Bacheler et al.: Influence of soak time and fish accumulation on the catches of reef fishes 
219 
tats (e.g., seagrass meadows and coral reefs), and are 
sometimes the only feasible method of sampling in 
these habitats because of their relatively low effect 
on substrate and benthic communities (Miller, 1990). 
Clearly, the catch of fishes or invertebrates in traps 
cannot continue to increase linearly with soak time 
because the space inside a trap is finite and will 
eventually become filled with animals to the point 
at which no additional individuals can enter (Ben- 
nett, 1974; Austin, 1977; Miller, 1990). Models have 
been developed to describe the relationship between 
catch per trap and soak time, with the intention of 
using those models to standardize catch for differ- 
ent soak times (Munro, 1974; Somerton and Mer- 
ritt, 1986; Zhou and Shirley, 1997). Unfortunately, 
these approaches to standardization do not account 
for landscape (e.g., depth) or environmental effects 
(e.g., water temperature) on catch. Nor do they help us 
understand how the catch of one species may be influ- 
enced by the catches of other species. 
One recent method to examine these drawbacks 
has been to model the catch of a species as a func- 
tion of soak time and fish accumulation (i.e., the to- 
tal number of individuals of all species caught by a 
fishing gear; Olin et al., 2004) after accounting for 
other variables that may influence catch. In other 
words, catch rates can first be standardized by all of 
the predictor variables in the model building process 
(Lo et al., 1992; Maunder and Punt, 2004), and then 
the specific effects of soak time and fish accumulation 
can be extracted and examined independently of other 
predictor variables (Li et al., 2011). For example, Li 
et al. (2011) showed that gill nets became saturated 
with Yellow Perch ( Perea flavescens ) because of the 
total number of individuals caught in a gill net but 
not because of increased soak times. These results in- 
dicate that Yellow Perch catch rates decline when, for 
instance, this species sees fish already caught in the 
gill net, and not for reasons associated purely with 
increased soak time (i.e., when all the Yellow Perch in 
an area are caught, which takes some time). 
In our study, we used a standardized catch approach 
to examine the influence of soak time and fish accumu- 
lation on the catches of several reef fish species from 
long-term fishery-independent, multispecies trap sur- 
veys occurring in the southeastern U.S. Atlantic (SEUS) 
from North Carolina to Florida. The inclusion of soak 
time and fish accumulation separated mechanisms that 
cause gear saturation into 2 groups: those mechanisms 
related to fish accumulation (e.g., agonistic behaviors 
or bait depletion) and those mechanisms related to the 
length of time a trap soaks (e.g., local depletion of the 
target species or loss of bait freshness). We developed 
a delta-generalized-additive model (delta-GAM) that 
was able to accommodate both nonlinearities between 
the response and predictor variables and zero-inflation 
(i.e., a high proportion of zero catches; Martin et al., 
2005). This approach allowed us to determine whether 
trap saturation occured because of either soak time or 
fish accumulation, or both, and then we used the model 
to predict relative abundance after accounting for the 
influence of soak time and fish accumulation. 
Materials and methods 
Study area 
In our study reef fish species associated with hard 
substrates were sampled on the continental shelf and 
continental shelf-break in the SEUS. The continental 
shelf and shelf-break in the SEUS are dominated by 
sand and mud substrates, within which areas of hard, 
rocky substrates (“hard bottom”) occur and a highly 
diverse reef fish assemblage associates. Hard bottom 
habitats range in complexity from flat limestone pave- 
ment, sometimes covered with a sand or gravel veneer, 
to high-relief rocky ledges (Schobernd and Sedberry, 
2009; Glasgow, 2010). Hard bottom areas often host di- 
verse epifauna that can provide food and shelter for 
reef fishes. The major oceanographic feature of the 
SEUS is the Gulf Stream, which influences outer sec- 
tions of the continental shelf as it flows northward. 
Consistently warm Gulf Stream waters along the outer 
SEUS shelf allow tropical and subtropical species to 
inhabit areas at least as far north as North Carolina 
(Miller and Richards, 1980). For our study, sampling 
occurred on continental shelf and shelf break habitats 
from approximately Cape Lookout, North Carolina, to 
St. Lucie Inlet, Florida (Fig. 1). 
Sampling approach 
The Marine Resources Monitoring, Assessment, and 
Prediction (MARMAP) Program of the South Carolina 
Department of Natural Resources has used chevron 
fish traps to index reef fish abundance since the late 
1980s. Since 2009, MARMAP funding for reef fish sam- 
pling has been supplemented by the cooperative South- 
east Area Monitoring and Assessment Program — South 
Atlantic (SEAMAP-SA) administered by the National 
Marine Fisheries Service. We analyzed MARMAP data 
from 1990 through 2011, during which time sampling 
with chevron fish traps was conducted in a consistent 
manner (as described later in this section). We also in- 
cluded in our analyses 2010-11 data from the South- 
east Fishery-independent Survey (SEFIS), which the 
National Marine Fisheries Service created in 2010 to 
increase fishery-independent sampling in the SELTS, 
because sampling methods were identical. Hereafter, 
the 2 sampling programs are referred to as “MARMAP/ 
SEFIS.” 
Hard bottom sampling stations included in the anal- 
yses were selected for sampling in 1 of 3 ways. First, 
most sites were selected randomly from the MARMAP/ 
SEFIS sampling frame that consisted of approximately 
2000 sampling stations on hard bottom habitat. Second, 
some stations in the sampling frame were sampled op- 
