Enhanced estimation of reference evapotranspiration using hybrid deep learning models and remote sensing variables
Effective water resources management and irrigation scheduling for agricultural sector highly depend on the precise estimation of reference evapotranspiration, ETo. This study aims to develop ETo estimation models using deep learning algorithms with remote sensing variables as the input variables at...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-06-01
|
| Series: | Agricultural Water Management |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S0378377425002483 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850137307449917440 |
|---|---|
| author | Tze Ying Fong Yuk Feng Huang Ren Jie Chin Chai Hoon Koo |
| author_facet | Tze Ying Fong Yuk Feng Huang Ren Jie Chin Chai Hoon Koo |
| author_sort | Tze Ying Fong |
| collection | DOAJ |
| description | Effective water resources management and irrigation scheduling for agricultural sector highly depend on the precise estimation of reference evapotranspiration, ETo. This study aims to develop ETo estimation models using deep learning algorithms with remote sensing variables as the input variables at Pulau Langkawi and Kuantan stations, located in Peninsular Malaysia. Support vector regressor (SVR) was found to satisfactorily estimate the daytime land surface temperature (LST) using a set of significant variables including meteorological and remote sensing variables. It was then used along with downward shortwave radiation and surface reflectance bands to estimate ETo. Both long short-term memory (LSTM) and gated recurrent unit (GRU) showed their equivalent capability in estimating ETo and achieved the highest R2 of 0.695 and 0.796, respectively. The proposed hybrid deep learning models, combined model of convolutional neural network (CNN) with LSTM and GRU, respectively, achieved higher accuracy compared to individual models. They managed to improve the accuracy of the prediction in most of the cases, with the highest R2 = 0.805 and the lowest prediction errors, MAE = 0.265 mm/day, RMSE = 0.343 mm/day and NRMSE = 0.096. It was shown that the incorporation of surface reflectance bands and auxiliary variables (day length, Julian day and solar zenith angle) enhanced the performance of the models. This study provides valuable insights into deep learning algorithms and further confirms the potential of remote sensing variables as an alternative data source for ETo estimation. |
| format | Article |
| id | doaj-art-5a5fd014e4a44b088d7bb59f042919ee |
| institution | OA Journals |
| issn | 1873-2283 |
| language | English |
| publishDate | 2025-06-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Agricultural Water Management |
| spelling | doaj-art-5a5fd014e4a44b088d7bb59f042919ee2025-08-20T02:30:54ZengElsevierAgricultural Water Management1873-22832025-06-0131510953410.1016/j.agwat.2025.109534Enhanced estimation of reference evapotranspiration using hybrid deep learning models and remote sensing variablesTze Ying Fong0Yuk Feng Huang1Ren Jie Chin2Chai Hoon Koo3Lee Kong Chian Faculty of Engineering and Science, Universiti Tunku Abdul Rahman, Jalan Sg. Long, Bandar Sg. Long, Cheras, Kajang, Selangor 43000, MalaysiaLee Kong Chian Faculty of Engineering and Science, Universiti Tunku Abdul Rahman, Jalan Sg. Long, Bandar Sg. Long, Cheras, Kajang, Selangor 43000, MalaysiaLee Kong Chian Faculty of Engineering and Science, Universiti Tunku Abdul Rahman, Jalan Sg. Long, Bandar Sg. Long, Cheras, Kajang, Selangor 43000, MalaysiaCorresponding author.; Lee Kong Chian Faculty of Engineering and Science, Universiti Tunku Abdul Rahman, Jalan Sg. Long, Bandar Sg. Long, Cheras, Kajang, Selangor 43000, MalaysiaEffective water resources management and irrigation scheduling for agricultural sector highly depend on the precise estimation of reference evapotranspiration, ETo. This study aims to develop ETo estimation models using deep learning algorithms with remote sensing variables as the input variables at Pulau Langkawi and Kuantan stations, located in Peninsular Malaysia. Support vector regressor (SVR) was found to satisfactorily estimate the daytime land surface temperature (LST) using a set of significant variables including meteorological and remote sensing variables. It was then used along with downward shortwave radiation and surface reflectance bands to estimate ETo. Both long short-term memory (LSTM) and gated recurrent unit (GRU) showed their equivalent capability in estimating ETo and achieved the highest R2 of 0.695 and 0.796, respectively. The proposed hybrid deep learning models, combined model of convolutional neural network (CNN) with LSTM and GRU, respectively, achieved higher accuracy compared to individual models. They managed to improve the accuracy of the prediction in most of the cases, with the highest R2 = 0.805 and the lowest prediction errors, MAE = 0.265 mm/day, RMSE = 0.343 mm/day and NRMSE = 0.096. It was shown that the incorporation of surface reflectance bands and auxiliary variables (day length, Julian day and solar zenith angle) enhanced the performance of the models. This study provides valuable insights into deep learning algorithms and further confirms the potential of remote sensing variables as an alternative data source for ETo estimation.http://www.sciencedirect.com/science/article/pii/S0378377425002483MODIS aquaLong short-term memoryGated recurrent unitDownward shortwave radiationSatelliteLand surface temperature |
| spellingShingle | Tze Ying Fong Yuk Feng Huang Ren Jie Chin Chai Hoon Koo Enhanced estimation of reference evapotranspiration using hybrid deep learning models and remote sensing variables Agricultural Water Management MODIS aqua Long short-term memory Gated recurrent unit Downward shortwave radiation Satellite Land surface temperature |
| title | Enhanced estimation of reference evapotranspiration using hybrid deep learning models and remote sensing variables |
| title_full | Enhanced estimation of reference evapotranspiration using hybrid deep learning models and remote sensing variables |
| title_fullStr | Enhanced estimation of reference evapotranspiration using hybrid deep learning models and remote sensing variables |
| title_full_unstemmed | Enhanced estimation of reference evapotranspiration using hybrid deep learning models and remote sensing variables |
| title_short | Enhanced estimation of reference evapotranspiration using hybrid deep learning models and remote sensing variables |
| title_sort | enhanced estimation of reference evapotranspiration using hybrid deep learning models and remote sensing variables |
| topic | MODIS aqua Long short-term memory Gated recurrent unit Downward shortwave radiation Satellite Land surface temperature |
| url | http://www.sciencedirect.com/science/article/pii/S0378377425002483 |
| work_keys_str_mv | AT tzeyingfong enhancedestimationofreferenceevapotranspirationusinghybriddeeplearningmodelsandremotesensingvariables AT yukfenghuang enhancedestimationofreferenceevapotranspirationusinghybriddeeplearningmodelsandremotesensingvariables AT renjiechin enhancedestimationofreferenceevapotranspirationusinghybriddeeplearningmodelsandremotesensingvariables AT chaihoonkoo enhancedestimationofreferenceevapotranspirationusinghybriddeeplearningmodelsandremotesensingvariables |