Skill of Operational Aerosol Forecast Models in Predicting Aerosol Events and Trends of the Eastern United States.

Global aerosol forecast models are now commonplace, providing predictions of dust storms, smoke event transport and even anthropogenic pollution events out to 6 days or even seasonally. These models often serve dual purposes providing analysis or reanalysis products for earth science applications as well as aerosol forecast guidance. The NASA Studies of Emissions & Atmospheric Composition, Clouds & Climate Coupling by Regional Surveys (SEAC4RS) coupled with the International Cooperative for Aerosol Prediction Multi Model Ensemble (ICAP-MME) provided an excellent opportunity to evaluate the nature and skill of global aerosol forecast models to simulate or predict the nature of a relatively "simple" anthropogenic pollution regime of the eastern United States. Generally, models are able to capture the relative distribution of haze events two days out, and reanalysis versions easily capture decadal trends. However, aerosol physics, chemistry, and meteorology uncertainties lead to the conclusion that these forecasts and reanalyses should be interpreted along the same semi-quantitative lines as most forecasters interpret meteorological model forecasts and analyses. Here, we systematically explore how differences in model configuration and data assimilation methodologies translate into differences in final model analysis and forecasts in a series of events observed during SEAC4RS and then extrapolate findings to the problem of decadal monitoring. We also explore the impact and ramifications of transient events such as biomass burning and dust impact forecast product interpretation.