Applying Stochastic Parameter Perturbations Directly into a WRF Microphysics Parameterization and Evaluating its Impacts Using GOES-16 Satellite Data

Due to the fact that cloud microphysics schemes contain many single value constants that derive from a large spread of values found in the observations, we decided to introduce perturbations directly into a few parameters of a WRF microphysics scheme. The stochastic parameter perturbation (SPP) technique of Berner et al (2017) with various spatial/temporal correlation patterns was applied within the Thompson-Eidhammer aerosol-aware microphysics scheme. Sensitivity experiments were performed in which perturbations affected: 1) the diagnosed cloud water gamma distribution shape parameter; 2) the Y‑intercept parameter of the graupel/hail distribution; 3) the activation of aerosols as cloud condensation and ice nuclei (CCN and IN, respectively); and 4) all of the above.

The choice to alter the model’s cloud water distribution shape parameter has a direct influence in the initiation of warm rain due to the impact of altering the mean size of cloud drops to inhibit rain formation in the case of shifting the distribution to smaller sizes or promoting the initiation of rain in the opposite case. The choice for altering the Y‑intercept parameter of the graupel/hail distribution can have profound influence in the development of convective weather due to dynamical feedbacks of the cold pool as well as the balance of hydrometeor mass held aloft in updrafts versus what reaches the surface as precipitation. The choice to alter the initiation of CCN and IN can directly impact the cloud droplet and ice number concentration that affects mean size of each hydrometeor as well as subsequent collection and conversion to rain and snow.

Preliminary sensitivity experiments will be presented along with comparisons of the forecast cloud fields to observed GOES‑16 ABI brightness temperatures. WRF was configured with a grid of high resolution (dx = 3 km) over most of the continental U.S. similar to the High Resolution Rapid Refresh (HRRR) model to determine if an ensemble of SPP-altered simulations improved convective weather forecasts.

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