Development of a New Method for Modeling Site Index Using the Digital Stereo Aerial Photo-Derived Spatial Distribution of Canopy Height

Abstract

The objective of this study was to develop a new method for modeling the site index using the spatial distribution of canopy height derived from digital stereo aerial photos. The study area was Shimanoto District, Kochi Prefecture, Shikoku Island, Japan, and the target species was sugi (Cryptomeria japonica). A data set of a digital canopy height model (DCHM), which was treated as dominant tree height, and tree age were prepared at 50-m resolution. DCHM was derived as the difference between digital surface height model (DSM) and digital elevation model (DEM). Topographic factors derived from digital terrain analysis using 50-m resolution DEM were incorporated into the data set. A model combining the site index and height curve was developed for predicting site index for a specific site using topographic factors and dominant tree height for specific ages at the site. Using the data set, the proposed combined model was parameterized by Bayesian calibration. Markov chain Monte Carlo sampling did not converge, and the goodness-of-fit of the developed model was poor, although the predicted height growth pattern and site index using estimated parameters were reasonable.