Improving drought monitoring using climate models with bias-corrected under Gaussian mixture probability models

Abstract

Global climate models (GCMs) are extensively used to calculate standardized drought indices. However, inaccuracies in GCM simulations and uncertainties inherent in the standardization methodology limit the precision of drought evaluations. The objective of this research is to remove bias in GCMs for improving drought monitoring and assessment. Consequently, this article proposes a new framework for drought index under the ensemble of GCMs—Multi-Model Quantile Mapped Standardized Precipitation Index (MMQMSPI). In accordance of Standardized Precipitation Index (SPI), the second stage derives a new index by assessing the feasibility of parametric and nonparametric models during standardization. In the application, we used 18 GCMs from the Coupled Model Intercomparison Project Phase 6 (CMIP6) data of precipitation across 32 grid points within the Tibetan Plateau region. The comparative findings reveal that the integration of KCGMD is the most suitable choice compared to other best-fitted univariate distributions in both features of the proposed framework. In this research, we assess the implications of evaluating future patterns of drought for the years 2015–2100 using seven different time periods and three different future scenarios. Temporal behavior clearly shows monthly variations in the pattern of MMQMSPI, and these variations differ on each time scale, but a drastic change can be seen over the long term, i.e., extreme dry and wet conditions, with a higher probability in all scenarios.