Dubey et al. (2025) Entropy theory-based performance appraisal of CMIP6 climate models in regional drought simulation over the Indus River basin: a multifactorial investigation

Identification

• Journal: Natural Hazards
• Year: 2025
• Date: 2025-10-17
• Authors: Amit Dubey, Deepak Swami, V. Gupta, Nitin Joshi
• DOI: 10.1007/s11069-025-07684-4

Research Groups

• School of Civil and Environmental Engineering, Indian Institute of Technology Mandi, Mandi, Himachal Pradesh, India
• Department of Civil Engineering, Indian Institute of Technology Jammu, Jammu, Jammu and Kashmir, India

Short Summary

This study assessed the historical performance of 16 CMIP6 climate models in simulating regional drought and associated meteorological variables over the Indian Indus River basin (1979–2014) using an entropy-based approach, identifying MIROC6, MPI-ESM1-2-HR, and NorESM2-LM as the most suitable models for future climate projections.

Objective

• To assess the performance of different CMIP6 models in simulating historical drought against observations over the Indian extent of the Indus River basin during 1979–2014.
• To identify the best performing CMIP6 models for simulating droughts and associated meteorological factors (precipitation, minimum and maximum temperature, potential evapotranspiration) using an entropy-based feature selection approach, for use in future climate extreme projections.

Study Configuration

• Spatial Scale: Indian extent of the Indus River basin (29°8’–36°59’ N, 72°28’–79°39’ E), at a spatial resolution of 0.12° × 0.12° (1956 grids).
• Temporal Scale: 1979–2014 (36 years).

Methodology and Data

• Models used: 16 Coupled Model Intercomparison Project Phase 6 (CMIP6) General Circulation Models (GCMs).
• Data sources:
  • Reference Data: Indian Monsoon Data Assimilation and Analysis (IMDAA) dataset from the National Centre for Medium Range Weather Forecasting (NCMRWF), Government of India, providing daily gridded precipitation, minimum, and maximum temperature at 0.12° × 0.12° resolution.
  • Simulated Data: Historical simulations of daily gridded precipitation, minimum, and maximum temperature from 16 CMIP6 GCMs, acquired from the Earth System Grid Federation (ESGF) portal.
  • Drought Index: Standardized Precipitation Evapotranspiration Index (SPEI) at a 3-month timescale (SPEI-3) was used for drought quantification.
  • Potential Evapotranspiration (PET) Estimation: Hargreaves-Samani (HS) approach.
  • Model Ranking: Symmetrical Uncertainty (SU), an entropy-based feature selection approach, was employed to rank GCMs based on their similarity to reference data. A 50th-percentile criterion was used for final model selection.

Main Results

• Spatial and density distributions of model biases exhibited high heterogeneity across different parameters.
• GCMs largely underestimated precipitation attributes, with bias distributions showing a negative skewness (higher concentration of negative biases).
• Minimum temperature was generally overestimated by GCMs, while maximum temperature was largely underestimated. Historical temperature simulations showed relatively lower biases compared to precipitation.
• Top-ranking GCMs for individual variables based on Symmetrical Uncertainty (SU) were:
  • Precipitation: INM-CM4-8, ACCESS-ESM1-5, NorESM2-LM, and INM-CM5-0.
  • Minimum Temperature: NESM3, ACCESS-ESM1-5, ACCESS-CM2, and MPI-ESM1-2-HR.
  • Maximum Temperature: ACCESS-ESM1-5, NESM3, MPI-ESM1-2-HR, and MPI-ESM1-2-LR.
  • SPEI (drought characteristics): KACE-1-0-G, NorESM2-LM, INM-CM5-0, and NorESM2-MM.
• A trade-off analysis using a 50th-percentile criterion identified MIROC6, MPI-ESM1-2-HR, and NorESM2-LM as the most suitable GCMs for effectively reproducing all considered parameters (precipitation, minimum and maximum temperature, PET, and SPEI) over the Indus River basin.

Contributions

• This study provides a detailed investigation into the capability of CMIP6 models to simulate precipitation, minimum and maximum temperature, potential evapotranspiration (PET), and drought (SPEI-3) over the complex topography of the Indian extent of the Indus River basin.
• It introduces a novel approach by comprehending drought simulation capability at a fine spatial resolution of 0.12 degrees, an aspect previously under-addressed in the region.
• It identifies a robust ensemble of GCMs (MIROC6, MPI-ESM1-2-HR, and NorESM2-LM) that perform well across all critical meteorological and drought-related variables, offering a solid foundation for reliable future climate and drought projections in the Indus River basin and similar complex terrains.

Funding

The authors declare that this research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Citation

@article{Dubey2025Entropy,
  author = {Dubey, Amit and Swami, Deepak and Gupta, V. and Joshi, Nitin},
  title = {Entropy theory-based performance appraisal of CMIP6 climate models in regional drought simulation over the Indus River basin: a multifactorial investigation},
  journal = {Natural Hazards},
  year = {2025},
  doi = {10.1007/s11069-025-07684-4},
  url = {https://doi.org/10.1007/s11069-025-07684-4}
}