Improving time upscaling of instantaneous evapotranspiration based on machine learning models

Evapotranspiration (ET) plays a crucial role in the global water and energy cycle. Upscaling instantaneous ET ([Formula: see text]) to daily ET ([Formula: see text]) is vital for thermal-based ET estimation. Conventional methods – such as the constant evaporative fraction method (ConEF), radiation-b...

Full description

Saved in:
Bibliographic Details
Main Authors: Danni Yang, Shanshan Yang, Jiaojiao Huang, Shuyu Zhang, Sha Zhang, Jiahua Zhang, Yun Bai
Format: Article
Language:English
Published: Taylor & Francis Group 2025-01-01
Series:Big Earth Data
Subjects:
Online Access:https://www.tandfonline.com/doi/10.1080/20964471.2024.2423431
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849760996722212864
author Danni Yang
Shanshan Yang
Jiaojiao Huang
Shuyu Zhang
Sha Zhang
Jiahua Zhang
Yun Bai
author_facet Danni Yang
Shanshan Yang
Jiaojiao Huang
Shuyu Zhang
Sha Zhang
Jiahua Zhang
Yun Bai
author_sort Danni Yang
collection DOAJ
description Evapotranspiration (ET) plays a crucial role in the global water and energy cycle. Upscaling instantaneous ET ([Formula: see text]) to daily ET ([Formula: see text]) is vital for thermal-based ET estimation. Conventional methods – such as the constant evaporative fraction method (ConEF), radiation-based method, and evaporative ratio method – often overlook environmental factors, leading to biased estimates of [Formula: see text] from [Formula: see text]. To resolve this issue, this study aimed to assess four machine learning (ML) algorithms—XGBoost, LightGBM, AdaBoost, and Random Forest—to integrate meteorological and remote sensing data for upscaling [Formula: see text] across 88 global flux sites. Each ML model was tested with eight different variable combinations. Results indicated that XGBoost exhibited the best performance, with a root mean square error (RMSE) generally below 13 W [Formula: see text] in estimating [Formula: see text] from [Formula: see text]. The best variable combination simultaneously considers evaporative fraction, available energy, meteorology factors, remote sensing albedo, normalized vegetation index, and leaf area index. Using this combination, the XGBoost model achieved an [Formula: see text] = 0.88 and an RMSE = 12.33 W [Formula: see text], outperforming the ConEF method ([Formula: see text] = 0.71 and RMSE = 18.86 W [Formula: see text]) and its expansions. These findings support the application of ML models in ET upscaling, enabling ET estimation across large spatiotemporal scales.
format Article
id doaj-art-d1ed48bee52f451ca63ceb641215c831
institution DOAJ
issn 2096-4471
2574-5417
language English
publishDate 2025-01-01
publisher Taylor & Francis Group
record_format Article
series Big Earth Data
spelling doaj-art-d1ed48bee52f451ca63ceb641215c8312025-08-20T03:06:10ZengTaylor & Francis GroupBig Earth Data2096-44712574-54172025-01-019112715410.1080/20964471.2024.2423431Improving time upscaling of instantaneous evapotranspiration based on machine learning modelsDanni Yang0Shanshan Yang1Jiaojiao Huang2Shuyu Zhang3Sha Zhang4Jiahua Zhang5Yun Bai6School of Computer Science and Technology, Qingdao University, Qingdao, ChinaSchool of Computer Science and Technology, Qingdao University, Qingdao, ChinaSchool of Computer Science and Technology, Qingdao University, Qingdao, ChinaSchool of Computer Science and Technology, Qingdao University, Qingdao, ChinaSchool of Geographic Sciences, Hebei Normal University, Shijiazhuang, ChinaAerospace Information Research Institute, Chinese Academy of Sciences, Beijing, ChinaHebei Technology Innovation Center for Remote Sensing Identification of Environmental Change, Shijiazhuang, ChinaEvapotranspiration (ET) plays a crucial role in the global water and energy cycle. Upscaling instantaneous ET ([Formula: see text]) to daily ET ([Formula: see text]) is vital for thermal-based ET estimation. Conventional methods – such as the constant evaporative fraction method (ConEF), radiation-based method, and evaporative ratio method – often overlook environmental factors, leading to biased estimates of [Formula: see text] from [Formula: see text]. To resolve this issue, this study aimed to assess four machine learning (ML) algorithms—XGBoost, LightGBM, AdaBoost, and Random Forest—to integrate meteorological and remote sensing data for upscaling [Formula: see text] across 88 global flux sites. Each ML model was tested with eight different variable combinations. Results indicated that XGBoost exhibited the best performance, with a root mean square error (RMSE) generally below 13 W [Formula: see text] in estimating [Formula: see text] from [Formula: see text]. The best variable combination simultaneously considers evaporative fraction, available energy, meteorology factors, remote sensing albedo, normalized vegetation index, and leaf area index. Using this combination, the XGBoost model achieved an [Formula: see text] = 0.88 and an RMSE = 12.33 W [Formula: see text], outperforming the ConEF method ([Formula: see text] = 0.71 and RMSE = 18.86 W [Formula: see text]) and its expansions. These findings support the application of ML models in ET upscaling, enabling ET estimation across large spatiotemporal scales.https://www.tandfonline.com/doi/10.1080/20964471.2024.2423431Constant evaporative fraction method (ConEF)evapotranspiration (ET)machine learning (ML)time upscaling
spellingShingle Danni Yang
Shanshan Yang
Jiaojiao Huang
Shuyu Zhang
Sha Zhang
Jiahua Zhang
Yun Bai
Improving time upscaling of instantaneous evapotranspiration based on machine learning models
Big Earth Data
Constant evaporative fraction method (ConEF)
evapotranspiration (ET)
machine learning (ML)
time upscaling
title Improving time upscaling of instantaneous evapotranspiration based on machine learning models
title_full Improving time upscaling of instantaneous evapotranspiration based on machine learning models
title_fullStr Improving time upscaling of instantaneous evapotranspiration based on machine learning models
title_full_unstemmed Improving time upscaling of instantaneous evapotranspiration based on machine learning models
title_short Improving time upscaling of instantaneous evapotranspiration based on machine learning models
title_sort improving time upscaling of instantaneous evapotranspiration based on machine learning models
topic Constant evaporative fraction method (ConEF)
evapotranspiration (ET)
machine learning (ML)
time upscaling
url https://www.tandfonline.com/doi/10.1080/20964471.2024.2423431
work_keys_str_mv AT danniyang improvingtimeupscalingofinstantaneousevapotranspirationbasedonmachinelearningmodels
AT shanshanyang improvingtimeupscalingofinstantaneousevapotranspirationbasedonmachinelearningmodels
AT jiaojiaohuang improvingtimeupscalingofinstantaneousevapotranspirationbasedonmachinelearningmodels
AT shuyuzhang improvingtimeupscalingofinstantaneousevapotranspirationbasedonmachinelearningmodels
AT shazhang improvingtimeupscalingofinstantaneousevapotranspirationbasedonmachinelearningmodels
AT jiahuazhang improvingtimeupscalingofinstantaneousevapotranspirationbasedonmachinelearningmodels
AT yunbai improvingtimeupscalingofinstantaneousevapotranspirationbasedonmachinelearningmodels