Data quality and uncertainty issues in flood prediction: a systematic review
Jinhui Yu, Yichen Li, Xiao Huang, Xinyue Ye
Abstract
The twenty-first century has witnessed an increase in the frequency and severity of flood events, amplifying the consequences of data quality and uncertainty in flood prediction. This study conducts a systematic literature review to critically examine the challenges associated with the quality and uncertainty of diverse data types related to flood prediction, evaluate existing mitigation strategies, and identify future research directions. The findings reveal that flood prediction relies on various data types, including hydrological data, rainfall patterns, infrastructure characteristics, and topographical information. These datasets often suffer from issues such as incompleteness, inconsistency, and accuracy deficits, further complicated by uncertainties arising from complex spatial features and environmental changes. The literature proposes a range of solutions, including the development of innovative methodologies, model construction, and comparative analysis, to address these challenges. Future research should place greater emphasis on underdeveloped regions, fostering more literature-based studies, exploring the application of artificial intelligence (AI) and the integration of other emerging technologies to better address data challenges, and developing novel data sources, such as real-time dynamic data and integrated datasets. This review offers critical insights into the development and practical application of flood prediction data while charting a course for future research in the field.