Íslenskar landbúnaðarrannsóknir - 01.09.1978, Side 146

144 ÍSLENZKAR LANDBÚNAÐARRANNSÓKNIR
ern to the castern cnd. High densilies of
(ish were lound in all parts of the lake, with
the highest heing in a dcep part near the
western end (transcct 2). A map showing
the transect courses is shown in Fig. 3.
Magnetic ttipe recordings were made of
both surweys.
ANALYSIS OF RECORDED HYD-
ROACOUS'I IC DA FA
Onc form of analysis of recorded hyd-
roacoustic data consisls of four major
parts: counting of detections, estimation
of effcclive sampling volumcs, ftsh density
calculations, and estimation ofítsh popu-
lalions. 'l'he íirst two parts are done by
viewing thc recorded echosoundings on an
oscilloscope and by physictdly counting
fish detections, while the last two parls are
calculated from the counts. I’lie
techniques used to estimate flsli popula-
tions have been rcported thoroughh
elsewhere (Nunnallee and Mathisen
1972a. Nunnallee and Mathisen 1972b .
NunnalleecI al. 1973 . Nunnallee 1974a .
Nunnallee 1974b). A briefoutline of the
procedure is given in the next several scc-
tions.
COUNTING OF FISH DETECTIONS
The basic method of counling echo re-
turns from fish is to play the recorded echo
soundings and to display them on an os-
cilloscope where individual fish detections
can be counted within any desired depth
interval. Depth is proportional to the
elapsed time after the transmitted pulse
when an echo of interest is detected (1.37
ms = 1 sonar meter). However, before
echoes can bbe counted, a threshold must
be calculated for all depths and only those
echo amplitudes exceeding the threshold
are counted. The intensity of sound per
unit area is reduced by geomeiric spread-
ing in proportion to the reciprocal of dept
(1 /D); therefore, an echo return is prop-
ortional to 1/D2. The time varied circuit
(TVG) of the echo sounder acts to in-
crease the sensitivity of the receixer in
proportion to depth, but this is only halfof
the total correction and the return from
any specific target size will slill appear to
decrease proporlional to 1/D. The count-
ing threshold then is computed to repres-
ent the theoretical echo amplitude from
the smallest detectable target at the
maximum depth of interest, if it were
moved from thal depth to the surface. In
this way, the counting threshold repres-
ents a depth normalized minimum target
amplilude sensitivity. Since thc threshold
is based on the smallesl detectablc targct
at the maximum depth of interest, it is
necessary only to measure the peak vol-
tage of the ambient noise at thatdepth and
subslitute il into the following cquation.
E. = threshold voltage
D = depth (metérs or as ms)
K = constant
As an example, if the peak voltage of the
ambient noise = 0,03 V at 50 m depth,
then by substitution into equation 1, k =
1.5 and E. can then be found for all shal-
lower depths. In practice, the counting
threshold is scribed on clear plastic and
placed over the oscilloscope screen.