Error Analysis of Regional Radar-gauge Synchronously Integrated Precipitation Estimation Equation

Research is focused on the error from A_BS regional radar precipitation estimation equation (Wan et al., 2010). It is pointed out that the presence of μ_i misestimate factor of single station is the fundamental reason for radar precipitation estimation error. This misestimate factor is caused by the relative bias between A_Bi (from single station observation) and A_BS (common coefficient from the whole region). Two important evaluation criterions for precipitation estimation on radar covered region are developed and their analytic expressions are given based on μ_i. The first criterion is the regional precipitation misestimate ratio μ_S, which is defined as the ratio of the sum of every single station's misestimate value (positive or negative) and the sum of the corresponding station's observed precipitation in the whole region. The misestimate ratio μ_S is used to evaluate the estimation error of the regional precipitation gross. The second criterion is the regional precipitation absolute misestimate ratio μ_|S|, which is given by the ratio of the sum of every single station's misestimate absolute value and the sum of the corresponding station's observed precipitation in the whole region. The absolute misestimate ratio μ_|S| is used to evaluate the gross absolute estimation error of the region and is the most widely used criterion for precipitation estimation error. The value of μ_|S| indicates the general polymerization and dispersion degree of A_Bi from A_BS of every station within the area S . Under the different raingauge network conditions of basic network, intensive network and grouping network, the difference and variance of μ_|S| during one same rain event is analyzed. It is revealed that due to the nonuniform distribution of A_Bi (the transform coefficient of Z_B-Q_G relationship on single station) in natural rainfall, the change in the estimated area S and its domain, raingauge network distribution and its density will all lead to different values of μ_|S|. This is the uncertainty of precipitation estimation error during radar-raingauge integration. In this paper a new data pair quality control method is proposed called single station misestimate factor μ_i control method. This method will decrease the dispersion degree of A_Bi from A_BS and increase the polymerization degree, thus the regional precipitation absolute misestimate ratio μ_|S| is diminished accordingly. Experimental evaluations from Wuhan radars during four rain events give a value μ_|S| of less than 30%, which is a 3-5% decrease compared with none quality control.