Document Type : Research Paper
Authors
1 Department of Irrigation and Reclamation Engineering, University of Tehran, Karaj, Iran.
2 Department of Geography & Environmental Sustainability, University of Oklahoma, Norman, Oklahoma, USA.A
3 Department of Civil Engineering, University of Texas at Arlington, Arlington, Texas, USA.
Abstract
Reliable multi-step-ahead forecasting of reference evapotranspiration (ETo) is critical for proactive water resource management, yet understanding the temporal memory of hydrological systems remains a challenge for black-box deep learning models. This study presents a novel, interpretable forecasting framework integrating temporal SHapley additive explanations (SHAP) with advanced recurrent neural networks to predict daily ETo up to 7 days in advance across three contrasting climatic zones in Iran; Birjand (arid), Mashhad (semi-arid), and Gorgan (humid). By benchmarking long short-term memory (LSTM), bidirectional LSTM (BiLSTM), and CNN-LSTM architectures, it is demonstrated that model complexity does not always guarantee superiority; the standard LSTM proved remarkably robust, achieving high short-term accuracy (R² > 0.93 for 1-day forecast) in arid regions. However, a distinct humid-climate penalty was observed, with forecast accuracy degrading more rapidly in Gorgan due to stochastic cloud dynamics. The application of temporal SHAP revealed climate-specific memory effects: in arid zones, wind speed exhibited a persistent influence extending back 5 days, acting as a long-term driver of evaporative demand, whereas humid regions were governed by short-term radiative pulses. Furthermore, analysis of extreme events and drought propagation showed that while the model successfully captures heatwave-driven peaks, its reliability decreases under severe evaporative stress (standardized ETo anomaly > 2). Cross-spatial generalization tests confirmed that models trained on arid data transfer effectively to humid regions (R² = 0.95), but the reverse transfer fails to capture extreme advective forcing. This study provides a transferable, physically interpretable blueprint for developing early warning systems in data-scarce regions.
Keywords
Achite, M. et al. (2025) 'A coupled extreme gradient boosting-MPA approach for estimating daily reference evapotranspiration', Journal of Theoretical and Applied Climatology, 156(2), 113. doi: https://doi.org/10.1007/s00704-024-05313-x
Akbarpour, A. et al. (2024) 'Performance analysis of finite element method in groundwater studies based on Web of Science using R Biblioshiny', Journal of Aquifer and Qanat, 4(2), pp. 131-148. doi: https://doi.org/10.22077/jaaq.2024.7481.1071
Ali, M. et al. (2025) 'Improving daily reference evapotranspiration forecasts: Designing AI-enabled recurrent neural networks based long short-term memory', Ecological Informatics, 85, p. 102995. doi: https://doi.org/10.1016/j.ecoinf.2025.102995
Allen, R. G. (2000) 'Using the FAO-56 dual crop coefficient method over an irrigated region as part of an evapotranspiration intercomparison study', Journal of Hydrology, 229(1), pp. 27-41. doi: https://doi.org/10.1016/S0022-1694(99)00194-8
Borna, R. et al. (2023) 'Mitigation of Climate Change Impact on Bioclimatic Conditions Using Different Green Space Scenarios: The Case of a Hospital in Gorgan Subtropical Climates', Forests, 14(10), p. 1978. doi: https://doi.org/10.3390/f14101978
Chang, Y. et al. (2025) 'Machine Learning for Reference Crop Evapotranspiration Modeling: A State-of-the-Art Review and Future Directions', Agronomy, 15(9). p. 2038. doi: https://doi.org/10.3390/agronomy15092038
Chen, J., Chen, K. and Zhu, Y. (2026) 'Hybrid CNN-LSTM-attention model for enhanced reference crop evapotranspiration estimation toward optimized irrigation scheduling', Journal of Agriculture and Food Research, 25, p. 102545. doi: https://doi.org/10.1016/j.jafr.2025.102545
Dikshit, A. and Davis, C. (2025) 'Finding the link between flash drought and bushfires', Nature Reviews Earth & Environment, 6(5), pp. 322-322. doi: https://doi.org/10.1038/s43017-025-00675-w
Elbeltagi, A. et al. (2025) 'An interpretable machine learning approach based on SHAP, Sobol and LIME values for precise estimation of daily soybean crop coefficients', Scientific Reports, 15(1), p. 36594. doi: https://doi.org/10.1038/s41598-025-20386-y
Gao, Z. et al. (2025) 'An Interpretable Hybrid TCN-BiLSTM Model for Reference Evapotranspiration Prediction', Water Resources Management, 39(11), pp. 5481-5503. doi: https://doi.org/10.1007/s11269-025-04213-7
