Diversity of Aerosol Optical Thickness in analysis and forecasting modes of the models from the International Cooperative for Aerosol Prediction Multi-Model Ensemble (ICAP-MME)

Abstract:

With the emergence of global aerosol models intended for operational forecasting use at global numerical weather prediction (NWP) centers, the International Cooperative for Aerosol Prediction (ICAP) was founded in 2010. One of the objectives of ICAP is to develop a global multi-model aerosol forecasting ensemble (ICAP-MME) for operational and basic research use. To increase the accuracy of aerosol forecasts, several of the NWP centers have incorporated assimilation of satellite and/or ground-based observations of aerosol optical thickness (AOT), the most widely available and evaluated aerosol parameter. The ICAP models are independent in their underlying meteorology, as well as aerosol sources, sinks, microphysics and chemistry. The diversity of aerosol representations in the aerosol forecast models results in differences in AOT. In addition, for models that include AOT assimilations, the diversity in assimilation methodology, the observed AOT data to be assimilated, and the pre-assimilation treatments of input data also leads to differences in the AOT analyses. Drawing from members of the ICAP latest generation of quasi-operational aerosol models, five day AOT forecasts and AOT analyses are analyzed from four multi-species models which have AOT assimilations: ECMWF, JMA, NASA GSFC/GMAO, and NRL/FNMOC. For forecast mode only, we also include the dust products from NOAA NGAC, BSC, and UK Met office in our analysis leading to a total of 7 dust models. AOT at 550nm from all models are validated at regionally representative Aerosol Robotic Network (AERONET) sites and a data assimilation grade multi-satellite aerosol analysis. These analyses are also compared with the recently developed AOT reanalysis at NRL. Here we will present the basic verification characteristics of the ICAP-MME, and identify regions of diversity between model analyses and forecasts. Notably, as in many other ensemble environments, the multi model ensemble consensus mean outperforms all of the models therein with respect to RMSE. Polluted and mixed fine-coarse environments of India, China, and the Sahel pose the most difficulty. Based on the evaluation of analyzed and forecast AOT from the ICAP models, we will discuss potential paths to improved aerosol forecasts from these systems.