Search for a command to run...
The most recent high-resolution global map of dust emission sources is provided by Ginoux et al. (2012), which utilizes an aerosol loading approach based on time series of MODIS Aerosol Optical Depth (AOD). However, advancements in remote sensing technology and analytical techniques have created a growing need for more accurate and up-to-date maps of global dust sources to enhance the understanding and management of this phenomenon. In this study, we first calculated the global mean Sentinel-5P Absorbing Aerosol Index (AAI) for the period 2018-2024. Regions with AAI values greater than 0.25 were identified as potential dust sources through histogram analysis validated by ground truth data. Next, areas without dust emission potential were excluded from the mean AAI map using a multi-stage masking process that considers land surface characteristics such as soil depth, permanent water bodies, and built-up areas. Validation results demonstrate strong performance, with a Precision of 84.7%, Recall of 80.7%, and F1-score of 82.6%, confirming the reliability of the global dust source map produced. The findings indicate that about 5% of the world's land area acts as a dust emission source, mainly located in North Africa (67%) and Asia (30%). Land use/land cover analysis reveals that global dust sources comprise deserts, vegetative, and hydrological categories, accounting for 65%, 26%, and 9%, respectively. Among these, sandy areas, rangelands, and intermittent water bodies exhibit the largest extent on a global scale, respectively. Natural and human factors contribute 65% and 35%, respectively, to the formation of global dust sources. The frequency of dust events from desert sources has experienced an increasing trend worldwide, but in the case of non-desert sources, it has decreased in some regions, such as the Middle East. This study focused on identifying major dust emission sources based on relatively high aerosol loads over time. Our results provide a new global dust atlas that can serve as a practical foundation for climate modeling and for formulating disaster risk reduction and management plans.