Predictive relationships between winter climate and hot season temperatures over India.

Climate change has impacted the nature and the severity of extreme temperatures over India in the recent past. In this study, the characterization in nature and change of hot season (March–May) temperatures across India's various geographical regions have been analysed over the last 69 years. Using canonical correlation analysis (CCA), we elucidate the relationships between various meteorological parameters from the preceding Indian winter (December–February) and the following hot season. A combination of detrended and raw data for CCA has revealed the asynchronous winter parameter's effects on hot season temperatures. The raw data explained the evolution of the relationship between the variables in the warming climate. In contrast, detrended data explains the predictive relationship between hot season temperature and other predictive variables which are not influenced by the trend in the data. The evolving nature of the global and Indian ocean as well as, El‐Niño Southern Oscillation, zonal circulation, and pressure systems, are evident in effecting Indian hot season temperatures. The post‐1980s warming is also tangible from the increasing strength of El Niño in the last few decades. The cross‐validated and mean square skill scores are moderate except for a few regions having an average correlation of ~.5 or more for the study periods. The minimum temperature has shown better skill than the maximum temperature, and the southern peninsular region has shown more predictability than the northern regions. Compared to raw data, detrended has a lesser prediction skill. The present study will help to predict the seasonal temperature along with the associated relationship and the warming effect post‐1980s. [ABSTRACT FROM AUTHOR]