Detailed analysis of the error associated with the rainfall retrieved by the NOAA/NESDIS Special Sensor Microwave/Imager algorithm 2. Rainfall over land

Abstract

This study has investigated the error associated with the estimation of instantaneous areal rain rate over land by the NOAA/NESDIS Office of Research and Application's Special Sensor Microwave/Imager (SSM/I) algorithm. By comparing with ground‐based rain observations from the Oklahoma Mesonet rain gage network, the error inherent in the SSM/I‐based areal rain estimate is quantified. Since the ground‐based gauges use point measurements to estimate areal rain, the gauges do not make a perfect estimate and contain an associated error. A formulation has been developed by which the difference between the SSM/I‐ and Mesonet‐based estimates is divided into two parts: one due to the error in the Mesonet and the other due to the error in the SSM/I. A separate formulation has also been developed by which the error in Mesonet‐based rain measurement is quantified according to the gage density and the statistics of rain (variance and spatial correlation). A nine‐month data set of 15‐min rain accumulation for over 100 gages from the Mesonet has been used to obtain the rain statistics and to compare with the SSM/I estimate over various spatial scales. Results show that the error for the instantaneous SSM/I rain rate over 0.5, 1.0, and 2.5‐degree boxes are approximately 150%, 100%, and 70% of the mean areal rain rate, respectively.