Extended Abstract (404K)7A.2
Forecasting the solid-to-liquid ratio of precipitation a cloud-resolving model
Jason A. Milbrandt, Environment Canada, Dorval, QC, Canada
The horizontal grid-spacing of numerical weather prediction models is continually increasing and high-resolution cloud-resolving models are quickly becoming important tools in operational meteorology. In these models, the precipitation is predicted almost entirely from the cloud microphysics parameterization scheme. The total frozen precipitation comes from the sum of the precipitation of the various ice-phase hydrometeor categories in the scheme. Typically, the precipitation rate from the model is given as the mass flux at the surface, which is used in turn to compute the accumulated liquid-equivalent precipitation. For some applications, the depth of the unmelted snow that is forecast is desired. Typically, this quantity is obtained by applying a solid-to-liquid ratio, either an assumed value or one obtained by some post-processing algorithm, to the forecast liquid-equivalent amount. The methods, however, are generally based on rules-of-thumb and do not always work well to forecast snow depth.
In this study, a method is proposed to obtain an estimate of the instantaneous bulk snow density from the cloud microphysics scheme. With this, the volume flux of unmelted precipitating snow at the surface can be obtained and the instantaneous solid-to-liquid ratio of precipitating snow can be determined. The approach is essentially to compute a mass-weighted average of the bulk densities of the precipitating ice-phase categories in the cloud scheme: pristine ice crystals, large crystals/aggregates, and graupel (rimed crystals). The bulk densities of pristine ice and graupel are prescribed constants, while the density of the large crystal/aggregate category is a diagnostic function of the maximum particle dimension, based on recent distrometer measurements. While this method is simple and still depends heavily on assumptions made in the microphysics scheme, it allows us to exploit information in the model about the growth environment at given points in time and space. For example, the solid-to-liquid ratio will be relatively large, reflecting a low bulk snow density, for a dry environment (i.e. no supercooled liquid water) in which snow is grows by deposition only and the ratio will increase further as the ambient temperature warms and the rate of aggregation increases. On the other hand, if riming occurs and graupel forms, the solid-to-liquid ratio will decrease.
A description of the method, as applied in the cloud microphysics scheme of the Canadian GEM-LAM model, will be presented along with discussion of some recent changes to the snow category in the microphysics scheme. Case studies from the west coast of Canada and from the Ontario-Quebec region will be presented to demonstrate the skill of the proposed method.
Session 7A, Microphysical Processes in Modeling
Wednesday, 3 June 2009, 8:00 AM-9:00 AM, Grand Ballroom East
Previous paper Next paper
Recorded presentation