Heery and Cope: Co-occurrence of bycatch and target species in the groundfish trawl fishery 
41 
Table 2 
Results from hierarchical agglomerative cluster analyses (HCAs) of groundfish species and simulated, 
random fake species (“fakies”). Only species that formed groups more readily than fakies are included. 
Species that exhibited the strongest association with other species in their respective group are set in italic 
type. Groups A, C, and D included species commonly associated with the continental slope, and group B 
included species associated with shelf habitats. Note that the “skates” category included Longnose Skate 
from 2002 to 2004, causing it to cluster with slope species. Longnose Skate were recorded under a distinct 
species code starting in 2005. 
Species 
2002 
2003 
2004 
2005 
2006 
2007 
2008 
2009 
Dover Sole 
Sablefish 
Pacific Hake 
Arrowtooth Flounder 
Skates 
Longnose Skate 
Rex Sole 
Petrale Sole 
English Sole 
Lingcod 
Spotted Ratfish 
Pacific Sanddab 
Pacific Spiny Dogfish 
Longspine Thornyhead 
Shortspine Thornyhead 
Pacific Flatnose 
Pacific Grenadier 
A 
A 
A 
A 
A 
A 
B 
B 
B 
B 
B 
C 
C 
A 
A 
A 
A 
A 
A 
B 
B 
B 
B 
A 
A 
A 
A 
A 
A 
A 
A 
B 
B 
B 
B 
A 
A 
A 
A 
A 
B 
B 
B 
B 
B 
B 
C 
C 
A 
A 
A 
A 
A 
B 
B 
B 
B 
B 
B 
C 
C 
D 
D 
A 
A 
A 
A 
A 
A 
B 
B 
B 
B 
B 
A 
A 
A 
A 
A 
A 
A 
A 
A 
A 
A 
A 
A 
coastwide basis. The 2 major clusters identified in the 
run with combined areas over all years were gener- 
ally recognized annually as well (Table 2). Dover Sole, 
Sablefish, Pacific Hake, and Arrowtooth Flounder con- 
sistently clustered together in all years. Skates were 
also grouped with these species in 2002 through 2004. 
From 2005 through 2009, Longnose Skate was instead 
clustered with the Dover Sole+Sablefish group. This oc- 
currence is likely due to a shift in the way skate spe- 
cies were recorded in the observer data. Before 2005, 
Longnose Skate were given the species code for the un- 
specified skate category. 
Shortspine Thornyhead and Longspine Thornyead 
were only part of the cluster of slope species in some 
years during the time series (Table 2). In 2002, 2004, 
2005, and 2006, they formed a cluster with each other 
and with no other species. Petrale Sole and English 
Sole formed a consistent group in all years. Depending 
on the year, they were also grouped with Lingcod, Spot- 
ted Ratfish, Rex Sole, Pacific Spiny Dogfish, and Pacific 
Sanddab (Citharichthys sordidus) (Table 2). 
When ubiquitous species that formed an identifi- 
able assemblage on a coastwide basis were removed 
from the data and the analysis of all years of data was 
rerun, no further groupings were identified to cluster 
more readily than were the fakies. The same was true 
when identified assemblages were removed from the 
data and analyses were conducted separately for each 
year. 
Rebuilding bycatch species 
We used 2 data treatments to discern co-occurrences of 
rebuilding species with other species in trawl catch. In 
the first treatment, simulated fakies were introduced 
at frequencies of occurrence that matched those of the 
respective rebuilding species. This step was taken to 
evaluate whether the rebuilding species grouped in a 
better-than-expected manner on the basis of frequency 
of occurrence alone. Bocaccio, Cowcod, Canary Rockfish, 
Darkblotched Rockfish, Pacific Ocean Perch, and Widow 
Rockfish all formed groups with target species more 
readily than did fakies that were simulated on the ba- 
sis of the percent occurrence of each rebuilding species. 
This observation indicates the tendency for rebuilding 
species to cluster more than expected from their low 
encounter rates. Yelloweye Rockfish, which occurred 
in 0.6% of tows, was the only rebuilding rockfish spe- 
cies that did not form groups more than the simulated 
fakies. 
With the second treatment, we evaluated whether 
rebuilding species formed clusters with any other spe- 
cies, using only the positive occurrence data for the re- 
building species in question (i.e., only tows where that 
rebuilding species was encountered). An example for 
Canary Rockfish is presented in Figure 3. In all cases, 
this step resulted in one large cluster that inevitably 
contained the rebuilding species in question, along with 
those species that occur most commonly in the bottom 
