Investigating modification coefficients of TRMM rainfall network in Mazandaran province

Precipitation is one of the basic elements of Hydrometeorological studies. Due to the limited number of rain-gauging stations and the lack of meteorological stations in highland area, the use of satellite rainfall products as an effective tool in predicting the regional spatial distribution of precipitation data has attracted the attention of researchers. But its disadvantage is the high estimation error of rainfall estimation, which requires accuracy evaluating before using. For this purpose, based on the rainfall data of 21 meteorological stations of Mazandaran province from 1998-2018, the performance of TRMM data and two interpolation methods of kriging and inverse distance weighting were investigated. The results showed that contrary to the proper correlation of TRMM data with observed data, this network has a high bias error. For example, in the case of annual rainfall data, the network bias error reached 70 mm, which is approximately 30% of the annual error. The modification coefficients of TRMM data were found between 0 and 2, that most of the coefficients were higher than one, which indicates the TRMM data in Mazandaran mostly has an underestimation error. The comparison of modified TRMM with the interpolation methods showed that using the modified network instead of interpolation methods in estimation of rainfall data reduces the mean square error of annual rainfall from 237 to 186 mm.