Libro Stable-Isotope-Hydrology

Title Author(s)
Figure
1
Stable Isotope Hydrology Roger E. Diamond
Stable Isotope Hydrology
i
The GROUNDWATER PROJECT ©The Author Free download from gw-project.org
Anyone may use and share gw-project.org links. Direct distribution of the book is strictly prohibited.
Author
ii
All rights reserved. This publication is protected by copyright. No part of this book may be
reproduced in any form or by any means without permission in writing from the authors
(to request permission contact: permissions@gw-project.org). Commercial distribution
and reproduction are strictly prohibited.
Copyright
Groundwater-Project (The GW-Project) works are copyrighted and can be downloaded for
free from gw-project.org. Anyone may use and share gw-project.org links to download
GW-Project’s work. It is neither permissible to make GW-Project documents available on
other websites nor to send copies of the documents directly to others. Kindly honor this
source of free knowledge that benefits you and all those who want to learn about
groundwater.
Copyright © 2022 Roger E. Diamond (The Author)
Published by the Groundwater Project, Guelph, Ontario, Canada, 2022.
Diamond, Roger E.
Stable Isotope Hydrology / Roger E. Diamond - Guelph, Ontario, Canada, 2022
102 pages
ISBN: 978-1-77470-043-3
DOI: https://doi.org/10.21083/978-1-77470-043-3
Please consider signing up for the GW-Project mailing list to stay informed about new book
releases, events and ways to participate in the GW-Project. When you sign up for our email
list it helps us build a global groundwater community. Sign up.
Citation: Diamond, Roger E., 2022, Stable Isotope Hydrology. The Groundwater Project,
Guelph, Ontario, Canada. doi: https://doi.org/10.21083/978-1-77470-043-3.
Domain Editors: Robert Kalin, Eileen Poeter, and John Cherry

Board: John Cherry, Paul Hsieh, Robert Kalin, Ineke Kalwij, Stephen Moran, Everton de
Oliveira and Eileen Poeter
Steering Committee: John Cherry, Allan Freeze, Paul Hsieh, Ineke Kalwij, Douglas Mackay,
Stephen Moran, Everton de Oliveira, Beth Parker, Eileen Poeter, Ying Fan, Warren
Wood, and Yan Zheng.
Cover Image: Roger Diamond, 2022. Schematic diagram of typical hydrogen and oxygen
isotope ratios for major water types in the global hydrological cycle.
iii
Dedication
To all those friends, family, colleagues, students, climbers, weeders and hackers
who have shared time with me outdoors, enjoying and puzzling over the wonders of
nature.
iv
Table of Contents
DEDICATION ............................................................................................................................................. IV
TABLE OF CONTENTS.................................................................................................................................. V
THE GROUNDWATER PROJECT FOREWORD ............................................................................................. VII
FOREWORD ............................................................................................................................................ VIII
ACKNOWLEDGEMENTS ..............................................................................................................................IX
1 INTRODUCTION .................................................................................................................................. 1
2 ISOTOPES AND ISOTOPOLOGUES........................................................................................................ 4
3 MEASUREMENT AND STANDARDS ..................................................................................................... 7
3.1 MASS SPECTROMETRY ........................................................................................................................... 8
3.2 LASER CAVITY SPECTROSCOPY.................................................................................................................. 9
4 ISOTOPE FRACTIONATION ................................................................................................................ 11
4.1 KINETIC FRACTIONATION ...................................................................................................................... 11
4.2 EQUILIBRIUM FRACTIONATION ............................................................................................................... 12
4.3 FRACTIONATION AND ENRICHMENT FACTORS ........................................................................................... 13
4.4 RAYLEIGH DISTILLATION........................................................................................................................ 15
5 METEORIC WATER LINES .................................................................................................................. 17
5.1 THE GLOBAL METEORIC WATER LINE ...................................................................................................... 17
5.2 LOCAL METEORIC WATER LINES............................................................................................................. 18
5.3 CALCULATION OF METEORIC WATER LINES .............................................................................................. 20
5.3.1 Least Squares Regression ........................................................................................................... 21
5.3.2 Reduced Major Axis Regression (RMA) ...................................................................................... 21
5.4 WEIGHTED REGRESSION LINE EQUATIONS................................................................................................ 21
6 STABLE ISOTOPE HYDROLOGY .......................................................................................................... 23
6.1 ENVIRONMENTAL FACTORS AFFECTING WATER ISOTOPES (ISOTOPE EFFECTS) ................................................. 27
6.1.1 The Temperature Effect ............................................................................................................. 27
6.1.2 The Latitude Effect ..................................................................................................................... 28
6.1.3 The Continental Effect................................................................................................................ 29
6.1.4 The Altitude Effect ..................................................................................................................... 31
6.1.5 The Amount Effect ..................................................................................................................... 31
6.2 THE DEUTERIUM EXCESS ...................................................................................................................... 33
6.3 EVENT VARIATION ............................................................................................................................... 35
6.4 MASS BALANCE .................................................................................................................................. 37
6.5 HYDROGRAPH SEPARATION ................................................................................................................... 39
6.6 GEOTHERMAL WATERS ........................................................................................................................ 43
6.7 PALEOWATERS .................................................................................................................................... 46
6.8 PLANT WATERS .................................................................................................................................. 53
7 CASE STUDIES ................................................................................................................................... 56
7.1 MOISTURE SOURCE REGION .................................................................................................................. 56
7.2 RECHARGE AREA ................................................................................................................................. 57
7.3 SELECTIVE RECHARGE OF HEAVY RAINS ................................................................................................... 59
7.4 RECHARGE ESTIMATION ....................................................................................................................... 60
7.5 PISTON FLOW OF GROUNDWATER DURING A STORM ................................................................................. 61
7.6 GROUNDWATER CIRCULATION AND AGE .................................................................................................. 64
v
7.7 DETECTING SURFACE WATER IN GROUNDWATER ...................................................................................... 64

7.8 DETECTING GROUNDWATER IN SURFACE WATER ...................................................................................... 66
7.9 RESIDENCE TIME ................................................................................................................................. 67
8 WATER SAMPLING FOR STABLE ISOTOPE ANALYSIS ......................................................................... 68
8.1 PRACTICAL CONSIDERATIONS ................................................................................................................. 68
8.2 PRECIPITATION ................................................................................................................................... 68
8.2.1 Spatial and Temporal Variation in Precipitation ........................................................................ 68
8.2.2 Standard Daily or Cumulative Monthly Collectors ..................................................................... 69
8.3 SURFACE WATER ................................................................................................................................ 70
8.4 GROUNDWATER.................................................................................................................................. 71
8.4.1 Natural Groundwater Sources: Springs and Seeps..................................................................... 71
8.4.2 Artificial Groundwater Sampling Points: Wells, Boreholes and Piezometers............................. 72
9 FURTHER READING ........................................................................................................................... 74
10 WRAP-UP ......................................................................................................................................... 75
11 EXERCISES......................................................................................................................................... 77
EXERCISE 1 - METEORIC WATER LINE CALCULATIONS ............................................................................................. 77
EXERCISE 2 - CALCULATION OF RECHARGE ALTITUDE .............................................................................................. 78
EXERCISE 3 - HYDROGRAPH SEPARATION .............................................................................................................. 80
12 REFERENCES ..................................................................................................................................... 81
13 EXERCISE SOLUTIONS ....................................................................................................................... 94
SOLUTION EXERCISE 1 - METEORIC WATER LINES .................................................................................................. 94
SOLUTION EXERCISE 2 - RECHARGE ALTITUDE........................................................................................................ 96
SOLUTION EXERCISE 3 - HYDROGRAPH SEPARATION ............................................................................................... 98
14 NOTATIONS ...................................................................................................................................... 99
15 ABOUT THE AUTHOR ...................................................................................................................... 101
vi
The Groundwater Project Foreword

The Year 2022 marks as an important year for groundwater because the United
Nations Water Members and Partners chosen the theme of this year’s March 22 World
Water Day to be: “Groundwater: making the invisible visible”. The goal of the
Groundwater Project (GW-Project) is in sync with this theme.
The GW-Project, a registered charity in Canada, is committed to contributing to
advancement in groundwater education and brings a unique approach to the creation and
dissemination of knowledge for understanding and problem solving. The GW-Project
operates the website https://gw-project.org/ as a global platform for the democratization
of groundwater knowledge, founded on the principle that:
“Knowledge should be free and the best knowledge should be free knowledge.” Anonymous
The mission of the GW-Project is promoting groundwater learning. This is

accomplished by providing accessible, engaging, high-quality, educational, copyrighted
materials, free-of-charge online in many languages, to all who want to learn about
groundwater. In short, providing essential knowledge tools for developing groundwater
sustainably for humanity and ecosystems.
This is a new type of global educational endeavor in that it is based on volunteerism
of professionals from different disciplines and includes academics, consultants and retirees.
The GW-Project involves many hundreds of volunteers associated with more than 200
hundred organizations from 27 countries and six continents, with growing participation.
The GW-Project is an on-going endeavor and will continue with hundreds of books
being published online over the coming years, first in English and then in other languages,
for downloading wherever the Internet is available. An important tenet of the GW-Project
books is a strong emphasis on visualization via clear illustrations that stimulate spatial and
critical thinking to facilitate absorption of information.
The GW-Project publications also include supporting materials such as videos,
lectures, laboratory demonstrations, and learning tools in addition to providing, or linking
to, public domain software for various groundwater applications supporting the
educational process.
The GW-Project is a living entity, so subsequent editions of the books will be
published from time to time. Users are invited to propose revisions.
We thank you for being part of the GW-Project Community. We hope to hear from
you about your experience with using the books and related material. We welcome ideas
and volunteers!
The GW-Project Steering Committee
January 2022
vii
Foreword
Isotope hydrology uses isotopic measurements of water constituents for
understanding of hydrology, and its most diverse applications are those directed at
groundwater. Given the importance of this topic, the Groundwater Project (GW Project)
has launched a series of books about isotopes, starting with Introduction to Isotopes and
Environmental Tracers as Indicators of Groundwater Flow by Peter Cook, which will be
followed by in-depth books. Stable Isotopes Hydrology by Roger Diamond is the first in depth
book of the series.
This book concerns the most fundamental isotopes for groundwater studies, the
non-radioactive elements of oxygen and deuterium in water molecules. The analysis of
oxygen and deuterium give water a voice. Since the first use of these isotopes for research
in the 1960s, their application has become routine in the pursuit of answers to many
groundwater questions. Use of these isotopes, and a few others, became more feasible as
water analysis laboratories became more common and the cost of running such tests ceased
to be restrictive. Subsequent tests showed a number of successful applications resulting in
increased use of the technique. Of the many isotopes useful in groundwater studies, the
stable isotopes of water are the easiest to use because the required sample size is small and
sampling methods are uncomplicated. In many circumstances, these water isotopes
identify the geographic or temporal origin of the water, such as paleowater, or distinguish
water zones of different origins in an aquifer. The small sample volume needed for analysis
enables study of aquifer-aquitard systems to address questions relevant to water resources
and contamination. Of the many techniques that have been added to the hydrogeologist’s
toolbox over the past half century, the stable isotopes of water are amongst the most
essential.
The author of this book, Roger Diamond, is a Senior Lecturer in Hydrogeology and
Geochemistry at the University of Pretoria in South Africa. He has published extensively
on a wide variety of applications of these isotopes in the southern part of Africa. Isotopic
tools have proven to be especially effective in this region where there is much complexity
in the hydrogeology and diversity in groundwater.
John Cherry, The Groundwater Project Leader

Guelph, Ontario, Canada, September 2022
viii
Acknowledgements
I deeply appreciate the thorough and useful reviews of and contributions to this
book by the following individuals:
❖ John Cherry, G360 Institute for Groundwater Research, University of Guelph,

Ontario, Canada;
❖ Robert Kalin, Department of Civil and Environmental Engineering, University of
Strathclyde, Glasgow, Scotland;
❖ Peter Cook, National Centre for Groundwater Research and Training, Flinders
University, South Australia;
❖ Ian Clark, Department of Earth Sciences, University of Ottawa, Canada;
❖ Warren Wood, Department of Earth and Environmental Sciences, Michigan State
University, USA;
❖ Johannes Barth, Department of Geography and Earth Sciences,
Friedrich-Alexander-University of Erlangen-Nürnberg;
❖ Everton de Oliveira, Hidroplan and Instituto Água Sustentável (Sustainable Water
Institute), Brazil; and,
❖ Ferdinando Manna, G360 Institute for Groundwater Research, University of
Guelph, Ontario, Canada.
I appreciate the contributions of Peter Cook and John Cherry on the genesis of this
book. I am grateful for Amanda Sills and the Formatting Team of the Groundwater Project
for their oversight and copyediting of this book. I thank Eileen Poeter (Colorado School of
Mines, Golden, Colorado, USA) for reviewing, editing, and producing this book.
ix
1 Introduction
Stable isotope hydrology is the study of water using the stable isotopes of the
hydrogen and oxygen making up the water molecule. Many other stable isotopes can also
be used to understand water, but are not part of the actual water molecule and are found
within the dissolved matter (carbonate, nitrate, sulphate among others) in water.
As explained later in this book, the abundances of the different isotopes of hydrogen
and oxygen vary as water moves through the water cycle. These variations are caused by
reactions and transitions (e.g., evaporation and condensation). By measuring these
variations, we are able to say something about the history of the water before it arrived at
the point where we sampled it (Figure 1).
Figure 1 - The author collecting samples of snow for stable isotope analysis for his
PhD, from Waaihoek Peak in the Hex River Mountains, Western Cape, South Africa.
The snow was cored to enable analysis of the layers of snow from consecutive
weather events and from the melted/evaporated surface layer between snow storms.
Many other chemical and isotopic methods are available to help us understand
water and the processes it undergoes. However, all of the methods other than stable
isotopes of water make use of dissolved substances or isotopes that occur in minor to trace
quantities. For example, Cl is typically in the range of 1 ppm to 10,000 ppm, NO3 is seldom
above 100 ppm, Sr isotopes occur at ppb levels, and 14C or 222Rn (radioactive isotopes) occur
at even smaller fractions of the water being sampled. These tracers can therefore more
easily be affected by precipitation, adsorption, dissolution and other reactions than the H
and O isotopes of water.
The stable isotopes of water constitute nearly 100 percent of the water molecule and
therefore suffer far less from reactions that might disturb the isotope abundances. For
example, weathering of minerals will release Mg or Sr and change the chemical or isotopic
1
abundance of such tracers as groundwater moves through an aquifer. The release of oxygen
into the groundwater in the form of HCO3, H4SiO4 or other species, although able to
exchange oxygen with H2O, will do so in such minor quantities relative to the circulating
groundwater, that the stable isotope abundances of the water will not be affected. This
behavior, where the species of interest conserves its composition through space and time,
is called conservative behavior.
The main exception to the conservative behavior of stable isotopes of groundwater
is in geothermal environments, where high temperatures increase the rate of reaction
between rock and water, shifting the isotope abundances, particularly oxygen, due to its
dominance in almost every mineral. Another exception to conservative behavior is where
dissolved gases, such as H2S or CO2 interact with H2O, either chemically or through isotopic
exchange, and thereby change the isotope composition of the water. These environments
aside, the other main way in which the stable isotopes in groundwater may change over
time is due to mixing with other groundwaters, which may or may not have different
isotope abundances.
In many groundwater systems though, neither geothermal or geochemical reactions
nor mixing occur, and the stable isotope composition of groundwater remains fairly
constant from recharge point to discharge location. In these instances, variations in the
stable isotope composition of groundwater must be due to variations in the recharge, or
input to the system. Therefore, in order to understand the hydrogeology, we look to both
surface and atmospheric waters to find the source of stable isotope variations. This means
that, as hydrogeologists, we need to consider the complete water cycle and not only the
portion occurring below ground, in order to make full use of stable isotopes. This is why
we talk about stable isotope hydrology and not stable isotope hydrogeology (Figure 2).
Figure 2 - Stable isotope hydrology interfaces with many

other branches of earth science and the fundamental
sciences, such as mathematics and chemistry.
2
The value of the preserved stable isotope compositions of groundwater is that they
can then tell us something about the conditions at and prior to recharge. This is useful for
understanding recharge and precipitation processes, including evaporation, intensity of
precipitation, altitude of recharge and temperature of condensation. For example, if the
groundwater is old, then it may provide insight into past climates and therefore forms one
avenue in which scientists conduct paleoclimatic research.
This book outlines stable isotope hydrology at an introductory level as well as
including some more advanced applications, aimed at students and professionals with a
good understanding of hydrogeology, but little prior knowledge of isotopes. A good
grounding in geology along with basic chemistry and physics is essential to being a good
hydrogeologist. Knowledge of subjects such as meteorology, hydrology and geochemistry
are also helpful at times. This book explains the occurrence of stable isotopes in water
molecules and how they change abundance through the water cycle. Sections include
measurement and reporting of data, calculation of meteoric water lines, the deuterium
excess and other ratios, trends and patterns. The main body of the book explains how to
use stable isotopes to perform various hydrogeological investigations and gives a series of
case studies to show how these are done. This book also contains practical advice on water
sampling for stable isotopes, some worked examples and exercises, as well as suggested
reading for those wishing to delve deeper.
The overall intention of this book is that the reader will not only understand the
theory, but also learn how to collect water samples, have them analyzed for stable isotope
compositions, interpret the data and make conclusions useful for hydrogeologists or other
earth scientists. This information, in turn, can be used by those managing natural resources,
conducting site investigations or site rehabilitation and many other aspects of the
ever-increasing field of environmental management.
3
2 Isotopes and Isotopologues

An isotope is an atom of an element with a specific number of neutrons. An element
is defined by the number of protons, for example hydrogen has 1 proton, helium has 2 and
uranium has 92. Each of these elements, however, has a few possible isotopes - the isotopes
of an element will all have the same number of protons, but a variable number of neutrons,
which determine the isotope. For example, hydrogen has 1 proton, but with different
numbers of neutrons: protium has 0 neutrons, deuterium has 1 neutron and tritium
(radioactive) has 2 neutrons (Figure 3). As the mass of protons and neutrons is very similar,
the mass of each of these hydrogen isotopes is approximately 1, 2 and 3 amus (atomic mass
units), respectively. These are depicted as 1H, 2H and 3H. Hydrogen notation is slightly
complicated by the alternative notation of H, D and T. Similarly, oxygen has 3 stable
isotopes, 16O, 17O and 18O, all with 8 protons, but 8, 9 and 10 neutrons.
Figure 3 - The isotopes of hydrogen, including the two stable isotopes, protium and deuterium, and the
radioactive isotope, tritium.
Isotopes can be stable, radioactive (decay to another element) or radiogenic

(product of decay), or both radioactive and radiogenic, if they are part of a decay chain
(Figure 4). In this book we will focus on only stable isotopes, which neither decrease nor
increase in abundance, as they are not part of any radioactive decay chain. Note that
radiogenic isotopes are "stable" from a physics definition, but as their abundance changes
due to being daughters of radioactive isotopes, they are not called stable in a geochemical
definition. As the abundance of stable isotopes on Earth is fixed, the ratios of their average
global abundances are constant. For example, for oxygen, which occurs as 16O, 17O and 18O,
as these are all stable, their relative abundances are 99.76 percent, 0.04 percent and 0.20
percent.
4
Figure 4 - Examples of the different types of isotopes, especially those of interest in hydrogeology.
Geochemists define stable isotopes as those that are not only physically stable, but also neither increase
(radiogenic) nor decrease (radioactive) in abundance.
Hydrogen is somewhat more complex, as 3H (tritium) is radioactive, with a half-life

of 12.3 years. Protium and deuterium, however, are both stable, and on average make up
99.98 percent and 0.02 percent of hydrogen in natural waters, respectively. As with oxygen,
you may notice that the lighter stable isotope is by far the most abundant. This is true for C
and N and in all other light stable isotopes, but not always for the heavier stable isotopes.
When displayed as a ratio, by convention the heavier isotope is always the numerator and
the lighter isotope the denominator, for example 18O/16O or 17O/16O.
An isotopologue is an isotopic species of a molecule. For example, a water molecule
can include 2 stable hydrogen isotopes, in 2 possible positions and 3 stable oxygen isotopes,
giving 9 possible isotopologues (Table 1). If the possibilities with tritium (the radioactive
isotope of hydrogen) were added, then the isotopologues increase to 18, but many of these
are virtually non-existent. For example, the theoretical abundance of TD17O would be the
theoretical abundances of each of the isotopes multiplied together, giving:
10-17  210-4  410-4 = 810-25  10-24

for modern waters, which means it exists in extremely small quantities!
5
Table 1 - List of stable isotopologues for water and their

theoretical abundances, as calculated from global average
abundances of these stable isotopes (Emiliani, 1987).
3
Isotopologues with tritium ( H, radioactive) have been excluded.
Isotopologue Mass Abundance (%)
1 1 16 18 99.732
HH O
1 1 18 20 0.200
HH O
1 1 17 19 0.038
HH O
1 2 16 19 0.015
HH O
1 2 18 21 0.00003
HH O
1 2 17 20 0.0000057
HH O
2 2 16 20 0.0000022
HH O
2 2 18 22 0.0000000045
HH O
2 2 17 21 0.00000000086
HH O
A short note on jargon. Isotope scientists often refer to heavier or lighter isotope
compositions as enriched or depleted, respectively. The convention here is that the words
"in heavier isotopes" have been left out. So, an "enriched sample" actually means a sample
that is enriched in heavier isotopes, and not a sample enriched in lighter isotopes. This
slightly confusing jargon can be avoided by referring to samples with more of the heavier
isotope (enriched) as having:
• heavier isotope compositions; and,
• more positive/higher delta values (delta, , will be explained below).
Similarly, for water with more of the lighter isotopes (depleted), we can say:
• lighter isotope compositions; and,
• more negative/lower delta values.
6
3 Measurement and Standards

Isotope ratios are, by convention, the ratio of the heavier isotope to the lighter one.
The differences in isotope ratios between materials are slight and are not easily measured
in an absolute sense. However, if measured as a relative difference between the sample and
a standard of known isotope ratio, then the precision increases substantially. Thus
measurements in stable, light-isotope mass spectrometers are made by comparing the
sample to a standard, using either a dual inlet or continuous flow system.
Stable isotope ratios are reported as deviation from an international standard.
For both hydrogen and oxygen in water, the standard is SMOW, Standard Mean
Ocean Water. SMOW was devised by Harmon Craig in 1961 (Craig, 1961) as an average of
previous ocean water samples from Epstein and Mayeda (1953) and Horibe and
Kobayakawa (1960), but no actual sample existed. Because of this, SMOW was defined
relative to NBS1 (National Bureau of Standards, standard 1), a United States administered
sample from the Potomac River. Because of the difficulties of not having an actual standard
to analyze, the International Atomic Energy Agency in Vienna commissioned the creation
of VSMOW in 1966, which was to mimic SMOW. Although VSMOW is not isotopically
identical to SMOW (Clark and Fritz, 1997), it is similar enough (within laboratory error) to
be treated as the same. Even though VSMOW may have been used in their laboratories to
enable calibration of their local laboratory standards (Gonfiantini, 1981; Sharp, 2007),
workers should not report the deviation from VSMOW, as this merely enables correction
to SMOW. Data corrected to VSMOW should be reported as deviation from SMOW.
Some natural waters are very isotopically different from ocean water, particularly
those at very high latitudes, altitudes and low temperatures, in other words, high
mountains or polar areas. For this reason, the Standard Light Antarctic Precipitation (SLAP)
standard was created. This standard should be used to calibrate isotopically light working
standards for use in high altitude or high latitude environments.
The Greek lower-case delta, , is used to denote the deviation of a sample from a
standard, as shown in Equation 1.
(𝑅𝑠𝑎𝑚𝑝𝑙𝑒 − 𝑅𝑠𝑡𝑎𝑛𝑑𝑎𝑟𝑑 )
𝛿 = (1)
𝑅𝑠𝑡𝑎𝑛𝑑𝑎𝑟𝑑
where:
R = isotope ratio such as 2H/1H (dimensionless)

