Real-time analysis and short-term forecasting of snowbands using a mesoscale model

High resolution mesoscale models, such as MM5, are potentially useful tools for predicting snowfall in the terminal area in the 1-12 hour range. However, despite improvements in model physics, there still exist significant discrepancies between the observations and model predictions of the timing, duration and amount of snowfall produced by snowbands. To improve the accuracy of the MM5 forecasts, high resolution datasets need to be analyzed and assimilated into the model initial conditions.

The objective of this work is to understand the mesoscale structures of the snowstorms and to improve the mesoscale model forecasts using high resolution observations. A number of case studies have been conducted. The model forecast fields are examined and compared with observations from various platforms, especially those from Doppler radars. Techniques that can effectively assimilate radar data into MM5 are explored.

Numerical experiments using a real-time, MM5-based, four-dimensional data assimilation (RTFDDA) system will be presented in the paper. The RTFDDA system was run operationally for seven weeks during the past winter. Preliminary results of radar data analysis and its impact on MM5 forecasts will be presented.