Applications of machine learning and deep learning in hydrology from a bibliometric perspective: a comprehensive review
Ying Nie, Kok Hwa Yu, Yang Wang, P. Liu
Abstract
Hydrology faces significant challenges, including water resource prediction and natural disaster mitigation. Consequently, integrating machine learning with hydrology is pivotal for enhancing research accuracy and efficiency. Based on the Web of Science database, this paper systematically analyzes the application of machine learning in hydrology using the VOSviewer tool and bibliometric methods. It maps the field’s development trajectory and across dimensions including publication timelines, high-impact paper characteristics, national contributions, keyword frequency, and research domains. These analyses reveal regional roles in global research and demonstrate the unique impact of machine learning-hydrology integration. The results indicate that since 2012, the application of machine learning, particularly deep learning models like CNNs and LSTMs, has rapidly grown in hydrology. China, the United States, and India are the top three contributing countries in this field, reflecting their leadership in the intersection of hydrology and machine learning research. Furthermore, keyword clustering further identifies current research focuses and trends. Enhanced integration and cross-disciplinary collaboration are crucial for tackling complex hydrological challenges. The application of machine learning provides powerful tools for hydrology, promising more extensive and diverse applications in the future.