| Atmospheric temperature and water vapor are parameters of the physical state in the atmosphere,and they are also important meteorological parameters in the research fields of meteorology and environmental analysis.At present,the commonly used methods for detecting temperature and humidity profiles are Radiosond balloon,Raman lidar and microwave radiometer.Radiosond balloon can provide high vertical resolution temperature and humidity profile,but its detection frequency is limited and time resolution is low;Raman lidar can also provide high vertical resolution temperature and humidity profile,but there are blind areas in the bottom layer,the penetration ability of cloud is limited,microwave radiometer has strong ability to penetrate cloud,rain,fog and resist solar radiation,so it can obtain more accurate brightness temperature.Through inversion,we can get strong temperature profile of high temperature and humidity below 10km all day and all day,but its vertical resolution is low.Through inversion,we can get strong all-weather and all-weather temperature and humidity profiles below 10km,but its vertical resolution is low.In order to make up for the limitation of single measurement and realize the complementary advantages of Raman lidar and microwave radiometer,the radiosonde temperature humidity profile is taken as the standard,and the temperature and humidity profile of Raman lidar and microwave radiometer is fused by using the splicing method and the Optimal Estimation Method(OEM),so as to achieve the high temperature and humidity profile of all day,all-weather and high precision.In this thesis,firstly,the temperature and humidity profile inversion based on Raman lidar and Microwave Radiometer Based on BP neural network method is carried out.The error and statistical analysis of temperature and humidity profile of Raman lidar and microwave radiometer are carried out.The results show that the temperature and water vapor mixing ratio of Raman lidar are strongly positive correlated with radiosonde data within the effective detection range,and the correlation are above 0.96;the correlation between the temperature profile of microwave radiometer and sounding data are above 0.95,while the water vapor mixing ratio profile are 0.89 in cloudy days and 0.96 in cloudless days.Secondly,the temperature and humidity profiles of Raman lidar and microwave radiometer are fused.First of all,the high-quality parts were linear stitched,and the effective height of temperature and humidity profiles after splicing reached 10km.The correlation between temperature and water vapor mixing ratio in cloudy days and sounding data are above 0.90,and in cloudless days are above 0.96.Then the OEM is used to optimize the spliced profiles.The cost function is constructed by using the temperature and humidity profiles as the initial value,the brightness temperature measured by microwave radiometer as the standard,and the covariance matrix of the two.Levenberg Marquardt method and Monte Carlo method are used to optimize the cost function,and the optimal cost function corresponds to the optimal temperature and humidity profile.The correlation between temperature and sounding data obtained by Levenberg Marquardt method are 0.9661,and the error is within 2K;the correlation between water vapor mixing ratio and sounding data are 0.9585,and the error is within 0.4g/kg.The correlation of temperature profile obtained by Monte Carlo method is 0.9669,and the error is within 0.3k;the correlation of water vapor mixing ratio is 0.9851,and the error is within 0.3g/kg.The research shows that the linear splicing method can correct the blind area and the part of the cloud which is not penetrated by the Raman lidar.The OEM can effectively optimize the results after splicing.The Levenberg Marquardt method is weaker than the Monte Carlo method,but the optimization speed is far better than the Monte Carlo method. |
PDF Full Text Request