A Study On Dynamic Correction For Radar-derived Quantitation Precipitation Estimation Based On The Optimal Z-I Relationship

The heavy rainfall is one of the main meteorological disasters in shaanxi province.In order to meet the requirement of cloud precipitation physics research,improve the level of the short-time forecasting andEarly-warning,and give full play to the role of theprecipitation forecast in industrial and agricultural production,weather modification,hydrological forecast,disaster prevention and reduction,etc,the research on high space-time resolution radar-derived quantitation precipitation estimation especially important.In this study,by using the radar mosaic of the seven CINRAD/CB weather radar and 1984 rain gauge in Shaanxi province from 2015 to 2016,dynamic methods for radar-derived quantitation precipitation estimation based on the optimal Z-I relationship has been studied.There are two dynamic methods established,the dynamic partitioning and hierarchical optimal Z-I relationship and the single-station optimal kalman fitering method.By choosing a wide range heavy rainfall process in Shaanxi province during 18-19 July 2016,the two methods above have been tested,simultaneously,comparative analyzed with three generic algorithm,the fixed Z-I relationship(Z=300I1.4),the single-station optimal Z-I relationship and partitioning and hierarchical optimal Z-I relationship.The mainresultsareasfollows:(1)Dynamic methods for radar-derived quantitation precipitation estimation are based on large scale,long time serial observation data statistics,so that can objectively reflect the climate difference,and the temporalevolutionof precipitation weather systems.(2)In the study of the optimal Z-I relationship,it's time saver and better accurate to use different interval of A and b than use the same interval.And in the study of the single-station optimal Z-I relationship,it's better to choose CTF1 as discriminant function than CTF2.(3)Under the influence of the Loess Plateau at Shaanxi northern and Qin-ba mountain district at Shaanxi southern,for the same rainfall intensity,the precipitation by south radars estimation is apparently less than by north radars,so partitioning and partitioning-stations have been used on building the optimal Z-I relationship.(4)Comparing with the fixed Z-I relationship,which is widely used presently,accuracy rates of the precipitation estimated by dynamic partitioning and hierarchical optimal Z-I relationship and the single-station optimal kalman fitering method have all increased significantly,and the latter has the highest accuracy.(5)For the precipitation under 1.5mm/h,the results all of five methods are generally large,but for the precipitation above 15mm/h,they are generally low.With the increase of precipitation grades,the Ts score of radar estimate precipitation decrease gradually.(6)With kalman filtering method having continuous correction,the duration of the correction takemore long time,the accuracy rates of the correction are more higher.