Radar-data assimilation into the Rapid Refresh (RR) and High Resolution Rapid Refresh (HRRR) models toward improved convective guidance for aviation

Assimilating radar data into mesoscale models has the potential to improve short-range convective forecasting through information provided about wind, temperature, water-vapor, and/or hydrometeor fields in small-scale features (convective storms and systems, mesoscale boundaries, etc.). Introducing quasi-balanced small-scale features at the initial time has the potential to improve forecasts out to ~6 hours, with applications to both terminal area forecasts and aviation route planning. In some situations, improvement in 12+ hour forecasts might be possible if the radar-data assimilation produces more realistic convective evolution earlier during the model integration.

Currently, radar-reflectivity and Doppler-velocity data are assimilated into the National Oceanic and Atmospheric Administration (NOAA) Rapid Refresh (RR) model, an hourly-updated implementation of the Advanced Research Weather Research and Forecasting (WRF-ARW) model with 13-km horizontal grid spacing. The RR's reflectivity-data assimilation occurs during diabatic digital filter initialization (DDFI). During the DDFI, latent heating arrays that are computed from gridded reflectivity data replace the temperature tendencies that would otherwise be computed from microphysical processes in the model.

NOAA's experimental High Resolution Rapid Refresh (HRRR) model, also an hourly-updated WRF-ARW model, produces convective storms explicitly through the use of 3-km horizontal grid spacing. The HRRR was an important component of the FAA's 2010 operational performance evaluation of CoSPA to assess its impact on air traffic management decision-making. Currently, the HRRR initialization only includes radar data through the initialization of its parent 13-km model. Reflectivity-data assimilation into the HRRR itself is currently being tested in retrospective HRRR experiments, in preparation for possible real-time implementation. Such data assimilation has the potential to produce more realistic convective evolution and more accurate forecasts immediately after the initialization period and throughout the first ~6 hours of the forecast. Microphysical temperature tendency on the 3-km grid is provided by latent heating fields computed from reflectivity data at multiple times during a relatively short (~1 hour) period of model integration. At the conference, HRRR and RR forecasts initialized with and without radar-data assimilation will be compared and verified.

This research is partially in response to requirements and funding by the Federal Aviation Administration (FAA). The views expressed are those of the authors and do not necessarily represent the official policy or position of the FAA.