MONITORING FOREST CANOPY ALTERATION AROUND THE WORLD
WITH DIGITAL ANALYSIS OF LANDSAT IMAGERY
Lee D. Miller* Darrel L. Williams
Professor of Forest Science Environmental Physical Scientist
Texas A&M University NASA/Goddard Space Flight Center
Abstract
Sufficient experience has gradually been accumulated to indicate that the greatest potential
for the sustained applications of Landsat-type imagery by the forestry profession is the mon-
itoring of areas where forest cover is subject to rapid alteration by either natural or manmade
activities. Preliminary tests of the hypothesis that forest canopy alteration can be readily
monitored with digital processing of Landsat MSS data have been completed. A collection of
small, individual research efforts involving forest sites from around the world are presented
to illustrate the types of forest canopy monitoring which can be undertaken, the approaches
available, and the potential results which can be achieved. All of these case studies deal with
digital analysis of Landsat imagery, but they vary considerably in complexity and types of
analysis procedures employed.
Overview
A substantial portion of the reported applications of satellite remote sensing to forest eco-
systems has dealt with attempts to identify tree species or forest types.! Sufficient experience
has gradually been accumulated to indicate that this may not be the most important applica-
tion of the Landsat-type satellites by the forestry profession. The normal progression of
events in the application of remote sensing to each new topic area has generally been “to de-
tect”, then “to identify’’ and finally ''to monitor". Usually the supportive research activity
has progressed from one plateau to another in this chain, often spending several years at each
level. Remote sensing research on the applications of satellite imagery to forestry has just
reached the third plateau (i.e., to monitor") and it promises to be the most fruitful.
Successional changes in forest land cover generally take 20 to 50 years or more. Airphotos
and their interpretation have proven quite successful for forest tree species and type inven-
tories, especially with the increased use of color photographic materials. Since ecological
succession from one forest type to another forest type occurs over long time-frames of tens
of years, the existing airphoto-based forest cover mapping programs will continue to be
“appropriate technology”. It seems unreasonable to assume that a viable, economically justi-
fiable program for the remote sensing of forest lands from space could be built around making
measurements at frequencies of once every five or ten years, regardless of how easy or
*A portion of the work reported upon was completed while serving as a Senior Post-doctoral Research
Associate of the National Research Council assigned to the Earth Resources Branch at the National Aero-
nautics and Space Administration’s (NASA) Goddard Space Flight Center.
IMead, R. A., and Meyer, M. P., 1977. Landsat digital data application to forest vegetation and land-use
classification in Minnesota. Proc. of the 1977 Machine Processing of Remotely Sensed Data Symposium,
IEEE Cat. No. 77CH1218-7 MPRSD. LARS, Purdue, Indiana. pp. 270-279.