Columbia, 1990) is acceptable as a primary
source of that data and, with the addition of
features specific to MacMillan Bloedel’s
needs (cadastre, forest inventory), will form
a new spatial database for the GIS.
Hardware and software were acquired that
will facilitate loading, enhancing and revising
of the TRIM data while maintaining accuracy
and topological integrity.
2. FOREST BASE MAPPING
2.1 Traditional Forest Mapping Procedure
Annual revision of the mapbase is an
important function as MacMillan Bloedel
harvests approximately 10,000 hectares of
forest each year and the resultant land
changes (new roads and cutblocks) must be
added. Although this process is a corporate
requirement, the responsibility for
documenting these changes rests with each
of the ten individual woodlands divisions.
These divisions have traditionally used time-
consuming field sketching techniques to
update paper copies of the current mapbase
that may or may not reflect the previous
revision period(s). There has not been any
standardized compilation techniques or
understanding of data quality at this level. In
addition, the logistics of coordinating the
data sources, acquiring the data and creating
source documents for digitization has been
complicated and labourious. This method for
performing the annual Inventory Revision has
been identified as inadequate for the current
mapbase and is totally unsuitable for the new
spatial database. Aerial photography has
never been implemented as an annual data
source due to the cost and difficulties in
coordinating flights over such a large area.
2.2 Forest Mapping and Satellite Remote
Sensing
Satellite imagery, both optical and radar, has
been identified as a viable data source for
MacMillan Bloedel's map revision for a
number of reasons. Firstly, the precise
correction of satellite data to acceptable
accuracies for operational mapping has
become possible due to the availability of the
448
digital TRIM data, especially the Digital
Elevation Model (DEM). Secondly, the
proposed revision process, by the very nature
of the data source and its unique handling
requirements, will be a centralized process
with standardized compilation techniques
performed by personnel with an appreciation
for data quality and integrity. Thirdly, spatial
coverage of the imagery will allow updating of
large areas in one session at a considerably
lower cost than with aerial photography.
Fourthly, the re-visit capabilities of the
satellites will permit revision on a more timely
basis. Lastly, radar imagery in particular will
permit all-weather acquisitions. This is an
important factor as many of MacMillan
Bloedel's operating areas are affected by the
cloud and fog conditions that are typical of
British Columbia's west coast climate.
The methodologies developed will allow
MacMillan Bloedel to revise it's new spatial
database in an economic and timely manner
with an accuracy that, while not necessarily
to TRIM mapping specifications, will be a
known quantity that can be documented and
applied to the decision making process. As
satellite data spatial resolution improves, data
compiled at less than the ideal specification
can be replaced.
3. DATA SETS
Remote sensing data were collected over
Louise Island in the Queen Charlotte
Archipelago (52?56'N, 131?45'W). The data
set consists of satellite Synthetic Aperture
Radar (SAR), airborne SAR, simulated satellite
SAR (using airborne SAR as input data), and
satellite optical data.
The optical data set includes four bands of
imagery from the French SPOT satellite.
SPOT satellites carry two high resolution
visible imaging systems (HRV). Each HRV
system has two imaging modes: 1) the 10m
spatial resolution panchromatic mode (PLA)
with a spectral range of 0.51 to 0.73 um; 2)
and the multi-spectral mode (MLA) operating
in 3 bands at 20 m spatial resolution and
spectral ranges of 0.50 to 0.59 ym, 0.61 to
0.68 um and 0.79 to 0.89 um. For the
presen:
27,16
1992.
The SZ
from t
satellite
system
Resolu!
Narrow
acquire
cm w
wavele
with a
inciden
range t
lines \
RADAF
directic
One EF
10, 1€
Mode (
SAR o
polarize
permitt
angle «
resoluti
mode oa
imaginc
close tc
(Figure
RADAF
images
polarize
data wv
SARPA
Wessel
to-9 x
altered
charact
The ba
Provinc
digital
planime
the NA
Figure |
PLA, (3
C/X SA