing a feedback into future evaporate. This build-up of deuterium excess has been discovered over large lakes and semi-constrained seas (e.g., deuterium excess values of up to +85‰ observed downwind of the North American Great Lakes: Machavaram and Krishnamurthy, 1995 other examples: Gat et al.,

238 1994; Bowen et al., 2012; Jasechko et al., 2014). An increase in the deuterium excess of near-surface atmospheric vapor may explain part of the observed difference in 18O and 2H-based evaporation/input ratios. However, the deuterium excess values required for convergence of 18O and

2H-based evaporation/input ratios (+15 ‰ and +60 ‰) are unreasonably high for some lakes in this region, given the deuterium excess of regional precipitation (+12.3; Table 4.1), suggesting that another model parameter must also be causing observed differences in 18O and 2H-based evaporation/input ratios.

Fourth, modifying the modelled isotopic composition of the atmosphere under fixed deuterium excess conditions (i.e., modifying δA under fixed deuterium excess) resulted in a convergence of 2H/1H- and 18O/16O-based evaporation/input ratios if the offset between vapour- precipitation was reduced to ~3.8‰ for δ18O and ~30‰ for δ2H. Model predictions of the isotopic composition of the near-surface atmosphere using an equilibrium offset (Horita and Wesolowski,

1994) suggest higher offsets between atmospheric vapour and precipitation of ~9‰ for δ18O and

~70‰ for δ2H. This finding suggests that the near-surface atmospheric vapor δ18O and δ2H value is higher than that of condensing vapor, conceptually consistent with known decreases in vapor δ18O values with increasing height above the land surface (e.g., Strong, 2012). Indeed, Strong (2012) show that atmospheric vapor δ2H values at >500 metres above the land surface are ~30‰ to ~150‰ lower than at the near surface. This finding shows that precipitation isotopic compositions are a poor determinant of near-surface atmospheric vapor and that treating δA values determined by an assumption of equilibrium with precipitation isotopic compositions provides a minimum value for near surface atmospheric vapor δ18O and δ2H values.

The above discussion rules out changes to the modelled (i) relative humidity, (ii) kinetic fractionation coefficient and (iii) atmospheric isotopic composition deuterium excess values as the cause of the discrepancy between 2H/1H- and 18O/16O-based evaporation/input ratios. I show that

239 increasing model input δA under fixed deuterium excess conditions is able to constrain 2H/1H- and

18O/16O-based evaporation/input ratios. This increase is conceptually consistent with a decrease in atmospheric vapor from the near surface to condensation altitudes. The magnitude of the increase in model input δA values under fixed deuterium excess required to converge 2H/1H- and 18O/16O-based evaporation/input ratios is consistent with observations (Strong, 2012). This research suggests that stable isotope mass balances should use atmospheric vapor isotopic compositions derived from

∗ equilibrium offset with precipitation (i.e., δA = δP – (훼푙∙푣 − 1)) only as a minimum value.

Conclusions

The primary objective of this study were to quantify Ugandan lake water balances using a stable isotope mass balance. The water balance of lakes throughout Uganda was indeed quantified using a stable isotope mass balance and showed that well flushed (evaporation/inflow ratio approaching zero) and terminal lakes (evaporation/input ratio approaching one) can be determined using a stable isotope mass balance. Results from δ18O and δ2H mass balances were discovered to produce inconsistent results with 2H-based evaporation/inflow ratios generally exceeding 18O-based evaporation/inflow ratios. Further analysis of input parameters to the Craig-Gordon evaporation model showed that the two tracers are synchronized when the modelled isotope composition of atmospheric vapour is increased from original predictions that were based upon precipitation isotopic compositions as a proxy for atmospheric vapor. This finding is conceptually consistent with a decrease in atmospheric vapor δ18O and δ2H from the near surface to condensation altitudes. This research suggests that stable isotope mass balances should use atmospheric vapor isotopic

∗ compositions derived from equilibrium offset with precipitation (i.e., δA = δP – (훼푙∙푣 − 1)) as a minimum value.

240 4.6 References

Araguás-Araguás, L., Froehlich, K., and Rozanski, K. (2000), Deuterium and oxygen-18 isotope composition of precipitation and atmospheric moisture, Hydrological Processes 14, 1341–

1355.

Bahati, G., Pang, Z., Ármannsson, H., Isabirye, E. M., and Kato, V. (2005), Hydrology and reservoir characteristics of three geothermal systems in western Uganda, Geothermics, 34, 568–591.

Bennett, K.E., Gibson, J.J., McEachern, P.M., 2008. Water yield estimates for critical loadings assessment: comparisons of gauging methods versus and isotopic approach. Can. J. Fish.

Aquat. Sci., 65, 83–99.

Bowen, G. J., Kennedy, C. D., Henne, P. D., and Zhang, T. (2012), Footprint of recycled water subsidies downwind of Lake Michigan, Ecosphere, 3, art53.

Brock, B. E., Yi, Y., Clogg-Wright, K. P., Edwards, T. W. D., Wolfe, B. B. (2009). Multi-year landscape-scale assessment of lakewater balances in the Slave River Delta, NWT, using water isotope tracers, Journal of Hydrology, 379, 81–91.

Craig, H., and Gordon, L.I. (1965), Deuterium and oxygen-18 variations in the ocean and the marine atmosphere, In E. Tongiorgi, ed, Proceedings of a Conference on Stable Isotopes in

Oceanographic Studies and Paleotemperatures, Spoleto, Italy: 9–130.

Froehlich, K. (2000), Evaluating the water balance of inland seas using isotopic tracers: the

Caspian Sea experience, Hydrological Processes, 14, 1371–1383.

Gat, J. R., Bowser, C. J., and Kendall, C. (1994), The contribution of evaporation from the

Great Lakes to the continental atmosphere: estimate based on stable isotope data, Geophysical Research

Letters, 21, 557–560.

241 Gat, J. R., Shemesh, A., Tziperman, E., Hecht, A., Geogopoulos, D., and Basturk, O. (1996),

The stable isotope composition of waters of the Eastern Mediterranean Sea, Journal of Geophysical

Research, 101, 6441–6451.

Gibson, J. J. (2002), Short-term evaporation and water budget comparisons in shallow arctic lakes using non-steady isotope mass balance, Journal of Hydrology, 264, 247–266.

Gibson, J. J., and Edwards, T. W. D. (2002), Regional water balance trends and evaporative- transpiration partitioning from a stable isotope survey of lakes in northern Canada, Global

Biogeochemical Cycles, 16.

Gibson, J. J., and Reid, R. (2014), Water balance along a chain of tundra lakes: a 20-year isotopic perspective, Journal of Hydrology, 519, 2148–2164.

Gibson, J. J., Edwards, T. W. D., and Prowse, T. D. (1996), Development and validation of an isotopic method for estimating lake evaporation, Hydrological Processes, 10, 1369–1382.

Gibson, J. J., Reid, R., and Spence, C. (1998), A six-year isotopic record of lake evaporation in the Canadian Subarctic, Hydrological Processes, 12, 1779–1792.

Gibson, J. J., Prepas, E. E., and McEachern, P. (2002), Quantitative comparison of lake throughflow, residency, and catchment runoff using stable isotopes: Modelling and results from a survey of boreal lakes, Journal of Hydrology, 262, 128–144.

Gonfiantini, R. (1986), Environmental isotopes in lake studies, In: Fritz, P., Fontes, J.Ch.

(Eds.), Handbook of Environmental Isotope Geochemistry, Elsevier, New York, 113–168.

Horita, J., and Wesolowski, D. J. (1994), Liquid-vapor fractionation of oxygen and hydrogen isotopes of water from the freezing to the critical temperature, Geochimica et Cosmochimica Acta, 58,

3425–3437. 242 Horita, J., Rozanski, K. and Cohen, S. (2008), Isotope effects in the evaporation of water: a status report of the Craig-Gordon model, Isotopes in Environmental and Health Studies, 44, 23–49.

Jasechko, S., Gibson, J. J., and Edwards, T. W. D. (2014), Stable isotope mass balance of the

Laurentian Great Lakes, Journal of Great Lakes Research, 40, 336–346.

Kato, V. (2000). Geothermal field studies using stable isotope hydrology: case studies in

Uganda and Iceland, United Nations University Reports, 10, 189–216.

Machavaram, M. V. and Krishnamurthy, R. V. (1995), Earth surface evaporative process: a case study from the Great Lakes region of the United States based on deuterium excess in precipitation, Geochemica et Cosmochimica Acta, 59, 4279–4293.

Ndembo, L., Travi, Y., Moyengo, L. M., and Wabakaghanzi, J. N. (2007), Isotopes in precipitations of Kinshasa area: Moisture sources and groundwater tracing, in Advances in Isotope

Hydrology and its Role in Sustainable Water Resources Management (IHS—2007), 169–176.

New, M., Lister, D., Hulme, M. and Makin, I. (2002), A high–resolution data set of surface climate over global land areas, Climate Research, 21, 1–25.

Russell, J. M., and Johnson, T. C. (2006), The water balance and stable isotope hydrology of

Lake Edward, Uganda-Congo. Journal of Great Lakes Research, 32, 77–90.

Strong, M. (2012), Variations in the stable isotope compositions of water vapor and precipitation in New Mexico : links to synoptic-scale weather, Ph.D. Dissertation, University of New

Mexico, Albuquerque, U.S.A., 266 pp.

Taylor, R. G., and Howard, K. W. (1996), Groundwater recharge in the Victoria Nile basin of east Africa: support for the soil moisture balance approach using stable isotope tracers and flow modelling. Journal of Hydrology, 180, 31-53. 243 Taylor, R. G., and Howard, K. W. (1998a), The dynamics of groundwater flow in the regolith of Uganda. International Contributions to Hydrogeology, 18, 97–114.

Taylor, R. G., and Howard, K. W. F. (1998b), Post-Palaeozoic evolution of weathered landsurfaces in Uganda by tectonically controlled deep weathering and stripping, Geomorphology, 25,

173–192.

Taylor, R. G., and Howard, K. W. (2000), A tectono-geomorphic model of the hydrogeology of deeply weathered crystalline rock: evidence from Uganda, Hydrogeology Journal, 8, 279–294.

Tindimugaya, C., Taylor, R. G., Atkinson, T. C., Barker, J., and Kulkarni, K. M. (2007),

Assessment of groundwater dynamics in Uganda using a combination of isotope tracers and aquifer hydraulics data, in Advances in Isotope Hydrology and its Role in Sustainable Water Resources

Management (IHS—2007), 169–176.

Turner, K. W., Wolfe, B. B., and Edwards, T. W. D. (2010), Characterizing the role of hydrological processes on lake water balances in the Old Crow Flats, Yukon Territory, Canada, using water isotope tracers, Journal of Hydrology, 386, 103–117.

Turner, K. W., Wolfe, B. B., Edwards, T. W., Lantz, T. C., Hall, R. I., and Larocque, G.

