CHAPTER 9

PHOTOGRAMMETRIC DETERMINATION OF SLOPE MOVEMENTS ON THE SLUMGULLION LANDSLIDE

by William K. Smith

Introduction

As part of the multidisciplinary study of the Slumgullion landslide, aerial photographs taken August 12, 1985, and August 25, 1990, were compared using photo-identifiable natural features to determine the displacements of different parts of the active slide between the two dates.

There is scant literature describing the use of photogrammetry to determine displacements of natural features on active landslides. At the La Frasse landslide, Switzerland, aerial photographs at scales of 1:25,000 and 1:40,000 were used to calculate displacements of natural points over four time periods between 1957 and 1982 (Bonnard, 1984). Baum and Fleming (1991) used aerial photogrammetry to determine the displacement of about 150 points on the Twin Lake landslide, Utah, and integrated these measurements with detailed field mapping to identify different elements of the slide. Baum and Reid (1992) determined the displacements of about two dozen ground points and about 100 points on roofs of houses on an active slide in Hawaii.

Theoretically, the use of natural photo-identifiable targets gives the interpreter an almost unlimited supply of potential data points, without the need for placing large arrays of artificial targets in the field. Furthermore, the photographs themselves provide a permanent record of the conditions existing at the time of exposure, which allows one to go back and get more data points or to remeasure questionable ones. One may also be able to make kinematic interpretations of the measurements with respect to physical features (such as faults or cracks) visible on the photographs. Another advantage is that photographs record positions of all points at an instant; significant movements may take place during an extensive conventional survey. Once corresponding points are identified, it is faster to locate them photogrammetrically than by conventional field surveying, although at the cost of some loss of accuracy. The principal advantage of the method, however, is that one can measure displacements on older photographs, provided that suitable targets on the moving slide can be identified, and that a sufficient number of photogrammetric control points on stable ground can be identified.

Methodology

A local Cartesian coordinate system was developed in 1990 by a survey to establish photogrammetric control for both the 1985 and 1990 photography and to provide a consistent reference system for other studies of the Slumgullion landslide. Measurements on both sets of photographs were based on this local system. Varnes and others (1993, this volume) provide more detailed descriptions of the coordinate system and the survey network.

The method for determining displacements at individual points consists of finding the same photo-identifiable points on both the 1985 and 1990 photographs, then measuring the coordinates of the points with a stereoplotting instrument with an attached microcomputer. Instruments used were a Kern DSR-11 analytical plotter and a computer-assisted Kern PG-2 mechanical plotter. In both instruments, a computer calculates and displays the coordinates and elevation of the measuring mark (floating dot) based on the input coordinates of photo control points and photogrammetric parameters. The photo-identifiable points, other than targeted control points, consisted entirely of the bases of trees and bushes.

The accuracy and reproducibility of coordinate measurements from photographs depend on a variety of factors. Based on remeasurement of 28 points on several models of the 1985 photography, the standard deviation of a horizontal position measurement, [((x)² + (y)² )^1/2 ], in this study is about 0.44 m, resulting in a standard deviation of a calculated vector displacement of 0.63 m (Smith, 1993). The reproducibility of the elevation measurements is not as good as that of horizontal position. Of the 28 remeasured points, 17 had elevation discrepancies greater than 0.5 m. These discrepancies occur because it is often difficult to place the floating dot accurately on the ground at the base of the tree as the surrounding ground detail is obscured by the tree's shadow on one side and the image of the tree on another. The difficulty increases with larger trees and steeper ground.

Results

A total of 310 natural points (trees and bushes) were identified and measured on the two sets of photographs. Details of the measurements are given by Smith (1993). Figure 1 is a map of the active portion of the slide showing the horizontal displacement vectors, and figure 2 shows the contours of net horizontal displacement over the 5-year period. Displacements from 1985 to 1990 ranged from less than 0.25 m to more than 25 m. The pattern of the overall movement on the landslide surface shows the largest displacements in the central, narrow part of the slide. The upper part of the slide is undergoing extending (active) flow while the lower part is undergoing compressive (passive) flow. Points just above the active toe are advancing at the rate of about 1 m per year, whereas material in the narrow, central portion of the slide is advancing at about 5.7 m per year. These displacements are similar to those measured at several points by Crandell and Varnes (1961): 2.5 ft per year at the toe and 20.0 ft per year in the central part. They also concluded that the rate of movement is fairly constant from year to year and season to season. Similar values were found at the toe and in the narrow part from the fault creep studies of Savage and Fleming (this volume).

Analysis of the displacement contours (fig. 2) indicates a zone of extending flow from the basin below the main scarp to the upper part of the narrow section and a zone of compressive flow from the narrow part to the active toe. It is possible to apply formulas from the theory of plasticity to the displacement measurements to calculate various components of the velocity and strain fields. Velocities, rather than the total displacements, are used to calculate normal- and shear-strain rates, principal-strain rates and their direction, and maximum-shear-strain rates. Details and results of this analysis are given in a forthcoming article (Smith, in press).

Conclusions

Photogrammetry is an effective tool for monitoring actively moving landslides and for analyzing the velocity or strain-rate fields. Photogrammetric methods allow one to use archival photographs to determine displacements over long periods of time, provided that a sufficient number of identifiable points on stable ground are visible on each set of photographs to control the stereo models, and also provided that corresponding objects on the active slide can be positively identified on both sets of photographs. With the equipment used in this study (computer-assisted PG-2), horizontal positions of suitable trees and bushes can be measured with a standard deviation less than 0.5 m. Field surveys are generally more precise than photogrammetric measurements, but they involve more personnel and require recoverable points that will not be disturbed between surveys. Also, one cannot obtain additional points to fill in detail in areas of particular interest.

Measured displacements can be analyzed using two-dimensional plastic-flow theory to obtain principal- and shear-strain rates and principal-strain-rate trajectories. Finally, the data indicate, that for future experimental or theoretical studies, this landslide might be suitably modeled as a purely cohesive material extruding by its own weight through a constriction on a sloping surface.

References Cited

Baum. R.L., and Fleming, R.W., 1991, Use of longitudinal strain in identifying driving and resisting elements of landslides: Geological Society of America Bulletin, v. 103, no. 8, p. 1121-1132.

Baum, R.L., and Reid, M.E., 1992, Geology, hydrology and mechanics of the Alani-Paty landslide, Manoa Valley, Oahu, Hawaii: U.S. Geological Survey Open-File Report 92-501, 87 p.

Bonnard, Christophe, 1984, Determination of slow landslide activity by multidisciplinary measurement techniques, in Kovari, K., ed., Field Measurements in Geomechanics: Rotterdam, Balkema, Proceedings of the International Symposium, Zurich, September 5-8, 1983, v. 1, p. 619-638.

Crandell, D.R., and Varnes, D.J., 1961, Movement of the Slumgullion earthflow near Lake City, Colorado, in Geological Survey Research 1961: U.S. Geological Survey Professional Paper 424-B, p. 136-139.

Smith, W.K., 1993, Photogrammetric determination of movement on the Slumgullion Slide, Hinsdale County, Colorado, 1985-1990: U.S. Geological Survey Open-File Report 93-597, 17 p., 2 pl.

Smith, W.K., in press, Measurement of landslide displacements by aerial photogrammetry, Slumgullion Slide, Colorado: Association of Engineering Geologists Bulletin .

Varnes, D.J., Smith, W.K., Savage, W.Z., and Varnes, K.L., 1993, Control and deformation surveys at the Slumgullion Slide, Hinsdale County, Colorado--A progress report: U.S. Geological Survey Open-File Report 93-577, 15 p., 1 pl.