CHAPTER 12

SLUMGULLION LANDSLIDE FAULT CREEP STUDIES

by William Z. Savage and Robert W. Fleming

Introduction

The active part of the Slumgullion landslide has been observed to be moving for nearly 100 years, and indirect evidence suggests that it may have been active for the last 300 years (Crandell and Varnes, 1961). Direct observations of the movement carried out during the summer months in the 1950's showed the sliding rate to be roughly constant. This remarkable behavior contrasts with that of most large translational landslides where movement rates increase or decrease dramatically in response to variations in seasonal precipitation and other factors that influence subsurface hydrology. Prior to this investigation, there were no long term measurements of velocity of the Slumgullion landslide capable of detecting small seasonal variations.

The virtual certainty of nearly constant movement, continued year after year, makes the Slumgullion landslide an excellent field laboratory for various rate and process studies. In this study, small variations in the rate of movement across faults on the landslide were measured with four creepmeters installed during 1993. Three of the instruments are tide gages adapted to sample fault displacements at 15-minute intervals. The creepmeters were located at the toe, on a strike-slip fault above the toe and within the slide, and on a major, bounding, strike-slip fault on the left flank in the narrow part of the landslide (fig. 1). This last location is known, for historical reasons, as the Camera Station. The fourth creepmeter, a specialized high-precision instrument, was installed downslope from the Camera Station and across the same bounding left-flank strike-slip fault. This instrument was placed as part of a geophysical study, and the results obtained with it are discussed elsewhere (see Gomberg and others, this volume).

Results of the Fault Creep Measurements

The toe creepmeter was emplaced on April 9, 1993, and removed on November 9, 1993. Figure 2A is a sketch of the toe creepmeter installation, and figure 2B shows the recorded displacements. The dashed line with triangles indicates displacements between the center of the wheel on the creepmeter and the post as determined manually with a measuring tape, and the solid line in figure 2B indicates displacements recorded by the creepmeter. The early creepmeter record between April and mid-May was compromised by a loose pulley wheel, but movement during that interval (manually established dashed line in fig. 2B) was faster than the summer and fall rates. From April 9 to May 18, the manually determined apparent displacement rate at the toe was 5.2 m/yr, and from May 18 to November 9, a displacement rate of approximately 1.0 m/yr was manually obtained for this part of the landslide. The high rate from April 9 to May 18 may have been caused, in part, by observed local failure on the toe near the post. Because of instrumental problems and consequent gaps, a good creepmeter record was obtained only for about 13 of the 21 weeks that the instrument was in place. From the record that started on July 13 and ended before September 15, the creepmeter-determined displacement rate was 1.40 m/yr, and for the period from September 15 to November 9, the creepmeter-determined rate was 1.46 m/yr. The creepmeter installed on July 13, 1993, across a strike-slip fault above the toe (the location labelled Elk Trail in fig. 1) also suffered from instrumental problems and was removed on January 3, 1994. A sketch of this type of creepmeter installation, and the trigonometric relations used to obtain displacements parallel to a strike-slip fault are shown in figures 3A and 3B. Figure 4 shows the rather discontinuous record of displacements recorded at this internal strike-slip fault. Again, the dashed line with triangles indicates displacements between the center of the wheel and the post determined with a measuring tape, and the solid line indicates displacements recorded by the creepmeter. Manually determined displacement rates averaged 0.76 m/yr from July 13 to January 3, with rates of 0.74 m/yr from July 13 to September 15, 1.00 m/yr from September 15 to November 9, and 0.41 m/yr from November 9 to January 3. The 16-day creepmeter record starting on July 13 showed a displacement rate of 0.93 m/yr, and the more complete creepmeter record starting on September 15 showed a displacement rate of 1.04 m/yr.

The record obtained at the Camera Station is the most complete. The installation at this location and the trigonometric relations used to obtain displacements parallel to the strike slip fault are as shown in figures 3A and 3B. Figure 5 shows the record of displacements. Again, triangles indicate manually obtained values. Shortly after installation on April 9, 1993, velocities underwent an acceleration, presumably caused by spring snowmelt and thawing of the landslide surface. Velocities have been slowly decreasing here since midsummer. Minimum velocities of 2.75 m/yr were recorded from April 9 to approximately April 20. Maximum velocities of 11.6 m/yr were recorded from approximately May 1 to May 18. Average velocities from April 9, 1993, to January 3, 1994, are 5.84 m/yr at this location. This value of nearly 6 m/yr is believed to be the largest average velocity at any location around the perimeter of the landslide. Crandell and Varnes (1961) reported a similar velocity (5.8 m/yr) for a marker near the Camera Station but about in the middle of the slide.

Concluding Discussion

Long-term movement data (Crandell and Varnes, 1961; Smith, this volume) showing rates of about 0.75 to 1 m/yr in the vicinity of the toe and the lower right flank and 6 m/yr near the Camera Station appear to be an average of our measured seasonally varying rates. Variation in velocity over the course of the year suggests that there are hydrologic controls on the movement rate. However, because the movement rate appears constant over long periods, the landslide must contain various internal mechanisms that operate to regulate the rate of sliding in the face of competing factors that increase and decrease the sliding rate. For example, subsurface water pressures might be equilibrated through a network of springs or by water flow into open cracks on the landslide. The springs would bleed off excess water causing water pressures to decrease, and the flow of surface water into open cracks could cause an increase in subsurface water pressures. Such mechanisms would nullify both the effects of frozen ground in the winter that might cause the movement to decelerate and the effects of large, intense summer thunderstorms that might cause the movement to accelerate.

References Cited

Crandell, D.R., and Varnes, D.J., 1961, Movement of the Slumgullion earthflow near Lake City, Colorado, in Short Papers in the Geologic and Hydrologic Sciences: U.S. Geological Survey Professional Paper 424-B, p. B136-B139.