A Novel Hybrid Deep Learning Framework for Evaluating Field Evapotranspiration Considering the Impact of Soil Salinity
Abstract Accurate evaluation of evapotranspiration (ET) is crucial for efficient agricultural water management. Data‐driven models exhibit strong predictive ET capabilities, yet significant limitations like naive extrapolation hamper wider generalization. In this perspective, we explore a novel hybr...
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2024-09-01
|
| Series: | Water Resources Research |
| Subjects: | |
| Online Access: | https://doi.org/10.1029/2023WR036809 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850211746315239424 |
|---|---|
| author | Yao Rong Weishu Wang Peijin Wu Pu Wang Chenglong Zhang Chaozi Wang Zailin Huo |
| author_facet | Yao Rong Weishu Wang Peijin Wu Pu Wang Chenglong Zhang Chaozi Wang Zailin Huo |
| author_sort | Yao Rong |
| collection | DOAJ |
| description | Abstract Accurate evaluation of evapotranspiration (ET) is crucial for efficient agricultural water management. Data‐driven models exhibit strong predictive ET capabilities, yet significant limitations like naive extrapolation hamper wider generalization. In this perspective, we explore a novel hybrid deep learning (DL) framework to integrate domain knowledge and demonstrate its potential for evaluating ET under the influence of soil salinity. Specifically, we integrated physical constraints from process models (Penman‐Monteith or Shuttleworth‐Wallace) and salinity‐induced stomatal stress mechanisms into the DL algorithm, and evaluated its performance by comparing four diverse scenarios. Results demonstrate that hybrid DL framework offers a promising alternative for ET estimation, achieving comparable accuracy to pure DL during training and validation. Nonetheless, due to the limited available measurements, data‐driven model may not adequately capture plant responses to salt stress, leading to significant prediction biases observed during independent testing. Encouragingly, the hybrid DL model (DL‐SS) integrating Shuttleworth‐Wallace and salinity‐induced stomatal stress mechanisms demonstrated enhanced interpretability, generalizability, and extrapolation capabilities. During testing, DL‐SS consistently showed optimal performance, yielding root mean square error (RMSE) values of 37.4 W m−2 for sunflower and 39.2 W m−2 for maize. Compared to traditional Jarvis‐type approaches (JPM and JSW) and pure DL model during testing, DL‐SS achieved substantial reductions in RMSE values: 51%, 33%, and 43% for sunflower, and 45%, 31%, and 35% for maize, respectively. These findings highlight the importance of integrating prior scientific knowledge into data‐driven models to enhance extrapolation capability of ET modeling, especially in salinized regions where conventional models may struggle. |
| format | Article |
| id | doaj-art-642ae68fdcab4cfeb814832cc61c1024 |
| institution | OA Journals |
| issn | 0043-1397 1944-7973 |
| language | English |
| publishDate | 2024-09-01 |
| publisher | Wiley |
| record_format | Article |
| series | Water Resources Research |
| spelling | doaj-art-642ae68fdcab4cfeb814832cc61c10242025-08-20T02:09:29ZengWileyWater Resources Research0043-13971944-79732024-09-01609n/an/a10.1029/2023WR036809A Novel Hybrid Deep Learning Framework for Evaluating Field Evapotranspiration Considering the Impact of Soil SalinityYao Rong0Weishu Wang1Peijin Wu2Pu Wang3Chenglong Zhang4Chaozi Wang5Zailin Huo6Center for Agricultural Water Research in China China Agricultural University Beijing ChinaCenter for Agricultural Water Research in China China Agricultural University Beijing ChinaCenter for Agricultural Water Research in China China Agricultural University Beijing ChinaCenter for Agricultural Water Research in China China Agricultural University Beijing ChinaCenter for Agricultural Water Research in China China Agricultural University Beijing ChinaCenter for Agricultural Water Research in China China Agricultural University Beijing ChinaCenter for Agricultural Water Research in China China Agricultural University Beijing ChinaAbstract Accurate evaluation of evapotranspiration (ET) is crucial for efficient agricultural water management. Data‐driven models exhibit strong predictive ET capabilities, yet significant limitations like naive extrapolation hamper wider generalization. In this perspective, we explore a novel hybrid deep learning (DL) framework to integrate domain knowledge and demonstrate its potential for evaluating ET under the influence of soil salinity. Specifically, we