Research On Particle Spectrum Retrievation And Application Based On Millimeter Wave Cloud Radar

The thermodynamic and dynamic characteristics of atmosphere determine the processes of generation,development,and extinction of cloud,and affect the distribution characteristics of microphysical parameters such as cloud particle number concentration,size and phase states etc.Similarly,the microphysical structure and vertical motion characteristics of clouds also reflect the development of atmospheric thermodynamic and dynamic structures and weather systems.Therefore,the fine detection of cloud macroscopics and microphysics as well as atmospheric vertical motion has great significance to reveal the laws of the generation,extinction and evolution of the weather systems.In the dissertation,cloud microphysical parameter vertical distribution observation experiments are carried out by using a vertical profile integrated observation station which includes a millimeter-wave cloud radar,a microwave radiometer,a wind profile radar,and a lidar.Based on performing quality control on power spectrum observation data of the millimeter-wave cloud radar from 2018 to 2020,cloud particle spectra and atmospheric vertical velocities are retrieved by using a cost function optimization method at first;secondly,the retrieval results are cross checked and validated with the retrieval results by using the observation data from both the millimeter-wave cloud radar and the lidar or by using the observation data from both the millimeter-wave cloud radar and the microwave radiometer;and finally,and the vertical distribution characteristics of cloud microphysical parameters of typical weather processes such as stratiform cloud precipitations,convective cloud precipitations,and snowfall are analyzed in detail by using the cloud microphysical parameters which are calculated by using the obtained cloud droplet spectra.The main research results are as follows:(1)The computational efficiency and discrimination accuracy are improved by using an improved neural network algorithm combining a principal component analysis method and a K-means clustering method to identify clear-air echoes from the millimeter-wave cloud radar.For point-shaped clear-air echoes,the maximum time scale is 15 seconds and the vertical spatial scale is between 60 and 450 m.If the millimeter-wave cloud radar is used to directly detect the point-shaped clear-air echoes,the temporal resolution of the millimeter-wave cloud radar should be preset smaller than 15 seconds.For the layered clear-air echo caused by turbulent atmosphere,the vertical distributions between radar echo intensities and atmospheric structure constants have good consistency.(2)When a cost function optimization algorithm based on power spectra of the millimeter-wave cloud radar is used to retrieve cloud particle spectra,a linear relationship between the deconvolution ratio DF and ? / Vm is used to reduce the parameter number of the cost function from three to one,therefore calculation efficiency is improved.A gradient zero method is used to achieve multi-peak spectral separation,which solves the problem of low particle spectrum accuracy retrieved by using double or multi-peak power spectra,so that the method can be adapted to the retrieval of the multi-peak particle spectra of precipitation events and snowfall.(3)Compared the retrieval method only using the millimeter-wave cloud radar with the retrieval method combining the millimeter-wave cloud radar and the lidar and with the retrieval method combining the millimeter-wave cloud radar and the microwave radiometer,the comparison results as following can be obtained: the cloud particle spectra retrieved based on the millimeter-wave cloud radar power spectra are close to the two combination retrieval methods,the combination method by using the millimeter-wave cloud radar and the lidar is more sensitive to small particles;while the retrieval method based on the millimeter-wave cloud radar has a better retrieval effect on large particles;and the combination method by using the millimeter-wave cloud radar and the microwave radiometer is greatly affected by liquid water content.The error sources of cloud particle spectra retrieved by the millimeter-wave cloud radar mainly comes from detection mode,cloud and precipitation attenuation,and power spectrum types.(4)During the precipitation of stratiform clouds,velocity spectrum width in a melting layer area increases obviously,and double-peak and multi-peak spectra are generated;in the upper part of the melting layer,and Bergeron effects and the “catalyzing cloud-feeding cloud”process of cold clouds are obvious;and wet snow below the melting layer melts into raindrops and forms precipitation.(5)In the process of convective cloud precipitation events,the vertical distribution of reflectivity factors obviously has a “bow-shaped” characteristic.In the case of weak convection,the melting layer exists,which does not exist in the case of strong convection.In the strong convection process,precipitation particles are directly carried into cloud body by vertical updraft with large velocity and a secondary precipitation is formed.(6)In the process of snowfall,before the formation of wet snow,at the ambient temperature of-20?,ice crystal particles in clouds drop into areas with rich supercooled water,increase rapidly according to a Bergeron cooling effect,and then enter a freezing growth stage and form snowflakes.