The Penn State/NCAR Mesoscale Model (MM5) was used in a two-dimensional configuration to investigate the sensitivities within the so-called Reisner2 bulk microphysical parameterization (BMP), which includes graupel and super-cooled water processes. Simulations were completed for the 11-12 February 1986 event during the Sierra project as well as for some simple bell-shaped terrain geometeries and freezing levels. A microphysical budget over the windward slope for each experiment illustrates the primary microphysical pathways that lead to the precipitation differences in the simulations.

A number of processes are examined as well as the impact of horizontal grid spacing (when using the same terrain forcing) and model time step. The microphysical processes examined include the sensitivities of ice initiation, CCN, slope intercept for snow (NOS), fall speeds, hydrometeor densities, autoconversions, secondary ice production. It was found that grid spacings of less than a few kilometers were important to properly simulate the mid-level convection within a shallow potentially unstable layer aloft during the Sierra event. There was little sensitivity to model time step, ice initiation, and secondary ice production. The model sensitivities were largest for NOS, fall speeds, and graupel density; however, as the barrier half width decreases from 50 to 10 km there is less sensitivity with the various ice parameters (NOS, fall speeds, etc...) since snow is more easily advected into the lee. An increase in the freezing level also results in less ice parameter sensitivity, but very large sensitivities develop with the cloud water autoconversion.