TY - JOUR
T1 - Extrapolability improvement of machine learning-based evapotranspiration models via domain-adversarial neural networks
AU - Shi, Haiyang
AU - Cai, Ximing
N1 - This research has been supported by the Dashboard for Agricultural Water use and Nutrient management (DAWN) project (Grant 2020-68012-31674) funded by the U.S. Department of Agriculture National Institute of Food and Agriculture. We thank Dr. Istv\u00E1n G\u00E1bor Hatvani and another anonymous reviewer for their insightful suggestions for improvement of this manuscript.
PY - 2025/4
Y1 - 2025/4
N2 - Machine learning-based evapotranspiration (ET) models capture complex nonlinear relationships but struggle with global extrapolation due to unbalanced data distribution, limiting accurate ET assessments crucial for understanding water and energy cycles. This study used Domain-Adversarial Neural Networks (DANN) to improve the geographical adaptability of ET models by mitigating site-level distributional discrepancies. DANN significantly enhanced ET prediction accuracy, with a median Kling-Gupta Efficiency (KGE) increase of 0.27 (p < 0.001) and with a range from 0.06 to 0.58 for the middle 90% values compared to the traditional Leave-One-Out (LOO) method. DANN proves particularly effective for isolated sites and biome transition zones, reducing errors and avoiding low-accuracy predictions. By leveraging data from resource-rich areas, DANN strengthens the reliability of global-scale ET products, especially in ungauged regions. Future evaluations and improvements are necessary, such as using additional accuracy metrics beyond KGE and focusing on sites located at the intersection of several climate types and sites with unique soil-vegetation-atmosphere processes. This study demonstrates the potential of domain adaptation techniques to enhance the generalization and extrapolation capabilities of machine learning in hydrological predictions.
AB - Machine learning-based evapotranspiration (ET) models capture complex nonlinear relationships but struggle with global extrapolation due to unbalanced data distribution, limiting accurate ET assessments crucial for understanding water and energy cycles. This study used Domain-Adversarial Neural Networks (DANN) to improve the geographical adaptability of ET models by mitigating site-level distributional discrepancies. DANN significantly enhanced ET prediction accuracy, with a median Kling-Gupta Efficiency (KGE) increase of 0.27 (p < 0.001) and with a range from 0.06 to 0.58 for the middle 90% values compared to the traditional Leave-One-Out (LOO) method. DANN proves particularly effective for isolated sites and biome transition zones, reducing errors and avoiding low-accuracy predictions. By leveraging data from resource-rich areas, DANN strengthens the reliability of global-scale ET products, especially in ungauged regions. Future evaluations and improvements are necessary, such as using additional accuracy metrics beyond KGE and focusing on sites located at the intersection of several climate types and sites with unique soil-vegetation-atmosphere processes. This study demonstrates the potential of domain adaptation techniques to enhance the generalization and extrapolation capabilities of machine learning in hydrological predictions.
UR - http://www.scopus.com/inward/record.url?scp=85218167520&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85218167520&partnerID=8YFLogxK
U2 - 10.1016/j.envsoft.2025.106383
DO - 10.1016/j.envsoft.2025.106383
M3 - Article
AN - SCOPUS:85218167520
SN - 1364-8152
VL - 187
JO - Environmental Modelling and Software
JF - Environmental Modelling and Software
M1 - 106383
ER -