Application of Tree-Based Intelligence Methods for Wind Speed Estimation at the East of Lake Urmia

Considering the importance of wind in terms of its risks and potentials, the study of this issue is so important. The use of modern intelligence methods is an appropriate approach to predict the wind speed in different regions. East Azerbaijan province is located in the east of Lake Urmia and is one of the windy regions of the country. Due to its proximity to the salt bed resulting from the drying up of Lake Urmia, it is exposed to environmental and health threats. In this study, to predict the wind speed at the target station (Tabriz) using wind speed data from 15 other synoptic stations, Tree-Based intelligence methods including Random Forest (RF), Extra Tree (ET), and M5 models are employed. The monthly wind speed data over eight years (2015–2022) are performed. The 75% of the dataset is considered for the training phase, and the remaining 25% is used for the testing phase. The accuracy and performance of the models in predicting wind speeds at the target station were compared using Pearson's Correlation Coefficient (CC), Root Mean Square Error (RMSE), Nash-Sutcliffe Efficiency (NSE), and Willmott's Index of Agreement (WI). The results of comparing the models showed that the Random Forest (RF) with CC = 0.74, RMSE = 0.48 m/s, NSE = 0.12 and WI = 0.76 has predicted the wind speed of Target station (Tabriz synoptic station) with better performance and more accuracy compared to other models.