18
When relatively more of the heavy isotope (e.g., O) is present in the sample than the
standard, then the  value will be greater than zero, whereas samples relatively depleted in
the heavy isotope will have negative  values. The 18O and 2H values of SMOW are equal
to 0. As the variations in isotope ratios are generally quite small, these  values are reported
7
in permil (parts per thousand), using the ‰ notation. The equation combining these
definitions is shown as Equation 2.
18
𝑂
( 16 ) 𝑠𝑎𝑚𝑝𝑙𝑒
𝑂
𝛿 18 𝑂𝑠𝑎𝑚𝑝𝑙𝑒−𝑆𝑀𝑂𝑊 = ( 18 − 1) × 1000 (2)
𝑂
( 16 ) 𝑆𝑀𝑂𝑊
𝑂
3.1 Mass Spectrometry

Mass spectrometers were first developed in the early part of the 20th century, but
the precision was only sufficient for application to natural stable isotope variations after
World War II. Thus, was born the field of stable isotope hydrology (Gat, 1981; Sharp, 2007).
Today there are many types of mass spectrometers, with specialized designs and features
enabling a wide variety of analyses to be performed on samples as diverse as rocks, gases,
proteins and teeth. The laboratory preparation procedures are equally diverse for all these
substances. Hydrogen and oxygen isotopes of water are typically analyzed in gas-source,
light-isotope ratio, mass spectrometers (IRMS) specially adapted for the stable and light
isotopes, which include carbon, nitrogen and sulfur.
Mass spectrometers work in the following general way (Figure 5). First, the sample
is introduced into the machine, either as a gas, or by vaporizing a liquid or solid sample.
This gas or vapor is then ionized using a variety of techniques, such as chemical ionization,
electrospray ionization or inductively coupled plasma. The ions are then accelerated and
focused down the flight tube and through a strong magnet, where electrical and magnetic
fields cause deflection of the ions. This deflection is dependent upon the charge of the ion,
as well as its mass. For ions of the same charge, the heavier ions will deflect less, as they
possess more momentum than the lighter ions. The last part of the process is the detector,
containing collector cups, amplifiers and other electronics to record, process the signal, and
output ratios of the various masses of ions being emitted from the source.
8
Figure 5 - Schematic diagram of a gas-source, light-isotope-ratio, mass spectrometer, including the two main
types of sample injection setups, continuous-flow and dual-inlet. The m/q (mass/charge) ratios are illustrative
of the paths taken by various ions. Ions with lower m/q ratios are deflected more by the magnet.
For water, the gases H2 (for 2H) and CO2 (for 18O) are used as the inputs to the
mass spectrometer. Sample preparation procedures are outlined in Schimmelman and
others (1993) and Socki and others (1992), although other methods are available.
IRMS makes use of either a dual-inlet system, or a continuous flow system. Water
has generally been converted to gases, so the dual-inlet system is mostly used. For the
dual-inlet system, the samples are alternated with the reference gases, providing a high
level of precision. For the continuous flow system, a sample is analyzed only once and the
precision is therefore not as good. Laboratory standards are inserted every ten to twenty
samples, allowing correction of any drift in the reference gases. These laboratory standards
are calibrated to the international standards, such as SMOW, allowing the researcher to
present their data in an internationally recognizable form. Paul and others (2007) discuss
normalization of the sample data to the standards.
3.2 Laser Cavity Spectroscopy

This method is more recent than mass spectrometry and was developed in the 1980's
(O'Keefe, 1988). The first form of this method was known as cavity ringdown spectroscopy
(CRDS) and is still used, but more recent developments have broadened the method and
9
its application. One of the more recent modifications is called off-axis integrated cavity
output spectroscopy (OA-ICOS) and reduces the need for highly accurate mirror alignment
(Los Gatos Research, 2020). As with mass spectrometry, CRDS is not only used for isotopic
measurements, as it can be set up to determine the quantities of different gases in a sample,
or when modified with a plasma source can analyze for heavy element concentrations
(Wang, 2007).
CRDS works in roughly the following way (Figure 6). A cavity containing an inert
gas (He, N, etc.) is injected with the gaseous phase of a sample. A short burst of laser light
is emitted into the cavity, which contains at least 2 high reflectivity mirrors. The light
bounces between the mirrors to give an effective path length of many kilometers, in which
the light decays due to mirror inefficiency and absorption by the sample gas. By monitoring
wavelengths which are both affected and not affected by the sample absorption, the
instrument can compare the decay of light due only to reflection losses and that due to
absorption by the sample as well. This means variations in laser intensity and power that
occur between measurements are corrected for. Multiple measurements are taken of the
time taken for the light to diminish (the ringdown time) and therefore the strength of the
sample absorption can be calculated. This absorption factor can be converted into an
isotopic composition, as the different isotopic species have different absorptivity.
Figure 6 - Schematic diagram of a laser cavity ringdown instrument.
As with mass spectrometry, laboratory standards are inserted every ten or so

samples, to allow final corrections to be made to the CRDS results, and allow reporting to
the international standards, SMOW or SLAP.
10
4 Isotope Fractionation
The ratios of one isotope to another, such as 18O/16O, vary slightly between different
materials or even different reservoirs of the same substance. The magnitude of these
variations depends on the element concerned, the compounds, reactions and
environmental conditions, but typically range up to around 10‰. The variations are
greatest for hydrogen isotopes, up to 1000‰, because the mass difference is 2-fold, or 100
percent, for 2H/1H. As the mass difference decreases, so the isotopic variations tend to
decrease, such as with oxygen 18/16, which is a mass difference of an eighth, or 12.5 percent.
These differences in isotope ratios come about through various processes or
reactions, including chemical reactions, physical reactions (changes of state), diffusion and
exchange. A chemical reaction is when two or more elements or compounds react to form
different compounds; a physical reaction is where an element or compound undergoes a
change of state (gas to liquid to solid); diffusion is when atoms or molecules disperse from
high to low concentration through other material; exchange is when atoms of the same
element swap places from one compound to another without causing any chemical
changes. In all of these processes, molecules or atoms bearing different isotopes will
proceed through these reactions at different speeds, creating differences in isotope ratios in
the different materials.
Preferential location of the heavier or lighter isotopes (or isotopologues) will change
the isotopic abundances from the normal ratio. For example, 1H2 will preferentially diffuse
and escape to space at the top of the atmosphere, compared to 1H2H or 2H2. Isotopic
exchange is where atoms of the same element swap places in different molecules, for
example oxygen in H2O and dissolved CO2 in the ocean. Preferential location of the lighter
or heavier isotope occurs due to different bond energies, related to mass, for example the
preferential condensation of 1H218O(v) relative to 1H216O(v) during cloud formation ((v)
indicates a vapor phase, and (l) a liquid phase). These processes of differential accumulation
of isotopes are known as fractionation.
4.1 Kinetic Fractionation

For physical, chemical and exchange reactions, the dissociation energy of molecular
bonds controls the reaction rate. A bond involving lighter isotopes has a lesser dissociation
energy and can break more easily than the same bond involving a heavier isotope. One can
think of this as requiring more energy to break the bond with a heavier than a lighter
isotope. For example, when a body of liquid water evaporates into air, the resultant water
vapor (H2O gas mixed in air) will be relatively enriched in the lighter isotopes and therefore
have lower values for the ratios 2H/1H and 18
O/16O than the source body. This isotopic
fractionation will be exaggerated if the vapor is being removed rapidly, as happens in
11
windy situations or at low humidity when water evaporates from natural water bodies,
and is known as kinetic fractionation. When the air is drier or the wind is stronger, removal
of 1H and 16O bearing water molecules will increase, because there is less opportunity for
the evaporated water to react backwards with the liquid water body. Bear in mind, that on
the other hand, a lack of mixing in the source water body will result in a lesser fractionation
effect, because the surface water layer will become depleted in 16O with time.
For diffusion, the diffusive velocity is inversely proportional to the mass of the
molecule and therefore molecules with lighter isotopes will diffuse faster. For the above
example, if there is no wind, but diffusion is high because humidity is low, then again, the
increased diffusion of water molecules containing 1H and 16O will increase the kinetic effect.
For kinetic fractionation there is no fixed difference in isotope ratios between the
source and receptor reservoirs, as this is dependent upon factors such as time, degree of
removal of one reservoir and degree of mixing of the reservoirs (i.e., wind or diffusivity),
as well as the type of reaction, or change of state, that is taking place.
4.2 Equilibrium Fractionation

If chemical, physical or exchange reactions are allowed to run to completion, then
there will be a fixed isotope difference between the source and receptor reservoirs, at a
given temperature. Temperature plays an important role in fractionation. The higher the
temperature, the less the degree of fractionation (Figure 7). Once equilibrium is reached,
reactions will continue, but backward and forward reactions will occur at an equal rate,
with no net effect on isotopic composition of either reservoir. For example, at a given
temperature, there is a fixed difference between the 18O/16O ratio in H2O(l) and H2O(v) in
equilibrium with each other. In this situation, the kinetic effects of reaction rate are not
important, and it is the relative preference for a heavier or lighter isotope within a chemical
bond that determines which isotopes are located where, with heavier isotopes favored in
bond positions with higher strength. To continue the above example, the hydrogen bonds
between water molecules are stronger between H218O--H216O than H216O--H216O, and so
evaporation will preferentially select for H216O in the vapor mass, resulting in the H2O(v)
having lower values for 18O/16O than the H2O(l).
12
Figure 7 - Temperature dependence of fractionation. Graph a) shows the

fractionation factor decreasing with rising temperature, and b) shows the difference
in isotopic composition between a water source and its evaporated moisture
becoming less at higher temperatures, assuming equilibrium conditions.
4.3 Fractionation and Enrichment Factors

To quantify the fractionation of isotopes between two phases or compounds, the
fractionation factor, , is used as shown in Equation 3.
𝑅𝑟𝑒𝑎𝑐𝑡𝑎𝑛𝑡
𝛼 = (3)
𝑅𝑝𝑟𝑜𝑑𝑢𝑐𝑡
where:
R = isotope ratio, such as 2H/1H (dimensionless)

For example, as expressed by Equation 4.
2
𝐻
( 1 ) 𝑤𝑎𝑡𝑒𝑟
𝐻
𝛼 2𝐻(𝑤𝑎𝑡𝑒𝑟−𝑣𝑎𝑝𝑜𝑟) = 2 (4)
𝐻
( 1 ) 𝑣𝑎𝑝𝑜𝑟
𝐻
This factor describes the partitioning of an isotope between two phases or
compounds, which is determined by the temperature, chemical bonds and other atomic
scale properties of the element. Importantly, the fractionation factor is temperature
dependent; in other words, at equilibrium, the isotope ratios in the reactant and product
vary with temperature (Figure 7), as was first outlined by Urey (1947). Chacko and others
(2001) provide more recent and detailed information on fractionation factors and their
calculation.
Calculating the fractionation factor for a given temperature is usually done
experimentally with enough measurements at different temperatures to develop an
equation such as Equation 5.
𝑙𝑛 𝛼𝑋−𝑌 = 𝑎𝑇 −2 + 𝑏𝑇 −1 + 𝑐 (5)
where:
αX-Y = fractionation factor for transition from X to Y (dimensionless)
13
X = subscript indicating reactant compound or state

Y = subscript indicating product compound or state
T = temperature (K)
2
a, b, c = fitting constants (K , K, dimensionless, respectively)
For most isotopic reactions, the fractionation factor, , is close to 1, meaning ln()
will be close to 0. If this is multiplied by 1000, then permil units are obtained.  values are
given in permil units. As long as  is indeed near 1, an enrichment factor (ε) can be
calculated (Figure 8). The enrichment factor expresses the difference in  values between
reactants and products (Clark, 2015), as shown in Equation 6.
𝜀 = 1000(𝛼 − 1) ≈ 1000𝑙𝑛(𝛼) (6)
Figure 8 - Graph a) shows the isotopic evolution of a body of water evaporating to vapor over time, at a
set temperature and under equilibrium conditions. In this case, the enrichment factor remains constant as
the remaining water and the instantaneous vapor evolve. Also, eventually the final total vapor body,
assuming no losses from the system, has the same isotope composition as the initial water body. Graph
b) shows the isotopic evolution of vapor, cloud and precipitation, for constant temperature and equilibrium
condensation.
Importantly the variation in equilibrium fractionation factors is systematic, because

lower temperatures always result in higher fractionation factors and vice versa. This
temperature dependence is the fact enabling paleoclimate work. The isotopic composition
of a material, such as ice in a glacier, depends on the temperature at which the snow or ice
crystallized. Calcite in a speleothem can also be used, as the calcite crystallization occurs at
a given temperature in the presence of groundwater, which once was precipitation. With
enough measurements to establish an accurate average, the isotopic composition of the
original precipitation will be known, and if an age can be calculated for the glacier or
speleothem, then the temperature at a specific time and place can be reconstructed.
14
4.4 Rayleigh Distillation

Around the turn of the 20th century, Lord Rayleigh proposed a law to describe the
changing composition of an original pool of mixed liquids as it is progressively distilled
(Lord Rayleigh, 1902). This law relied on the fact that the differential rate of evaporation of
the two liquids is known. This was before stable isotopes had been discovered, but it turns
out to perfectly predict the theoretical rate at which a reservoir of water will change isotope
composition as it loses water to another reservoir through evaporation, freezing, etcetera.
The equation Rayleigh proposed is show here as Equation 7 (Clark, 2015).
𝑟𝑓𝑖𝑛𝑎𝑙 = 𝑟𝑖𝑛𝑖𝑡𝑖𝑎𝑙 × 𝑓 (𝛼−1) (7)
where:
rfinal = isotope composition of the pool of water with fraction f remaining
(dimensionless)
rinitial = original isotope composition of the pool (dimensionless)
f = fraction (1 to 0) of water that remains (dimensionless)
 = equilibrium fractionation factor for the reaction at a given temperature
for example, 2H/1H in liquid to water vapor (evaporation) at 25 °C
(dimensionless)
If an imaginary block of moisture laden atmosphere (a vapor body) has water

removed by condensation to form a cloud, there will be an isotopic difference between the
vapor and cloud exactly equivalent to the equilibrium fractionation factor, for both H and
O isotopes. However, this only holds for the first batch of condensation, because the
remaining vapor will immediately change its isotope composition due to the fractionation
factor causing heavier isotopes to preferentially condense into cloud droplets. This leaves
the vapor body isotopically lighter, which means future condensate will be lighter
(Figure 8).
If we envision someone fishing tries to catch the biggest fish in the sea, even though
she consistently catches fish that are 1 kg heavier than the average weight of all the fish (the
fractionation factor), then as those fish are removed, the average weight of fish in the sea
decreases, and so will the weight of each fish in her catch. Sadly, this analogy for Rayleigh
distillation is also the true story of fish in the oceans over the last century or so.
Rayleigh distillation is applicable to equilibrium situations, such as cloud
formation, where vapor and cloud droplets are in contact long enough for equilibrium
fractionation to apply. In settings with orographically driven rainout in a high precipitation
location such as the Azores in the north Atlantic which receives 1500 mm/a (Antunes et al.,
2019) or for heavy frontal precipitation in California (Mix et al., 2019), Rayleigh distillation
gives a reasonable approximation of the stable isotope behavior. In higher latitudes and
15
deeper continental locations, for example the Canadian Rockies (Figure 9), an
orographically adjusted Rayleigh model is better for predicting isotope composition of
precipitation than a linear Rayleigh distillation (Sinclair et al., 2011).
Figure 9 - Regional investigation of the applicability of Rayleigh distillation models to predict stable
isotope content of precipitation. This study used three models to predict the stable isotope
composition of snowfall in the Canadian Rockies, at two sites, the Opabin and Haig Glaciers. Actual
measurements from snow pits are precipitation weighted means. The three models are the
NARR-RM (North American Regional Reanalysis - Rayleigh model), the LRM (linear Rayleigh model)
and the ORM (orographic Rayleigh model). The ORM can be seen to produce values closest to the
actual precipitation seen in the snow pits (after Sinclair et al., 2011).
For a fractionating process, such as a chemical reaction or diffusion in a porous

medium like an aquifer, Rayleigh distillation overestimates the changes to the original
reservoir, or the reactants. This is due to the rate limiting effects of transport through the
aquifer matrix or kinetic reaction effects. Use of a lower fractionation factor partly resolves
this problem, but calculation of this lower fractionation factor is complex (Druhan and
Maher, 2017).
16
5 Meteoric Water Lines

5.1 The Global Meteoric Water Line
Stable isotope data from precipitation samples fall on a line known as a meteoric
water line. This line is usually plotted with data from local, regional or global precipitation
samples, but can include surface and groundwater samples. The Global Meteoric Water
Line (GMWL) was first recognized by Craig (1961), based on fresh surface water samples
from around the world (Figure 10), and is represented by Equation 8.
2H = 8 18O + 10 (8)
Figure 10 - The Global Meteoric Water Line (GMWL) as discovered by Craig (1961).
The spread of data along the GMWL is influenced by several meteorological

processes or factors, such as humidity and temperature, but it is probably rainout of
atmospheric moisture, as air masses move from the tropics to the poles, that accounts for
the bulk of the variation in  values (Rozanski et al., 1993; Yurtsever and Gat, 1981). As a
moisture laden air mass moves from the tropics to the poles, moisture is removed by
precipitation and the temperature tends to be lower. These lower temperatures not only
encourage further condensation causing more precipitation, but enhance the removal of
heavy isotopes by increasing the equilibrium isotope fractionation factors. As the process
of rainout is governed by condensation, which is an equilibrium process, 2H-18O co-vary,
but with a factor of 8 difference. This factor of 8 difference is due to the ratio of the
17
fractionation factors for H and O during rainout, which in turn is due to the difference in
mass of 2H/1H being 8 times greater than the difference in mass of 18O/16O.
18𝑂 − 16𝑂 2 1 2𝐻 − 1𝐻 2 −1
16𝑂 = 16
= 8
and 1𝐻 = 1
= 1
Thus, the variations in 2H will be roughly 8 times those of 18O. The difference in
mass causes a difference in energy needed to break the bonds, and that is the ultimate cause
of fractionation. However, temperature changes the 2H/1H and 18
O/16O equilibrium
fractionation factors differently, so the gradient of 8 steepens in colder regions and lessens
in warmer regions (Clark, 2015). The value of 10 is the intercept, or the 2H value when 18O
is 0.
The GMWL was updated by Rozanski and others (1993) and then again by
Araguas-Araguas and others (2000) to 2H = 7.96 18O + 8.86.
5.2 Local Meteoric Water Lines

A meteoric water line for a region is known as the Local Meteoric Water Line
(LMWL) and is usually essential for interpreting stable isotope data. It is ideally generated
by using only precipitation samples (not surface or groundwater samples), and samples
that capture all of the precipitation in a region over several years, to be representative of
unusual weather events and climatic oscillations, such as El Niño or the North Atlantic
Oscillation.
The LMWL can be thought of as representing the default or input stable isotope
values to the local hydrological cycle. It is different from the GMWL because the local area
has a limited range of climatic conditions and weather systems, thus produces a limited
and more specific range of stable isotope values compared to the global dataset
(Araguas-Araguas et al., 2000). Comparison of groundwater, surface water and particular
precipitation events to the LMWL allows investigation of processes such as water-rock
interaction, evaporation, recharge and mixing (Jasechko, 2019) as illustrated in Figure 11.
18
Figure 11 - A conceptual diagram showing typical isotope compositions of various water bodies and
differences in the LEL (local evaporation line), LMWL (local meteoric water line) and GMWL (global meteoric
water line) for a specific region. These differences in slope and position allow interpretation of stable isotope
data so as to begin formulating hypotheses about water sources, sinks and interactions in a study region.
Most LMWLs have slopes of < 8, usually around 5 to 7 and a notably limited range
of 2H and 18O values compared to the GMWL. When several LMWLs for areas with
different climates are drawn, these lines lie semi-parallel, but are displaced 'up' or 'down'
on a 2H-18O plot and stack adjacent to each other to form the GMWL, as shown in
Figure 12. The GMWL is the cumulative result of all the LMWLs for regions of different
climate, with degree of rainout being the main discriminant for the position of each LMWL.
LMWLs for higher latitude regions tend to plot lower on a 2H-18O diagram whereas
LMWLs for lower latitude and more arid regions tend to plot higher on the diagram.
19
Figure 12 - Local meteoric water lines (LMWLs) for various regions of the world, and the GMWL.
5.3 Calculation of Meteoric Water Lines

A meteoric water line is a best fit line (i.e., a regression line) for a set of points in x-y
space. The most common regression analysis is known as the least squares method. This
assumes the x-variable is independent and accurately known, whereas the y-variable
depends upon the x-value and has errors and random variations. An example of an
independent x-variable would be time or distance, and a dependent y-variable could be
temperature, precipitation or an isotope ratio. For analysis of one stable isotope ratio, say
2H against one of these independent variables, like time, a least squares regression is a
suitable representation of the relationship. However, where both variables are dependent,
such as 2H and 18O, neither one should be treated as more certain than the other and so
the reduced major axis form of a structural regression is suitable (Ma, 2019).
20
5.3.1 Least Squares Regression

The straight-line regression is of the form shown in Equation 9.
y = mx + c (9)
where:
m = gradient, i.e., slope of the line (dimensions of y-axis values over

dimensions of x-axis values)
c = y-intercept (dimensions of y-axis values)
x, y = regression variables
the least squares regression approach produces Equation 10.
𝑆𝑃𝑥𝑦
𝑚 =
𝑆𝑆𝑥
𝑐 = 𝑦̅ − 𝑚𝑥̅
𝑛
𝑆𝑆𝑥 = ∑(𝑥𝑖 − 𝑥̅ )2
𝑖=1
𝑛
𝑆𝑃𝑥𝑦 = ∑(𝑥𝑖 − 𝑥̅ )(𝑦𝑖 − 𝑦̅) (10)

𝑖=1
where:
SSx = sum of the squared deviations from the mean of x (squared dimensions
of the x variable)
SPxy = sum of product of deviations from the mean of x and deviations from
mean of y (dimensions are product of x and y dimensions)
5.3.2 Reduced Major Axis Regression (RMA)

The RMA regression line is calculated in a manner similar to that shown in Section
5.3.1, with the single difference that the gradient, m, is calculated as shown in Equation 11.
𝑆𝑆𝑦
𝑚𝑅𝑀𝐴 = √
𝑆𝑆𝑥
(11)
𝑛
2
𝑆𝑆𝑦 = ∑(𝑦𝑖 − 𝑦
̅)
𝑖=1
5.4 Weighted Regression Line Equations

As noted by Hughes and Crawford (2012), weighting of isotopic values for monthly
cumulative precipitation by the precipitation amount produces regression lines (meteoric
water lines) with higher gradients, as a result of minimizing the influence of evaporated
21
samples from low precipitation events. The difference in gradient and intercept between
weighted and unweighted regression lines depends on the dataset and the regression
method, and although the differences are generally minor, they can be significant (Boschetti
et al., 2019). These meteoric water lines better characterize the average precipitation and
especially heavier events that are more likely to play an important role in hydrological
processes such as groundwater recharge (Li et al., 2018). These weighted regressions use
methods similar to those presented in Section 5.3, but include a precipitation term in the
statistical quantities as shown in Equation 12.
𝑛
𝑆𝑆𝑥 = ∑(𝑟𝑎𝑖𝑛𝑖 )(𝑥𝑖 − 𝑥̅ )2

𝑖=1
𝑛
𝑆𝑆𝑦 = ∑(𝑟𝑎𝑖𝑛𝑖 )(𝑦 − 𝑦̅)2 (12)

𝑖=1
𝑛
𝑆𝑃𝑥𝑦 = ∑(𝑟𝑎𝑖𝑛𝑖 )(𝑥𝑖 − 𝑥̅ )(𝑦𝑖 − 𝑦̅)

𝑖=1
It is useful to be aware that some computer statistical programs will perform least
squares regression as their standard method to produce a best fit line. Thus, it is essential to
either: 1) know what the program is doing; or, 2) set up the calculations in a spreadsheet
using the formulae of Equations 10 - 12.
22
6 Stable Isotope Hydrology

