Shukla et al. (2025) Ranking and comparison of temperature-, mass transfer- and radiation based daily reference evapotranspiration models by using compromise programming index (CPI) and global performance indicator (GPI)

Identification

• Journal: Theoretical and Applied Climatology
• Year: 2025
• Date: 2025-12-16
• Authors: Abhishek Shukla, Arvind Singh Tomar, Kusum Pandey, Dinesh Kumar Vishwakarma, Siham Acharki, Ali Raza, Makrand Dhyani, Ozgur Kisi, Ahmed Z. Dewidar, Ahmed Al-Othman, Mohamed A. Mattar
• DOI: 10.1007/s00704-025-05910-4

Research Groups

• Department of Irrigation and Drainage Engineering, College of Technology, G. B. Pant University of Agriculture and Technology, Udham Singh Nagar, Pantnagar, Uttarakhand, India
• G. B. Pant National Institute of Himalayan Environment, Garhwal Regional Centre, Srinagar, Uttarakhand, India
• Division of Irrigation & Drainage Engineering, College of Agricultural Engineering and Technology, Sher-e-Kashmir University of Agricultural Sciences and Technology of Kashmir, Shalimar, Srinagar, Jammu and Kashmir, India
• International Water Resource Institute (IWRI), Mohammed VI Polytechnic University (UM6P), Benguerir, Morocco
• School of Geography, Nanjing Normal University, Nanjing, China
• Graphic Era Hill University, Dehradun, Uttarakhand, India
• Department of Civil Engineering, University of Applied Sciences, Lübeck, Germany
• Department of Civil Engineering, Ilia State University, Tbilisi, Georgia
• School of Civil, Environmental and Architectural Engineering, Korea University, Seoul, South Korea
• Prince Sultan Bin Abdulaziz International Prize for Water Chair, Prince Sultan Institute for Environmental, Water and Desert Research, King Saud University, Riyadh, Saudi Arabia
• Department of Agricultural Engineering, College of Food and Agriculture Sciences, King Saud University, Riyadh, Saudi Arabia

Short Summary

This study evaluates 30 empirical reference evapotranspiration (ET o ) models against the FAO56 Penman–Monteith method in two contrasting Indian agro-climatic zones using multi-criteria decision-making tools, finding that model performance is highly region-specific and requires localized validation.

Objective

• To evaluate temperature-based, mass transfer-based, and radiation-based ET o models at daily scales against the FAO56-Penman–Monteith (FAO56-PM) benchmark in two contrasting Indian climates (humid subtropical Dehradun and semi-arid Ludhiana).
• To apply Compromise Programming Index (CPI) and Global Performance Indicator (GPI) for systematic model ranking and to address inconsistencies arising from multiple statistical performance indices.
• To identify robust, practical, and data-efficient models suitable for reliable ET o estimation in the Indian region, particularly under data-scarce conditions.

Study Configuration

• Spatial Scale: Two distinct agro-climatic zones in India: Dehradun, Uttarakhand (humid subtropical climate, 78.01295039° E, 30.34600616° N, 516.50 m above mean sea level) and Ludhiana, Punjab (semi-arid climate, 75.812336° E, 30.900833° N, 247 m above mean sea level).
• Temporal Scale: Daily basis over a 31-year period (1989–2019) for Dehradun and a 30-year period (1990–2019) for Ludhiana.

Methodology and Data

• Models used:
  • Benchmark Model: FAO56 Penman–Monteith (FAO56-PM).
  • Empirical Models Evaluated: 30 models, comprising 10 temperature-based, 10 mass transfer-based, and 10 radiation-based equations.
  • Evaluation Metrics: Mean Bias Error (MBE), Root Mean Square Error (RMSE), Normalized Root Mean Square Error (NRMSE), Percent Bias (PBIAS), Kling-Gupta Efficiency (KGE), Index of Agreement (IA), and Coefficient of Determination (R²).
  • Multi-criteria Decision-Making Tools: Compromise Programming Index (CPI) and Global Performance Indicator (GPI) for integrated model ranking, validated using a bootstrap resampling approach.
• Data sources:
  • Daily meteorological data for Dehradun from the ICAR-Central Soil and Water Conservation Research and Training Institute.
  • Daily meteorological data for Ludhiana from the India Meteorological Department (IMD), Pune.
  • Key parameters included maximum and minimum temperature (°C), maximum and minimum relative humidity (%), wind speed at 2 m height (m/s), sunshine hours (h/day), and evaporation (mm/day).

Main Results

• Dehradun (Humid Subtropical):
  • CPI Ranking: DORJ (temperature-based) ranked first (CPI = 0.00), followed by MAKK (radiation-based, CPI = 0.778) and PRTA (radiation-based, CPI = 0.891).
  • GPI Ranking: JORI (radiation-based) ranked first (GPI = 0.712), followed by PRTA (radiation-based, GPI = 0.711) and IRRN (radiation-based, GPI = 0.704).
• Ludhiana (Semi-Arid):
  • CPI Ranking: HSM-3 (temperature-based) ranked first (CPI = 0.578), followed by HSM-1 (temperature-based, CPI = 0.708) and HSM-2 (temperature-based, CPI = 0.729).
  • GPI Ranking: HSM-1 (temperature-based) ranked first (GPI = 0.691), followed by HSM-2 (temperature-based, GPI = 0.680) and HSM-3 (temperature-based, GPI = 0.672).
• General Performance: Temperature- and radiation-based models consistently outperformed mass transfer-based equations across both regions.
• Weakest Performers: The TURC, SAIF, and BERT models consistently ranked among the weakest across both climatic conditions and evaluation methods.
• Climatic Dependence: No single empirical method consistently outperformed across diverse climatic settings, highlighting the necessity of localized calibration and validation.
• Bootstrap Validation: Bootstrapped rankings confirmed the stability and robustness of the CPI and GPI results, with top-performing models (DORJ, HSMs, JORI, PRTA, MAKK) and consistently weak performers (TURC, BRWE, DALT, MEYE) maintaining their relative ranks.

Contributions

• This study provides a comprehensive, multi-criteria evaluation framework for ranking empirical ET o models, integrating statistical indices with Compromise Programming Index (CPI) and Global Performance Indicator (GPI) to overcome limitations of single-criterion assessments.
• It identifies robust, computationally efficient, and data-efficient alternatives to the FAO56-PM model for specific humid subtropical and semi-arid climates in India, crucial for data-scarce regions.
• The research emphasizes the critical need for localized calibration and validation of ET o models due to their strong region-specific performance, offering valuable guidance for improved irrigation scheduling and water resource management.

Funding

• Ongoing Research Funding program - Research Chairs (ORF-RC-2025-5514), King Saud University, Riyadh, Saudi Arabia.

Citation

@article{Shukla2025Ranking,
  author = {Shukla, Abhishek and Tomar, Arvind Singh and Pandey, Kusum and Vishwakarma, Dinesh Kumar and Acharki, Siham and Raza, Ali and Dhyani, Makrand and Kisi, Ozgur and Dewidar, Ahmed Z. and Al-Othman, Ahmed and Mattar, Mohamed A.},
  title = {Ranking and comparison of temperature-, mass transfer- and radiation based daily reference evapotranspiration models by using compromise programming index (CPI) and global performance indicator (GPI)},
  journal = {Theoretical and Applied Climatology},
  year = {2025},
  doi = {10.1007/s00704-025-05910-4},
  url = {https://doi.org/10.1007/s00704-025-05910-4}
}