Research On Quality Control Method Of Sounding Air Temperature Observation

With the improvement of people’s quality of life,China has paid more and more attention to the prevention of meteorological natural disasters.As an important means of preventing meteorological natural disasters,high-altitude meteorological research is to improve the quality of second-level sounding meteorological observation data.Quality control technology is an important part of improving the quality of meteorological observation data.Therefore,improving the comprehensive performance of quality control of second-level sounding air temperature observation data is a basic requirement for preventing natural meteorological disasters.Based on this,this paper starts from the vertical distribution characteristics of the second-level sounding air temperature observation data in the barometric dimension,and considers the complex environment faced by the balloon sounding process,and builds a single station quality control model.The validity,generalization and stability of the model were analyzed by quantifying the effect of quality control through hypothesis testing methods.The main contents are as follows:This paper analyzes the vertical distribution characteristics of the second-level sounding air temperature observation of Xuzhou Station in the past 20 years by using the double-weight standard deviation discrimination technology and comprehensive static testing technology.The construction of a single station quality control model provides a theoretical basis.Based on this,the cubic spline interpolation(CSI)algorithm and cross-checking ideas are introduced into sounding quality control to build a CSI single-station sounding quality control model.By comparing with the system model,study its applicability at different altitudes and its stability on the seasonal time scale;Based on the feasibility and applicability test of the model,this paper analyzes the high-level analysis of the complex environment,considers the noise brought by the complex environment to the temperature observation data,and uses the ensemble empirical mode decomposition(EEMD)to improve the CSI quality control model.A method of EEMD-CSI sounding quality control is proposed.The second-level exploratory air temperature in Xuzhou,Shanghai,Beijing and Guiyang from 2006 to 2016 was selected as the observation data,and combined with the mean absolute error(MAE)and root mean square error(RMSE)to analyze the quality control effect of the method.Multiple case analysis shows that the vertical distribution characteristics of the secondlevel sounding air temperature observation data at different altitudes change significantly,so the sounding temperature observation data is divided into four layers: 1000-460 h Pa;460-90 h Pa;90-60 h Pa;60-10 h Pa.Compared with the system model,the CSI quality control model shows higher error detection and generalization capabilities at the 460-90 h Pa level,but at the other levels,the quality control effect and stability on the time scale are average.After the algorithm is improved,the EEMD-CSI single-station sounding air temperature quality control method performs better error detection and prediction capabilities at all levels,and the quality control effect in the plain area is significantly better than that in the high altitude area.