PubAg

Main content area

Disentangling the causes of canopy height increase in managed and unmanaged temperate deciduous forests using multi-temporal airborne laser scanning

Author:
Senécal, Jean-Francois, Doyon, Frédérik, Messier, Christian
Source:
Remote sensing of environment 2018 v.217 pp. 233-243
ISSN:
0034-4257
Subject:
Bayesian theory, branches, canopy, canopy height, chronosequences, data collection, deciduous forests, forest management, lidar, models, open space, remote sensing, saplings, stem elongation, stochastic processes, temperate forests, tree crown, tree mortality, uncertainty, Quebec
Abstract:
Airborne laser scanning (ALS) is a tool that can be used to monitor canopy height changes when data acquisitions are done at successive times. However, ALS is subject to some limitations in forested environments. One of these is that height growth measured between two times can be due to vertical crown growth of trees or to lateral growth of branches. The latter process gives canopy height change values unrelated to actual stem elongation. Lateral growth is a source of uncertainties in height change analyses because canopy height models do not contain information on the source of the growth.Lateral infilling of open space in the canopy and vertical crown growth are important processes in the gap dynamics of temperate deciduous forests. Small gaps are expected to close more often by lateral growth of branches while larger gaps can sometimes close by the vertical crown growth of seedling and saplings. Yet, few studies have examined growth patterns in and out of gaps in temperate deciduous forests.We studied both unmanaged and managed temperate deciduous forests in southern Quebec, Canada using a multi-temporal ALS dataset. A Bayesian stochastic model of three-dimensional dynamics was developed to classify whether measured height growth was due to lateral growth or vertical crown growth. A generalized additive model was also constructed to relate vertical crown growth to the height of trees.Between 10.6% and 13.7% of the canopy height model's pixels changed value due to lateral growth. There was no clear temporal trend for the frequencies of lateral growth in the managed forests chronosequence. Frequencies of lateral growth were also comparable between unmanaged and managed sites. Treetops of dominant trees grew slowly while crown edges of those trees grew both vertically and laterally. Our results show that canopy structure recovery in managed forests occurs through faster vertical crown growth of trees <10 m tall and lower tree mortality compared to unmanaged forests.
Agid:
6159578