(2014), Controls on water balance of shallow thermokarst lakes and their relations with catchment characteristics: a multi‐year, landscape‐scale assessment based on water isotope tracers and remote sensing in Old Crow Flats, Yukon (Canada), Global Change Biology, 20, 1585–1603.

Yi, Y., Brock, B. E., Falcone, M. D., Wolfe, B. B., Edwards, T. W. D. (2008), A coupled isotope tracer method to characterize input water to lakes, Journal of Hydrology, 350, 1–13.

Zuber, A. (1983), On the environmental isotope method for determining the water balance of some lakes, Journal of Hydrology, 61, 409–427.

244

APPENDICIES

245

The stable isotopic composition of Earth’s large lakes

246

δ18O δ2H δ18O δ2H Lake Lake (SMOW) (SMOW) (SMOW) (SMOW) Abhe 3.73 -3.5 Ladoga -9.53 Abiyata 10.00 64.2 Ladoga -10.45 -77.2 Abiyata 7.09 46.9 Ladoga -10.31 -75.6 Abiyata 8.36 Lucern -12.68 Abiyata 7.56 Malawi 1.70 10.6 Abiyata 7.52 Malawi 1.66 11.0 Abiyata 8.37 51.9 Malawi 1.60 11.7 Abiyata 7.99 56.1 Malawi 1.60 12.3 Afdera 6.61 29.0 Malawi 1.90 12.3 Afdera 6.62 27.9 Malawi 1.85 12.4 Afdera 5.90 23.9 Malawi 1.89 13.0 Afdera 6.28 29.0 Malawi 1.94 12.5 Afdera 6.49 29.4 Malawi 1.94 13.0 Afdera 6.66 29.2 Malawi 2.08 13.2 Afdera 6.71 28.5 Malawi 2.16 13.2 Afdera 6.87 28.8 Malawi 2.08 13.2 Afdera 6.79 28.3 Malawi 2.14 13.2 Afdera 6.57 27.3 Malawi 2.14 13.5 Afdera 5.28 25.3 Malawi 2.08 13.6 Albert 5.20 37.0 Malawi 2.02 14.0 Aral Sea 3.77 7.0 Malawi 2.09 13.4 Aral Sea 3.57 7.9 Malawi 2.07 13.4 Aral Sea 3.79 9.0 Malawi 2.13 13.5 Aral Sea 3.97 9.2 Malawi 2.06 13.6 Aral Sea 3.98 10.3 Malawi 2.00 13.2 Aral Sea 3.90 -0.1 Manasarovar -11.34 -83.7 Aral Sea 3.84 8.9 Manasarovar -9.36 -75.3 Aral Sea 4.07 9.4 Manasarovar -4.96 -56.2 Aral Sea 3.88 7.5 Manasarovar -3.81 -49.2 Aral Sea 3.89 10.6 Manasarovar -3.30 -44.9 Aral Sea 3.89 9.3 Manasarovar -3.34 -51.6 Aral Sea 3.56 8.0 Manasarovar -2.22 -43.6 Aral Sea 3.25 3.9 Mar Chiquita 3.20 16.0 Aral Sea 3.40 2.2 Mar Chiquita 3.00 16.0 Aral Sea 1.80 -8.4 Mar Chiquita 3.20 17.0 Aral Sea 4.30 7.0 Mar Chiquita 3.30 19.0 Aral Sea 3.60 4.0 Mar Chiquita 3.20 19.0 Aral Sea 3.00 3.2 Mar Chiquita 3.10 16.0 Aral Sea 0.80 -10.4 Mar Chiquita 3.00 18.0 Aral Sea 0.80 -8.8 Mar Chiquita 3.30 20.0 Aral Sea 1.50 -10.6 Mar Chiquita 3.20 18.0 Aral Sea 1.79 -5.9 Mar Chiquita 3.20 18.0 Aral Sea 2.40 -3.9 Mar Chiquita 3.10 18.0 Aral Sea 2.20 -7.6 Mar Chiquita 3.20 16.0 Aral Sea 2.89 0.4 Mar Chiquita 3.10 17.0 Aral Sea 1.60 -7.7 Mar Chiquita 3.00 18.0 247 δ18O δ2H δ18O δ2H Lake Lake (SMOW) (SMOW) (SMOW) (SMOW) Aral Sea -5.30 -47.9 Mar Chiquita 2.90 18.0 Aral Sea 3.60 5.2 Mar Chiquita 3.20 19.0 Aral Sea 1.80 2.5 Mar Chiquita 3.20 18.0 Aral Sea 4.60 4.9 Mar Chiquita 3.20 18.0 Aral Sea -0.50 -19.4 Mar Chiquita 3.20 17.0 Aral Sea 3.20 1.7 Mar Chiquita 3.20 18.0 Aral Sea 3.90 3.6 Mar Chiquita 3.00 18.0 Aral Sea 3.70 4.2 Mar Chiquita 3.20 18.0 Aral Sea 3.80 4.1 Mar Chiquita 3.30 20.0 Aral Sea 3.70 4.6 Mar Chiquita 3.30 18.0 Athabasca -16.40 Mar Chiquita 3.10 18.0 Athabasca -18.30 Mar Chiquita 3.10 17.0 Athabasca -17.00 Mar Chiquita 3.20 17.0 Athabasca -15.30 -131.0 Mar Chiquita 2.90 18.0 Awasa 8.25 54.5 Mar Chiquita 3.30 18.0 Awasa 7.80 53.0 Mar Chiquita 3.30 20.0 Awasa 7.80 53.0 Mar Chiquita 2.10 13.0 Awasa 7.91 54.5 Mead -14.50 -113.5 Awasa 7.85 53.8 Mead -12.80 -100.6 Awasa 7.92 55.3 Mead -12.73 -101.7 Awasa 8.10 57.6 Mead -12.57 -99.3 Awasa 8.18 54.7 Mead -13.54 -107.3 Awasa 8.14 55.9 Mead -13.36 -106.9 Awasa 8.20 55.6 Mead -13.53 -107.5 Awasa 8.27 54.7 Mead -13.82 -108.5 Awasa 8.20 54.3 Mead -14.34 -112.4 Awasa 8.21 56.0 Mead -13.63 -107.8 Awasa 8.22 56.0 Mead -13.67 -106.9 Awasa 8.24 56.3 Mead -13.50 -107.7 Awasa 8.22 56.1 Mediterranean 2.19 8.4 Awasa 8.26 56.3 Mediterranean 1.76 7.4 Awasa 8.26 57.2 Mediterranean 1.73 8.0 Awasa 8.21 57.1 Mediterranean 1.74 8.3 Awasa 8.21 56.8 Mediterranean 1.75 7.5 Awasa 8.25 55.4 Mediterranean 2.20 8.1 Awasa 7.68 46.6 Mediterranean 2.38 8.2 Awasa 7.58 48.4 Mediterranean 1.84 7.8 Awasa 7.48 51.4 Mediterranean 2.04 10.4 Awasa 6.60 44.8 Mediterranean 1.84 9.1 Awasa 5.36 38.3 Mediterranean 10.3 Awasa 6.77 45.2 Mediterranean 2.00 8.6 Awasa 6.65 46.1 Mediterranean 2.13 9.7 Awasa 6.74 45.9 Mediterranean 2.42 5.6 Awasa 6.92 46.2 Mediterranean 2.16 8.2 Awasa 5.46 39.2 Mediterranean 1.53 8.7 Awasa 6.74 43.4 Mediterranean 1.48 8.7 248 δ18O δ2H δ18O δ2H Lake Lake (SMOW) (SMOW) (SMOW) (SMOW) Awasa 7.31 45.6 Mediterranean 1.50 8.1 Baikal -15.90 -123.7 Mediterranean 1.51 8.7 Baikal -15.90 -125.5 Mediterranean 1.48 8.0 Baikal -15.80 -121.6 Mediterranean 1.54 7.4 Baikal -15.90 -123.0 Mediterranean 1.39 6.4 Baikal -16.00 -123.4 Mediterranean 1.55 9.1 Baikal -16.00 -122.5 Mediterranean 1.41 7.8 Baikal -15.80 -122.5 Mediterranean 1.52 6.6 Baikal -15.90 -123.1 Mediterranean 1.46 8.4 Baikal -15.90 -122.2 Mediterranean 1.46 8.1 Baikal -15.80 -122.6 Mediterranean 1.90 7.6 Baikal -15.90 -123.3 Mediterranean 1.43 7.8 Baikal -14.40 -118.1 Mediterranean 1.63 8.4 Baikal -15.80 -123.7 Mediterranean 2.11 8.2 Baikal -15.90 -124.4 Mediterranean 1.74 8.1 Baikal -15.80 -123.5 Mediterranean 1.64 7.8 Baikal -15.90 -121.6 Mediterranean 1.61 6.9 Baikal -15.80 -122.9 Mediterranean 1.55 7.8 Baikal -15.90 -123.4 Mediterranean 1.68 7.5 Baikal -15.90 -123.0 Mediterranean 1.63 7.9 Baikal -15.90 -122.7 Mediterranean 2.19 8.4 Baikal -15.80 -124.2 Mediterranean 1.76 7.4 Baikal -15.90 -123.8 Mediterranean 1.73 8.0 Baikal -15.80 -124.0 Mediterranean 1.74 8.3 Baikal -15.90 -123.2 Mediterranean 1.95 7.5 Baikal -15.90 -124.2 Mediterranean 2.20 8.1 Baikal -15.80 -123.4 Mediterranean 2.38 8.2 Baikal -15.80 -123.7 Mediterranean 1.84 7.8 Baikal -15.80 -122.1 Mediterranean 2.04 10.4 Baikal -15.70 -123.0 Mediterranean 1.84 9.1 Baikal -15.80 -121.3 Mediterranean 10.3 Baikal -15.80 -123.2 Mediterranean 2.00 8.6 Baikal -15.70 -123.4 Mediterranean 2.13 9.7 Baltic Sea -8.20 -61.0 Mediterranean 2.42 8.6 Baltic Sea -7.70 -60.0 Mediterranean 2.16 8.2 Baltic Sea -7.30 -57.0 Mediterranean 1.53 8.7 Baltic Sea -7.40 -57.0 Mediterranean 1.48 8.7 Baltic Sea -7.30 -56.0 Mediterranean 1.50 8.1 Baltic Sea -6.90 -56.0 Mediterranean 1.51 8.7 Baltic Sea -6.90 -55.0 Mediterranean 1.48 8.0 Baltic Sea -7.00 -55.0 Mediterranean 1.54 7.4 Baltic Sea -7.20 -54.0 Mediterranean 1.39 6.4 Baltic Sea -6.80 -54.0 Mediterranean 1.55 9.1 Baltic Sea -6.70 -54.0 Mediterranean 1.41 7.8 Baltic Sea -6.90 -54.0 Mediterranean 1.52 6.6 Baltic Sea -6.90 -54.0 Mediterranean 1.46 8.4 249 δ18O δ2H δ18O δ2H Lake Lake (SMOW) (SMOW) (SMOW) (SMOW) Baltic Sea -7.00 -53.0 Mediterranean 1.46 8.1 Baltic Sea -6.50 -52.0 Mediterranean 1.90 7.6 Baltic Sea -6.70 -51.0 Mediterranean 1.43 7.8 Baltic Sea -6.40 -51.0 Mediterranean 1.63 8.4 Baltic Sea -6.50 -51.0 Mediterranean 2.11 8.2 Baltic Sea -6.70 -51.0 Mediterranean 1.74 8.1 Baltic Sea -6.30 -51.0 Mediterranean 1.64 7.8 Baltic Sea -6.30 -50.0 Mediterranean 1.61 7.0 Baltic Sea -6.40 -49.0 Mediterranean 1.55 7.8 Baltic Sea -6.60 -49.0 Mediterranean 1.68 7.5 Baltic Sea -6.10 -49.0 Mediterranean 1.63 7.9 Baltic Sea -6.40 -47.0 Mediterranean 1.51 8.9 Baltic Sea -5.80 -47.0 Mediterranean 1.71 7.6 Baltic Sea -5.80 -46.0 