NH006-01
Exploring the potential of satellite data to characterize debris flow and landslide hazards within mountainous terrain

Monday, 7 December 2020: 20:30
Virtual
Dalia Kirschbaum, NASA Goddard Space Flight Center, Hydrological Sciences Laboratory, Greenbelt, MD, United States, Robert Emberson, Universities Space Research Association, Greenbelt, MD, United States, Pukar Man Amatya, Universities Space Research Association, Columbia, MD, United States, Thomas Stanley, Universities Space Research Association, GESTAR, Greenbelt, MD, United States, Sana Khan, Earth System Science Interdisciplinary Center, College Park, MD, United States and Hakan Tanyas, Universities Space Research Association Greenbelt, Greenbelt, United States
Recorded Presentation
Abstract:
The global coverage and temporal frequency that satellites provide offers a unique opportunity to estimate hydrometerological hazards and impacts of extreme rainfall within complex terrain, particularly in remote areas without in situ observations. However, defining when and where satellite-derived data and model products are relevant is largely contingent on the spatiotemporal sampling, the hazard characteristics and the needs from the research or applications community. In this presentation we present several new ways to map, model and assess debris flow hazards using a range of satellite data that may help to augment or replace in situ assessments in some cases.

A Semi-automatic Landslide Detection (SALaD) algorithm has been developed to map landslides using commercial high resolution or public optical data products. Using post-event imagery,SALaD combines image segmentation and machine learning to identify landslides that have likely resulted from a major trigging event. Once landslides have been identified in a landscape, we can use high resolution imagery to map debris flow runout from rainfall-triggered landslide events to provide a more complete picture of the impacts associated with mass wasting. Rainfall data can be compared across in situ and satellite products to better understand the characteristics of the extreme triggering event. Finally, we can leverage satellite-derived information from precipitation, topography, soil moisture and temperature along with in situ data such as rock strength to train a model designed to provide a broader picture of potential landslide hazards in near real-time using the Landslide Hazard Assessment for Situational Awareness (LHASA) model. Using these empirical observations, satellite data and simple, generalizable models, this work addresses how satellite data be applied to characterize population and infrastructure that are critically exposed to debris flow hazards.

Previous Abstract | Next Abstract >>