Machine Learning-Based Estimation of Daily Reference Evapotranspiration in Vojvodina, Serbia
Earth · 26.05.2026 · 7 (3) · 88
DOI: 10.3390/earth7030088 ↗Kurz gefasst
Fachartikel in der Zeitschrift Earth, online erschienen am 26.05.2026. Verfasst von Milica Stajić und 8 weiteren Autor:innen. Die Arbeit nennt 53 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Provincial Secretariat for Higher Education and Scientific Research of AP Vojvodina, Interreg’s Danube Region Programme project “Development of a harmonized water balance modelling system for the Danube River Basin (Danube Water Balance)”.
Abstract (Originalsprache)
Reference evapotranspiration (ET0) is most commonly estimated using the FAO-56 Penman–Monteith (PM) equation. However, its application is often limited by the lack of required meteorological parameters. Due to their flexibility, ability to operate with limited input, and high accuracy in estimating ET0, machine learning models have become increasingly relevant in scientific research, offering a practical alternative under limited data conditions. In this study, artificial neural networks (ANNs) were applied to estimate daily ET0 using meteorological data from the Novi Sad station in Vojvodina (Serbia). The dataset consisted of eight meteorological variables relevant to evapotranspiration processes. Analysis showed that some variables had a stronger influence on ET0 prediction than others. To evaluate their combined effect, a series of ANN models with different input combinations were developed and tested. The random forests, gradient boosting and k-nearest neighbors models were used as a benchmark, and model performance was evaluated using R2, NSE, RMSE, and MAE. The highest accuracy was achieved when all variables were included, providing the model with maximum information. The best performance was obtained using a two-hidden-layer architecture with 32 and 16 neurons, resulting in R2 = 0.97, NSE = 97.07%, RMSE = 0.23 mm/day, and MAE = 0.21 mm/day. The results showed that a limited number of input variables can be used to estimate ET0 with high accuracy, achieving an R2 value of 0.95 using only three input variables. Therefore, the findings of this study may contribute to more accurate and cost-effective irrigation scheduling and water balance estimation, providing practical benefits for agricultural water management and farmers in Serbia.
Förderung
- Provincial Secretariat for Higher Education and Scientific Research of AP Vojvodina
- Interreg’s Danube Region Programme project “Development of a harmonized water balance modelling system for the Danube River Basin (Danube Water Balance)”
Zitieren
APA
Stajić, M., Mirčetić, D., Bezdan, A., Savić, R., Antić, S., Santrač, N., Salvai, A., Lakićević, M., & Blagojević, B. (2026). Machine Learning-Based Estimation of Daily Reference Evapotranspiration in Vojvodina, Serbia. Earth, 7(3), 88. https://doi.org/10.3390/earth7030088
Vancouver
Stajić M, Mirčetić D, Bezdan A, Savić R, Antić S, Santrač N, et al. Machine Learning-Based Estimation of Daily Reference Evapotranspiration in Vojvodina, Serbia. Earth. 2026;7(3):88. doi:10.3390/earth7030088
BibTeX
@article{Staji2026,
author = {Stajić, Milica and Mirčetić, Dejan and Bezdan, Atila and Savić, Radovan and Antić, Sanja and Santrač, Nikola and Salvai, Andrea and Lakićević, Milena and Blagojević, Boško},
title = {Machine Learning-Based Estimation of Daily Reference Evapotranspiration in Vojvodina, Serbia},
journal = {Earth},
year = {2026},
volume = {7},
number = {3},
pages = {88},
doi = {10.3390/earth7030088}
}
EUVIMED