Goodarzi, A. et al. (2025) 'Prediction of pan evaporation across diverse climates and scenarios using temporal attention clockwork recurrent neural networks coupled with long short-term memory', Water Cycle, 6, pp. 241-253. doi: https://doi.org/10.1016/j.watcyc.2025.03.002
Hosseini, F., Prieto, C. and Álvarez, C. (2025) 'An explainable AI approach for interpreting regionally optimized deep neural networks in hydrological prediction', Journal of Hydrology, 661, p. 133689. doi: https://doi.org/10.1016/j.jhydrol.2025.133689
Huang, F. and Zhang, X. (2024) 'A new interpretable streamflow prediction approach based on SWAT-BiLSTM and SHAP', Environmental Science and Pollution Research, 31(16), pp. 23896-23908. doi: https://doi.org/10.1007/s11356-024-32725-z
Jiao, P. et al. (2024) 'Uncertain effect of component differences on land evapotranspiration', Journal of Hydrology: Regional Studies, 55, p. 101904. doi: https://doi.org/10.1016/j.ejrh.2024.101904
Khalili, S., Fayaz, R. and Zolfaghari, S. A. (2022) 'Analyzing outdoor thermal comfort conditions in a university campus in hot-arid climate: A case study in Birjand, Iran', Urban Climate, 43, p. 101128. doi: https://doi.org/10.1016/j.uclim.2022.101128
Khan, S. et al. (2025) 'Climate Impact on Evapotranspiration in the Yellow River Basin: Interpretable Forecasting with Advanced Time Series Models and Explainable AI', Remote Sensing, 17(1), p. 115. doi: https://doi.org/10.3390/rs17010115
Lakhiar, I. A. et al. (2025) 'A review of evapotranspiration estimation methods for climate-smart agriculture tools under a changing climate: vulnerabilities, consequences, and implications', Journal of Water and Climate Change, 16(2), pp. 249-288. doi: https://doi.org/10.2166/wcc.2024.048
Li, M. et al. (2024) 'Prediction of reference crop evapotranspiration based on improved convolutional neural network (CNN) and long short-term memory network (LSTM) models in Northeast China', Journal of Hydrology, 645, p. 132223. doi: https://doi.org/10.1016/j.jhydrol.2024.132223
Li, X. et al. (2024) 'A Framework for Quantifying the Uncertainty in Upscaling Evapotranspiration From Homogeneous to Heterogeneous Underlying Surface', IEEE Transactions on Geoscience and Remote Sensing, 62, pp. 1-24. doi: https://doi.org/10.1109/TGRS.2024.3453589
Long, J. et al. (2026) 'Reconstruction of drought propagation pathways: A global analysis of multitype propagation chains and nonlinear mechanisms', Global and Planetary Change, 256, p. 105144. doi: https://doi.org/10.1016/j.gloplacha.2025.105144
Lundberg, S. M. and Lee, S.-I. (2017) 'A unified approach to interpreting model predictions', Advances in Neural Information Processing Systems, 30, pp. 4765-4774. doi: https://doi.org/10.48550/arXiv.1705.07874
Maity, R. et al. (2024) 'Revolutionizing the future of hydrological science: Impact of machine learning and deep learning amidst emerging explainable AI and transfer learning', Applied Computing and Geosciences, 24, p. 100206. doi: https://doi.org/10.1016/j.acags.2024.100206
Mardani, M. et al. (2025) 'A bibliometric analysis of research trends on the application of remote sensing in precipitation estimation with an emphasis on spatio-temporal analysis in Iran', Iranian Journal of Rainwater Catchment Systems, 13(2), pp. 101-118. dor: https://dor.isc.ac/dor/20.1001.1.24235970.1404.13.2.1.3
Necesito, I. V. et al. (2025) 'Modeling daily evapotranspiration time series based on Non-Linear Autoregressive Exogenous (NARX) method and climate variables for a data-deficient region', PloS One, 20(2), p. e0318675. doi: https://doi.org/10.1371/journal.pone.0318675
Nikoo, M. R. et al. (2026) 'Assessing the fidelity of multi-satellite precipitation estimates for drought monitoring in a mountain water tower to arid basin system', Journal of Arid Environments, 232, p. 105519. doi: https://doi.org/10.1016/j.jaridenv.2025.105519
Nkiaka, E. et al. (2024) 'Quantifying the effects of climate and environmental changes on evapotranspiration variability in the Sahel', Journal of Hydrology, 642, p. 131874. doi: https://doi.org/10.1016/j.jhydrol.2024.131874
Nourani, V. et al. (2025a) 'Ensemble machine learning-based extrapolation of Penman-Monteith-Leuning evapotranspiration data', Ecological Indicators, 170, p. 113012. doi: https://doi.org/10.1016/j.ecolind.2024.113012
Nourani, V. et al. (2025) 'Advances in multi-source data fusion for precipitation estimation: remote sensing and machine learning perspectives', Earth-Science Reviews, 270, p. 105253. doi: https://doi.org/10.1016/j.earscirev.2025.105253
