International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sci
2. PRACTICAL WORK
2.1 Fieldwork
The Fraser River drains 250,000 km? of western Canada, rising
in the Coast, Columbia and Rocky Mountains and flowing
1.300 km to the Fraser delta in the vicinity of Vancouver,
British Columbia. The river regime is typified by both high and
low flows created by annual snow melt cycles. Chilliwack is a
small town located 100 miles east of Vancouver where the river
is both an important habitat for salmon and a resource for
aggregate extraction for the construction industry. Several field
sites close to Chilliwack have been studied intensively by the
research centre at The Department of Geography, University of
British Columbia (Church, 2004) and one of these field sites,
known as “Queens Bar”, was selected for this particular study,
(Figure 1) ;
Figure 1- Queens Bar, British Columbia, Canada
Fieldwork at the Queens site was carried out in March and April
2002, before the spring melt and during a period when river
flows were at their seasonal lowest. An area of exposed river
bed that occupied an area 120 x 80 metres was selected. Criteria
for selection included the presence of a range of sub-regions
occupied by particle mixtures with differing size distributions
(facies). Intensive fieldwork was carried out on site during a
two week period and generated the following data sets:
1. A high resolution digital elevation model (DEM)
covering the entire study area. Two resolutions of DEM
were measured: | m resolution DEM within a small
central portion (60 by 40 m) and; a 2 m resolution DEM
for the remaining part of the 120 x 80 m study area;
2. Conventional grain-size parameters were obtained
within 96 sampling units covering the entire test area,
each being 10x10 m (Figure 2).
3 Ground observations identified the boundaries of fifteen
sub-areas which could be classified as type sites using
descriptive terminology (i.e. sand, sandy gravel, pebble,
clean gravel, cobble).
Both the 1 and 2 m resolution DEMs ( Dataset 1) were measured
using a motorised Total Station (Leica TCA1105), complete
with the remote control unit (RCS1100). The DEMs at the two
resolutions consisted of approximately 3,000 points. The plan
view of the test area (Figure 2) includes each data point
measured and demonstrates the difficulty of measuring points
along what is perceived to be a straight line in the field.
The bed characteristics were measured within the 96 10 x 10m
blocks, using conventional Wolman (1954) sampling techniques
ences, Vol XXXV, Part B7. Istanbul 2004
(Dataset 2). This laborious work entails identifying 100
particles randomly within each quadrat and manually measuring
the *b" axis by passing the pebble through a sizing sieve. This
was a large undertaking, requiring the assistance of fifteen field
workers for a whole day. A clear weakness with this sampling
strategy was the transgression of sediment type boundaries
(facies) by the rigid 10 m quadrat boundaries. In an attempt to
overcome this weakness the irregular boundaries between a
series of facies types were measured and digitised using the
motorised Total Station also (Dataset 3). The sediment
characteristics within each sub area were determined also.
Additional surveying was carried out to determine the positions
of 20 photo-control targets distributed across the test area.
These targets were constructed locally by laminating à simple
black and white target design printed on a standard A4 laser
printer. Two sizes of target were used, ! 10 x 110mm and 180 x
180 mm. Angles and distances were measured to each target
using the Leica TCA1105 located at two survey stations. These
data were combined in a least squares “variation of coordinates”
estimation to derive the best coordinate estimates for these
targeted points.
Multi-scale digital imagery was acquired from a helicopter
using a Kodak DCS460 hand-held 35mm digital camera, with a
resolution of 3060 x 2048 pixels. Unfortunately weather
conditions were marginal during photo-acquisition, with low
cloud and light drizzle creating poor lighting conditions.
Imagery was acquired at 1:10,000, 1:5,000 and 1:3,000 photo-
scales using a 24mm lens and 1:2,000 and 1:1,000 photo-scales
using an 85mm lens. At such a diverse range of scales, stereo
coverage could be achieved using a varied number of frames
ranging between 2 images through to a block consisting of 24
images. The latter proved to be un-achievable because of the
time required to store each frame on the PCMCIA card, but
sufficient imagery was acquired at all scales except 1:2,000, to
provide coverage of one area of the site. Most of the 1:2,000
scale images were blurred, arising from long exposure times
caused by the poor lighting conditions. Additional oblique
imagery proved valuable in maintaining coverage at other
scales, image redundancy originally acquired for purpose of
camera calibration.
e
3
>
10 x 10m qudrat
2m DEM point
1m DEM paint
Figure 2- DEM points and 10m sampling quadrats
1080
On
cor
phe
usi
cali
Au
a D
DE
pro
LM
was
ach;
3.1
Onc
"sup
