Research On Quality Control Method For Surface Temperature Observations Based On Spatial Dimension Analysis

Under the background of informatization,numerical weather prediction(NWP)has become an important means of meteorological prediction.The quality control of surface temperature observation data is the basis of data assimilation,which is helpful to improve the accuracy of numerical weather prediction.Based on the analysis of spatial distribution characteristics,spatial correlation and collaborative correlation of temperature observation data,a series of ground air temperature spatial quality control methods were established according to the perspectives of model effectiveness,running time cost of the model and collaborative prediction,and the test was carried out.The main contents are as follows:The hypothesis test method proposed by Professor Hubbard was adopted to randomly add artificial errors to the original temperature data,and the optimal quality control parameters were selected to identify the suspected error data,so as to provide evaluation indexes for the quality control model.In view of the validity of the model,on the basis of analyzing the spatial correlation of surface temperature observation data in various regions and considering the observation differences of temperature data in each reference station and central station,a surface temperature quality control method based on b-spline fitting(BSF_QC)is proposed;In order to reduce the running time cost of BSF_QC method in areas with high site density,a B spline fitting surface air temperature quality control method based on spatial regression test(SRT_BSF_QC)was proposed to screen the reference stations with the highest correlation by using spatial regression test;Considering the strong coupling between temperature and humidity,humidity data was taken as the supplement of temperature data,a B spline fitting surface temperature quality control method(CCA_BSF_QC)based on collaborative prediction was proposed for areas with few adjacent stations or missing historical temperature observation data.The analysis results of multiple groups of tests show that,as a local unbiased estimation method,BSF_QC method has excellent prediction performance and error detection performance compared with SRT and IDW methods,and can be effectively used for the interpolation of temperature data of observation stations,or for the detection of suspected error values in temperature data.For regions with high site density,if the spatial correlation of air temperature in the region is high,SRT_BSF_QC method can reduce the time cost of model operation on the basis of ensuring the validity of the model.If the spatial correlation of air temperature in the region is low,reference stations should be selected as many as possible to ensure the quality control effect of SRT_BSF_QC method.The CCA_BSF_QC method is applicable to areas with low site density or areas with more missing historical temperature data.After dimensionless processing of data,humidity data is used as a supplement to temperature data,which not only improves the diversity of data,but also improves the accuracy of quality control model.