The global water cycle is extremely complex if all the interactions with geological
and biological materials are taken into account. The water cycle includes processes as
diverse as weathering and volcanic eruptions in the geosphere, and organic decay and
drinking by animals in the biosphere. Fortunately, the flows of water involved in those
geological and biological interactions are orders of magnitude less than the major flows of
water, such as evaporation from the oceans or precipitation over land, except for
transpiration by plants (Jasecko et al., 2013). This global flow of water is known as the
hydrological cycle. A simplified quantification of the hydrological cycle is shown in
Figure 13.
Figure 13 - A simplified quantification of the global hydrological cycle (after Reeburgh, 1994). Pools are
reservoirs that store water such as the atmosphere, glaciers, lakes, rivers, soil water, groundwater and the
ocean. Fluxes between pools are indicated by arrows.
Most of the fluxes from one pool to another are key points where the isotopic
composition of parcels of water change. An approximation of the range of isotope
compositions across the hydrological cycle is depicted in Figure 14.
23
Figure 14 - Estimates of typical stable isotope compositions of a variety of water sources, reservoirs and
flows:  H in yellow,  O in red. Values are approximations based on information in relevant literature so,
2 18
the values may differ widely from actual values for specific locations. The range of isotopic compositions
shown here is less than those that occur in field settings, because these are averaged values. Nonetheless
they provide an indication of the range of  H and  O values that occur in the hydrological cycle. This wide
2 18
range of values, especially where there is a large change in climate, elevation or distance, is what makes
stable isotopes useful for tracing water sources and movement.
With such complexity in the hydrological cycle, it is perhaps surprising that the
variation in 2H-18O is very defined, as shown in Figure 10 from the landmark article in
Science by Harmon Craig in 1961. The most important feature of note in Figure 10 is that
most precipitation has 2H and 18O values less than zero. This is primarily because
evaporation from the oceans produces vapor that is depleted in the heavier isotopes relative
to sea water (which is very similar to SMOW and has 2H and 18O values close to zero).
Evaporation from the oceans is a two-stage process. Initially, equilibrium
fractionation occurs where water evaporates into a saturated boundary layer over the ocean
surface. Then, from this saturated layer, diffusion and wind continuously remove vapor
and mix it upward into the atmosphere, causing an additional kinetic fractionation (Craig
and Gordon, 1965). Thus water vapor in the atmosphere is more depleted in the heavier
isotopes than would occur if only equilibrium fractionation was taking place (Clark and
Fritz, 1997). The measured values of 18O in vapor over the oceans vary from about -10
to -15‰, as latitude increases (temperature decreases). These values are about 4‰ less than
the equilibrium values would be. The values for 2H are similarly lower than theoretical
equilibrium values and range from -70 to -100‰ (Sharp, 2007).
24
Generation of atmospheric water vapor occurs mainly over the warmer oceans and
it has been estimated that around 65 percent is generated between 30°S and 30°N (Peixoto
and Oort, 1983). Once an air mass is cooled, either by advection over a cool surface in colder
climates or by convection, the air can become saturated and condensation may commence.
Condensation is dependent not only on temperature and humidity, but on the availability
of condensation nuclei of the correct type (Sumner, 1988). Condensation generally proceeds
slowly in response to reduced pressure or further cooling and this takes place in the
presence of the vapor. In other words, cloud droplets are surrounded by vapor when they
form and continuously exchange with the vapor, thus condensation is an equilibrium
fractionation process.
Only general statements can be made about the causes of the distribution of isotope
ratios in meteoric water because the hydrological cycle is complex. Attempts have been
made to develop equations or conceptual models that can predict isotopic values for
precipitation, but these have not been adequate (Sharp, 2007 and Yurtsever and Gat, 1981).
However, several key factors have been identified by the first workers to interpret stable
isotopes in water samples, for instance Friedman (1953), Epstein and Mayeda (1953) and
Craig (1961), culminating in Dansgaard (1964). Some of these have been alluded to in
previous sections. Subsequent workers continued to develop the understanding of these
key factors, until they became widely accepted as a fundamental part of isotope hydrology
(Gat, 1996). These factors, or isotope effects, are known as the temperature effect, latitude
effect, continental effect, altitude effect and amount effect. With decreasing temperature,
increasing latitude, increasing distance from the coast, increasing altitude or increasing
amount of rain in a precipitation event, the isotopic composition of the precipitation
becomes lighter (depleted in 2H and 18O). The cumulative result of all these effects, and
others specific to local settings, can be seen in the global distribution of stable isotope
compositions in groundwater, superbly collated by Jasechko (2019) and shown in Figure 15.
25
Figure 15 - Global distribution of  O in groundwater: a) shows groundwater only; b) contains both groundwater and precipitation data. This graph summarizes nearly
18
48,000 measurements, from hundreds of publications over several decades. Image from Jasechko (2019).
26
6.1 Environmental Factors Affecting Water Isotopes (Isotope Effects)

6.1.1 The Temperature Effect
The temperature effect is the positive correlation between local air temperature and
local precipitation stable isotope  values. Neither the temperature nor the stable isotopes
are instantaneous values, but rather are the long-term (annual) means, to avoid sharp
fluctuations that reflect only temporary factors related to a particular storm (cloud base
elevation, precipitation intensity, and so on). The temperature effect results in water having
more negative  values in areas with lower temperatures, and is due to several factors. First,
evaporation from colder oceans, in mid-latitude (and polar) regions, is subject to a greater
fractionation factor, because of the colder temperature, and produces vapor with more
negative  values than those from warmer oceans. Second, condensation in colder clouds
will produce more complete condensation of the vapor present, thereby incorporating more
lighter isotopes. Third, the progressive rainout from low to high latitudes occurs in parallel
to decreases in temperature, and so results in a higher correlation between T and .
Araguás-Araguás and others (2000) noted this T- correlation for Vienna's
precipitation, as shown in Figure 16. The correlation was improved by statistical smoothing
and manipulation, mainly to remove the seasonal variation so that the interannual changes
in T and  values could be displayed, and to show the change in T or  versus the average,
rather than the absolute T or  values.
Figure 16 - The relationship between surface air temperature and  O in Vienna from 1960-1996. Statistical
18
removal of seasonality was accomplished by creating 12-month running means for T and  O, after which
18
18
ΔT and Δ O were calculated by subtracting the monthly running averages from the long-term means, and
finally the curves were smoothed by reapplying a 12-month running mean (after Araguás et al., 2000).
27
The temperature effect manifests relatively well in high latitudes, but is not as
strong in tropical regions, where the amount effect dominates (Jasecko, 2019; Rozanski et
al., 1993; Yang et al., 2011).
6.1.2 The Latitude Effect

Most evaporation occurs over the tropical oceans (an estimated 65 percent as
reported by Peixoto and Oort, 1983) because the higher temperature of the sea allows
greater evaporation than in middle or high latitude ocean areas. Consequently,
atmospheric moisture evolves isotopically as it moves away from the tropics. Condensation
and rainout favor the removal of the heavier isotopes and so the precipitation at higher
latitudes has more negative  values. Evaporation occurs from the mid-latitude oceans and
because the temperatures are colder, the fractionation factors are greater, resulting in vapor
relatively more depleted in the heavier isotopes than vapor forming above the tropical
oceans.
The latitude effect is only easily perceptible over the continental to global scale, as
the myriad local variations in moisture source, humidity, precipitation amount, altitude
and so forth, create a large amount of "noise" when looking for the latitude- trend. As such,
the latitude effect is mainly of interest in modeling global patterns of precipitation (Jasecko,
2019; Rozanski et al., 1993), but can be detected regionally in some cases (Laonamsai et al.,
2020) as shown in Figure 17.
28
Figure 17 - Stable oxygen isotopes for precipitation in Thailand based on monthly cumulative samples
from 2013 to 2015. Rainout causes depletion of heavy isotopes with distance, which: a) in the case of
the SW monsoon is the typical negative correlation with latitude; while b) in the case of the NE monsoon,
is a positive correlation. During the NE monsoon, the weather systems move south-westward, or right to
left in b), creating a continental effect where isotopic rainout over Vietnam and Laos lowers the delta
values of rainfall with decreasing latitude (after Laonamsai et al., 2020).
6.1.3 The Continental Effect

Progressive rainout is the main cause of increasingly negative  values for
precipitation that is further and further inland (Figure 18). The continental effect is seen as
a decrease in delta values of precipitation with distance from the coast, as shown in Table 2.
29
In some cases, where winter precipitation occurs, cooler air inland may also reduce the
amount of evaporation and isotopic change that occurs as rain drops fall through
unsaturated air below the cloud. These colder inland temperatures will also increase the
equilibrium fractionation factor that applies during condensation, thus removing heavier
isotopes more effectively from the vapor and resulting in precipitation further inland being
even lighter isotopically.
Figure 18 - Schematic diagram illustrating both latitude and continental effects. Generalized values of  O are
18
given, and these apply to both the evolution of moisture away from the tropics towards the poles (latitude effect)
or evolution of moisture inland from the coast to the interior of a continent (continental effect). In both cases,
progressive rainout (depletion of atmospheric moisture) is the key driver of the isotope composition.
In other cases, rainout can be so effective at removing the heavier isotopes that the
temperature effect, due to seasonality, is overridden by the continental effect. For example,
precipitation during summer (June to September) on the Tibetan Plateau is 6‰ (18O)
lighter than in winter when temperatures are 10 C colder (Araguás-Araguás et al., 1998).
The continental effect is best observed over continental scales, but does operate at smaller,
regional scales (Jasecko, 2019; Laonamsai et al., 2020).
Table 2 - Some examples of the continental effect from around the world. The effect is given
as a gradient, i.e., the change in mean  values of precipitation per 1000 km of distance.
Gradient
Species ∆‰ Location Reference
𝟏𝟎𝟎𝟎𝒌𝒎
D -13 Europe: Belgium to Poland - summer Rozanski et al., 1982
D -33 Europe: Belgium to Poland - winter Rozanski et al., 1982
18
 O -1.6 Europe: Poland to Russia Rozanski et al., 1993
18
 O -3.8 Europe: Poland to Russia Rozanski et al., 1993
18
 O -3 to -4 North America: Atlantic to Rockies Clark and Fritz, 1997
18
 O -10 Canada: Pacific to Prairies Yonge et al., 1989
18
 O -0.75 Amazon: Atlantic to Andes Salati et al., 1979
30
6.1.4 The Altitude Effect

As with the continental effect, the altitude effect is caused mainly by rainout, in this
case triggered by orographic uplift, as well as a decrease in temperature, resulting in greater
fractionation factors, which will drive rainout of heavier isotopes and cause a faster shift to
lighter isotopes with altitude. Also, rain falling at higher elevations will have less distance
to travel to the ground and less chance for evaporative enrichment, in which the lighter
isotopes preferentially evaporate. The altitude effect is shown as a decrease in  values per
100 m elevation gain in Table 3.
Table 3 - Some examples of the altitude effect from around the world. The effect is given here as a gradient:
the change in  values of precipitation per 100 m increase in elevation. The range of elevation over which
the precipitation was sampled is also given.
18
 O Gradient Altitude
Location Country ∆‰ Reference
masl
𝟏𝟎𝟎 𝒎
Mount Cameroon Cameroon -0.16 0-4000 Gonfiantini et al., 2001
Eastern Andes Bolivia -0.24 200-5200 Gonfiantini et al., 2001
Hérault France -0.27 500-1800 Ladouche et al., 2009
Whole Island Taiwan -0.20 0-2500 Peng et al., 2010
Fuego Volcano Guatemala -0.67 800-1200 Mulligan et al., 2011
Table Mountain South Africa -0.075 100-1100 Diamond & Harris, 2019
Mount Shasta California, USA -0.21 1000-3100 Peters et al., 2018
The altitude effect for an area is derived by collecting precipitation in at least two
locations close to each other, but at different elevations. If using only two locations, then
several precipitation seasons (several years) of data should be collected so as to average out
the differences between years or unusual years caused by climatic variability. It is
important to collect all rain that falls, and to weight the data for each sample (be they daily
or monthly precipitation samples) by precipitation amount, to get a representative average
isotope composition. If more precipitation stations are used, then sampling can be
undertaken over a shorter period, but ideally there should be several precipitation stations
sampled over several rainy seasons, including climatically different years, for example El
Niño and La Niña fluctuations.
The altitude effect creates large enough variations in stable isotope compositions to
allow stable isotopes to be used as tracers over fairly small distances (kilometers to tens of
kilometers). This is commonly used to delineate recharge locations for groundwater
discharging at springs or boreholes at lower elevations (Blarasin et al., 2020; Diamond and
Harris, 2019; Jasecko, 2019).
6.1.5 The Amount Effect

The amount effect also has a close relationship with rainout. The amount effect
manifests as a shift to lighter isotope compositions for heavier precipitation events
(Figure 19). First, heavy individual rainstorms will tend to remove more of the vapor and
cloud droplets in the air, and so with increasing precipitation in one location, the isotopic
31
signature should become lighter. Second, the air below the cloud base will gradually
become more saturated and colder, as rain and air from higher in the cloud descends (the
downdraft), both of which will reduce evaporative enrichment of the later rain drops. The
amount effect is known to be more pronounced in low latitudes (Dogramaci et al., 2012;
Rozanski et al., 1993; Yang et al., 2011).
Figure 19 - Oxygen isotopes of precipitation in winter (triangles) and summer (circles) for
a) Guangzhou and b) Changsha from June 2006 to May 2009. The negative correlation with
precipitation amount is apparent and has been quantified in the equations displayed on the
graphs (from Yang et al., 2011).
Of the 5 isotope effects, temperature and rainout are the main underlying processes
that drive the various 'effects'. It is important to note that all these effects and their
underlying causes occur in a highly complex natural system where many variables
contribute to the final isotopic composition of a rainwater sample. Other than temperature
32
and rainout, factors such as humidity, storm track, preceding atmospheric conditions and
source region also modify the isotopic composition. Isotope content of rainwater varies by
the minute in a rainstorm (Muller et al., 2015) and between rain events, as is typical of most
meteorological phenomena (e.g., temperature, cloud formations, precipitation amount,
storm duration). Averaging the isotope composition of precipitation over longer periods,
such as a month, has been found to be the most useful way of understanding the variation
in isotopic signatures in an area (Yurtsever and Gat, 1981).
6.2 The Deuterium Excess

Kinetic fractionation during evaporation from the ocean surface takes place because
of diffusion of water vapor molecules from a saturated boundary layer at the sea surface
and into the open atmosphere. The 1H1H16O isotopologue diffuses faster than all the others
and so the vapor in the open atmosphere is depleted in the heavier isotopes. If the
atmosphere was saturated, then isotope exchange would occur fully between the sea and
water vapor in the air, resulting in isotopic equilibrium.
Importantly, there is a difference between the diffusion fractionation factors for
2
H-1H and 18
O-16O and so the relative depletion of 2H and of 18
O is not the same. This
difference between the rate of 2H depletion and 18O depletion changes with the degree of
saturation, or relative humidity, h. In other words, at lower humidity, 2H is not affected by
evaporation as much as 18O, which means vapor at lower humidity will be further from the
GMWL. As h increases, more isotope exchange will occur and the fractionation the system
will come closer to equilibrium (Clark and Fritz, 1997).
This means the position of a LMWL is partly a function of the relative humidity of
the source region, where evaporation generated the vapor mass. Evaporation under low
relative humidity will generate lines displaced further left of the GMWL. Figure 20 shows
how evaporation from sea water under 85 percent relative humidity conditions and then
condensation at equilibrium generates water with isotopic compositions that plot along the
GMWL. At various values of relative humidity, vapor and the resulting water samples will
be displaced from the GMWL to a greater or lesser degree and will have an intercept on the
2H axis that is larger or smaller than 10. This intercept, known as the deuterium excess
(D-excess, or d) of a water sample is calculated with Equation 13.
D-excess = 2H - 818O (13)
where:
2H and 18O = values for the water sample
33
Figure 20 - Different relative humidity in source regions during evaporation (red dotted lines)
create moisture masses with different deuterium excess (d) values due to kinetic fractionation. Sea
water lies at the 0,0 point on the graph. Evaporation of sea water at 100 percent humidity would
produce vapor somewhere along the red dotted line pointing to 100 percent humidity, depending
on the temperature. Higher temperatures would have lower fractionation factors and so the vapor
would be closer to the sea water point. At lower humidity, kinetic fractionation affects evaporation,
resulting in vapor plotting somewhere along the other red dotted lines, as shown for 85 percent
and 50 percent humidity. The average global humidity at the sea surface is approximately 85
percent, so the GMWL intercepts the  H axis at a d value of 10. For a hypothetical region with
2
evaporation occurring under 50 percent relative humidity conditions, d ~50 as indicated by the
extrapolation of the dashed blue line to  O = 0. During condensation and precipitation, the d
18
value remains similar for the vapor and the rain, even though rain or vapor samples will have very
different  H and  O values along the blue dashed lines. This may not be the case for very low
2 18
precipitation amounts and low humidity, for example in deserts (modified from Clark and Fritz,
1997).
The D-excess (d) is a proxy of the humidity of the source region, with lower humidity
causing higher d values. An example of this is shown in Figure 21.
34
Figure 21 - Hydrogen and oxygen isotopes for precipitation on Okinawa Island, south of Japan, from
April 2008 to April 2011. The distinct separation of the winter precipitation from the rest of the year is
shown by calculating a MWL for the winter precipitation only. The winter data has a higher d-excess
than the rest of the year, suggesting less humid conditions during evaporation in the moisture source
region, which was postulated to be due to local evaporation around the island into a relatively dry
continental air mass, compared to the other precipitation that was sourced in the tropical western
Pacific (from Uemura et al., 2012).
6.3 Event Variation

As stated in the introduction, knowledge of the whole hydrological cycle is often
needed to unravel groundwater questions, and although the cycle has no end or beginning,
the isotopic signature of precipitation is a good point to begin. This is particularly true for
groundwater, as precipitation precedes recharge. Most hydrogeological studies make use
of bulk isotope values for precipitation, typically monthly cumulative samples, as per the
IAEA GNIP (International Atomic Energy Agency, Global Network for Isotopes in
Precipitation) program, allowing calculation of a MWL, assessment of seasonality in
isotope values, and generation of the mean annual isotope composition. This approach is
sufficient for many groundwater applications.
However, some studies have analyzed the isotopic composition daily
(Sanchez-Murillo et al., 2016), per event (Qu et al., 2014), within-event (Harris et al., 2010)
and even down to 30 second increments (Munksgaard et al., 2012). The results show a
tremendous variation in isotope composition of rain over short timeframes. From one
location and a set of nine rain events totaling 15 days over 8 months, Munksgaard and
others (2012) reported -140 to +13‰ in 2H and -19.6 to +2.6‰ in 18O, capturing as much
35
variation as the monthly cumulative samples across the whole continent of Australia over
a period of 40 years.
Although the isotopic variation of a rain event is mostly dependent upon the type
of event and its duration, typical variations are on the order of 10-50‰ for 2H and 1-5‰
for 18O (Han et al., 2020; Harris et al., 2010; Muller et al., 2015). Different studies have noted
different trends. Harris and others (2010) found less  variation in a shorter (5 hour), heavier
rain event associated with a cold front from the South Atlantic passing over Cape Town as
opposed to a larger, more complex frontal system lasting more than a day. Qu and others
(2014) similarly found more variation in  values during larger storms for the subtropical
monsoon climate of eastern China, and Han and others (2020) saw variability decrease
during an event (Figure 22).
Figure 22 - Variation of oxygen isotope composition with time, during a single rain
event of several hours is revealed by analyzing 10-minute increment samples of the
precipitation. This event exhibits an L-shaped trend in isotope composition, as well
as stabilization of the variations with time (after Han et al., 2020).
From a review of previous work, Muller and others (2015) noted several intra-event
patterns of change in isotope composition over time. These were increasing, decreasing,
V-shaped, W-shaped, L-shaped and no-trend. The most common of these are the
36
increasing, decreasing, V-shaped and L-shaped. They also found a general relationship that
 values are inversely proportional to rain rate, in other words, heavier rain is associated
with lighter isotopes. This is not quite the same as the amount effect, where lighter isotopes
are associated with larger total precipitation amounts (and not higher precipitation rate).
For example, Sanchez-Murillo and others (2016), working in Costa Rica, where one would
expect an amount effect due to its tropical latitude, took daily precipitation samples and
found humidity and condensation levels to best correlate with isotope composition, and
not precipitation amount. In contrast, Han and others (2020) found relative humidity to
have no effect on isotope values, but saw that fresh input of an air mass into the weather
system can renew the system with heavy isotopes and make the  values of the event
increase. Adar and others (1998) hypothesized that rain from different parts of the storm
system can also cause variations in the isotope composition of rain.
In understanding infiltration, runoff and other parts of the water cycle, a single
sample for the total precipitation of a whole event is clearly going to obscure some of the
detail. In a desert setting in Israel, Adar and others (1998) saw that some events of < 1 mm
precipitation triggered runoff. If one wants to understand the detail on a site,
high-frequency, sub-event rain monitoring may be needed, and if trying to understand a
whole catchment, then sufficient spatial coverage may also be needed. Fortunately, the low
cost and high speed of analysis with laser cavity instruments makes this type of science
possible for researchers even on moderate budgets. The potential for application to
understanding vadose zone processes, ecological issues, triggers for fracture flow, and
other poorly understood aspects of hydrogeology, are tantalizing.
6.4 Mass Balance

The principle of mass balance can be applied widely across many disciplines, such
as meteorology, oceanography and geochemistry. In essence it applies when a set of inputs
merge perfectly and completely into a set of outputs. Quite often a system is simplified into
only 2 or 3 inputs and a single output. For example, precipitation, groundwater and surface
water could be the inputs to a river, which is the output. Assuming conservation of mass,
in other words, no losses along the way, then the river quantity and quality must comprise
the sum of the quantity multiplied by the quality, of all of the inputs.
For most calculations, the quantities can be flow rates in liters or cubic meters per
unit time, or something similar. The use of moles is necessary when dealing with different
compounds, but this is beyond the scope of this text and is seldom needed when dealing
with water only. For isotope compositions, it is generally fine to represent them as delta
values, but they are most accurately calculated using isotope ratios (Hayes, 2004). The
errors introduced by using delta values are caused by rounding errors and mole fraction
mixing, both of which are worse when there are large differences in delta values (personal
communication from Robert Kalin 2022). These discrepancies are mostly below analytical
37
or other errors, and therefore probably only need to be considered when using high
precision data (Steur et al., 2020).
Mass balance can be presented as an equation, here using stable isotopes ( values)
as the quality parameter in Equation 14.
(Q1  1) + (Q2  2) + (Q3  3) + ... = Qf  f (14)
where:
Q = flow rate (MT-1)
 = delta value (as in Equations 1 and 2) of isotope species of interest (e.g.,

H/1H) of each input (1, 2, 3, etc.)
2
1, 2, 3... = subscript for input streams

f = subscript for output stream
It is important that the stable isotope compositions differ enough in the end-members for
the method to work.
The simplest application of mass balance in stable isotope hydrology is to evaluate
the groundwater contribution to surface water during precipitation events, often called
storms. In this simple case, the mass balance equation is expressed by Equation 15.
(Qg  g) + (Qp  p) = Qs  s (15)
where:
g = subscript indicating groundwater

p = subscript indicating precipitation
s = subscript indicating resultant total stream flow
The isotopic composition of groundwater and pre-storm-event stream flow is
assumed to be the same, as the river is assumed to be totally fed by baseflow, in other
words, g = pre-storm-s. Note that these assumptions may often be close to the truth, but
may be quite incorrect in other cases. The H and O isotope compositions for the
precipitation (p), groundwater (dg) and storm flow (s) are measured, as is the total river
flow during the storm or flood event (Qs). This leaves 2 unknowns, Qg and Qp, but these are
related because they both add up to the streamflow. In equation form, this is as shown in
Equation 16.
Qg + Qp = Qs
(16)
 Qp = Qs - Qg
Then, by substituting into the original mass balance equation (Equation 15) and rearranging
the terms, results in Equation 17.
Qg = Qs(s - p) / (g - p) (17)
38
In this way, with only one flow measurement (stream discharge), and three isotope
values (precipitation, pre-storm streamflow and storm streamflow), we can calculate the
baseflow and precipitation contributions to storm flow in a stream. In reality, due to the
variability of isotope values in both precipitation and streamflow, multiple samples
produce a more reliable result. For precipitation, instead of multiple samples, a single
cumulative sample of all the precipitation during the storm can be taken, but this would
most likely miss out on a lot of the finer detail and perhaps some critical changes in the
contributions to flow that are known to occur (Xie et al., 2016).
This technique is suitable for small catchments with simple geology. Where
groundwater contributions from the vadose and phreatic zones may differ, or where there
is a large upstream catchment and the river flow may be expected to change stable isotope
values during the storm due to varied isotope composition coming downstream, this
method will need adjustment and more measurements will be required (Das et al., 2020).
Mass balance on a part of the Gariep River, the second largest river in southern
Africa, after the Zambezi, was demonstrated by Diamond and Jack (2018) using only
surface waters, as groundwater input was insignificant over short reaches. The stable
isotopes were measured upstream on two tributaries and downstream after being well
mixed at the location of a flow-gauging station. The contributions from the two tributaries
were then calculated using the above equations.
6.5 Hydrograph Separation

