Evaluation of microphysics and cumulus parameterization schemes in the HWRF model using satellite observations

The ability of several cloud microphysical and cumulus parameterization schemes in the operational HWRF model to accurately simulate the spatial characteristics and temporal evolution of the cloud and water vapor fields is assessed through comparison of observed and simulated satellite infrared and microwave brightness temperatures. The Community Radiative Transfer Model in the Unified Post Processor is used to generate the simulated satellite brightness temperatures. Bulk cloud characteristics such as the spatial extent of cloud cover are examined using brightness temperature probability distribution functions, traditional point statistics, and neighborhood verification methods such as the Fractions Skill Score (FSS). The inner core hurricane structure and the spatial distribution of deep convective areas are also assessed using simulated microwave observations. In addition, the observed and simulated infrared brightness temperatures are input to the Advanced Dvorak Technique (ADT) to assess the accuracy of the simulated cloud-inferred tropical cyclone intensity and the organization and location of deep convection in the eye wall and surrounding areas. Lastly, dropsonde observations are used to assess the accuracy of the thermodynamic fields within the tropical cyclone environment.

The analysis focuses on a set of cycled forecast experiments performed using an updated version of the 2015 operational HWRF model for the case of Hurricane Edouard (2014) in the Atlantic Ocean. A positive water vapor bias that increases with forecast lead-time is evident in the water-vapor sensitive infrared brightness temperatures simulated for all experiments. Preliminary results also indicate that the vertical and horizontal distribution of clouds is very sensitive to which parameterization scheme is used. In particular, differences in the spatial structure of the upper-level cloud field and the inferred strength of deep convection can be especially large when advection of cloud species is enabled. The cumulative impact of these differences on both the large and small-scale phenomena that impact the track and intensity of the hurricane will also be discussed.