Mediterranean 1.51 8.9 Baltic Sea -5.80 -46.0 Mediterranean 1.54 8.0 Baltic Sea -5.60 -45.0 Mediterranean 1.19 7.5 Baltic Sea -5.30 -42.0 Mediterranean 1.21 7.9 Baltic Sea -5.30 -41.0 Mediterranean 1.20 8.2 Baltic Sea -5.30 -40.0 Mediterranean 1.33 7.5 Baltic Sea -4.10 -36.0 Mediterranean 1.13 7.8 Baltic Sea -4.30 -35.0 Mediterranean 0.99 7.5 Baltic Sea -1.70 -15.0 Mediterranean 1.38 7.8 Baltic Sea -5.0 Mediterranean 1.39 7.7 Baringo 8.70 47.8 Mediterranean 1.62 7.6 Baringo 8.40 48.0 Mediterranean 1.55 8.5 Baringo 6.60 36.0 Mediterranean 1.50 7.7 Beysehir -1.60 -16.0 Mediterranean 1.43 8.4 Beysehir -1.40 -23.0 Mediterranean 1.82 8.4 Beysehir -1.60 -21.0 Mediterranean 1.54 8.3 Beysehir -1.50 -22.0 Mediterranean 1.41 6.9 Beysehir -1.70 -20.0 Mediterranean 1.80 7.5 Beysehir -1.30 -19.0 Mediterranean 1.63 7.5 Beysehir -3.70 Mediterranean 1.61 8.6 Beysehir -0.60 -13.0 Mediterranean 1.55 7.5 Beysehir -0.70 -13.0 Mediterranean 1.57 7.1 Beysehir -0.70 -16.0 Mediterranean 1.30 7.8 Beysehir -0.60 -14.0 Mediterranean 1.45 7.9 Biwa -6.79 -42.2 Mediterranean 1.49 6.5 Biwa -6.49 -40.2 Mediterranean 1.51 7.8 Biwa -6.84 -42.8 Mediterranean 1.38 7.4 Biwa -6.50 -40.3 Mediterranean 1.84 8.0 Biwa -7.36 -41.7 Mediterranean 1.83 7.3 Biwa -6.81 -41.1 Mediterranean 1.99 7.5 Biwa -6.70 -41.2 Mediterranean 1.99 8.3 Biwa -7.83 -54.4 Mediterranean 1.68 7.2 Biwa -6.94 -46.2 Mediterranean 1.61 7.1 250 δ18O δ2H δ18O δ2H Lake Lake (SMOW) (SMOW) (SMOW) (SMOW) Biwa -7.24 -47.2 Mediterranean 1.60 8.0 Biwa -6.30 -37.9 Mediterranean 1.53 8.1 Biwa -6.89 -44.6 Mediterranean 1.94 8.2 Biwa -7.79 -51.3 Mediterranean 1.77 7.3 Biwa -7.37 -50.8 Mediterranean 1.76 8.4 Biwa -6.29 -41.0 Mediterranean 1.80 7.7 Black Sea -3.59 -27.7 Mediterranean 2.03 7.0 Black Sea -3.49 -26.7 Mediterranean 2.01 8.0 Black Sea -3.42 -27.2 Mediterranean 2.17 7.5 Black Sea -3.37 -26.9 Mediterranean 2.08 7.0 Black Sea -3.38 -26.6 Mediterranean 1.58 7.6 Black Sea -3.29 -25.7 Mediterranean 1.49 8.0 Black Sea -3.27 -25.3 Mediterranean 1.72 7.1 Black Sea -2.98 -23.6 Mediterranean 1.47 7.9 Black Sea -2.97 -23.9 Mediterranean 1.95 8.4 Black Sea -2.97 -24.2 Mediterranean 1.69 7.9 Black Sea -2.93 -24.6 Mediterranean 2.17 7.8 Black Sea -2.84 -23.2 Mediterranean 1.63 7.0 Black Sea -2.63 -23.0 Mediterranean 1.95 7.1 Black Sea -2.69 -22.8 Mediterranean 1.74 7.3 Black Sea -2.71 -22.4 Mediterranean 1.37 7.2 Black Sea -2.67 -22.5 Mediterranean 1.82 8.1 Black Sea -2.63 -22.5 Mediterranean 1.69 8.4 Black Sea -2.65 -22.3 Mediterranean 1.97 8.2 Black Sea -2.69 -22.0 Mediterranean 2.05 7.4 Black Sea -2.68 -21.9 Mediterranean 1.69 8.3 Black Sea -2.63 -22.2 Mediterranean 1.61 7.9 Black Sea -2.63 -21.9 Mediterranean 2.37 7.4 Black Sea -2.64 -21.8 Mediterranean 1.83 7.8 Black Sea -2.61 -21.7 Mediterranean 1.56 7.5 Black Sea -2.61 -22.4 Mediterranean 1.80 7.3 Black Sea -2.58 -22.1 Mediterranean 1.74 7.8 Black Sea -2.57 -21.8 Mediterranean 1.49 8.4 Black Sea -2.59 -21.9 Mediterranean 1.64 8.3 Black Sea -2.57 -21.6 Mediterranean 1.54 8.7 Black Sea -2.56 -21.2 Mediterranean 1.49 8.1 Black Sea -2.60 -21.0 Mediterranean 1.48 8.5 Black Sea -2.62 -21.5 Mediterranean 1.50 7.9 Black Sea -2.65 -21.6 Mediterranean 1.27 7.9 Black Sea -2.67 -21.9 Mediterranean 1.31 8.1 Black Sea -2.60 -21.3 Michigan -5.70 -43.9 Black Sea -2.58 -21.0 Michigan -5.81 -43.3 Black Sea -2.56 -20.9 Michigan -5.79 -43.8 Black Sea -2.57 -20.5 Michigan -5.90 -43.4 Black Sea -2.63 -20.7 Michigan -5.84 -43.6 Black Sea -2.68 -21.6 Michigan -5.77 -43.9 251 δ18O δ2H δ18O δ2H Lake Lake (SMOW) (SMOW) (SMOW) (SMOW) Black Sea -2.65 -23.0 Michigan -5.78 -43.2 Black Sea -2.49 -21.6 Michigan -5.86 -43.3 Black Sea -2.49 -20.8 Michigan -5.78 -44.0 Black Sea -2.50 -20.4 Michigan -5.74 -43.9 Black Sea -2.50 -20.2 Michigan -5.76 -43.6 Black Sea -2.48 -19.7 Michigan -5.87 -43.9 Black Sea -2.42 -20.7 Michigan -5.76 -43.3 Black Sea -2.39 -21.1 Michigan -5.86 -44.2 Black Sea -2.41 -20.3 Michigan -5.75 -43.3 Black Sea -2.35 -20.4 Michigan -5.79 -43.6 Black Sea -2.27 -20.4 Michigan -5.73 -43.9 Black Sea -2.31 -20.1 Michigan -5.76 -43.6 Black Sea -2.31 -19.9 Michigan -5.82 -43.8 Black Sea -2.35 -19.9 Michigan -5.82 -43.8 Black Sea -2.41 -19.3 Michigan -5.82 -44.1 Black Sea -2.34 -19.0 Michigan -5.79 -43.9 Black Sea -2.29 -18.8 Michigan -5.79 -44.4 Black Sea -2.30 -19.1 Michigan -5.83 -44.5 Black Sea -2.25 -18.7 Michigan -5.78 -44.2 Black Sea -2.23 -19.3 Michigan -5.83 -43.9 Black Sea -2.20 -19.0 Michigan -5.87 -44.0 Black Sea -2.21 -18.7 Michigan -5.90 -43.7 Black Sea -2.16 -18.2 Michigan -5.88 -43.5 Black Sea -2.17 -17.8 Michigan -5.83 -44.1 Black Sea -2.16 -17.5 Michigan -5.84 -44.0 Black Sea -2.14 -17.8 Michigan -5.89 -44.4 Black Sea -2.08 -17.8 Michigan -5.90 -44.0 Black Sea -2.08 -17.4 Michigan -5.78 -44.6 Black Sea -2.08 -17.2 Michigan -5.81 -44.3 Black Sea -2.05 -17.2 Michigan -5.94 -44.9 Black Sea -2.05 -17.6 Michigan -5.85 -44.4 Black Sea -2.06 -17.5 Michigan -5.87 -44.4 Black Sea -2.06 -17.6 Michigan -5.78 -43.9 Black Sea -2.03 -17.5 Michigan -5.78 -44.3 Black Sea -2.01 -17.3 Michigan -5.84 -44.3 Black Sea -2.02 -17.2 Michigan -5.78 -43.8 Black Sea -1.95 -17.3 Michigan -5.81 -44.7 Black Sea -2.00 -17.1 Michigan -5.87 -44.3 Black Sea -2.03 -16.8 Michigan -5.84 -44.5 Black Sea -1.89 -16.9 Michigan -5.77 -44.7 Black Sea -1.90 -16.8 Michigan -5.90 -43.9 Black Sea -1.85 -15.7 Michigan -5.74 -44.1 Black Sea -1.79 -15.5 Michigan -5.83 -44.7 Black Sea -1.75 -15.5 Michigan -5.78 -44.8 Black Sea -1.75 -16.0 Michigan -5.81 -44.7 Black Sea -1.78 -16.2 Michigan -5.85 -44.8 252 δ18O δ2H δ18O δ2H Lake Lake (SMOW) (SMOW) (SMOW) (SMOW) Black Sea -1.79 -16.4 Michigan -5.85 -44.4 Black Sea -2.63 -21.6 Michigan -5.81 -44.2 Black Sea -2.62 -22.1 Michigan -5.84 -44.2 Caspian Sea -2.00 -25.0 Michigan -5.85 -44.1 Caspian Sea -2.30 -25.2 Michigan -6.02 -44.5 Caspian Sea -1.87 -26.2 Michigan -5.78 -44.4 Caspian Sea -1.78 -23.8 Michigan -5.92 -44.4 Caspian Sea -1.75 -23.1 Michigan -5.89 -44.0 Caspian Sea -1.83 -22.3 Michigan -5.95 -45.3 Caspian Sea -1.86 -20.8 Michigan -5.87 -44.5 Caspian Sea -1.87 -19.4 Michigan -5.86 -44.6 Caspian Sea -1.73 -21.9 Michigan -5.89 -44.4 Caspian Sea -1.73 -21.2 Michigan -5.90 -44.0 Caspian Sea -1.70 -19.8 Michigan -5.81 -44.3 Caspian Sea -1.74 -18.7 Michigan -5.79 -44.2 Caspian Sea -1.54 -22.5 Michigan -5.79 -44.2 Caspian Sea -1.57 -21.3 Michigan -5.86 -44.2 Caspian Sea -1.61 -20.2 Michigan -5.74 -44.4 Caspian Sea -1.56 -19.6 Michigan -5.94 -45.7 Caspian Sea -1.48 -19.6 Michigan -5.91 -44.7 Caspian Sea -1.40 -18.7 Michigan -5.91 -44.4 Caspian Sea -1.44 -19.4 Michigan -5.83 -45.0 Caspian Sea -1.46 -18.5 Michigan -5.82 -43.7 Caspian Sea -1.44 -16.9 Michigan -5.94 -44.1 Caspian Sea -1.57 -18.3 Michigan -5.95 -44.1 Caspian Sea -1.65 -17.3 Michigan -5.85 -44.8 Caspian Sea -1.73 -16.2 Michigan -5.80 -44.4 Caspian Sea -1.62 -14.8 Michigan -5.92 -44.6 Chad -0.89 -2.8 Naivasha 6.56 36.3 Chad 0.35 2.4 Naivasha 6.30 40.4 Chad 0.30 3.8 Naivasha 6.60 36.0 Chad 0.85 10.5 Naivasha 3.60 23.0 Chad 1.69 8.0 Naivasha 4.10 24.0 Chad 2.54 12.9 Naivasha 4.40 18.0 Chad 4.67 25.2 Naivasha 4.20 20.0 Chad 5.01 25.7 Naivasha 4.90 33.0 