9710

• UCREFRP Catalog Number: 9710
• Author: Connolly, P.J., I.G. Jezorek, K.D. Martens and E.F. Prentice.
• Title: Measuring the performance of two stationary interrogation systems for detecting downstream and upstream movement of PIT-tagged salmonids.
• USFW Year: 2008.
• USFW - Doc Type: North American Journal of Fisheries Management
• Copyright Material: YES

4]4

CONNOLL Y ET AL.

column coverage could have a high detection efficien-
cy and good precision for both upstream- and
downstream-moving fish, but if one antenna fails, the
capability of deciphering directional movement is
much diminished.
Because we wanted information on the movement of
resident and anadromous fish in relatively remote
locations, we developed a stationary, continuously
operating PIT tag interrogation system for use in free-
flowing streams. Although traps and weirs can be used
to obtain similar life history information. these methods
are expensive to operate because of staffing needs, and
can be difficult to operate year round due to high flow
and debris loads. The antennas that we constructed
could be placed in a variety of configurations and are
highly adaptable to the challenges of stream environ-
ments. High gradient or high velocity will make for
more difficult and riskier deployments. Our antennas
proved to stand up to rigorous conditions of flow and
debris loads, but several did become dislodged in
Rattlesnake Creek upon extreme conditions of flow in
winter 2006 (estimated maximum flow, 41.1 m3js)
after almost continuous detection ability through the
previous two winters. The Beaver Creek system nm
from September 2004 until unusually high flows in
spring of 2006 (>8.86 m3js) disabled most of the
antennas. Because of redundancy in arrays and
redundancy of antennas within a.ridYS, the retention of
some or most antennas allowed some level of
continuous monitoring of fish movement, though
ability to determine detection efficiency and direction
of fish movement was not possible when at least one
upstream and one downstream antenna was not
maintained.
Since beginning our project in Rattlesnake Creek in
2001, we have improved the anchoring systems. We
replaced nylon cord tied to anchors with heavy
webbing with metal cam buckles. To secure the
systems we are currently operating and deploying, we
increased the number of anchor points. Despite
upgrades in gear and amount of anchoring, we thought
it futile and too expensive to attempt to build a system
that could withstand all flows, especially flows that
initiate movement of the bed load that the system may
be anchored to. An interrogation system that is too
formidable may actually be harmful to a stream if it
causes debris jams and subsequent redirection of
stream flow.
Study goals, target species, and budget will dictate
the specific designs of interrogation systems. We
wanted an interrogation system that would differentiate
between upstream and downstream movement, and we
wanted to be able to estimate detection efficiency and
the precision of the estimate without the use of known

numbers of passing fish. We developed a protocol that
determined whether to include or exclude a fish that
was detected on a single antenna. While incorrect
assignment was possible, we believe that the adopted
protocol minimized it. It was also possible that the
calculations of detection efficiency underestimated the
number of tagged fish passing the antennas that did not
get detected on any of the antennas, especially during
high flows. Use of a known tagged fish population
passing the interrogation site to assess our derived
efficiency estimates was not feasible, because of cost
and permitting restrictions. Where possible. the use of a
known population of PIT-tagged fish, such as salmonid
smolts with strong one-way migratory tendencies,
would likely prove helpful. If direction of movement
is known and efficiency of detection can be empirically
determined, it becomes much simpler and more direct
to derive estimates of total fish passing the detector site
and to assign weights to particular life history
strategies.
Our efficiency calculations were derived with
passage information from trout (primarily rainbow
trout and steel head but also cutthroat and brook trout).
The fish in Rattlesnake Creek were resident juvenile
and adult trout from Rattlesnake and Indian creeks and
the White Salmon River. Historically, the White
Salmon subbasin supported anadromous salmonids,
but was blocked by Condit Dam at rkm 5.0 in 1913. In
Beaver Creek, there was a mixture of resident and
anadromous juvenile and adult rainbow trout and
steelhead and a few resident brook trout. We combined
some salmonid species in our analysis, but this may not
be justified in other studies. Smolt trap studies have
shown differences in capture efficiencies for various
salmonids (Thedinga et al. 1994) and between hatchery
and wild fish (Roper and Scamecchia 1996). Some of
the differences in capture efficiencies may be the result
of trap avoidance, but may also be related to position in
the water column. This may be particularly important
in large streams where fish may have deeper water
columns available to them.
At the time of this study, the electrical functionality
of the transceivers and the electrical properties of the
cables connecting the antennas to the transceivers
limited the size of the antennas that we could use to
about 3 m long X 1.25 m wide. This size limitation is
changing with new technology. The antenna size
limitation was addressed in our study design by adding
a second antenna within an array, thus allowing for
more complete spanning of the channel width. We
found that a three-array system allowed a good
measure of redundancy in case of mishaps. Some of
the problems that we or others have encountered
include power disruption (e.g., AC outages, battery