Open AccessA new merged analysis of precipitation utilizing satellite and reanalysis data
Author(s) -
Sapiano M. R. P.,
Smith T. M.,
Arkin P. A.
Publication year - 2008
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 - 10.1029/2008jd010310
Subject(s) - precipitation , satellite , latitude , environmental science , climatology , data set , interpolation (computer graphics) , meteorology , special sensor microwave/imager , northern hemisphere , quantitative precipitation estimation , computer science , microwave , geology , geography , geodesy , animation , telecommunications , computer graphics (images) , artificial intelligence , brightness temperature , engineering , aerospace engineering
Many merged multi‐source global analyses of precipitation exist, including the Global Precipitation Climatology Project (GPCP) analysis and the CPC Merged Analysis of Precipitation. The multi‐source nature of these data sets allows them to use the most accurate type of inputs available to produce the best estimate of precipitation for any given place and time. However, studies have shown that the oceanic satellite estimates used in these data sets are less accurate at high latitudes when compared to reanalysis data. This study describes the Multi‐Source Analysis of Precipitation (MSAP), a new 2.5° gridded global analysis of precipitation from 1987 to 2002 using Optimum Interpolation (OI) based on the Defense Meteorological Satellite Program (DMSP) Special Sensor Microwave/Imager (SSM/I) and the forecast precipitation from the ERA‐40 reanalysis. The main goal of this new analysis is to produce a spatially consistent estimate using the same set of inputs over all regions and times rather than to have the lowest mean squared error. An advantage of the OI methodology is that it optimally merges the inputs based on pre‐defined weights and errors associated with the analysis that are directly estimated from the technique. Validation against other gridded data sets as well as tropical ocean and high‐latitude land gauges show that MSAP performs particularly well at high latitudes when compared to the satellite‐only part of GPCP. However, it contains negative biases in parts of the Northern Hemisphere because of the ERA‐40 data and large positive biases over tropical land areas due to issues with the SSM/I estimates. In the future, this new approach can be applied using precipitation estimates from the next generation reanalysis systems such as the JRA‐25, NASA's MERRA, and the ERA Interim reanalysis.