Effects analysis and probability forecast (EAPF) of real-time management on urban flooding: A novel bidirectional verification framework

Government departments usually prepare and implement contingency plans to address frequent urban flooding caused by short-term heavy rainfall. Previous studies focused on the evaluation of the static impact of the policies on urban floods, while there is a lack of research on the effect of off-design conditions, real-time feedback and treatments of the flood events on urban flood mitigation, which is detrimental to the optimization of management strategies of the cities. To quantify the effects of real-time management on flood mitigation in Fuzhou City, China, this study proposed a framework (EAPF) for evaluation and risk prediction. First, we collected data on the locations, rainfall intensity, inundation time, and the triggers of the waterlogging events from 2017 to 2021. Second, based on the vigilance analyses, a structural equation model (SEM) was constructed to quantitatively evaluate the mitigation effects of management on waterlogging. Finally, a probability prediction model of dynamic drainage capacity was proposed for flood simulation caused by the rainwater grate blockage. The results indicate that the environmental factors were the decisive triggers affecting the severity of waterlogging, and increasing the frequency of management events effectively reduced the probability of blocking. The correlation between the number of management events and blocking flood events was -0.42, while a decrease in vigilance increased the possibility of flooding caused by overdue treatment. The proposed hydrological waterlogging model, which considered blockages, exhibited a Nash-Sutcliffe efficiency (NSE) coefficient exceeding 0.9 under deterministic conditions. The probability prediction model verified the mitigating effect of management on the blockages and urban flooding, and its results were consistent with those of the SEM. Our study contributes to improving the reliability of waterlogging prediction and optimizing the management flow in the developing cities.