Identifying global dust source areas using high-resolution land surface form

Dust originates from specific, landscape-scale (approximate to 1-10 km) features; however, dust erodibility is commonly parameterized in global climate models using grid-cell-scale (approximate to 1-5 degrees) relationships. We examine surface characteristics from the high-resolution (90 m at equator) global Shuttle Radar Topography Mission digital elevation model and aerosol optical thickness (AOT) measurements from the Multiangle Imaging Spectroradiometer (MISR) instrument to find landscape-scale characteristics common to dust producing regions. Regions with climatologically high aerosol optical thickness are associated with extremely low-slope landscapes at approximate to 5 km scale that also have high-elevation variance, likely due to aeolian features. We refer to these landscape characteristics as levelness'' and residual landscape roughness,'' and we extrapolate these relationships globally to identify geomorphologically based dust erodibility functions on the basis of levelness and residual landscape roughness criteria. We test these with the Dust Entrainment and Deposition model embedded in the Model for Atmospheric Transport and Chemistry global atmospheric tracer transport model, driven by reanalysis meteorology. The modeled global dust spatial distributions calculated with these erodibility parameters agree well with observations of MISR AOT and deposition and in the modeling framework used here represent an improvement over existing parameterizations in predicting the spatial patterns of dust sources in the Sahara and the relative amount of emission of sources in the Dust Belt to those in the Southern Hemisphere; however, agreement with observations is less good in high-relief areas such as the Asian dust source regions, possibly because of the coarse resolution of the meteorological fields used to drive the model.