264 
Fishery Bulletin 109(3) 
labile 1 
Number of stations sampled and hauls completed by region, year, and season, for trawl hauls and environmental variables 
obtained from CTD casts. Region: NC=north coast of California, GF=Gulfofthe Farallones; season: S=summer, F=fall; variables: 
TMP=water temperature, SAL=salinity, DEN=water density, PAR=photosynthetically available radiation, CHL=chlorophyll-a 
concentration, DEP=water depth, DIS= distance of station from shore. 
Region 
Year 
Season 
Stations 
Hauls 
TMP 
SAL 
DEN 
PAR 
CHL 
DEP 
DIS 
NC 
2000 
S 
1 
1 
1 
1 
1 
0 
1 
1 
1 
F 
2 
2 
2 
2 
2 
2 
0 
2 
2 
2001 
S 
2 
2 
2 
2 
2 
2 
2 
2 
2 
F 
6 
6 
5 
5 
5 
1 
5 
6 
6 
2002 
S 
6 
6 
5 
5 
5 
5 
5 
6 
6 
F 
10 
11 
10 
10 
10 
10 
10 
11 
11 
2003 
S 
11 
11 
11 
11 
11 
11 
11 
11 
11 
F 
11 
13 
13 
13 
13 
13 
13 
13 
13 
2004 
S 
17 
17 
17 
17 
17 
16 
17 
17 
17 
F 
21 
21 
21 
21 
21 
20 
21 
21 
21 
2005 
S 
22 
24 
24 
24 
24 
24 
24 
24 
24 
F 
17 
17 
16 
16 
16 
16 
16 
17 
17 
GF 
2000 
S 
4 
6 
4 
4 
4 
2 
4 
6 
6 
F 
6 
8 
7 
7 
7 
5 
2 
8 
8 
2001 
S 
5 
7 
6 
6 
6 
6 
6 
7 
7 
F 
3 
5 
4 
4 
4 
1 
4 
5 
5 
2002 
S 
6 
10 
9 
9 
9 
9 
9 
10 
10 
F 
5 
8 
6 
6 
6 
6 
6 
8 
8 
2003 
S 
5 
8 
6 
6 
6 
6 
6 
8 
8 
F 
7 
10 
10 
10 
10 
9 
10 
10 
10 
2004 
S 
5 
8 
5 
5 
5 
5 
5 
8 
8 
F 
4 
6 
5 
5 
5 
5 
5 
6 
6 
2005 
S 
5 
6 
6 
6 
6 
6 
6 
6 
6 
F 
4 
5 
5 
5 
5 
5 
5 
5 
5 
be undersampled because they avoid the net (Emmett 
et al., 2006), although at times they were captured in 
large numbers. Macroinvertebrates — primarily jellyfish, 
squid, ctenophores, and salps — although often abundant 
in the catch, were not consistently identified or counted 
and therefore are not included here. To account for dif- 
ferences in tow distance and duration, fish abundance 
was standardized to a volume of 10 6 m 3 for all hauls — 
a standard that is about equal to a typical tow of 30 
minutes at 5.6 km/h (3.0 knots). 
Individual hauls were often sparse in diversity, some- 
times containing only one or two species. For this rea- 
son, it was necessary to combine all hauls into larger 
sample groupings to run statistical tests and produce 
meaningful ordinations. Because sampling effort was 
not equal among stations, regions, or years, averaging 
the standardized abundance of each species (rather 
than pooling) was the appropriate method to cumulate 
hauls. For regional and seasonal comparisons, hauls 
at each station were averaged across all six years to 
obtain species abundance (averaged for each station) for 
each season and region. For interannual comparisons, 
hauls were grouped more broadly, by averaging for re- 
gion and season within each of the six years separately. 
The analytical methods we used were robust to the in- 
clusion of rare species and unaffected by zero values in 
the community matrix, so that the full species matrix 
was used throughout the study, thus avoiding arbitrary 
omissions. Standardized abundances were square-root 
transformed to mildly reduce the disproportionately 
large influence of highly abundant species in the com- 
munity analysis. 
Assemblage structure: tests and ordinations 
We used multivariate statistical tests and ordinations to 
search for patterns of community structure in space and 
time. PRIMER analytical software (vers. 6.1.6, PRIMER- 
E Ltd, Plymouth, U.K.) with PERMANOVA+ (Anderson 
et al., 2008) was used for all multivariate routines. We 
first tested for differences among main effects (regions, 
years, and seasons) and interaction terms by using a 
type-III permutational multivariate analysis of variance 
(PERMANOVA) with hauls averaged by region, year, and 
season in a three-way crossed design. PERMANOVA is a 
semiparametric group difference test directly analogous 
to multivariate analysis of variance but with pseudo-F 
ratios and F-values generated by resampling (permuta- 
tion) the resemblance measures of the actual data; thus 
it is less sensitive to assumptions of parametric tests 
that are frequently violated by community data sets 
(Anderson, 2001; Anderson et al., 2008). For all biotic 
