Oxygen isotopes by gyvwpsjkko


									                              Oxygen isotopes

                                      Ricky Taylor
                                     December 2003

Oxygen occurs mainly in the form of the 16O isotope. However, there are also very
small quantities of the stable 18O isotope occurring naturally. The 18O isotope has two
more neutrons than 16O, and as a result is more about 6% heavier.

Although quantities of 18O relative to 16O are small, they can be measured. As the
different isotopes have different masses, it is possible to separate them by passing a
stream of the ionised sample (usually in a gas form) through a powerful magnetic
field. The heavier ions have greater momentum than the lighter ones, and hence are
deflected to a lesser extent than the lighter ones. This splits the stream of ions into its
heavy and its light components – each of which can be measured using a Faraday
collector. This is the principle of a mass spectrometer - the instrument used to
measure the proportion of different isotopes within a sample.

The abundance of the 18O isotope is expressed as a percentage of the 18O isotope
relative to 16O (Indicated by “R” in the equation below). For oxygen, on average, the
natural abundance 18O relative to 16O is 0.204% (Clark and Fritz, 1997). However,
when comparing different samples, the differences in the abundances of the isotopes
are very small. When doing measurements at this level, machine calibration errors
become critical. To alleviate this problem the isotope being measured is measured
relative to an internationally accepted standard. The quantity of the isotope is then
expressed as the ratio of the isotope in the sample relative to that in the standard. This
relative proportion (δ) is multiplied by 1000 to reduce the number of decimal places
being used.

The units are "permil" or "parts per thousand" which is written as “‰”.

The international standard used for the measurements of 18O and 2H is the VSMOW
(which stands for the Vienna Standard Mean Ocean Water). It is against this that all
samples are compared.

If δ is positive the sample is "enriched" and if δ is negative the sample is "depleted"
relative to the standard.


As oxygen occurs commonly in the form of H2O, and as our main interest is water,
most of the use of 18O is in this form. Where this is the case the 18O isotope is
frequently used in conjunction with 2H – the heavy isotope of hydrogen which is also
known as Deuterium. The ratios of 18O to 16O and of 2H to 1H "fingerprint" water.
But, for a “fingerprint” to form, it is necessary for the ratio of the isotopes to be
altered by natural environmental processes. Under specific conditions, there is a
separation of the heavier and lighter isotopes. This is the process called

It is the difference in neutron numbers in the isotopes that enables fractionation to
occur. The number of neutrons affects the strength of the bonds between molecules.
The molecules with heavier isotopes have stronger bonds and more energy is required
to separate these molecules than is the case with molecules containing the lighter
isotopes. For water molecules this is important whenever there is a change in state
between ice, water and vapour. It takes more energy to change heavy isotopes from ice
to water or from water to vapour than it does for the lighter isotopes. Conversely, it
needs less energy for heavy molecules to change from vapour to water or water to ice.
This process fractionates the isotopically heavy and light water molecules

As this process is energy dependent, temperature affects the rate of fractionation of
water into its heavier and lighter forms. In a closed system, an equilibrium is set up,
with water molecules moving in both directions between two states. However if the
system is not closed, the fractionation is not in equilibrium and as it will be biased in
one direction. A process of distillation (Rayleigh distillation) occurs. In the natural
environment there is air circulation, which is continually shifting the vapour
evaporating from the surface of water – and so this is not a closed system.
 O/16O ratios can also be used in CH2O, CO2, Sulphates, NO3-, carbonates, silicates
and –OH minerals (Clark & Fritz 1997 )

There are many applications for the study of water. Some of the main ones are

        Separation of hydrograph
        Identification of the providence of groundwater
        Palaeoclimate studies.

For all applications it is necessary to understand how the 18O isotope behaves, and also
what the dynamics are of the hydrological system being studied – and at what stages
fractionation is taking place.

The Global Meteoric Water Line (GMWL)
In 1961, Craig investigated the global distribution of 18O and 2H in freshwaters and
discovered that there is a relationship between the abundance of 18O and of 2H (Clarke
& Fritz, 1997). Fractionation occurs more at cold temperatures than at warmer
temperatures. Craig expressed this as a generalised graph showing the relationship
between δ 2H and δ18O (both relative to VSMOW). (See Figure 1)

Figure 1. The relationship between 18O and 2H gives the GMWL (figure from Clark & Fritz, 1997)

The general trend is that the waters of cold regions tend to be isotopically depleted
relative to the waters of the warm regions. When δ 2H is plotted as a function of δ 18O
the relationship is linear (at the global scale). This relationship is known as the
Global Meteoric Water Line (GMWL).

The GMWL provides a reference for the comparison of local differences in water and
thus for interpreting the provenance of groundwaters.

