Study on variational quality control of Huber model based on observation errors of NOAA19 satellite data channels

Satellite data assimilation is an important method for improving the quality of the initial field in numerical weather prediction. However, the restriction of observation error in satellite data will lead to low effective assimilation rate of some available satellite data, which in turn reduces the contribution of satellite data to the assimilation analysis. The variational quality control scheme adjusts the weight of data to ensure that observations of different quality are utilized appropriately, thereby effectively improving the performance of the assimilation analysis. Based on the Huber-VarQC variational quality control scheme, which more appropriately characterizes the non-Gaussian observation error of satellite data. It optimizes the relevant parameters for different satellite channels according to their specific error characteristics, allowing the assimilation system to adjust the weight of observation in the assimilation analysis based on the different observation error feature of each channel. This enhances the utilization and assimilation efficiency of satellite data and improves the quality of the assimilation analysis. The experimental results indicate that the Huber-VarQC scheme can effectively capture the "fat-tailed" distribution characteristics of observation error in satellite data across different channels. Statistically analyzing satellite observation error by channel and optimizing the Huber-VarQC variational quality control scheme can maximize the practical application potential of this approach and enhance the contributions of satellite data to the analysis. The channel separation Huber-VarQC scheme can allocates appropriate weight to the data based on the error characteristics of each channel, and then increasing the effective assimilation rate of polar orbit meteorological satellite microwave observation data. This approach not only incorporates beneficial information from the data but also mitigates the negative impact of harmful information on the assimilation analysis. As a result, it enhances the effective assimilation rate of satellite data and improves its positive contribution to the assimilation analysis, particularly demonstrating significant improvements in the forecasting of micro- and meso-scale heavy precipitation events.