Scholarlink

<p>Rainfall-runoff (R-R) modelling is crucial for hydrological analysis and water resources management. However, accurately simulating the complex and nonlinear relationship between rainfall and runoff using conventional methods is often challenging. Therefore, soft computing methods, which can handle uncertainty and imprecision, have been widely applied for rainfall-runoff modelling. This paper compares the performance of various soft computing methods in river basin rainfall-runoff modelling, including Backpropagation Neural Network (BPN), Support Vector Machine (SVM), Genetic Algorithm (GA), Radial Basis Function (RBF), and Fuzzy Logic (FL). Monthly rainfall and runoff data from four river basins in India are used as case studies. The results show that BPN is outstanding compared to the other methods in accuracy and robustness, followed by RBF and SVM. FL and GA have relatively lower performance but provide more flexibility and interpretability for rainfall-runoff modelling. The study concludes that soft computing methods are effective and reliable tools for river basin rainfall-runoff modelling analysis, and they can be further improved by incorporating more hydrological knowledge and data. During the testing period, all models exhibited a notable average deviation of approximately 13.1% (BPN), 43.6% (RBF), 19.2% (FL), 23.5% (SVM), and 19.2% (GA) from the actual values.</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

Soft Computing Techniques for Rainfall-Runoff Modeling and Analysis in River Basin

  • 期刊论文

  • 发表日期：

    2025年3月6日

查看全文
  • Pradeep Kumar Mishra,
  • Rashmi Dwivedi

摘要

Rainfall-runoff (R-R) modelling is crucial for hydrological analysis and water resources management. However, accurately simulating the complex and nonlinear relationship between rainfall and runoff using conventional methods is often challenging. Therefore, soft computing methods, which can handle uncertainty and imprecision, have been widely applied for rainfall-runoff modelling. This paper compares the performance of various soft computing methods in river basin rainfall-runoff modelling, including Backpropagation Neural Network (BPN), Support Vector Machine (SVM), Genetic Algorithm (GA), Radial Basis Function (RBF), and Fuzzy Logic (FL). Monthly rainfall and runoff data from four river basins in India are used as case studies. The results show that BPN is outstanding compared to the other methods in accuracy and robustness, followed by RBF and SVM. FL and GA have relatively lower performance but provide more flexibility and interpretability for rainfall-runoff modelling. The study concludes that soft computing methods are effective and reliable tools for river basin rainfall-runoff modelling analysis, and they can be further improved by incorporating more hydrological knowledge and data. During the testing period, all models exhibited a notable average deviation of approximately 13.1% (BPN), 43.6% (RBF), 19.2% (FL), 23.5% (SVM), and 19.2% (GA) from the actual values.