Retrieval of leaf area index and canopy closure from high-spatial resolution CASI data over the BOREAS flux tower sites

This study is to investigate two approaches, spectral classification and canopy model inversion, to map the overstory canopy closure and LAI from CASI images with the spatial resolution of 1 m by 1 m obtained over four flux tower sites in the BOREAS study area in the summer of 1996. The study sites are old jack pine (SOJP) and young jack pine (SYJP) in the southern study area and NYJP and old black spruce (NOBS) in the northern study area. The results for NYJP are reported in this paper. In the classification approach, a K-means classifier is employed to identify 40 classes using three bands: green, red, and near-infrared for the NYJP site. These classes are then merged into two classes: overstory and understory, with an overall classification accuracy of 0.89. In the second approach the authors investigate the retrieval of LAI and canopy closure by inversion of a simple forest-light interaction model (FLIM) based on the spatially-degraded CASI images (10 m by 10 m). A visual evaluation shows that the created LAI and canopy closure images reflect the spatial distribution of the different strata of the forest stands in the NYJP site. The canopy closure obtained from the canopy model inversion corresponds to that from the classification with an R/sup 2/ of 0.46. The average LAI and its standard deviation over the transact are 1.08 and 0.4, which conforms with the field-measured data (1.2 and 0.3).