T122 Hydrogeomorphic Processes in Hillslopes, Rivers, and Landscapes (Posters)

Monday, 1 November 2010
8:00 AM - 6:00 PM
Hall D

USING A NATURAL EXPERIMENT TO UNDERSTAND GULLY EROSION RATES AND MECHANISMS

Paper number: T122-5
Program: Technical Sessions
Time: 8:00 AM – 6:00 PM

Authors

Abstract

Gully erosion is a typical feature of landscapes in the high plains of the western United States. In recent years, researchers have been trying to decipher the mechanisms that control these features and the time scales of the mechanisms (e.g. annual vs. decadal processes), with an eye toward developing and testing mathematical models for their formation. To accomplish this, well-constrained case studies are needed. A dendritic network of deeply incised gullies that flow into West Bijou Creek near the town of Strasburg, Colorado, provides a natural experiment for exploring the mechanisms of gully erosion. The gully networks are anomalously steep, dropping 100 m in 1.5 km, in the otherwise low relief of the high plains of eastern Colorado. These gullies have vertical knickpoints up to 2 meters tall that appear to have migrated upstream from their confluence with West Bijou Creek. This site is constrained by a long historical record of aerial photos, a dense network of rain gages, and high precision topographic data including airborne lidar and repeat terrestrial lidar scans (TLS).

Our research to date has elucidated the rates and mechanisms that are contributing to knickpoint erosion in the gullies. Historical photo analysis from 1937 to present suggests a headward knickpoint erosion rate of decimeters per year to as much as 2 m per year. Repeated TLS in 2010 between the months of April and June demonstrated a knickpoint erosion rate of ~0.5 m, suggesting that the long term rate is observed on an annual timescale not just over the decadal period. Moreover, the geometry of the TLS data demonstrate that the knickpoint erosion can be attributed to vertical slab failure with both overland flow, and groundwater piping potentially influencing block failure. This observation has also been confirmed with time lapse photography of the knickpoint.

Site rain gage data has allowed us to investigate the relationship between rainfall intensity and overland flow. We have established that hourly rainfall intensities of at least 10 mm/hr are needed to generate overland flow, and the duration of the storm does not appear to influence overland flow generation until the threshold intensity is met. This indicates that in-channel gully erosion is primarily triggered by hortonian overland flow, when infiltration capacity is overcome by rainfall intensity.