CHALLENGES IN PREDICTING THE RESPONSE OF DUNE FIELDS TO ENVIRONMENTAL CHANGE

This talk discusses progress, problems and prospects in predicting dune field response to environmental change. Two important pressures are guiding future research agendas: (1) climate change, and (2) the growing demand for science that directly serves the public’s interest. Both pressures should drive renewed efforts to develop predictions of future changes in global dune systems. Reactivation of now vegetation-stabilized dune fields is one of the most significant changes that would benefit from prediction, and also provides a strong case example. The socio-economic effects of reactivation could be devastating. However, presently the risk is poorly quantified. How can we improve predictions of reactivation? Each of the three common research approaches (Quaternary reconstruction, process measurement and modeling) shows strengths and weaknesses. Quaternary studies have advanced considerably with new geochronology tools (e.g., optical dating), but study timescales are massive and the understanding of past landscape-climate connections may not inform us about a future that may drift beyond past dynamics. Process measurements have also advanced, but the gap between what can be measured in the field and what is needed to parameterize and reduce the uncertainty of model predictions continues to grow wider. Models have made a big impact in terms of visualizing dune field dynamics, including the role of vegetation, but the parameterizations and virtual landscapes that emerge remain largely untested.

Present literature demonstrates that predictions of future changes in real dune fields are very rare; it is difficult to produce viable work. Continuing with the status quo is unlikely to inform society on a reasonable timescale about how these landscapes might change in the future. However, through greater awareness and coordination between research groups and some humble introspection, a new course could be charted. It may be possible to develop predictions that are continually tested and refined to rein in uncertainty. Modeling will be required to formulate predictions, but without real data from Quaternary reconstructions and process measurements the value of those predictions will be low. The grand challenge is to develop models and measurements that are designed to tightly integrate and work in tandem.