Improving dust emission characterization in dust models using dynamic high-resolution geomorphic erodibility map

Dust is known to affect the earth radiation budget, biogeochemical cycle, precipitation, human health and visibility. Despite the increased research effort, dust emission modeling remains challenging because dust emission is affected by complex geomorphological processes. Existing dust models overestimate dust emission and rely on tuning and a static erodibility factor in order to make simulated results comparable to remote sensing and ground-based observations. In most of current models, dust emission is expressed in terms of threshold friction speed, which ultimately depends mainly upon the percentage clay content and soil moisture. Unfortunately, due to the unavailability of accurate and high resolution input data of the clay content and soil moisture, estimated threshold friction speed commonly does not represent the variability in field condition. In this work, we attempt to improve dust emission characterization by developing a high resolution geomorphic map of the Middle East and North Africa (MENA), which is responsible for more than 50% of global dust emission. We develop this geomorphic map by visually examining high resolution satellite images obtained from Google Earth Pro and ESRI base map. Albeit subjective, our technique is more reliable compared to automatic image classification technique because we incorporate knowledge of geological/geographical setting in identifying dust sources. We hypothesize that the erodibility is unique for different geomorphic landforms and that it can be quantified by the correlation between observed wind speed and satellite retrieved aerosol optical depth (AOD). We classify the study area into several key geomorphological categories with respect to their dust emission potential. Then we quantify their dust emission potential using the correlation between observed wind speed and satellite retrieved AOD. The dynamic, high-resolution geomorphic erodibility map thus prepared will help to reduce the uncertainty in current dust models associated with poor characterization of dust sources. The baseline dust scheme used in this study is the Dust Entrainment and Deposition (DEAD) model, which is also a component of the community land model (CLM). Proposed improvements in the dust emission representation will help to better understand the accurate effect of dust on climate processes.