THE CROOKSTON LANDSLIDE OF SEPTEMBER 4, 2003

On September 4, 2003, a landslide occurred along the Red Lake River bank in Crookston, Minnesota, USA. It created a 1200 ft long scarp with a maximum vertical displacement of about 6 ft, crossing an area of 60 acres. It caused more than $500,000 property damage. The scarp was 20 ft away from U.S. Trunk Highway 2 at the nearest spot. Fortunately no one was hurt. A similar landslide in this area occurred in 1933 under similar conditions, indicating similar events possible in the future.

Geologically the site was covered by 200~300 ft of clays, tills, and granular sediments from past glaciation associated with the formation of Glacial Lake Agassiz. The landslide involved the Red Lake Falls Formation at the bottom, and the Huot Formation on the top. The Red Lake Falls Formation is a hard, ground moraine till deposited by the retreating glaciers. The Hout Formation, which is soft and highly plastic, was deposited after an additional glacial event occurred. It was lacustrine clay but over-consolidated later.

The triggering factor of the landslide was the first rain following a three-month drought. During the drought, both river level and groundwater table were low. Decreased river level reduced hydrostatic support of the water to the toe area. Reduced groundwater table increased in-situ effective stresses which led the soils closer to failure. The first rain after the drought further weakened the soils. Combination of these three factors eventually triggered the landslide.

The landslide belongs to rotational category, with two planes of failure: one very deep, and the other fairly shallow. Using slice method, the factor of safety (FS) under several conditions was estimated. For river level at its average elevation of 839 ft, FS was 1.02. For a 5 ft drop of the river level due to drought, FS would be 0.99. To enhance stability of the site, by transferring soil from the head to the toe alone, FS could be improved to 1.14 at average river level; if external rocks would be added to the toe, FS could be improved to 1.24. Other methods, such as piling, would be needed to further improve the FS without creating new areas of potential slides.