Heery and Cope: Co-occurrence of bycatch and target species in the groundfish trawl fishery 
45 
Table 4 
Species clusters produced by hierarchical cluster analysis and based on data that were stratified into 1° latitudinal intervals. 
The clusters are shaded or enclosed in solid or dashed lines to make the table easier to read. Group A (light gray), C (dark 
gray) and G (surrounded by dashed line) represent slope species, and group B (medium gray), D (enclosed in solid lines), E 
(surrounded by dashed line), and F (surrounded by dashed line) represent primarily shelf species. Only species that formed 
identifiable clusters were included in the table. 
Latitudinal intervals (in ° 
North) 
34- 
35- 36- 37- 
38- 
39- 
40- 
41- 
42- 
43- 
44- 
45- 
46- 
47- 
48- 
Species 
35° 
36° 37° 38° 
39° 
40° 
41° 
42° 
43° 
44° 
45° 
46° 
47° 
48° 
49° 
Dover sole 
A 
AAA 
A 
A 
A 
A 
A 
A 
A 
A 
A 
A 
A 
Sablefish 
A 
AAA 
A 
A 
A 
A 
A 
A 
A 
A 
A 
A 
A 
Shortspine Thornyhead 
A 
A A C 
A 
A 
A 
A 
A 
A 
C 
A 
A 
Longspine Thornyhead 
A A C 
A 
A 
A 
A 
C 
C 
c 
Pacific Grenadier 
C 
Pacific Halibut 
B 
A 
A 
Pacific Cod 
A 
Petrale Sole 
A 
B B 
B 
B 
B 
B 
B 
B 
B 
B 
B 
A 
English Sole 
B B 
B 
B 
B 
B 
B 
B 
F 
B 
B 
A 
Lingcod 
B 
B 
B 
B 
B 
B 
F , 
B 
A 
Bocaccio 
B 
B 
Sanddabs 
B 
B 
JT i 
Pacific Sanddab 
B 
F ) 
B 
F 
Spotted Ratfish 
A 
B B 
D 
D 
B 
A 
B 
B 
B 
A 
Pacific Spiny Dogfish 
B B 
D 
A 
A 
A 
Longnose Skate 
A 
ABA 
A 
D 
D 
D 
A 
A 
Skates 
A 
A 
A 
Pacific Hake 
A 
ABA 
D 
D 
D 
D 
A 
A 
A 
A 
A 
Rex Sole 
A 
B B 
D 
D 
D 
D 
D 
A 
A 
B 
B 
B 
A 
Arrowtooth Flounder 
D 
D 
D 
D 
D 
A 
A 
A 
A 
A 
A 
Splitnose Rockfish 
A 
D 
G 
G 
Pacific Ocean Perch 
G 
G 
Redbanded Rockfish 
G 
Greenstriped Rockfish 
E 
Canary Rockfish 
E 
in stark contrast to more recent fishery trends. Bo- 
caccio, Canary Rockfish, and Yelloweye Rockfish were 
declared overfished in the early 2000s. Landings of 
these species, particularly the latter 2 fishes, have 
since been tightly regulated, and many of the areas 
where they were known to have been caught have 
been closed to fishing. As a result of that regulation 
and of declining population trends, these species were 
extremely rare in fisheries catch observed from 2002 
through 2009. 
Despite their rarity, all rebuilding species except for 
Yelloweye Rockfish formed nonrandom groupings rela- 
tive to their observed percent occurrence. This find- 
ing indicates that even though the ability to predict 
bycatch of rebuilding species is limited because of their 
rare occurrence, there may be a possibility to pre- 
dict species composition specifically from tows where 
bycatch of rebuilding species occurs. A possible excep- 
tion may be the Yelloweye Rockfish, which grouped 
randomly even at low encounter rates. Further re- 
search is needed to identify whether bycatch of this 
species can be tied to other variables beyond species 
composition. 
When all species were included in the analysis, rela- 
tionships between rebuilding species and other ground- 
fish were not evident at large spatial scales. However, 
some groupings occurred when data were used from a 
smaller geographic range. Specifically, rebuilding rock- 
fish formed groups most readily when the data were 
analyzed separately for each port group. It was difficult 
to tell whether the clusters formed in these cases were 
an artifact of error introduced by a smaller sample size, 
or whether the data were simply isolated temporally 
and spatially in a way that effectively allowed the rec- 
ognition of the temporal and spatial structure of spe- 
cies assemblages. The propensity of rebuilding species 
to form clusters appeared to be a function of sample 
size. When the subset of data used had a small sample 
size but included catch of rebuilding species, bycatch of 
those species became less rare. That said, it is possible 
that because of the use of departure port, rather than 
standard latitudinal intervals, to stratify the data, the 