Hydrographs are charts showing the amount of water flowing in a river over time.
A hydrograph reflects the sum of the contributions to flow. Catchment water balance
theory tells us the river flow at a point has a set of possible contributions, which, when
there is no precipitation occurring and no flows provided by human activities, can only be
groundwater, from both the phreatic and vadose zone, the latter sometimes called soil
water or interflow. During a precipitation event or storm, the possible contributors also
include direct precipitation (onto the river surface) and overland flow or runoff. In general,
direct precipitation is negligible and interflow (from the vadose zone) is often assumed to
be negligible, and so this leaves runoff and groundwater as the main sources of river flow
during a storm as illustrated in Figure 23.
39
Figure 23 - Schematic diagram of the main flows of water in a natural catchment. Inclusion or exclusion of
each flow in a model depends on the degree of detail in the model, data availability and the catchment itself.
Hydrograph separation involves apportioning the river storm flow into

components. In the past, this was done graphically by comparing the baseflow (fairly
constant flow) and quick flow (rapidly rising and falling component of flow due to a storm)
portions of the hydrograph and assuming that quick flow was mostly due to runoff. The
assumption that the precipitation from a rain event is transferred to the stream via runoff
was shown to be inaccurate in many parts of the world, especially after analysis of
conservative tracers (e.g., H and O isotopes, Cl) in the water (Sklash and Farvolden, 1979;
Hooper and Shoemaker, 1986). Although some hydrologists still use graphical methods to
attempt hydrograph separation, many are of the opinion that this is as misleading as it is
beneficial, with Beven (2001) stating "the best method of dealing with hydrograph
separation is to avoid it altogether".
Natural tracers, such as dissolved ions and isotopes, both stable and radioactive,
provide a different way of performing hydrograph separation. Each component or possible
contributor to river flow can be characterized, such as groundwater and precipitation. For
a 2-component system, the total river flow is then compared to these using a mixing line
between the components and can provide an estimate of the proportions of groundwater
and rainwater in the river. Isotopes in particular, due to their non-reactivity in aquifers, are
ideal tracers, better than most dissolved species, as the latter often undergo various
reactions and transitions. Certain conditions should be met to apply this method
successfully, as stated by Sklash and Farvolden (1979).
40
“(1) The isotopic content (18O, D or T) of the event component is significantly

different from that of the pre-event component.
(2) The event component maintains a constant isotopic content.
(3) The groundwater and vadose water are isotopically equivalent or vadose
water contributions to runoff are negligible due to hydrogeologic constraints.
(4) Surface storage contributes minimally to the runoff event.”
If these guidelines are applied (Laudon and Slaymaker, 1997), hydrograph
separation using natural tracers may be reasonably successful. However, the study area
may deliver complexities that cannot be easily resolved, or the study itself may be
inadequate, rendering this simple approach less valuable (Klaus and McDonnell, 2013). To
be more specific, Hooper and Shoemaker (1986) noted isotope compositions of various
components to be indistinguishable at times and snowmelt contributions to vary over time,
in a study in New Hampshire. The co-variation of 2H and 18O has been used to dismiss
use of both of these tracers, however, Lyon and others (2009) find that each element gives
a different result. Use of both is not as robust as two totally independent tracers, for
example 18O and SiO2, but it is better than using only one. Also, as mentioned, the stable
isotopes are least likely to undergo changes from reactions with the aquifer, other water
types or the atmosphere.
One of the major challenges is estimating the isotopic compositions of the different
components, also called end-members. Pu and others (2017) observed how sensitive their
calculations were to their end-member estimates, and Penna and Van Meerveld (2019) used
the observed variations in values for end-members from 148 studies to calculate the typical
error due to natural variations of isotopes in groundwater, precipitation, streamflow and
other components. They computed a value of 26 % for the variation, or error, in the
estimates of final flow components. Another common issue is the number of components.
In many cases flow originates from more than groundwater and precipitation. Snowmelt,
glacier melt, surface storage (pre-existing water that becomes a flow component after
precipitation) and vadose zone water may all contribute significant flows during an event.
In these cases, a 3-component hydrograph separation is more likely to capture the main
sources of flow (Klaus and McDonnell, 2013), such as groundwater, meltwater and runoff
for a study on the Ganga River in the Himalaya (Maurya et al., 2011) or glacier melt,
snowmelt and runoff at the Gangotri Glacier, also in the Himalaya (Rai et al., 2019). The
latter study used a 2-component model on non-rainy days to determine that glacier melt
and snowmelt were two distinguishable components, and therefore were able to apply the
more complex 3-component model in rainy weather.
The results of isotopic hydrograph separation are therefore variable in accuracy,
and often incorporate hydrochemical parameters, as does much other work in stable
isotope hydrology. However, many studies have undertaken isotopic analysis of the
41
various hydrological flows in catchments in order to estimate the groundwater

contribution to surface water flow (Jasecko, 2019). The results show that the ratio of
groundwater to runoff as contributions to total river flow vary from catchment to
catchment, and also within the same catchment, but under differing weather conditions
(Camacho Suarez et al., 2015). Some studies, for example Iqbal (1998), continue to show that
precipitation tends to recharge groundwater, driving an increase in baseflow, causing quick
flow to have an isotopic composition closer to groundwater than event water, as shown in
Figure 24.
Figure 24 - a) Stable isotopes for the Cedar River catchment, Iowa, USA (Iqbal, 1998). The final storm
discharge in the river (blue squares) is assumed to comprise two components, namely groundwater discharge
(purple star) and precipitation (turquoise triangles). The proportions that these two components contribute to
the storm discharge can be calculated by measuring the graphed distances along a straight line between all
3 water types. In this case the distance between the pre-storm data point and the storm water discharge data
points is about 0.15 of the total distance between the pre-storm and the rainwater data points, whereas the
distance between the storm discharge and the rainwater data points is about 0.85 of the total distance,
indicating the groundwater contributes roughly 85 percent of the storm discharge. The spread of data and
number of samples in this case suggest this method is applicable. In contrast, scenarios where this method is
not applicable are shown in the insets: b) too much scatter of the data resulting in high standard deviation of
the estimate; c) too little variation of the data such that, statistically, their means that are not significantly
different; and, d) too few samples producing statistically uncertain results.
Another study, shown in Figure 25, used data from a seasonal period to show that
quick flow is mostly event water. It was theorized that this was due to steep slopes, poor
aquifers and compaction of soils from agriculture in the basin (Tekleab et al., 2014).
42
Figure 25 - The results of a classic two-component hydrograph separation performed using

D and  O over a seasonal period for the a) Chemoga and b) Jedeb catchments in
18
Ethiopia. The proportion of event water in the river flow was calculated to be an average of
71 percent and 64 percent in the two catchments respectively (from Tekleab et al., 2014).
Hydrograph separation is a worthy pursuit, and although some progress has been
made in improving the procedure, more integration of physical, chemical and isotopic
evidence is likely to bring greater understanding (Kirchner, 2003).
6.6 Geothermal Waters

Heated groundwater can be divided into two types, although there is a continuum
between the two. The term thermal (or hot) springs usually applies to groundwater that is
noticeably warmer than the mean annual temperature of an area, so the temperature
definition varies with the local climate. For temperate climates, 15 °C is a typical threshold
between shallow unheated groundwater and geothermally heated water. The threshold is
closer to 20 °C for Mediterranean climates, around 25 °C in the tropics, and up to 35 °C in
hot deserts. The depth to which groundwater must circulate in order to be heated again
varies from region to region, depending on the geothermal gradient, which is influenced
by the age and thickness of the crust and the degree of volcanic activity in the area.
However, there is usually a clear distinction between relatively low geothermal gradients
of around 10 to 30 °C/km in tectonically quiet areas (most of the earth's surface), versus the
much higher geothermal gradients in volcanically active areas where boiling water and
steam occur within tens to hundreds of meters of the surface. These latter areas are usually
referred to as geothermal.
43
From an isotopic perspective, thermal springs are usually caused by deep

circulation of meteoric water, usually at temperatures below those likely to cause isotope
exchange with rocks. As such, studies evaluating recharge and flow paths, and other
hydrogeologic issues, can be conducted using the stable isotopes, as they are conservative,
for example as shown in Figure 26 (Chandrajith et al., 2012; Diamond and Harris, 2000;
Durowoju et al., 2019). Most of the concepts in this book are applicable in these situations.
Figure 26 - Stable isotope data for cold and hot springs (< 62 °C) in eastern Sri Lanka. The overlap
of isotope compositions for hot and cold springs is typical for lower temperature thermal springs.
Surface water typically has heavier isotope composition of the groundwater as displayed in the inset
(after Chandrajith et al., 2012).
In contrast, geothermal waters usually experienced isotope exchange with the host
rocks resulting in the isotope composition changes (Lelli et al., 2021). Figure 27 shows how
isotope exchange, caused by the high temperatures of geothermal fields (> 100 °C), with
silicates, which generally have positive 18O values, is the cause of variation in isotope
composition. If mixing with juvenile (magmatic) water was the cause, the D values of the
geothermal water would have changed and the mixing lines from the various geothermal
fields would converge on the juvenile water type.
44
Figure 27 - Stable isotope compositions for water and steam at geothermal locations around the world, in
comparison to the GMWL (Global Meteoric Water Line). Mixing with juvenile (magmatic) water is not the
cause of variation from meteoric water, but rather the variation is caused by isotope exchange with silicates.
This is revealed by major changes in oxygen isotopes, while changes in hydrogen are negligible because
of its lesser abundance in minerals (after Panichi and Gonfiantini, 1981; D'Amore and Panichi, 1985;
Stewart, 1985).
In the case of Wairakei and some of the other fields, the shift in 18O is not large.
This likely occurs because the rocks already exchanged oxygen with the groundwater over
many years and now the system is close to the equilibrium fractionation value with the
meteoric water, so further exchange is minor. Extending this concept, the factors
responsible for determining the degree of isotope exchange between water and rock are:
original isotopic composition of the rocks, original isotopic composition of the meteoric
water, quantity of rocks, quantity of water, temperature, water circulation path, the
water/rock contact-surface area, and time (Panichi and Gonfiantini, 1981; Daniele et al.,
2020).
Groundwater temperatures in geothermal fields are very high, so it is expected that,
given the chance, significant evaporation will occur. This evaporation can either be at the
surface, or in the subsurface, and there are ways to model both the amount of evaporation
and the process. These processes include single-stage or continuous steam separation, the
former being when steam is released only once the groundwater-steam mixture arrives at
surface, the latter being where steam is continuously removed from the water as it boils
underground. Work has been done monitoring the changes in hydrochemistry and isotopes
45
during exploitation of geothermal fields to provide estimates on the sustainability of the

field (Prasetio et al., 2021).
Finally, stable isotopes of water and another compound, such as CH4 or CO2, can be
used to estimate the temperature underground. For example, if the hydrogen isotope
composition of water and methane is measured at surface, and these are assumed to be in
equilibrium, then the temperature of reaction can be estimated from the known change in
fractionation factor with temperature. Several assumptions need to be fulfilled, but the
method can be made more robust by using other geothermometers, such as oxygen isotopes
in water and sulphate (Panichi and Gonfiantini, 1981).
6.7 Paleowaters
Paleowaters in groundwater include any water that was recharged under a different
climate from the present one or formed under different environmental conditions to those
now found at the site. For hydrogeologists, who are more concerned with circulating
groundwater, waters involved with the genesis of rocks and minerals are generally of less
interest as they are not an active part of the hydrological cycle and therefore do not
represent a sustainable water resource. As such, connate water, formation water, juvenile
water, interstitial water and crystallization water are excluded from the usual definition of
paleowater.
Paleowater groundwater resources are called fossil aquifers and are sometimes
used for water supply. An example is the Nubian Aquifer System of the eastern Sahara,
shown in Figure 28. Roughly speaking, actively circulating groundwater is likely to be less
than 10,000 years old, paleo groundwaters are in the range of tens to hundreds of thousands
of years old, and connate, crystallization, formation, interstitial and juvenile waters are
more likely to be millions to billions of years old.
46
Figure 28 - Contrast in isotope patterns for modern precipitation (shown by color-coded country outlines)
versus paleowaters (shown by dashed contour lines) across the Sahara Desert. The west to east decrease in
 H for paleowaters was used as evidence for an shift in westerly rain systems toward the equator during the
2
last ice age (after Abouelmagd et al., 2012).
A common requirement for working on paleowaters is the ability to date the

groundwater age. Tritium, although commonly used in young groundwater, is of no use in
paleowaters as its half-life is too short (12.3 years). Carbon-14 (half-life 5730 years) is the
most widely used dating method, but, as explained in another Groundwater Project book,
Introduction to Isotopes and Environmental Tracers as Indicators of Groundwater Flow (Cook,
14
2020), this is far from simple. The quantity of C in the groundwater is given as pmc
(percent modern carbon), from which the age can be calculated, however, there are
complications which can cause this calculation give the wrong age. On the one hand the 14C
can be removed by preferential precipitation or diluted by contributions of inorganic
carbon dissolving out of the aquifer matrix, and on the other hand modern carbon can enter
the system and mix with the old groundwater or contaminate the water during sampling.
In general, therefore, although the analytical precision can theoretically date materials to
60,000 years, the uncertainties in the groundwater environment reduce this method's age
maximum to around 30,000 years (Clark and Fritz, 1997; Fontes, 1981).
47
Other dating methods for paleowaters can also be used, for example 36Cl with a
half-life of 301,000 years. This and other methods have advantages and disadvantages and
have certain age ranges over which they may give reliable ages. Groundwater dating is a
vast field and the reader is encouraged to consult other books on these topics, if they want
to know more about it (Clark, 2015; Cook, 2020).
Although accurate groundwater dating is the surest way to know one is dealing
with paleowater, discovering groundwater with a stable isotope composition different
from groundwater known to be young, often alerts one to the presence of paleowater (Van
Gelden et al., 2014). In most cases, the paleowater is in deeper, confined aquifers with little
connection to unconfined aquifers with modern recharge (McIntosh et al., 2006), but there
are instances where the geometry of aquifers and flow regime can result in the paleowater
occurring at shallow depths (Malov and Tokarev, 2019).
Identifying the presence of paleowater from a difference in stable isotope
composition with the young groundwater of the region is useful. One of its uses is to
understand the sustainability of groundwater, because use of paleowater is generally
assumed to be non-sustainable. This is because paleowaters were either recharged during
a wetter climate and are not being recharged at all in the current climate, or are recharged
very slowly, hence their great age, and therefore only sustainable at very low abstraction
rates. Al-Charideh and Kattaa (2016) investigated groundwater in the regional, deep
Cretaceous aquifer across western Syria, and were able to classify groundwater as
renewable or semi-renewable when the stable isotope composition was similar to modern
precipitation (18O of -7.0 to -7.2‰), or as non-renewable when the stable isotope
composition was significantly more negative, suggesting recharge during a previous
climate (18O of -8.0‰). Additionally, 14C values for the latter groundwater were < 15 pmc,
confirming this as older groundwater.
A study by Smith and others (2002) in the Great Basin region (centered on Nevada,
USA) identified locations where 2H of groundwater was more than 10‰ lighter than
winter precipitation. They proceeded to rule out sites adjacent to mountains, where the
altitude effect could be the cause of the isotopically lighter groundwater values, and then
classified areas as having possible paleowater (10-19‰ lighter) and probable paleowater
(>20‰ lighter), recharged during the Pleistocene.
In previously glaciated parts of the world where meltwater from receding glaciers
recharged groundwater as the previous glacial period ended, differences in paleowater and
modern groundwater isotope compositions can be large. However, in some cases the
paleowater in far northern latitudes are not from glacial meltwater and can therefore have
differences to modern waters that are similar to those in mid-latitudes or tropical areas. For
example, in Germany, Van Gelden and others (2014) noted the similarity of modern
precipitation (-9.8‰ and -66‰) and the shallow, unconfined aquifer (-9.6 to -8.9‰ and -68
to -65‰) compared to the confined Benkersandstein (-10.6‰ and -74‰) giving roughly a
48
1‰ and 10‰ difference in 18O and 2H respectively, between modern (shallow
groundwater or precipitation) and paleowater. In contrast, in the Great Lakes region of
North America, McIntosh and others (2006) found current precipitation and surficial
aquifers to have 18O of -7 to -4.5‰ in northern Indiana and Ohio, and -11 to -8‰ in
northern Michigan, whereas the Silurian-Devonian carbonate aquifers were in the -13
to -15‰ range, exhibiting a difference of 5-10‰ in 18O, as shown in Figure 29.
Figure 29 - Groundwater regime and stable isotope systematics for northern Ohio and Indiana states in the
USA, south of the Great Lakes. Various groundwater samples are plotted on the 2H - 18O graph according
to their dominant dissolved ions. The estimated isotope composition for glacial meltwater ranges from -25 to
-11‰ for 18O. Stable isotope ranges for paleoprecipitation are shown as dotted boxes with associated ages.
The CaMgHCO3 waters are associated with the shallow, unconfined glacial drift aquifers and tend to have
stable isotope values somewhat more negative than modern precipitation, but plotting close to the LWML and
GMWL. The other water types are found in confined aquifers, at depth, and their chemistry is influenced by
rock dissolution and mixing with formation brines. The existence of isotopically very negative groundwater at
shallow depths in the Silurian-Devonian carbonate aquifer suggests this groundwater was recharged by
melting glaciers at the end of the last ice age and is largely isolated from modern recharge by aquitards (after
McIntosh et al., 2006).
In much of the southern interior of Canada and parts of the northern plains of the
USA, there is abundant paleowater within the Quaternary deposits overlying bedrock. As
the Pleistocene glaciers were melting (about 10,000 to 20,000 years before present), run-off
from these glaciers formed lakes with deposits rich in silt and clay across broad plains.
These deposits have exceptionally low permeability (with thicknesses from 5 to 40 m) and
function as aquitards with the water table positioned near the ground surface. Studies of
these aquitards show that where the thickness is greater than ~20 m, the groundwater is
comprised mostly or entirely of water derived from the Pleistocene glaciers (Figure 29).
This conclusion is based on stable isotopes indicated by cold weather isotopic signatures.
One of the most extensive deposits of these glacial lake sediments occurs in
Manitoba, Canada, and extends southward into North Dakota, USA. A uniform oxygen
isotope value of –25‰ was obtained at depths of 20 to 30 m in a thick deposit of clay found
in the southernmost part of the glacial Lake Agassiz plain (Remenda et al., 1994). The lake
occupied parts of North Dakota and southern Manitoba at the end of the last glacial
49
maximum and received water from the ice margin and the interior plains region of Canada.
The value of –25‰ is characteristic of meltwater impounded in the southern basin of Lake
Agassiz. This value corresponds to an estimated mean air temperature of –16 °C, compared
with the modern mean air temperature of 0 °C for this area.
Groundwater from thick, late Pleistocene-age clay deposits elsewhere in Canada
shows the same uniform oxygen isotope value at similar depths, including a glacial till in
southern Saskatchewan and a glaciolacustrine deposit in northern Ontario. These aquitards
are at about 50 °N latitude, span a distance of 2000 kilometers, and like the Lake Agassiz
sites, have a groundwater velocity of less than a few millimeters per year. Hence, there is a
substantial volume of this paleowater in the southern interior regions of Canada, but almost
all of this water occurs in aquitards where flow rates are minimal, rather than in aquifers.
This is because water of glacial origin, that initially filled the aquifers, has flushed out since
the glaciers disappeared. An exception is the regional aquifer beneath the lake sediments
in southwestern Ontario where the aquitard is an effective cap, impeding the flow of
paleowater (Husain et al., 2004). The glacial age of these paleowater occurrences has been
confirmed by carbon dating. However, given the characteristically negative stable isotope
values for the paleowater in this region, the most cost-effective method for dating
paleowater is through the use of stable isotopes.
The differences in stable isotope composition for modern water and paleowater
may be known in a region. Sultan and others (2000) applied such regional knowledge to
the unconsolidated aquifers close to the Nile River, south of Cairo, near Wadi El Tarfa.
These aquifers had 18O of -2 to +5.2‰ and 2H of -10 to +34‰, which ruled out upwards
leakage from the Nubian Aquifer, known to have mean values of -11‰ and -80‰,
respectively (Figure 30). They concluded that recharge of the shallow aquifers was from
flash flood events with rare, but appreciable recharge into coarse sediments and fractured
rock underlying alluvial channels. As such, these aquifers can be sustainably used, if the
frequency and magnitude of recharge from these events is known.
50
Figure 30 - Stable isotope analyses of various waters in Egypt and rain from the Negev Desert in Israel, with
the GMWL and a trend line for Wadi El-Tarfa groundwater, in order to determine the likely source of Wadi
El-Tarfa groundwater. Upwards leakage of paleowater from the Nubian Sandstone Aquifer of the Western
Desert was ruled out due to its lighter isotope composition. The position of the Wadi El-Tarfa samples below
the GMWL is suggestive of evaporation. The dashed trendline has a gradient of 5.7, indicating evaporation
from an open water body, such as a temporary pond after heavy rain. The extrapolation of the trendline into
the cluster of Sidi Barrani rain water suggests evaporated modern rainfall recharges the Wadi El-Tarfa
groundwater, meaning abstraction is potentially sustainable at a rate similar to the recharge (after Sultan et
al., 2000).
In nearby Libya, Al Faitouri and Sandford (2015) conducted a survey of three

sandstone aquifers within the Nubian Aquifer System, to establish paleoclimatic and
recharge dynamics. However, the stable isotope characteristics of these three aquifers can
also be used to detect leakage or mixing of groundwater, by monitoring the pumped
groundwater for changes in stable isotope composition over time. Initial results suggest
that younger groundwater is being incorporated into the deeper aquifers as the
potentiometric head of the deeper aquifers has been lowered and therefore flow induced
from the shallower, younger groundwater.
Work on the Nubian Aquifer System, as well as other paleowaters in the Saharan
region, has extended to paleoclimatic reconstruction. Using d-excess values calculated by
Sonntag and others (1979) for 20,000-year old groundwater beneath the Sahara, Merlivat
and Jouzel (1979) were able to, with the aid of an isotopic precipitation model, calculate the
humidity under which evaporation took place over the ancient ocean, finding a result of 90
percent as compared to 80 percent for the current climate. Abouelmagd and others (2012)
51
showed how the average modern precipitation can get isotopically heavier from west to
east (W to E), but the paleowater gets isotopically lighter from W to E (Figure 28). However,
some modern rain events show isotopic depletion from W to E, matching the fossil
groundwater. These were produced with westerly winds and so it was inferred that the
paleoclimatic recharge was due to the westerlies (mid-latitude cyclones or frontal
depressions) shifting toward the equator during the previous ice age.
In contrast to the findings in the Sahara, Plummer and others (2012) found high
D-excess values for paleowater in the vicinity of Chesapeake Bay, USA, indicating lower
relative humidity during moisture formation (Figure 31). This is consistent with the colder
temperatures and icy conditions expected over mid-latitudes during glacial periods. The
higher humidity detected for paleowater recharge conditions over the Sahara and the lower
humidity over the American mid-latitudes agree with the general model that current polar
conditions (cold and dry) occupied mid-latitudes and current temperate conditions (wet
and rainy) occupied low latitudes during the previous ice age.
Figure 31 - Stable isotope values for groundwater from various aquifers in the
Chesapeake Bay region. Distinct groups are apparent for the various aquifers,
reflecting differences in climate, and particularly the d values, which is an
indicator of the humidity level in the region that provided the moisture source.
Higher d values indicate lower humidity in the moisture source region (after
Plummer et al., 2012).
Stable isotopes of groundwater may also be used indirectly in paleoclimatic and

other reconstructions. Andrews (2006) provides a review of the use of groundwater-fed
riverine tufas (carbonate deposits) for these purposes. The method used the 18O of the
52
carbonates, corrected for fractionation at the temperature of precipitation from water, to

determine the 18O of the groundwater at recharge. High resolution sampling of tufas (sub
annual scale) tends to provide more information on the temperature of precipitation
(precipitation), whereas decadal scale sampling gives more information on the
groundwater isotope composition, and is susceptible to the well-known isotope effects of
continentality, amount and temperature. These allow some inference regarding the types
of events responsible for groundwater recharge. For north-western Europe, tufas revealed
centennial scale cooling prior to the end of the last ice age, reaching a minimum
temperature around 8200 years before present. Tufas are therefore a good alternative or
complement to paleoclimatic methods using speleothems.
Canavan and others (2014) performed analyses of the 2H of volcanic glass in the
central Andes. Primary volcanic glass has only 0.1-0.3 weight percent H2O, whereas after
hydration (thousands of years) the glass contains about 5 weight percent H2O, and the
hydration seals in the isotopic content of the hydration water with a layer of "silica gel".
This means the hydrogen is largely sourced from meteoric water and the original 2H of
hydrating water can be calculated after applying the fractionation factor for glass-H2O.
With this paleowater 2H signature, and applying known altitude isotope gradients, the
elevation at the time of eruption (given that hydration is achieved shortly thereafter) can
be calculated. This in turn can be used to infer uplift rates and continental geodynamics of
a region.
It should be apparent that stable isotopes of paleowater, and associated materials
(e.g., speleothems, tufas, volcanic glasses) are a mine of potential information. However,
the corrections and complexities require attention, to avoid misinterpretation. Geological
dating, in particular, is very useful in this area, but attention to fractionation, mixing,
dilution and other factors is necessary. As is often the case, use of hydrochemistry or other
isotope systems can assist in painting a fuller picture.
6.8 Plant Waters

