Publication date: September 2016
Source:Journal of Archaeological Science, Volume 73
Author(s): Eelco J. Rohling
Quantitative estimates of climate variability are increasingly important in interpretations of archaeological turnovers in arid regions. Variations in lake levels or lake-water oxygen isotope ratios (δ18O) are often used to infer droughts or humid periods, along with speleothem δ18O, pollen, and windblown dust records. Key examples are the centennial-scale Holocene events associated with the end of the Bronze Age (∼1200 BCE), the end of the Copper Age (∼4000 BCE), and the onset of Neolithic expansion (∼6200 BCE). Whether explicitly stated or only implied, causality between archaeological turnovers and inferred droughts is often ascribed to a disturbance to food resources, which means a disturbance to the agricultural potential of the study region. In the present study, a simple framework of equations is presented for evaluation of this causality. It quantitatively reveals significant complications. In one example, substantially improved crop-growing potential is found to coincide with dropping lake levels, which reflect significant net drought. The complications mainly arise from: (1) control of annually averaged climate conditions on lake changes versus control of seasonal conditions on the yield potential of fields; and (2) changes in the ratios between the overall catchment area of a lake or field, and the surface area of the lake or field itself. The results demonstrate that lake records per se do not satisfactorily reflect agricultural potential, but also that this gap may be bridged with targeted information collection about the regional setting. In particular, improved results may be obtained from detailed assessments of change in the catchment ratios of the lake(s) and field(s) that are being studied (e.g., using digital elevation models), along with expert opinions on field irrigation potential. The scenarios presented here then allow initial field-based assessments and hypothesis formulation to prompt more sophisticated modelling.