AbstractThe growing menace of global warming and restrictions on access to water in each region is a huge threat to global hydrological sustainability. Hence, the perspective at which hydrological studies are currently being carried out across the world to quantify and understand the water cycle modeling requires a further boost. In the past few decades, the theoretical understanding of machine learning (ML) algorithms for solving engineering issues, and the application of this method to practical problems have made very significant progress. In the field of hydrology, ML has been using for a better understanding of hydrological complexities. Then, using ML-based approaches for hydrological simulation have been a popular method for runoff modeling in recent years; it seems necessary to understand the application of ML in runoff modeling fully. Current research seeks to have an overview for rainfall–runoff modeling using ML approaches in recent years, including integrated and ordinary ML techniques (such as ANFIS, ANN, and SVM models). The main hydrological topics in this review study include surface hydrology, streamflow, rainfall–runoff, and flood modeling via ML approaches. Therefore, in this study, the author has critically reviewed the characteristics of machine learning models in runoff simulation, including advantages and disadvantages of three widely used machine learning models.
Geomorphology-based genetic programming approach for rainfall–runoff modeling
