Describing forest canopy gaps efficiently, accurately, and objectively: New prospects through the use of terrestrial laser scanning
Describing forest canopy gaps efficiently, accurately, and objectively: New prospects through the use of terrestrial laser scanning
复制标题
DOI：
10.1016/j.agrformet.2015.06.006
复制
发表时间：
2015-11
影响因子：
6.2
通讯作者：
D. Seidel;C. Ammer;K. Puettmann
中科院分区：
文献类型：
--
作者：
D. Seidel;C. Ammer;K. Puettmann
林冠间隙是森林景观的重要生态组成部分。研究间隙的一个限制是缺乏有效、准确和客观的方法来表征间隙的大小和形状。本研究旨在探讨各种方法来克服这一限制。利用地面激光扫描仪测量了一个针叶林和一个落叶林（共12个）的6个人工树冠间隙。利用这些测量的点云，人工导出间隙大小作为基线，以评估使用全自动划定边缘线进行间隙大小计算的准确性。此外，我们将这些结果与模拟的传统间隙测量结果进行了比较，这些测量结果基于关于间隙形状（椭圆）的假设或基于不同数量的距离测量（间隙中心和Stand边缘之间）。以人工圈定为参考，自动圈定产生的圈定间距略小，相对均方根误差在3.4%到5.3%之间，具体取决于圈定间距的大小。所有模拟的传统方法（具有不同数量的测量和形状假设）都会产生较大的误差。然而，当超过16次测量来描述间隙形状时，通过增加样本量而获得的精度迅速下降。为了进一步讨论缝隙形状，我们提出了一种利用激光点云计算冠层缝隙边缘线分形维数的方法。最后，我们讨论了通过更详细地描述三维林隙形状来加深我们对森林林隙相关过程的理解的其他方法。
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.