Large eddy simulation (LES) of the Martian convective boundary layer (CBL) with a Mars-adapted version of the Weather Research and Forecasting model (MarsWRF) is used to examine aerosol dust radiative-dynamical feedback upon turbulent mixing. The LES is validated against spacecraft observations and prior modeling. To study dust redistribution by coherent dynamical structures within the CBL, two radiatively-active dust distribution scenarios are used—one in which the dust distribution remains fixed and another in which dust is freely transported by CBL motions. In the fixed dust scenario, increasing atmospheric dust loading shades the surface from sunlight and weakens convection. However, a competing effect emerges in the free dust scenario, resulting from the lateral concentration of dust in updrafts. The resulting enhancement of dust radiative heating in upwelling plumes both generates horizontal thermal contrasts in the CBL and also increases buoyancy production, jointly enhancing CBL convection. We define a dust inhomogeneity index (DII) to quantify how much dust is concentrated in upwelling plumes. If the DII is large enough, the destabilizing effect of lateral heating contrasts can exceed the stabilizing effect of surface shading such that the CBL depth increases with increasing dust optical depth. Thus, under certain combinations of total dust optical depth and the lateral inhomogeneity of dust, a positive feedback may exist among dust optical depth, the vigor and depth of CBL mixing, and—to the extent that dust lifting is controlled by the depth and vigor of CBL mixing—the further lifting of dust from the surface.

Changes in aerosol optical depth, both positive and negative, are observed across the globe during the 21rst Century. However, attribution of these changes to specific sources is largely uncertain as there are multiple contributing natural and anthropogenic sources that produce aerosols either directly or through secondary chemical reactions. Here we show that satellite-based changes in small-mode AOD between 2002 and 2019 observed in data from MISR can primarily be explained by changes, either directly or indirectly, in combustion emissions. We quantify combustion emissions using MOPITT total column CO observations and the adjoint of the GEOS-Chem global chemistry and transport model. The a priori fire emissions are taken from the Global Fire Emission Data base with small fires (GFED4s) but with fixed a priori for non-fire emissions. Aerosol precursor and direct emissions are updated by re-scaling them with the monthly ratio of the CO posterior to prior emissions. The correlation between modeled and observed AOD improves from a mean of 0.15 to 0.81 for the four industrial regions considered and from 0.52 to 0.75 for the four wildfire-dominant regions considered. Using these updated emissions in the GEOS-Chem global chemistry transport model, our results indicate that surface PM2.5 have declined across many regions of the globe during the 21rst century. For example, PM2.5 over China has declined by ~30% with smaller fractional declines in E. USA and Europe (from fossil emissions) and in S. America (from fires). These results highlight the importance of forest management and cleaner combustion sources in improving air-quality.