 |
| Radar remote sensingto supporttropical forest management |
This text describes an investigation into the potential of radar remote sensing for application to tropical forest management. The information content of various radar images is compared and assessed with regard to the information requirements of parties involved in tropical forest management at the global, national and local spatial levels.
The study distinguishes between the use of radar remote sensing for
application to forest resource assessment and forest resource monitoring. Both
assessment and monitoring are essential components of procedures for sustainable
forest management. The radar data studied are of tropical forest areas near the
township of Mabura Hill in Guyana and the city of San José del Guaviare in Colombia.
Mabura Hill is comprised of differing intact, primary forest types and forests that have
been subjected to industrial selective logging. San José del Guaviare, on the other
hand, is characterised by the presence of secondary forests and a variety of
non-forest cover types. The available radar data set includes high resolution airborne
radar images with differing wavelengths (i.e. X-, C-, L- and P-band) and polarizations,
time-series images acquired by the first European remote sensing satellite ERS-1 and
a collection of low altitude, nadir-looking, X-band scatterometer measurements.
The study makes use of three fundamentally different information sources from the
radar return signal: its strength or back-scatter, polarization and phase, and spatial
variability or texture. Results show that back-scatter values computed from L- and
P-band radar data and textural attributes computed from high resolution X- and
C-band radar data make modest to good and complementary bases for region-based
classification of tropical land cover at the level of primary forest types. Textural
attributes and back-scatter values computed per region from mono-temporal ERS-1
images make modest bases for classifying at the levels of primary forest, logged-over
forest, secondary forest and non-forest and poor bases for classifying at the level of
primary forest types.
Roads are usually the most easily observable indicators of
foregoing and/or forthcoming (selective) logging and other human activities in ERS-1
images. Detection of change in road networks by means of ERS-1 images would
make a good first step in forest resource monitoring at the national spatial level, in
particular. Textural attributes enable the ranking of forest types according to the
degree of canopy roughness. Specific textural attributes also allow for quantification
of canopy architectural properties.
Despite differences in measurement scale, the
canopy roughness of the land cover types studied was found to appear similarly in
the texture of the available space borne and short wavelength airborne radar images.
Keywords: remote sensing, radar, tropical rain forest, forest resource assessment,
forest resource monitoring, sustainable forest management.
Download Link
Posts
No comments:
Post a Comment