Turtle Mountain is the site of Canada’s deadliest rockslide called the Frank Slide, which occurred in 1903 and killed over 70 people.  At approximately 4 am on April 29, 1903, 30 million cubic metres (90 million tonnes) of rock travelled down into the valley bottom, burying part of the town of Frank, a mining town built at the toe.

This photo gives an appreciation for the condition of the mountain prior to the slide and the location of the town of Frank.  It also is valuable in providing a visual indication of what this volume of slide looks like.

After the Frank Slide numerous studies were undertaken in order to determine the causes and contributing factors for the slide.  The overwhelming reason why the slide occurred was the unstable structure of the mountain.  It was considered that the mountain was likely at a very tenuous state of stability and a number of factors contributed to the timing of the slide, as listed above.

This figure (taken from the Frank Slide Interpretive Centre) shows the anticline fold and relevant structural features.  The portion of the mountain that is the focus of the current studies, South Peak, is situated on the eastern limb of the anticline.

Although the area below North Peak (peak on the right) was the original source of concern in 1911, studies showed that the geological structure was different than that for the Frank Slide and therefore the same sliding mechanism did not exist.  In 1931, John Allan (founder of the Alberta Geological Survey) discovered a series of deep cracks around South Peak (yellow dash) and based on a review of the orientation of the cracks and bedrock structure, estimated that a slide with a volume of approximately 5 million cubic metres could detach and fail in a similar manner as to the 1903 slide.  At this time some very rudimentary monitoring was undertaken (tape measurement between painted marks on crack sides).  In the 1980’s a series of monitoring points were also installed but monitoring discontinued prior to 1990.

Since the initial studies by Allan there has been the assumption in the geotechnical community that the landslide would originate on bedding along the eastern limb of the anticline and a failure surface would propagate upslope and through the western side of the mountain peak.  Recently we have re-evaluated this assumption and consider that the lower slope and upper peak (~ 0.3 million m3) are likely moving independently and the peak may be either sliding or rotating in response to the lower slope movements along bedding.

Aerial view from the north showing the series of deep cracks on the west side of the mountain in relation to the anticline.

Views of the cracks from the ground with people for scale.

The modern monitoring program was initiated in 2003 on the centennial of the 1903 Frank Slide and involved a variety of consultants, researchers and academic institutions.  The project ended on March 31, 2005 and has since received a provincial and national engineering award of excellence.

The current sensor configurations measures deformation (in millimetres of movement and degrees of tilt), microseismicity (to measure tremors within the mountain) and climate (to understand the impacts of climatic fluctuations on the rock mass and instrumentation).

All of the current instrumentation is situated on the South Peak of the mountain around the network of deep fissures.  The layout was designed to allow to aid in the determination of the movement mechanism (either sliding, toppling or a combination of the two), to provide redundancy (overlapping series of different sensors so that movement could be verified across various sensor types) and to be robust (with stand significant climatic influences and vandalism)

Aerial view looking east across South Peak showing location of deformation sensors that can be seen from this view.

There were two large storm events in the summer of 2005 where significant precipitation was followed by freezing air temperatures.  The initial hypothesis is that ice formed rapidly in the cracks and wedged the rock apart.  Both of these events registered deformations (up to 20 mm on September 10) on two of the extensometers.

Tiltmeters have been reliable but very sensitive to temperature fluctuation and humidity.  This being said, a number of tiltmeters have shown long term creep of the rock mass, not just the episodic events as seen on the extensometers.

Very small seasonal rock mass expansion/contraction trends, as well as actual deformations are observed on the crackmeters.  The September 2005 event is the same one registered on the extensometers.

An arrange of six surface passive microseismic stations were installed on the east side of the mountain as shown.  The aim of this array was to detect and source located microseismic events within the array that my be due to slope movements or continued collapse of the underground coal workings (abandoned).  These stations were installed in winter 2003/2004 and to this point have not provided reliable results due to a variety of issues, including sensor frequency and assumptions on the rock mass characteristics (and associated velocity models).

As indicated previously, the source location of events has not been successful at this time.  Interesting patterns have been observed that appear to be rock falls and the approximate location determined by the strength of the signal relative to the sensor locations.

Sensors on the west side of the mountain send data to a radio and network interface in a town called Blairmore and this data is then sent via a wireless link to the main data storage server at the Frank Slide Interpretive Centre.  Sensors on the east side of the mountain send data directly to the Frank Slide Centre via a wireless link.  Data is then accessed via the internet.

In addition to the continuous data stream from the sensor network a variety of new and emerging techniques are being used to characterize movements at Turtle Mountain.  These consist of ground and air based laser scanning and spaceborne radar.

Spaceborne Interferometric Synthetic Aperture Radar (InSAR) is a technique that slowly starting to be utilized by the landslide community for characterizing ground movements.  This technique uses SAR data from repeat pass polar satellites (Radarsat-1, ERS-1, ERS-2, Envisat, JERS, ALOS) to compare changes in phase that relate to ground deformations.

