Isotope geochemistry is the very definition of technically sweet, it's the metaphorical Babel Fish of environmental research. That is - it's a set of tools so mindbogglingly useful they almost feel as if they were given to us with intent. In the days before Ernest Rutherford, environmental systems had to be taken on their own terms and certain questions (such as the age of rocks, and where things came from) were unanswerable. Then with isotopic measurements came a new toolbox - a way of distinguishing between chemically identical elements, but those posessed with different weights. On its own this doesn't sound like much, but the sweetest aspect is that individual isotopes seem perfectly tuned to individual technical problems.
It's a surprisingly big discipline, and it's one that sits on the periphery of physics, chemistry and geology. But as mass spectrometers (the basic tool of isotope science) have become more powerful, cheaper, and common, this little taught discipline has moved into mainstream discussions. The big reveal scene in An Inconvenient Truth uses an oxygen isotope graph, and carbon dating gets thrown around as a topic for discussion whenever archaeology is mentioned. Typically, we use isotope geochemistry for two purposes - to determine the age of a process, and to pinpoint reservoir differences in materials that appear superficially identical. Recently I've been taking a look at isotope hydrology, the question of how to use this information to approach something as dynamic as water systems...
Water is composed of hydrogen and oxygen, and each has several common isotopes. Hydrogen possesses the stable isotopes H1 and H2 (deuterium), and the unstable isotope H3(tritium). Whereas, for oxygen the isotopes of interest are O16,O17 and O18, all stable. To measure the abundance of these elements, the main tool is a mass spectrometer, and it's performed relative to a standard. For all of these isotopes we use the same standard – Standard Mean Ocean Water - and list these measurements as (for example) δ18O and δ2D. In this notation, positive values indicate an enrichment of the heavy isotope relative to seawater, and negative indicating depletion.
Firstly, I want to discuss how isotopes and their relative behaviour can be used as a process signature. This is usually through the mechanism of isotopic fractionation, and in the case of water this takes place during evaporation and precipitation. In any situation where there is a significant relative mass difference between two isotopes the isotope with faster moving small masses tends to preferentially break its chemical bonds when excited. In water, this means during evaporation the process leaves the heavier H2 and O18 in the liquid phase. Concentrating heavier isotopes in the oceans. Fractionation occurs again during precipitation, concentrating the heavier isotopes in the initial liquid phase. The first rains are the heaviest. Put these processes together, and rainwaters tend to be depleted in O18 and H2 relative to seawater, and the further from the equator (where most evaporation occurs) the greater the depletion. A useful tool for telling us about precipitation and weather systems. But even further, the initial fractionation at evaporation is temperature dependent, with even minor temperature increases in measurably increased fractionation of heavy isotopes into the water vapour. The makes δ18O an invaluable indicator of palaeo-climate, sensitive even to subtleties of seasonal change. At any given site it has been demonstrated that there is measurable relative decrease in the δ18O of winter precipitation relative to summer.
The implications of this model for different reservoirs of water are vast, polar precipitation tends to be significantly depleted in heavy isotopes, whereas evaporating bodies such as the Dead Sea are enriched in them. Further, in deep lakes such as Tanganyika and Baikal there is vertical variation in δ18O with the heavier isotopes enriching with depth, with implications for measuring the degree of vertical mixing. This is such a powerful tool it has become essential in every aspect of climate science. These sorts of isotopic tools have proved sufficiently useful that modelling using relative proportions of each isotope has been applied to almost every major hydrological process.
After the stable isotopes, which pinpoint climate and water bodies, the radioactive ones allow us to determine ages. Dating water is tricky, since these isotopes usually give us a clock that is ticking from the moment it becomes a closed system. It usually works on the principle of assuming the initial proportions of parent and daughter isotopes, and measuring the final proportions empirically allowing us to deduce the age from the disparity. This gets harder when the system isn't inherently closed. In the case of Carbon dating of organic remains, we have an organism that stopped breathing at a certain point. Unfortunately water never stops being dynamic. Yet, certain isotopes seem ideally suited to solving these problems, and with a few assumptions we can start dating water bodies such as confined aquifers, or glacial lakes.
The unstable isotope of hydrogen H3 (tritium) incorporates directly into water molecules making it a sensible first choice. Like most radiogenic isotopes in Earth surface processes, tritium is normally produced by the interaction of cosmic rays with atmospheric gases at an approximately constant rate, in this case its parent isotope is nitrogen. This continuous background of isotope production, followed by its gradual decay (tritium decays into He3 over a half life of 12.33 years) means its abundance in any Earth surface body can be used to determine its age, as is the case with Carbon 14.
Unfortunately, there are problems with the use of H3 in dating water bodies: its low mass means isotope fractionation is likely, its short half life limits its use to periods up to approximately 50 years (a rule of thumb is that an isotope is useful for up to 4 times its halflife), and most importantly - we, the human race, ruined it. Between 1950 and 1963 (the year of the test ban treaty) atmospheric nuclear testing massively elevated tritium levels throughout the atmosphere. While the immediate aftermath of this period made tritium a very useful presence/absence indicator of processes such as oceanic mixing and the infiltration of a young component into older groundwaters, the decreasing abundance since means its present application as a dating tool is limited.
Carbon dating using C14 (also produced from cosmic ray interactions with nitrogen) is an alternative. Carbon has distinct advantages: its half-life is 5730 years (ideal for human timescales) and it is readily absorbed in meteoric water as dissolved CO2. In dating a body as active as water a number of assumptions are important: that the whole body is of roughly one age with limited mixing with other water bodies, that the carbon has been chemically conservative since its dissolution. While the first of these problems can be alleviated by accompanied use of other isotopes (eg H3 or δ18O) the second is more intractable. For example, when CO2 dissolves in water it produces carbonic acid (H2CO3), and this is not helpful when it encounters limestone host rocks - massive reservoirs of ancient carbon. The carbonic acid attacks the limestone, leading to a chemical and isotopic exchange between the two carbon reservoirs, diluting it and thus increasing the derived age of the water. While such problems can be modelled, this limits C14 as a tool for groundwaters and this is not the only such problem.
A more promising isotope is chlorine 36. Produced from argon interacting with cosmic rays this has a half-life of 301,000 years, making it ideal for groundwater dates up to 1 million years. The benefit of chlorine is it behaves extremely conservatively when dissolved in waters, and thus dates derived from its decay are robust - the reason the seas are saturated with sodium chloride is that chlorine likes to stay dissolved in water. This makes it an ideal tool for dating geological and ancient groundwaters, but less productive in oceanic waters where the high salinity dilutes its signature. On its own, this still has problems, but in combination these three three tools can prove formidable in dating groundwaters.
So, that is an introduction... The key isotopes that can be applied to solving problems in hydrology. But a passing examination of the table of nuclides shows that there are far, far more isotopes than this. Over the last half century geology and chemistry have been developing ways to use each of these isotopes to model different systems. Strontium isotopes have proved essential in determining the origins of sediments and the structure of the Earth's mantle. Uranium isotopes have developed into a tool for dating almost everything, at any timescale. It's unbelievably useful to have a chemical tool that only changes under very specific circumstances. When the human race gets itself, properly, to the outer planets we are going to need a lot of mass spectrometers to work out what the hell is going on in their geological processes.