McElroy et al.: Comparison of length distributions from a longline survey and a trawl survey for 2 groundfish species 239 
Table 3 
Bhattacharyya coefficients (BCs) representing the degree 
of overlap in length distributions between the NOAA 
Northeast Fisheries Science Center bottom longline sur- 
vey (LLS) and bottom trawl survey (BTS) for Atlantic cod 
(Gadus morhua) and white hake (Urophycis tenuis) sam- 
pled in the Gulf of Maine during 2014—2018 by season and 
year and for all years combined. Data were selected from 
strata with rough bottom for the LLS and from strata used 
in the stock assessments for these species for the BTS 
(for strata used in the assessments, see NEFSC, 20138a, 
2013b). The BC may range from 0.00 (no overlap) to 1.00 
(identical distributions). 
Atlantic cod White hake 
Spring Fall 
Year Spring Fall 
2014 0.86 0.62 0.53 0.58 
2015 0.91 0.75 0.64 0.58 
2016 0.90 0.87 0.73 0.67 
2017 0.77 0.83 0.84 0.60 
2018 0.50 0.55 0.63 0.50 
2014-2018 0.96 0.89 0.77 0.65 
Discussion 
The intent of this study was to evaluate the presence of 
large cod (2100 cm TL) in the rough-bottom habitat of the 
GOM, as indicated by a comparison of catches of cod and 
white hake in the LLS and BTS conducted by the NEFSC 
in 2014-2018. Possible aggregation of large cod in rough- 
bottom habitat due to presumed density-dependent range 
contraction could warrant reconsideration of the flat- 
topped BTS gear selectivity currently used in the GOM 
cod stock assessment. Using white hake as a control, we 
evaluated whether the LLS and BTS data support use of a 
dome-shaped selectivity curve for cod in the BTS by com- 
paring length distributions of the 2 species in each survey. 
Our results indicate that large cod were not more prev- 
alent in the rough-bottom habitat sampled for the LLS 
than in the BTS sampling area; therefore, no evidence 
was found to support a dome-shaped selectivity pattern 
for cod in the BTS. On the other hand, our approach pro- 
vided a unique opportunity for validation of the selectivity 
curve used for white hake. Large white hake (290 cm TL) 
were sparse in the BTS but present in the LLS, support- 
ing the continued use of a dome-shaped selectivity for the 
BTS in the assessment for this species. 
Patterns of the size distributions for the 2 gadoid spe- 
cies in this study are distinct between the LLS and BTS, 
which sample different compositions of habitats. The 
largest divergence in the distributions for both species 
occurred at small sizes (<35 cm TL), which were present in 
trawl catches but poorly represented in longline catches. 
These differences were expected given the contrasting 
capture mechanisms of each gear. In particular, hook and 
bait size, as well as mouth gape, swimming ability, and 
other behaviors, limit the longline selectivity for small 
fish, and trawls capture most fish that cannot outswim or 
pass through the net (Engas et al., 1996; Millar and Fryer, 
1999). The efficiency of trawl gear at catching small fish 
increases when a lined codend is used, and the codend of 
the trawl net used in the BTS is lined. Cod at the upper 
end of the size range had similar distributions between 
the 2 surveys, with few large cod observed in catches with 
either gear. In contrast, large white hake were more abun- 
dant in the LLS than in the BTS. This pattern was rela- 
tively consistent among years despite some interannual 
and seasonal variation. Both gear types could have addi- 
tional selectivity characteristics that contribute to the size 
patterns observed. However, the presence of large white 
hake in the LLS, which is concentrated on rough-bottom 
habitats, confirms that this survey is capable of catching 
large gadoids that are not selected by trawl gear in the 
areas sampled for the BTS. Therefore, given the compa- 
rable low catches of large cod in both surveys, our results 
indicate that BTS gear selectivity or ineffective sampling 
in rough-bottom habitat are not likely the cause of the 
absence of large cod in the BTS in recent years. 
Commercial landings data support the observation of 
the absence of large cod in BTS catches in the GOM over 
a 40-year time frame. Big cod were infrequent through- 
out the period in both the BTS and commercial landings 
data sets, and when the occurrence of big fish increased, 
the length frequencies were still comparable between 
BTS catches and landings. Although the proportion of 
large white hake in commercial landings has varied over 
the past 20 years, these fish nevertheless represented a 
much greater proportion of the annual commercial land- 
ings than of the BTS catches in the same period. The size 
of fish in commercial landings for both species could be 
influenced by changes in market demand, fishing costs, 
and regulatory changes. The general expectation is that 
the market demand for large cod is always high for this 
quintessential New England seafood. Regulatory rules 
and low quotas could have been disincentives for cod land- 
ings in recent years and have contributed to the paucity of 
large cod in the commercial landings; however, these mar- 
ket and regulatory factors would not account for the lack 
of large cod in the BTS time series. In contrast, catches 
of large white hake have been consistently low across the 
BTS time series, and their proportional occurrence in the 
commercial landings is very high. These differences in 
the occurrence of large white hake have provided some 
of the justification for the application of a dome-shaped 
selectivity to the BTS time series in the assessment for 
white hake (NEFSC, 2013b). Although neither gear type 
provides complete selectivity across all fish lengths, data 
from the LLS provide complementary and independent 
validation that habitat-related availability to the BTS is 
likely for large white hake. 
The size distribution of white hake in the LLS may 
reflect preference for rough-bottom habitat, particularly 
for large individuals, and fine-scale differences in station 
locations relative to rough-bottom habitat. Studies charac- 
terizing the habitat of white hake have been limited, but 
