Mohammed et al. (2026) Performance evaluation of CMIP6 climate models for rainfall and erosivity in the thamirabharani basin, India

Identification

• Journal: Theoretical and Applied Climatology
• Year: 2026
• Date: 2026-03-09
• Authors: Jeilani Mohammed, Shashi Mesapam
• DOI: 10.1007/s00704-026-06071-8

Research Groups

• Department of Civil Engineering, National Institute of Technology, Warangal, Telangana, India.

Short Summary

This study evaluates the performance of 35 CMIP6 Global Climate Models (GCMs) for rainfall and erosivity in the Thamirabharani River Basin, India, identifying the best-performing models through a multi-criteria decision-making framework. The research projects significant increases in seasonal and annual rainfall (up to 93.3%) and rainfall erosivity (up to 71.7%) by the end of the century under high-emission scenarios, implying heightened risks for soil erosion and water resource management.

Objective

• To evaluate the performance of 35 CMIP6 Global Climate Models (GCMs) in simulating rainfall over the Thamirabharani River Basin, India, using a multi-criteria decision-making framework.
• To project future rainfall and rainfall erosivity under Shared Socioeconomic Pathways (SSPs) 4.5 and 8.5 scenarios using the identified best-performing GCMs.

Study Configuration

• Spatial Scale: Thamirabharani River Basin, India (approximately 4,400 km²), located between 77° 9′ 5″ and 78° 9′ 23″ E longitude and 8° 27′ 7″ and 9° 11′ 39″ N latitude, evaluated at nine specific grid points.
• Temporal Scale:
  • Historical observation/evaluation period: 1950–2014.
  • Future projection periods: Near future (2021–2040), Mid future (2041–2070), Far future (2071–2100).

Methodology and Data

• Models used: 35 CMIP6 Global Climate Models (GCMs) from the NASA Earth Exchange Global Daily Downscaled Projections (NEX-GDDP) dataset. The top five identified models are ACCESS-CM2, CanESM5, MIROC6, NorESM2-MM, and BCC-CSM2-MR.
  • Performance evaluation used six statistical indicators: Pearson Correlation Coefficient (CC), Nash–Sutcliffe Efficiency (NSE), Normalized Root Mean Square Error (NRMSE), Percent Bias (PBIAS), Kling–Gupta Efficiency (KGE), and Skill Score (SS).
  • Weighting schemes for indicators: Equal weight, Entropy-based weight, and Principal Component Analysis (PCA)-derived weight.
  • Multi-criteria decision-making (MCDM) techniques: PROMETHEE-II, TOPSIS, VIKOR, MOORA, and Compromise Programming (CP).
  • Group Decision Method (GDM) for integrated ranking.
  • Rainfall erosivity calculated using the Wischmeier and Smith (1978) equation: R = Σ(1.735 * 10^(1.5log10[Pi²/P] - 0.08188)), where R is in MJ·mm·ha⁻¹·h⁻¹·yr⁻¹, Pi is monthly rainfall (mm), and P is annual rainfall (mm).
• Data sources:
  • Observed rainfall data: India Meteorological Department (IMD) gridded daily rainfall dataset (0.25° × 0.25° resolution) for 1950–2014.
  • Future climate projections: NASA NEX-GDDP-CMIP6 dataset (0.25° × 0.25° resolution) for SSP4.5 and SSP8.5 scenarios.

Main Results

• The best-performing GCMs identified for the Thamirabharani River Basin are ACCESS-CM2, CanESM5, MIROC6, NorESM2-MM, and BCC-CSM2-MR, with ACCESS-CM2 consistently achieving the top rank.
• Projected Rainfall Increases (relative to 1950–2014 IMD baseline, for the far future 2071–2100):
  • Under SSP4.5: South-West Monsoon rainfall is projected to increase by 69.0%, North-East Monsoon by 53.2%, and annual rainfall by 52.4%.
  • Under SSP8.5: South-West Monsoon rainfall is projected to increase by 93.3%, North-East Monsoon by 80.6%, and annual rainfall by 72.8%.
• Average annual rainfall is projected to increase from approximately 900–950 mm (IMD period) to 1400–1500 mm under SSP4.5 (far future) and to exceed 1500 mm under SSP8.5 (far future).
• Projected Rainfall Erosivity Increases (relative to 1950–2014 IMD baseline, for the far future 2071–2100):
  • Under SSP4.5: Rainfall erosivity is projected to increase by 45.0%.
  • Under SSP8.5: Rainfall erosivity is projected to increase by 71.7%.
• Average rainfall erosivity increases from 808.18 MJ·mm·ha⁻¹·h⁻¹·yr⁻¹ (IMD period) to 1179.00 MJ·mm·ha⁻¹·h⁻¹·yr⁻¹ (SSP4.5 far future) and 1415.51 MJ·mm·ha⁻¹·h⁻¹·yr⁻¹ (SSP8.5 far future).

Contributions

• Developed a comprehensive, multi-level, and spatially explicit evaluation framework for 35 CMIP6 GCMs, addressing spatial heterogeneity often overlooked in basin-averaged assessments.
• Reduced subjectivity in GCM selection by integrating six statistical performance indicators with three weighting schemes (Entropy, Equal, PCA) and five Multi-Criteria Decision-Making (MCDM) techniques (PROMETHEE-II, TOPSIS, VIKOR, MOORA, CP).
• Employed a Group Decision-Making (GDM) framework to robustly integrate rankings from various methods, leading to a more reliable identification of best-performing GCMs for the Thamirabharani River Basin.
• Provided detailed projections of future rainfall and rainfall erosivity under SSP4.5 and SSP8.5 scenarios for the basin, highlighting increased risks of soil erosion and water resource challenges.

Funding

No specific funding was reported for this research.

Citation

@article{Mohammed2026Performance,
  author = {Mohammed, Jeilani and Mesapam, Shashi},
  title = {Performance evaluation of CMIP6 climate models for rainfall and erosivity in the thamirabharani basin, India},
  journal = {Theoretical and Applied Climatology},
  year = {2026},
  doi = {10.1007/s00704-026-06071-8},
  url = {https://doi.org/10.1007/s00704-026-06071-8}
}