Validation of Surface Temperature Derived From MERRA-2 Reanalysis Against IMD Gridded Data Set Over India

Abstract

The first detailed validation of maximum temperature of Modern-Era Retrospective analysis for Research and Application Version 2 (TMERRA-2) against Indian Meteorological Department (TIMD) has been carried out for 35 years (1981–2015) over India. For this purpose, India has been divided into seven different zones, i.e Western Himalaya (WH), Northwest, North Central, Northeast (NE), West Peninsula India, East Peninsula India, and South Peninsula India. The descriptive statistics and correlation between TMERRA-2 and TIMD have been determined for monthly, seasonal, and annual basis. A significant correlation (>0.9) has been found for monthly TMERRA-2 and TIMD with a root-mean-square error value closer to 1 except for WH where a high root-mean-square error value of 18.2 is obtained. Seasonal analysis also indicates a significant correlation for all the zones except for WH and NE with a correlation value of <0.3 during monsoon season; this may be due to sparse network, cold climate, and heterogeneity due to topography. Percent bias indicates that TMERRA-2 generally overestimates the TIMD monthly observations for all the zones, that is, Northwest, North Central, NE, West Peninsula India, East Peninsula India, and South Peninsula India by 4.1%, 2.4%, 1.6%, 0.5%, 0.2%, and 0.8%, respectively, except WH where an underestimation (−82.5%) is determined. Thus, after calibration, MERRA-2 Reanalysis maximum temperature may be used for further study of extreme weather events. © 2019. The Authors.