Synthesis of radar rainfall data

Abstract

A method of generating synthetic radar‐rainfall data is described. The data are generated by imposing random noise on a given, high‐quality radar‐rainfall field. Certain conditions are imposed on the resultant rainfall field so that the noise parameters are prespecified. The conditions pertain to the second order statistics of the generated rainfall fields: the mean, the variance, the correlation, and the variance of the logarithmic ratio of the resultant field to the original field. Accuracy of the generation method is evaluated from implementing a test case using Global Atmospheric Research Program Atlantic Tropical Experiment radar data. The method can be used in a number of different, mainly hydrologic, applications. These include validation of radar and rain gage data merging procedures, testing of various methods for computation of mean areal precipitation, and sensitivity analysis of rainfall‐runoff models. Copyright 1985 by the American Geophysical Union.