Frontiers in Forests and Global Change (Mar 2020)

Using Process Based Snow Modeling and Lidar to Predict the Effects of Forest Thinning on the Northern Sierra Nevada Snowpack

  • Sebastian A. Krogh,
  • Sebastian A. Krogh,
  • Patrick D. Broxton,
  • Patricia N. Manley,
  • Adrian A. Harpold,
  • Adrian A. Harpold

DOI
https://doi.org/10.3389/ffgc.2020.00021
Journal volume & issue
Vol. 3

Abstract

Read online

Reductions in snow accumulation and melt in headwater basins are increasing the water stress on forest ecosystems across the western US. Forest thinning has the potential to reduce water stress by decreasing sublimation losses from canopy interception; however, it can also increase snowpack exposure to sun and wind. We used the high-resolution (1 m) energy and mass balance Snow Physics and Lidar Mapping (SnowPALM) model to investigate the effect of two virtual forest thinning scenarios on the snowpack of two adjacent watersheds (54 km2 total) in the Lake Tahoe Basin, California, where forest thinning is being planned. SnowPALM realistically represents small-scale snow-forest interactions to simulate the impact of virtual thinning experiments in which trees <10 and <20 m are removed. In general, thinning results in an overall increase in peak snow water equivalent and snowmelt. Areas around sheltered tree clusters have the largest increases of snowmelt due to decreases of canopy sublimation, while more open and exposed areas show a small decrease due to increases in snowpack sublimation. At the 30-m forest stand scale, existing forest structure controls the efficacy of thinning, where forest stands with mean leaf area index (LAI) >3 m2/m2 and 5–15-m tall show the largest increases in snow accumulation (up to 450 mm) and melt volume (up to 650 mm). Despite the role of tree- and stand-scale thinning on snowmelt, macroscale effects were limited to slightly larger increases in melt volumes at mid to low elevation slopes (<2,300 masl) and south facing areas per unit of LAI removed. A decision support tool using machine learning (random forest) was developed to synthesize SnowPALM results, and was applied to neighboring watersheds. These results will inform ongoing forest management practices in California, and improve our understanding of the effects of snow-forest interactions at scales relevant to water management.

Keywords