Get reliable science

Attention has been directed towards both the impacts of future climate change on environmental systems and dunefield activity in the past, but there has been relatively little consideration of potential dune mobility in a future and possibly warmer world. This paper considers the use and limitations of four Global Circulation Models (GCMs) (Hadcm3, Hadcm2, CSIRO-mk2b and CGCM1), in combination with simple dune mobility indices to predict the activity of the Kalahari dunefield. It is clear that uncertainties surround GCM resolution and accuracy, mobility index robustness for the calculation of intra-annual dune mobility and data collection for mobility index calibration. Macro-scale studies that look at large areas of the world over long time scales are well suited to GCM and mobility index use, but dune mobility can be variable within a dunefield, and it is the extreme sand transporting events, occurring at high temporal resolutions, that are the most important for short term studies. To investigate intra-annual changes in dune mobility over a specific dunefield techniques such as downscaling, weather generators and probability curve fitting can help provide climate predictions for smaller areas over shorter time frames. However, these methods introduce uncertainty of their own, and they often rely on the accuracy of original GCM predictions or the climate parameter relationships observed at present. Analysis of intra-annual changes also requires mobility indices that can model monthly mobility patterns well, although existing indices have only been used for calculating annual dune mobility potential. When they are used for intra-annual predictions, the lack of lag response between precipitation decreases and the assumed vegetation dieback leads to an exaggerated amplitude pattern of dune mobility throughout the year. Calibration of dune mobility indices to dune mobility observed on the ground is therefore important but is hampered by a lack of observed measurements for individual months. Solutions are available to overcome some of the outlined problems, but they can provide their own set of uncertainties, which combine to further reduce the confidence given to future dune mobility predictions.