Schetselaar, Ernst
Paleoproterozoic metamorphic rocks in the study area display a range in lithologies from granulite
facies mafic and quartzofeldspathic gneisses, granitoids, and mid to upper amphibolite supracrustal
assemblages of Paleoproterozoic age (1857 - 1827 Ma; Wodicka and Scott 1997). Regional
tectonostratigraphic subdivisions in the study area include two major subdivisions. A footwall
othogneiss complex referred to as structural Level-2 and a hanging wall supracrustal assemblage of
psamites, pellites and quartzites with minor marbles and peridotite, which constitutes structural Level-
3. The supracrustal suite in this area is an outlier of the Markham Bay assemblage. This cover unit is
also variably intruded by monzogranites from the Cumberland batholith suite. The boundary between
these two regional classes is likely to be a regional scale découlment or thrust which has juxtaposed the
supracrustal Level-3 rocks against Level-2 orthogneisses and tonalites (St-Onge et. al. 1996a, 1996b).
For this study the Level-2 package is referred to as ‘basement’ and the Level-3 rocks as ‘cover’
(Figure-3).
Figure-3: General tectonostratigraphic subdivision of bedrock map units. Blue areas underlain by structural Level-
3: supracrustal rocks of the Markham Bay assemblage (garnetiferous psammite, semipelite and quartzite)
intruded by 1.85 Ga Cumberland Batholith monzogranites. Pink areas represent structural Level-2:
orthogneisses (orthopyroxene — biotite tonalite, diorite and monzogranites). The boundary between Level-2
and Level-3 rocks is a Paleoproterozoic thrust. The thrust, and local associated imbricates, was subsequently
polydeformed through two ductile folding events that result in dome and basin structures (central study area)
and complex mushroom shaped interference patterns (southwest quadrant of the study area). Dips of the
Level-2/Level-3 boundary thrust are generally steeper than topography. Average range of dips is 45° to 75°.
30 Km A
3 DESCRIPTION OF THE DATA AND PREPROCESSING
The study employed five main datasets (Figure-4). I) Digital topographic data in the form of CDED
(Canadian Digital Elevation Data) gridded at 60-70 meters per pixel. This data was derived form 1:250
000 scale topographic contours. IT) Geological map data in GIS format at 1:100 000 scale compiled
from recent mapping. III) Digitized bedrock traverse lines used for the training locations, IV) 7 bands
Landsat 5 Thematic Mapper (TM) imagery from August 15, 1995 and V) Total Field Magnetic data
flown at 800 meter flight line spacing and gridded to 200 meters cells.
Vertices from traverse lines were used to sample the ‘basement’ and ‘cover’ classes from the
geological map. These points were saved with coordinate and class information in a sequential ASCII
file that could be used during the training phase. All data sets were gridded to a common pixel size
(25.853 meters) and clipped to a common image dimension (2243 rows x 2173 columns). In addition a
data mask was created from the addition of lakes, quaternary overburden and vegetated areas. The
mask is represented as black areas on the TM colour composite image (Figure-4C) and as white areas
on the classified images (figure-5). The vegetation component of the mask was developed using a GVI
International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B7. Amsterdam 2000. 1327