Sand dunes as giant rain gauges in the desert?

Reconstructing past precipitation fluctuations in dryland regions is challenging because of the nature of evidence left behind. Pore moisture within sand dunes provides a novel archive of palaeomoisture availability that has yet to be fully utilised. A recent review paper, led by Dr Abi Stone, outlines the status and future of this archive. Written with the late Mike Edmunds, the paper celebrates Mike’s great contribution to this area. Here, Abi reflects on their collaboration.


Rain gauge in the Kalahari. Picture by Abi Stone.

Understanding palaeoclimatic conditions in dryland regions is a key goal for climate science, particularly the responses of these environmentally sensitive regions at resolutions of decades to millennia. This allows the geosciences community to understand the nature of the response of drylands to factors that force the climate system, which enables better predictions of future dryland environmental change and variability. Arid and semi-arid regions make up around a third of the earth’s surface, are home to up to 2.1 billion people, and are areas particularly sensitive to future changes in precipitation and temperature.

However, reconstructing past rainfall trends and fluctuations in drylands is not an easy task. The negative moisture balance means that many of the pieces of evidence for precipitation (climatic proxies) used in other environments are not well preserved in drylands. For example, pollen does not survive in dry sediments and degrades on exposure to air. In addition, there are few records from calcium-carbonate deposits in caves (speleothems) found in more humid areas, such as those in China, that have notably produced impressive time-series of palaeoprecipitation and monsoon dynamics (in excess of 50,000 years).

This is because drylands often lack sufficient levels of precipitation for the speleothems to grow continuously. The distinctively dryland climate proxies of sand dunes and lake shorelines, such as those I cut my teeth on as a researcher during my DPhil, have their own shortfalls as palaeomoisture indicators. Their accumulation is mediated by a number of factors, of which changing moisture availability is just one.

Mike Edmunds was one of the pioneering researchers in the 1970s who discovered that the solute concentrations within the pore moisture of sediments above the water table (the unsaturated zone, USZ) were inversely correlated to rainfall amount. This work, based in Cyprus was followed by research in Senegal in the 1990s. By 2002, Mike had set out the potential of the unsaturated zone as a climate archive with Scott Tyler from the University of Nevada, Reno.

It was on a coach in 2013 driving around the chalk aquifer sites of Dorset, helping Mike to co-ordinate the introductory field-trip for Oxford’s Water Science, Policy and Management MSc class, that we got chatting about the current state of research into palaeoclimate using the USZ. After talking about our data from the Kalahari, I asked Mike if he thought it would be useful to write a review of the latest progress. He agreed in his usual enthusiastic manner and we started to plan the scope of the paper there and then.

We chose the journal Earth Science Reviews in order to try to bring this technique to the attention of the geoscience communities interested in palaeoclimatic reconstruction, but who may not spend so much time with their heads in the hydrogeological literature. Throughout the process of reviewing the literature and drafting the paper, I looked forward to my meetings with Mike, most often over coffee in my office of the time, at St John’s College. This looked out over St Giles and Mike was always a fan of watching the Oxford world go by. It was a contrast to the tranquillity of his own study which overlooked a wonderful garden at his home in Appleton, which he shared with Kathy.


Mike’s schematic, started in propelling pencil and meticulously completed in ink pen.

The main message of the paper is that yes, sand dunes do act as giant rain gauges in the desert. The length of that record depends on how big your sand dune is and how quickly the rainfall is trickling through it. Almost 40 studies worldwide have now used this approach to find out something about past climate. The schematic figure gives a summary explanation of how an USZ hydrostratigraphy accumulates. Mike revised an earlier conceptual diagram, sketching it using his trademark propelling pencil. Hydrostratigraphy is a fancy way of saying that the signal in the water in different layers of sand with increasing depth below the surface stores information about what climate was like when that water was near the surface in the past (and before it started soaking down through the sand)



Schematic with summary explanation.

I was in Namibia in May 2016 when the proofs of the final paper arrived for me to check. This was to collect more of the bright red-orange dune sand from the Kalahari above the Stampriet Artesian Basin, and investigate more about the nitrate Mike and I had uncovered a few years before. I had to explain the delay in checking the proofs. The reply had some tragicomedy attach to it, with the Global Journals Production Company (for Elsevier) asking if they could send the proofs to my co-author, rather than wait for my return. If only! I can only guess that they also found it hard to believe that he was gone.



Sampling sand dunes in the Kalahari desert. Picture by Abi Stone.

In the paper we draw global examples together to highlight that hydrostratigraphies are useful for three timescales (duration and resolution):
(i) centennial-length, decadal resolution: records of past rainfall and moisture availability and useful to identify how land-use change modifies rates of water infiltration, such as in irrigated parts of the Thar Desert, India.
(ii) millennial-length, decadal resolution: most notably in the Badain Jaran Desert in China, where the record spans 2,000 years, demonstrating the same climatic shifts recorded in tree-rings and even written historical documents from former Chinese dynasties.
(iii) multi-millennial length, low-resolution records, such as in Nevada in the United States, where the shift from wetter to arid conditions occurred before the start of the Holocene (~11,700 years ago).

You can read more in Stone, A. E. C., Edmunds, W. M. (2016) Unsaturated zone hydrostratigraphies: a novel archive of past climates in dryland continental regions. Earth Science Reviews 157, 121-144.

Oxford University and the British Geological Survey will host the first W. Mike Edmunds Memorial Lecture at Christ Church on 3 November 2016. For further details visit the online booking page.