North-Central Section - 54th Annual Meeting - 2020

Paper No. 4-6
Presentation Time: 9:55 AM

LANDSLIDES IN NORTHEAST MINNESOTA: INVENTORY MAPPING AND SUSCEPTIBILITY ASSESSMENT


RICHARD, Emilie M.1, DAHLY, Derek T.1, REHWINKEL, Rayann W.1, GRAN, Karen B.1, BRECKENRIDGE, Andy J.2, DELONG, Stephen B.3, DELONG, Whitney M.4, ENGLE, Zachary T.3, JENNINGS, Carrie E.5 and WICKERT, Andrew D.6, (1)Department of Earth and Environmental Sciences, University of Minnesota-Duluth, 1114 Kirby Drive, Duluth, MN 55812, (2)Department of Natural Sciences, University of Wisconsin - Superior, Belknap and Catlin, P.O. Box 2000, Superior, WI 54880, (3)Earthquake Science Center, U.S. Geological Survey, Moffett Field, CA 94035, (4)Department of Geography, Environment & Society, University of Minnesota, Minneapolis, ND 55455, (5)Freshwater, 2424 West Territorial Rd., St. Paul, MN 55114; Department of Earth & Environmental Sciences, University of Minnesota - Twin Cities, Minneapolis, MN 55455, (6)Deptartment of Earth & Environmental Sciences and SAFL, University of Minnesota, 310 Pillsbury Drive SE, Minneapolis, MN 55455

Newly available high-resolution lidar data of northeastern Minnesota have enabled new assessment of the role of geology and topography on slope stability in a post-glacial landscape. Following a record storm event on June 19th-20th in 2012, during which 20-25 centimeters of rainfall occurred over 48 hours, extensive flooding triggered hundreds of mass-wasting events that damaged local infrastructure and contributed large volumes of sediment to tributaries of Lake Superior. We developed a landslide inventory using repeat 1-meter airborne lidar data from 2011 (pre-flood) and 2012 (post-flood), as well as Google Earth imagery at a scale of 1:3,000. Field reconnaissance verified ~700 of ~1500 remotely mapped landslides: 97% of field-checked remotely-mapped slides were accurately identified. This research focuses on the assessment of slope instability across northeast Minnesota as part of a larger, collaborative, nearly state-wide landslide inventory and susceptibility mapping project. We conducted a multivariate logistic regression analysis of slope, aspect, elevation, relief, depth to bedrock, soil K-factor, lithology, land cover, and distance to stream to develop a susceptibility map for three watersheds. Only pre-flood data were used in the susceptibility modeling to capture landscape conditions prior to the majority of (2012 storm-related) mass-wasting events in the inventory. Data extracted from our landslide inventory were randomly sampled into 80% training data and 20% test data for cross-validation.

Our inventory and field observations document that slope failures in northeastern Minnesota are concentrated in glacial sediments along river valleys and the Lake Superior shoreline. We hypothesize that these are primarily driven by hydrologic factors such as channel incision, toe-cutting, lateral migration, sapping, and wave action. Slope failures are more frequent in areas containing glacial-lake deposits that were destabilized in response to the 2012 flood. Head scarps initially triggered by the flood in these areas continue to re-activate and propagate upslope several years later. The outcomes of this project will allow stakeholders and decision-makers to reduce risks from future landslides.