Chad 5.14 23.7 Naivasha 6.60 36.0 Chad 5.51 29.8 Nam Co -7.57 -73.0 Chad 6.40 31.8 Nam Co -7.03 -66.7 Chad 6.80 34.2 Nasser -1.17 1.2 Chad 6.97 41.9 Nasser -1.15 0.7 Chad 7.21 42.6 Nasser -1.11 -0.1 Chad 7.88 45.0 Nasser -0.99 1.1 Chad 8.06 43.8 Nasser -0.57 4.1 Chad 8.06 41.1 Nasser 0.06 10.5 Chad 8.60 54.1 Nasser 0.17 8.3 253 δ18O δ2H δ18O δ2H Lake Lake (SMOW) (SMOW) (SMOW) (SMOW) Chad 10.61 57.5 Nasser -1.35 -0.1 Chad 12.37 64.9 Nasser -1.30 1.1 Chad 11.33 57.3 Nasser -1.20 2.1 Chad 9.52 47.7 Nasser -0.62 5.7 Chad 11.60 59.2 Nasser -1.22 0.6 Chad 12.69 69.8 Nasser -1.23 1.1 Chad 13.01 70.0 Nasser -1.27 -0.1 Chad 12.99 67.6 Nasser -1.24 0.3 Chad 13.11 73.7 Nasser -1.25 1.0 Chad 13.51 72.7 Nasser -1.41 1.1 Chad 13.88 69.8 Nasser -0.90 5.1 Chad 14.05 74.7 Nasser -0.76 4.4 Chad 14.97 76.7 Nasser -0.64 3.8 Chad 0.11 -1.3 Nasser -0.84 3.0 Chad 3.57 22.5 Nasser -0.85 3.5 Chad 4.09 25.7 Nasser -0.80 3.0 Chad 3.95 22.0 Nasser -0.78 3.2 Chad 4.09 22.7 Nasser -0.73 2.4 Chad 4.07 20.0 Nasser 0.69 11.9 Chad 4.59 23.9 Nasser 0.46 10.1 Chad 4.94 24.9 Nasser 0.27 9.5 Chad 5.46 26.6 Nasser 0.19 9.2 Chad 5.21 27.9 Nasser 0.34 9.4 Chad 5.63 31.1 Nasser 2.16 19.6 Chad 5.91 30.1 Nasser 2.11 18.8 Chad 5.93 31.5 Nasser 2.01 20.0 Chad 5.98 33.8 Nasser 2.12 21.0 Chad 5.85 33.8 Nasser 2.12 20.2 Chad 6.27 37.7 Nasser 2.11 20.1 Chad 6.35 36.7 Nasser 2.10 18.9 Chad 6.45 36.7 Nasser 2.11 20.8 Chad 6.65 36.7 Nasser 2.18 21.2 Chad 6.72 35.5 Nasser 1.56 15.8 Chad 6.82 38.7 Nasser 2.41 21.9 Chad 7.04 38.9 Ngangla Ringco -4.22 -56.6 Chad 7.32 36.2 Nicaragua -2.00 -9.0 Chad 7.72 35.7 Oahe -14.17 -112.6 Chad 7.67 37.7 Oahe -14.06 -113.5 Chad 7.84 39.1 Oahe -14.02 -113.1 Chad 7.09 40.6 Oahe -14.39 -116.7 Chad 7.29 44.6 Oahe -14.45 -117.4 Chad 7.39 43.8 Oahe -14.24 -116.8 Chad 7.54 44.3 Oahe -14.28 -116.0 Chad 7.96 43.1 Oahe -14.23 -115.8 Chad 8.11 42.6 Oahe -14.57 -117.4 Chad 8.36 42.1 Oahe -14.03 -115.5 254 δ18O δ2H δ18O δ2H Lake Lake (SMOW) (SMOW) (SMOW) (SMOW) Chad 8.71 43.1 Oahe -14.21 -116.0 Chad 8.21 45.8 Okanagan -11.1 -104.0 Chad 7.98 49.5 Okanagan -12.0 -105.0 Chad 8.53 49.5 Okanagan -11.6 -105.0 Chad 8.90 52.9 Okanagan -10.6 -105.0 Chad 9.07 51.4 Okanagan -11.4 -103.0 Chad 9.12 50.9 Okanagan -10.7 -102.0 Chad 8.95 47.2 Okanagan -9.9 -101.0 Chad 9.10 47.7 Okanagan -11.5 -102.0 Chad 9.42 49.9 Okanagan -11.3 -101.0 Chad 9.22 55.8 Okanagan -11.6 -101.0 Chad 9.62 57.3 Okanagan -11.6 -102.0 Chad 9.54 56.1 Okanagan -11.8 -102.0 Chad 10.11 57.5 Okanagan -11.7 -101.0 Chad 10.24 50.7 Okanagan -11.2 -99.0 Chad 10.61 56.6 Okanagan -11.7 -108.0 Chad 10.74 54.8 Okanagan -10.7 -108.0 Chad 11.04 53.4 Okanagan -11.4 -108.0 Chad 10.91 57.0 Okanagan -11.9 -109.0 Chad 11.21 59.5 Okanagan -11.8 -105.0 Chad 11.55 62.7 Okanagan -10.7 -106.0 Chad 11.80 64.7 Okanagan -11.2 -104.0 Chad 12.02 70.6 Okanagan -10.7 -102.0 Chad 12.14 70.8 Okanagan -10.7 -98.0 Chad 12.29 68.6 Okanagan -11.8 -101.0 Chad 12.47 68.8 Okanagan -11.5 -101.0 Chad 12.08 58.7 Okanagan -11.8 -103.0 Chad 13.36 71.5 Okanagan -11.5 -101.0 Chad 13.88 75.4 Okanagan -11.9 -102.0 Chad 14.28 77.6 Okanagan -12.1 -104.0 Chad 14.95 79.9 Okanagan -11.9 -102.0 Chamo 8.54 50.4 Okanagan -11.8 -102.0 Chamo 6.55 45.1 Okanagan -11.7 -103.0 Chamo 6.63 45.2 Okanagan -11.3 -100.0 Chamo 6.60 45.6 Okanagan -11.9 -107.0 Chamo 7.59 49.5 Okanagan -11.7 -109.0 Chamo 7.46 50.1 Okanagan -11.6 -109.0 Chamo 8.12 50.9 Okavango Delta -4.72 -34.5 Chamo 8.23 53.0 Okavango Delta -4.54 -30.2 Chamo 8.31 49.5 Okavango Delta -4.06 -32.2 Chamo 9.33 55.0 Okavango Delta -4.11 -29.1 Chamo 8.13 53.7 Okavango Delta -3.91 -29.1 Chamo 7.12 47.9 Okavango Delta -3.77 -28.2 Chamo 7.80 51.2 Okavango Delta -1.48 -16.5 Dabusun -0.61 -45.9 Okavango Delta -2.31 -28.0 Dabusun -0.21 -43.5 Okavango Delta -0.85 -14.8 255 δ18O δ2H δ18O δ2H Lake Lake (SMOW) (SMOW) (SMOW) (SMOW) Dabusun 0.17 -40.7 Okavango Delta -0.81 -11.9 Dabusun 0.54 -39.6 Okavango Delta -0.15 -3.3 Dabusun 0.56 -45.6 Okavango Delta 0.11 0.4 Dabusun 0.37 -47.8 Okavango Delta 0.56 -2.7 Dabusun 1.57 -40.8 Okavango Delta 0.65 -5.1 Dabusun 2.34 -38.6 Okavango Delta 0.71 -4.2 Dagze Co -6.38 -69.5 Okavango Delta 0.79 -4.2 Dead Sea 4.30 4.0 Okavango Delta 0.99 -4.2 Dead Sea 3.74 2.2 Okavango Delta 0.79 -2.2 Dead Sea 4.15 5.7 Okavango Delta 1.22 -0.5 Dead Sea 4.31 -1.7 Okavango Delta 1.36 -3.1 Dead Sea 4.35 1.9 Okavango Delta 1.57 -1.1 Dead Sea 4.46 -0.7 Okavango Delta 1.43 2.1 Dead Sea 5.00 2.0 Okavango Delta 2.06 4.4 Dead Sea 5.10 5.0 Okavango Delta 2.11 3.2 Dead Sea 4.90 3.0 Okavango Delta 2.28 3.5 Dead Sea 0.10 4.6 Okavango Delta 2.43 4.4 Dead Sea 0.80 4.8 Okavango Delta 2.51 6.1 Dead Sea 0.20 4.8 Okavango Delta 2.66 5.2 Dead Sea -0.10 4.4 Okavango Delta 2.34 1.2 Dead Sea 0.10 4.4 Onega -10.94 Dead Sea -0.40 Onega -9.95 Dead Sea 0.10 4.7 Ontario -6.61 -49.2 Dead Sea -0.50 4.9 Ontario -6.37 -48.6 Dead Sea 0.70 4.5 Ontario -6.58 -48.9 Dead Sea -1.20 4.4 Ontario -6.42 -49.0 Dead Sea 0.10 4.5 Ontario -6.57 -48.8 Dead Sea 0.50 3.9 Ontario -6.50 -49.2 Dead Sea 0.20 4.5 Ontario -6.58 -49.4 Dead Sea 1.10 4.7 Ontario -6.67 -48.7 Dead Sea -0.50 4.5 Ontario -6.48 -49.1 Dead Sea -0.40 4.4 Ontario -6.53 -49.1 Dead Sea -1.90 4.5 Ontario -6.45 -49.2 Dead Sea -0.80 4.5 Ontario -6.58 -49.2 Edward 4.30 29.0 Ontario -6.62 -48.6 Edward 4.50 31.0 Ontario -6.56 -49.2 Edward 4.20 29.0 Ontario -6.68 -49.2 Edward 4.20 30.0 Ontario -6.48 -48.6 Egridir -1.90 -18.0 Ontario -6.59 -49.2 Egridir -1.60 -22.0 Ontario -6.57 -49.2 Egridir -2.30 -20.0 Ontario -6.65 -49.0 Egridir -1.30 -19.0 Ontario -6.68 -48.7 Egridir -2.30 -19.0 Ontario -6.58 -49.0 Egridir -2.70 -22.0 Ontario -6.64 -48.9 Egridir -2.90 -23.0 Ontario -6.50 -48.5 Egridir -3.20 -23.0 Ontario -6.63 -49.1 256 δ18O δ2H δ18O δ2H Lake Lake (SMOW) (SMOW) (SMOW) (SMOW) Egridir -3.00 -25.0 Ontario -6.58 -49.1 Egridir -2.30 -22.0 Ontario -6.61 -49.4 Elephant -8.8 -73.0 Ontario -6.62 -49.3 Butte Elephant -7.6 -67.0 Ontario -6.47 -48.8 Butte Elephant -7.4 -66.0 Ontario -6.68 -49.0 Butte Elephant -7.2 -65.0 Ontario -6.47 -49.2 Butte Elephant -7.6 -68.0 Ontario -6.57 -49.0 Butte Elephant -7.8 -68.0 Ontario -6.54 -49.6 Butte Elephant -7.0 -65.0 Ontario -6.60 -51.6 Butte Elephant -7.1 -64.0 Ontario -6.63 -49.7 Butte Elephant -7.0 -65.0 Ontario -6.62 -49.7 Butte Elephant -6.6 -63.0 Ontario -6.32 -47.9 Butte Elephant -7.8 -64.0 Ontario -6.67 -49.5 Butte Elephant -7.8 -65.7 Ontario -6.58 -49.1 Butte Erie -6.47 -47.2 Ontario -6.62 -49.3 Erie -6.54 -47.6 Ontario -6.56 -48.8 Erie -6.42 -47.0 Ontario -6.67 -49.4 Erie -6.57 -47.8 Ontario -6.48 -49.1 Erie -6.47 -47.5 Ontario -6.59 -49.4 Erie -6.56 -47.0 Ontario -6.45 -48.6 Erie -6.70 -48.6 Ontario -6.57 -49.2 Erie -6.44 -47.2 Ontario -6.56 -49.2 Erie -6.71 -49.2 Ontario -6.59 -48.8 Erie -6.53 -47.5 Ontario -6.57 -49.1 Erie -6.75 -48.8 Ontario -6.62 -49.1 Erie -6.53 -47.4 Ontario -6.64 -49.2 Erie -6.43 -48.5 Ontario -6.48 -48.5 Erie -6.48 -47.8 Ontario -6.58 -49.2 Erie -6.53 -47.2 Ontario -6.65 -49.5 Erie -6.41 -47.4 Ontario -6.60 -49.1 Erie -6.55 -47.3 Ontario -6.61 -49.1 Erie -6.50 -47.3 Ontario -6.68 -49.4 Erie -6.37 -47.4 Ontario -6.45 -48.2 Erie -6.80 -44.6 Ontario -6.56 -49.1

