Describing forest canopy gaps efficiently, accurately, and objectively: New prospects through the use of terrestrial laser scanning

Abstract

Canopy gaps are an important ecological component in forested landscapes. One limitation to investigating gaps is the lack of efficient, accurate, and objective methods to characterize gap size and shape. This study aimed at investigating various methodologies to overcome this limitation. Six man-made canopy gaps were measured in a coniferous and a deciduous Stand (total of twelve) using a terrestrial laser scanner. Using the point clouds from these measurements, gap sizes were manually derived as a baseline to assess the accuracy of using fully automatic delineations of edge-lines for gap size calculations. Furthermore, we compared these results to those obtained from simulated conventional gap measurements that are based on assumptions regarding the gap shape (ellipse) or on a varying number of distance measurements (between gap center and Stand edge). Using the manual gap delineations as a reference, automatic delineations yielded slightly smaller gap sizes with a relative root mean square error between 3.4% and 5.3%, depending on gaps size. All simulated conventional approaches (with various numbers of measurements and shape assumptions) yielded larger errors. However, the gain in accuracy by increasing the sample size declined rapidly when more than 16 measurements were taken to describe the gap shape. To further the discussion about gap shape, we developed an approach to calculate the fractal dimension of the canopy gap edge-line from laser point clouds. Finally, we discuss other approaches to deepen our understanding of gap related processes in forests by means of a more detailed description of the three-dimensional gap shape.