Elucidating Hidden and Enduring Weaknesses in Dust Emission Modeling
Adrian Chappell, Nicholas P. Webb, Mark Hennen, Charles S. Zender, Philippe Ciais, Kerstin Schepanski, Brandon L. Edwards, Nancy P. Ziegler, Yves Balkanski, Daniel Tong, John Leys, Stephan Heidenreich, R. G. Hynes, David Fuchs, Zhenzhong Zeng, Matthew Baddock, Jeffrey A. Lee, Tarek Kandakji
Abstract
Abstract Large‐scale classical dust cycle models, developed more than two decades ago, assume for simplicity that the Earth's land surface is devoid of vegetation, reduce dust emission estimates using a vegetation cover complement, and calibrate estimates to observed atmospheric dust optical depth (DOD). Consequently, these models are expected to be valid for use with dust‐climate projections in Earth System Models. We reveal little spatial relation between DOD frequency and satellite observed dust emission from point sources (DPS) and a difference of up to 2 orders of magnitude. We compared DPS data to an exemplar traditional dust emission model (TEM) and the albedo‐based dust emission model (AEM) which represents aerodynamic roughness over space and time. Both models overestimated dust emission probability but showed strong spatial relations to DPS, suitable for calibration. Relative to the AEM calibrated to the DPS, the TEM overestimated large dust emission over vast vegetated areas and produced considerable false change in dust emission. It is difficult to avoid the conclusion that calibrating dust cycle models to DOD has hidden for more than two decades, these TEM modeling weaknesses. The AEM overcomes these weaknesses without using masks or vegetation cover data. Considerable potential therefore exists for ESMs driven by prognostic albedo, to reveal new insights of aerosol effects on, and responses to, contemporary and environmental change projections.