Plants use water in their leaves, (and sometimes stems and other parts) for
photosynthesis to create simple sugars and in turn all the compounds in their structure.
They obtain this water, and a mixture of nutrients dissolved in the water, from the soil by
absorption into their roots (Dawson et al., 2002). Describing this as use of soil water is an
oversimplification. In fact, water for plant use is available from many sources, including
surface water (e.g., as used by rice, mangroves, riparian plants), water in the vadose zone
(e.g., water adhered to soil particles, gravitational water, and capillary water), as well as
groundwater in the phreatic zone. This is further complicated by the fact that, with time,
these sources may merge, separate or dry up.
53
Stable isotopes of H and O provide a way to discover which pools of water are being
used by plants, and have the added advantage of being fairly inexpensive to analyze
(especially with laser cavity instruments) and requiring small samples, therefore causing
less disturbance to the soil and plants (Ehleringer and Dawson, 1992).
Caution is necessary with the use of the IRIS (isotope ratio infra-red spectroscopy)
or laser instruments, as trace quantities of organic compounds in plant waters may distort
the analysis and produce errors of more than 10‰ for both 2H and 18O. Adequate
filtration to remove organic contaminants is necessary (West et al., 2010).
The basic principle of the method is to measure the isotope compositions of plant
waters and potential source waters then compare the  values to determine which pool of
water the plant is using. Although this approach may work in simple situations, research
over the last 40 years has shown there are many complications that may require
unravelling. For example, most earlier studies assumed there was no fractionation during
water uptake at the roots (Dawson and Ehleringer, 1991; Ehleringer and Dawson, 1992),
although Alison and Hughes (1983) noticed fractionation and ascribed it to a 2-stage
process of soil water evaporation followed by root uptake of the vapor with its evaporated
isotope signature, leaving isotopically heavier water in the soil. More recent studies, have
shown there is appreciable fractionation, up to 10‰ for 2H (Ellsworth and Williams, 2007),
correlated with the strength of the halophytic character of a species, or the magnitude of
salt stress experienced by an individual plant (Gat et al., 2007). Non-halophytic plants may
also display fractionation during water uptake, but this is dependent upon soil type and,
again, the drier the soil, the more fractionation (Vargas et al., 2017). It is fairly clear that
fractionation between the soil and the plant is mainly a concern in dry or salty
environments, typical of deserts or in areas that experience seasonal droughts.
Large variation has also been noted in the isotope composition in different parts of
the plant. This is to be expected, especially near the leaves, where transpiration causes large
fractionation effects. The degree of fractionation is correlated with the transpiration rate,
which is in turn influenced by higher temperature and precipitation (Pan et al., 2020) and
lower relative humidity or dry conditions (Yakir et al., 1990; Bodé et al., 2019). As a result,
there will be a variation in isotope composition between the water in a plant's leaves and
elsewhere in the plant, with a range of  values occurring as water in the stems mixes with
the fractionated leaf water. Also, over time, as transpiration varies, the degree of
fractionation in the leaves will vary, and therefore any part of the plant exposed to
fractionated water in the leaves will also vary. For example, Goldsmith and others (2018)
found the variation in  values within one tree to be similar to that between different trees.
Because of these complexities, a simple 2-point model with soil water and
groundwater at each end, and one value for plant water which hopefully sits on the mixing
line between the 2 source waters, is often not applicable. Instead, models have been
developed to incorporate this complexity, whether it be calculation of soil water  values
54
(Goldsmith et al., 2018), incorporation of species-specific traits (Yakir et al., 1990),

transpiration and leaf- and soil-water potentials (Cook and O'Grady, 2006) (Figure 32) or
tree-water deficit, fine root distribution and soil water potentials (Nehemy et al., 2020). The
choice of model then becomes a factor to consider in this kind of work (Wang et al., 2019).
Figure 32 - Stable isotope values for groundwater, soil water

and leaf water in 4 species of tree in northern Queensland,
Australia. Cc = Corymbia clarksoniana, Ls = Lophostemon
suavolens, Ep = Eucalyptus platyphylla, Mv = Melaleuca
viridiflora. The stable isotope values were used in conjunction
with transpiration rates, leaf and soil water pressure potentials
to estimate the proportions of soil water and groundwater used
by the 4 species (after Cook and O'Grady, 2006).
In spite of these challenges, stable isotope studies of plants have wide application
in agriculture and forestry (Penna et al., 2020) and have produced some important findings,
such as the proportions of groundwater versus soil water use (Dawson and Ehleringer,
1991; Cook and O'Grady, 2006), surface water versus soil water use in rice (Mahindawansha
et al., 2018). They have also made it obvious that much care needs to be taken when using
18O values in old or fossilized plant material to determine paleoclimates (Gat et al., 2007).
55
7 Case Studies
Some examples of the application of stable isotopes to groundwater are included
here. Most of these examples include work with precipitation and surface water,
emphasizing how, in most cases, groundwater work involves aspects of the rest of the
water cycle. The excerpts from these studies are brief, merely to illustrate the concept.
Before implementing these methods, it is important to read more widely about the topic,
including reading of the full article used for the example.
7.1 Moisture Source Region

Uemura and others (2012) monitored precipitation weekly on Okinawa Island, to
the south of Japan. They found a negative correlation between air temperature and 2H and
18O values, which is the converse of the usual positive correlation where temperature and
 values decrease in concert (Figure 33). The reason for this negative correlation was related
to the source regions for precipitation, which had a greater effect on the isotope
composition than temperature at the precipitation site. In summer, precipitation has lower
 values than in winter. When considering atmospheric circulation, they noticed that
summer precipitation is sourced from a distant oceanic location, allowing greater rainout
and therefore more negative  values. Winter precipitation is sourced by local evaporation
from the warm Kuroshio Current into dry Asian continental air, allowing little rainout prior
to precipitation at the site. The higher D-excess values for winter are also typical for
evaporation into drier air. The finding that precipitation on Okinawa has different sources
in winter and summer may be important when predicting how climate change may affect
precipitation and therefore water resources on the island.
56
Figure 33 - Graphs of a) 2H and  O, b) d-excess and c) precipitation and

18
temperature over a 3-year period on Okinawa Island. The negative correlation

between T and  values is caused by differences in moisture source region for
summer and winter, and the degree of rainout as moisture travels from those
different regions to the island (after Uemura et al., 2012).
7.2 Recharge Area

Diamond and Harris (2000) made a study of the hot springs (Figure 34) in Western
Cape, South Africa, including collection of precipitation near springs. They found the stable
isotope values of the springs to be significantly more negative than the rain falling on the
ground near each spring (Figure 35). Isotope exchange with rocks was ruled out due to the
relatively low temperature of the spring water (the maximum being 64 °C), thus a deep,
long flow path through the fractured quartzites of the Table Mountain Group was
postulated. The flow paths for each spring included recharge at high altitude (up to
2000 m), which could explain the relatively negative isotope compositions, caused by the
altitude effect. The postulated high elevation of recharge was consistent with a high head
required to drive a long flow path with deep circulation that could heat meteoric water to
57
the spring water temperatures. The existence of long groundwater flow paths is important
for protecting recharge areas because any impact on the water source may not be observed
at the discharge location until many years later.
Figure 34 - The Citrusdal hot spring (43 °C) in the Cape Fold Belt of South Africa.
Groundwater flows through the Cape Supergroup in fractures and in some cases
circulates deeply enough for water to be heated. This photograph shows one of the
discharge points of the spring - several other fractures within a few meters of this
point also discharge hot water.
58
Figure 35 - Stable isotope compositions of cumulative monthly precipitation samples from

4 precipitation stations and single samples or averages of monthly samples from the thermal
springs of the Cape Fold Belt in South Africa. The tendency of the spring water to have lighter
isotopic composition suggests recharge at higher altitude, made possible by deep
groundwater circulation through large fold structures, which have wavelengths of many
kilometers. Data from Diamond and Harris (2000), LMWL from Harris and others (2010) and
GMWL from Craig (1961).
7.3 Selective Recharge of Heavy Rains

Vogel and Van Urk (1975) sampled groundwater and precipitation in the semi-arid
interior of southern Africa. They found stable isotopes in precipitation were highly
variable, whereas stable isotopes in groundwater of each region was quite consistent.
Further, the groundwater tended to be more isotopically negative than precipitation,
indicating that recharge took place during precipitation events with more negative 
values, which are typically the larger rainfall events (as per the amount effect). Similar
findings have been made by many workers, for example Dogramaci and others (2012) for
another semi-arid region, the Pilbara in Western Australia (Figure 36) and Li (2018) near
Beijing in China. Gat (1996) termed this phenomenon "selection".
59
Figure 36 - Stable isotope data for the Hamersley Basin in Western Australia. The
LMWL was calculated for only larger rainfall events (> 20 mm). The near coincidence
of this line with the data from shallow alluvium and fractured rock aquifers suggests
recharge for these aquifers is during large precipitation events. The low gradient of
the saline waters line is evidence for substantial evaporation of these waters,
providing evidence that the salinity is from concentration of salts resulting from the
evaporation process. The ultimate source of the salts, from rock weathering or
marine aerosol, would have to be proven by evaluating the balance of salts, such as
dominance of bicarbonate over chloride, or magnesium over sodium (after
Dogramaci et al., 2012).
7.4 Recharge Estimation

Yidana and others (2016) evaluated water fluxes through the vadose zone in the
Nabogo catchment of the White Volta Basin, Ghana. They sampled rain, surface water, pore
water (vadose zone) and groundwater (phreatic zone) and analyzed these samples for their
stable isotope composition and chloride content (Figure 37). They used Cl mass balance to
estimate the fraction of infiltrating water remaining at each sampling interval and therefore
estimate evapotranspiration (ET). The Cl content was still changing at their maximum
sampling depth of 3 m, suggesting there is more ET taking place at greater depths. Their
estimated recharge was 13 to 20 percent of precipitation. They also used differences in the
stable isotope composition of phreatic zone groundwater, vadose zone porewater and
precipitation to estimate evaporation, according to the Craig-Gordon model (1965) as
outlined by Dogramaci and others (2015), which was found to be about 40 percent of
precipitation. This suggested that transpiration makes up 40 to 47 percent of the water
losses prior to recharge (i.e., with 13 to 20 percent of precipitation recharged and 40 percent
60
evaporated, then the remainder (40 to 47 percent) is transpired by plants). The use of Cl
mass balance is sufficient to estimate total ET and hence recharge, but when combined with
an evaporation-only estimate based on stable isotopes, transpiration can also be calculated.
Figure 37 - Stable isotope values from the Nabogo catchment in the White Volta Basin of Ghana. Distinct
clustering of the precipitation and groundwater, and the low gradient of the groundwater trendline (LGWL) is
strong evidence for evaporation of precipitation prior to recharge of groundwater (after Yidana et al., 2016).
7.5 Piston Flow of Groundwater During a Storm

Midgley and Scott (1994) conducted a study of streams in the Jonkershoek area near
Cape Town, South Africa, where they collected stream water and precipitation samples
simultaneously. Interestingly, they found that although the streamflow rose as the rain
event proceeded, the stable isotope composition of the stream water did not match that of
the rain (Figure 38).
61
Figure 38 - Stable isotope data for stream water (solid circles) and precipitation
(open circles) in relation to a precipitation event and associated streamflow. The
lack of correspondence of the streamflow isotope composition to the precipitation
suggests groundwater as the source of streamflow, especially at Bosboukloof
where the stream discharge isotope composition is steady. Lambrechtsbos
stream water shows some change in isotope composition toward the isotopic
composition of the precipitation early in the precipitation event, suggesting either
runoff or recent interflow is contributing to the total streamflow during that period
(after Midgley and Scott, 1994).
Using the stable isotope compositions and a mass balance approach, they
determined that < 5 percent of the streamflow was direct precipitation runoff. They did not
take groundwater samples, but rather assumed that the streamflow prior to the rain event
was purely baseflow and therefore representative of the groundwater. As discussed in
62
Section 6.5 “Hydrograph Separation”, this assumption is not always valid. Nonetheless,
they concluded groundwater was being driven, in a piston flow fashion, by the recharging
precipitation, towards the stream and discharging at increased rates during the
precipitation event (Figure 39).
Figure 39 - The Elandspad River, Western Cape, during a) and after b) the passage of a cold
front of the South Atlantic Ocean in July 2012. The two pictures were taken 16 hours apart. In
the nearby Jonkershoek catchment, Midgley and Scott (1994) found stable isotopes of the flood
flow matched that of groundwater and not precipitation, suggesting that rapid increase of
baseflow from fractures in the quartzite of the Table Mountain Group were largely responsible
for the flood peak.
63
7.6 Groundwater Circulation and Age

Yin and others (2011) measured stable isotope ratios in groundwater and lake water
in the Habor Lake Basin of the Ordos Plateau in north-central China. Their results for the
groundwater hydrochemistry revealed a range of ionic compositions, but no clear
separation between aquifers. The stable isotope values, however, varied according to
depth, with the deepest samples (up to 300 m) showing the most negative isotope values
(Figure 40). Shallow Cretaceous aquifer groundwater and Quaternary aquifer groundwater
have values between modern precipitation and modern lake water. This suggests
substantial evaporation prior to recharge, as the lake water represents a highly evaporated
isotope composition. The deep Cretaceous aquifer groundwater was probably recharged
during the wetter climate of the glacial period at the end of the Pleistocene, during which
the precipitation would have been isotopically lighter than the present, and evaporation
would have been less during recharge, because of the cooler, more humid conditions than
today.
Figure 40 -  H -  O diagram for water samples from the Lake Habor area in north-central China.
2 18
Differences in the stable isotope composition of shallow and deep groundwater, and comparison with
precipitation and lake water allow inferences to be made about the age and recharge conditions of the
various groundwaters as explained in the text of Section 7.6 (after Yin et al., 2011).
7.7 Detecting Surface Water in Groundwater

Grimmeisen and others (2017) conducted a multi-tracer study, using chemical
parameters (especially Cl- and NO3-) and the stable isotopes of H, N and O, in and around
the city of As-Salt in Jordan (Figure 41). The stable water isotopes alone provided good
evidence for leakage from various possible "surface water" sources, in this case human
sources such as water mains and sewers. The distinctive isotope composition of these urban
water sources is due to the city's water supply being the King Abdullah Canal, fed largely
from highly evaporated surface water in Lake Tiberias and the Yarmouk River, with only
64
a portion from the Mukheiba wells. Mixing with local groundwater is apparent in some of
the boreholes that were monitored, with a seasonal fluctuation in locations where leakage
from urban water sources affects some of the local groundwater more in the dry summer
(May to November) and less in the wetter winter. The mixing line allows estimation of the
proportion of leakage water comprising local groundwater, which the authors calculated
as up to 70 percent.
Figure 41 - Isotope composition of various groundwater sources (GW1-5) and the urban water (clean water
and waste water) over several seasons at As-Salt, Jordan. The changes in groundwater from unmixed/fresh
(GW1-2) to the most mixed/polluted (GW3 & GW5 in May-Nov) is revealed by the change in stable isotope
composition, which can be used to estimate leakage (after Grimmeisen et al., 2017).
Sunguro and others (2000) undertook a study of groundwater in the dolomites of

the Mcheka Formation of the Palaeoproterozoic Lomagundi Belt in northern Zimbabwe,
known informally as the Lomagundi dolomites. Spring water had relatively negative 
values, compared to precipitation, indicating recharge during heavy precipitation events
(as discussed in Section 7.3 “Selective Recharge”), whereas boreholes, particularly those in
65
areas of active irrigation, tended to have water with more positive  values, suggestive of
evaporative isotope enrichment and therefore irrigation return flows to the aquifer. This
further highlights the vulnerability of the aquifer to pollution by agricultural waters.
7.8 Detecting Groundwater in Surface Water

Camacho Suarez and others (2015) used stable isotopes in combination with
hydrochemistry to compare water types and discharge rates in the Kaap catchment,
Mpumalanga, South Africa. They made use of multi-component hydrograph separations
with steady-state mass balance equations, as per Uhlenbrook and others (2002). The results
of their two-component hydrograph separation yielded runoff contributions of 5 to 36
percent of direct runoff as a proportion of streamflow for 4 different rain events. In addition
to this, they conducted a principal component analysis (PCA) and found 2D and K to be
poorly correlated (i.e., to vary independently) and were thus able to use these two
parameters to conduct a three-component separation, as shown in Figure 42. This revealed
a large difference in relative contributions of shallow versus deep groundwater to
streamflow. In particular, for events where prior precipitation has wetted the catchment,
the shallow groundwater and runoff contributions are greater than for events occurring
after a dry period.
Figure 42 - Three-point hydrograph separation based on  H and K values for 4 rain events in the Kaap
2
catchment, Mpumalanga, South Africa. Different events revealed different proportions of groundwater
and runoff in the streamflow, depending on the weather and ground conditions before and during each
event (after Camacho Suarez et al., 2015).
66
7.9 Residence Time

Paulsson and Widerland (2020) investigated 4 pit lakes (waterbodies in abandoned
open pit mines) in central Sweden, including measuring stable isotope compositions for
snow, rain, groundwater and the pit lakes (Figure 43). Their study was complicated by the
fact that the lakes are stratified, and that for 6 months of the year, the lake surface is frozen.
Nonetheless, they applied the hydrocalculator (http://hydrocalculator.gskrzypek.com)
method of Skrzypek and others (2015), which is based on the Craig-Gordon evaporation
model (Craig and Gordon, 1965). The hydrocalculator uses measured, calculated and
estimated values for the initial and final isotope compositions, relative humidity,
temperature, as well as equilibrium and kinetic fractionation factors and enrichment and
adjustment factors (Biggs et al., 2015; Dogramaci et al., 2015) to estimate the amount of
evaporation that occurred from the lake surface, based on the change in 2H and 18O
values. With adjustments for the lake stratification and the volume of water in each pit lake,
they used the estimated evaporation value to calculate the residence time of water in each
lake. The residence times varied from 3 to 14 years as the lakes are quite different in size
and shape. These values can easily have a factor of 2 error, due to natural variability and
complexity, and the assumptions in the models, but these results are useful to help quantify
the broad differences between the hydrological systems of each pit lake.
Figure 43 - Stable isotope data for snow, rain, groundwater and pit lakes (named) in central Sweden. The
distinct isotope compositions of rain and snow, and of each of the pit lakes allow calculation of residence
time of water in the pit lakes (from Paulsson and Widerland, 2020). The large negative isotope compositions
are due to the high latitude of the study area.
67
8 Water Sampling for Stable Isotope Analysis

8.1 Practical Considerations
Stable isotopes of water have a major advantage over many other water analytes, in
that they comprise virtually all of the sample, rather than being a minor or trace constituent,
such as Cl- or 3H. This means that contamination is less of a concern during sampling and
storage. However, stable isotopes are sensitive to evaporation, so ensuring the sample is
well sealed is essential.
Sample bottles need not be specialized. They can be glass or plastic, but must seal
completely. Sample volumes required for IRMS or CRDS are only a few milliliters, so any
bottle more than 50 ml will suffice. Duplicates are useful if loss or breakage may occur.
Duplicates that are labeled so it is not obvious that they are duplicates (e.g., given a fake
sample number) are useful as a check of the laboratory's precision.
Answering hydrogeological questions almost always requires knowledge of the
whole water cycle, so it is usually necessary to sample precipitation and surface water, in
addition to groundwater (Orlowski et al., 2016).
8.2 Precipitation
From a hydrogeological perspective, the main reason to sample precipitation is to
better understand recharge. The isotope composition of precipitation provides an input
value for water cycle or catchment related studies. The isotope composition of recharge
may differ from that of precipitation, primarily due to evaporation during infiltration,
however, knowing the original precipitation value is useful for many applications.
8.2.1 Spatial and Temporal Variation in Precipitation

Precipitation varies isotopically over time and space. Spatial variation increases
with the size of the study area, the climatic complexity (erratic precipitation or multiple
types of weather systems) and terrain complexity (mainly altitude change). It is the
responsibility of the investigator to ensure capture of such variation. Temporal variation
occurs by the minute in rain events (Muller et al., 2015) as well as over many years
(Diamond and Harris, 2019). The more completely that all of the rain is captured and the
longer the period of sampling (years to decades), the more accurate the average isotope
composition. Over decades of sampling, climate change is likely to have an impact so
long-term averages may reflect rising or falling trends, rather than steady means.
Sampling of precipitation can be challenging. The ultimate goal is to capture all
precipitation that would otherwise have fallen on the ground and to measure how much
fell, according to local or internationally accepted measurement methods (Bosman, 1981).
Challenges arise due to many possible factors, including strong winds that can blow over
precipitation collectors or blow the precipitation past the opening, heavy rain that can fill
68
the collector in under 24 hours, low precipitation in hot and dry climates that can evaporate
from the collector, snow and ice that can block openings, and so forth. Wildlife can also
pose hazards, such as insects clogging tubes or drinking the captured water, birds
defecating into the collector, lizards falling in the collector and dying, monkeys breaking
parts of the equipment or a rhinoceros inadvertently having a scratch against the collector
and totally destroying it without noticing! Fortunately, interference by organisms tends to
be uncommon or to not affect the isotope ratios. The last possible issue is human tampering,
which is a very real concern in some places.
8.2.2 Standard Daily or Cumulative Monthly Collectors

Standard rain gauges tend to be designed to be emptied at 08h00 daily, with a
typical maximum volume that represents 100 mm precipitation. Using a manned
precipitation collection site where this happens, typically at a weather station that records
other parameters, is ideal, as any problems will be noticed and can be rectified soon.
However, often precipitation is monitored in possible groundwater recharge areas which
can be remote mountain tops or upper valley slopes, where no personnel are available.
Such settings require a cumulative collector that is emptied monthly. The cumulative
collector needs to achieve two main goals: 1) to capture all the rain that falls; and, 2) to
prevent evaporation from occurring. For the first goal, the tank of the collector needs to be
large enough to contain all the precipitation for the wettest month of the wettest year. For
the second goal, the simplest method is to use a narrow tube of approximately 4 mm
internal diameter to connect the collection funnel with the storage tank (Gröning et al.,
2012). This is wide enough to let water flow down freely but narrow enough to prevent
moisture from easily moving back to the atmosphere (Figure 44). Other methods include
using oil to cover the water but this generally becomes messy, not only in the field, but also
in sample bottles and even back at the laboratory. The evaporation effect by using a narrow
tube was shown to be less than use of oil (Gröning et al., 2012). Shielding the storage tank
from the sun will also help prevent moisture loss.
Sampling at frequencies less than monthly (more than one month's rain per sample)
is not recommended. Too many things can go wrong and the loss of one sample result in
the loss of data for a greater period of time and could considerably derail research.
69
Figure 44 - A cumulative precipitation collector, designed to be emptied monthly. These are ideal for remote
locations, such as this mountain top, Blesberg in the Groot Swartberg Mountains, Eastern Cape, South Africa.
8.3 Surface Water

