240 
Fishery Bulletin 119(4) 
results from most of the studies indicate a preference for 
mud, fine-grained sand, or soft bottom (Scott, 1982; Chang 
et al., 1999). However, these studies relied upon fishery- 
independent trawl data. Ames (2012) looked at historical 
seasonal movements based on interviews with fishermen 
and described fishing grounds not only as mostly mud bot- 
tom but also as around the ledges, rises, and banks of the 
northern GOM. Habitat and distribution of white hake 
changes with life stage (Macdonald et al., 1984; Chang 
et al., 1999), possibly contributing to some of the observed 
differences in catch composition for white hake between 
gears. Differences in survey catches could also be driven 
by behavioral factors that influence availability of the 
large white hake to different survey gears or by an inter- 
action between behavior, habitat, and the response to fish- 
ing gear presence. The use of baited hooks may attract 
large white hake that avoid trawls behaviorally or have a 
habitat preference for structure. 
Another possibility is that small-scale habitat varia- 
tion (e.g., a few boulders or a wreck in a muddy area) may 
facilitate trawl gear avoidance. In a study in the Gulf of 
St. Lawrence, white hake were captured in channel bot- 
toms, but the highest rate of capture was on deepwater 
slopes (Nozéres et al.°). This preference for slope habitat is 
consistent with what was observed in the LLS, which has 
a bottom-type stratification derived from depth data and 
includes many of the sloped edges of the ledges, banks, 
and other structured habitats in the GOM (McElroy et al., 
2019). The actual bottom type can vary greatly because 
the LLS classification is broad within the 2 categories of 
rough and smooth. The results of our study indicate at 
least some presence proximal to structured or slope habi- 
tats in the deeper waters of the GOM for large white hake, 
a habitat preference that may influence their availability 
to the BTS. 
Habitat preference of Atlantic cod may have less influ- 
ence on their availability to longline gear compared with 
their availability to bottom trawl gear. Cod are known for 
their preference for sand, gravel, rocks, ledges, slopes, and 
wrecks (Scott, 1982; Fahay et al., 1999; Wieland et al., 2009). 
However, they are also captured on other bottom types, 
such as mud (Macdonald et al., 1984; Wieland et al., 2009). 
Scott (1982) found higher concentrations of cod on sand and 
gravel bottoms on the Scotian Shelf but reported there was 
not a high preference. In a study in which multiple gear 
types were used, Wieland et al. (2009) found that cod prefer 
rough-bottom habitat, with some variation by season and 
gear type potentially influenced by seasonal movements. 
The authors also found no relationship between bottom 
type and cod size but cautioned that the study could not 
explore other factors, such as current strength or prey dis- 
tribution, that could affect cod distribution. The results 
of the analysis presented here between the BTS and LLS 
8 Nozéres, C., J. Gauthier, H. Bourdages, and Y. Lambert. 2015. 
Review of white hake (Urophycis tenuis) in the northern Gulf 
of St. Lawrence in support of a recovery potential assessment. 
DFO Can. Sci. Advis. Sec. Res. Doc. 2015/076, 56 p. [Available 
from website.] 
indicate that there was no difference in the size composi- 
tion for the upper end of the size distribution. Cod appear to 
lack a size-related habitat preference sufficient enough to 
limit their availability to the BTS or to limit the survey's 
ability to capture the full size range of the individuals pres- 
ent. It remains possible that size-related habitat preference 
is a function of population density and patterns observed 
now reflect the current low abundance. It seems reasonable 
that under conditions of reduced abundance, cod would be 
expected to concentrate in their preferred habitats (e.g., 
rocky habitat), but this notion was not supported by the 
catches of the LLS. 
The LLS is an example of successful cooperative 
research that can benefit assessments and management. 
Although the LLS time series is short compared with that 
of the BTS at this time, it will be long enough in the com- 
ing years to be evaluated within a stock assessment model. 
Analysis of data from the LLS may result in additional 
indices of abundance for a few of the common fish species 
in the region. The emphasis on more structured habitat 
will particularly benefit species strongly associated with 
those habitats, including a number of data-poor species 
and those considered species of concern by the National 
Marine Fisheries Service, such as cusk (Brosme brosme) 
and thorny skate (Amblyraja radiata), which are both 
well represented in LLS catches. This fixed-gear survey 
represents a supplemental source of biological and demo- 
graphic information that can be examined in conjunction 
with data from the BTS to gain a deeper understanding of 
the status of fish communities in the GOM and to advance 
research and management. Greater diversity of compara- 
ble surveys in which different gear types are used also pro- 
vides data for a broader suite of species, further benefiting 
ecosystem-based management approaches. 
Conclusions 
The choice of selectivity for a fishery-independent survey 
is critical to a stock assessment because it can influence 
model outputs and subsequent management decisions. 
The results of the analysis in this study provide a unique 
exploration of selectivity patterns for 2 groundfish spe- 
cies through a comparison of catches from 2 overlapping 
surveys with different gear types and sampling stratifi- 
cations. Our results indicate that the size selectivity for 
white hake differs between the 2 surveys, with large white 
hake readily observed in the LLS but not well represented 
in BTS catches. These fish may not be available to the BTS 
because of habitat or other factors, such as gear avoidance. 
Our findings validate the continued use of a dome-shaped 
selectivity for the BTS in the white hake stock assessment. 
In contrast, results indicate that catches were similar 
between the 2 surveys for cod in the upper end of the size 
distribution. A lack of large cod in the LLS, which focuses 
on sampling rough-bottom habitat, indicates that a similar 
absence of large cod in BTS catches is not due to habitat- 
related availability. Therefore, our findings do not support 
the hypothesis of habitat-related bias in the catches of cod 
