A new method for modeling the heterogeneity of forest structure

Despite the critical ecological roles of structural complexity, ecologically relevant quantitative measures of structural complexity that allow comparisons among forest stands are still lacking. The objective of this study was to develop a method that allows comparisons of structural heterogeneity among stands. To encompass a broad range of potential structural complexities, we simulated three spatial point patterns each for tree-size distributions from five inventoried natural stands. Forest structure was modeled and analyzed by simulating point patterns of trees and constructing triangular networks to connect neighboring tree tops to one another. This method is based on the concept of spatial tessellation of tree positions, where point patterns are converted into 2-dimensional nearest neighbor triangles. A structural complexity index (SCI) was defined as the sum of the areas of 3-dimensional triangles (with x, y, and z coordinates) divided by the sum of the areas of 2-dimensional triangles. Vertical gradients were defined as the maximum size difference among the trees forming a triangle, that is, the greater the difference, the greater the structure and the larger the SCI value. Patch-types were defined as classes of structural gradients at different positions within the canopy. More patch-types are also indicative of more structurally heterogeneous stands. Both the SCI and the number of patch-types and patch-size heterogeneity related closely to conventional descriptors of forest structure. While the number of patch-types and patch-type heterogeneity related more to the vertical component of structural complexity, SCI integrated well with both the vertical and horizontal structure. SCI and the concept of patch-types complement one another and can be used to quantitatively compare the structure of different stands. The applicability of this modeling approach in characterizing the structural heterogeneity of forests across spatial scales is discussed.