Stable isotopes in surface water bodies may vary spatially and temporally, and
capturing this variation is important to generate a meaningful average, as well as to
understand seasonality, stratification, inflow sources and other hydrologically important
variations. From a hydrogeological perspective, the isotopic variations in a surface water
body may help determine the degree or nature of surface water-groundwater interaction.
Temporal changes in surface water isotope composition are not as sudden as with
precipitation, which changes by the minute, but it is equally important to capture them
because these isotopic changes are often associated with large changes in flow. A flood may
only last a few hours in a small catchment, but may be a substantial percentage of the mean
annual runoff. This is particularly true for arid environments where watercourses may not
flow much during most of the year such that the mean annual runoff may occur in a single
event. In more temperate catchments where the mean annual runoff is the result of many
storms throughout the year, temporal changes are important because seasonal, and other
changes, in weather or land use affect flow dynamics and water quality. Increasingly erratic
and stormy weather is predicted due to global climate change and this will require a
preparedness to capture such variations in order to obtain representative samples (IPCC,
2014).
The extent of spatial variation of isotopes in surface water bodies depends largely
on the water body type. Rivers tend to be fairly well mixed, so the spatial variations are
70
restricted to upstream and downstream as influenced by inputs and outputs. Lakes are
more complex because they may be stratified and may vary temporally between net inflow
or net outflow. Reservoirs (man-made) have similar properties to lakes. Wetlands are even
more complex in that they may be seasonal or episodic, having no surface water in dry
periods that may last months to years. The role of groundwater is usually very important
in wetlands.
The degree to which a study needs to capture the temporal and spatial variations in
surface water is dependent on the goals of the study. Researchers must be aware that
having only one sample from a single location is not as reliable as having multiple samples.
The value of one sample for representing the larger system, depends on the system.
8.4 Groundwater
8.4.1 Natural Groundwater Sources: Springs and Seeps
Natural sampling points for groundwater are most easily sampled. If totally
natural, then the water is taken directly from the source (Figure 45). For springs
(particularly hot springs) that have been capped or developed into resorts, it is best to
sample as close to the source as possible. This may require finding the inlet to a hot pool,
which is often an enjoyable sampling experience!
Figure 45 - Sibusiso Mbonani sampling gas from a CO2-rich cold-water spring near the uMtamvuna River
on the Bongwana Fault, kwaZulu-Natal, South Africa. The water is sampled easily, but the gas is collected
by first filling the sample bottle with water, then inverting it and capturing enough gas to fully displace all
water and then sealing the bottle whilst still underwater. Stable isotopes of the CO 2 suggest dissolution of
carbonate rocks at depth as the source of CO2 (Harris et al., 1997), but more work remains to be done on
the groundwater source and flow path.
71
8.4.2 Artificial Groundwater Sampling Points: Wells, Boreholes and Piezometers

Purging
Depending on the borehole construction and groundwater flow regime,
groundwater may pass through the borehole casing and be relatively fresh, or it may not
be part of the active flow system and become stagnant. If this condition is known, sampling
procedures can be adjusted, but in many cases, the flow regime and in particular, the
borehole construction, are not known. It is therefore advisable, to always purge a borehole
before sampling.
Borehole purging requires removing enough water from the borehole to provide a
fresh sample of water from the aquifer, which is representative of the natural groundwater.
The rule of thumb is to purge 3 times the borehole volume before sampling. Calculating
this value requires knowledge of the borehole diameter (this is usually the same as that
seen at surface but may decrease with depth) and the height of the water column in the
borehole. This height is the difference between the water level and the bottom of the
borehole. Unfortunately, in many cases the depth of the borehole is not known. If this is the
case, and the bottom cannot be sounded, then a rough guess can be made based on nearby
boreholes in the aquifer being targeted to estimate the number of liters stored in the
borehole and therefore, how much purging must be done.
In other cases, boreholes are equipped with pumps and the owner refuses to move
these, and one has little idea of the water level and borehole depth. In this case, again, it is
best to estimate values to calculate the volume to be purged. On the other hand, in many
of these equipped boreholes, the groundwater is being pumped frequently and so a full
purge is not as essential as in a research borehole that may have been standing unpumped
for months or years.
If one has a thermometer, or other field probes, then these can also be used to
determine when fresh aquifer water is being pumped because such readings should
stabilize after sufficient purging. With temperature probes, it is useful to be aware of the
difference between the air temperature (at which the probe will start reading) and the
groundwater temperature. It may seem the groundwater is changing temperature rapidly
at first, but that is the probe adjusting to the groundwater temperature.
Groundwater Heterogeneity
Groundwater is often assumed to be well mixed such that one sample will be
representative of an aquifer. This is not always the case, and various investigations have
discovered heterogeneity on a scale from centimeters upwards. Pollutants are well known
to be heterogeneously distributed due to point source releases, accumulation near the water
table, or other factors (Lasagna and De Luca, 2016; Ronen et al., 1987), but natural water
quality parameters also vary substantially, even when the aquifer matrix appears relatively
homogenous (Baloyi and Diamond, 2019; Petrella and Celico, 2012). The causes of these
72
differences are complex and beyond the discussion here. However, it is important to
understand how well a sample, or set of samples, represents the aquifer, bearing in mind
issues such as heterogeneity. For example, a single sample may contain water that is not
representative because of the variation from place to place, including materials and
conditions between boreholes, or vertically within one borehole. This sample is not
representative of the aquifer because it only characterizes one of the various water types
present. On the other hand, the process of sampling, especially pumping at high rates from
a borehole open to, or screened in, a large interval of the aquifer, can mix water from
different levels that may have quite distinct compositions. This sample is not representative
because it averages some or all of the water types present. For some investigations, often
for regional studies, these sampling errors may not be an issue, but for others, such as
characterizing contamination at a site, isolating discrete intervals and sampling them to
preserve the natural heterogeneity is important. Technology to sample discrete horizons
has been developed in more recent years (Cherry et al., 2007).
From a stable isotope perspective, many of the variations in water chemistry may
not be accompanied by related changes in stable isotope composition. However, some
hydrochemical changes may be associated with stable isotope changes, such as evaporation
prior to recharge causing an increase in salinity, and an associated increase in  values.
Leaking water mains (or sewers) with hydrochemistry (or pollution levels) different from
that of the local groundwater, may or may not reflect in the stable isotopes, depending on
the source of municipal water relative to the source of local groundwater. Petrella and
Celico (2012) mapped stable isotopes from the water table downward in a carbonate aquifer
in Italy and found the upper 12 m or so to contain varied compositions, while below that
the groundwater was better mixed. They concluded the level at which mixing is achieved
depends upon the physical nature of the aquifer as well as the hydraulic conditions, for
example the precipitation history and water level fluctuations.
73
9 Further Reading
For me, the materials listed in this section serve as substantial resources of
information about stable isotope behavior, data analysis, interpretation and presentation,
and will be useful in any aspirant stable isotope hydrologist's library, real or digital. I list
them in chronological order to preserve the natural development of ideas.
Isotopic Variations in Meteoric Waters by Harmon Craig. Science, volume 133, pages
1702-1703, 1961, https://www.science.org/doi/10.1126/science.133.3465.1702.
Stable isotopes in precipitation by Willi Dansgaard. Tellus, volume 16, pages 436-468, 1964,
https://www.tandfonline.com/doi/abs/10.3402/tellusa.v16i4.8993.
Stable Isotope Hydrology - Deuterium and Oxygen-18 in the Water Cycle by Joel Gat and
Roberto Gonfiantini (editors). International Atomic Energy Agency, Technical
Reports Series #210, 339 pages, 1981,
https://inis.iaea.org/search/search.aspx?orig_q=RN:13677657.
Isotopic Patterns in Modern Global Precipitation by Kazimierz Rozanski, Luis
Araguás-Araguás and Roberto Gonfiantini. In: Climate Change in Continental
Isotopic Records, Swart PK, Lohmann KC, McKenzie J and Savin S (editors),
Geophysical Monograph 78, American Geophysical Union, pages 1-36, 1993,
https://agupubs.onlinelibrary.wiley.com/doi/abs/10.1029/GM078p0001.
Oxygen and hydrogen isotopes in the hydrologic cycle by Joel Gat. Annual Reviews in Earth &
Planetary Sciences, volume 24, pages 225-262, 1996,
https://www.annualreviews.org/doi/abs/10.1146/annurev.earth.24.1.225.
Stable Isotopes in Plant Ecology. Todd Dawson, Stefania Mambelli, Agneta Plamboeck,
Pamela Templer and Kevin Tu. Annual Review of Ecology and Systematics, volume
33, pages 507-559, 2002,
https://www.annualreviews.org/doi/abs/10.1146/annurev.ecolsys.33.020602.095451.
Hydrograph separation using stable isotopes: Review and evaluation by Julian Klaus and Jeffrey
McDonnell. Journal of Hydrology, volume 505, pages 47-64, 2013,
https://water.usask.ca/hillslope/documents/pdfs/2013/13-
08%20Klaus2013JOH_505_47-64.pdf.
Groundwater Geochemistry and Isotopes by Ian Clark. CRC Press, Boca Raton, 438 pages,
2015, https://www.taylorfrancis.com/books/mono/10.1201/b18347/groundwater-
geochemistry-isotopes-ian-clark.
Global Isotope Hydrogeology -- Review by Scott Jasecko. Reviews of Geophysics, volume 57,
pages 835-965, 2019,
https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2018RG000627.
74
10 Wrap-up
Research using the stable isotopes of water was originally focused on the major
flows of water in the hydrological cycle. This included evaporation from the oceans,
condensation from vapor to cloud, and precipitation in different environments across the
globe. It was during this time that the effects of temperature, latitude, altitude,
continentality, precipitation amount, source region, humidity and other parameters on
isotope composition of precipitation became known. Once these global patterns of isotope
composition were fairly well understood in the oceans, lakes, rivers and groundwater,
researchers started more detailed regional and local investigations. These focused works
provided insight regarding transpiration, recharge, groundwater flow, hydrograph
separation, mixing of water masses and other processes.
By the late 20th century, stable isotope studies were integrating information from
geology, hydrochemistry, other isotope systems (both stable and radioactive) as well as
biological work. Topics being addressed included geothermal studies, paleowaters,
paleoclimatic interpretation, and plant water sources.
Moving into the 21st century, the arrival of laser cavity instruments reduced the time
and cost of isotope analysis, previously only done by mass spectrometer. Oxygen and
hydrogen isotopes in water could now be routinely analyzed by mining companies,
agricultural interests and local government, generating huge datasets that could be used to
monitor for leaking canals or pipes, movement of pollution plumes and determination of
the sustainability of groundwater abstraction (Figure 46). Isotope hydrology had become a
mature science, used by researchers at universities while being applied by natural resource
managers.
75
Figure 46 - Artesian flow from a future production borehole on the Coega Fault,
tapping the Table Mountain Group, Eastern Cape, South Africa. Stable isotope
information can help manage such large, deep groundwater reserves sustainably.
The forefront of stable isotope hydrology now includes analysis of global datasets
to quantify changes in groundwater storage, applications to urban water issues and use of
triple oxygen isotopes (16O, 17O and 18O) to re-evaluate the existing meteoric water lines and
kinetic fractionation effects. At the same time, the greater accessibility to information via
the internet and the lower cost of isotopic analysis has allowed the benefits of this science
to penetrate less-wealthy parts of the world and assist in understanding developing world
issues, such as community water supply, sanitation, environmental protection and
sustainability.
Isotope hydrology is poised to continue expanding, both in research settings and
for application to day-to-day water management. This book was written to educate and
motivate the next generation of isotope hydrologists, so they can advance the science and
tackle the swathe of environmental challenges that need to be addressed for humanity to
continue to inhabit a healthy and beautiful planet.
76
11 Exercises
Exercise 1 - Meteoric Water Line Calculations
Meteoric water lines are not essential for every stable isotope study, but they are
needed frequently, particularly when comparing surface water or groundwater to
precipitation. The calculation of a meteoric water line was covered theoretically in this
book, but this exercise provides readers with the opportunity to develop a MWL.
Precipitation and stable isotope data for the University of Cape Town for 2010 are
given in the table below. Using the data:
1. Calculate the local meteoric water line using the RMA (reduced major axis)
regression method and calculate the Pearson's r correlation coefficient; and,
2. Calculate the precipitation-weighted RMA regression line and its Pearson's r.
Stable isotope data for UCT precipitation

Precipitation 18
 O D
Location Date
mm ‰ ‰
UCT Jan 2010 9 +0.75 +9.7
UCT Feb 2010 12 -0.54 -5.8
UCT Mar 2010 7 -0.83 +0.4
UCT Apr 2010 44 -2.25 -12.1
UCT May 2010 277 -2.78 -8.4
UCT Jun 2010 222 -3.31 -17.0
UCT Jul 2010 166 -3.69 -15.0
UCT Aug 2010 121 -1.74 -4.9
UCT Sep 2010 100 -2.24 -3.3
UCT Oct 2010 102 -2.15 -3.3
UCT Nov 2010 73 -2.39 -6.2
UCT Dec 2010 17 -5.06 -39.9
Click for solution to Exercise 1
77
Exercise 2 - Calculation of Recharge Altitude

Determine the recharge altitude of the Kirstenbosch spring, which is 160 masl, by working
through the 4 items listed here.
1. Calculate the mean isotopic composition of the spring water based on the
measurements provided in this table.
Stable isotope data for the Kirstenbosch spring

D 18
 O
date
∆‰ ∆‰
March
-13.7 -3.01
2010
November
-9.4 -3.03
2010
May 2011 -12.4 -3.71
October
-12.9 -3.04
2011
May 2012 -10.2 -3.48
September
-10.1 -2.37
2012
2. Calculate the weighted mean stable isotope compositions (ignore missing values) for
the University of Cape Town (135 masl) and the Table Mountain Cableway (1074 masl)
using the 3-year dataset provided in the table on the following page.
3. Use the weighted means caclulated in item 2 to calculate an altitude effect for Table
Mountain.
4. Calculate the recharge altitude of the Kirstenbosch spring.
78
Stable isotope data for precipitation at UCT and the Table Mountain Upper Cableway Station.
UCT TMC
D 18
 O Precipitation D 18
 O Precipitation
‰ ‰ mm ‰ ‰ mm
J +9.7 +0.75 9
F -5.8 -0.54 13
M +0.4 -0.83 7
A -12.1 -2.25 45
M -8.4 -2.78 275 -15.0 -3.14 280
J -17.0 -3.31 201 -24.4 -4.35 170
2010
J -15.0 -3.69 166 200

A -4.9 -1.74 121 150
S -3.3 -2.24 101 100
O -3.3 -2.15 103 -16.3 -3.04 120
N -6.2 -2.39 73 -13.1 -2.65 90
D -39.9 -5.06 17 -16.0 -2.35 60
J -5.2 -0.15 16 -1.2 -0.75 30
F 13.9 +3.19 3 20
M -23.6 -4.12 16 50
A -5.8 -2.73 111 -14.8 -4.61 90
M -12.3 -3.69 145 -14.7 -4.17 142
J -18.3 -4.82 191 -16.9 -4.85 150
2011
J -14.6 -4.68 49 -11.6 -4.71 70

A -3.8 -2.63 209 -14.4 -4.64 165
S -0.1 -1.51 141 -9.2 -3.51 115
O -17.0 -3.48 41 -13.6 -3.84 100
N -12.0 -3.21 72 -17.4 -4.64 100
D -1.6 -2.05 62 -9.6 -3.51 90
J +5.4 +0.78 46 -4.7 -2.93 130
F -2.5 -0.68 21 -1.0 -2.05 80
M -11.9 -3.07 69 -12.9 -3.91 110
A -9.8 -3.12 133 -13.7 -4.34 130
M -10.4 -3.08 143 155
J -6.4 -4.15 230 -8.7 -2.32 200
2012
J -8.8 -2.67 300 -7.6 -2.93 240

A -13.5 -3.25 214 -20.8 -4.82 145
S -10.8 -2.36 172 -22.4 -5.75 215
O -7.1 -3.16 44 123
N +0.1 -1.08 60 44
D +6.9 -0.17 10 20
Click for solution to Exercise 2
79
Exercise 3 - Hydrograph Separation

In the most simplified case, the surface water flow from a catchment may comprise
only baseflow (groundwater discharge to the river). One notch of added complexity is that
the surface water may comprise a mixture of baseflow and precipitation. For this scenario
in Yongan watershed in Zhejiang, eastern China, estimate the fractions of groundwater and
precipitation contributing to surface water flow using the data from Hu and others (2020)
provided in the table below. Comment on the significance of the errors (as shown in the
table) to the estimated proportions.
Stable isotope data for key flows in Yongan, Zhejiang, China

2 18
Sample Type  H (‰)  O (‰)
Precipitation -33.9 ± 24.1 -6.4 ± 2.9
River water -42.8 ± 6.1 -6.9 ± 1.1
Groundwater -44.7 ± 3.3 -7.2 ± 0.9
Click for solution to Exercise 3
80
12 References
Abouelmagd, A., M. Sultan, A. Milewski, A.E. Kehew, N.C. Sturchio, F. Soliman, R.V.
Krishnamurthy, and E. Cutrim, 2012, Toward a better understanding of
palaeoclimatic regimes that recharged the fossil aquifers in North Africa: Inferences
from stable isotopes and remote sensing data. Palaeogeography, Palaeoclimatology,
Palaeoecology, volume 329-330, 13 pages, doi:10.1016/j.palaeo.2012.02.024.
Adar, E.M., A. Dody, M.A. Geyh, A. Yair, A. Yakirevich, and A.S. Issar, 1998, Distribution
of stable isotopes in arid storms I. Relation between the distribution of isotopic
composition in rainfall and in consequent runoff. Hydrogeology Journal, volume 6,
16 pages, doi:10.1007/s100400050133.
Al-Charideh, A. and B. Kattaa, 2016, Isotope hydrology of deep groundwater in Syria:
renewable and non-renewable groundwater and paleoclimate impact. Hydrogeology
Journal, volume 24, 20 pages, doi:10.1007/s10040-015-1324-4.
Al Faitouri, M. and W.E. Sanford, 2015, Stable and radio-isotope analysis to determine
recharge timing and paleoclimate of sandstone aquifers in central and southeast
Libya. Hydrogeology Journal, volume 23, 11 pages, doi:10.1007/s10040-015-1232-7.
Alison, G.B. and M.W. Hughes, 1983, The use of natural tracers as indicators of soil-water
movement in a temperate semi-arid region. Journal of Hydrology, volume 60,
17 pages, doi:10.1016/0022-1694(83)90019-7.
Andrews, J.E., 2006, Palaeoclimatic records from stable isotopes in riverine tufas:
Synthesis and reviews. Earth Science Reviews, volume 75, 20 pages,
doi:10.1016/j.earscirev.2005.08.002.
Antunes, P., D.F. Boutt, and F.C. Rodrigues, 2019, Orographic distillation and
spatio-temporal variations of stable isotopes in precipitation in the North Atlantic.
Hydrological Processes, volume 33, 19 pages, doi:10.1002/hyp.13362.
Araguás-Araguás, L., K. Froehlich, and K. Rozanski, 1998, Stable isotope composition of
precipitation over southeast Asia. Journal of Geophysical Research, volume 103,
number D22, 22 pages, doi:10.1029/98JD02582.
Araguás-Araguás, L., K. Froehlich, and K. Rozanski, 2000, Deuterium and oxygen-18
isotope composition of precipitation and atmospheric moisture. Hydrological
Processes, volume 14, 15 pages,
doi:10.1002/1099-1085(20000615)14:8<1341::AID-HYP983>3.0.CO;2-Z.
Baloyi, R.S. and R.E. Diamond, 2019, Variable water quality of domestic wells emphasizes
the need for groundwater quality monitoring and protection: Stinkwater,
Hammanskraal, Gauteng. Water SA, volume 45, 9 pages, doi:10.4314/wsa.v45i2.08.
Beven, K.J., 2001, Rainfall-Runoff Modelling - The Primer. John Wiley and Sons, Ltd.,
New York, 488 pages.
Biggs, T.W., C.T. Lai, P. Chandan, R.M. Lee, A. Messina, R.S. Lesher, and N. Khatoon,
2015, Evaporative fractions and elevation effects on stable isotopes of high elevation
81
lakes and streams in arid western Himalaya. Journal of Hydrology, volume 522,
11 pages, doi:10.1016/j.jhydrol.2014.12.023.
Blarasin, M., A. Cabrera, I. Matiatos, V. Lutri, L. Maldonado, D. Giacobone, E. Matteoda,
F.B. Quinodoz, J.G. Albo, C.E. Felizzia, and J. Felizzia, 2020, Application of isotope
techniques to enhance the conceptual hydrogeological model and to assess
groundwater sustainability in the Pampean Plain in Cordoba, Argentina. Isotopes in
Environmental and Health Studies, volume 56, 16 pages,
doi:10.1080/10256016.2020.1796658.
Bodé, S., L. de Wispelaere, A. Hemp, D. Verschuren, and P. Boeckx, 2019, Water-isotope
ecohydrology of Mount Kilimanjaro. Ecohydrology, volume 13, issue 1, 19 pages,
doi:10.1002/eco.2171.
Boschetti, T., J. Cifuentes, P. Iacumin, and E. Selmo, 2019, Local meteoric water line of
northern Chile (18-30°S): An application of error-in-variables regression to the
oxygen and hydrogen stable isotope ratio of precipitation. Water, volume 11, issue 4,
791 pages, doi:10.3390/w11040791.
Bosman, H.H., 1981, Raingauges: Quality pays. Water SA, volume 7, 2 pages.
Camacho Suarez, V.V., A.M. Saraiva Okello, J.W. Wenninger, and S. Uhlenbrook, 2015,
Understanding runoff processes in a semi-arid environment through isotope and
hydrochemical hydrograph separations. Hydrology and Earth System Sciences,
volume 19, 17 pages, doi:10.5194/hess-19-4183-2015.
Canavan, R.R., B. Carrapa, M.T. Clementz, J. Quade, P.G. DeCelles, and L.M.
Schoenbohm, 2014, Early Cenozoic uplift of the Puna Plateau, Central Andes, based
on stable isotope paleoaltimetry of hydrated volcanic glass. Geology, volume 42,
4 pages, doi:10.1130/G35239.1.
Chacko, T., D.R. Cole, and J. Horita, 2001, Equilibrium oxygen, hydrogen and carbon
isotope fractionation factors applicable to geological systems, editors, Valley J.W. and
D.R. Cole, Stable Isotope Geochemistry, Reviews in Mineralogy and Geochemistry,
volume 43, 81 pages, doi:10.2138/gsrmg.43.1.1.
Chandrajith, R., J.A. Barth, N.D. Subasinghe, D. Merten, and C.B. Dissanayake, 2012,
Geochemical and isotope characterisation of geothermal spring waters in Sri Lanka:
Evidence for steeper than expected geothermal gradients. Journal of Hydrology,
volume 476, 10 pages, doi:10.1016/j.jhydrol.2012.11.004.
Cherry, J.A., B.L. Parker, and C. Keller, 2007, A new depth-discrete multilevel monitoring
approach for fractured rock. Groundwater Monitoring and Remediation, volume 27,
14 pages, doi:10.1111/j.1745-6592.2007.00137.x.
Clark, I., 2015, Groundwater Geochemistry and Isotopes, CRC Press, Boca Raton,
456 pages, doi:10.1201/b18347.
Clark, I.D. and P. Fritz, 1997, Environmental Isotopes in Hydrogeology, CRC Press, Boca
Raton, 342 pages, doi:10.1201/9781482242911.
82
Cook, P., 2020, Introduction to isotopes and environmental tracers as indicators of

