Numerical Experiments in Mesoscale Prediction over Southeast Australia

Abstract

A new movable fine-mesh model (MFM) developed at the Australian Numerical Meteorology Research Centre, for mesoscale data assimilation and forecasting, is described. For the present, it is a 10-layer, hydrostatic, primitive equations model with 60 km horizontal resolution nested in a one-way interactive sense within a coarse-mesh primitive equations model of 250 km horizontal resolution. The MFM model has been designed to take advantage of recent local increases in data availability and computer power which have hitherto prevented much attention being given to numerical forecasts of mesoscale features in Australia. Important features of the model include an efficient semi-implicit time differencing algorithm, the use on option of a vertical mode initialization procedure, and the incorporation of realistic topographic forcing. The performance of the MFM model has been assessed on eight case studies over southeastern Australia. Two forecasts are discussed here in detail. The firm is an example involving the development and translation of a mid-tropospheric mesoscale ?cold pool?. Such situations are difficult to anticipate and often result in significant rainfall occurrence over the region. The second is a ?Southerly Bustee? event, a severe coastal windstorm which affects New South Wales about 30 times each year in the warmer months.