GEOLOGIC EFFECTS OF THE HEBGEN LAKE EARTHQUAKE, SOUTHWESTERN MONTANA, AUGUST 17, 1959, AN EYEWITNESS ACCOUNT

On August 17, 1959, as a USGS geologic field assistant to Irving Witkind, I was located close to the epicenter of the largest recorded earthquake in the lower forty-eight Rocky Mountains (M_w7.3-7.5). Two fault scarps, the Red Canyon and Hebgen faults, formed coseismically with displacements approaching 20 feet. The faults offset trails and initiated waterfalls where they crossed creeks. Subsequent erosion has subdued the scarps, obliterated the falls, and partially filled basal grabens. The faults mostly cut colluvium, and abundant synthetic step-faults developed on the hanging wall. Slickensides on footwall bedrock indicated sub-vertical movement. A load roar to the west foretold the collapse of nearly 40 million yd³ of Archean rock onto the Rock Creek campground, killing 26 people. The Madison Canyon landslide dammed the Madison River, creating Quake Lake. A steeply-dipping dolomite rib at the base of the mountain, which had restrained overlying weathered schist and gneiss, ruptured during the quake, and the rock mass rose 400 feet up the opposite slope of the canyon. There was little intermingling of the rock types—a frontal zone of dolomite now lines the opposite northwest slope. The basin containing Hebgen Lake tilted northward more than 20 feet during the quake as seiche waves overtopped Hebgen dam several times. In places old scarps suggest prehistoric movement along the same trend as the Red Canyon fault. Many landslides in colluvium disrupted US Highway 287 along the north shore of Hebgen Lake. At the Cliff Lake campground, two campers were killed by a falling boulder. Hebgen dam cracked during the earthquake, but did not fail because its concrete core was jammed between a rigid southwest dolomite buttress and a downward-slumping colluvial cone on the northeast. Fearing that Ennis, Montana, 35 miles to the north, would flood if Quake Lake breeched its landslide dam, the US Army Corps of Engineers began channellizing through the Madison Canyon landslide. Several days after the quake, a 3,000-foot-long earthflow in the headwaters of Kirkwood Creek, with a 150-foot high head scarp, many pressure ridges, and tilted and falling trees, developed, and was still moving at the end of the summer. Shaking and compaction of water-saturated alluvial fan deposits resulted in solifluction features including sand blows along fractures and sinkholes as much as 15 feet deep.