| Radar quantitative precipitation estimation is a hot field in application of radar data, high spatial and temporal resolution of quantitative precipitation estimation analysis of products have great significance in weather forecast, analysis of cloud and precipitation physics research, numerical model calibration and verification, artificial influence on weather, climate change, industrial and agricultural production, flood control and flood forecasting. The raindrop size distribution is a reflection of the precipitation microscopic characteristics, and the temporal and spatial variation of the raindrop size distribution is one of the main reasons for the quantitative estimation of precipitation deviation. Based on the characteristics of multi echo parameters, the dual polarization weather radar has a certain adaptability to the temporal and spatial variations of the raindrop size distribution, but its estimation accuracy is still affected by the change of the raindrop size distribution. In addition, the estimation error of the polarization parameters ZDR is also the cause of the precipitation deviation of the dual polarization radar, and the variation of the raindrop size distribution has an influence on the estimation precision of the parameters to a certain extent.In this paper, we use the Gamma distribution function to simulate the distribution characteristics of the raindrops, and use the scattering theory to inverse the polarization parameters of the dual polarization radar. In his paper we analyze the raindrop size distribution rules of polarization radar parameters vary in a wide range of conditions, and put forward a kind of improved rainfall estimation formula:divide the differential reflectivity factor into some intervals to receive a series of rainfall estimation formula. At the same time, put forward three kinds of methods to divide ZDR:(1) using statistical t test to detect the precipitation coefficient with the change of ZDR, discover the points of marked mutation, and use the points to divide ZDR; (2) Divided according to a specific step; (3) Based on the fitting relationship between the median volume diameter and ZDR in order to divide it. Then use the simulated raindrop size distribution to analysis the three method, and compare the results.The first method use the statistical knowledge, based on the basic principle of t-test, sliding sampling the coefficient by 3 step size, calculate the t test statistic between two adjacent samples in order to judge the statistic by significance, if significant, then use the point to divide ZDR. This method has the best result in improving the accuracy in precipitation.The second method is to divide ZDR by step size of 0.1 dB、0.2 dB、0.3 dB、0.4 dB、0.5 dB in order to get a series of rainfall estimation formula. This method is simple and convenient, and it is easy to transplant. Using 0.1 dB to divide ZDR is the most effective way. However, the overall effect is slightly less than that of the method.The third method is based on the fitting formula and has definite physical meaning, it is the most effective when using 0.2 mm to divide Do. The overall effect is almost the same as the first method. The three are both effectively in improving precipitation measurement. However in some intervals, because of the weaken of relevance, the deviation of precipitation cannot be totally eliminated.Finally, according to the characteristics of the raindrop size distribution, use the method of signal vector superposition to simulate precipitation echoes and obtained simulated precipitation echo which contain the characteristics of raindrop number density and scale. To lay a foundation for further research on measuring precipitation deviation from the perspective of quality control in the future. |
PDF Full Text Request