Cooperative Institute for Research in Environmental Sciences

A climatology of short-range weather forecasts from the High Resolution Rapid Refresh model

Eric James (1), Curtis Alexander (1), Brian Jamison (2), Stan Benjamin (3), Steve Weygandt (3)

(1) CIRES, (2) CIRA, (3) NOAA ESRL

The High Resolution Rapid Refresh (HRRR) model is being run hourly in real-time at the Global System Division (GSD) of the Earth System Research Laboratory (ESRL). The model is run out to fifteen forecast hours over a domain covering the entire conterminous United States (CONUS) at a spatial resolution of three kilometers, allowing the use of explicit convection. Initial and boundary conditions are obtained from the 13-km Rapid Refresh (RAP). In order to learn more about the systematic biases in the HRRR, we have initiated a long-term effort to map the temporal and spatial error patterns for an extensive history of HRRR forecasts. To facilitate this study, we established a real-time archive of two-dimensional HRRR forecast grids starting in late January 2012. In this presentation, we will stratify HRRR forecasts by time of day and lead time in order to discern the model’s skill in representing the diurnal cycle of boundary layer structures and moist convection. We will also group forecasts by synoptic weather regimes and by sub-regions of the model domain. Forecasts of low-level winds and cloudiness will be evaluated in the context of the HRRR’s utility for use in renewable energy applications. Evaluation of HRRR convective forecasts in regional and diurnal perspectives will be presented and ultimately used to calibrate time-lagged ensemble forecasts of convection. We will provide a description of HRRR time-lagged ensemble convective forecasts and their use as an estimate of thunderstorm likelihood for aviation and severe-weather applications.