This is a topic I plan to flesh out a little over the next month or two. I want to explore a couple of topics that interest me:
– uncertainty in geotechnique; and
– scale effects in geomorphology (in time and space)
First, by way of a quick snapshot, there are a couple dozen large landslides in the Thompson River corridor south of Ashcroft, BC, Canada. Here’s an overview map:
Here are a couple of pictures of what these landslides look like:
The landslides in this map are up to about 1 km wide by 1 km long. They tend to have fairly low travel angles, ranging from about 7 to 18 degrees, averaging about 10-ish. Such relatively low travel angles are, to me, an indication of a brittle failure mechanism, or progressive failure. In order to have travel angles that flat, one must have either low strength, high pore pressures, or both.
The landslides occur in glaciolacustrine laminated silts and clays. Here’s a picture of a typical exposure:
The geological record in the region retains a record of at least four glacial episodes, and I believe three of them are known to be present in this reach of the Thompson, with post-glacial (glaciolacustrine) sediments separated by glacial drift (till, glaciofluvial, etc), resulting in a fairly complex sediment package.
On the topic of geotechnical uncertainty…
There has been some really excellent work published about these landslides, in particular several papers by former doctoral student Eshraghian at University of Alberta, working under the supervision of Norbert Morgenstern, David Cruden and Derek Martin. In their work, they’ve drawn conclusions about failure mechanisms that have substantial support from the physical evidence. However, I think the door remains open for alternate interpretations, with considerable evidence available to support such alternatives. Here I suggest that liquefaction of silt under high seepage pressures likely played a significant role, perhaps moreso than the role of high plastic clay with low residual strength. More on this some time later. But the point I’m working on is that even with the benefit of considerable effort by world class researchers, it is not always possible to solve complex geotechnical problems.
On the topic of scale….
These particular landslides are known to be influenced by seasonal changes in groundwater behaviour. The following image shows 4 years of observations of groundwater levels in nested piezometers, along with river levels in the adjacent Thompson:
What we see is that groundwater levels at different depths follow the river level, with the effect being more pronounced closer to the ground surface. By extension, we will see upward gradients throughout the year, with these increasing rapidly following drawdown after the summer peak flood. Coincidentally, major landslide movements tend to occur during that period. Also, major landslide activity tends to occur in years with above average peak flood, as shown in the following graph:
During the period of relatively careful record, landslides ONLY occur following major floods, and DO NOT occur in years with low peak floods.
The point I’m trying to come around to, however, is that in the case of these landslides, behaviour of the river, and associated landslide activity, is not obviously controlled by local weather, snowmelt, rainfall etc, but rather is controlled by larger area factors that determine behaviour of the river at this point. This shows the absence of a relationship between local weather (rainfall and temps) and river levels:
Here is a map showing location of the affected area within the overall river watershed:
You can see the site is near the outlet of the river into the Fraser River. We can look at snow pack thickness in the uplands of the watershed:
And find there is a pretty good correlation between average peak snowpack and peak flood:
Anyway, I will work both these angles a little bit, but will likely leave this topic along for a couple weeks or more.