Detection of red attack stage mountain pine beetle infestation with high spatial resolution satellite imagery

The on-going mountain pine beetle outbreak in British Columbia, Canada, has reached historic proportions. There is an operational need for efficient and cost-effective methods to identify red attack trees in these areas. In this paper, we examine the use of an unsupervised clustering 4-m multispectral IKONOS imagery for the detection of mountain pine beetle red attack at sites with low and medium levels of attack. Independent validation data were collected from aerial photography and were used to determine the accuracy with which mountain pine beetle red attack could be detected using the multispectral IKONOS imagery. Concentric buffers, in 1-m increments to a maximum of 4 m, were applied to red attack pixels to characterize attribute accuracy as a function of positional accuracy. When a one-pixel buffer (4 m) is applied, the accuracy with which mountain pine beetle red attack could be detected using the multispectral IKONOS imagery was 71% (low attack) and 92% (medium attack). Analysis of red attack trees that were omitted in the analysis of the multispectral IKONOS image indicated that detection of red attack was most effective for larger tree crowns (diameter > 1.5 m) that were less than 11 m from other red attack trees. These results demonstrate that the unsupervised classification of mountain pine beetle red attack using multispectral IKONOS imagery is an operationally viable approach.