Slope Analysis
For this project we returned to Riverside Ave in Burlington, Vt. USA to examine the slope failure that occurred on 31 Oct 2019. From field measurements and remote observation we can assess the soil cohesion using an infinite slab on a linear slope model.
Field Observation 21 Oct 2020
The team collected measurements pertaining to the "Halloween" landslide from the base of the failure to determine simple model parameters. Field observations include: debris pile contents of large cement blocks, fallen trees, trash, and quartz rich sandy soils. The slide spans from the edge of a parking lot above to the debris pile at the base which extends nearly to the Winooski River. Measurements taken in situ include a slope angle of 30° - to confirm the slope distance between two trees was measured at 124in, while the vertical displacement between their stumps is 60in, quick field calculation showed this slope angle to be 28.9385°. The debris field height measured between 7ft 2in and 10ft at various points along it's perimeter (these values are held separately do to their high relative variance, and will used later to constrain min and max debris field volumes).
Remote observations from LiDAR equipped drones show the the landslide in Fig 3. This image was overlain on the the topography on the site provided by Dr. Bierman. After importing the .tif into Inkscape with included scale, distance can be calculated for the vertical projection of the failure area (Fig 4.). The landslide length was averaged over two measurements at 168ft long. The width was measured to be 172ft wide. The debris field is approximately the same width, and averaged 103ft long. While old topographic maps show an elevation change of 100-120ft, the GoogleEarthPro elevation data records 114.8291ft of relief (pre-landslide, though the run out is relatively flat, this is a safe assumption).
For ease of calculation, all measurements were converted to metric and averaged where appropriate. For an example on slope shear stress and strength, measurement and data, and model calculations, see My analysis and the Assignment Worksheet
Results
The model runs have assumptions about the internal angle of friction as well as densities of the dry and wet soils. While the soil cohesion estimates are informed by USGS standards, without sampling extensively throughout the debris, these runs will remain well educated guesses. With fallen trees in the lower end of the landslide (the upper trees were cleared for lot expansion and back filled with cement, sand, gravel, and rubbish) and the soil showing high traces of quartz, a rough estimate of cohesion is 4 kPa. If we assume our mean slope of ~30°, the model predicts 9 slide surface areas of water intake before failure (about 30,000 m^2) (See Fig. 5)
While the slope was extremely saturated when it failed, the poor cohesion of the soils due to dumping, clearcutting, and toploading certainly played a large role in the landslide. More information could be gleaned by measuring the cohesion and densities of the debris field, analyzing past images for tree cover, and a stronger measurement of slope inclination.
