15R2O Impact of Aerosol Forcing on UFS-Based Seasonal-Length Predictions

NOAA is in the process of developing a new physically-based modeling system – the Seasonal Forecast System (SFS) – for seasonal-length prediction out to 12 months lead time; SFS is planned to eventually replace the CFSv2 for operations. For long-range seasonal prediction in a coupled atmosphere-ocean framework, it is difficult to properly represent the time-varying aerosols which are crucial for meteorological forcing (e.g., radiation and clouds). In real-time applications, monthly-varying climatology (i.e., “fix files”) are often used, although it is understood that the true Earth system can have aerosol forcings that, in any given year, deviate from climatology. The SFS development strategy relies heavily on hindcasts, as these historical forecasts can be evaluated against truth. The use of hindcasts allows for observed (or reanalysis) datasets to be strategically inserted into portions of a coupled modeling system, in order to reduce errors arising from that component. Here, we replace climatological aerosol forcing with observed MERRA2 monthly mean aerosols in 12-month-long UFS hindcasts initialized every May 1^st from 2000 through 2023 (24 total cases); these “observed aerosol” (OA) runs are compared against a control set of runs that uses “climatological aerosols” (CA). Not surprisingly, it is found that the OA runs have improved atmospheric prediction relative to CA. More importantly, however, the OA runs allow for better tuning of other components of SFS (e.g., atmospheric physics parameterizations) by minimizing errors due to aerosol forcing.