GSA Annual Meeting in Indianapolis, Indiana, USA - 2018

Paper No. 277-6
Presentation Time: 2:55 PM

NO STONE UNTURNED, LITERALLY: HUNTING FOR PALEOFLOOD DEPOSITS IN THE TENNESSEE RIVER GORGE


HARDEN, Tessa, U.S. Geological Survey, 2130 SW 5th Avenue, Portland, OR 97201 and O'CONNOR, Jim E., U.S. Geological Survey, 2130 SW 5th Ave., Portland, OR 97201

The geomorphology of the Tennessee River Gorge and adjacent floodplain preserves deposits from large late-Holocene Tennessee River floods that can be used to improve flood frequency analyses. Large colluvial sandstone boulders, which tumbled from the surrounding cliffs, dot the slopes of the gorge and provide flow obstructions during large floods. Flow separation around many of these boulders promotes sediment deposition and preservation if appropriately oriented. The stable bedrock boundary and narrow valley of the gorge allow for sensitive stage-discharge relationships so that reliable flow magnitudes can be estimated from the elevation of the flood deposits. Preliminary stratigraphic analysis and geochronology show evidence of 3-4 floods in the last 3,000 years similar in magnitude to the 1867 peak of record (460,000 ft3/s at nearby Chattanooga, Tennessee) that inundated much of the city. Evidence of the largest late-Holocene flood on the Tennessee River indicates it may have an estimated discharge about twice the 1867 flood and probably occurred in the last 400 years. Flood sediment preserved in caves and small alcoves in limestone bedrock immediately upstream from the gorge constriction provide additional evidence of large late-Holocene floods.

The USGS in cooperation with the Nuclear Regulatory Agency are using the paleoflood data developed for the Tennessee River to improve procedures for flood-risk assessments for nuclear power plants in the United States. The magnitude, timing and number of floods can be directly incorporated into flood frequency analyses using the recently released Guidelines for Determining Flood Flow Frequency - Bulletin 17C. This study is one of the first studies in the eastern U.S. to use paleoflood data to improve flood frequency estimates for rare, low annual exceedance probability floods.