groundwater flow. The Groundwater Project, Guelph, Ontario, Canada, 74 pages,
https://gw-project.org/books/introduction-to-isotopes-and-environmental-tracers-as-i
ndicators-of-groundwater-flow/.
Cook, P.G. and A.P. O'Grady, 2006, Determining soil and groundwater use of vegetation
from heat pulse, water potential and stable isotope data. Oecologia, volume 148,
11 pages, doi:10.1007/s00442-005-0353-4.
Craig, H., 1961, Isotopic variations in meteoric waters. Science, volume 133, 2 pages,
doi:10.1126/science.133.3465.1702.
Craig, H. and L.I. Gordon, 1965, Deuterium and oxygen 18 variations in the ocean and the
marine atmosphere, editor, Tongiorgi, E., Stable Isotopes in Oceanographic Studies
and Palaeotemperatures. Proceedings of the Spoleto Conferences in Nuclear Geology,
Pisa, 122 pages.
D'Amore, F. and C. Panichi, 1985, Geochemistry in geothermal exploration. Energy
Research, volume 9, 22 pages, doi:10.1002/er.4440090307.
Daniele, L., M. Taucare, B. Viguier, G. Arancibia, D. Aravena, T. Roquer, J. Sepúlveda, E.
Molina, A. Delgado, M. Muñoz, and D. Morata, 2020, Exploring the shallow
geothermal resources in the Chilean Southern Volcanic Zone: Insight from the
Liquiñe thermal springs. Journal of Geochemical Exploration, volume 218, 15 pages,
doi:10.1016/j.gexplo.2020.106611.
Dansgaard, W., 1964, Stable Isotopes in Precipitation. Tellus, volume 16, 33 pages,
doi:10.3402/tellusa.v16i4.8993.
Das, P., A. Mukherjee, S.A. Hussain, M.S. Jamal, K. Das, A. Shaw, M.K. Layek, and P.
Sengupta, 2020, Stable isotope dynamics of groundwater interactions with Ganges
River. Hydrological Processes 2021, volume 35, issue 1, 15 pages,
doi:10.1002/hyp.14002.
Dawson, T.E. and J.R. Ehleringer, 1991, Streamside trees that do not use stream water.
Nature, volume 350, 3 pages, doi:10.1038/350335a0.
Dawson, T.E., S. Mambelli, A.H. Plamboeck, P.H. Templer, and K.P. Tu, 2002, Stable
isotopes in plant ecology. Annual Review of Ecology and Systematics, volume 33,
53 pages, doi:10.1146/annurev.ecolsys.33.020602.095451.
Diamond, R.E. and C. Harris, 1997, Oxygen and hydrogen isotope composition of
Western Cape meteoric water. South African Journal of Science, volume 93, 4 pages,
https://hdl.handle.net/10520/AJA00382353_74.
Diamond, R.E. and C. Harris, 2000, Oxygen and hydrogen isotope geochemistry of
thermal springs of the Western Cape, South Africa: recharge at high altitude?. Journal
of African Earth Sciences, volume 31, 15 pages, doi:10.1016/S0899-5362(00)80002-0.
Diamond, R.E. and C. Harris, 2019, Annual shifts in O- and H-isotope composition as
measures of recharge: the case of the Table Mountain Springs, Cape Town, South
Africa. Hydrogeology Journal, volume 27, 16 pages, doi:10.1007/s10040-019-02045-5.
83
Diamond, R.E. and S. Jack, 2018, Evaporation and abstraction determined from stable
isotopes during normal flow on the Gariep River, South Africa. Journal of
Hydrology, volume 559, 16 pages, doi:10.1016/j.jhydrol.2018.02.059.
Dogramaci, S., G. Skrzypek, W. Dodson, and P.F. Grierson, 2012, Stable isotope and
hydrochemical evolution of groundwater in the semi-arid Hamersley Basin of
subtropical northwest Australia. Journal of Hydrology, volume 475, 13 pages,
doi:10.1016/j.jhydrol.2012.10.004.
Dogramaci, S., G. Firmani, P. Hedley, G. Skrzypek, and P.F. Grierson, 2015, Evaluating
recharge to an ephemeral dryland stream using a hydraulic model and water,
chloride and isotope mass balance. Journal of Hydrology, volume 521, 13 pages,
doi:10.1016/j.jhydrol.2014.12.017.
Druhan, J.L. and K. Maher, 2017, The influence of mixing on stable isotope ratios in
porous media: A revised Rayleigh model. Water Resources Research, volume 53,
24 pages, doi:10.1002/2016WR019666.
Durowoju, O.S., M. Butler, G.E. Ekosse, and J.O. Odiyo, 2019, Hydrochemical processes
and isotopic study of geothermal springs within Soutpansberg, Limpopo Province,
South Africa. Applied Sciences, volume 9, issue 8, 18 pages,
doi:10.3390/app9081688.
Ehleringer, J.R. and T.E. Dawson, 1992, Water uptake by plants: perspectives from stable
isotope composition. Plant, Cell and Environment, volume 15, 10 pages,
doi:10.1111/j.1365-3040.1992.tb01657.x.
Ellsworth, P.Z. and D.G. Williams, 2007, Hydrogen isotope fractionation during water
uptake by woody xerophytes. Plant Soil, volume 291, 15 pages,
doi:10.1007/s11104-006-9177-1.
Emiliani, C., 1987, Dictionary of the Physical Sciences. Oxford University Press, New
York.
Epstein, S. and T. Mayeda, 1953, Variation of 18O content of waters from natural sources.
Geochimica et Cosmochimica Acta, volume 4, 12 pages,
doi:10.1016/0016-7037(53)90051-9.
Fontes, J.Ch., 1981, Palaeowaters, editors, Gat J.R. and R. Gonfiantini, Stable Isotope
Hydrology, Technical Reports Series 210, International Atomic Energy Agency,
Vienna, 30 pages,
http://www-naweb.iaea.org/napc/ih/documents/IAEA%20Monographs/TRS%20210%
20Stable%20Isotope%20Hydrology1981.pdf.
Friedman, I., 1953, Deuterium content of natural waters and other substances. Geochimica
et Cosmochimica Acta, volume 4, 15 pages, doi:10.1016/0016-7037(53)90066-0.
Gat, J.R., 1981, Historical Introduction, editors, Gat J.R. and R. Gonfiantini, Stable Isotope
Hydrology, Technical Reports Series 210, International Atomic Energy Agency,
Vienna, 6 pages,
84
Gat, J.R., 1996, Oxygen and hydrogen isotopes in the hydrological cycle. Annual Reviews
in Earth and Planetary Sciences, volume 24, 38 pages,
doi:10.1146/annurev.earth.24.1.225.
Gat, J.R. and E. Matsui, 1991, Atmospheric water balance in the Amazon Basin: an
isotopic evapotranspiration model. Journal of Geophysical Research, volume 96,
issue D7, 10 pages, doi:10.1029/91JD00054.
Gat, J.R., D. Yakir, G. Goodfriend, P. Fritz, P. Trimborn, J. Lipp, I. Gev, E. Adar, and Y.
Waisel, 2007, Stable isotope composition of water in desert plants. Plant Soil,
volume 298, 15 pages, doi:10.1007/s11104-007-9321-6.
Gillespie, J., S.T. Nelson, A.L. Mayo, and D.G. Tingey, 2012, Why conceptual groundwater
flow models matter: a trans-boundary example from the arid Great Basin, western
USA. Hydrogeology Journal, volume 20, 15 pages, doi:10.1007/s10040-012-0848-0.
Goldsmith, G.R., S.T. Allen, S. Braun, N. Engbersen, C.R. González-Quijano, J.R. Kirchner,
and R.T. Siegwolf, 2018, Spatial variation in throughfall, soil, and plant water
isotopes in a temperate forest. Ecohydrology, volume 12, 11 pages,
doi:10.1002/eco.2059.
Gonfiantini, R., 1981, The  notation and the mass-spectrometric measurement techniques,
editors, Gat J.R. and R. Gonfiantini, Stable Isotope Hydrology, International Atomic
Energy Agency, Technical Reports Series #210, Vienna. Chapter 4, 50 pages,
https://inis.iaea.org/collection/NCLCollectionStore/_Public/13/677/13677657.pdf?r=1.
Gonfiantini, R., M.A. Roche, J.C. Olivry, J.C. Fontes, and G.M. Zuppi, 2001, The altitude
effect on the isotopic composition of tropical rains. Chemical Geology, volume 181,
21 pages, doi:10.1016/S0009-2541(01)00279-0.
Grimmeisen, F., M.F. Lehmann, T. Liesch, N. Goeppert, J. Klinger, J. Zopfi, and N.
Goldscheider, 2017, Isotopic constraints on water source mixing, network leakage
and contamination in an urban groundwater system. Science of the Total
Environment, volume 583, 12 pages, doi:10.1016/j.scitotenv.2017.01.054.
Gröning, M., H.O. Lutz, Z. Roller-Lutz, M. Kralik, L. Gourcy, and L. Pöltenstein, 2012, A
simple rain collector preventing water re-evaporation dedicated for delta O-18 and
delta H-2 analysis of cumulative precipitation samples. Journal of Hydrology,
volume 448, 6 pages, doi:10.1016/j.jhydrol.2012.04.041.
Han, T., M. Zhang, S. Wang, D. Qu, and Q. Du, 2020, Sub-hourly variability of stable
isotopes in precipitation in the marginal zone of East Asian monsoon. Water,
volume 12, issue 8, 14 pages , doi:10.3390/w12082145.
Harris, C., C. Burgers, J. Miller, and F. Rawoot, 2010, O- and H-isotope record of Cape
Town rainfall from 1996 to 2008, and its application to recharge studies of Table
Mountain groundwater, South Africa. South African Journal of Geology, volume 113,
24 pages, doi:10.2113/gssajg.113.1.33.
85
Harris, C., W.D. Stock, and J. Lanham, 1997, Stable isotope constraints on the origin of
CO2 gas exhalations at Bongwan, Natal. South African Journal of Geology,
volume 100, 6 pages, https://journals.co.za/doi/pdf/10.10520/EJC-92945afa1.
Hayes, J.M., 2004, An Introduction to Isotopic Calculations. Woods Hole Oceanographic
Institution, Woods Hole, Massachusetts, USA, 10 pages,
https://www.whoi.edu/cms/files/jhayes/2005/9/IsoCalcs30Sept04_5183.pdf.
Hooper, R.P. and C.A. Shoemaker, 1986, A comparison of chemical and isotopic
hydrograph separation. Water Resources Research, volume 22, 11 pages.
Horibe, Y. and M. Kobayakawa, 1960, Deuterium abundance of natural waters.
Geochimica et Cosmochimica Acta, volume 20, 11 pages,
doi:10.1016/0016-7037(60)90078-8.
Hu, M., Y. Zhang, K. Wu, H. Shen, M. Yao, R.A. Dahlgren, and D. Chen, 2020, Assessment
of streamflow components and hydrologic transit times using stable isotopes of
oxygen and hydrogen in waters of a subtropical watershed in eastern China. Journal
of Hydrology, volume 589, 11 pages, doi:10.1016/j.jhydrol.2020.125363.
Hughes, C.E. and J. Crawford, 2012, A new precipitation weighted method for
determining meteoric water line for hydrological applications demonstrated using
Australia and global GNIP data. Journal of Hydrology, volume 465, 8 pages,
doi:10.1016/j.jhydrol.2012.07.029.
Husain, M.M., J.A. Cherry, and S.K. Frape, 2004, The persistence of a large stagnation
zone in a developed regional aquifer, southwestern Ontario. Canadian Geotechnical
Journal, volume 41, 16 pages, doi:10.1139/t04-040.
IPCC, 2014, Intergovernmental Panel on Climate Change, Climate Change 2014: Synthesis
Report. Contribution of Working Groups I, II and III to the Fifth Assessment Report
of the Intergovernmental Panel on Climate Change [Core Writing Team, editors,
Pachauri R.K. and L.A. Meyer]. IPCC, Geneva, Switzerland, 151 pages,
https://archive.ipcc.ch/pdf/assessment-report/ar5/syr/SYR_AR5_FINAL_full_wcover.
pdf.
Iqbal, M.Z., 1998, Application of environmental isotopes in storm-discharge analysis of
two contrasting stream channels in a watershed. Water Research, volume 32,
10 pages, doi:10.1016/S0043-1354(98)00070-0.
Jasecko, S., 2019, Global Isotope Hydrogeology - Review. Reviews of Geophysics,
volume 57, 101 pages, doi:10.1029/2018RG000627.
Jasecko, S., Z.D. Sharp, J.J. Gibson, S.J. Birks, Y. Yi, and P.J. Fawcett, 2013, Terrestrial
water fluxes dominated by transpiration. Nature, volume 496, 4 pages,
doi:10.1038/nature11983.
Kirchner, J.W., 2003, A double paradox in catchment hydrology and geochemistry.
86
Klaus, J. and J.J. McDonnell, 2013, Hydrograph separation using stable isotopes: Review
and evaluation. Journal of Hydrology, volume 505, 18 pages,
doi:10.1016/j.jhydrol.2013.09.006.
Ladouche, B., A. Luc, and D. Nathalie, 2009, Chemical and isotopic investigation of
rainwater in southern France (1996-2002): Potential use as input signal for karst
functioning investigation. Journal of Hydrology, volume 367, 15 pages,
doi:10.1016/j.jhydrol.2009.01.012.
Laonamsai, J., K. Ichiyanagi, K. Kamdee, A. Putthividhya, and M. Tanoue, 2020, Spatial
and temporal distributions of stable isotopes in precipitation over Thailand.
Lasagna, M. and D.A. De Luca, 2016, The use of multilevel sampling techniques for
determining shallow aquifer nitrate profiles. Environmental Science Pollution
Research, volume 23, 18 pages, doi:10.1007/s11356-016-7264-2.
Laudon, H. and O. Slaymaker, 1997, Hydrograph separation using stable isotopes, silica
and electrical conductivity: an alpine example. Journal of Hydrology, volume 201,
20 pages, doi:10.1016/S0022-1694(97)00030-9.
Lelli, M., T.G. Kretzschmar, J. Cabassi, M. Doveri, J.I. Sanchez-Avila, F. Gherardi, G.
Magro, and F. Norelli, 2021, Fluid geochemistry of the Los Humeros geothermal field
(LHGF - Puebla, Mexico): New constraints for the conceptual model. Geothermics,
volume 90, 17 pages, doi:10.1016/j.geothermics.2020.101983.
Li, J., Z. Pang, Y. Kong, S. Wang, G. Bai, H. Zhao, D. Zhou, F. Sun, and Z. Yang, 2018,
Groundwater isotopes biased toward heavy rainfall events and implications on the
local meteoric water line. Journal of Geophysical Research: Atmospheres,
volume 123, 8 pages, doi:0.1029/2018JD028413.
Los Gatos Research, 2020, Off-Axis Integrated Cavity Output Spectroscopy,
http://www.lgrinc.com/advantages.
Lyon, S.W., S.L. Desilets, and P.A. Troch, 2009, A tale of two isotopes: differences in
hydrograph separation for a runoff event when using D and 18O. Hydrological
Processes, volume 23, 7 pages, doi:10.1002/hyp.7326.
Ma, Y.Z., 2019, Quantitative Geosciences: Data Analytics, Geostatistics, Reservoir
Characterization and Modeling. Springer, Cham, Switzerland, 401 pages,
doi:10.1007/978-3-030-17860-4.
Mahindawansha, A., N. Orlowski, P. Kraft, Y. Rothfuss, H. Racela, and L. Breuer, 2018,
Quantification of plant water uptake by water stable isotopes in rice paddy systems.
Plant Soil, volume 429, 22 pages, doi:10.1007/s11104-018-3693-7.
Malov, A.I. and I.V. Tokarev, 2019, Using stable isotopes to characterize the conditions of
groundwater formation on the eastern slope of the Baltic Shield (NW Russia). Journal
of Hydrology, volume 578, 10 pages, doi:10.1016/j.jhydrol.2019.124130.
Maurya, A.S., M. Shah, R.D. Deshpande, R.M. Bhardwaj, A. Prasad, and S.K. Gupta, 2011,
Hydrograph separation and precipitation source identification using stable water
87
isotopes and conductivity: River Ganga at Himalayan foothills. Hydrological

McIntosh, J.C. and L.M. Walter, 2006, Paleowaters in Silurian-Devonian carbonate
aquifers: Geochemical evolution of groundwater in the Great Lakes region since the
Late Pleistocene. Geochimica et Cosmochimica Acta, volume 70, 26 pages,
doi:10.1016/j.gca.2006.02.002.
Merlivat, L. and J. Jouzel, 1979, Global climatic interpretation of the deuterium-oxygen 18
relationship from precipitation. Journal of Geophysical Research, volume 84, C8,
5 pages, doi:10.1029/JC084iC08p05029.
Midgley, J.J. and D.F. Scott, 1994, The use of stable isotopes of water (D and 18O) in
hydrological studies in the Jonkershoek Valley. WaterSA, volume 20, 4 pages,
https://researchspace.csir.co.za/dspace/bitstream/handle/10204/805/Midgley_1994.pdf
?sequence=1&isAllowed=y.
Mix, H.T., S.P. Reilly, A. Martin, and G. Cornwell, 2019, Evaluating the roles of rainout
and post-condensation processes in a landfalling atmospheric river with stable
isotopes in precipitation and water vapor. Atmosphere, volume 10, issue 2, 86 pages,
doi:10.3390/atmos10020086.
Muller, C.L., A. Baker, I.J. Fairchild, C. Kidd, and I. Boomer, 2015. Intra-event trends in
stable isotopes: Exploring midlattitude precipitation using a vertically pointing micro
rain radar. Journal of Hydrometeorology, volume 16, 20 pages,
doi:10.1175/JHM-D-14-0038.1.
Mulligan, B.M., M.C. Ryan, and T.P. Cámbara, 2011, Delineating volcanic aquifer recharge
areas using geochemical and isotopic tools. Hydrogeology Journal, volume 19,
13 pages, doi:10.1007/s10040-011-0766-6.
Munksgaard, N.C., C.M. Wurster, A. Bass, and M.I. Bird, 2012, Extreme short-term stable
isotope variability revealed by continuous rainwater analysis. Hydrological
Nehemy, M.F., P. Benettin, M. Asadollahi, D. Pratt, A. Rinaldo, and J.J. McDonell, 2020,
Tree water deficit and dynamic source water partitioning. Hydrological Processes,
volume 35, 22 pages, doi:10.1002/hyp.14004.
O'Keefe, A. and D. Deacon, 1988, Cavity ringdown optical spectrometer for absorption
measurements using pulsed laser sources. Review of Scientific Instruments,
volume 59, 8 pages, doi:10.1063/1.1139895.
Orlowski, N., P. Kraft, J. Pferdmenges, and L. Breuer, 2016, Exploring water cycle
dynamics by sampling multiple water isotope pools in a developed landscape in
Germany. Hydrology and Earth System Sciences, volume 20, 22 pages,
doi:10.5194/hess-20-3873-2016.
Pan, Y-X., X-P. Wang, X-Z. Ma, Y-F. Zhang, and R. Hu, 2020, The stable isotopic
composition variation characteristics of desert plants and water sources in an
88
artificial revegetation ecosystem in northwest China. Catena, volume 189, 10 pages,

doi:10.1016/j.catena.2020.104499.
Panichi, C. and R. Gonfiantini, 1981, Geothermal Waters, editors, Gat J.R. and R.
Gonfiantini, Stable Isotope Hydrology, Technical Reports Series No 210, International
Atomic Energy Agency, Vienna, 32 pages,
Paul, D., G. Skrzypek, and I. Fórizs, 2007, Normalization of measured stable isotopic
compositions to isotope reference scales - a review. Rapid Communications in Mass
Spectrometry, volume 21, 9 pages, doi:10.1002/rcm.3185.
Paulsson and Widerland, 2020, Pit lake oxygen and hydrogen isotopic composition in
subarctic Sweden: A comparison to the local meteoric water line. Applied
Geochemistry, volume 118, 10 pages, doi:10.1016/j.apgeochem.2020.104611.
Peixoto, J.P. and A.H. Oort, 1983, The atmospheric branch of the hydrological cycle and
climate. In: Variations in the Global Water Budget, Street-Perrott, et al (editors), D
Reidel Publishing Company, 61 pages, doi:10.1007/978-94-009-6954-4_2.
Penna, D., J. Geris, L. Hopp, and F. Scandellari, 2020, Water sources for root water uptake:
using stable isotopes of hydrogen and oxygen as a research tool in agricultural and
agroforestry systems. Agriculture, Ecosystems and Environment, volume 291,
6 pages, doi:10.1016/j.agee.2019.106790.
Penna, D. and H.J. van Meerveld, 2019, Spatial variability in the isotopic composition of
water in small catchments and its effect on hydrograph separation. Wiley Reviews
Water, volume 6, 33 pages, doi:10.1002/wat2.1367.
Peng, T.R., C.H. Wang, C.C. Huang, L.Y. Fei, C.T. Chen, and J.L. Hwong, 2010, Stable
isotope characteristics of Taiwan's precipitation: A case study of western Pacific
monsoon region. Earth and Planetary Science Letters, volume 289, 10 pages,
doi:10.1016/j.epsl.2009.11.024.
Peters, E., A. Visser, B.K. Esser, and J.E. Moran, 2018, Tracers reveal recharge elevations,
groundwater flow paths and travel times on Mount Shasta, California. Water,
volume 10, 22 pages, doi:10.3390/w10020097.
Petrella, E. and F. Celico, 2012, Mixing of water in a carbonate aquifer, southern Italy,
analysed through stable isotope investigations. International Journal of Speleology,
volume 42, 9 pages, doi:10.5038/1827-806X.42.1.4.
Plummer, L.N., J.R. Eggleston, D.C. Andreasen, J.P. Raffensperger, A.G. Hunt, and G.C.
Casile, 2012, Old groundwater in parts of the upper Patapsco aquifer, Atlantic
Coastal Plain, Maryland, USA: evidence from radiocarbon, chlorine-36 and helium-4.
Hydrogeology Journal, volume 20, 26 pages, doi:10.1007/s10040-012-0871-1.
Prasetio, R., Y. Daud, J. Hutabarat, and H. Hendarmawan, 2021, Fluid geochemistry and
isotope compositions to delineate physical processes in Wayang Windu geothermal
89
reservoir, Indonesia. Geosciences Journal, volume 25, 17 pages,