 O behaviour – and where in the hydrocycle fractionation occurs
Fundamental to the use of isotopes, is the need to understand the processes in which
fractionation occurs, and then at what stages these occur within the hydrological cycle.

As water moves through the hydrological cycle, so it is subjected to various
conditions in which fractionation of the isotopes takes place. This starts with
seawater – which has δ 18O and δ 2H values, which are close to those of VSMOW
(Clark & Fritz, 1997). Water evaporates from the surface of the sea. This is a process
that depletes the evaporating water – making it isotopically lighter than the water left
behind. The following meteoritic processes then affect this water:

A)      Global processes:

        (i)      Temperature change effects

Dansgard (1964) plotted the relationship between temperature (taken as the mean
annual air temperature for a particular station) and the δ 18O value. From this he
derived the graph and equation shown in figure 2. This clearly shows the very strong
relationship between temperature and δ 18O. This is a response to greater fractionation
taking place at low temperatures than at high temperatures. However, there are
deviations from this and they are from other effects.

Figure 2: Dansgaard’s temperature-δ 18O relationship for precipitation.(from Clarke & Fritz 1997)

As the δ 18O values reflect the temperature – the δ 18O value can be used as a proxy to
show palaeo-temperatures. These can be used to show changes that occurred between
glacial and interglacial conditions, and also smaller changes – such as the “Little Ice
Age”. For instance, the δ 18O value related to temperature has been used as a proxy to
generate mean sea temperatures for the Quaternary Period (figure 3).

Figure 3: The record of the 18O composition of seawater during the Quaternary Period. The
measurement of 18O is from carbonate deposit (from Imbrie et al., 1984, in Clarke and Fritz, 1997).

        (ii)     Distance from sea – “rainout” and the “continental” effect.

After the water has evaporated from the sea, and is in the atmosphere, it moves inland
and condenses, to fall as rain. This process results in rainfall that is isotopically
enriched relative to the water vapour. However, as the vapour moves further and

further inland – there is a progressive depletion of heavy isotopes – resulting in the
degree of heavy isotope depletion increasing with distance from the sea. Thus,
although the rainfall is enriched relative to the water vapour from which it is created,
the rainfall is also progressively depleted with the distance from the sea. Thus the
heavy isotopes in the atmosphere are said to become "rained out", giving rise to the
"continental" effect. (Figure 4)

Figure 4: The "continental effect" where the rain becomes isotopically depleted with distance from the
sea. (Figure from Clark & Fritz, 1997). The insert graph shows the drop of mean annual air
temperature (MAAT) relative to distance over the same route.

        (iii)    Latitude effects

Temperature has a very strong effect on fractionation. The colder the conditions; the
greater the amount of fractionation that occurs. This is shown very clearly when
plotting δ 18O values on a global scale. The higher the latitude the more depleted the δ
  O value. This is shown in figure 5.

Figure 5: This illustrates the iso-lines showing that precipitation at higher latitudes has more negative
δ 18O values than at lower latitudes. (after Clark & Fritz, 1997).

As the water vapour to liquid water process is modified by temperature, variations in δ
  O and δ 2H occur wherever there is a temperature gradient. Thus there are
correlations with latitude, altitude, and season – the colder the temperature the greater
the depletion of the heavier isotopes in the precipitation.

So we end up with local rainfall containing specific values for the ratio of 18O and 2H,
which differ from the GMWL. Not only can this be site specific, but also there can be
seasonal differences and even differences between and within specific rainfall events
at that site.

B)       Local processes:

         (i)      Altitude effects

Temperature decreased with altitude. This is again a temperature driven effect.
Figure 6 shows the effect of altitude.

Figure 6: This graph shows the altitude effect on local rainfall in the Italian Alps. The higher the
elevation, the colder the conditions and hence the more negative the δ 18O values. The two lines
indicate seasonal differences. (From Clark & Fritz, 1977).

         (ii)     Seasonal effects

The seasonal change from warm summers to cold winters is reflected in the increasing
depletion of δ 18O. (See figure 7)

Figure 7: This shows the seasonal variation in δ 18O at different stations in North America. Those with
the greatest seasonal temperature differences, show the greatest seasonal variation in δ 18O. The data
are from: Puerto Rico (18.4° N), North Carolina (35.3° N), Ohio (40.4° N), Manitoba (50.6° N) and
North West Territories (74.7° N). (From Clark & Fritz, 1997).

Determination of the provenance of the water and hydrograph separation

In aquifers where there is local recharge, and where groundwater is of recent origin,
we expect much of the water to have the 18O/16O and 2H/1H isotopic compositions
similar to the annual mean of that of the rainfall for that area. Once the rain has
fallen, the isotopic composition may be altered by extreme evaporation processes, and
by chemical interactions associated with weathering in the ground, otherwise the
isotopic composition should only be changed by the mixing of water. It is, however,
not altered by transpiration losses, as transpiration is not a fractionating process.

