Qian et al. (2025) A Multi-Scale Comprehensive Evaluation for Nine Evapotranspiration Products Across Mainland China Under Extreme Climatic Conditions

⚠️ Warning: This summary was generated from the abstract only, as the full text was not available.

Identification

• Journal: Agriculture
• Year: 2025
• Date: 2025-09-14
• Authors: Long Qian, Lifeng Wu, Ning Dong, Tianjin Dai, Xingjiao Yu, Xuqian Bai, Qiliang Yang, Xiaogang Liu, Junying Chen, Zhitao Zhang
• DOI: 10.3390/agriculture15181945

Research Groups

Information not provided in the given text.

Short Summary

This study comprehensively evaluates nine evapotranspiration (ET) products across grid, basin, and site scales in China under varying climatic conditions from 2003 to 2014, finding that while products like GLEAM perform well, their accuracy significantly decreases under extreme conditions, a limitation largely overcome by integrating daily ET products into machine learning models.

Objective

• To evaluate the applicability and accuracy of nine different evapotranspiration (ET) products across grid, basin, and site scales in China from 2003 to 2014, under varying climatic conditions including extreme temperatures, vapor pressure deficit (VPD), and drought, and to explore methods for improving ET estimation.

Study Configuration

• Spatial Scale: Grid, basin (e.g., Songhua River, Hai River Basins), and site scales across China.
• Temporal Scale: 2003 to 2014.

Methodology and Data

• Models used:
  • Evapotranspiration (ET) products evaluated: MODIS/Terra Net Evapotranspiration 8-Day L4 Global 500m SIN Grid (MOD16A2), Global Land Evaporation Amsterdam Model V4.2a (GLEAM), Synthesized Global Actual Evapotranspiration Dataset (Syn), Reliability Ensemble Averaging (REA) product, Penman–Monteith–Leuning Evapotranspiration V2 (PMLv2).
  • Evaluation methods: Three-cornered hat (TCH) method, water-balance-based ET (WB-ET).
  • Improvement models: Random Forest (RF) machine learning model.
• Data sources:
  • Nine global/regional ET products.
  • Water-balance-based ET (WB-ET) for basin-scale comparison.
  • Daily ET products used as inputs for machine learning models.
  • Climatic conditions data (extreme temperatures, vapor pressure deficit, drought).

Main Results

• At the grid scale, most ET products (except MOD16A2) showed high consistency, with GLEAM exhibiting the highest comparability. TCH analysis indicated GLEAM and Syn had low uncertainties in multiple basins, while REA and PMLv2 had the smallest uncertainties in the Songhua River and Hai River Basins.
• At the basin scale, ET products closely aligned with water-balance-based ET (WB-ET), with GLEAM achieving the smallest root mean square error (RMSE) of 22.94 mm/month.
• At the site scale, ET product accuracy significantly decreased under extreme climatic conditions. The coefficient of determination (R2) dropped from approximately 0.60 to below 0.30, and the mean absolute error (MAE) increased by 110.30% under extreme high temperatures and 101.40% under extreme high vapor pressure deficit (VPD). Drought conditions caused slight instability, with MAE increasing by approximately 12.00–40.00%.
• Using a small number of daily ET products as inputs for machine learning models, such as Random Forest (RF), greatly improved ET estimation, achieving an R2 of 0.91 overall and 0.81 under extreme conditions. GLEAM was identified as the most important product for RF in ET estimation.

Contributions

• Provides a comprehensive, multi-scale, and multi-climatic condition evaluation of nine widely used ET products in China.
• Quantifies the significant decrease in ET product accuracy under extreme climatic events (high temperatures, high VPD, drought) at the site scale.
• Demonstrates the effectiveness of machine learning models (Random Forest) in substantially improving ET estimation, even under extreme conditions, by leveraging existing daily ET products.
• Offers essential guidance for selecting and improving ET products to enhance agricultural water-use efficiency and sustainable irrigation practices.

Funding

Information not provided in the given text.

Citation

@article{Qian2025MultiScale,
  author = {Qian, Long and Wu, Lifeng and Dong, Ning and Dai, Tianjin and Yu, Xingjiao and Bai, Xuqian and Yang, Qiliang and Liu, Xiaogang and Chen, Junying and Zhang, Zhitao},
  title = {A Multi-Scale Comprehensive Evaluation for Nine Evapotranspiration Products Across Mainland China Under Extreme Climatic Conditions},
  journal = {Agriculture},
  year = {2025},
  doi = {10.3390/agriculture15181945},
  url = {https://doi.org/10.3390/agriculture15181945}
}