Determining the Distribution and Abundance of Shelf Rockfish:
A Cooperative Research Study in the Southern California Bight
Colin Jones1, John Harms2, Jim Benante1, Aaron Chappell3, Allan Hicks2, and John Wallace2
1
Pacific States Marine Fisheries Commission, Seattle, WA
2 Fishery Resource Analysis and Monitoring Division, Northwest Fisheries Science Center, NOAA, Seattle, WA
3 Fishery Resource Analysis and Monitoring Division, Northwest Fisheries Science Center, NOAA, Newport, OR
Introduction
We conduct an annual fishery-independent hook and line survey to monitor groundfish
within the Southern California Bight (SCB). The survey targets rocky, high-relief habitats
that are generally not well-sampled using other survey techniques such as bottom trawls
and acoustic backscatter.
The primary objective of the hook and line survey is to collect catch-per-unit-effort (CPUE)
data and biological information to generate annual abundance indices for several species
of shelf rockfish within the SCB, including bocaccio (Sebastes paucispinis), cowcod (Sebastes
levis), the vermilion rockfish complex (e.g., S. miniatus and S. crocotulus), and greenspotted
rockfish (S. chlorostictus). The survey was developed in 2003 and is a collaborative effort
among Pacific States Marine Fisheries Commission (PSMFC), Northwest Fisheries Science
Center (NWFSC), and southern California’s sportfishing industry.
Methods
• The survey employs a fixed site design with 121 sites sampled annually throughout the
SCB. The region is subdivided into 20 sampling areas to ensure sampling coverage
throughout the region (Figure 1).
• Sampling consists of three deckhands each making five coordinated drops of a five-hook
gangion at each site (Figure 3). Maximum possible catch of all species is 75 fish per site.
• Deckhands use stopwatches to measure the amount of time the gangion spends on the
seafloor available to demersal rockfish. Maximum fishing time is five minutes per drop.
• Biologists record species, length, weight, sex, age (via otolith), and genetic (via finclip)
information for each rockfish caught. A subset of this information is collected from nonrockfish species, and these fishes are discarded when possible.
A towable camera sled (Figure 7) was
developed to provide visual observations
of the seafloor at survey sites. In 2014, a
GoPro camera was added to sled to
provide a second visual observation, and
improve video quality. Data is used to
analyze habitat attributes, and compare
habitat with catch information.
• Implemented within the SCB in 2001 to
protect shelf habitat where cowcod are
most abundant, and prohibit most
bottom-fishing deeper than 20 fathoms.
• Two boundary areas were developed,
CCA West (4,200 sq mi) and CCA East
(100 sq mi) (Figure 1).
• Upon request from Pacific Fishery
Management Council (PFMC) and other
governing bodies, the hook and line
survey started sampling 42 fixed sites
within CCAs in 2014 (yellow dots,
Figure 1), and will add 40 more sites in
2015.
• These new sites will allow us to
determine the efficacy of CCAs for not
only cowcod, but other important
rockfish in the region (bocaccio,
vermilion, etc.).
• The depth range sampled is 37m – 229m.
• The survey gear is hook and line gear deployed by rod and reel. A sampling rig (or
gangion) is affixed to the end of each line (Figure 2).
Underwater Video
Cowcod Conservation Areas
(CCAs)
Figure 4. Deckhands and biologists displaying
cowcod caught during hook and line survey.
Bocaccio Length Frequency
Length frequency analysis of survey data suggest a bocaccio population characterized by a
few strong or moderately-strong year classes (Figure 5). The large 1999 year class that
dominated catch in 2004 has diminished in size over time due to natural and fishing
mortality as well as emigration to depths and latitudes not sampled during the survey.
The emergence and progression of the 2003 and 2005 year classes are also visible along
with the sharp spike of the 2010 year class. This information can provide analysts with an
early confirmation of year class strength before cohorts are fully recruited to the sport and
commercial fisheries. The ability to track specific cohorts over time suggests the survey
maintains generally constant selectivity over a broad range of sizes.
• Catch data are modeled within a Generalized Linear Model (GLM) framework to help
generate abundance indices for various species (see Results).
San Miguel Isl.
Santa Barbara
The figures to the right show
the median proportion of
hooks deployed at each site
which caught a bocaccio,
vermilion/sunset rockfish,
and greenspotted rockfish.
While greenspotted and the
vermilion complex show a
relatively flat trendline, the
bocaccio data track the
abundance indices
generated for the 2013 and
2015 stock assessments and
show the strength of the
2010 year class.
Santa Barbara Channel
Santa Rosa Isl.
Pt. Hueneme
Santa Cruz Isl.
Anacapa Isl.
Figure 8. Images from GoPro of the seafloor off
60 Mile Bank, CA, showing rockfish, brittle stars,
and complex rocky habitat.
Catch Rate by
Species
• A variety of environmental and sampling information is collected to help develop
probabilistic models used to predict species presence/absence on survey hooks.
Point Conception
Figure 7. Custom-designed camera sled,
winch, and A-frame as mounted on the F/V
Aggressor.
