2009 Portland GSA Annual Meeting (18-21 October 2009)

Paper No. 10
Presentation Time: 4:15 PM

WASHINGTON GEOLOGICAL SURVEY EMBARKS ON A STATEWIDE LANDSLIDE PRECIPITATION FORECASTING MODEL


SARIKHAN, Isabelle, NORMAN, Dave and CONTRERAS, Trevor A., Division of Geology and Earth Resources, Washington Department of Natural Resources, 1111 Washington Street SE, PO Box 47007, Olympia, WA 98504-7007, ISABELLE.SARIKHAN@dnr.wa.gov

During the December 2007 and January 2009 storms in Western Washington, the need for adequate statewide landslide forecasting was apparent. Widespread highway and road closures and building damage related to landslides was unforeseen by most emergency responders and operations. Current cumulative precipitation and rainfall intensity-duration threshold models have poorly forecasted landslides outside metropolitan areas and do not give adequate community based warning.

The DNR Washington Geological Survey has embarked on creating a statewide landslide forecasting model, based on data collected from the Washington Geological Survey’s GIS Statewide Landslide Database and from the December 2007 and January 2009 storms. Landslide initiation times have been recorded for both storm events and will a critical component on determining threshold values. The model will mesh precipitation thresholds with a statewide landslide susceptibility map, based on a modified SHALSTB model tied with lithology susceptible to precipitation events. Lithologies identified as susceptible during high precipitation events will be coded to a higher risk category, creating a more realistic model. During the past two storms, landslides clustered in specific lithologic units. In the December 2007 storm, shallow landslide densities were ten times higher in Crescent basalt (and related rocks) in the Chehalis River Headwaters than the surrounding marine sedimentary rocks, even with equal or higher precipitation. The lithologic effects in the January 2009 storm were similar; shallow landslide densities in Whatcom County concentrated over four times greater in the Chuckanut Formation sandstones than in the surrounding Darrington Formation phyllites. Landslide attributes, such as process, slope gradient, and slope morphology will be added into the equation to determine what characteristics lithologic units are most susceptible to. Downslope deposition areas for long reaching landslides, such as debris flows, will be mapped as a higher hazard, giving residences and communities adequate warning of future dangers. The model will also give emergency managers and the Department of Transportation the ability to determine high risk areas for landslide impacts to critical structures and infrastructure.