Ice Phase Parameterization in a Numerical Weather Prediction Model

Abstract

This article presents the implementation of simple parameterization schemes for ice phase microphysics, with snow as a diagnostic variable. The microphysical schemes are used within a standard parameterization scheme for stratiform and convective precipitation in a numerical weather prediction model with a 50-km mesh width. The aim is to improve precipitation forecasts, improve forecasts of temperature changes due to latent heat transfer, and to supply the weather service with a tool for prediction of snow. Another objective is to determine to what extent the simplifications of the parameterization are affecting the model?s forecast. The schemes are compared to a condensation scheme with no ice phase included. Parallel runs show that the impact on the general dynamics is limited compared to control runs. Precipitation from the ice phase parameterization model has larger maximum intensity and is often of less horizontal extent than the fields in the control runs. Prediction of snow is improved. Introducing cloud ice as a prognostic variable has an insignificant effect on the development of the forecast as compared to a scheme when cloud ice is diagnosed. Introduction of a probability function for ice gives small changes in distribution of snow and rain, but the scale of the changes is beyond the resolution of the observational network.