Data Assimilation and Ensemble Modelling for the Improvement of Short Range Quantitative Precipitation Forecasts (DAQUA)

点击翻译按钮获取中文摘要

We propose the improvement of short range quantitative precipitation forecasting by regional high resolution weather forecast models. We will combine improved regional ensemble modelling with best member selection based on the most recent remote sensing information, with further broadening and narrowing of the distribution by a new evolutionary approach, followed by improved data assimilation and further forecast integration. The former step will introduce a Monte Carlo technique to invert a highly nonlinear and critically discontinuous microphysical problem in a novel way in cloud and rain assimilation. Physical initialisation techniques, nudging techniques and variational approaches with the most timely available information from remote sensing including Radar reflectivities, cloud parameters, and water vapour content will be employed. With this stacked procedure we expect to substantially reduce the influence of phase errors in the background field, which currently impede the successful assimilation of observations for short range precipitation forecasting by regional numerical weather forecast models. The project will establish an advanced capability for ensemble forecasting in the German research community. Validation efforts will explore and quantify the sources of uncertainty in forecasts especially under convective conditions. The differing requirements for the forecasting System under different meteorological conditions will be explored in the first instance by examining a set of case studies of convective storms in environments with orography of varying degrees of steepness. Some of the case studies will be orientated around the likely location of the field experiment, to aid in planning of the operations and to prepare for real-time forecasting.