Development of a new hybrid ensemble method for accurate characterization of future drought using multiple global climate models

Drought as a natural disaster, can have devastating effects on various sectors. This study proposes a new weighting scheme called weighted aggregation (WA) and introduces the Multi-model weighted drought severity index (MMWDSI) as an improved indicator for drought assessment. The methodology of MMWDSI involves ensemble modeling using the WA scheme, incorporating the K-components Gaussian mixture model (K-CGMM) for appropriate distribution fitting. It employs a multi-stage statistical procedure that considers point-to-point variations and the past performance of climate models through the Taylor Skill Score in the initial stage. Subsequent stages involve linear models and K-CGMM for prediction and standardization. Similar to other standardized drought indices, the proposed index allows for inferring the probabilistic behavior of extreme events, such as extreme drought or extreme wet conditions, and assessing trends using various statistical techniques. For the application of the index, historical precipitation data from 1961 to 2014 was utilized from 32 grid points on the Tibetan Plateau as the reference dataset. Additionally, simulations from 18 models of the Coupled Model Intercomparison Project phase 6, both past and future, were employed for the estimation procedure. The findings demonstrate that the developed weighting scheme surpasses the Equal Weighted Averaging approach. In conclusion, the MMWDSI approach proves to be a flexible and effective method that enhances accuracy in drought monitoring.