Physics-augmented deep learning models for improving evapotranspiration estimation in global land regions

Existing evapotranspiration (ET) estimation models face inherent limitations when relying solely on physics-based or data-driven paradigms. To address this issue, we propose three data-physics hybrid modeling methods for improving instantaneous ET estimation in this study. A Physics-Data Learning (P...

Full description

Saved in:
Bibliographic Details
Main Authors: Binrui Liu, Xinguang He, Wenkai Lyu, Lizhi Tao
Format: Article
Language:English
Published: Elsevier 2025-08-01
Series:Agricultural Water Management
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S0378377425003488
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849254141301358592
author Binrui Liu
Xinguang He
Wenkai Lyu
Lizhi Tao
author_facet Binrui Liu
Xinguang He
Wenkai Lyu
Lizhi Tao
author_sort Binrui Liu
collection DOAJ
description Existing evapotranspiration (ET) estimation models face inherent limitations when relying solely on physics-based or data-driven paradigms. To address this issue, we propose three data-physics hybrid modeling methods for improving instantaneous ET estimation in this study. A Physics-Data Learning (PDL) model is first formed by adding a complementary physical variable generated by Penman–Monteith (PM) equation to a deep learning (DL) model along with driving variables to regress latent heat flux. Building on the PDL, a Physics-Augmented Learning (PAL) model is then formulated by introducing a physics-augmented term into the loss function. Finally, a Physics-Augmented Residual Learning (PARL) model is developed by using the residual learning technique to deeply integrate the PM and pure DL baseline models. Using the FLUXNET dataset, three proposed models are compared with the baselines on ten vegetation types (VTs) across the globe. The results show that all proposed models improve the accuracy of two baselines and reduce the uncertainty of pure DL to different extents. Among them, the PARL achieves the highest accuracy and robustness, with NSE (RMSE) ranging from 0.71–0.82 (22.40–43.14 W/m2) across ten VTs. The PAL ranks second and effectively mitigates the PDL’s sensitivity to imperfect physical knowledge. Although three proposed models show better extrapolation ability than the pure DL under conditions of limited data, the PARL stands out for its superior generalization under four created extreme climate scenarios. These results confirm the potential of data-physics hybrid modeling in ET estimation, which is conducive to supporting efficient irrigation water management.
format Article
id doaj-art-7bfacda45f2c4a24a2a73489f85515b2
institution Kabale University
issn 1873-2283
language English
publishDate 2025-08-01
publisher Elsevier
record_format Article
series Agricultural Water Management
spelling doaj-art-7bfacda45f2c4a24a2a73489f85515b22025-08-20T03:56:08ZengElsevierAgricultural Water Management1873-22832025-08-0131710963410.1016/j.agwat.2025.109634Physics-augmented deep learning models for improving evapotranspiration estimation in global land regionsBinrui Liu0Xinguang He1Wenkai Lyu2Lizhi Tao3Key Laboratory of Geospatial Big Data Mining and Application, College of Geographic Sciences, Hunan Normal University, Changsha 410081, China; National Centre for Groundwater Research and Training, and College of Science and Engineering, Flinders University, Adelaide, SA 5001, AustraliaKey Laboratory of Geospatial Big Data Mining and Application, College of Geographic Sciences, Hunan Normal University, Changsha 410081, China; Key Laboratory of Computing and Stochastic Mathematics of Ministry of Education, College of Mathematics and Statistics, Hunan Normal University, Changsha 410081, China; Corresponding author at: Key Laboratory of Geospatial Big Data Mining and Application, College of Geographic Sciences, Hunan Normal University, Changsha 410081, China.Key Laboratory of Geospatial Big Data Mining and Application, College of Geographic Sciences, Hunan Normal University, Changsha 410081, ChinaKey Laboratory of Computing and Stochastic Mathematics of Ministry of Education, College of Mathematics and Statistics, Hunan Normal University, Changsha 410081, China; Key Laboratory of Poyang Lake Wetland and Watershed Research of Ministry of Education, and School of Geography and Environmental Science, Jiangxi Normal University, Nanchang 330022, ChinaExisting evapotranspiration (ET) estimation models face inherent limitations when relying solely on physics-based or data-driven paradigms. To address this issue, we propose three data-physics hybrid modeling methods for improving instantaneous ET estimation in this study. A Physics-Data Learning (PDL) model is first formed by adding a complementary physical variable generated by Penman–Monteith (PM) equation to a deep learning (DL) model along with driving variables to regress latent heat flux. Building on the PDL, a Physics-Augmented Learning (PAL) model is then formulated by introducing a physics-augmented term into the loss function. Finally, a Physics-Augmented Residual Learning (PARL) model is developed by using the residual learning technique to deeply integrate the PM and pure DL baseline models. Using the FLUXNET dataset, three proposed models are compared with the baselines on ten vegetation types (VTs) across the globe. The results show that all proposed models improve the accuracy of two baselines and reduce the uncertainty of pure DL to different extents. Among them, the PARL achieves the highest accuracy and robustness, with NSE (RMSE) ranging from 0.71–0.82 (22.40–43.14 W/m2) across ten VTs. The PAL ranks second and effectively mitigates the PDL’s sensitivity to imperfect physical knowledge. Although three proposed models show better extrapolation ability than the pure DL under conditions of limited data, the PARL stands out for its superior generalization under four created extreme climate scenarios. These results confirm the potential of data-physics hybrid modeling in ET estimation, which is conducive to supporting efficient irrigation water management.http://www.sciencedirect.com/science/article/pii/S0378377425003488Residual learningPhysical knowledgePenman–Monteith equationGeneralization capabilityExtreme eventsLimited data
spellingShingle Binrui Liu
Xinguang He
Wenkai Lyu
Lizhi Tao
Physics-augmented deep learning models for improving evapotranspiration estimation in global land regions
Agricultural Water Management
Residual learning
Physical knowledge
Penman–Monteith equation
Generalization capability
Extreme events
Limited data
title Physics-augmented deep learning models for improving evapotranspiration estimation in global land regions
title_full Physics-augmented deep learning models for improving evapotranspiration estimation in global land regions
title_fullStr Physics-augmented deep learning models for improving evapotranspiration estimation in global land regions
title_full_unstemmed Physics-augmented deep learning models for improving evapotranspiration estimation in global land regions
title_short Physics-augmented deep learning models for improving evapotranspiration estimation in global land regions
title_sort physics augmented deep learning models for improving evapotranspiration estimation in global land regions
topic Residual learning
Physical knowledge
Penman–Monteith equation
Generalization capability
Extreme events
Limited data
url http://www.sciencedirect.com/science/article/pii/S0378377425003488
work_keys_str_mv AT binruiliu physicsaugmenteddeeplearningmodelsforimprovingevapotranspirationestimationingloballandregions
AT xinguanghe physicsaugmenteddeeplearningmodelsforimprovingevapotranspirationestimationingloballandregions
AT wenkailyu physicsaugmenteddeeplearningmodelsforimprovingevapotranspirationestimationingloballandregions
AT lizhitao physicsaugmenteddeeplearningmodelsforimprovingevapotranspirationestimationingloballandregions