Determining the threshold of issuing flash flood warnings based on people’s response process simulation

Abstract. The effectiveness of flash flood warnings depends on the people’s response processes to the warnings. And false warnings and missed events cause the people’s negative responses. It is crucial to find a way to determine the threshold of issuing the warnings that reduces the false warning ratio and the missed event ratio, especially for uncertain flash flood forecasting. However, most studies determine the warning threshold based on the natural processes of flash floods rather than the social processes of warning responses. Therefore, an agent-based model (ABM) was proposed to simulate the people’s response processes to the warnings. And a simulation chain of "rainstorm probability forecasting - decision on issuing warnings - warning response processes" was conducted to determine the warning threshold based on the ABM. Liulin Town in China was selected as a case study to demonstrate the proposed method. The results show that the optimal warning threshold decreases as the forecasting accuracy increases. And as the forecasting variance or the variance of the forecasting variance increases, the optimal warning threshold decreases (increases) for low (high) forecasting accuracy. Adjusting the warning threshold according to the people’s tolerance levels of the failed warnings can improve warning effectiveness, but the prerequisite is to increase the forecasting accuracy and decrease the forecasting variance. The proposed method provides valuable insights into the determination of warning threshold for improving the effectiveness of flash flood warnings.