Abstract

Flash floods, characterized by rapid streamflow response to rainfall, pose a significant natural hazard, particularly in small tropical watersheds (< 40 km2). Understanding the role of rainfall event characteristics, including amount, intensity, and spatial structure, is crucial for addressing and predicting flash floods. This study employs the WRF-Hydro® model with high-resolution (250-m, hourly) rainfall data and the Random Balance Designs – Fourier Amplitude Sensitivity Testing (RBD-FAST) method to investigate how rainfall impacts streamflow, specifically peak flow events, in seven watersheds on Oʻahu, USA. Analyzing storm events from 2015 to 2020, we examined peak flow responses to corresponding rainfall event characteristics and estimated their contributions to model efficiency. In addition, (1) random redistribution of rainfall and (2) spatial shifting of rainfall were experimented with to assess the sensitivity of peak flow to rainfall event characteristics. Not only the rainfall amount and intensity but heavy rainfall areas (>= 25 mm) within an event also exerted a significant impact on peak flow, while other spatial features contributed varying degrees of influence. Notably, spatially shifting rainfall for at least 250-m in any direction highly affected event peak streamflow, emphasizing the importance of rainfall amount, intensity, heavy rainfall areas, total rainfall areas, and connectivity among rainfall areas. Given the significance of rainfall's spatial heterogeneity, these findings underscore the benefits of incorporating rainfall spatial characteristics in probabilistic flood forecasting and the mitigation of flood risks. This research contributes valuable insights for enhancing flood prediction strategies in small tropical watersheds, providing a basis for informed decision-making and risk management.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call