Detection of spruce beetle (Dendrotonus rufipennis) infestations using aerial photographs

Normal colour (Kodak 2448) and colour infrared (Kodak 2443) aerial photographs (1:2000) of spruce trees infested with the spruce beetle (Dendroctonus rufipennis) were obtained during the summer and fall of 1982. The study site was located southeast of Prince George, B.C., on the north shore of Narrow Lake. The 70 mm stereo-transparencies were studied visually, the health status of each tree was determined, and the classifications compared to actual health categories as determined from extensive ground truthing. Densities of the individual dye-layers were measured, ratios of these densities calculated, and the data transformed as described by Moore (1980) and Hall et al. (1983). The densitometric variables were analysed using both analysis of variance and discriminant analysis computer programs available through the University of British Columbia Computing Centre. Analysis of the healthy, strip-attack, 1981-attack and dead tree images led to the following conclusions. The advantages of colour infrared film do not surpass those of the normal colour film, and vice versa. The decision of which film type to use should be determined by the photo-interpreter. Visual analysis of the health classes studied in this research is sufficient, without densitometric analysis. If densitometry is used, the basic dye-layer measurements and ratios are adequate. The Moore transformations do not provide any further information. The ratio between the green- and red-sensitive dye-forming layers, obtainable from both normal colour and colour infrared films, was the one variable that consistently separated the images of healthy, attacked and dead trees in the analyses of variance. This variable also figured prominently in the discriminant analyses. An optimal photographic scale for infestation detection and the minimum amount of time after attack required by a tree to show signs of strain on aerial photographs have yet to be determined. These questions must be answered before this survey technique could be termed an operational procedure in spruce beetle management.Forestry, Faculty ofGraduat