257 δ18O δ2H δ18O δ2H Lake Lake (SMOW) (SMOW) (SMOW) (SMOW) Erie -6.75 -46.6 Ontario -6.63 -49.0 Erie -6.38 -47.6 Ontario -6.53 -48.9 Erie -6.57 -49.2 Ontario -6.70 -49.3 Erie -6.43 -47.7 Ontario -6.55 -49.0 Erie -6.41 -47.3 Ontario -6.60 -49.3 Erie -6.63 -47.7 Ontario -6.70 -53.0 Erie -6.61 -47.9 Ontario -6.70 -51.0 Erie -6.85 -47.1 Ontario -6.60 -50.0 Erie -6.45 -47.6 Ontario -6.60 -51.0 Erie -6.48 -47.7 Ontario -6.60 -52.0 Erie -6.44 -47.9 Powell -14.61 -113.0 Erie -6.56 -48.0 Powell -14.82 -116.4 Erie -6.57 -48.0 Powell -15.46 -119.1 Erie -6.43 -47.5 Powell -15.02 -114.3 Erie -6.45 -47.7 Powell -15.12 -114.8 Erie -6.64 -47.4 Powell -14.94 -113.1 Erie -6.40 -47.8 Powell -14.92 -114.3 Erie -6.60 -47.1 Powell -15.14 -115.5 Erie -6.61 -47.8 Powell -14.87 -113.4 Erie -6.46 -47.8 Powell -14.72 -113.6 Erie -6.36 -47.8 Powell -15.35 -119.4 Erie -6.56 -47.5 Powell -15.44 -120.2 Erie -6.45 -48.0 Powell -15.36 -121.1 Erie -6.67 -48.3 Powell -15.37 -121.4 Erie -6.72 -47.9 Powell -15.37 -121.8 Erie -6.44 -47.7 Powell -15.44 -121.8 Erie -6.44 -48.2 Powell -15.29 -121.7 Erie -6.61 -48.0 Powell -15.33 -121.3 Erie -6.38 -47.5 Powell -15.47 -122.9 Erie -6.77 -48.2 Powell -15.23 -121.3 Erie -6.35 -47.5 Powell -15.61 -122.8 Erie -6.52 -47.6 Powell -15.64 -122.7 Erie -6.33 -47.5 Powell -15.58 -122.9 Erie -6.51 -47.3 Powell -15.63 -123.2 Erie -6.79 -50.8 Powell -15.52 -123.3 Erie -6.47 -48.0 Powell -15.45 -122.6 Erie -6.84 -50.6 Powell -15.46 -122.6 Erie -6.53 -47.6 Powell -15.51 -122.4 Erie -6.55 -47.5 Powell -15.45 -122.3 Erie -6.76 -51.3 Powell -15.60 -122.6 Erie -6.53 -46.6 Powell -15.57 -122.7 Erie -6.39 -48.4 Powell -15.54 -123.3 Erie -6.55 -47.9 Powell -15.29 -119.8 Erie -6.37 -48.5 Powell -14.76 -118.4 Erie -6.49 -46.7 Powell -14.82 -118.9 Erie -6.42 -48.4 Powell -14.92 -119.3 258 δ18O δ2H δ18O δ2H Lake Lake (SMOW) (SMOW) (SMOW) (SMOW) Erie -6.61 -48.8 Powell -15.38 -120.9 Erie -7.48 -54.9 Powell -15.12 -120.1 Erie -6.57 -47.8 Powell -15.47 -121.3 Erie -7.37 -55.2 Powell -15.41 -121.3 Erie -6.55 -48.7 Powell -15.41 -121.6 Erie -7.37 -55.1 Powell -15.15 -121.2 Erie -7.11 -53.7 Powell -15.31 -121.6 Erie -6.51 -48.3 Powell -15.25 -120.3 Erie -6.54 -48.5 Powell -15.33 -121.1 Erie -7.22 -54.2 Powell -15.43 -121.1 Erie -6.63 -48.6 Powell -15.44 -121.2 Erie -7.18 -53.9 Powell -15.12 -120.0 Erie -7.24 -53.5 Powell -15.25 -120.2 Erie -6.88 -50.5 Powell -15.32 -120.8 Erie -6.92 -50.5 Powell -14.87 -119.4 Erie -7.18 -52.3 Powell -14.82 -116.9 Erie -7.20 -53.4 Powell -15.14 -117.7 Erie -6.91 -51.0 Powell -15.07 -117.7 Erie -7.06 -54.0 Powell -15.52 -120.2 Erie -6.90 -50.9 Powell -15.45 -120.9 Erie -7.07 -53.4 Powell -15.58 -121.1 Erie -6.96 -51.1 Powell -15.62 -121.7 Erie -7.15 -53.8 Powell -15.58 -121.9 Erie -6.86 -50.8 Powell -15.59 -121.4 Erie -6.37 -47.7 Powell -15.59 -121.5 Erie -6.31 -48.3 Powell -15.55 -121.0 Erie -6.31 -48.1 Powell -15.54 -120.7 Erie -6.48 -47.0 Powell -15.52 -120.7 Erie -6.45 -47.6 Powell -15.57 -121.3 Erie -6.41 -47.8 Powell -15.56 -121.0 Erie -6.35 -48.4 Powell -15.45 -120.7 Erie -6.54 -49.4 Powell -15.52 -120.5 Erie -6.50 -49.3 Powell -15.48 -120.8 Erie -6.33 -48.3 Powell -15.43 -120.5 Erie -6.46 -48.3 Powell -15.53 -120.9 Erie -6.52 -49.3 Powell -15.53 -121.0 Erie -6.41 -48.1 Powell -15.56 -120.9 Erie -6.56 -49.6 Powell -15.56 -120.9 Erie -6.47 -49.2 Powell -15.47 -120.5 Erie -6.42 -49.3 Powell -15.48 -120.9 Erie -6.42 -48.6 Powell -15.54 -120.9 Erie -6.43 -49.1 Powell -15.60 -120.7 Erie -6.65 -51.6 Powell -15.56 -121.3 Erie -7.21 -53.5 Powell -15.59 -120.5 Erie -6.72 -52.3 Powell -15.49 -121.3 Erie -7.23 -54.0 Powell -15.46 -120.9 259 δ18O δ2H δ18O δ2H Lake Lake (SMOW) (SMOW) (SMOW) (SMOW) Erie -6.75 -51.8 Powell -15.48 -120.7 Erie -7.30 -54.7 Powell -15.35 -120.4 Erie -6.85 -52.0 Powell -15.43 -120.8 Erie -7.32 -54.1 Powell -15.37 -120.3 Erie -6.83 -52.5 Powell -15.35 -120.8 Erie -7.18 -54.4 Powell -15.06 -118.6 Erie -6.90 -52.2 Powell -15.31 -121.3 Erie -7.25 -53.6 Powell -15.37 -120.8 Erie -6.40 -48.8 Powell -15.42 -121.3 Erie -6.37 -49.2 Powell -15.28 -120.3 Erie -6.45 -49.0 Powell -15.35 -120.3 Erie -6.58 -49.4 Powell -15.28 -121.3 Erie -6.48 -49.3 Powell -15.43 -120.5 Erie -6.63 -50.3 Powell -15.25 -120.7 Erie -6.30 -48.0 Powell -15.37 -120.6 Erie -6.40 -52.0 Powell -15.37 -120.4 Erie -6.40 -47.0 Powell -15.19 -120.9 Erie -6.40 -46.0 Powell -15.33 -120.5 Erie -6.50 -50.0 Powell -15.23 -120.9 Erie -6.50 -46.0 Powell -15.34 -120.4 Erie -6.50 -48.0 Powell -15.21 -119.9 Erie -6.50 -50.0 Powell -14.99 -118.3 Erie -6.70 -52.0 Powell -14.88 -117.3 Erie -6.70 -49.0 Powell -15.17 -117.9 Erie -6.50 -50.0 Powell -14.95 -116.2 Erie -6.50 -51.0 Powell -14.93 -116.2 Erie -6.50 -48.0 Powell -15.12 -117.2 Erie -6.70 -49.0 Powell -15.16 -118.1 Erie -6.60 -45.0 Powell -14.86 -117.0 Erie -6.70 -51.0 Powell -14.89 -117.5 Erie -6.70 -52.0 Powell -15.20 -118.3 Erie -6.80 -53.0 Powell -15.32 -118.5 Erie -6.60 -49.0 Powell -14.86 -117.3 Erie -6.50 -48.0 Powell -15.24 -117.9 Erie -6.70 Powell -15.18 -118.2 Erie -6.60 -49.0 Powell -14.88 -117.1 Erie -6.80 -51.0 Poyang -10.61 -68.4 Erie -6.80 -56.0 Poyang -9.64 -57.4 Erie -6.10 -46.5 Poyang -9.28 -55.5 Garda -7.30 -55.1 Poyang -8.35 -49.2 Garda -7.20 -54.0 Poyang -7.85 -51.5 Garda -7.00 -53.4 Poyang -7.74 -43.9 Garda -7.40 -55.7 Poyang -6.83 -41.3 Garda -7.20 -55.0 Poyang -6.24 -42.2 Garda -9.20 -69.9 Poyang -6.42 -41.6 Garda -7.80 -59.0 Poyang -6.51 -38.5 260 δ18O δ2H δ18O δ2H Lake Lake (SMOW) (SMOW) (SMOW) (SMOW) Garda -7.40 -56.2 Poyang -6.39 -38.1 Garda -7.30 -54.1 Poyang -6.28 -38.4 Garda -7.30 -55.0 Poyang -6.09 -34.1 Garda -7.20 -54.4 Poyang -6.16 -33.0 Garda -7.30 -55.5 Poyang -6.37 -29.9 Garda -7.20 -56.2 Poyang -6.48 -30.9 Garda -7.40 -56.9 Poyang -6.54 -32.3 Garda -7.30 -54.3 Poyang -6.53 -33.3 Garda -7.40 -54.8 Poyang -6.46 -33.7 Garda -7.40 -55.4 Poyang -6.30 -34.6 Garda -7.60 -56.4 Poyang -6.33 -33.3 Garda -7.30 -54.8 Poyang -6.48 -38.9 Garda -7.30 -54.3 Poyang -6.33 -39.7 Garda -7.40 -55.3 Poyang -6.30 -38.1 Garda -7.30 -54.2 Poyang -6.20 -33.0 Garda -7.00 -53.0 Poyang -6.37 -32.3 Garda -7.30 -54.3 Poyang -6.37 -33.6 Garda -7.30 -54.8 Poyang -6.45 -32.8 Garda -7.10 -53.0 Poyang -6.29 -32.6 Garda -7.40 -54.1 Poyang -6.46 -31.3 Garda -7.30 -54.4 Poyang -6.45 -33.9 Garda -7.30 -54.2 Poyang -6.27 -34.7 Garda -7.50 -54.7 Poyang -6.30 -32.0 Garda -7.30 -54.5 Poyang -6.50 -32.9 Garda -7.40 -55.3 Poyang -6.11 -33.6 Garda -7.20 -53.9 Qarhan Salt 6.63 -15.6 Garda -7.40 -55.1 Qianhai Hu 0.97 4.4 Garda -7.10 -53.8 Qianhai Hu 1.26 3.1 Garda -7.48 -55.0 Qianhai Hu 2.48 11.9 Garda -7.46 -55.0 Qianhai Hu 2.69 11.9 Garda -7.19 -54.8 Qianhai Hu 2.80 12.5 Garda -7.23 -54.3 Qianhai Hu 2.60 14.0 Garda -7.15 -53.9 Qianhai Hu 2.78 14.8 Garda -7.24 -53.5 Qianhai Hu 2.85 15.6 Garda -7.31 -54.2 Qianhai Hu 2.69 15.6 Garda -7.35 -55.3 Qianhai Hu 2.76 16.9 Garda -7.36 -54.2 Red Sea 0.98 5.4 Garda -7.30 -54.7 Red Sea 1.13 6.2 Garda -7.23 -53.9 Red Sea 1.33 8.2 Garda -7.26 -54.3 Red Sea 1.85 11.3 Garda -7.36 -55.0 Red Sea 1.95 11.5 Garda -7.34 -54.5 Red Sea 1.14 7.0 Garda -7.30 -55.1 Red Sea 1.16 7.1 Garda -7.32 -55.9 Red Sea 1.16 7.2 Garda -7.34 -54.9 Red Sea 1.19 7.7 Garda -7.33 -55.9 Red Sea 1.22 7.2 261 δ18O δ2H δ18O δ2H Lake Lake (SMOW) (SMOW) (SMOW) (SMOW) Garda -7.26 -55.0 Red Sea 1.36 8.1 Garda -7.35 -55.3 Red Sea 1.38 7.7 Garda -7.27 -55.5 Red Sea 1.55 9.3 Garda -7.31 -55.1 Red Sea 1.57 9.0 Garda -7.38 -55.9 Red Sea 1.59 9.3 Garda -7.41 -55.5 Red Sea 1.62 9.5 Garda -7.38 -55.5 Rukwa 4.50 27.0 Garda -7.23 -54.7 Rukwa 4.30 27.0 Garda -7.38 -54.7 Rukwa 4.40 27.0 Garda -7.33 -54.4 Rukwa 4.10 23.0 Garda -7.21 -54.7 Sakakawea -15.35 -122.0 Garda -7.27 -55.6 Sakakawea -15.35 -122.5 Garda -7.45 -56.2 Sakakawea -15.60 -123.7 Garda -7.26 -55.9 Sakakawea -15.70 -123.3 Garda -7.43 -56.4 Sakakawea -15.75 -126.2 Garda -7.37 -55.4 Sakakawea -15.33 -123.1 Garda -7.47 -55.8 Sakakawea -15.22 -122.7 Garda -7.17 -54.6 Sakakawea -15.34 -125.6 Garda -7.39 -55.9 Sakakawea -15.37 -125.7 Garda -7.40 -56.1 Sakakawea -15.39 -124.1 Garda -7.42 -56.5 Sakakawea -15.46 -125.4 Garda -7.32 -54.7 Sakakawea -15.37 -122.6 Garda -7.12 -54.5 Sakakawea -15.63 -124.7 Garda -7.39 -55.1 Sakakawea -15.77 -126.0 Garda -7.29 -55.3 Sakakawea -15.53 -122.6 Garda -7.35 -55.8 Salton Sea -1.95 -43.0 Garda -7.35 -55.8 Salton Sea -5.30 -60.0 Garda -7.33 -55.9 Salton Sea -8.83 -82.7 Garda -7.32 -55.0 Sambhar Salt -5.50 Garda -7.29 -55.7 Sambhar Salt -1.00 Garda -7.31 -55.5 Sambhar Salt 4.50 Garda -7.34 -55.7 Sambhar Salt 9.60 Garda -7.29 -55.4 Sambhar Salt 19.10 Garda -7.22 -54.3 Sambhar Salt 19.70 Garda -7.24 -55.1 Sambhar Salt 21.40 Garda -7.29 -55.3 Sambhar Salt 24.00 Garda -7.18 -54.9 Shala 7.29 53.2 Garda -7.28 -55.4 Shala 7.92 55.1 Garda -7.33 -55.1 Shala 7.40 51.9 Garda -7.46 -56.2 Shala 6.22 45.7 Garda -7.46 -55.7 Shala 7.36 50.9 Garda -7.45 -55.8 Shala 7.52 48.4 Garda -7.10 -55.1 Shala 7.66 Garda -7.12 -54.3 Shala 7.49 Garda -7.13 -55.0 Shala 7.73 Garda -7.10 -54.9 Shala 7.77 262 δ18O δ2H δ18O δ2H Lake Lake (SMOW) (SMOW) (SMOW) (SMOW) Garda -7.16 -54.5 Shala 5.36 Garda -6.99 -53.8 Shala 7.82 Garda -7.15 -54.0 Shala 7.23 Garda -7.08 -55.1 Shala 6.96 Garda -7.14 -54.6 Shala 8.30 54.4 Garda -7.36 -55.7 Shala 8.28 54.3 Garda -7.28 -55.2 Shala 8.49 54.5 Garda -7.29 -55.7 Shala 7.78 51.6 Garda -7.17 -55.0 Shala 7.05 52.5 Garda -7.33 -54.3 Superior -8.58 -65.6 Garda -7.27 -55.2 Superior -8.61 -66.8 Garda -7.21 -55.1 Superior -8.66 -66.7 Garda -7.30 -54.8 Superior -8.60 -65.9 Garda -7.31 -54.8 Superior -8.66 -66.9 Garda -7.37 -54.7 Superior -8.70 -66.8 Garda -7.34 -54.3 Superior -8.64 -65.6 Garda -7.34 -54.3 Superior -8.60 -66.7 Garda -7.43 -55.7 Superior -8.63 -65.1 Garda -7.47 -54.6 Superior -8.66 -66.8 Garda -7.30 -54.7 Superior -8.59 -65.9 Garda -7.38 -55.3 Superior -8.65 -67.2 Garda -7.46 -56.0 Superior -8.66 -65.6 Garda -7.42 -55.1 Superior -8.63 -66.8 Garda -7.29 -54.0 Superior -8.61 -66.6 Garda -7.24 -54.1 Superior -8.72 -65.0 Garda -7.19 -55.6 Superior -8.74 -66.8 Garda -7.39 -56.3 Superior -8.56 -65.3 Garda -7.30 -54.8 Superior -8.64 -66.8 Garda -7.30 -55.4 Superior -8.58 -66.8 Garda -7.32 -55.7 Superior -8.61 -64.6 Garda -7.44 -55.1 Superior -8.61 -67.4 Garda -7.41 -55.5 Superior -8.65 -64.6 Garda -7.54 -55.6 Superior -8.53 -67.4 Garda -7.51 -55.8 Superior -8.62 -65.3 Garda -7.58 -55.9 Superior -8.52 -67.2 Garda -7.44 -56.0 Superior -8.61 -65.4 Garda -7.41 -55.0 Superior -8.47 -67.0 Garda -7.35 -55.6 Superior -8.65 -64.4 Garda -7.39 -55.1 Superior -8.64 -66.7 Garda -7.43 -56.0 Superior -8.58 -66.5 Garda -7.40 -55.5 Superior -8.55 -64.7 Garda -7.25 -54.8 Superior -8.67 -64.8 Garda -7.26 -54.1 Superior -8.49 -67.4 Garda -7.37 -54.8 Superior -8.49 -67.0 Garda -7.51 -56.2 Superior -8.64 -64.9 Garda -7.47 -55.4 Superior -8.63 -67.0 263 δ18O δ2H δ18O δ2H Lake Lake (SMOW) (SMOW) (SMOW) (SMOW) Garda -7.42 -56.1 Superior -8.65 -67.3 Garda -7.37 -54.4 Superior -8.69 -64.9 Garda -7.41 -55.7 Superior -8.55 -67.1 Garda -7.42 -54.5 Superior -8.60 -66.9 Garda -7.25 -54.5 Superior -8.74 -64.9 Garda -7.21 -53.5 Superior -8.60 -67.0 Garda -7.23 -54.4 Superior -8.71 -65.1 Garda -7.27 -55.0 Superior -8.52 -67.1 Garda -7.44 -56.1 Superior -8.65 -65.0 Garda -7.46 -55.8 Superior -8.55 -67.1 Garda -7.43 -54.8 Superior -8.55 -67.1 Garda -7.42 -55.5 Superior -8.59 -64.8 Garda -7.47 -55.5 Superior -8.62 -66.9 Garda -7.17 -54.6 Superior -8.61 -65.1 Garda -7.23 -55.2 Superior -8.53 -66.6 Garda -7.34 -55.1 Superior -8.60 -64.8 