A COMPARATIVE STUDY OF AERIAL PHOTOGRAPHS AND LIDAR IMAGERY FOR LANDSLIDE DETECTION IN THE PUGET LOWLAND, WASHINGTON

Landslides in western Washington cause thousands of dollars in damage each year. Accurate and precise remote sensing techniques are a necessary first step in creating useful landslide inventories for future land use planning and engineering mitigation decisions. Both aerial photos and LIDAR (LIght Distance and Ranging) imagery were evaluated and compared on the basis of the accuracy of landslide location and the precision of landslide boundary definition for a five-mile stretch of heavily forested coast along Hood Canal, Kitsap County, which is characterized by numerous slides occurring in late Pleistocene glacial and non-glacial sediments. Independent landslide inventories were developed from each remote sensing dataset and were followed by field observations of approximately half of the identified slides. Results suggest that the two methods yield similar results for the identification of slides; however, other factors such as the type of sliding, precision of slide boundary definition, vegetation, cost, imagery availability, and user efficiency show varied results. Photogrammetry is effective for locating small, shallow slides occurring along coastal bluffs as well as larger, deep-seated slides, costs only $25/mi² to fly, works well when vegetation cues indicate recent sliding, and is available throughout Washington State in multi-year intervals. However, photogrammetric interpretation often fails in precisely defining slide boundaries in forested terrain, is not manipulatable with respect to shadows and vertical exaggeration, and is less efficient in the field and office than LIDAR imagery. LIDAR imagery is highly effective for precisely defining landslide boundaries where the laser return from the ground is good, is efficient in the field and office, is easier to interpret than aerial photographs, and is highly manipulatable with respect to shadows and vertical exaggeration. However, LIDAR imagery does not always show smaller, shallower slide features, lacks vegetation cues, costs about $500/mi² to fly, and has limited availability in Washington State. Results from this comparison suggest that both methods have strengths and weaknesses with regard to generating landslide inventories and that the best approach is to use both methods in a complementary fashion.