Asch and Collie: Changes in a benthic megafaunal community due to disturbance from bottom fishing 
443 
differences among depths, disturbance categories, and 
years, data on the abundance of these five species are 
not presented in this manuscript. However, information 
on trends in the abundance of these five taxa can be 
referenced in Asch (2006). 
All other noncolonial taxa were collectively exam- 
ined as a multivariate data set. Differences in the 
species composition of these noncolonial organisms 
among depth strata, years, bottom-fishing disturbance 
categories, and sites were explored by constructing 
nonmetric multidimensional scaling (MDS) plots in 
PRIMER 6 (Plymouth Marine Lab, Plymouth, UK). 
MDS plots were constructed with a Bray-Curtis simi- 
larity matrix containing information on the mean 
number of individuals of each taxon per photograph, 
aggregated across each transect (Clarke and Warwick, 
2001). To ensure that patterns in the MDS plot were 
not dictated solely by trends in the abundance of a few 
ubiquitous species, data were square-root transformed 
before this analysis and all other multivariate statis- 
tical routines (i.e., analysis of similarity [ANOSIM] 
and similarity of percentages [SIMPER]). ANOSIM 
tests were used to evaluate whether differences in 
noncolonial species composition between disturbance 
categories and years were statistically significant 
(Clarke and Warwick, 2001). In cases where signifi- 
cant differences were detected, the SIMPER routine in 
PRIMER 6 was used to determine the mean percent 
dissimilarity that each species contributed to differ- 
ences between disturbance categories (Clarke and 
Warwick, 2001). 
In addition to examining the effect of bottom-fish- 
ing-induced disturbance on species composition, its 
influence on two measures of noncolonial diversity was 
investigated. Jackknife species richness and Simpson’s 
index of diversity (1 — A') were calculated according to 
the methods described in Krebs (1999). These two 
diversity indices were selected, because 1) they are 
either unaffected by differences in sample size (e.g., 
Simpson’s index) or incorporate estimates of the num- 
ber of rare species that are unlikely to be sampled 
(e.g., jackknife species richness) (Clarke and Warwick, 
2001), and 2) together these indices measure several 
aspects of biodiversity (e.g., richness, heterogeneity, 
and evenness). The diversity indices were computed 
for five levels of disturbance at both shallow and deep 
sites. Data were binned into five disturbance levels 
to achieve sufficient sample sizes because the jack- 
knife species richness index requires that the size 
of samples is large enough to include at least half 
of the species in an area (Krebs, 1999). Only data 
from shallow sites collected during years when the 
NMFS satellite vessel monitoring program was in 
operation (i.e., 1998-2000) were analyzed. Linear and 
polynomial regressions were performed to evaluate 
the relationship between bottom-fishing disturbance 
and these two measures of noncolonial diversity. The 
most parsimonious regression model was selected on 
the basis of the results of the analysis of deviance test 
included in the R statistical package. 
Results 
Colonial epifauna 
At shallow sites, the cover of five out of six taxa of 
colonial epifauna differed significantly between undis- 
turbed and disturbed sites located, respectively, inside 
and outside of CA-II (Fig. 2). Hydroids were the only 
taxon of colonial epifauna whose percent cover was not 
significantly affected by disturbance at shallow sites. 
During most years, structurally complex taxa of colonial 
epifauna, such as sponges and bushy bryozoans, were 
more abundant at shallow, undisturbed sites, whereas 
encrusting taxa, such as encrusting bryozoa, cultch, and 
F. implexa, exhibited a higher percent cover at shallow, 
disturbed sites (Fig. 2). With the exception ofF. implexa 
whose cover at each shallow site remained fairly stable 
throughout the time series, all other colonial epifauna 
taxa demonstrated significant between-year variations 
in abundance at shallow sites. Several of these taxa 
showed marked changes in abundance beginning in 1997 
and 1998, indicating that it took at least two years for 
these species to respond to the establishment of CA-II 
or to the increased bottom-fishing effort at sites that 
remained open to mobile fishing gear. For example, 
sponges were absent from all photographs taken at 
the shallow site in CA-II during 1994 and 1996 but 
were seen in 97% of the photographs taken in this area 
in 1997 (Figs. 2 and 3A). Similarly, bushy bryozoans 
experienced a temporary spike in their percent cover 
at the shallow, undisturbed site in 1997. The cover of 
encrusting bryozoans was fairly similar at all shallow 
sites at the initiation of this study, but then began to 
increase in disturbed areas beginning in 1998. By the 
year 2000, 62% of the substratum at shallow, disturbed 
sites was covered by encrusting bryozoa (Fig. 3B). Tem- 
poral variations in the abundance of hydroids and cultch 
are unlikely to reflect changes in bottom-fishing effort 
because similar between-year changes were seen at all 
shallow sites regardless of closure status. 
The only two taxa that clearly exhibited significant 
differences between disturbed and undisturbed sites at 
the deep depth stratum were F. implexa and hydroids, 
both of which consistently had an elevated percent cover 
in undisturbed areas (Fig. 4). Bushy bryozoans also 
had a higher cover in 1997 and 1998 in deep areas 
that experienced little to no bottom fishing. However, 
this pattern was reversed in 1996 when deep, disturbed 
sites exhibited slightly greater cover of bushy bryozoans 
than deep, undisturbed sites. Cultch and encrusting 
bryozoans at deep sites were frequently more abundant 
in disturbed areas, but this trend was inconsistent be- 
tween years, causing ANOVA results to be only margin- 
ally significant (i.e., P=0.08 and P=0.07 for cultch and 
encrusting bryozoans, respectively). Sponge abundance 
was depressed at all deep sites from 1996 through 1998, 
and its mean percent cover never exceeded 0.5% during 
these years. Temporal variations in colonial epifauna 
cover appeared to be of lesser importance at deep sites 
than at shallow sites because encrusting bryozoans were 
