Canadian Forest Service Publications
Forest volume estimation using a canopy reflectance model in a multiple-forward-mode, Kananaskis, Alberta. 2003. Pilger, N.; Peddle, D.R.; Hall, R.J. 10 pages in Proceedings of the 25th Canadian Remote Sensing Symposium / 11th Congress of the Association québécoise de télédétection. 14-17 October 2003. Université de Montréal, Montréal, Québec. CD-ROM.
Issued by: Northern Forestry Centre
Catalog ID: 37063
Availability: PDF (download)
Three dimensional canopy geometric optical reflectance models provide a physical-structural basis to the analysis of satellite imagery, representing a potentially more robust, objective and accurate approach for obtaining forest cover type and structural information for volume estimation compared to existing image analysis methods. In this study, the Geometric Optical Mutual Shadowing (GOMS) reflectance model was run in multiple-forward-mode (MFM) and used with digital multispectral IKONOS satellite imagery to estimate mean tree height by area and stand volume for a Rocky Mountain study site in Kananaskis, Alberta. Stand volume was estimated as a function of mean tree height and basal area over a series of 100m2 plots. Average tree height per plot was obtained from MFM model runs, with basal area per hectare calculated as a function of diameter at breast height (dbh). Stem counts were derived from the MFM density and horizontal crown radius model output, with MFM derived dbh estimated using a relationship between field measurements of dbh and height over specific species locations. Allometric relationships between these variables were used with MFM modeled tree heights to obtain dbh over larger areas. Results indicated that MFM modeled tree height was within 1.7m of field measured heights for conifers, and within 2.7m for deciduous species, with stem counts estimated to within 5 trees per 100m2 plot area. This result was combined with model derived basal area to estimate stand level volume to within 3.3 m3/100m2 for Trembling aspen, and 0.68 m3/100m2 for Lodgepole pine plots without the requirement for extensive ground based field measurements. This modeling approach can be used as a stand-alone capability or it may be integrated with existing air photo or satellite based forest inventories with distinct advantages over current methods.