Open AccessDetailed analysis of the error associated with the rainfall retrieved by the NOAA/NESDIS SSM/I algorithm: 1. Tropical oceanic rainfall
Author(s) -
Li Qihang,
Ferraro Ralph,
Grody Norman
Publication year - 1998
Publication title -
journal of geophysical research: atmospheres
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.67
H-Index - 298
eISSN - 2156-2202
pISSN - 0148-0227
DOI - 10.1029/98jd00680
Subject(s) - sampling (signal processing) , environmental science , remote sensing , satellite , autocorrelation , meteorology , atmosphere (unit) , special sensor microwave/imager , statistics , algorithm , mathematics , microwave , computer science , geology , physics , brightness temperature , telecommunications , filter (signal processing) , astronomy , computer vision
A general formulation has been established for the total error associated with passive microwave retrieval of space‐time rainfall. The total error consists of two independent components: the sampling error and the algorithm error. The former arises from the discrete sampling scheme of a polar‐orbiting satellite, and the later is the result of the imperfect algorithm used to convert the measured radiation to instantaneous rain rate. Rainfall data over the Tropical Ocean and Global Atmosphere/Coupled Ocean and Atmosphere Response Experiment (TOGA/COARE) region has been used to represent oceanic rainfall conditions to evaluate the error of the National Oceanic and Atmospheric Administration/National Environmental Satellite, Data, and Information Service (NOAA/ NESDIS) Special Sensor Microwave/Imager (SSM/I) rainfall algorithm. Rainfall statistics that affect the sampling error have been obtained from radar data, which include mean, variance, and autocorrelation of areal rain. The magnitude of the variance and the strength of the autocorrelation are quantitatively related to the size of the averaging area. Instantaneous measurement errors on the FOV scale and the large scale (2.5° box) are obtained by analyzing simultaneous observations of SSM/I and radar. Based on these parameter values, the sampling error, the algorithm error, and the total error are computed as a function of sampling frequency, size of averaging area, and other factors. This study has demonstrated the feasibility of a procedure to quantify the total retrieval error of space‐time rainfall by focusing on TOGA/COARE area (equivalent in size to a 2.5° box). The same procedure can potentially be applied to other grid boxes to produce a global error field if rain statistics and measurement errors are estimated for those boxes. The accuracy of the global error field will depend on the accuracy of the rain statistics and the measurement error values used in the individual grid boxes.