8208

• UCREFRP Catalog Number: 8208
• Author: Osmundson, D. B., P. Nelson, K. Fenton and D. W. Ryden.
• Title: Relationships Between Flow and Rare fish Habitat in the '15-Mile Reach' of the Upper Colorado River.
• USFW Year: 1995.
• USFW - Doc Type: \
• Copyright Material: NO

available range within the river. Second, within the upper portion, they further concentrate in the
Grand Valley including the 15-mile reach. Within the 15-mile reach, there are specific subreaches
that are further selected. In the study reported here, we utilize these subreaches as study sites and
within these identify the preferred mesohabitat types.
Data Collection
Aerial Video and Habitat Mapping
On-the-ground mapping was used to quantify habitat area at various flows. Carter et al. (1985)
used this method to determine habitat changes with changes in flow in a stretch of the Colorado
River in the Debeque-to-Rifle area. In our study, prints from aerial video were used for base maps
for the habitat mapping rather than aerial photography. Aerial video was used because of the large
savings in cost of acquisition and processing (acquisition cost was approximately 65% that of aerial
photography). Bliesner and Lamarra (1995) are currently using this same videography-habitat
mapping technique on the San Juan River. Bureau of Reclamation (BR) acquired the aerial video
for our study and later quantified mapped habitat areas. A continuous image of the river was
recorded on eleven dates using a video camera attached to the front of a helicopter. Dates were
selected to provide a range of different discharge levels. At the time of each flight, USFWS was
responsible for mapping habitat within the sites; to do so, numerous vantage points along shore
and on the water were used. Color prints of video taken during the previous flight were used as
base maps on which to draw habitat boundaries.
Consistency in mapping technique was a crucial element in use of this methodology. Roper and
Scamecchia (1995) found that variation among individual mappers in classifying stream habitat
types was related to at least three factors: (1) the level of definition required in classification (e.g.,
pools in general versus specific types of pools), (2) the level and uniformity of observer training,
and (3) the stream channel characteristics. In general, they found that consistency among mappers
was poor without extensive and uniform training, and that repeatability was reduced when
elaborate habitat classification schemes (many types) were used despite the training provided.
To maintain consistent technique in our study, and thereby reduce one potential source of error, one
person (the lead author) was responsible for all habitat mapping during the study. This person also
identified habitats at radio-tagged fish locations in the same area during the earlier Osmundson and
Kaeding (1989) habitat use study. River area was broken into eight generic mesohabitat types:
pools, eddies, riffles, rapids, slow runs, fast runs, backwaters and flooded gravel pits (see Appendix
III, Table II for habitat type criteria). Aside from breaking runs into fast (>2 ft/sec) and slow (<2
ft/sec) types and dropping shorelines as a separate type, habitat classification followed that of the
previous habitat use study (see Osmundson and Kaeding 1989).
Mapping took two days to accomplish per discharge level; thus, the video imagery was acquired
while either the 15- or 18-mile reach was being mapped. Sites in the other reach were mapped the
day prior to or following the day of the video flight.
In the laboratory, habitat delineations were transferred from the field maps to hard-copy mosaics of
the video images taken at the time of mapping. Some adjustment to the positioning of site bound-
aries was done at this time for those habitats that could be clearly discerned on the images
(backwaters and gravel pits and in some cases riffles and rapids). No adjustment of boundaries
10