A preliminary study of reflectivity data resolution enhancement technology for Geostationary Doppler weather radar

Geostationary Doppler Weather Radar(NEXRAD in-space:NIS) is an advanced instrument concepts and technologies for future spaceborne atmospheric radars which can provide hourly monitoring of the life cycle and acquire 3-dimensional information on the intensity and vertical motion of hurricane rainfall of hurricanes and tropical storms. Contrast to current TRMM Precipitation Radar (PR), NIS has lots of characteristics better than that of TRMM, such as high time resolution, wider surveillance domain, no effects of satellite platform movement. However, due to its large distance far from the earth, the NIS horizon resolution of ground reflectivity data has been decreased to more than 12 kilometers, which would result to large bias of precipitation estimation and limit the detection of small-scale weather system such as Severe storms, Tornado and Mesocyclone. Therefore, it's necessary to enhance the horizon resolution for NIS ground data. NIS echo power in each range bins contribute to all of scatters in one radar resolution volume which could be taken as the results of all scatters echoes' power weighting, and these weighting functions in angular and range are in aspect relate to weather radar's effective antenna beam function, the pulse shape of transmitter and filter response of receiver which are known or can be measured. Now we divide one resolution volume into several sub ones ,as a result NIS echo power can also be taken as weighting results of these sub resolution volumes, if enough measurements could be acquired, we can use deconvolution technology to retrieval these sub resolution volume echo power and improve NIS reflectivity data resolution. In this paper, a linear inversion mathematics formula between sub resolution volume echo power and multiple partially correlated radar measurements in horizon is present based on radar meteorology integrated equation, its solution is pursued by means of a super resolution numerical procedure based on the truncated singular value decomposition (TSVD) method using known radar antenna weight function. Numeric simulation uses the APR-2 reflectivity measurement data as initial true high resolution (400m) data to generate NIS measurement data of low resolution (12km ) through the integral equation above , sub resolution volume reflectivity reconstruction through TSVD algorithm is based on these NIS measurement data adding Gaussian noise. Three parameters of amplitude scope difference, standard error and correlation coefficient among these three types of data and a method of subjective image contrast are used to evaluate the resolution enhancement performance. Results show that these three performance parameters of super-resolution reconstructed data from NIS simulation measurement are better than those of reconstructed before and are all close to those of simulated reflectivity data of high resolution(4km) , more fine weather structure among these reflectivity data could be present after reconstruction.