Forest cover type mapping and spruce budworm defoliation detection using simulated imagery

Forest cover-type mapping and spruce budworm defoliation detection were attempted using a June, 1983 SPOT simulation iinuge of an area in northern Wisconsin (Sawyer County). The study site contained a dicerse mix of cocer types typical of the region, including sugar maple-yellow birch, white pine-red oak-red maple, white birch-balsam fir, aspen, and red pine stands (uplands), and black spruce, tamarack, and white cedar-balsam fir stands (low-lands). The imagery is being evaluated using both manual and computer-assisted interpretation techniques. Results, to date, of the interpretations show the potential for substantial improvement in the accuracy and specificity of forest cocer type mapping, under Lake States' conditions, as compared to Landsat M S S data. Detail to level 11 of the Anderson-U. S.G. S. classijikation system was obtained manually with both the panchromatic and multispectral film products. The excellent spectral and textural information apparent in the inultispectral data allowed additional visual discrimination to level 111 (species) for forest types throughout most of the scene. The digital interpretation (supervised) was extremely accurate at lecel I. Level 11 dqferentiation was achieved; however, discrimination of detail to level I11 was limited to lowland species only. Budworm defoliation detection was limited, by the date of the imagery, to dgferentiation of mortality from the precious years' defoliation. Acquisition of optimally timed SPOT imagery will likely allow further discriminatiqn of forest damage caused by agents such as the spruce budworm. Ocerall, it appears that SPOT satellite data will likely provide forest managers with information that currently is obtained from medium and high altitude photography. The data's inherent amenability to computer processing in a geographic information system, wherein manual and digital analysis techniques can be integrated, is an added advantage. . . source evaluation procedures has been demoncover-type mapping and spruce budworm defoliastrated by researchers for a diverse range of applition detection. In that we are continuing our research with the SPOT simulation data described herein, this paper '* Presented at the 1984 SPOT Symposium, Scottsdale, should be considered as merely a progress report Arizona, 20-23 May 1984. on our research. Here we limit our discussion to the PHOTOGRAMMETRIC ENGINEERING AND REMOTE SENSING, Vol. 51, No. 8, August 1985, pp. 1115-1122. 0099-1112/85/5108-1115$02.25/0 O 1985 American Society for Photogramxnetry and Remote Sensing PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING, 1985