Abstract
For intercomparison purposes and validation of different satellite‐based precipitation estimates in the framework of the first WetNet Precipitation Intercomparison Project (PIP‐1), two surface precipitation data sets have been used. The Comprehensive Pacific Rainfall Data Base (CPRDB) has been assimilated specifically for the calibration and verification of satellite rainfall algorithms over parts of the tropical Pacific ocean. The CPRDB consists of daily rain gauge data from more than 250 sites in the tropical Pacific, most of which are atolls and low‐lying islands. The data set evaluated by the Global Precipitation Climatology Centre (GPCC), a central element of the World Climate Research Programme (WCRP) Global Precipitation Climatology Project (GPCP), provides areal mean monthly precipitation totals on a 2.5° grid on a global scale for climate research and the verification of climate models. Although the merged GPCC data set is global in extent, it is based on an objective analysis of rain gauge measurements only over land, whereas over the tropical to mid‐latitude oceans estimates derived from satellite images are included and the remaining gaps are filled with results of numerical weather prediction models.
For the verification and validation of satellite‐based precipitation estimates with conventional measurements, the CPRDB can be used over parts of the Pacific Ocean, whereas the preliminary GPCC data set, which is based on about 6700 stations world‐wide, can serve as a reference over land. After a description of the CPRDB its usefulness for the PIP‐1 purposes is demonstrated. For the GPCC data set, the station bias, processing of the data, the objective analysis method for the calculation of the areal mean precipitation totals, as well as error estimates of the gridded results are described. Intercomparisons with precipitation estimates based on satellite data used in the GPCP, which are carried out separately for land and ocean areas, show in part substantial differences. They indicate both problems in estimating precipitation from satellite remote sensing, and uncertainties in the GPCC analysis of the conventional measurements due to the sparse data density in some regions.
