Premium
One shape fits all, but only in the aggregate: Diversity in sub‐stand scale diameter distributions
Author(s) -
Zenner Eric K.,
Peck JeriLynn E.,
SaghebTalebi Khosro
Publication year - 2018
Publication title -
journal of vegetation science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.1
H-Index - 115
eISSN - 1654-1103
pISSN - 1100-9233
DOI - 10.1111/jvs.12634
Subject(s) - beech , homogeneity (statistics) , spatial ecology , neighbourhood (mathematics) , species richness , spatial distribution , mathematics , statistics , scale (ratio) , smoothing , physical geography , ecology , geography , cartography , forestry , biology , mathematical analysis
Questions To what extent are the sub‐stand size class distributions at the neighbourhood patch scale masked at the stand level via spatial smoothing? What are the implications for naturalness assessment of obscuring, at the stand level, structures that arise from endogenous processes at the neighbourhood scale? Location Elborz Mountains, south of the Caspian Sea, Iran. Methods Natural uneven‐aged old‐growth Oriental beech ( Fagus orientalis ) forest subject to fine‐scale disturbances was stem‐mapped in a 1.2‐ha plot of live trees >7.5 cm DBH; neighbourhoods were delineated across multiple scales using the tree‐centred floating neighbourhood approach of spatial tessellation to connect trees to their natural neighbours. Cluster analysis was used to identify an optimal number of recurring diameter distribution types (DDT) based on the proportion of trees in each of the five size classes. The richness of DDTs was assessed and the extent of DDTs was modelled as a function of tree size. Results The number of member trees and spatial extent increased with neighbourhood scale. While the stand‐level tree size class distribution exhibited a rotated sigmoid shape, the nine most distinct DDTs were characterized by four different distribution shapes (negative exponential, increasing‐q, rotated sigmoid and unimodal) and varied considerably in the abundance of trees in different size classes. Each individual tree contributed to at least three of the nine DDTs, which were nonetheless spatially distinct and varied in extent with size class composition. As expected, all metrics exhibited spatial smoothing (i.e., increasing homogeneity with increasing scale) as small, heterogeneous tree neighbourhoods were expanded into larger neighbourhoods with more similar average composition, which converged on a rotated sigmoid shape at the broadest scale. Conclusions The stand‐level rotated sigmoid distribution reflected spatial smoothing across more heterogeneous fine‐scale tree neighbourhood patches. Further studies are needed to assess if a better measure of natural forest structure may be the diversity of sub‐stand structures and the contribution of individual trees to that diversity, rather than an aggregated metric that averages across the hallmark heterogeneity of neighbourhood dynamics.