Examining multi-scale snow vegetation relationships using object-derived measures of canopy structure from terrestrial laser scanning, Grand Mesa, Colorado
Abstract
Forest canopies alter the energy balance of snow-covered regions as well as intercept snowfall. Understanding the impact of vegetation structure on snow processes in the sub-canopy and in open areas is important for scientists and water managers to develop parameters for water storage prediction models. To investigate the snowpack's response to canopy structure, we use Grand Mesa SnowEx 2017 data to implement a study on snow and canopy vertical metrics at different scales (from single trees to patches) and at different distances from the canopy. We assess the snowpack accumulation in relation to vegetation metrics such as foliage height diversity (FHD) and canopy height at sub-canopy scales and at different distances from the canopy. We use terrestrial laser scanning (TLS) data collected from sites in Grand Mesa, Colorado, US. The snow-off and snow-on TLS data were acquired during September 2016 and February 2017, respectively. The point clouds are classified into ground, snow, and vegetation using the CAractérisation de NUages de POints (CANUPO) method in CloudCompare software. We use the Marker-controlled watershed (mcwatershed) algorithm in R for singletree segmentation. Tree patches are also defined by aggregating the singletree segments. Furthermore, we compute snow depth using the Multiscale Model to Model Cloud Comparison (M3C2) algorithm for 1m projection and 10cm normal scale. Both Pearson's correlation and mutual information between snow depth and vegetation metrics are calculated to have a proper metric for linear and nonlinear related variables. We plot different buffers from the edges of the canopy and single trees to see how vegetation metrics influence snow depth gradients farther from the edge and how tree height affects snow depth at different distances from the edge. The results show that the snow depth and vegetation metrics absolute correlation for the sub-canopy varies from ~ 0.4 to 0.7 among the field sites. In addition, snow depth correlation at different distances from the canopy edge varies per distance and TLS site. Patches of trees show a higher correlation with mean snow depth rather than for single trees. We also show that removing shrubs using canopy height information improve the overall results.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2019
- Bibcode:
- 2019AGUFM.C33D1613H
- Keywords:
-
- 0736 Snow;
- CRYOSPHERE;
- 0758 Remote sensing;
- CRYOSPHERE;
- 0794 Instruments and techniques;
- CRYOSPHERE;
- 0798 Modeling;
- CRYOSPHERE