PREDICTING DEBRIS FLOW INUNDATION VIA SEMI-EMPIRICAL MODELING IN SE ALASKA

Debris flows occur with little warning and can harm lives, damage property, and disrupt infrastructure when inundating areas. Predicting debris flow inundation has increasingly become an objective for emergency planners in order to mitigate the impact of future flows. This study’s objective is to produce a long-term (Holocene time scales) landslide inundation map for the SE Alaska region, using a model based on semi-empirical geometric scaling to compute depositional area.

In August of 2015, SE Alaska experienced dozens of debris flows triggered by intense precipitation peaking at 26mm/hr. Three individuals lost their lives to one of the debris flows, which motivated scientists to assess the storm event and landscape, in order to mitigate future disasters. A landslide susceptibility map for initiation was produced as a result of this event, but an inundation map has yet to be made. Debris flows from the 2015 storm were observed to have an abundance of coarse woody debris (CWD, defined as > 15 cm diameter) at the front of deposits. The entrainment of CWD by debris flows typically results in shorter runout lengths, as elongated particles interlock and dam water and sediment behind them.

We used the program “DFLOWZ” to model debris flow inundation using semi-empirical relationships among volume (V), cross-sectional area (A), and planimetric area (B), defined as A = aV^2/3 and B = bV^2/3, where a and b are empirical constants. Preliminary field measurements have calibrated coefficient values for both equations, where a = 0.17 and b = 9.1. Relative to debris flows from other regions of the world, SE Alaska experiences inundation with significantly smaller planimetric areas and slightly larger cross-sectional areas. Utilizing DFLOWZ with Monte Carlo simulations, potential inundation zones are delineated by modeling hypothetical debris flows with incremental volumetric content. By applying this method to multiple potential debris flows, highlighted by the aforementioned susceptibility map of the study area, a complete inundation map for the region will be produced, quantifying the long-term likelihood of each point in the landscape being inundated by a debris flow. This map will have important implications for hazard mitigation as well as understanding how debris flows transport sediment and wood through the landscape.