COS 150-8
Multi-scale effects of fire severity on snowpack dynamics in montane coniferous forest

Friday, August 14, 2015: 10:30 AM
344, Baltimore Convention Center
Jens T. Stevens, Department of Environmental Science, Policy and Management, University of California, Berkeley, CA

Temperate montane forests contribute disproportionately to water supply in the arid landscapes of western North America. A large proportion of precipitation in these mountain ranges falls as snow during the winter months, yet climate change forecasts predict decreasing winter snowpack depth and duration in the future. Concurrently, the frequency and severity of wildfire is increasing in many of these forests. Fire severity is likely to have strong effects on snow accumulation and melt rates by changing canopy interception and landscape albedo, but these effects have not been quantified to date. I tested how fire severity and forest canopy cover influence snowpack depth at multiple scales by conducting snow surveys at three recently burned forest sites in the Sierra Nevada of California during winter 2014. I measured snow depth across a gradient of fire severity between 2 and 6 times per site, for a total of 11 site visits, and recorded the overhead canopy cover directly above the sampling point. I modeled snow depth for each site visit as a function of landscape-scale fire severity class, point-scale overhead canopy cover, and topography, using linear mixed-effects models.


Canopy cover had divergent effects on snowpack depth depending on the spatial scale. At the landscape scale, a model of snow depth as a function of fire severity class indicated that snow depth decreased significantly at higher levels of fire severity, after controlling for the effects of site visit date, overhead canopy cover and topography. Model-estimated snow depth decreased by approximately 19 cm moving from sample points in unburned stands to sample points in high-severity stands. However, at the point-scale, where overhead canopy cover was the predictor variable, snow depth was significantly higher without overhead canopy cover, after controlling for the effects of site visit date, fire severity and topography. Model-estimated snow depth increased by approximately 8 cm moving from sample points underneath a live tree canopy to sample points without tree cover. These results suggest that live tree cover reduces snowpack depth at the scale of individual trees, via snowfall interception and sublimation. However, in more severely burned landscapes with lower tree cover, increased solar radiation and albedo from charred trunks likely increases snowmelt rates. Thus, management strategies that promote heterogeneity in canopy cover and/or fire severity may also promote prolonged water storage in snowpack.