［ 2017年03月02日 ］
RIKEN International Symposium on Data Assimilation 2017
"What observations do we need for convective-scale data assimilation?"
Regional models with a grid spacing of a few kilometres ("convective-scale models") explicitly represent deep convection and hazards related to convective systems are often primary forecast events. This results in low predictability and calls for the assimilation of spatially dense and temporally frequent observations. In developed countries, there is a wide variety of observed information available, but little knowledge exists on what types of observations are most beneficial and how to use remote-sensing observations related to clouds and precipitation efficiently in Ensemble Kalman Filters. The HErZ research group at LMU tries to answer these questions in a three-fold effort: (1) The use of ensemble sensitivity analysis to estimate the potential impact of different observed variables; (2) The use of ensemble forecast sensitivity observation impact (EnFSOI) to estimate the actual impact of observations in the pre-operational LETKF system of Deutscher Wetterdienst and (3) work on an improved use of cloud-affected and cloud-related satellite observations that constitute a potentially very valuable source of information on convective systems. The presentation summarizes these efforts and provides preliminary findings on what observations are potentially most beneficial for convective-scale data.
|名前:Martin Weissmann||所属:LMU Munich, Germany|