TESTING THE LIMITS OF PHOTOGRAMMETRY TO MONITOR EROSION OF ARCHEOLOGICAL SITES IN GRAND CANYON

Gully erosion of archeological sites is an acute problem along the Colorado River corridor in Grand Canyon, but monitoring and studying it through conventional topographic surveys is expensive in this setting and can exacerbate erosion problems through human trampling. Thus we are testing very high resolution photogrammetry as a tool for non-destructive monitoring, and are pushing the methods limits as we try to resolve geomorphic features such as gully walls and knickpoints that are decimeter or centimeter in scale. We are also monitoring the integrity of erosion-control features emplaced by the National Park, as well as trying to detect changes in these features over time in successive collections of remotely sensed data.

Ground-control panels were placed evenly across each of four study sites in western Grand Canyon, and ground-control coordinates were collected at each panel at a 2.5 cm accuracy level. 1:1600 black-and-white stereo photography was collected and scanned to produce images with a pixel size of 1.95 cm. Aerial triangulation was preformed using 8-12 ground-control points and 150 to 200 tie points per stereo model, resulting in a triangulation solution with residual errors less than the image pixel size of 1.95 cm. The root-mean-square error of ground-control points not used in the aerial triangulation indicate horizontal and vertical errors of less than 5 cm. Digital terrain model (DTM) extraction was preformed at 20 cm grid spacing and the photogrammetric method's accuracy was tested by comparison to DTMs derived from total-station ground surveys. At a 90% confidence level, Z values are accurate to + or - 18 cm. This appears to be accurate enough for resolving sub-meter scale features, but is only at the threshold of effectively resolving the smaller features of interest such as gully knickpoints and cm-scale changes of features over time. As technology and software continue to improve, photogrammetry will provide highly detailed, non-intrusive monitoring of cultural sites and geomorphic phenomenon, and it may also be cost efficient in hard-to-access sites such as those in Grand Canyon.