Scholarlink

<p>Evaporation, a critical hydrological factor, profoundly affects various sectors, including agriculture, hydrology, and water resource management. Accurate daily evaporation estimation remains challenging, especially in data-scarce regions like the temperate Caspian region (e.g.Sari, Iran). This study explores the effectiveness of novel optimization algorithms—Innovative Gunner Algorithm (AIG), Wavelet (W), Firefly Algorithm (FFA), and Bat Algorithm (BA)—integrated with a Support Vector Regression (SVR) model to predict daily evaporation rates from 2012 to 2022. The core innovation is a combined SVR model optimized using these algorithms. Data from Sari, Iran, representing a moderate Caspian climate, was used to evaluate model performance across nine scenarios reflecting varying data availability. Models were rigorously assessed using multiple metrics, including Root Mean Square Error (RMSE) and the Nash–Sutcliffe coefficient (NS). The Wavelet-SVR (WSVR) model consistently outperformed others, achieving a high correlation (r = 0.980), a low RMSE of 0.133&#xa0;mm/day (a 15% improvement over standard SVR), and a high Nash–Sutcliffe efficiency (NS = 0.97). The AIG-SVR also showed promise, offering a viable alternative to traditional metaheuristic approaches. These results indicate that the WSVR model is an accurate and efficient tool for predicting evaporation, crucial for improving water resource management in similar climates. The study highlights the impact of data availability on model performance, suggesting strategies for optimized data usage.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Estimation of Daily Evaporation Using a New Metaheuristic Hybrid SVR Method in the Moderate Caspian Climate of Iran

  • 期刊论文

  • 发表日期：

    2025年9月4日

查看全文
  • Reza Chamanpira,
  • Hamidreza Babaali,
  • Reza Dehghani

摘要

Evaporation, a critical hydrological factor, profoundly affects various sectors, including agriculture, hydrology, and water resource management. Accurate daily evaporation estimation remains challenging, especially in data-scarce regions like the temperate Caspian region (e.g.Sari, Iran). This study explores the effectiveness of novel optimization algorithms—Innovative Gunner Algorithm (AIG), Wavelet (W), Firefly Algorithm (FFA), and Bat Algorithm (BA)—integrated with a Support Vector Regression (SVR) model to predict daily evaporation rates from 2012 to 2022. The core innovation is a combined SVR model optimized using these algorithms. Data from Sari, Iran, representing a moderate Caspian climate, was used to evaluate model performance across nine scenarios reflecting varying data availability. Models were rigorously assessed using multiple metrics, including Root Mean Square Error (RMSE) and the Nash–Sutcliffe coefficient (NS). The Wavelet-SVR (WSVR) model consistently outperformed others, achieving a high correlation (r = 0.980), a low RMSE of 0.133 mm/day (a 15% improvement over standard SVR), and a high Nash–Sutcliffe efficiency (NS = 0.97). The AIG-SVR also showed promise, offering a viable alternative to traditional metaheuristic approaches. These results indicate that the WSVR model is an accurate and efficient tool for predicting evaporation, crucial for improving water resource management in similar climates. The study highlights the impact of data availability on model performance, suggesting strategies for optimized data usage.