Improved rainfall and temperature satellite dataset in areas with scarce weather stations data: case study in Ancash, Peru

Abstract

Rainfall and temperature variables play an important role in understanding meteorology at global and regional scales. However, the availability of meteorological information in areas of complex topography is difficult, as the density of weather stations is often very low. In this study, we focused on improving existing satellite products for these areas, using Tropical Rainfall Measuring Mission (TRMM) and Global Precipitation Measurement (GPM) data for rainfall and Modern Era Retrospective Analysis for Research and Applications Version 2 (MERRA-2) data for air temperature. Our objective was to propose a model that improves the accuracy and correlation of satellite data with observed data on a monthly scale during 2012-2017. The improvement of rainfall satellite data was performed using 4 regions: region 1 Santa (R1Sn), region 2 Marañón (R2Mr), region 3 Pativilca (R3Pt) and region 4 Pacific (R4Pc). For temperature, a model based on the use of the slope obtained between temperature and altitude data was used. In addition, the reliability of the TRMM, GPM and MERRA-2 data was analyzed based on the ratio of the mean square error, PBIAS, Nash-Sutcliffe efficiency (NSE) and correlation coefficient. The final products obtained from the model for temperature are reliable with R2 ranging from 0.72 to 0.95 for the months of February and August respectively, while the improved rainfall products obtained are shown to be acceptable (NSE≥0.6) for the regions R1Sn, R2Mr and R3Pt. However, in R4Pc it is unacceptable (NSE<0.4), reflecting that the additive model is not suitable in regions with low rainfall values.