Comparison of Canopy Cover Estimations From Airborne LiDAR, Aerial Imagery, and Satellite Imagery

Abstract

Canopy cover is an important forest structure parameter for many applications in ecology, hydrology, and forest management. Light detection and ranging (LiDAR) is a promising tool for estimating canopy cover because it can penetrate forest canopy. Various algorithms have been developed to calculate canopy cover from LiDAR data. However, little attention was paid to evaluating how different factors, such as estimation algorithm, LiDAR point density and scan angle, influence canopy cover estimates; and how LiDAR-derived canopy cover differs from estimates using traditional methods, such as field measurements, aerial and satellite imagery. In this study, we systematically compared canopy cover estimations from LiDAR data, quick field measurements, aerial imagery, and satellite imagery using different algorithms. The results show that LiDAR-derived canopy cover estimates are marginally influenced by the estimation algorithms. LiDAR data with a point density of 1 point/m <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> can generate comparable canopy cover estimates to data with a higher density. The uncertainty of canopy cover estimates from LiDAR data increased drastically as scan angles exceed 12°. Plot-level canopy cover estimates derived from quick field measurements do not have strong correlation with LiDAR-derived estimations. Both the aerial imagery-derived and satellite imagery-derived canopy cover estimates are comparable to LiDAR-derived canopy cover estimates at the forest stand scale, but tend to be overestimated in sparse forests and be underestimated in dense forests, particularly for the aerial imagery-derived estimates. The results from this study can provide practical guidance for the selection of data sources, sampling schemes, and estimation methods in regional canopy cover mapping.