Presented on April 2, 2025, by

Dr. Yu-Fen Huang
Junior Researcher
Tsang Stream Lab
Department of Natural Resources and Environmental Management (NREM)
University of Hawaiʻi at Mānoa

ABSTRACT: 

Flash floods pose significant risks in mountainous watersheds due to the rapid streamflow response to rainfall. This presentation explores the role of rainfall spatial structure in shaping hydrological responses in small watersheds (< 40 km²) on Oʻahu. We investigated how specific rainfall characteristics beyond rainfall amount and intensity—including rainfall area, connectivity of rainfall clusters, and spatial dispersion—affect high-flow events. Using the WRF-Hydro® model with high-resolution rainfall data and the Random Balance Designs – Fourier Amplitude Sensitivity Testing (RBD-FAST) method, we analyzed rainfall-streamflow relationships across seven watersheds on Oʻahu, Hawai‘i, from 2015 to 2020. Key findings highlight that heavy rainfall areas (≥25 mm/hr), rainfall cluster connectivity, and dispersive indices significantly influence peak flow magnitude. Additionally, small shifts in rainfall location notably impact peak flow, underscoring the importance of precise rainfall spatial data. Our results emphasize the need to incorporate rainfall spatial structure and its associated uncertainties into flood prediction models to improve streamflow forecasting and flood risk assessment.