Within the groundwater, deviations from the precipitation composition may be
important – indicating a mixing of water sources and possibly an extended residence
time (Fetter, 1994). The measuring of isotopes can shed light on the recharge and
movement of water in such systems. This is especially the case in localities where
there is a large seasonal temperature difference resulting in marked seasonal changes
in δ18O. This effect, combined with precipitation rates, groundwater discharge
measurements and the δ18O for the discharging water, has been used to calculate
recharge periods for groundwater within a catchment. Haldorsen et al. (1997) report
on a study in Norway where the δ18O of the groundwater has a value indicating that
most of the recharge is from the winter precipitation.

The difference in δ18O, in rainfall compared with that of groundwater, can also be
used to provide estimates of the amount of water loss through evaporation relative to
that lost through transpiration (especially under irrigated conditions).

Another application is that of hydrograph separation in stream flow – to determine the
contribution of groundwater by separating the base flow component from the storm
flow. This is based on the difference in isotopic composition of groundwater when
compared to that of a given storm. This can be used when the hydrograph is
composed only of the above two components (Figure 8). But, if there are additional

components – such as separate inputs of deep groundwater and of shallow
groundwater in addition to the storm flow, then it is possible to use a chemical
concentration in addition to the isotope composition to provide additional
information. This can be done only if the chemical (such as SiO2) is conservative and
its concentrations are not linked to the isotope being measured. In this case three
mass-balance equations are used - and solved iteratively to provide estimates for the
relative contribution of the three component sources of the water (Figure 9). This is
the basis of the EMMA (End Member Mixing Analysis) system described by
McDonnell (2001). The relative composition of the water can be shown graphically
using ternary diagrams.

Figure 8: The use of oxygen isotopes to separate groundwater from surface runoff water in a stream
hydrograph. (Figure from Clark & Fritz, 1997)

Figure 9: This figure uses the same data as in Figure 3, but there has been separation of the
groundwater and soil water components by using SiO2 concentrations to give a third dimension.
(Figure from Clark & Fritz, 1997)

This method has also has been used to indicate the contribution of fossil water in
places. (Figure 10)

Figure 10: A ternary diagram showing how the oxygen isotope signature combined with the presence
of chlorine (from salinity) has been used to investigate the extent of mixing of palaeowater from the
Canadian Shield. (Figure from Clark & Fritz, 1997)


Global temperature changes are reflected in changes in δ18O in water that has been
through a fractionation process – i.e. polar ice and old groundwater deposits. The
isotopes of oxygen are used for palaeoclimate research in ice cores obtained by
drilling into ice deposits near the poles. The δ18O values reflect the temperature
regime that existed when the ice was deposited. See figure 11.

Figure 11: δ18O values for ice cores. The vertical scale is the depth in the ice core. The sections
record temperatures for a period dating back about 125 000 years.(From Clark and Fritz, 1997).

The shifts in δ18O with time in polar ice cores provide both long-term and seasonal
markers – with the deposition of more depleted waters in cold winter conditions and
less depleted in warm summer conditions.

δ18O in carbonates

δ18O in carbonate deposits (such as stalagmites) has also been used as a tool to show
palaeotemperatures. This does rely on having an independent and accurate method for
dating the deposit. Establishment of the chronsequence has been done using 14C or
U/Th. The independent dating is necessary as the rapid exchangesthat occur within
the various carbonate phases (largely due to pH differences) make it unsuitable as a
dating tool. However, once the date of the sample has been established, δ18O is used,
as in the ice, to derive the temperature regime under which the sample was deposited.

Some speculative uses of oxygen isotopes for understanding
hydrodynamics within the Greater St Lucia Wetland Park:
Oxygen isotopes provide a “fingerprint” for water. Used in conjunction with other
measuring techniques, oxygen isotopes could be usefully applied to various problems
within the GSLWP. The following are some speculations about possible applications:

   1       Frontal/cyclonic rain patterns at St Lucia

Rain at St Lucia comes from two main directions – the frontal rain from the south-
west and the cyclonic rain from the north-east. The former has its origins in the
southern oceans and passes over 1500 km of land before reaching St Lucia. I would
expect the frontal rain to be δ18O depleted because of its cold origin, and then to be
further depleted by “rainout” as it passes over land. Conversely, the cyclonic rain has
a warm tropical origin, and passes over much less land before reaching St Lucia. I
would expect the rain to be relatively much less δ18O depleted.

If this is the case, then it should be possible to detect the difference between rain
events of the two forms. As the cyclonic rainfall is very seasonal – and often is
related to a few discrete cyclonic events in one season – then this could possibly be
used to identify the source of various pulses of water within the groundwater. This
could be used to obtain estimates of rate of passage of the water through sand aquifers
in the GSLWP.

