| Article ID | Journal | Published Year | Pages | File Type |
|---|---|---|---|---|
| 4459169 | Remote Sensing of Environment | 2013 | 10 Pages |
•We develop proxies for canopy cover and canopy closure from LiDAR data•We compare these proxies with Ellenberg indicator values for light (EIVlight)•LiDAR-based canopy closure better correlated with EIVlight than canopy cover•The method could estimate understory light availability across large landscapes
Canopy cover and canopy closure are two closely related measures of vegetation structure. They are used for estimating understory light conditions and their influence on a broad range of biological components in forest ecosystems, from the demography and population dynamics of individual species to community structure. Angular canopy closure is more closely related to the direct and indirect light experienced by a plant or an animal than vertical canopy cover, but more challenging to estimate. We used airborne laser scanner (ALS) data to estimate canopy cover for 210 5-m radius vegetation plots in semi-open habitats and forests in protected nature areas in Denmark. The method was based on the area of Thiessen (Voronoi) polygons generated from the ALS points. We also estimated angular canopy closure by transforming ALS points from Cartesian to spherical coordinates, and calculating the percentage of azimuth and zenith angle intervals which contained points. We compared these estimates with field-based estimates using densiometer for 60 vegetation plots in forest. Finally, we compared ALS-based estimates of canopy cover and canopy closure to field-based estimates of understory light, based on the average Ellenberg indicator values for light for the plant species present in a given plot. The correlations of Ellenberg values with ALS-based canopy closure were higher (r2: 0.47) than those with ALS-based canopy cover (r2: 0.26) and densiometer readings (r2: 0.41) for the forest sites. ALS-based canopy closure is thus a reasonable indicator of understory light availability and has the advantage over field-based methods that it can be rapidly estimated for extensive areas.
