What question did this study set out to answer?

February 20, 2026

Development of multi-model ensemble of CMIP6 GCMs using MCDM techniques for future climate projections in Mahanadi River Basin

Key Points

The aim is to rank Global Climate Models for climate projections in the Mahanadi River Basin using a multi-criteria approach.
Evaluated 13 CMIP6 GCMs based on historical data and observations.
Employed four statistical performance indicators for ranking: Correlation Coefficient, Normalized Root Mean Square Deviation, Absolute Normalized Mean Bias Deviation, and Skill Score.
Utilized Principal Component Analysis to determine weights for performance indicators.
Six MCDM methods were applied including Compromise Programming and Cooperative Game Theory.
Identified MPI-ESM1-2-HR, MPI-ESM1-2-LR, and NorESM2-LM as top performers across GCM rankings.
RAC ensemble provided the most reliable annual precipitation projection of 1315 mm, close to observed data.
Projected a 44.42% increase in annual precipitation and significant rises in maximum (13.08%) and minimum (20.93%) temperatures under SSP5-8.5.

Abstract

This study proposes a multi-criteria approach to rank Global Climate Models (GCMs) based on Coupled Model Intercomparison Project (CMIP6) data and develops effective Multi-Model Ensembles (MMEs) for climate projections in the Mahanadi River Basin. The ranking framework evaluates 13 CMIP6 GCMs by comparing their historical data against observations using four statistical performance indicators: the Correlation Coefficient, Normalized Root Mean Square Deviation, Absolute Normalized Mean Bias Deviation, and Skill Score. Principal Component Analysis (PCA) objectively determined the weights for these indicators. Six Multi-Criteria Decision-Making (MCDM) methods, such as Compromise Programming (CP), Weighted Average Method (WAM), PROMETHEE-II, VIKOR, Cooperative Game Theory (CGT), and TOPSIS, were employed to rank the GCMs. The results consistently highlighted MPI-ESM1-2-HR, MPI-ESM1-2-LR, and NorESM2-LM as top performers across the methods for the CMIP6 (NEX-GDDP) dataset. MMEs were built based on the performance (Representation of the Annual Cycle (RAC) and Upgraded Reliability Ensemble Averaging (UREA)). The RAC ensemble gives the most reliable result, with its historical annual precipitation (1315 mm) closest to the observed IMD data (1320 mm). The RAC ensemble projects the most extreme changes in the far future (2071 – 2100) under SSP5-8.5, showing a 44.42% increase in annual precipitation and significant rises in maximum (13.08%) and minimum (20.93%) temperatures. Seasonal analysis reveals that precipitation increases are primarily confined to the monsoon season (JJAS), with the most significant warming occurring in the pre-monsoon season (MAM). Spatially, coastal areas are projected to receive higher precipitation, whereas western and central areas may experience lower precipitation. This performance-based MME framework provides a structured and transparent basis for comparing climate projections, supporting water resources management, adaptation planning, and climate policy within the Mahanadi River Basin.

Mark Helpful

Bookmark

Relay