Content uploaded by C. Phillip Goodyear
Author content
All content in this area was uploaded by C. Phillip Goodyear on Aug 11, 2014
Content may be subject to copyright.
CSIRO PUBLISHING
www.publish.csiro.au/journals/mfr Marine and Freshwater Research, 2003, 54, 409–417
Spatio-temporal distribution of longline catch per unit effort,
sea surface temperature andAtlantic marlin
C. Phillip Goodyear
415 Ridgewood Road, Key Biscayne, FL 33149, USA. Email: philgoodyear@cox.net
Abstract. Atlantic blue and white marlin are currently overfished, primarily as a result of bycatch in pelagic long-
lines directed at other species. One possible management measure to reduce fishing mortality on these species
would be to restrict fishing effort in times and places with exceptionally high marlin catch per unit effort (CPUE).
The International Commission for the Conservation of AtlanticTunas maintains a database of catch and catch-effort
statistics of participating nations. These data were analysed to determine whether the distribution of CPUE is suffi-
ciently heterogeneous in time and space that such measures might provide meaningful management alternatives.The
resulting distributions of catch rates were also contrasted with monthly average sea surface temperatures to examine
the possible association between temperature and CPUE. The results show spatio-temporal heterogeneity in catch
rates that may be partly explained by seasonal changes in sea surface temperatures. The time–area concentrations
of high CPUE differ between the species. This observed heterogeneity might be exploited to develop alternatives
for reducing fishing mortality for future management of the fisheries, but additional research is needed to refine
the spatial scale of the analysis and to more fully understand the factors contributing to the observed distribution.
Extra keywords: blue marlin, bycatch, catch effort, seasonal variations, white marlin.
Introduction
The most recent stock assessment for Atlantic blue and white
marlin (Makaira nigricans and Tetrapturus albidus respect-
ively) found them to be overfished, primarily as a result
of bycatch on pelagic longlines targeting tunas and sword-
fish (Anonymous 2000). Bycatch fishing mortality poses a
special problem because management institutions are gener-
ally more focused on the target species, and tend to ignore
bycatch species even to the point of near extinction (Casey
and Myers 1998). This problem is especially severe when
the population of the bycatch species is more susceptible to
overfishing than the population of one or more of the tar-
get species, which is the likely situation for Atlantic marlins
(Goodyear 2001). Because of the current condition of these
stocks, the International Commission for the Conservation of
Atlantic Tunas (ICCAT) Scientific Committee on Research
and Statistics recommended that management take steps to
reduce the catch of both species as much as possible. These
recommendations, among other possible actions, included the
establishment of time–area closures to restrict fishing effort
in times and places with exceptionally high marlin catch rates.
This notion is based on the premise that species preferences
for physical attributes of their habitat may be exploited to
reduce catchability in the population as a whole.
These features might include factors such as water depth
that are strictly a function of geography, or be a function
of variables such as water current or temperature that may
vary seasonally. For example, de Sylva (1990) argued that
temperature is one of the chief determinants of the distribu-
tion in abundance of blue marlin, and Hinton and Nakano
(1996) model the depth distribution of Pacific blue marlin
as a function of the difference between temperature at depth
and the sea surface temperature (SST). This paper presents a
preliminary analysis of ICCAT catch-effort (task ii) data to
determine if the monthly average catch per unit effort (CPUE)
distributions are sufficiently heterogeneous that time–area
closures might provide meaningful management alternatives,
and examines their relationship to monthly average SST.
Methods
The catch-effort records for blue and white marlin were extracted from
the ICCAT catch-effort longline database (ICCAT task ii data through
revision updr40 of the database). Each record provided latitude and lon-
gitude, country of origin for the vessel, year, month, number of hooks
deployedand the number of each species caught. Inspection of the result-
ing data indicated a variety of metrics used to quantify the effort and
the record set was further limited to those records where the effort was
measured in numbers of hooks. The CPUE was then calculated as the
ratio of the number of fish caught to the number of hooks deployed for
each species.
The population of both blue and white marlin in the Atlantic is
known to have declined during the time interval represented by the
data. In addition, the fleets of different nations target different species,
and there are temporal trends in species targeted within various nations.
© CSIRO 2003 10.1071/MF01255 1323-1650/03/040409
410 Marine and Freshwater Research C. P. Goodyear
Catch-effort data for 1982–1997 were selected for analysis. A general
linear model (GLM) was employed with year and country as fixed fac-
tors to standardize the CPUE series by removing the year, country, and
the year ×country effects from the data series. This was achieved by
substituting the ‘standardized’ GLM residuals for the raw catch per
hook values. The resulting data were placed into bins of months and
latitude–longitude geographic cells. Monthly means were derived for
each latitude–longitude cell. These values were sorted to obtain the
cumulative frequency distribution of the mean standardized CPUE for
each species by latitude, longitude and month. The data for each time–
area cell was then plotted with a pattern identifying its location within
the cumulative frequency distribution.
