Quality Analysis And Application Research Of The Ground Meteorological Microwave Radiometer Data

At present,the temperature and humidity meteorological elements still are obtained by rawinsonde sounding in the world.But,there is limited by the spacing of sounding stations and detection time,the sounding data is insufficient on the spatio-temporal continuity.The value of reference of sounding data could greatly reduce when there are strong convective weathers that burst strongly and are short duration.The microwave radiometer(MWR) have the advantage of continuous detection,it's data possesses the characteristic of high density of time and space.It can be as supplement for the rawinsonde sounding and play an important role in the monitoring and warning of severe weathers.However,because of the restriction of radiation principle,use of different areas,the differences in the process of factories and models,the MWR's data is restricted in the application of meteorological services.In order to give full play to the advantages of the MWR and provide services for the meteorological operations and research,it is of great importance to study the MWR's data roundly and explore the cause of the errors from the principle of radiation and an effective data control method.First of all,making the rawinsonde sounding data as a standard,with the method of time-space consistency,the quantitative analysis of the HTG3-MWR's data of July and August in Beijing and May and June in Hainan is done,the errors distribution of temperature, dew point,relative humidity and water vapor density is gotten in the different regions,different seasons,different time and different weather backgrounds.Then,in order to explain the reasons of errors between the two types of data,the radiative transfer characteristics of the clear sky,the fog and the rainy day are analyzed by simulation calculation in the model of Liebe.Based on the above results,making Hainan as the example,function relational expressions that are about the MWR's detecting altitude,temperature,water vapor density are, relative humidity and the errors are calculated in different weatherbackgrounds.The deviation of meteorological elements revised model is builted based on the BP neural network.The correction of the MWR's data by itself comes true.70% of the MWR's data in Hainan is randomly selected as training samples,10 deviation revised models in different weather backgrounds which are good in function are gotten.The remaining 30% as the validation samples is used, except the rainy day,it is founded that the correlation coefficients increase slightly and the average deviations improve obviously after correction. The average deviation of dew point is from 3? to 1?,the average deviation of water vapor density is from about 1g/m3 to0.26g/m3,the average deviation of relative humidity is from 17% to 3%.Because of the samples of rainy day are few,the availability of the deviation revised model about rainy day isn't high and the revision effect is also not good.But with the increase of examples,the above phenomenon can be improved.The next,detecting about the rationality of brightness temperature each channels of the HTG3-MWR is finished and the inversion algorithm is also improved through adding cloud information.So,there are some scientific advices and theoretical guidance about data quality control and software improvements for reference.Using the rawinsonde sounding data,the brightness temperature is calculated by radiative transfer equations and the model of Liebe.After system error correction,the consistency and continuity of the MWR's detecting brightness temperature are done,it proves that checking the status of the channels applies to the MWR's oxygen and water vapor channels.Making the sounding data of Yunnan from September to November in 2015 as example,the cloud information added in the process of radiometer inversion temperature humidity vertical profile.After adding cloud information to calculate,the correlation coefficient of temperature is from 0.997 to0.998,it's average deviation is from 0.96? to 0.76?.The correlation coefficient of relative humidity is from 0.838 to 0.946,it's average deviation is from 14.14% to7.08%.On the whole,the correlation coefficients and errors of temperature and relative humidity in each altitude layer get improved,relative humidity improvement is more obvious.