McElroy et al.: Comparison of length distributions from a longline survey and a trawl survey for 2 groundfish species 233 
Fish 2100 cm TL (%) 
% © D AV 4O A Nv ok A DD OM VM wo SW we A 
PEE IIL BP LPF FFF, SF HF SF ri ras 
Year 
oS 
oS 
— 
i 
= 
(S) 
(o) 
Oo 
2 
Al 
f= 
2 
LL 
A 
% © D Av AO AD OW Ob \ 
O”? LO” gO” AY SN? SV? DD YD 
DB D> D> D> D>." BP 
S © 2 dO 
HP DB? MP? GY YP VPS 
PSS? AS? GH HP OP 
Year 
Fish 290 cm TL (%) 
x 
x 
RS 
Year 
Fish 290 cm TL (%) 
5 
0 T T T T T T T T 
Ne 
% © DAV AN KA DD GOO DMO OM 
SLI IN PPP PF PPP PF gS 
Year 
Percentage of large Atlantic cod (Gadus morhua) (2100 cm in total length [TL]) (left panels) and white hake (Urophycis tenuis) 
(290 cm TL) (right panels) captured in the Gulf of Maine (A and B) in the NOAA Northeast Fisheries Science Center bottom 
trawl survey (BTS) during 1963-2018 and (C and D) in the commercial fishery, based on commercial landings (CL) data used in 
the stock assessments, during 1982-2018 for Atlantic cod and during 1989-2018 for white hake. Because of changes over time 
in size regulations of the fishery for the Gulf of Maine Atlantic cod stock, the percentages represented are relative to the total 
number of individuals >50 cm TL. The percentages of white hake are relative to fish of all sizes because there is no size limit 
for this stock. 
distribution of cod has changed over the last 50 years and 
that some portion of the stock (e.g., large cod, >100 cm TL) 
has shifted to habitat that cannot consistently be sampled 
as part of the BTS. 
Although the BTS research vessel has an otter trawl 
outfitted with large roller (rockhopper) gear 40.6 cm in 
diameter (Politis et al.°) to accommodate the variety of bot- 
tom habitat types in the sampling area, towing of the trawl 
gear in highly structured habitat is limited on the basis of 
the likelihood of the gear getting stuck on the rugged bot- 
tom. Consequently, organisms that inhabit rough-bottom 
habitats are less available to BTS sampling. There is a 
concern, particularly among industry stakeholders, that 
large cod may have a reduced distribution and are lim- 
ited to hard-bottom habitats that are difficult to sample 
with mobile trawl gear. If large adult cod have moved to 
hard-bottom locations, it would result in a change in their 
vulnerability to the survey gear and could explain the 
truncation of the size distribution evident in the BTS data. 
Although adult cod are known to inhabit areas with sub- 
strates of sand, gravel, or mud, they do have an affinity 
for rocky bottoms, and their distribution among habitats 
can vary among seasons and sizes (Scott, 1982; Fahay 
® Politis, P. J., J. K. Galbraith, P. Kostovick, and R. W. Brown. 2014. 
Northeast Fisheries Science Center bottom trawl survey proto- 
cols for the NOAA ship Henry B. Bigelow. Northeast Fish. Sci. 
Cent. Ref. Doc. 14-06, 138 p. [Available from website.] 
et al., 1999). If the spatial distribution of large cod in the 
GOM has changed, it could indicate that BTS selectivity 
for cod may be more dome-shaped, like the selectivity for 
white hake, rather than flat-topped, which is the type of 
selectivity used in the stock assessment for cod (NEFSC, 
2013a, 2017). The consequences of incorrectly specifying 
a flat-topped selectivity curve when a dome-shaped curve 
is warranted would be an underestimation of stock abun- 
dance and biomass. 
In 2014, the NEFSC started a bottom longline survey 
(LLS) to address concerns about the presence of fish spe- 
cies in rough-bottom habitat in the GOM that may experi- 
ence low encounter rates with the trawl survey. The LLS 
was designed following a stratified random design similar 
to that of the BTS, but the LLS survey area was further 
stratified by bottom type, with classification of smooth 
and rough bottom (McElroy et al., 2019). Because of the 
concentrated effort of the LLS in rough-bottom habitats, 
data from the LLS provide a perspective on availability of 
groundfish in the GOM complementary to the BTS. 
With respect to cod, the LLS stratification design offers 
an opportunity to evaluate the presence of large cod in 
the rough-bottom areas of the GOM. If data from the LLS 
indicate that large cod are more prevalent in the rough- 
bottom habitats than across the BTS sampling area, it 
would provide support for considering dome-shaped selec- 
tivity for the BTS in the GOM cod stock assessment. On 
the other hand, if large cod are not caught at a higher rate 