Harrison Reef
Santa Monica Bay
Santa Rosa Flats
San Pedro Bay
14 Mile Bank
Central Coast
Santa Catalina Isl.
CCA West
South Coast
125 Fathom Bank
San Clemente Isl.
Biological Samples Collected
9 Mile Bank
CCA East
60 Mile Bank
Maturity
Figure 5. Bocaccio length frequency from 2004 to 2013.
Figure 1. Map of the SCB showing hook and line survey sites from 2014, and boundaries of 20
sampling areas and two CCAs. Red dots represent 121 fixed sites, and yellow dots indicate 42
new sites sampled within CCAs.
To generate species-specific abundance indices, catch data were modeled at the level of
the individual hook in a GLM framework using a binomial distribution: either a hook
caught the species of interest, or it did not. Based upon our experience with the survey in
the field and some relationships identified in earlier statistical analyses, we examined
several sampling and environmental parameters as potential explanatory variables within
the model. We used Akaike’s Information Criteria (AIC) to assist in selecting the final
model. This model selection process is re-visited for each species and with each new year
of data. Some of the covariates in the final model for the 2013 bocaccio index included
site, fishing time, vessel, hook position, angler, and wave height. After backtransformation, the index is presented as the probability of any random survey hook
capturing the species of interest (bocaccio) (Figure 6). Per our 2012 CIE (Center for
Independent Experts) review and 2013 SSC (Scientific and Statistical Committee) review,
we are reviewing our methods of analysis. Future work will be further exploring catch
rate information to a number of covariates that might provide us with a more informative
means of determining an index of abundance then we have historically been providing.
Mainline: 80 lb. spectra
16’’
between
tippets
Barrel Swivel: Size 1/0
Gangion: 60 lb. monofilament
Shrimp Flies:
•Red and yellow bucktail bristles w/ red
collar
•VMC long shank, kirby style “J” hook,
size 5/0
•Bait: 1 frozen squid strip per hook
6” tippet
Tippet/Leader:
•6” long
24’’
between
swivel &
sinker
•Tie a 6” dropper loop; cut loop at
midpoint to form 2 strands, & then
remove one of the strands
•Tie the shrimp fly directly to tippet
0.25
Figure 3. Deckhands and biologists engaged in typical
sampling operations aboard the F/V Aggressor.
0.20
Barrel Swivel: Size 1/0
Index
16’’
between
bottom
tippet &
barrel
swivel
Catch Modeling
Sinker Breakaway: 30 lb. monofilament; can be
changed if greater strength is needed
Sinker: square shape with rounded bottom;
1,2,3,4, & 5 lb. sinkers can be used
(a)
0.05
Seabird 19plus CTD and R.M. Young
anemometers record oceanographic and
weather conditions during each sampling
event.
0.00
00
2004
nts
Figure 2. Diagram of the sampling
gangion used during the hook and line
survey.
Example of mature vermilion rockfish ovary at
macroscopic stage 2 (vitellogenic), and its
corresponding histological image (microscopic ova
stages 1-7 present)
Genetics
Survey personnel are currently working
with the NWFSC Genetics & Evolution
Program to develop separate indices of
abundance for vermilion and sunset
rockfish by analyzing the finclips
collected from each of the vermilion
rockfish complex specimens during
sampling (Figure 9). This genetic
information is used along with age and
biological data collected during the
survey for use in a biological profile of
the two species to support stock
assessments.
Vermilion rockfish
(Sebastes miniatus)
S
V
S
V
S
Sunset rockfish
(Sebastes crocotulus)
V
S
V
V
V
V
V
800 bp
600 bp
300 bp
S = Sunset rockfish, V = Vermilion rockfish
Figure 9. Gel electrophoresis analysis of a
Restriction Fragment Length Polymorphism (RFLP).
Acknowledgements
0.15
0.10
Hook and Line staff annually collect
ovaries from vermilion/sunset rockfish,
bocaccio, greenspotted rockfish, cowcod,
and canary rockfish. We assist scientists
from the SWFSC collecting ovaries from
chilipepper, yellowtail, swordspine,
squarespot, speckled, and rosy rockfish.
2005
2006
2007
2008
2009
2010
2011
2012
2013
Year
Figure 6. Index of relative abundance for bocaccio from 2004 – 2014.
2014
The hook and line survey would like to acknowledge the captains and crew of the F/V Aggressor
and Mirage for their hard work and dedication over the last 11 years, and the F/V Toronado for last
year’s work with us. We also recognize Ian Stewart, now with the IPHC, for his ingenuity in
developing the first GLM-based approach for modeling rod and reel data. We thank Anna Elz,
Jennifer Hempelmann, and the NWFSC Molecular Genetics Lab for their work analyzing the
thousands of tissue samples we have collected. Curt Whitmire consulted on GIS and mapping
issues. Victor Simon configured and coordinated the scales and oceanographic sensors used during
survey operations. We also thank Aimee Keller for her guidance and suggestions that have
significantly improved this project. Melissa Head analyzed all of our maturity samples. The PSMFC
Ageing Lab analyzed our otolith specimens.