Wednesday, 31 January 2024
Hall E (The Baltimore Convention Center)
Floods frequently threaten human life and property in southern Guam’s villages. Southern Guam’s mountainous topography, steep slopes, impermeable volcanic bedrock, and clay-based soils all contribute to a regime of surface runoff during high intensity rainfall events. Nonetheless, there is currently a poor understanding of the timing and mechanisms responsible for flash flooding in Guam. This investigation serves to identify the intensity and duration of rainfall events responsible for high streamflow events. The XGBoost machine learning algorithm and logistic regression are used to identify short timescale precipitation accumulation (two hours or less) as a dominant contributor to large peak flows; this is consistent with “flash” flooding. Traditional flood models and river forecasting tools employed in the contiguous United States are unhelpful in forecasting such flash floods, especially given the unique tropical setting and underlying geology of Guam. This gap in understanding motivates an analysis of the meteorological mechanisms behind these events. Understanding these mechanisms will help forecasters issue appropriate flood advisories, watches, and warnings. Circular statistics are used to examine temporal patterns in flash flooding as they are distributed (a) about the calendar year and (b) within a calendar day. Flash floods are overwhelmingly a feature of the wet season (July-December), with the dry season seeing very few high flow events. Although tropical cyclones (TC) account for many of the largest flows, only a fifth of all high-flow events can be attributed to TCs. Other mechanisms are revealed through the examination of events about the calendar day. Namely, mid-day island-effect convection (IEC) and an early morning peak in precipitation correspond well to the timing of extreme precipitation and streamflow events. Other mechanisms are also discussed. Finally, this study includes an historical analysis of the flood products issued by Weather Forecast Office (WFO) Guam.

