STOCHASTIC MODELING OF DESERT DUST EMISSION: BRIDGING THE SCALE GAP

Despite the improvement of wind erosion and dust emission models for vegetated surfaces in recent years, little is known about how spatial variability in the landscape influences their location and magnitude. As a result, most current models are highly scale-dependent. A new stochastic spatially-explicit wind erosion and dust flux model (SWEMO) allows estimation of wind erosion and dust flux across a landscape by incorporating coarse- and fine-scale spatial variability of pertinent landscape parameters. This approach provides a powerful basis for trying to understand how vegetation and soil interact in the landscape to create the most important dust sources. By explicitly incorporating fine-scale variability a Monte Carlo framework, SWEMO can accommodate the inherently nonlinear nature of wind erosion and dust flux. In particular, results from SWEMO indicate that the variance of cover and plant size parameters exerts primary control on estimates of wind erosion and dust flux. The inclusion of subgridcell variability highlights the importance of small but intense emission surfaces on landscape-scale measurements of dust and nutrient emission and indicates that accurate scale-independent model estimates of dust emission require some knowledge of landscape variability in source areas.