Toward a physically based canopy metric for estimating snow accumulation in forests

Abstract:

Spatial variation of snow accumulation in forested basins depends on both the structure of the forest canopy as well as properties of the snow falling into the forest. While the use of airborne LiDAR has led to breakthroughs in characterizing canopies for estimating snow interception and accumulation in forests, many LiDAR-derived canopy structure metrics are at scales that are larger than those of forest accumulation processes. Higher-resolution point-clouds from small-footprint UAV LiDAR allow for finer scale analysis of canopy structure metrics. This study analyzes branch-scale canopy structure relevant to snow particle fall vectors through the canopy and examines the corresponding spatial relationship with snow accumulation below the canopy, using observations from a field campaign in the Canadian Rockies. Six UAV LiDAR surveys paired with on-the-ground snow or topographic surveys were conducted at an instrumented, variable-density needleleaf forest in Marmot Creek Research Basin between February and May 2019. Differential snow depth maps were calculated and verified using ground observations. Common canopy structure metrics including LAI, canopy closure, and several laser penetration metrics were calculated at branch scales, along with their covariance with increases in snow depth. Footprint analysis was conducted for each canopy metric. LAI and canopy closure displayed much larger footprints than the scale of snow depth changes, while laser penetration metrics had much smaller footprints. The physical relevance of metric footprints and corresponding trade-offs of each metric are discussed. A framework is proposed to estimate sub-canopy snow accumulation using small-footprint metrics and estimates of snow particle fall vector distributions. This may better characterize forest canopies for estimating snow accumulation with metrics which are more physically relevant and can be measured using increasingly useful UAV-LiDAR remote sensing.