The annual monthly average SST by 1◦of latitude and longitude
was obtained for 1982–1997 from the International Research Institute
for Climate Prediction at Columbia University (Reynolds and Smith
1994). These data were averaged by 1◦of latitude and longitude for
each month over the years available, and the resulting data were plotted
using a colour code to represent temperature. Also, the standardized
CPUE by year, month and 5◦latitude–longitude cell were paired with
the corresponding SST to evaluate the relationship between CPUE and
temperature for those time–area strata with positive catches of each
species.
Results
The ICCAT longline data set provided 11 430 catch per hook
records in time-area strata with positive catches of blue mar-
lin, and 9143 catch per hook records in time-area strata with
Table 1. Results of general linear model analysis of the International Commission
for the Conservation ofAtlantic Tunas (task ii)Atlantic blue marlin longline catch
per unit effort (catch per hook) with country and year as main effects
Tests of between-subjects effects
Source Type III sums df Mean square F Significance
of squares
Model 3.253E – 03a82 3.967E – 05 34.630 0.000
Country 9.718E – 05 11 8.835E – 06 7.712 0.000
Year 9.417E – 04 15 6.278E – 05 54.801 0.000
Country ×year 1.978E – 03 55 3.596E – 05 31.386 0.000
Error 1.300E – 02 11 348 1.146E – 06
Total 1.625E – 02 11 430
aR2=0.200 (adjusted R2=0.194).
Table 2. Results of general linear model analysis of the International Commis-
sion for the Conservation ofAtlantic Tunas (task ii) Atlantic white marlin longline
catch per unit effort (catch per hook) with country and year as main effects
Tests of between-subjects effects
Source Type III sums df Mean square F Significance
of squares
Model 7.537E – 05a82 9.191E – 07 18.832 0.000
Country 8.558E – 06 11 7.780E – 07 15.941 0.000
Year 1.697E – 06 15 1.131E – 07 2.318 0.003
Country ×year 5.920E – 06 55 1.076E – 07 2.205 0.000
Error 4.422E – 04 9061 4.881E – 08
Total 5.176E – 04 9143
aR2=0.146 (adjusted R2=0.138).
positive catches of white marlin, for the time interval 1982–
1997 (the most recent compilation available at the time of the
analysis).As expected, the GLM results showed year, country
and year ×country effects to be significant for both species
(Tables 1 and 2) and the CPUE series was standardized to
remove these effects. Because the geographic resolution for
most of the data was to 5◦latitude and longitude, this scale
was selected for analysis. The averages of the standardized
CPUE in each latitude–longitude-month cell were sorted by
value, and the magnitudes corresponding to the 95th, 80th,
and 50th percentiles of the distribution were determined. The
resulting information was then plotted for each month with
the relative magnitude of the catch rate in each time-area cell
indicated by its associated pattern along with the mean SST
(Figs 1 and 2).
Areas on Figs 1 and 2 that are not enclosed by a 5◦box
did not have positive catches for the particular species within
the ICCAT longline catch-effort data selected for analysis.
Fishing effort exists in many of these cells and some mar-
lins are probable caught as a result. However, it seems likely
that such areas would tend to be toward the lower end of the
actual cumulative frequency distribution of catch rates and
would not distort the time–area pattern of higher catch rates
observed in these figures.
Longline effort and Atlantic marlin Marine and Freshwater Research 411
Seasonal changes in the distribution of SST are evident
in Figs 1 and 2, as are seasonal shifts in the areas of highest
standardized CPUE. There is a northward expansion of the
time–area cells with high CPUE values during the summer
in the northern hemisphere and a southward movement in the
summer in the southern hemisphere for both species. Most
of the cells with the highest catch rates for blue marlin tend
to be in the areas where mean SST is about 26–27◦C. White
marlin seem to have a somewhat broader thermal preference,
with high catch rates in waters with average SST from about
22–29◦C. However, it is also clear that SST does not com-
pletely explain the catch rate distributions for either species.
