THE ROLE OF VEGETATION CANOPY ON POST-FIRE DEBRIS-FLOW RAINFALL INTENSITY-DURATION THRESHOLDS

A post-fire debris-flow warning system in southern California uses empirical rainfall intensity-duration thresholds to forecast debris flow activity from burned watersheds. To increase process understanding of why such thresholds for burn areas are reduced relative to unburned sites, we investigated the role vegetation canopy exerts in modulating the rainfall reaching the ground. In three recent southern California fires, we installed a pair of tipping-bucket rain gages: one gage was installed in a burned area devoid of vegetation to measure gross rainfall, while the other was installed beneath neighboring unburned vegetation canopy to capture total throughfall.

Vegetation canopy significantly reduced the rainfall amount and intensity reaching the ground. During winter 2008-09, the canopy gages received an average of 42% less rainfall than the burn gages. For the largest storm recorded, canopy throughfall accounted for between 35-78% of gross rainfall. The 1-hour intensity at the canopy gages was reduced by an average of 3 mm/hr, and a maximum of 19 mm/hr; a 67% reduction. During the 2007-09 winters, about 1/3 of storms produced intensities at the burn gages that exceeded the intensity-duration thresholds that empirically forecast debris flow activity, although no debris flows were generated at our sites. At shorter durations throughfall intensities were well below the threshold, but for durations greater than 3 hours some throughfall began to match or exceed the intensity threshold. Canopy removal by fire effectively increases the recurrence interval at the ground level similar to events exhibiting higher intensities, lower frequencies, and generally greater rainfall totals. Based on the 1-hour intensity at one of our sites, for example, the effective recurrence interval of rainfall reaching the ground surface without vegetation canopy more than tripled. In the subsurface, at one site with co-located soil moisture probes, we also observed that soil in the burned area responded faster with higher moisture content values than nearby soil moisture beneath the canopy. Given a better understanding of the modulating role that canopy exerts on rainfall infiltrating into the ground we may be able to better forecast post-fire rainfall response for debris-flow initiation based upon pre-fire vegetation conditions.