For Turtle Mountain we have incorporated data from 2000 to October 2006 to generate a deformation map utlizing the Permanent Scatter INSAR technique (PS-INSAR). This technique takes scenes of data from different time periods, identified permanent scatters/coherent targets) and tracks changes over time. This allows to better distinguish potential errors in the data (such as athmospheric effects) and allows to track movements with the resolution of a few millimeters per year.  In this image the green area has been set as the stable area and the other movements are in relation to this stable area.  There were three main areas (1 to 3) that we identified.  Areas 2 and 3 appear to represent shallow movement of talus on a side slope.  Area 1 is overlying abandoned underground coal mine workings (discussed on next slide).

There is a very distinct area highlighted with the InSAR in an area underlain by coal mine workings that were abandoned in 1918.  There is some surface expression of the subsidence but this provides the first reliable meaurement of the rate of settlement over the crown of the openings.

Another technique utilized is airborne Light Detection and Ranging (LiDAR).  This technique uses either a plane or helicopter with positioning very accurately tracked using GPS.  A laser system is positioned on the base of the aircraft and light pulses are emitted and returned on a very high density.  By knowing very accurately the position of the aircaft and the distance between to the ground for each light return, a very high resolution elevation dataset can be collected.  Positional accuracy of the data is typically quoted as 0.3 m vertical and 0.5 m horizontal (based on a collection density of ~ 1 point per metre) but can be improved with higher density collection.

Above is an example of a high resolution DEM (0.5 m) generated from the LiDAR data for the South Peak of Turtle Mountain that has allowed to viewing of structural features and mapping.  As well these views have highlighted potential failure modes previously not appreciated.

A powerful advantage of airborne LiDAR data is the ability to remove vegetation and view the bare ground surface.  This is achieved by distinguishing between returns that hit trees/buildings (first returns) and the ground.  The data density if so great that if first returns can be identified they can be removed and the ground points can be used the interpolate the ground surface, allowing generation of a digital terrain model.  The above images show the example from turtle Mountain with the image of the left with trees removed.

The bare earth model has allowed for visible identification of the subsidence features above the abandoned coal mine workings that underlie the 1903 slide mass also allows for visible identification of collapse pits that lie outside the 1903 slide but within the trees.  These features typically can only be seen when on the ground, in the forested areas.

In addition to the subsidence features, displacement features below South Peak and other areas of the mountain have been identified for the first time.  The photo shows the area circled in yellow on the slide.  This is a large rock slide that is above the fault and below Third Peak.  Future monitoring will focus on this area.

The bare earth model has also allowed for distinction of other larger scale features on other parts of the mountain.  While the main focus of studies over the past 100 years has been on the 1903 slide (right) and South Peak, we are now seeing features with the bare earth model that were not obvious with the vegetation obscuring the slope.  The feature on the left appears to be some sort of historical movement that may have been rapid and of similar magnitude to the 1903 slide.

The outlined area is that shown on the bare earth model on the previous slide (image from Google Earth).

Another airborne technique utilized to map deformations is photogrammetry.  In 1981 a series of 24 targets were installed around South Peak but only flown (at 1:2,000 scale) until 1985.  We had the targets reflown in 2005 and analysis completed by Mike Chapman at Ryerson University.  These new results agreed very well with the trends observed on the modern instrumentation and visual observations with deformations with averaged rates of movement of up to 3.5 mm/year over the 24 year time frame.

A group from Simon Fraser University undertook the first ground based laser scanning campaign on the mountain in summer 2007.  The aim was to generate high resolution digital elevation models of portions of the peak to undertake structural mapping.

The images above represent scans obtained from the North Peak of the mountain towards the rubbly, unstable area at the head of the 1903 Frank Slide.  With a data density of >>> you can clearly see the large blocks and trees in this zone.  While very high resolution data can be collected, there are issues with instrument range and limitations regarding access to suitable set up areas that limit the ability of this technique to scan large areas but it is very promising for smaller zones.  This is discussed in more detail in the paper by Sturzenegger et al (2007)

We are currently working with other groups in Europe to bring over very new technologies that are not currently being utilized in North America.  A group from Italy (LiSALAB) has developed a ground based InSAR system (left) that can do repeat short interval scans of slopes up to a three kilometers away to generate deformation maps at a higher resolution and frequency than can be obtained from spaceborne InSAR.  Also, as this technique can be used at night and in fog, it has distinct advantages over laser techniques.  This technique has been used on volcanoes in Italy and the application shown is for a rock slide on a fjord in Western Norway.   The reflector shown on the right is for another ground based radar system used at another site in Western Norway artificial targets are used to monitor point source movements on a large rock slide.

Another technique being utilized for monitoring in Western Norway (the Aknes/Tafjord Project) is short range laser monitoring.  This technique, taken from the off shore oil rigs in the North Sea, shoots a laser on one side of a crack to a heated plate on the other side.  This allows for year round monitoring and can track millimeter level deformations of point sources.  This system requires a power source for the laser and heated plate but has shown very good results over one year at the Aknes site in Norway.

The development of a terrestrial laser scanning system that can shoot over two kilometers and repeatedly survey for movements of a large area on the east side of Turtle Mountain is currently being investigated. A three year contract has been given to the Geomatics Engineering Department at the University of Calgary (Alberta) to develop and test this system.  Expected completion in spring 2010.

Thank you for your attention.  More information on the work at Turtle Mountain and the Alberta Geological Survey can be found at the link above.