Research On Quality Control Algorithm Of Surface Air Temperature Observation Data In Different Geographical Environment

To improve the accuracy of numerical weather prediction,data assimilation technology has become an important tool to solve the problem,and an accurate and reliable data source is a prerequisite for data assimilation,so effective quality control of data is especially important.In addition,high-quality ground-based meteorological observations are one of the main prerequisites for the study of climate change characteristics,which ensure the reliability of the input variables in the analytical models and are important for the analysis of climate characteristics and their evolution patterns.The current ground-based meteorological observations are mainly affected by the observation method,observation environment and station density,resulting in a large number of suspicious and missing values.In this paper,considering the topography,spatial and temporal correlation,station distribution density and extreme weather characteristics,we construct quality control models for meteorological observation stations in different geographical environments respectively,and analyze the prediction effect and error detection effect,and the specific work is as follows:(1)Based on the fact that regions with complex geographical environments such as hills and plateaus are susceptible to external natural factors such as topography and landscape,traditional quality control models such as IDW and SRT cannot meet the meteorological operational needs of these types of regions,a ground temperature quality control model(MVO-MQ)applicable to multiple terrain is proposed.The model is based on the multi quadric equations,using the cosine similarity to analyze the spatial correlation between stations,and obtaining the optimal parameters of the model by virtue of the multi verse optimizer algorithm(MVO)to ensure that the model is in the best condition.The experimental results show that the MVO-MQ algorithm has high accuracy and error detection in plains and hilly areas,but performs poorly in highland areas,and the overall generalizability of the model still needs to be improved.(2)In order to further improve the prediction accuracy of the MVO-MQ model in multi-terrain and solve the problem of poor quality control in plateau areas,a multi-terrain surface air temperature quality control model(EC-MSM)based on error compensation was proposed.The model integrates the temporal correlation and spatial correlation between stations in the study area,measures the spatio-temporal correlation between stations with sample cross correlation function,and selects more reasonable nucleation nodes as the input variable of the quality control model.Meanwhile,multi verse optimizer algorithm and multi quadric equations based on the improvement of error compensation are combined to reduce the interference of external natural factors on the quality model.The experimental results show that the EC-MSM algorithm has higher prediction accuracy and error detection effect in different terrains,especially the most prominent improvement of error detection effect in the western plateau region,but the region is affected by extreme weather and the data false detection rate is higher,which makes the error detection effect still has more room for improvement.(3)In addition to the western plateau,there are also sparse station distribution and frequent extreme weather in other regions of the west.In order to further improve the error detection effect of the temperature observations quality model in this scenario,a single station quality method based on variational mode decomposition is proposed.Control Model(MVK).Firstly,the temporal and spatial variation characteristics of temperature in the western region are analyzed to obtain the information of areas with frequent extreme temperature events,and the stations in the selected area are taken as the research objects of the MVK algorithm,and the extreme data and routine data in the data are marked by the percentile threshold method.Secondly,the MVK algorithm is used to decompose-predict-reconstruct the station temperature series;finally,the hypothesis testing method is used to determine the quality control parameters corresponding to the extreme data and the conventional data respectively.The experimental results show that the quality control parameters in the temperature data are determined by classification,which can not only improve the error detection effect of the quality control model,but also preserve the extreme weather information to the greatest extent.