Analysis of landslide frequencies and characteristics in a natural system, coastal British Columbia

• Published in: Earth surface processes and landforms Vol. 29; no. 11; pp. 1321 - 1339
• Main Authors: Guthrie, R. H., Evans, S. G.
• Format: Journal Article
• Language: English
• Published: Chichester, UK John Wiley & Sons, Ltd 01.10.2004; Wiley
• Subjects: Bgi / Prodig; debris slides; debris flows; frequency-magnitude; Geomorphology; mass movement; Physical geography; sediment yield; Slopes; Canada; Vancouver; British Columbia; Mudflow; Spatial distribution; Watershed; Mass movement; Aerial photography; Landslide; Sediment budget; Slope dynamics; Geographical information system; Island
• ISSN: 0197-9337; 1096-9837
• DOI: 10.1002/esp.1095

Two hundred and one debris slides and debris flows were analyzed in a 286 km2 study area on the west coast of Vancouver Island, British Columbia, Canada. The study area remains essentially untouched by humans and therefore affords a natural setting in which to examine slope processes. Landslides were identified and characterized on aerial photographs from 1:15 000 to 1:31 680, and were then mapped and transferred to a GIS for analysis. Based on detailed landslide surveys, we propose a new method to accurately determine volume of landslides of this type by measured total area. Results indicate average denudation rates of 56 m3 y−1 km−2, and higher natural rates of failure than analogous regions in coastal British Columbia. In contrast, the landslide rates are substantially less than those from forested watersheds. Landslide distribution is spatially clustered in air photograph epochs, and we propose intense storm cells within regional events as the causal mechanism. Further, failures occurred preferentially over the West Coast Crystalline Complex (by 1·4 times), a metamorphic assemblage of gabbros, schists and amphibolites, but 1·5 times less often over the Island Plutonic Suite, a granitic intrusive formation. The former result represents a new finding, while the latter corroborates findings of previous authors. We examined magnitude–frequency relationships of the data set and present for the first time a strong argument that the rollover effect is not merely an artefact, but is instead a consequence of the physical characteristics of the landslides themselves. We subsequently analyzed magnitude–frequency relationships from two other complete data sets from coastal British Columbia and produced a family of curves corroborating this result. Copyright © 2004 John Wiley & Sons, Ltd.