Recent Improvements To Ceiling and Visibility Forecasts In The RAP and HRRR Models

Abstract

The 13-km Rapid Refresh (RAP) and 3-km convection-allowing High-Resolution Rapid Refresh (HRRR) are hourly-updating forecast models that support operational aviation weather forecasting and other short-range forecasting interests within the contiguous United States. Since 2010, the forecast accuracy of operationally restrictive ceilings and visibilities (hereafter C/V) has steadily improved in the RAP and HRRR, and is largely attributable to better representation of grid-scale cloud cover through the microphysics parameterization, and to improvements in model analysis fields via hybrid ensemble–variational assimilation of temperature and water-vapor observations. This presentation will discuss ongoing work to sustain continued C/V improvements in future versions of the RAP and HRRR. An end-to-end array of experimental model-system developments, encompassing new data-assimilation, model physics, and model postprocessing innovations, will be presented. These developments include: (1) the transition of the RAP/HRRR hydrometeor analysis from a non-variational to a hybrid ensemble–variational system, (2) application of a subgrid-scale cloud parameterization in RAP/HRRR, and (3) use of a simplified, cloud-fraction-based ceiling diagnostic during model postprocessing. Drawing from both real-time and retrospective forecasts results, the merits of these model-system developments will be evaluated, and their pathway to eventual operational implementation will be discussed.