Íslenskar landbúnaðarrannsóknir - 01.09.1978, Page 67

DESIGN OF A SALMON COUNTER 65
Fig. 7. Summated hourly count for Ellidaár and
Grimsá 1976.
1. The effect of dense runs of salmon.
Hellawell (1973) found no photographic
evidence of simultaneous passage of sal-
mon over the counter electrodes. How-
ever, data gathered in the last 2 years
show that the salmon in the Ellidaár, and
probably in most Icelandic streams, mig-
rate in dense runs during the night (Fig.
6). Up to 100 signals have been recorded
in 1 hr. Since the sensor mat can be up to
10 m in with, it cannot be ruled out that a
number of fish may pass over at short
intervals. There is a ,,dead“ time follow-
ing each count, and the counter will only
distinguish between fish that pass over
1-2 sec. apart.
2. Swimming behavior.
False counts may arise if the fish do not
swim straight over the mat, but hesitate or
swim in a zig-zag pattern. This seems to
occur especially where stream flow is very
slow.
3. Wandering.
In late summer and autumn there is the
problem of salmon milling in the stream.
The data from the Ellidaár show that
there is an increase in the total activity
recorded by the counter, although migra-
tion of fish into the stream is decreasing.
This increases considerably the chance of
false counts and errors in the summed
total. If one accepts that the cumulative
number of fish ascending the stream fol-
lows a sigmoid curve, there is a cutout
point in Ellidaár between 80 and 90%
beyond wich the counter cannot be used.
In other streams where the counter is
placed at a location where all fish are
moving upstream and no spawning is
taking place, this restriction is removed.
9