2007 GSA Denver Annual Meeting (28–31 October 2007)

Paper No. 20
Presentation Time: 1:30 PM-5:30 PM

A TEMPORAL DENDROGEOMORPHOLOGICAL ASSESSMENT OF THE RATTLESNAKE GULF LANDSLIDE, TULLY VALLEY, NEW YORK


TAMULONIS, Kathryn1, KAPPEL, William2 and JORDAN, Teresa1, (1)Earth and Atmospheric Sciences, Cornell University, Snee Hall, Ithaca, NY 14853, (2)U.S. Geol Survey, 30 Brown Road, Ithaca, NY 14850-1573, klt32@cornell.edu

The Tully Valley, New York, is a 6 mile long glacial trough located in the eastern Finger Lakes region and has a landslide history dating back to 9,870 14C yr B.P. In 1993, the largest landslide in the state since the early 1900's occurred on the west wall of the valley. Presently, two slow-moving earth slides are located in tributary valleys to the Tully Valley. Precipitation, ground water, and displacement measurements and a dendrogeomorphologic study indicate that Rattlesnake Gulf landslide movement occurs on two scales: shallow displacement immediately following precipitation, and deep seated movement related to the moving five year precipitation average and prolonged high stream discharge eroding the landslide toe.

The Rattlesnake Gulf landslide is composed of rotated, laminated silt and clay dipping 14°-55°SW. Data records from 2006-2007 show that shallow displacement corresponds to increased ground water elevation in the shallow water table, which lags precipitation events by approximately one to two days. Movement is greatest adjacent to the toe, and there is an eight month lag time for shallow stress release to translate up slope from the toe to near the head of the landslide.

Dendrogeomorphology relates dendrochronololgy to landform evolution as tree ring growth patterns provide dates for significant geomorphic events. Thirty-four tree cores and sections were collected from the landslide, as well 18 control samples. Tree rings were measured to determine years when ring growth changed from concentric to eccentric, and it was assumed this change was triggered by landslide displacement. With this data, an event index was generated to temporally analyze landslide activity. A multiple Fourier function regression model correlates the event index, five year moving precipitation average, and time. This implies there is a multi-year lag between precipitation and deep ground water discharge into the landslide, which causes deep seated movement. Model cyclicity also suggests that landslide activity has nearly a 70-year cycle, with destabilizing years in 1927 and 2000. Drought periods followed by persistent above average precipitation and corresponding high stream discharge eroding the landslide toe, as well as harvesting of mature trees above and within this landslide, may be cyclic displacement triggers.