doi:10.1007/s12303-020-0039-2.
Pu, T., D. Qin, S. Kang, H. Niu, Y. He, and S. Wang, 2017, Water isotopes and hydrograph
separation in different glacial catchments in the southeast margin of the Tibetan
Plateau. Hydrological Processes, volume 31, 17 pages, doi:10.1002/hyp.11293.
Qu, S., M. Zhou, P. Shi, H. Liu, W. Bao, and X. Chen, 2014, Differences in oxygen-18 and
deuterium content of throughfall and rainfall during different flood events in a small
headwater watershed. Isotopes in Environmental and Health Studies, volume 50,
10 pages, doi:10.1080/10256016.2014.845565.
Rai, S.P., D. Singh, N. Jacob, Y.S. Rawat, M. Arora, and B. Kumar, 2019, Identifying
contribution of snowmelt and glacier melt to the Bhagirathi River (Upper Ganga)
near snout of the Gangotri Glacier using environmental isotopes. Catena, volume 173,
13 pages, doi:10.1016/j.catena.2018.10.031.
Lord Rayleigh, 1902, On the distillation of binary mixtures. The London, Edinburgh and
Dublin Philosophical Magazine and Journal of Science, volume 4, 17 pages,
doi:10.1080/14786440209462876.
Reeburgh, W.S., 1994, Global water reservoirs, fluxes and turnover times,
http://www.ess.uci.edu/~reeburgh/fig8.html.
Remenda, V., J.A. Cherry, and T.W. Edwards, 1994, Isotopic composition of old ground
water from Lake Agassiz: Implications for Late Pleistocene Climate. Science,
volume 266, number 5193, 4 pages, doi:10.1126/science.266.5193.1975.
Ronen, D., M. Magaritz, H. Gvirtzman, and W. Garner, 1987, Microscale chemical
heterogeneity in groundwater. Journal of Hydrology, volume 92, 6 pages,
doi:10.1016/0022-1694(87)90094-1.
Rozanski, K., C. Sonntag, and K.O. Munnich, 1982, Factors controlling stable isotope
composition of European precipitation. Tellus, volume 34, 9 pages,
doi:10.3402/tellusa.v34i2.10796.
Rozanski, K., L. Araguas-Araguas, and R. Gonfiantini, 1993, Isotopic patterns in modern
global precipitation, editors, Swart P.K., K.C. Lohmann, J. McKenzie, and S. Savin,
Climate Change in Continental Isotopic Records, American Geophysical Union,
Geophysical Monograph 78, chapter 1, 36 pages.
Salati, E., A. Dall'Olio, E. Matsui, and J.R. Gat, 1979, Recycling of water in the Amazon
Basin: an isotopic study. Water Resources Research, volume 15, 9 pages,
doi:10.1029/WR015i005p01250.
Sanchez-Murillo, R., C. Birkel, K. Welsch, G. Esquivel-Hernandez, J. Corrales-Salazar, J.
Boll, E. Brooks, O. Roupsard, O. Saenz-Rosales, I. Katchan, R. Arce-Mesen, C.
Soulsby, and L.J. Araguas-Araguas, 2016, Key drivers controlling stable isotope
variations in daily precipitation of Costa Rica: Caribbean Sea versus Eastern Pacific
Ocean moisture sources. Quaternary Science Reviews, volume 131, 237 pages,
doi:10.1016/j.quascirev.2015.08.028.
90
Schimmelmann, A. and M.J. DeNiro, 1993, Preparation of organic and water hydrogen for
stable isotope analysis: Effects due to reaction vessels and zinc reagent. Analytical
Chemistry, volume 65, 4 pages, doi:10.1021/ac00054a024.
Sharp, Z., 2007, Principles of Stable Isotope Geochemistry. Pearson Prentice Hall,
https://www.abebooks.com/9780130091390/Principles-Stable-Isotope-Geochemistry-S
harp-0130091391/plp.
Sinclair, K.E., S.J. Marshall, and T.A. Moran, 2011, A Lagrangian approach to modelling
stable isotopes in precipitation over mountainous terrain. Hydrological Processes,
volume 25, 11 pages, doi:10.1002/hyp.7973.
Sklash, M.G. and R.N. Farvolden, 1979, The role of groundwater in storm runoff. Journal
of Hydrology, volume 43, 21 pages, doi:10.1016/0022-1694(79)90164-1.
Smith, G.I., I. Friedman, G. Veronda, and C.A. Johnson, 2002, Stable isotope compositions
of waters in the Great Basin, United States: 3. Comparison of groundwaters with
modern precipitation. Journal of Geophysical Research, volume 107, issue D19,
16 pages, doi:10.1029/2001JD000567.
Socki, R.A., H.R. Karlsson, and E.K. Gibson Jr., Extraction technique for the determination
of oxygen-18 in water using preevacuated glass vials. Analytical Chemistry,
volume 64, 3 pages, doi:10.1021/ac00031a026.
Sonntag, C.E., E.P. Klitzoch, K.O. Löhnert, K.O. Münnich, O. Junghaus, K. Thorweihe, K.
Weistroffer, and F.M. Swailen, 1979, Paleoclimatic information from deuterium and
oxygen 18 in C14 date north Saharan groundwaters: groundwater formation in the
past. In: Isotope Hydrology 1978, International Atomic Energy Agency, Vienna,
Austria, 13 pages.
Steur, P.M., H.A. Scheeren, D.D. Nelson, J.B. McManus, and H.A. Meijer, 2020,
Simultaneous measurement of 13C, 18O and 17O of atmospheric CO2 - Performance
assessment of a dual-laser absorption spectrometer. Atmospheric Measurement
Techniques, 26 pages, doi:10.5194/amt-14-4279-2021.
Stewart, M.K., 1985, Isotopic geochemistry at Wairakei - Hydrological implications.
Institute of Nuclear Sciences, INS-R -345, Lower Hutt, New Zealand, 19 pages.
Sultan, M., N.C. Sturchio, H. Gheith, Y. Abdel Hady, and M. El Anbeawy, 2000, Chemical
and isotopic constraints on the origin of Wadi El-Tarfa groundwater, Eastern Desert,
Egypt. Groundwater, volume 38, 9 pages, doi:10.1111/j.1745-6584.2000.tb02710.x.
Sumner, G., 1988, Precipitation: Process and Analysis. John Wiley and Sons, Inc.
Singapore,
https://www.abebooks.com/9780471905349/Precipitation-Process-Analysis-Sumner-G
raham-0471905348/plp.
Sunguro, S., K. Froehlich, B.Th. Verhagen, and W. Wagner, 2000, Geohydrology and
isotope hydrology of the Lomagundi dolomite, Zimbabwe. In Sililo et al (editors):
Groundwater: Past Achievements and Future Challenges, Balkema, Rotterdam.
91
Tekleab, S., J. Wenninger, and S. Uhlenbrook, 2014, Characterisation of stable isotopes to

identify residence times and runoff components in two meso-scale catchments in the
Abay/Upper Blue Nile basin, Ethiopia. Hydrology and Earth System Sciences,
volume 18, 17 pages, doi:10.5194/hess-18-2415-2014.
Uemura, R., N. Yonezawa, K. Yoshimura, R. Asami, H. Kadena, K. Yamada, and N.
Yoshida, 2012, Factors controlling isotopic composition of precipitation on Okinawa
Island, Japan: Implications for palaeoclimate reconstruction in the East Asian
Monsoon region. Journal of Hydrology, volume 475, 9 pages,
doi:10.1016/j.jhydrol.2012.10.014.
Uhlenbrook, S., M. Frey, C. Leibundgut, and P. Maloszewski, 2002, Hydrograph
separations in a mesoscale mountainous basin at event and seasonal timescales.
Water Resources Research, volume 38, 4 pages, doi:10.1029/2001WR000938.
Urey, H.C., 1947, The thermodynamic properties of isotopic substances. Journal of the
Chemical Society of London, 20 pages, doi:10.1039/JR9470000562.
Van Gelden, R., A. Baier, H.L. Subert, S. Kowol, L. Balk, and J.A. Barth, 2014, Pleistocene
paleo-groundwater as a pristine fresh water resource in southern Germany - evidence
from stable and radiogenic isotopes. Science of the Total Environment, volume 496,
9 pages, doi:10.1016/j.scitotenv.2014.07.011.
Vargas, A.I., B. Schaffer, L. Yuhong, and L. da Silviera Lobo Sternberg, 2017, Testing plant
use of mobile vs immobile soil water sources using stable isotope experiments. New
Phytologist, volume 215, 13 pages, doi:10.1111/nph.14616.
Vogel, J.C. and H. van Urk, 1975, Isotopic composition of groundwater in semi-arid
regions of southern Africa. Journal of Hydrology, volume 25, issues 1-2, 4 pages,
doi:10.1016/0022-1694(75)90036-0.
Wang, C., 2007, Plasma-cavity ringdown spectroscopy (P-CRDS) for elemental and
isotopic analysis. Journal of Analytical Atomic Spectroscopy, volume 22, 17 pages,
doi:10.1039/B701223C.
Wang, J., N. Lu, and B. Fu, 2019, Inter-comparison of stable isotope mixing models for
determining plant water source partitioning. Science of the Total Environment,
volume 666, 9 pages, doi:10.1016/j.scitotenv.2019.02.262.
West, A.G., G.R. Goldsmith, P.D. Brooks, and T.E. Dawson, 2010, Discrepancies between
isotope ratio infrared spectroscopy and isotope ratio mass spectrometry for the stable
isotope analysis of plant and soil waters. Rapid Communications in Mass
Spectrometry, volume 24, 7 pages, doi:10.1002/rcm.4597.
Xie, Y., P.G. Cook, and C.T. Simmons, 2016, Solute transport processes in
flow-event-driven stream-aquifer interaction. Journal of Hydrology, volume 538,
11 pages, doi:10.1016/j.jhydrol.2016.04.031.
Yakir, D., M.J. de Niro, and J.E. Ephrath, 1990, Effects of water stress on oxygen, hydrogen
and carbon isotope ratios in two species of cotton plants. Plant, Cell and
Environment, volume 13, 7 pages, doi:10.1111/j.1365-3040.1990.tb01985.x.
92
Yang, X., T. Yao, W. Yang, W. Yu, and D. Qu, 2011, Co-existence of temperature and
amount effects on precipitation 18O in the Asian monsoon region. Geophysical
Research Letters, volume 38, issue 21, article L21809, 6 pages,
doi:10.1029/2011GL049353.
Yidana, S.M., O.F. Fynn, D. Adomako, L.P. Chegbeleh, and P.M. Nude, 2016, Estimation
of evapotranspiration losses in the vadose zone using stable isotopes and chloride
mass balance method. Environmental Earth Sciences, volume 75, 208 pages,
doi:10.1007/s12665-015-4982-6.
Yin, L., G. Hou, Y. Dou, Z. Tao and Y. Li, 2011, Hydrogeochemical and isotopic study of
groundwater in the Habor Lake Basin of the Ordos Plateau, NW China.
Environmental Earth Sciences, volume 64, 10 pages, doi:10.1007/s12665-009-0383-z.
Yonge, C.J., L. Goldenberg, and H.R. Krouse, 1989, An isotopic study of water bodies
along a traverse of southwestern Canada. Journal of Hydrology, volume 1989,
11 pages, doi:10.1016/0022-1694(89)90075-9.
Yurtsever, Y. and J.R. Gat, 1981, Atmospheric Waters, Gat J.R. and R. Gonfiantini, Stable
Isotope Hydrology. International Atomic Energy Agency, Vienna, Technical
Report 20, 40 pages,
https://inis.iaea.org/collection/NCLCollectionStore/_Public/13/677/13677657.pdf.
93
13 Exercise Solutions
Solution Exercise 1 - Meteoric Water Lines
Step-by-step calculation is shown in the tables and calculations that follow.
Calculation of a meteoric water line

( Oi − ̅̅̅̅̅
18
Precipitation 18
 O D 𝟏𝟖 𝐎)
( Oi- ̅̅̅̅̅
2
(Di − 𝐃
18 2 ̅̅̅̅)
Location Date 𝟏𝟖 𝐎)
mm ‰ ‰ (Di − 𝐃
̅̅̅̅)
UCT Jan 2010 9 0.75 9.7 8.619 342.87 54.36
UCT Feb 2010 12 -0.54 -5.8 2.709 9.10 4.96
UCT Mar 2010 7 -0.83 0.4 1.838 84.95 12.50
UCT Apr 2010 44 -2.25 -12.1 0.004 10.78 0.21
UCT May 2010 277 -2.78 -8.4 0.353 0.17 -0.25
UCT Jun 2010 222 -3.31 -17.0 1.264 66.97 9.20
UCT Jul 2010 166 -3.69 -15.0 2.263 38.23 9.30
UCT Aug 2010 121 -1.74 -4.9 0.199 15.34 1.75
UCT Sep 2010 100 -2.24 -3.3 0.003 30.43 -0.30
UCT Oct 2010 102 -2.15 -3.3 0.001 30.43 0.20
UCT Nov 2010 73 -2.39 -6.2 0.042 6.85 -0.53
UCT Dec 2010 17 -5.06 -39.9 8.261 966.17 89.34
18 18
Total Mean Mean SS- O SS-D SP- O-D
1150 -2.19 -8.82 25.56 1602.30 180.74
SS = sum of squares; SP = sum of products.
For the best fit line by RMA (reduced major axis) regression:
𝑆𝑆𝐷 1602.30
the gradient = √ =√ = 7.92
𝑆𝑆18 𝑂 25.56
𝐷 − (7.92 ̅̅̅̅̅̅
the intercept = ̅̅̅̅ 18 𝑂) = −8.8 − (7.92 (−2.19)) = 8.49
therefore, the LMWLUCT-2010 is D = 7.9218O + 8.49
Pearson’s r, the sample correlation coefficient

𝑆𝑃18 𝑂𝐷 180.74
= 𝑟18 𝑂𝐷 = = = 0.893
√𝑆𝑆18 𝑂√𝑆𝑆𝐷 √25.56√1602.30
94
Calculation of a precipitation-weighted meteoric water line
𝟏𝟖 𝐎) 
2
𝟏𝟖 𝐎)
rf  ( Oi - ̅̅̅̅̅
̅̅̅̅)2
rf  (18Oi - ̅̅̅̅̅
rf  (Di - 𝐃
̅̅̅̅)
rf   O
(Di - 𝐃
rf  D
18
18
Rain-
18
Precipitation  O D fraction
Location Date mm ‰ ‰ (rf)
UCT Jan 2010 9 0.75 9.7 0.0078 0.0059 0.0759 0.067 2.68 0.43
UCT Feb 2010 12 -0.54 -5.8 0.0104 -0.0056 -0.0605 0.028 0.09 0.05
UCT Mar 2010 7 -0.83 0.4 0.0061 -0.0051 0.0024 0.011 0.52 0.08
UCT Apr 2010 44 -2.25 -12.1 0.0383 -0.0861 -0.4630 0.000 0.41 0.01
UCT May 2010 277 -2.78 -8.4 0.2409 -0.6696 -2.0233 0.085 0.04 -0.06
UCT Jun 2010 222 -3.31 -17.0 0.1930 -0.6390 -3.2817 0.244 12.93 1.78
UCT Jul 2010 166 -3.69 -15.0 0.1443 -0.5326 -2.1652 0.327 5.52 1.34
UCT Aug 2010 121 -1.74 -4.9 0.1052 -0.1831 -0.5156 0.021 1.61 0.18
UCT Sep 2010 100 -2.24 -3.3 0.0870 -0.1948 -0.2870 0.000 2.65 -0.03
UCT Oct 2010 102 -2.15 -3.3 0.0887 -0.1907 -0.2927 0.000 2.70 0.02
UCT Nov 2010 73 -2.39 -6.2 0.0635 -0.1517 -0.3936 0.003 0.43 -0.03
UCT Dec 2010 17 -5.06 -39.9 0.0148 -0.0748 -0.5898 0.122 14.28 1.32
1
SS  rf -
SP  rf -
SS  rf -
O -D
D
O
Total Mean Mean Sum Sum Sum
8
1150 -2.19 -8.82 1.00 -2.73 -9.99 0.909 43.98 5.08
SS = sum of squares; SP = sum of products; rf = rain fraction (i.e., weighted by precipitation).
For the best fit line by RMA (reduced major axis) regression:
𝑆𝑆 𝑟𝑓𝐷 43.98
the gradient = √ = √0.909 = 6.95
𝑆𝑆 𝑟𝑓18 𝑂
the intercept = ∑ 𝑟𝑓. 𝐷 − 6.95 ∑ 𝑟𝑓. 18 𝑂 = −9.99 − (6.95 (−2.73)) = 8.96
the precipitation weighted LMWLUCT-2010 is D = 6.9518O + 8.96
Pearson’s r, the sample correlation coefficient
𝑆𝑃𝑟𝑓18 𝑂𝐷 5.08
𝑟𝑟𝑓18 𝑂𝐷 = = = 0.805
√0.909√43.87
√𝑆𝑆𝑟𝑓 𝑂√𝑆𝑆𝑟𝑓𝐷
18
Return to Exercise 1
95
Solution Exercise 2 - Recharge Altitude

1. The mean stable isotope composition for the Kirstenbosch spring is:
∑𝑛𝑖 𝛿𝑖
𝛿̅ =
𝑛
where:
i = value for a given sample

n = total number of samples
the means are -11.45‰ and -3.11‰ for 2H and d18O respectively.
2. The weighted means for the precipitation stations are:

𝑛
𝛿̅ = ∑ 𝛿𝑖 × 𝑟𝑓𝑖
𝑖
where:
 = denotes a year
i =  value for a given month
rfi = fraction of the total precipitation
(over 3 years) for that month, calculated as:
𝑟𝑖
𝑟𝑓𝑖 =
𝑟𝑡
where:
ri = rain amount for a month
rt = total rain for the period of the study
For months with missing data, the total precipitation amount must exclude those months.
Final answers of -9.15‰ and -2.90‰ for 2H and 18O at University of Cape Town
and -14.34‰ and -3.73‰ similarly for the Table Mountain Cableway.
3. The altitude effect per isotope pair is calculated as:

Δ𝛿 ̅
𝛿𝑇𝑀𝐶 ̅
− 𝛿𝑈𝐶𝑇
=
𝑚 ℎ 𝑇𝑀𝐶 − ℎ𝑈𝐶𝑇
where:
Δ𝛿⁄𝑚 = change in  values per meter of altitude
x = mean stable isotope composition for a station
hx = elevation of that station
giving answers of -0.0055‰/m for 2H and -0.00089‰/m for 18O. These answers are more
reasonably displayed for 2H and 18O as -0.55‰/100 m and -0.089‰/100 m.
96
4. The recharge elevation for Kirstenbosch can be calculated by addition of the isotopic
altitude effect to the starting elevation at UCT, as follows:
̅ − 𝛿𝑈𝐶𝑇
(𝛿𝐾𝑏 ̅ )
ℎ𝐾𝑏 = ℎ𝑈𝐶𝑇 +
(Δ𝛿⁄𝑚)
or by using the same formula with the hTMC and TMC values, which will produce a
subtraction, both of which should give the same answer. Using 2H gives 591 m and 18O
gives 368 m, the average of which is 479 m. This means the recharge area is 319 m above
the spring at 160 m. This is obviously an average elevation as recharge occurs over an area
and not at one point.
97
Solution Exercise 3 - Hydrograph Separation

Using the mass balance equation for a 2-component system:
𝑄𝑝 × 𝛿𝑝 + 𝑄𝑔 × 𝛿𝑔 = 𝑄𝑠 × 𝛿𝑠
where:
Q = flow
 = stable isotope composition
p = precipitation
g = groundwater
s = surface water
and recognizing:
𝑄𝑝 + 𝑄𝑔 = 𝑄𝑠
and if fractions are used instead of actual flow:

𝑄𝑠 = 1 , then 𝑄𝑝 + 𝑄𝑔 = 1
∴ 𝑄𝑝 = 1 − 𝑄𝑔
∴ (1 − 𝑄𝑔 )𝛿𝑝 + 𝑄𝑔 𝛿𝑔 = 𝛿𝑠
∴ 𝛿𝑝 − 𝑄𝑔 𝛿𝑝 + 𝑄𝑔 𝛿𝑔 = 𝛿𝑠
∴ 𝑄𝑔 (𝛿𝑔 − 𝛿𝑝 ) = 𝛿𝑠 − 𝛿𝑝
(𝛿𝑠 −𝛿𝑝 )
∴ 𝑄𝑔 = ,
(𝛿𝑔 −𝛿𝑝 )
which, for 2H gives 0.82 for groundwater, which translates to the surface water flow being
82 percent groundwater and 18 percent precipitation derived. For 18O, the answers are 63
percent groundwater and 37 percent precipitation.
The measurement errors (given by Hu et al., 2020), especially for precipitation, are
so large, it is possible precipitation has the same isotope composition as groundwater, in
which case hydrograph separation is not possible. Similarly, the river water could have the
same isotope composition as the groundwater, again negating the ability to perform a
hydrograph separation. More measurements would lower the error and improve the
estimate.
98
14 Notations
‰ parts per thousand (permil)
1,2,3,... subscript for input streams
a,b,c fitting constants (K2, K, dimensionless, respectively)
α equilibrium fractionation factor for the reaction at a given temperature

(dimensionless)
X,Y fractionation factor for transition from X to Y (dimensionless)
c y-intercept (dimensions of y-axis values)
d deuterium excess (D-excess)
 deviation of stable isotope composition from a standard
ε enrichment factor
f fraction (1 to 0) of water that remains (dimensionless)
f subscript for output stream
g subscript indicating groundwater
m gradient, i.e., slope of the line (dimensions of y-axis values over dimensions of
x-axis values)
mRMW gradient, i.e., slope of the Reduced Major Axis Regression (dimensions of y-axis
values over dimensions of x-axis values)
p subscript indicating precipitation
Q flow rate (MT-1)
R isotope ratio such as 2H/1H (dimensionless)
rfinal isotope composition of the pool of water with fraction f remaining

(dimensionless)
rinitial original isotope composition of the pool (dimensionless)
99
s subscript indicating resultant total stream flow
SPxy sum of product of deviations from the mean of x and deviations from mean of y
(dimensions are product of x and y dimensions)
SSx sum of the squared deviations from the mean of x (squared dimensions of the x
variable)
SSy sum of the squared deviations from the mean of y (squared dimensions of the y
variable)
T temperature (K)
X subscript indicating reactant compound or state
x, y regression variables
Y subscript indicating product compound or state
100
15 About the Author

Dr. Roger Diamond studied geology at the
University of Cape Town, receiving his Master
of Science in 1997, after which he worked in
Australia as a gold exploration geologist and
then a hydrogeologist for the Western
Australian state government. After returning
to South Africa in 2002, he worked for the
Western Cape government and as a consultant
for Environmental Science Associates in
various aspects of environmental
management, before returning to University of
Cape Town in 2010 to study for a Doctor of
Philosophy degree, which was awarded in 2014. He then moved to Pretoria where he
lectures hydrogeology and geochemistry at the University of Pretoria, focusing his research
on water quality and hydrochemistry, including work on nitrates, uranium, radon, and of
course stable isotopes! He has been an active member of the Groundwater Division of the
Geological Society of South Africa for many years, and takes part in several environmental
and conservation actions, especially removal of invasive alien plants with the Mountain
Club of South Africa. When not looking at rocks, he is out climbing up them.
101
Please consider signing up for the GW-Project mailing list to stay informed about new book
releases, events and ways to participate in the GW-Project. When you sign up for our email
list it helps us build a global groundwater community. Sign up.
102

Libro Stable-Isotope-Hydrology

Uploaded by

Copyright:

Available Formats

Libro Stable-Isotope-Hydrology

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Libro Stable-Isotope-Hydrology

Uploaded by

Copyright:

Available Formats

Title Author(s)

Stable Isotope Hydrology

Copyright © 2022 Roger E. Diamond (The Author)

Published by the Groundwater Project, Guelph, Ontario, Canada, 2022.

Domain Editors: Robert Kalin, Eileen Poeter, and John Cherry

7.7 DETECTING SURFACE WATER IN GROUNDWATER ...................................................................................... 64

The Groundwater Project Foreword

The mission of the GW-Project is promoting groundwater learning. This is

John Cherry, The Groundwater Project Leader

❖ John Cherry, G360 Institute for Groundwater Research, University of Guelph,

Figure 2 - Stable isotope hydrology interfaces with many

2 Isotopes and Isotopologues

Isotopes can be stable, radioactive (decay to another element) or radiogenic

Hydrogen is somewhat more complex, as 3H (tritium) is radioactive, with a half-life

10-17  210-4  410-4 = 810-25  10-24

Table 1 - List of stable isotopologues for water and their

3 Measurement and Standards

Stable isotope ratios are reported as deviation from an international standard.

R = isotope ratio such as 2H/1H (dimensionless)

3.1 Mass Spectrometry

3.2 Laser Cavity Spectroscopy

Figure 6 - Schematic diagram of a laser cavity ringdown instrument.

As with mass spectrometry, laboratory standards are inserted every ten or so

4.1 Kinetic Fractionation

4.2 Equilibrium Fractionation

Figure 7 - Temperature dependence of fractionation. Graph a) shows the

4.3 Fractionation and Enrichment Factors

R = isotope ratio, such as 2H/1H (dimensionless)

X = subscript indicating reactant compound or state

𝜀 = 1000(𝛼 − 1) ≈ 1000𝑙𝑛(𝛼) (6)

Importantly the variation in equilibrium fractionation factors is systematic, because

4.4 Rayleigh Distillation

𝑟𝑓𝑖𝑛𝑎𝑙 = 𝑟𝑖𝑛𝑖𝑡𝑖𝑎𝑙 × 𝑓 (𝛼−1) (7)

If an imaginary block of moisture laden atmosphere (a vapor body) has water

For a fractionating process, such as a chemical reaction or diffusion in a porous

5 Meteoric Water Lines

2H = 8 18O + 10 (8)

The spread of data along the GMWL is influenced by several meteorological

5.2 Local Meteoric Water Lines

5.3 Calculation of Meteoric Water Lines

5.3.1 Least Squares Regression

m = gradient, i.e., slope of the line (dimensions of y-axis values over

𝑆𝑃𝑥𝑦 = ∑(𝑥𝑖 − 𝑥̅ )(𝑦𝑖 − 𝑦̅) (10)

5.3.2 Reduced Major Axis Regression (RMA)

5.4 Weighted Regression Line Equations

𝑆𝑆𝑥 = ∑(𝑟𝑎𝑖𝑛𝑖 )(𝑥𝑖 − 𝑥̅ )2

𝑆𝑆𝑦 = ∑(𝑟𝑎𝑖𝑛𝑖 )(𝑦 − 𝑦̅)2 (12)

𝑆𝑃𝑥𝑦 = ∑(𝑟𝑎𝑖𝑛𝑖 )(𝑥𝑖 − 𝑥̅ )(𝑦𝑖 − 𝑦̅)

6 Stable Isotope Hydrology

6.1 Environmental Factors Affecting Water Isotopes (Isotope Effects)

6.1.2 The Latitude Effect

6.1.3 The Continental Effect

6.1.4 The Altitude Effect

6.1.5 The Amount Effect

6.2 The Deuterium Excess

D-excess = 2H - 818O (13)

2H and 18O = values for the water sample

6.3 Event Variation

6.4 Mass Balance