A Monte Carlo Method for Radar-Derived Surface Rainfall Estimation Based on Disdrometer Data in Locations without Gauge Data

A technique is presented for using radar reflectivity in areas lacking rain gauge data to calculate a likely range for surface rainfall rate based on disdrometer data collected in a similar atmospheric environment. Instead of computing a deterministic Z-R relationship, probability distribution functions (PDFs) of rain rate that change as a function of reflectivity are constructed. The PDFs account for the variety of drop size distributions that may be found within echoes of equal reflectivity thereby providing a measure of the uncertainty in the retrieval. Random measurement error in the radar reflectivity measurement is incorporated into the PDFs. Radar data along the lowest available scan in a volume is extrapolated downward to within 100 m of the ground/ocean in order to provide accurate rain rates at the surface. The extrapolation is achieved using range-height indicator scan data to generate another set of PDFs for surface reflectivity as a function of reflectivity at the height of the radar data. Given a reflectivity data point and its height, the corresponding surface rain rate is captured by a two-dimensional PDF that combines uncertainty associated with instrument error, the Z-R relationship, and downward extrapolation of the reflectivity value to the surface. The method was applied to radar data collected by the S-PolKa radar system during the Dynamics of the Madden-Julian Oscillation (DYNAMO) field campaign and validated using disdrometer data near the radar.