Variational And Hybrid Assimilation Of Satellite Cloud Observation Based On WRFDA

Given the importance of cloud data for precipitation simulation, in this study the cloud liquid and cloud ice mixing ratios are added as control variable in assimilation system to assimilate satellite cloud water/ice path. The cloud assimilation WRFDA-3DVAR system is developed and the cloud assimilation Hybrid scheme is built. The main work is as follows:Firstly, experiments with and without assimilating satellite cloud water/ice path are conducted to assess the impact of cloud liquid/ice water path data assimilation on short-term regional numerical weather prediction. It is shown that assimilating cloud satellite cloud water/ice path has a positive impact on temperature, wind as well as surface temperature and humidity. In addition, it also increases the accuracy of precipitation forecasts.Secondly, given the sensitivity of cloud and precipitation simulation to microphysical scheme, sensitive experiments using different mixed-phase microphysics schemes (i.e. LIN, SBU-YLIN, WDM6, and WSM6) with and without satellite cloud water/ice path are performed to investigate the impacts of different microphysics schemes on numerical weather prediction with cloud water/ice mixing ratios changed by the assimilation. Diagnostic analysis suggests that satellite cloud assimilation provides better dynamic condition and more realistic vapor content, which result in improvements of precipitation in model simulations. It is also found that the largest improvement occurs in the simulation with WDM6 scheme while the smallest improvement appears in results with SBU-YLIN scheme. The possible reason for different effects of cloud assimilation with different microphysical schemes is the difference between the time of cloud information maintained in model, the enhancement of ascending motion and the change of moisture condition.Thirdly, in order to improve the characteristic of background error covariance of cloud condensates, the cloud water/ice mixing ratios are added as extension control variable to the Hybrid system of WRFDA. Single observation experiments and 3-h rapid update cycle assimilation experiments of strong precipitation process are conducted. Results indicates that the background error covariance of cloud condensates and other model variables in the updated Hybrid data assimilation system is flow-dependent. In addition this new Hybrid data assimilation system can improve the simulation of dynamic and moisture condition so that the precipitation forecasts are improved markedly.