Previsual detection of two conifer-infesting adelgid species in North American forests.
Conference Paper
Overview
abstract
The balsam woolly adelgid, Adelges piceae, and hemlock woolly adelgid, A. tsugae (Homoptera: Adelgidae), are invasive pests of coniferous forests in both the Eastern and Western United States. Balsam woolly adelgid is capable of attacking and killing native North American firs, with Fraser fir (Abies fraseri (Pursh) Poir.) in the East and subalpine fir (A. lasiocarpa (Hook.) Nutt.) in the West being particularly susceptible to infestation. Hemlock woolly adelgid is capable of infesting native hemlocks and is a serious pest in forests of the Eastern United States where it is causing significant mortality to both eastern (Tsuga canadensis (L.) Carr.) and Carolina hemlock (T. carolinianaEngelm.). Infestations by either of these insects may take several years to kill the host tree. Damage by hemlock woolly adelgid frequently causes needles to discolor from deep green to grayish green. Discoloration of needles is also one of the symptoms used to diagnose infestations of balsam woolly adelgid. Traditional methods for assessing damage by these adelgid species include field surveys and aerial detection surveys. However, because infestations frequently occur in remote locations and can take years to build up, stand damage may accrue prior to visual detection of the infestations. Branch-level, spectral data of the foliage from trees were collected for several categories of infestation. In the Western United States, data were collected from subalpine fir infested with balsam woolly adelgid in northern Idaho. In the Eastern United States, data were collected from eastern hemlock in western North Carolina. Trees were sampled using a hand-held spectroradiometer. The measured radiance spectra were converted to percentage of reflectance and comparisons made between the infestation categories. Separation of the infestation levels occurred in a progressive pattern moving from noninfested to newly (or lightly) infested to heavily infested trees. Results suggest that previsual detection of this group of invasive insects may be possible with appropriate spatial and spectral sensor resolution.
