DOES DEFORESTATION INFLUENCE THE ACTIVITY OF DEEP-SEATED LANDSLIDES? OBSERVATIONS FROM THE FLOOD OF 1997 IN THE CENTRAL KLAMATH MOUNTAINS, NORTHERN CALIFORNIA

Many deep, slow-moving landslides, as well as debris slides (shallow, rapid slides) were activated on the Klamath National Forest during the flood of 1997. A large proportion of these developed on pre-existing landslide deposits, in particular, within deforested areas (logged or burned by wildfire). Our data reveal a density of 1997 landslides within deforested parts of landslide deposits 3.5 times greater than in forested parts, and a sediment delivery rate 3.3 times greater. This strongly suggests a causal link between deforestation and increased landsliding. A total of 1543 active 1997 landslides (including road-related landslides) were inventoried across 1.7 million acres by a combination of air photo and field work. Of these, 415 occurred on landslide deposits, with 270 classed as debris slides, 97 deep landslides, and 48 combinations. These 415 slides delivered about 1.3 million cubic yards of sediment to streams (34% of the total). Many of the 270 debris slides were associated with reactivation of deep landslides. Exclusive of road-related landslides, the density of 1997 landslides on forested landslide deposits was 0.28 landslides/square mile, and the sediment delivery rate was 1.93 cubic yards/acre. For deforested landslide deposits, the density was 0.99 landslides/square mile and the delivery rate was 6.39 cubic yards/acre. Considerable controversy exists regarding the effect which deforestation has on the activity level of deep landslides. One school maintains that forest cover influences landsliding mainly through the mechanical reinforcement provided by roots, and since failure planes on deep landslides typically pass well below the rooting zone, root reinforcement has minimal stabilizing effect. Further, this school suggests that deforestation does not significantly raise peak groundwater levels. These views lead to timber harvest practices such as maintaining trees on steep toes of landslide deposits, where roots provide reinforcement against debris slides, and logging the gentle slopes at the heads of these deposits. If hydrologic effects are in fact important, such an approach could actually facilitate landsliding. In light of these issues, it remains critical to accurately map landslide deposits, and systematically assess reactivation potential prior to manipulating vegetation.