   2       Sibaya – hydrological modelling

Lake Sibaya is a water body that has no surface outlet. All water gained is either from
direct rainfall or from inflows from the extensive groundwater aquifer surrounding the
lake. Losses are through evaporation and groundwater seepage from the lake to the

For a hydrological model of the lake it is possible to work out these inputs and
outputs. The groundwater gains and losses can be modelled and the rainfall and
evaporation measured. To provide some verification of the model, it would be
necessary to have an independent variable that can be measured. This could possibly
be δ18O. The values of δ18O of incoming groundwater, rain and lake water can be
measured, and then compared to the expected values of δ18O in the lake. I would
expect the δ18O value in the lake to be higher than that of the groundwater – as a result
of evaporation losses that have occurred since the lake separated from the sea. This
occurred about 5000 years ago (Miller, 2001). At some stage an equilibrium would
have been reached, and since then the δ18O value would have been relatively constant
– showing only small fluctuations related to wet and dry periods.

   3       Mkhuze Swamps – water dynamics

The Mkhuze Swamps is hydrologically complex. Water inputs are in the form of
rainfall, surface runoff in the Mkhuze River, bed flow in the sand of the Mkhuze
riverbed, and groundwater from the eastern margins of the swamp. Water losses are
mainly through transpiration (a non-fractionating process), and outflow into St Lucia.
Evaporation is possibly negligible as there are few open water surfaces, and the high
transpiration losses are likely to saturate the atmosphere immediately above the
swamp. We have little knowledge about the water volume in the swamp, and are
unable to measure outflows (mainly because flows are dispersed at the mouth, the lake
water level changes frequently, and wind action “pumps” the water in and out). Our
management needs are to know the rate of passage of water through the swamp and
the amount of water flowing out of the swamp into St Lucia. The latter is needed as
an input for the Lake hydrological model. Can we use δ18O differences enhance our
understanding of the swamp hydrology?

We can possibly use similar measurement of δ18O, as with Sibaya, to provide insights.
The groundwater outputs can be modeled, and its δ18O measured. The river inflows
can be measured, as well as its δ18O values. Samples within the swamp could possibly
be used to show patterns of flows and mixing of water from the eastern aquifers and
rainfall with the Mkhuze river waters. For this it may be possible to assume that there
is no δ18O change due to evaporation (this can be checked experimentally). By
balancing flows and water levels it may be possible to gain an appreciation of the
sensitivity of the system to its various input & output components, and then install
devices to measure these components

In addition, it may be possible to follow plugs of Mkhuze water (derived from
cyclonic as opposed to frontal rainfall), as they flow through the swamp – to provide
information of flow rates through the swamp.

Sample collection and storage.
 The collection of samples is straightforward – as long as the sample does not
evaporate it can be stored indefinitely.

Normally 100ml of water are collected. No special precautions are required when
collecting the sample, and no filtering is necessary. It should be stored in a tightly-
sealed screw cap bottle – which can be plastic or glass. Waterproof tape should be
bound around the cap to ensure no evaporation takes place.

Analysis costs:
At QUADRU, the Quaternary Dating Research Unit of the CSIR in South Africa,
the measurement of 18O in water costs € 27 per sample, and the measurement of
both 13C and 18O in carbonates costs € 45 per sample.

Clark, I and Fritz, P. (1997). Environmental isotopes in Hydrogeology. Lewis
        Publishers. New York.
Craig, H. (1961). Isotopic variations in meteoric waters. Science, 133: 1702-1703
Dansgard, W. (1964). Stable isotopes in precipitation. Tellus, 16:436-468.
Fetter, C. W. (1994). Applied Hydrogeology. Prentice-Hall. New Jersey.
Haldorsen, S., Riise, G. and Swensen, B. (1997). Environmental isotopes as tracers
        in catchments. In: Saether, O. and De Caritat, P. (eds.). Geochemical
        Processes, Weathering and Groundwater Recharge in Catchments. A. A.
        Balkema, Rotterdam.
Imbrie, J. J, Hays, J. D., Martinson, J. D., MacIntyre, A., Mix, A. C., Morley, J.
        J., Pisas, N. G., Prell, W. L. and Shackleton, N. J. (1984). The orbital theory
        of climate: support from a revised chronology of the marine δ18O record. In:
        Berger et al. (Eds). Milankovitch and climate, D Reidel, Dordrecht Publishing
        Co., Netherlands (not seen)
McDonnell, J. (2001). The use of isotope techniques in catchment hydrology
        studies. http://cof.orst.edu/cof/fe/watershd/shortcourseEMMA.
Miller, W. R. (2001). The bathymetry, sedimentology and seismic stratigraphy of
        Lake Sibaya – northern KwaZulu-Natal. Bulletin 131, Council for
        Geoscience, South Africa.
RHT 12/12/2003


To top