Controls on aerosol wet deposition from satellite-based (re-)analysis products

Abstract:

Aerosol wet deposition is the key aerosol loss mechanism globally, yet is not well-understood relative to aerosol sources and transformations. The difficulty in generating appropriate observational data sets is one important barrier to the study of aerosol wet removal. In this study, we combine two independent products based on satellite measurements. Aerosol optical depth (AOD) is obtained from the ECMWF Monitoring Atmospheric Composition and Climate (MACC) project, which is a re-analysis product that assimilates MODIS-retrieved aerosol optical depth. Rainfall is obtained from the Tropical Rainfall Measurement Mission (TRMM) Multi-satellite Precipitation Analysis version 7 (TMPA-7). The latter product is available only from 50°N to 50°S, which sets our region of study. The data used is from 2011-12, is averaged to 6-hr intervals and has a horizontal resolution of 0.25°x0.25°. Our approach involves constructing a Lagrangian advection scheme that predicts aerosol AOD at the next time step (i.e. 6 hr in the future) based on current time step AOD and winds, and neglecting all aerosol sources and sinks. Predicted AOD is then compared with MACC reanalysis AOD conditioned on Lagrangian parcels that experienced rainfall during that interval, with AOD decreases attributed to wet deposition. Aerosol wet deposition is often parameterized in models as a function of rainfall rate using a power law. We evaluate the validity of such a power law relationship, and, when valid, compute the power law exponent globally, and by region (including continental and maritime locations) to reveal seasonal and geographic variability. Assuming precipitation is modulated by aerosol, at least in some regimes, then it follows that wet deposition also depends on AOD, and we quantify the strength of this coupling. This same approach could be used to study wet deposition of trace gases such as CO and ozone, as these are also available from the MACC re-analysis.