Subseasonal Forecasting with an Icosahedral, Vertically Quasi-Lagrangian Coupled Model. Part II: Probabilistic and Deterministic Forecast Skill

Abstract

Abstract Subseasonal forecast skill of the global hydrostatic atmospheric Flow-Following Icosahedral Model (FIM) coupled to an icosahedral-grid version of the Hybrid Coordinate Ocean Model (iHYCOM) is evaluated through 32-day predictions initialized weekly using a four-member time-lagged ensemble over the 16-yr period 1999–2014. Systematic biases in forecasts by the coupled system, referred to as FIM–iHYCOM, are described in a companion paper (Part I). This present study (Part II) assesses probabilistic and deterministic model skill for predictions of surface temperature, precipitation, and 500-hPa geopotential height in different seasons at different lead times ranging from 1 to 4 weeks. The coupled model appears to have reasonable agreement with reanalysis in terms of simulated weekly variability in sea surface temperatures, except in extratropical regions because the ocean model cannot explicitly resolve eddies there. This study also describes the ability of the model to simulate midlatitude tropospheric blocking frequency, Madden–Julian oscillation patterns, and sudden stratospheric warming events—all of which have been shown to be relevant on subseasonal time scales. The metrics used here indicate that the subseasonal forecast skill of the model is comparable to that of several operational models, including the National Oceanic and Atmospheric Administration’s (NOAA’s) operational Climate Forecast System version 2 and the European Centre for Medium-Range Weather Forecasts model. Therefore, FIM–iHYCOM—as a participant in NOAA’s Subseasonal Experiment—is expected to add value to multimodel ensemble forecasts produced through this effort.