Some areas with mean temperatures lower than what appears
to be optimum have catch rates in the upper 50th percentile of
the cumulative frequency distribution. The process of aver-
aging both CPUE and temperature may have obscured the
underlying relationship between temperature and abundance
shown in Figs 1 and 2. However, most time–area cells within
what appears to be the optimum temperature ranges for each
species and near the centre of the geographical distributions,
have catch rates in the lower 50th percentile of the cumulative
frequency distribution.
The frequency distributions of positive catches and mean
CPUE by mean monthly SST in the 5◦square at the time
of capture also indicate the presence of both species over a
broad range of temperatures (Figs 3 and 4). The frequency
distributions are influenced by different total effort in each
temperature stratum, and are unlikely to reflect the actual dis-
tribution of either species with respect to temperature. The
mean CPUE for blue marlin shows a bimodal distribution,
with one peak at 15◦C and another at 30◦C, the highest tem-
perature at which blue marlin were caught. Mean CPUE for
white marlin was similarly distributed, with a peak at 29–
30◦C and another at about 23◦C, but the effect was not as
pronounced as with blue marlin. Although CPUE for white
marlin was somewhat higher at intermediate temperatures, it
was fairly similar from 15◦Cto28
◦C.
Discussion
Hinton and Nakano (1996) model the depth distribution of
blue marlin as a function of water temperature at the surface.
The results here cannot test their hypothesis, but since CPUE
declines as SST falls below about 15◦C, they do support the
notion that temperature is an important determinant of the
distribution of both blue and white marlin in the Atlantic.
However, the effect does not appear to be nearly so strong as
hypothesized by de Sylva (1990). The relatively high CPUE
observed in areas with relatively low mean SST may be partly
explained by the fact that fishermen often selectivelyset along
oceanographic features such as current edges that have tem-
peratures higher than the average for the surrounding water.
Some of this variability might also reflect the expression of
sexually dimorphic and/or size–age differences in thermal
preferences. However, it seems clear that both species occur
over a relatively wide range of environmental temperatures.
It is also clear that other factors contribute to the distribution,
as reflected in the catch rate pattern. Local depletions in areas
of high effort may cause downward bias in CPUE where it
might otherwise be high. Also, there may be some residual
gear effects not completely removed by the standardization.
Nonetheless, it is likely that other factors, perhaps related to
prey abundance or current patterns, also play important roles
in marlin habitat. The distributions of standardized catch rates
derived from the ICCAT catch-effort data do not presuppose
which environmental factors contribute to the observed distri-
butions, but rather reflect the outcome of all factors involved.
As a result, the actual time–area distributions of high catch
rates provide the most reliable indicators of preferred habi-
tats, and best identify potential times and areas for closure or
other restrictions on fishing to reduce marlin bycatch.
Closed areas have been used in many parts of the world
to control bycatch mortality (Alverson et al. 1994). Marine
reserves (also called marine protected areas) have also been
promoted as management tools to enhance conservation
of fishery resources (Bohnsack 1994; Shackell and Wilson
1995; Hutchings 1995, 1996; Allison et al. 1998; Lauck et al.
1998). Hutchings (1996) noted that marine reserves might
have considerable merit in reducing bycatch when assessed
against the effectiveness of other forms of regulatory control
such as bycatch limitations and catch quotas. Often, how-
ever, the concept of marine reserves is restricted to closures
of particular habitats or portions of habitats. Kenchington
(1990) noted that such marine reserves would be of little
use for species such as billfish that have both pelagic larvae
and highly pelagic adults. Among other options for reduc-
ing bycatch, Alverson et al. (1994) observed that time–area
control of fishing activity offers an opportunity to reduce
unwanted bycatch.
The main objective of employing time-area strategies is to
take advantage of variation in the catchability of target and
bycatch species (Murawski 1992). The current results indi-
cate the presence of spatial heterogeneity in the distribution
of catch rates of marlins in the Atlantic that may provide
the opportunity for time–area closures to selectively protect
these species from overfishing. These findings are similar
to previous results that only considered data from the US
longline fishery (Goodyear 1999). However, as with the pre-
vious study, considerable additional research will be required
before it can be concluded that time–area restrictions on the
fishery would provide meaningful conservation measures.
Murawski (1992) pointed out that bycatch reduction plans
involving time-area manipulation of the fishery should be
economically viable and the proposed programme must be
effectively implemented and enforced. The effectiveness of
such controls obviously depends on the degree of overlap
between the bycatch and the targeted species (Adlerstein and
Trumble 1992). The identification of realistic management
412 Marine and Freshwater Research C. P. Goodyear
Fig. 1. Spatio-temporal distribution of Atlantic blue marlin longline catch per unit area and mean sea surface temperatures. Patterns indicate the
position of monthly 5◦latitude–longitude cells in the cumulative frequency distribution of standardized longline catch per unit area. The colour
scheme depicts monthly mean sea surface temperatures.
