Yılmaz et al. (2025) Evaluation of NEX-GDDP-CMIP6 to simulate precipitation using multi-criteria decision-making analysis over Türkiye

Identification

• Journal: Theoretical and Applied Climatology
• Year: 2025
• Date: 2025-10-14
• Authors: Muhammet Yılmaz, Kadir Diler Alemdar, Fatih Tosunoğlu
• DOI: 10.1007/s00704-025-05805-4

Research Groups

• Erzurum Technical University, Erzurum, Turkey

Short Summary

This study evaluates the performance of 27 NEX-GDDP-CMIP6 models in simulating precipitation over 28 river basins in Türkiye using a novel multi-criteria decision-making approach, finding that a Best Multi-Model Ensemble (BMME) consistently outperforms individual models and simple Multi-Model Ensembles.

Objective

• To propose an innovative methodology for selecting a Best Multi-Model Ensemble (BMME) for precipitation simulation.
• To comprehensively evaluate and rank the performance of 27 individual NEX-GDDP-CMIP6 models, a simple Multi-Model Ensemble (MME), and the proposed BMME in simulating historical precipitation across 28 river basins in Türkiye.

Study Configuration

• Spatial Scale: Türkiye, covering 28 river basins with a total surface area of 783,562 km². Model data (NEX-GDDP-CMIP6) has a spatial resolution of 0.25° × 0.25° (approximately 25 km), while reference data (ERA5-Land) has a resolution of 0.1° × 0.1°.
• Temporal Scale: Historical period from 1960 to 2014 (55 years), using daily precipitation data.

Methodology and Data

• Models used:
  • 27 individual NEX-GDDP-CMIP6 General Circulation Models (GCMs).
  • Multi-Model Ensemble (MME) calculated as the simple arithmetic mean of all 27 NEX-GDDP-CMIP6 models.
  • Best Multi-Model Ensemble (BMME) determined using an integrated PROMETHEE (Preference Ranking Organization Method for Enrichment Evaluations) method with the K-means clustering Elbow Method.
  • Model evaluation metrics: Correlation Coefficient (CC), Percentage Bias (PBias), Normalized Root Mean Square Error (nRMSE), Kling-Gupta Efficiency (KGE), and Mean Absolute Error (MAE).
• Data sources:
  • Reference data: ERA5-Land dataset for precipitation, provided by the European Centre for Medium-Range Weather Forecasts (ECMWF).
  • Model data: NASA Earth Exchange Global Daily Downscaled Projections (NEX-GDDP-CMIP6) from CMIP6 models.

Main Results

• NEX-GDDP-CMIP6 models generally show strong capabilities in simulating daily precipitation over Türkiye, with a median Correlation Coefficient (CC) of 0.8.
• Models exhibit low performance in capturing precipitation characteristics in the northern parts of Türkiye (e.g., East Black Sea, West Black Sea, Coruh basins), often underestimating precipitation (e.g., PBias of -60.03% for East Black Sea).
• Individual models MPI-ESM1-2-LR, INM-CM5-0, IPSL-CM6A-LR, EC-Earth3-Veg-LR, GDFL-CM4-gr2, GDFL-CM4, BCC-CSM2-MR, FGOALS-g3, NorESM2-LM, MPI-ESM1-2-HR, and CanESM5 generally demonstrated high performance across basins.
• The PROMETHEE method identified MPI-ESM1-2-LR, MIROC-ES2L, INM-CM5-0, and EC-Earth3-Veg-LR as generally providing sufficient performance in simulating precipitation.
• The proposed BMME, determined by the PROMETHEE-based Elbow Method, consistently outperformed both individual NEX-GDDP-CMIP6 models and the simple MME across all 28 basins in Türkiye.
• The simple MME was identified as the second-best performing model in most basins (24 out of 28).
• NEX-GDDP-CMIP6 models generally exhibit a systematic negative bias (underestimation) in daily average precipitation across most basins, with exceptions in Gediz, Büyük Menderes, Akarçay, Asi, and Susurluk.
• Mean Absolute Error (MAE) values ranged from 0.33 mm/day (Konya Closed) to 2.66 mm/day (East Black Sea).
• Normalized Root Mean Square Error (nRMSE) values ranged from 0.12 mm/day (Asi) to 0.55 mm/day (East Black Sea).
• Kling-Gupta Efficiency (KGE) scores ranged from -0.12 (East Black Sea) to 0.82 (Asi).

Contributions

• Proposes a novel, integrated methodology for Best Multi-Model Ensemble (BMME) selection by combining the PROMETHEE method with the K-means clustering Elbow Method.
• Provides a systematic, objective, and comprehensive framework for ranking and selecting climate models, addressing the existing research gap in determining the optimal number of GCMs for BMME.
• Offers a regionally tailored model selection process for 28 river basins in Türkiye, enhancing the accuracy and reliability of precipitation simulations for climate change impact assessments.
• Demonstrates the superior performance of the proposed BMME over traditional Multi-Model Ensembles (MME) and individual NEX-GDDP-CMIP6 models in simulating historical daily precipitation.
• Provides valuable information for researchers and policymakers to reduce uncertainties in model selection for climate change studies in Türkiye.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Citation

@article{Yılmaz2025Evaluation,
  author = {Yılmaz, Muhammet and Alemdar, Kadir Diler and Tosunoğlu, Fatih},
  title = {Evaluation of NEX-GDDP-CMIP6 to simulate precipitation using multi-criteria decision-making analysis over Türkiye},
  journal = {Theoretical and Applied Climatology},
  year = {2025},
  doi = {10.1007/s00704-025-05805-4},
  url = {https://doi.org/10.1007/s00704-025-05805-4}
}