A 'POST-MORTEM' EXAMINATION OF LANDSLIDE OCCURRENCE RELATED TO THE DECEMBER 2007 STORM IN WESTERN WASHINGTON

In 2006, scientists associated with Washington State's Forest Practices Adaptive Management Program designed a study to evaluate the effectiveness of Washington State's Forest Practices Rules with respect to landslide occurrence. The study was implemented following the storm event of December 2007 that resulted in widespread landsliding in the Willapa Hills and flooding that closed I-5. Field crews surveyed over 240 square kilometers (59,300 acres) of forested uplands in southwestern Washington. The surveyed area was stratified into 788 harvest polygons. Within that area, 943 landslides were identified that delivered to public resources (primarily streams). Field crews also surveyed 989 km of forest roads, which they stratified into 2631 segments. Crews identified 209 road-related landslides that delivered to pubic resources. This study incorporates a randomized block design to control for spatial variability in causal factors such as precipitation, which exert a strong control on landsliding and has a poorly constrained spatial distribution. Generalized linear models were used to evaluate relationships between landslides and other causal mechanisms, including topography. Preliminary results suggest relationships between landslide density and precipitation, slope, and stand age. This presentation will focus on issues encountered in the analysis and methods used to overcome them.