Ozupek, E. et al. (2025) 'Explainable artificial intelligence to explore the intrinsic characteristics of climatic parameters governing meteorological drought forecasting: opening the black box', Stochastic Environmental Research and Risk Assessment, 39(8), pp. 3201-3222. doi: https://doi.org/10.1007/s00477-025-03007-y
Pang, J. and Zhang, H. (2023) 'Global map of a comprehensive drought/flood index and analysis of controlling environmental factors', Natural Hazards, 116(1), pp. 267-293. doi: https://doi.org/10.1007/s11069-022-05673-5
Qian, H., Wang, W. and Chen, G. (2024) 'Assessing forecast performance of daily reference evapotranspiration: A comparison of equations, machine and deep learning using weather forecasts', Journal of Hydrology, 644, p. 132101. doi: https://doi.org/10.1016/j.jhydrol.2024.132101
Rezvani Moghaddam, P. et al. (2016) 'Saffron agronomy and technology (Book of Abstracts: 2013-2016)', Saffron Agronomy and echnology, 4(SUPPLEMENT), pp. 1-78. doi: https://doi.org/10.22048/jsat.2016.39250
Sreeparvathy, V., Debdut, S. and Mishra, A. (2025) 'A review of advances in flash drought research: Challenges and future directions', Earth's Future, 13(8), p. e2025EF006603. doi: https://doi.org/10.1029/2025EF006603
Talebi, H., Samadianfard, S. and Valizadeh Kamran, K. (2025) 'Estimation of daily reference evapotranspiration implementing satellite image data and strategy of ensemble optimization algorithm of stochastic gradient descent with multilayer perceptron', Environment, Development and Sustainability, 27(2), pp. 3707-3729. doi: https://doi.org/10.1007/s10668-023-04037-8
Tang, M. et al. (2025) 'Multi-step-ahead forecasting of daily reference evapotranspiration using hybrid deep learning models for the Taklamakan Desert oasis', Journal of Hydrology: Regional Studies, 61, p. 102663. doi: https://doi.org/10.1016/j.ejrh.2025.102663
Tosan, M. et al. (2024) 'Analysis of the global research trend of saffron (Crocus sativus L.) between 2000-2023', Saffron Agronomy and Technology, 12(2), pp. 115-138. doi: https://doi.org/10.22048/jsat.2024.443037.1524
Tosan, M. et al. (2026a) 'Spatiotemporal performance and error analysis of satellite precipitation products over a topographically complex semi-arid region in Iran', Journal of Mountain Science, 23, pp. 118-138. doi: https://doi.org/10.1007/s11629-025-9984-6
Tosan, M. et al. (2026b) 'The Transparency Revolution in Geohazard Science: A Systematic Review and Research Roadmap for Explainable Artificial Intelligence', CMES - Computer Modeling in Engineering and Sciences, 146(1), pp. 1-41. doi: http://dx.doi.org/10.32604/cmes.2025.074768
Umutoni, L. and Samadi, V. (2024) 'Application of machine learning approaches in supporting irrigation decision making: A review', Agricultural Water Management, 294, p. 108710. doi: https://doi.org/10.1016/j.agwat.2024.108710
Wang, J. et al. (2026) 'Estimation and mechanism analysis of global evapotranspiration based on a physics-informed deep-learning model', Journal of Hydrology, 664, p. 134351. doi: https://doi.org/10.1016/j.jhydrol.2025.134351
Wang, Y. and Zha, Y. (2024) 'Comparison of transformer, LSTM and coupled algorithms for soil moisture prediction in shallow-groundwater-level areas with interpretability analysis', Agricultural Water Management, 305, p. 109120. doi: https://doi.org/10.1016/j.agwat.2024.109120
Yang, F. et al. (2025) 'Enhancing groundwater predictions by incorporating response lag effects in machine learning models', Journal of Hydroinformatics, 27(2), pp. 338-356. doi: https://doi.org/10.2166/hydro.2025.295
Ye, S. et al. (2025) 'Explainable transfer learning for subsurface soil moisture prediction', Journal of Hydrology, 661, p. 133473. doi: https://doi.org/10.1016/j.jhydrol.2025.133473
Yin, X. et al. (2025) 'Quantifying the time-varying period and time lag features of groundwater response: Dynamic impacts of precipitation-fed groundwater recharge', Ecological Indicators, 176, p. 113648. doi: https://doi.org/10.1016/j.ecolind.2025.113648
Yonaba, R. et al. (2024) 'Accuracy and interpretability of machine learning-based approaches for daily ETo estimation under semi-arid climate in the West African Sahel', Earth Science Informatics, 18(1), p. 87. doi: https://doi.org/10.1007/s12145-024-01591-1
Zamanipoor, M. and Rahnama, M. R. (2025) 'Assessing the Role of Local Climate Zones in Shaping Land Surface Temperature in Cold Semiarid Metropolises: Evidence from Mashhad, Iran', Weather, Climate, and Society, 17(3), pp. 501-516. doi: https://doi.org/10.1175/WCAS-D-24-0124.1
Zhang, H. et al. (2025) 'Understanding Evapotranspiration Driving Mechanisms in China with Explainable Machine Learning Algorithms', International Journal of Climatology, 45(6), p. e8774. doi: https://doi.org/10.1002/joc.8774
)