Non-stationary multivariate bias-corrected CMIP6 climate projections of daily precipitation and temperature over India [Part2: Temperature Data]

Systematic biases and the coarse spatial resolution of global climate model (GCM) outputs limit their applicability for regional-scale studies. To address this, we present a daily bias-corrected climate projection dataset for India developed using a non-stationary multivariate bias correction method, rank resampling for distributions and dependences (R2D2). Unlike traditional univariate approaches (e.g., QM, CDF-t), R2D2 corrects model biases while preserving inter-variable dependence. Using precipitation and temperature outputs from 13 CMIP6 GCMs, we generate a 0.25° resolution dataset covering the historical period (1951–2014) and future projections (2015–2100) under four SSP scenarios (SSP1-2.6, SSP2-4.5, SSP3-7.0, SSP5-8.5). The dataset is evaluated against observations using multivariate statistical properties that capture dependence structures, including Pearson correlation, Kendall’s tau, the Clausius–Clapeyron relationship between precipitation and temperature, and the frequency of Compound Hot and Dry Events (CDHEs). This resource supports diverse applications in climate impact research, particularly studies of precipitation–temperature dependence, compound extremes, hydrological modelling, and drought risk assessment.