Real-time and Retrospective 3?4-week Forecasting For The NOAA Subseasonal Experiment With The Fim-hycom Coupled Model

Abstract

Multi-year retrospective tests have now been performed for relatively new atmospheric-oceanic models (including the FIM-HYCOM coupled model) toward calibration of experimental multi-model forecasts for NOAA’s Subseasonal Experimental (SubX) forecast program. The coupled atmosphere (FIM) and ocean (HYCOM) modeling system was developed primarily for subseasonal prediction (Green et al. 2017 and Bleck et al. 2015, both in Mon. Wea. Rev.). Both component models operate on a common icosahedral horizontal grid and use adaptive hybrid vertical coordinates. FIM-HYCOM also uses a unique physical parameterization suite including an evolved version of the Grell-Freitas (2014, G-F) scale-aware convective parameterization. The sensitivity of (simulated) earth system processes key for subseasonal prediction to physics, resolution, and numerics is shown through different sets of FIM-HYCOM coupled model 32-day hindcast experiments (3200 runs each) over a 16-year period. Such results include the finding that using the G-F convection scheme instead of the Simplified Arakawa-Schubert scheme improves predictive skill of Madden-Julian Oscillation indices; simulated blocking frequency is not sensitive to resolution; and that FIM-HYCOM can produce successful forecasts of stratospheric warming events 14-20 days in advance. FIM-HYCOM performance for the first 6 months (July-December 2017) of the real-time SubX period will be described, including for 2-m temperature and precipitation anomalies. Ongoing work will also be reported on for adding inline aerosol/chemistry treatment to the coupled FIM-HYCOM model and for advanced approaches to subgrid-scale clouds. This physics configuration from this project will be a candidate for testing in the upcoming FV3-coupled model now in development in NOAA.