Can you measure pest infestation over wide inaccessible areas?
The study area was located on the Manti-LaSal National Forest in
east central Utah on a high plateau known as the Wasatch Plateau.
The area was experiencing extensive mortality in the Engelmann spruce
type due to an ongoing spruce beetle outbreak. Ground and aerial
survey information on the outbreak collected during the same time
as the acquired Landsat TM imagery was available.
Subpixel Classifier did!
Project area within the Wasatch Plateau
The purpose of the project was to investigate the utility and potential
application of sub-pixel processing and analysis for forest health
protection applications. The scope of the effort included not only
the project, but also an evaluation of how subpixel analysis could
potentially fit into the workflow associated with a project involving
standard image processing.
Three dates of Landsat Thematic Mapper images, from 1985, 1988,
and 1993, were acquired to have imagery of the area prior to the
outbreak, at the initial stage of the outbreak, and during the ongoing
outbreak of spruce beetle infestation. The imagery was processed
Subpixel Classifier software. The results were compared with
existing ground survey reports conducted in 1992 and 1994 with aerial
photographs flown of the area in 1991.
· Sketch mapping and ground survey have limited
utility in wilderness areas where ground access is difficult and
aerial mapping too costly
· The original study area consists of separated
stands of spruce in different states and environments different
facing slopes and meadows, no greater than 60 percent canopy closure
on average in major spruce stands, and abundant exposed bedrock
· The project required an examination of
scene-to-scene migration of the derived reference signatures
"Based on the preliminary results
subpixel analysis can
be applied to forest pest management situations. However, the application
should be handled on a pest-by-pest basis, as it is definitely not
a blanket approach to all the pests that impact the forest health.
It should be viewed as a supplement to existing means of detecting
and evaluating pest infestations.
Subpixel Classifier MOI classification on a southern portion
of the study area using signatures derived from 2 spectral clusters
with a Mean Material Fraction (MMF) of 0.50. Note the correspondence
between the classification and the sketch map ploygons. Polygons
and trees per acre mortality counts are from 1993 aerial sketch
mapdata of the Wastach Plateau area.
"There are several application areas in forest health monitoring
where subpixel analysis can play a role. In terms of large area
assessment, it could be used to monitor areas where aerial sketch
mapping and ground surveys are not conducted. Signatures could be
developed using known areas of infestation, then migrated and applied
to satellite imagery covering remote areas. In cases where pest
infestations are too extensive to map with aerial sketch mapping,
subpixel analysis of satellite imagery could be used to supplement
"For more species specific applications, Abies species and
some Pinus species hold discolored needles for long time periods
following pest attack. Such pests would allow for the potential
development of unique signatures specific to the infestation."
Images, information, and quoted text in the Pest Infestation section
are courtesy of USDA Forest Service and investigation participants:
Forest Health Technology Enterprise Team, Forest Health Protection,
Manti-LaSal National Forest and the Remote Sensing Applications
Center. For futher information see "An
Evaluation of the Utility of Sub-Pixel Analysis of Thematic Mapper
Imagery for the Spruce Beetle Outbreak on the Manti-LaSal National