Landslide densities associated with rainfall,stand age,and topography on forested landscapes,southwestern Washington,USA

Elevated landslide rates in forested landscapes can adversely impact aquatic habitat and water quality and remove and/or degrade soil resources required for forest regeneration. As a result, understanding the associations between management actions, natural factors, and landslide rates is important information needed for land managers. An unusual and powerful storm in early December, 2007, caused record flooding and thousands of landslides across southwest Washington and northwest Oregon, USA, and provided a rare opportunity to examine the effects of both natural factors and forest management practices on landslide density. Landslide inventory data were collected from both aerial photos and systematic field surveys to provide a broad survey database that was used to develop estimates of landslide density and to examine associations between landslide density, precipitation, topography, and forest stand age across a 152,000 ha forested landscape in the Willapa Hills, Washington. We estimated the probability of detecting landslides on aerial photos for six strata defined by forest stand age and a broad range of rainfall intensity, expressed as percent of the 100-year, 24-h, maximum rainfall. Key findings are that landslide detection probability decreased with increasing stand age, but was similar across rainfall intensities. The overall fraction of field-detected landslides that were not detected on 1:12,000-scale aerial photos was 39%. Very few landslides occurred in the 0–100% of 100-year rainfall category, regardless of stand age or slope gradient class. At higher rainfall intensities, significantly higher landslide densities occurred on steep slopes (>70% gradient) compared to lower gradient slopes, as expected. Above ∼150% of 100-year rainfall, the density of landslides was ∼2–3 times larger in the 0–5 and 6–10 year stand age categories than in the 11–20, 21–30, 31–40, and 41+ categories. The effect of stand age was strongest at the highest rainfall intensities. Our results demonstrate that ground-based landslide inventory data are required in order to correct for detection bias from aerial photos, develop reasonable estimates of landslide density across environmental gradients such as rainfall magnitude and topography, and make unbiased interpretations of relationships between forest management associations and landslide occurrence.