Last week's ice storm in the south gave a good opportunity to see what the RUC precip type algorithm was doing. What we noticed was that the freezing rain indicator was very good, with fairly accurate coverage. However, much of the same area also had a positive ice pellet indicator, which was not accurate. This is from a general problem regarding the MM5 microphysics used in the RUC, in which there is too much graupel production near 0 deg C. There is a fix that NCAR has been working on for this that we are starting to test at FSL now.

Any comments from users about their experience with the RUC precip type fields?

There is a composite precip type product from the RUC available at the NCAR weather site at http://www.rap.ucar.edu/weather/model.html (Greg Thompson's page) that you can look at. The "mixed" category is often from the freezing rain/sleet mix that happened in last week's storm, but it can also be snow/rain or other mixtures. These mixtures are possible in the RUC precip type forecasts because of the explicit treatment of ice phase as well as water phase with the explicit forecasts for rain water, snow, and graupel. In other words, the precip type from the RUC is not just diagnosed from the surface temp and model sounding temp, but is a result of the mixed phase microphysics running directly in the model.

Please post any comments about your experience in using these RUC fields.