Garda -6.88 -54.5 Superior -8.51 -66.8 Garda -7.27 -55.6 Superior -8.67 -66.6 Garda -7.02 -54.0 Superior -8.57 -64.7 Garda -7.21 -55.3 Superior -8.74 -65.1 Garda -7.39 -55.3 Superior -8.56 -67.0 Garda -7.16 -55.5 Superior -8.56 -67.2 Garda -7.33 -55.9 Superior -8.61 -65.0 Garda -7.26 -55.6 Superior -8.57 -67.1 Garda -7.36 -54.8 Superior -8.63 -64.8 Garda -7.38 -56.0 Superior -8.69 -67.2 Garda -7.29 -55.2 Superior -8.55 -66.8 Garda -7.38 -54.9 Superior -8.68 -65.0 Garda -7.21 -55.5 Superior -8.60 -66.8 Garda -7.23 -55.9 Superior -8.62 -64.8 Garda -7.40 -56.2 Superior -8.54 -67.5 Geneva -12.38 -87.5 Superior -8.65 -64.9 Geneva -12.43 -88.5 Superior -8.58 -67.4 Geneva -12.23 -88.2 Superior -8.75 -64.9 Geneva -12.38 -85.2 Superior -8.62 -65.7 Geneva -12.17 -85.6 Superior -8.66 -67.2 Geneva -12.32 -88.5 Superior -8.65 -64.8 Geneva -12.41 -89.5 Superior -8.62 -66.8 Geneva -12.29 -86.2 Superior -8.50 -66.9 Geneva -12.21 -85.9 Superior -8.63 -65.1 Great Bear -17.90 Superior -8.73 -64.7 Great Bear -18.55 -157.6 Superior -8.72 -65.3 Great Bear -18.88 -155.7 Superior -8.57 -64.6 Great Bear -18.53 -153.9 Superior -8.64 -64.9 Great Bear -18.44 -155.7 Superior -8.63 -65.4 Great Bear -18.40 -152.5 Superior -8.70 -64.8 264 δ18O δ2H δ18O δ2H Lake Lake (SMOW) (SMOW) (SMOW) (SMOW) Great Bear -18.37 -155.7 Superior -8.60 -65.4 Great Bear -18.45 -155.7 Superior -8.77 -64.9 Great Bear -18.83 -154.8 Superior -8.49 -64.9 Great Bear -18.40 -152.3 Superior -8.59 -65.4 Great Bear -18.72 -153.6 Superior -8.72 -65.1 Great Bear -18.75 -151.4 Superior -8.69 -65.6 Great Bear -18.92 -155.4 Superior -8.68 -65.8 Great Bear -18.48 -153.4 Superior -8.62 -64.8 Great Bear -18.52 -153.6 Superior -8.71 -64.9 Great Bear -19.18 -155.7 Superior -8.63 -65.7 Great Bear -18.80 -155.5 Superior -8.55 -65.0 Great Bear -18.94 -155.6 Superior -8.68 -65.2 Great Bear -18.92 -154.4 Superior -8.57 -65.5 Great Bear -18.46 -153.4 Superior -8.73 -65.3 Great Bear -18.57 -153.8 Superior -8.67 -65.5 Great Bear -18.61 -155.0 Superior -8.82 -65.5 Great Bear -18.65 -154.7 Superior -8.64 -66.3 Great Bear -18.40 -151.5 Superior -8.79 -66.0 Great Bear -18.49 -153.8 Superior -8.70 -65.1 Great Bear -18.62 -152.6 Superior -8.54 -66.0 Great Bear -18.59 -154.2 Superior -8.68 -65.3 Great Bear -18.73 -155.5 Superior -8.68 -64.7 Great Bear -18.86 -155.7 Superior -8.71 -65.2 Great Bear -18.92 -157.3 Superior -8.68 -65.1 Great Bear -18.83 -157.3 Superior -8.60 -65.4 Great Bear -18.90 -153.2 Superior -8.64 -64.9 Great Salt -6.28 -72.8 Superior -8.65 -65.6 Great Salt -5.69 -69.8 Superior -8.54 -66.4 Great Salt -4.88 -67.8 Superior -8.66 -65.1 Great Salt -4.93 -64.8 Superior -8.70 -65.2 Great Salt -4.06 -60.9 Superior -8.54 -65.4 Great Salt -4.15 -60.9 Superior -8.63 -65.5 Great Salt -3.89 -60.6 Superior -8.59 -64.9 Great Salt -4.40 -61.3 Superior -8.64 -65.2 Great Salt -5.48 -70.9 Superior -8.66 -65.3 Great Salt -5.45 -71.1 Superior -8.67 -65.4 Great Salt -4.96 -65.9 Superior -8.55 -65.5 Great Salt -3.94 -64.4 Superior -8.71 -64.9 Great Salt -3.68 -61.4 Superior -8.53 -65.8 Great Salt -3.51 -61.4 Superior -8.71 -65.0 Great Salt -3.70 -60.9 Superior -8.59 -65.7 Great Salt -4.00 -63.1 Superior -8.62 -65.1 Great Salt -5.53 -72.8 Superior -8.59 -65.0 Great Salt -5.53 -69.8 Superior -8.56 -66.2 Great Salt -4.72 -66.4 Superior -8.72 -65.2 Great Salt -4.69 -64.5 Superior -8.56 -65.2 265 δ18O δ2H δ18O δ2H Lake Lake (SMOW) (SMOW) (SMOW) (SMOW) Great Salt -4.20 -60.9 Superior -8.55 -64.8 Great Salt -4.08 -60.6 Superior -8.62 -65.8 Great Salt -4.57 -60.6 Superior -8.49 -65.9 Great Salt -4.14 -62.5 Superior -8.62 -65.0 Great Salt -8.44 -83.4 Superior -8.69 -65.1 Great Salt -7.03 -78.9 Superior -8.54 -65.4 Great Salt -5.05 -65.8 Superior -8.56 -65.9 Great Salt -4.93 -64.1 Superior -8.51 -65.8 Great Salt -4.15 -60.3 Superior -8.59 -65.2 Great Salt -3.94 -60.5 Superior -8.53 -66.0 Great Salt -4.43 -60.2 Superior -8.59 -65.3 Great Salt -4.42 -62.2 Superior -8.70 -64.0 Great Slave -17.89 -140.1 Superior -8.80 -66.0 Great Slave -17.58 -141.7 Superior -8.70 -66.0 Great Slave -17.85 -144.2 Superior -8.70 -64.0 Great Slave -18.10 -145.2 Superior -8.80 -66.0 Great Slave -18.17 -138.3 Superior -8.70 -66.0 Great Slave -17.44 -136.5 Superior -8.80 -64.0 Great Slave -17.90 Superior -8.80 Huron -7.18 -52.9 Superior -8.80 -67.0 Huron -6.90 -54.3 Superior -8.80 -66.0 Huron -6.86 -54.1 Superior -8.80 -66.0 Huron -7.02 -52.1 Superior -8.70 -66.0 Huron -6.96 -54.0 Superior -8.70 Huron -7.00 -53.8 Superior -8.70 -66.0 Huron -7.04 -53.4 Superior -8.70 -65.0 Huron -6.97 -53.0 Superior -8.70 -65.0 Huron -7.04 -54.3 Superior -8.80 -65.0 Huron -7.07 -54.3 Superior -8.80 -67.0 Huron -7.03 -53.1 Superior -8.80 -68.0 Huron -6.92 -54.4 Superior -8.80 -67.0 Huron -7.08 -53.4 Superior -8.70 -68.0 Huron -7.01 -54.2 Tahoe -5.20 -59.0 Huron -7.00 -52.9 Tahoe -5.80 -59.0 Huron -7.05 -54.0 Tahoe -5.20 -56.0 Huron -7.09 -54.1 Tahoe -5.80 -56.0 Huron -7.06 -54.0 Tana 3.33 30.3 Huron -7.15 -53.9 Tana 4.15 33.5 Huron -7.12 -51.5 Tana 4.35 36.3 Huron -7.04 -54.0 Tana 4.57 36.6 Huron -7.06 -54.1 Tana 4.95 37.8 Huron -7.04 -52.4 Tana 5.11 38.4 Huron -6.97 -54.2 Tana 5.98 44.8 Huron -7.19 -53.4 Tana 5.32 42.8 Huron -6.94 -53.9 Tana 4.87 37.9 Huron -7.00 -53.5 Tana 4.88 38.0 266 δ18O δ2H δ18O δ2H Lake Lake (SMOW) (SMOW) (SMOW) (SMOW) Huron -6.91 -54.1 Tana 4.87 37.8 Huron -7.07 -54.8 Tana 4.90 37.7 Huron -7.04 -53.2 Tana 4.11 35.8 Huron -7.10 -54.9 Tana 3.77 32.9 Huron -7.08 -53.2 Tana 3.93 34.5 Huron -6.90 -54.1 Tana 4.26 36.7 Huron -7.12 -52.8 Tana 6.30 46.0 Huron -6.93 -54.2 Tana 6.70 45.0 Huron -7.09 -53.7 Tana 6.80 50.0 Huron -7.01 -54.2 Tana 3.10 27.5 Huron -7.08 -54.0 Tana 3.20 29.0 Huron -7.05 -52.9 Tana 3.50 33.4 Huron -7.15 -53.9 Tana 3.85 32.5 Huron -7.06 -53.4 Tana 5.48 41.3 Huron -7.16 -53.7 Tana 5.72 43.1 Huron -7.07 -53.7 Tana 6.17 44.1 Huron -7.05 -53.8 Tana 6.52 48.7 Huron -7.10 -54.6 Tana 4.30 34.6 Huron -7.09 -53.4 Tana 3.74 29.5 Huron -7.02 -54.7 Tana 3.60 28.8 Huron -7.01 -54.4 Tana 3.46 28.8 Huron -7.01 -53.6 Tana 3.57 29.7 Huron -7.11 -53.3 Tana 3.55 29.8 Huron -7.06 -54.7 Tana 3.70 29.6 Huron -7.09 -54.0 Tana 3.68 30.7 Huron -7.05 -53.7 Tana 3.56 30.5 Huron -6.94 -54.7 Tana 3.67 30.8 Huron -7.00 -54.0 Tana 4.11 32.6 Huron -6.93 -53.9 Tana 3.67 32.1 Huron -7.18 -54.9 Tana 4.24 34.3 Huron -7.01 -54.4 Tana 4.27 35.1 Huron -7.08 -53.6 Tana 3.88 31.4 Huron -7.13 -54.7 Tana 4.30 34.0 Huron -7.10 -54.3 Tana 4.03 33.5 Huron -7.15 -54.3 Tana 4.05 34.2 Huron -7.03 -53.8 Tana 4.41 35.4 Huron -6.99 -55.2 Tana 4.31 36.2 Huron -7.08 -54.6 Tana 4.38 37.0 Huron -7.03 -54.3 Tana 5.16 37.8 Huron -7.06 -53.0 Tana 4.41 37.6 Huron -7.04 -53.9 Tana 3.91 35.4 Huron -6.96 -54.6 Tana 3.98 31.5 Huron -7.13 -53.5 Tanganyika 3.50 23.8 Huron -7.16 -53.8 Tanganyika 3.51 23.8 Huron -7.15 -54.6 Tanganyika 3.57 23.9 Huron -7.13 -55.0 Tanganyika 3.71 24.7 267 δ18O δ2H δ18O δ2H Lake Lake (SMOW) (SMOW) (SMOW) (SMOW) Huron -7.09 -53.2 Tanganyika 3.82 25.3 Huron -7.00 -54.1 Tanganyika 3.95 26.2 Huron -6.93 -54.6 Tanganyika 4.01 26.6 Huron -7.09 -53.6 Tanganyika 4.08 26.9 Huron -7.18 -54.0 Tanganyika 4.14 27.5 Huron -7.21 -54.5 Tanganyika 4.15 27.5 Huron -7.18 -55.1 Tanganyika 4.18 27.5 Huron -7.29 -53.5 Tanganyika 4.16 27.8 Huron -7.16 -54.9 Tanganyika 4.18 27.9 Huron -7.15 -53.6 Tanganyika 4.19 28.0 Huron -7.16 -54.8 Tanganyika 4.18 27.9 Huron -6.94 -53.9 Tanganyika 4.19 28.0 Huron -7.10 -53.1 Tanganyika 4.21 28.0 Huron -6.94 -54.0 Tanganyika 4.21 28.0 Huron -7.18 -53.5 Tanganyika 4.21 27.9 Huron -6.99 -53.8 Tanganyika 3.52 23.5 Huron -6.89 -53.7 Tanganyika 3.80 26.1 Huron -7.11 -53.6 Tanganyika 4.14 27.6 Huron -7.06 -52.6 Tanganyika 4.16 27.8 Huron -7.11 -54.9 Tanganyika 4.20 27.9 Huron -7.11 -55.4 Tanganyika 3.26 23.7 Huron -7.12 -53.2 Tanganyika 3.96 26.2 Huron -7.07 -54.2 Tanganyika 4.08 27.1 Huron -7.12 -53.5 Tanganyika 4.17 27.2 Huron -7.20 -54.5 Tanganyika 4.18 27.9 Huron -7.20 -54.5 Tanganyika 4.17 23.7 Huron -7.14 -53.2 Tanganyika 4.12 24.7 Huron -7.20 -54.6 Tanganyika 4.02 26.6 Huron -7.25 -54.8 Tanganyika 3.48 27.7 Huron -7.11 -53.9 Tanganyika 3.45 23.8 Huron -7.35 -55.6 Tanganyika 3.62 24.5 Huron -7.06 -53.3 Tanganyika 3.88 25.7 Huron -7.20 -53.0 Tanganyika 3.96 26.5 Huron -7.20 -54.0 Tanganyika 4.17 27.6 Huron -7.00 -55.0 Tanganyika 3.45 23.8 Huron -7.20 -59.0 Tanganyika 3.50 23.8 Huron -7.30 -54.0 Tanganyika 3.51 23.9 Huron -7.30 -54.0 Tanganyika 3.57 24.6 Huron -7.20 -54.0 Tanganyika 3.56 24.6 Huron -7.20 -54.0 Tanganyika 3.62 24.7 Huron -7.20 -52.0 Tanganyika 3.66 24.6 Huron -7.30 -53.0 Tanganyika 3.64 24.7 Huron -7.20 Tanganyika 3.68 24.8 Huron -7.30 -56.0 Tanganyika 3.54 24.6 Huron -7.30 -54.0 Taro Co -5.63 -68.5 Huron -7.30 -58.0 Taupo -5.31 -33.0 268 δ18O δ2H δ18O δ2H Lake Lake (SMOW) (SMOW) (SMOW) (SMOW) Huron -7.40 -55.0 Titicaca -4.15 -52.0 Huron -7.40 -55.0 Titicaca -4.25 -53.0 Huron -7.20 -55.0 Titicaca -4.15 -50.4 Huron -7.20 -53.0 Titicaca -2.90 -47.3 Huron -7.10 -53.0 Titicaca -2.92 -47.3 Huron -7.10 -51.0 Titicaca -2.85 -47.1 Huron -7.30 -56.0 Titicaca -3.30 -46.3 Huron -7.40 -55.0 Titicaca -3.40 -48.5 Huron -6.40 -47.0 Titicaca -3.35 -46.4 Huron -7.50 -53.0 Titicaca -4.70 -54.8 Huron -8.60 Titicaca -4.60 -51.5 Huron -8.50 -62.0 Titicaca -4.70 -51.8 Huron -7.70 -58.0 Tonlé Sap -6.09 Huron -7.40 -53.0 Tonlé Sap -4.32 Huron -7.40 -58.0 Tonlé Sap -5.95 Huron -7.50 -52.0 Tonlé Sap -8.60 Huron -7.40 -53.0 Tonlé Sap -7.78 Huron -7.40 -53.0 Tonlé Sap -8.60 Huron -7.30 -54.0 Tonlé Sap -6.22 Huron -7.40 Tonlé Sap -3.61 Huron -7.40 Tonlé Sap -5.82 Huron -7.40 -55.0 Tonlé Sap -8.60 Huron -6.60 -53.6 Tonlé Sap -7.76 Issyk-Kul -0.97 -10.9 Tonlé Sap -8.47 Issyk-Kul -0.69 -10.9 Tonlé Sap -6.09 Issyk-Kul -0.58 -10.7 Tonlé Sap -4.32 Issyk-Kul -0.56 -8.9 Turkana 5.80 37.0 Issyk-Kul -0.73 -9.5 Turkana 6.10 40.0 Issyk-Kul -0.68 -7.9 Turkana 5.60 39.0 Issyk-Kul -0.60 -6.7 Turkana 4.80 30.0 Jackson -17.97 -138.3 Turkana 5.80 40.0 Jackson -17.89 -137.5 Turkana 5.50 37.0 Kainji -17.5 Turkana 5.80 38.0 Kainji -3.7 Turkana 6.10 42.0 Kainji -2.5 Turkana 5.30 35.0 Kainji -2.2 Valencia 22.0 Kainji -2.9 Van 1.00 -6.6 Kainji -30.4 Van 0.90 -6.8 Kainji -31.9 Victoria 3.50 Kainji -29.9 Winnipeg -11.00 Kainji -28.8 Winnipeg -10.52 -79.8 Kainji -25.9 Winnipeg -10.40 -79.0 Kainji -23.8 Yamdruk-tso -5.48 -68.0 Kainji -20.0 Yellowstone -16.64 -134.5 Kainji -11.6 Zhari Namco -6.67 -75.2 Kainji -9.8 Ziway 6.70 49.0 269 δ18O δ2H δ18O δ2H Lake Lake (SMOW) (SMOW) (SMOW) (SMOW) Kainji -8.6 Ziway 8.16 57.4 Kainji -12.8 Ziway 6.50 47.3 Kainji -12.8 Ziway 7.09 52.2 Kainji -25.8 Ziway 8.40 58.9 Kivu 0.11 10.0 Ziway 6.49 46.3 Kivu -0.29 10.8 Ziway 6.90 49.3 Kivu -0.19 12.3 Ziway 6.64 47.4 Kivu -0.15 12.1 Ziway 6.64 48.6 Kivu 0.00 12.2 Ziway 6.53 48.9 Kivu 0.47 10.8 Ziway 6.29 47.3 Kivu 0.68 11.5 Ziway 6.37 47.6 Kivu 0.70 13.6 Ziway 5.43 41.6 Kivu 0.59 16.4 Ziway 5.78 42.4 Kivu 1.36 19.9 Ziway 5.06 37.3 Kivu 1.56 20.3 Ziway 6.74 49.3 Kivu 1.85 21.5 Ziway 6.78 50.2 Kivu 2.50 24.9 Ziway 7.00 46.8 Kivu 2.51 22.0 Ziway 6.69 49.1 Kivu 3.05 24.3 Ziway 5.15 40.5 Kivu 3.15 25.4 Ziway 4.38 32.4 Kivu 3.27 24.8 Ziway 6.70 49.0 Kivu 3.47 25.7 Kivu 3.24 27.4 Kluane -21.13 -168.2 Kluane -22.51 -176.7 Kluane -22.94 -180.6 Kluane -22.86 -179.3 Kluane -22.95 -177.5 Kluane -22.86 -178.1 Kluane -22.58 -177.5 Kluane -22.81 -177.2 Kluane -22.79 -178.9 Kluane -22.84 -178.1 Kluane -22.52 -179.0 Kluane -22.47 -176.0 Kluane -22.49 -176.6 Kluane -22.54 -175.0 Kluane -22.16 -175.9

270