Document Type : Research Paper
Authors
1
Department of Water Engineering, Faculty of Agriculture and Natural Resources, Water Management Research Center, University of Mohaghegh Ardabili, Ardabil, Iran.
2
Department of Water Engineering, Faculty of Agriculture and Natural Resources, Water Management Research Center, University of Mohaghegh Ardabili, Ardabil , Iran.
10.22059/jwim.2025.383457.1182
Abstract
Evapotranspiration (ETo) is crucial for irrigation planning, crop performance simulation, hydrological water balance, crop water requirements determination, and irrigation system design. Soft computing models, such as neural networks and fuzzy logic systems, have been designed to overcome empirical models' shortcomings and accurately estimate ET. The advancement of remote sensing techniques in recent years has significantly enhanced the application of ET in agricultural and hydrological fields due to soft computing's ability to function effectively with less data and adapt to various climate conditions. We assessed the effectiveness of RF, MLR, and SVM models in estimating ETo in the Ardabil Plain area. meteorological stations' data in the study area were combined and formed into a random time series for constructing the model using RS and ETo. For the final model evaluation, the data from the fifth station were exclusively employed. The evaluation metrics employed included RMSE, R2, and NSE. The results for R2, NSE, and RMSE for the RF model were 0.7, 0.558, and 10.76, respectively for SVM, 0.71, 1, and 13.6, and for MLR 0.71, 0.688, and it was 21, which compares the results of the RF model with higher accuracy than other models. The random forest model's reliability for predicting ETo using RS datasets in data-poor areas is demonstrated by its high accuracy and stability in the present study.
Keywords
Main Subjects
)