Somerton et al: The effects of wave-induced vessel motion on the herding of Limanda aspera 
31 
2.8 - 
2.6 - 
o 
“i-1-1-1-1-1-r 
1.2 1.4 1.6 1.8 2.0 2.2 2.4 
Wave height (m) 
0.6 - 
o 
1.2 1.4 1.6 1 8 2.0 2.2 2.4 
Wave height (m) 
Figure 7 
(A) Swept-area biomass, in metric tons (t), of yellow- 
fin sole (Limanda aspera ) estimated by the bottom 
trawl survey conducted in the eastern Bering Sea 
by the National Marine Fisheries Service during 
2005-2016, plotted against the sum of the estimated 
wave and swell heights, in meters, averaged over 
all stations where yellowfin sole were caught. The 
slope of this regression line is significant (P=0.028). 
(B) Deviations between the swept-area biomass 
estimates and the biomass predicted by the stock 
assessment model for yellowfin sole in relation to 
wave height. The slope of the regression line is sig¬ 
nificant (P=0.044). 
significant, the increase in the off-bottom distance at 
the center and corner portions of the footrope indicate 
that escapement is likely also affected by sea state. 
Besides their effect of changing mean distance off- 
bottom, the vertical oscillations undoubtedly have a 
pronounced effect on fish escapement and herding be¬ 
haviors. The studies of both O’Neill et al. (2003) and 
Politis et al. (2012) show that the oscillations we mea¬ 
sured in terms of off-bottom distance also produce os¬ 
cillations in trawl speed. Such oscillations were shown 
to be the principal mechanism for the increase in es¬ 
capement from the codend mesh with increased sea- 
state, both by changing mesh geometry and by altering 
fish behavior (O’Neill et al., 2003). However, we have 
no direct evidence of how either the vertical oscilla¬ 
tions of the footrope and bridles on the survey trawl 
or the likely oscillations in speed influence the escape¬ 
ment or herding behavior of yellowfin sole. 
Influence of wave height on the herding of yellowfin sole 
Based on the results of the yellowfin sole herding ex¬ 
periment, the ratio of the standardized catches from 
long and short bridle tows clearly declined with in¬ 
creasing wave height. Because the paired design of 
the bridle experiment was intended to minimize all in¬ 
fluences other than bridle length, the decline in catch 
ratio shows that sampling efficiency declined with in¬ 
creasing wave height. We suspect this result is due to 
diminished herding. 
We proposed a model to predict how catch ratio 
could vary as a function of wave height to explore pos¬ 
sible mechanisms to explain the observed change in 
catch ratio. This model focuses on the change in bridle 
contact area with increasing SDH described earlier. 
Although the model does predict a decrease in catch 
ratio with increasing SDH, it explains only half of the 
variability in catch ratio that is explained by a simple 
linear regression on SDH. One reason for the poor fit 
is that the model considers herding to be a product of 
the bridle contact area and the efficiency of herding 
within this area. Therefore, parameters affecting the 
bridle area (L off ) are confounded with parameters af¬ 
fecting the efficiency of herding (h) within the bridle 
contact area, and a priori estimates of at least one is 
needed to obtain a tractable solution. The values we 
used for these two parameters were obtained from 
an experiment conducted in relatively calm condi¬ 
tions (Somerton and Munro, 2001) and may well be 
biased when applied to rougher sea-state conditions. 
Furthermore, the estimate of L o{{ was obtained from 
an experiment in which only standard length bridles 
were considered (Somerton and Munro, 2001). We ap¬ 
plied this value to bridles both longer and shorter than 
the standard length with some hesitancy, although in a 
similar study in which off-bottom distances were mea¬ 
sured along 3 lengths of bridle on a bottom trawl, this 
value was found to be reasonable (Somerton, 2003). In 
addition to these issues, we were concerned that the 
herding process during high waves may be fundamen¬ 
tally different from that in calm water. For example, 
when the bridle experiences increased tension and lifts 
off the bottom, we assumed in our model, that the ef¬ 
fect is to momentarily shorten the length of the bridle 
in contact with the bottom and thereby decrease the 
area of herding. However, the vertical motion itself 
