Performance evaluation of Google Earth Engine based precipitation
datasets under different climatic zones over India
Abstract
Satellite-based as well as reanalysis-based datasets are widely
available and useful in detecting spatial and temporal variability of
rainfall at a finer resolution. These products have been widely used in
weather forecasting, hydrological and climate studies. However, the
accuracy of these satellite products varies spatially and across
different datasets. In this study, the accuracy of five satellite-based
precipitation products with different spatial resolutions i.e., CHIRPS,
ERA5, TRMM, GPM, and Terra Clim available on Google Earth Engine (GEE)
were compared with India Meteorological Department (IMD) grided data in
six climate zones in India. The statistics such as RMSE, R
2, MSE, and PBIAS were computed. It was observed that
the performance of each product varies in different climatic zones. The
GPM was observed to have high accuracy in arid, semi-arid, and tropical
wet zones. TRMM showed a good match in tropical wet & dry, tropical
wet, and semi-arid zones. Terra Clim and ERA5 showed high accuracy in
humid subtropical and montane regions respectively. It was also observed
that CHRIPS was found to be least suitable in all the climate zones
across India. The findings from the present studies will serve as a
guiding document for the researcher to select appropriate datasets for
different applications such as drought monitoring, precipitation
anomaly, hydrological models, or other related studies in India.