Identifying ancient water availability through phytolith analysis: An experimental approach

Publication date: September 2016
Source:Journal of Archaeological Science, Volume 73
Author(s): Emma Jenkins, Khalil Jamjoum, Sameeh Nuimat, Richard Stafford, Stephen Nortcliff, Steven Mithen
Water management was critical to the development of complex societies but such systems are often difficult, if not impossible, to recognise in the archaeological record, particularly in prehistoric communities when water management began. This is because early irrigation systems are likely to have been ephemeral and as such would no longer be visible in the archaeological record. We conducted a three year crop growing experiment in Jordan to test the hypothesis that phytoliths (opaline silica bodies formed in plants) can be used to detect the level of past water availability and hence be used as a source of information for inferring past water management. Over a three year period we grew native land races of six-row barley (Hordeum vulgare) and durum wheat (Triticum durum) at three crop growing stations in Jordan with the crops being subjected to different irrigation regimes. Seeds were sown in the autumn and the crops harvested in the spring. The plants were then exported to the University of Reading for phytolith processing. Our results show that while there were unknown factors that influenced phytolith production between years, at the higher levels, the ratio of ‘fixed’ form phytoliths (those formed as a result of genetically determined silicon uptake) to ‘sensitive’ form phytoliths (those whose silicon uptake is environmentally controlled) can be used to assess past water availability. Our study is the first large scale experimental project to test this method and take into account multiple variables that can affect phytolith production such as soil composition and chemistry, location, climate and evapotranspiration rates. Results from the cereals grown at two of the crop growing stations, Deir ‘Alla and Ramtha, which received between 100 mm and 250 mm rainfall per annum, demonstrate that if the ratio of fixed to sensitive phytolith forms is >1, the level of past water availability can be predicted with 80% confidence. Results from the crops grown at the other growing station, Kherbet as-Samra, which received less than 100 mm of rainfall per year show that if the ratio of fixed to sensitive forms is >0.5, the level of past water availability can be predicted with 99% confidence. This demonstrates that phytolith analysis can be used as a method to identify past water availability.