Large Variability of Springtime African Dust in Recent Decades: A Consistent Characterization from Multiple Remote Sensing Observations

Dust affects key components of the climate system, including energy, water, and carbon cycles. The emission and transport of dust is strongly modulated by surface and meteorological conditions that change in a changing climate. Model simulations of dust variability depend strongly on how the dust lifecycle is represented in models, which currently has large uncertainties. In this study, we acquired an observational understanding of the interannual variability of African dust over the last two decades by analyzing remote sensing measurements from multiple satellites (MODIS, CALIOP, MISR, and IASI) and AERONET. We separated dust from non-dust aerosol by using remote sensing measurements of particle size (e.g., fine-mode fraction and optical depth at the thermal infrared wavelength) and shape (e.g., depolarization ratio and non-spherical fraction). These measurements of dust show a consistent interannual variability of dust optical depth in North Africa and tropical North Atlantic Ocean, which is mainly driven by that in spring. The interannual variability has become more pronounced since 2009 when Sahel rainfall varies substantially from year to year. A composite analysis between high-dust years and low-dust years shows that the springtime dust interannual variability in source regions is mainly controlled by surface wind variability, which is further associated with the northward progression of West African Monsoon or the anomaly of springtime rainfall over equatorial West Africa. Extended analysis of the dust-monsoon association with historical dust records from AVHRR remote sensing and Barbados in-situ sampling (up to 36-years) suggests that the variability of West African Monsoon progression could explain 30-40% of dust variability over tropical North Atlantic Ocean. In NASA GEOS model, more than half of dust emission variability in spring are explained by the West African Monsoon variability.