Utilizing Ensemble Forecast Sensitivity to Observation (EFSO) Metrics to Assess Simulated Observation Impacts

Ensemble Forecast Sensitivity to Observations (EFSO) methods are useful in identifying the overall impacts of observations that are assimilated into global forecasting models, revealing observations that negatively impact the forecast. This can be greatly beneficial to Observing System Simulation Experiments (OSSEs), confirming the OSSE framework yields similar observation impacts to real-world conditions while helping to optimize usage of proposed new observations. EFSO tools developed by Hotta et al. (2017, MWR) and T.-C. Chen (2018, UMD thesis) are applied to an OSSE assessing the impact of simulated Geostationary Hyper-Spectral infrared Sounders (Geo-HSS) simulated from the Goddard Earth Observation System version 5 global Nature Run and assimilated into the Q1FY15 version of the National Centers for Environmental Prediction Global Forecast System. Impacts of these new observations before and after EFSO-based optimization of Geo-HSS channel selection will be discussed, including a Southern Hemisphere forecast skill dropout case attributed to the addition of Geo-HSS.