integrated physical constraints from process models (Penman‐Monteith or Shuttleworth‐Wallace) and salinity‐induced stomatal stress mechanisms into the DL algorithm, and evaluated its performance by comparing four diverse scenarios. Results demonstrate that hybrid DL framework offers a promising alternative for ET estimation, achieving comparable accuracy to pure DL during training and validation. Nonetheless, due to the limited available measurements, data‐driven model may not adequately capture plant responses to salt stress, leading to significant prediction biases observed during independent testing. Encouragingly, the hybrid DL model (DL‐SS) integrating Shuttleworth‐Wallace and salinity‐induced stomatal stress mechanisms demonstrated enhanced interpretability, generalizability, and extrapolation capabilities. During testing, DL‐SS consistently showed optimal performance, yielding root mean square error (RMSE) values of 37.4 W m−2 for sunflower and 39.2 W m−2 for maize. Compared to traditional Jarvis‐type approaches (JPM and JSW) and pure DL model during testing, DL‐SS achieved substantial reductions in RMSE values: 51%, 33%, and 43% for sunflower, and 45%, 31%, and 35% for maize, respectively. These findings highlight the importance of integrating prior scientific knowledge into data‐driven models to enhance extrapolation capability of ET modeling, especially in salinized regions where conventional models may struggle.https://doi.org/10.1029/2023WR036809evapotranspirationsoil salt stressphysical constraintsdeep learninghybrid modelextrapolation capability |
| spellingShingle | Yao Rong Weishu Wang Peijin Wu Pu Wang Chenglong Zhang Chaozi Wang Zailin Huo A Novel Hybrid Deep Learning Framework for Evaluating Field Evapotranspiration Considering the Impact of Soil Salinity Water Resources Research evapotranspiration soil salt stress physical constraints deep learning hybrid model extrapolation capability |
| title | A Novel Hybrid Deep Learning Framework for Evaluating Field Evapotranspiration Considering the Impact of Soil Salinity |
| title_full | A Novel Hybrid Deep Learning Framework for Evaluating Field Evapotranspiration Considering the Impact of Soil Salinity |
| title_fullStr | A Novel Hybrid Deep Learning Framework for Evaluating Field Evapotranspiration Considering the Impact of Soil Salinity |
| title_full_unstemmed | A Novel Hybrid Deep Learning Framework for Evaluating Field Evapotranspiration Considering the Impact of Soil Salinity |
| title_short | A Novel Hybrid Deep Learning Framework for Evaluating Field Evapotranspiration Considering the Impact of Soil Salinity |
| title_sort | novel hybrid deep learning framework for evaluating field evapotranspiration considering the impact of soil salinity |
| topic | evapotranspiration soil salt stress physical constraints deep learning hybrid model extrapolation capability |
| url | https://doi.org/10.1029/2023WR036809 |
| work_keys_str_mv | AT yaorong anovelhybriddeeplearningframeworkforevaluatingfieldevapotranspirationconsideringtheimpactofsoilsalinity AT weishuwang anovelhybriddeeplearningframeworkforevaluatingfieldevapotranspirationconsideringtheimpactofsoilsalinity AT peijinwu anovelhybriddeeplearningframeworkforevaluatingfieldevapotranspirationconsideringtheimpactofsoilsalinity AT puwang anovelhybriddeeplearningframeworkforevaluatingfieldevapotranspirationconsideringtheimpactofsoilsalinity AT chenglongzhang anovelhybriddeeplearningframeworkforevaluatingfieldevapotranspirationconsideringtheimpactofsoilsalinity AT chaoziwang anovelhybriddeeplearningframeworkforevaluatingfieldevapotranspirationconsideringtheimpactofsoilsalinity AT zailinhuo anovelhybriddeeplearningframeworkforevaluatingfieldevapotranspirationconsideringtheimpactofsoilsalinity AT yaorong novelhybriddeeplearningframeworkforevaluatingfieldevapotranspirationconsideringtheimpactofsoilsalinity AT weishuwang novelhybriddeeplearningframeworkforevaluatingfieldevapotranspirationconsideringtheimpactofsoilsalinity AT peijinwu novelhybriddeeplearningframeworkforevaluatingfieldevapotranspirationconsideringtheimpactofsoilsalinity AT puwang novelhybriddeeplearningframeworkforevaluatingfieldevapotranspirationconsideringtheimpactofsoilsalinity AT chenglongzhang novelhybriddeeplearningframeworkforevaluatingfieldevapotranspirationconsideringtheimpactofsoilsalinity AT chaoziwang novelhybriddeeplearningframeworkforevaluatingfieldevapotranspirationconsideringtheimpactofsoilsalinity AT zailinhuo novelhybriddeeplearningframeworkforevaluatingfieldevapotranspirationconsideringtheimpactofsoilsalinity |