Longline effort and Atlantic marlin Marine and Freshwater Research 413
Fig. 1. (continued)
414 Marine and Freshwater Research C. P. Goodyear
Fig. 2. Spatio-temporal distribution of Atlantic white marlin longline catch per unit area and mean sea surface temperatures. Patterns indicate the
position of monthly 5◦latitude–longitude cells in the cumulative frequency distribution of standardized longline catch per unit area. The colour
scheme depicts monthly mean sea surface temperatures.
Longline effort and Atlantic marlin Marine and Freshwater Research 415
Fig. 2. (continued)
416 Marine and Freshwater Research C. P. Goodyear
2400
1800
1200
600
0
3.6
2.4
1.2
0.0
10 12 14 16 18 20 22 24 26 28 30
Sea surface temperature (ºC)
Standardized CPUE
Positive cells
Frequency
N/10 000 hooks
Fig. 3. Frequency of monthly 5◦cells with positive blue marlin catch
and mean standardized catch per unit effort (CPUE) by mean sea surface
temperature at the time of capture, 1983–1996.
2000
1500
1000
500
0
2.4
1.6
0.8
0.0
10 12 14 16 18 20 22 24 26 28 30
Sea surface temperature (ºC)
Standardized CPUE
Positive cells
Frequency
N/10 000 hooks
Fig. 4. Frequency of monthly 5◦cells with positive white marlin catch
and mean standardized catch per unit effort (CPUE) by mean sea surface
temperature at the time of capture, 1983–1996.
measures must specify reasonable, contiguous geographic
areas where pelagic longline closures would be both prac-
tical and of the greatest benefit to increased billfish survival,
hopefully with minimum effect on other fisheries.
Further evaluations of possible time-area closures for the
protection of Atlantic marlins should consider the distribution
of the ratios of marlin to target species in order to identify
those cells where closures would have the least impact on
directed fisheries. The potential gains of increased marlin
survival and any concomitant reductions in catches of target
species must also be quantified, including the evaluation of
the impacts of the effort displaced by time–area closures. The
identification of viable management options will also require
additional research to refine the spatial scale of the analysis.
Because most of the existing data is pooled by 5◦latitude
and longitude, such an effort will require new analyses of the
raw data from each of the nations contributing to the ICCAT
database. It might also prove valuable to evaluate the potential
to simultaneously optimize time-area closures for all ICCAT
species that might benefit from reduced fishing mortality;
however, such efforts pose daunting challenges.
Acknowledgments
I thank K. Davy and R. Nelson for helpful comments on
the draft manuscript. The Billfish Foundation supported
this work.
References
Adlerstein, S., and Trumble, R. (1992). Management implications of
changes in bycatch rates of Pacific halibut and crab species caused
by diel behavior of groundfish in the Bering Sea. In ‘Proceed-
ings of the Symposium on Fish Behavior in Relation to Fishing
Operations, Bergen, Norway, 11–13 June 1992.’ pp. 211–15.
(ICES: Copenhagen.)
Allison, G. W., Lubchenco, J., and Carr, M. H. (1998). Marine reserves
are necessary but not sufficient for marine conservation. Ecological
Applications 8(Suppl.), S79–92.
Alverson, D. L., Freeburg, M. H., Murawski, S. A., and Pope, J. G.
(1994). ‘A Global Assessment of Fisheries Bycatch and Discards.’
UN Food and Agriculture Organization. FAO Fisheries Technical
Paper 339. (FAO: Rome.)
Anonymous (2000). Proceedings of the fourth billfish workshop.
International Commission for Conservation of Atlantic Tunas,
Collective Volume of Scientific Papers 53, 375 pp.
Bohnsack, J.A. (1994). How marine fishery reserves can improve fish-
eries. Proceedings of the Gulf and Caribbean Fisheries Institute 43,
217–41.
Casey, J. M., and Myers, R. A. (1998). Near extinction of a large, widely
distributed fish. Science 281, 690–2.
de Sylva, D. P. (1990). Distributional changes in billfishes (Istiophori-
dae) and sea surface temperatures – a possible early warning system
to monitor global greenhouse climate warming. In ‘Planning the
Future of Billfishes: Research and Management in the 90’s and
Beyond, Part 2: Contributed Papers.’(Ed. R. H. Stroud.) pp. 137–43
(National Coalition for Marine Conservation, Inc.: Leesburg, VA.)
Goodyear, C. P. (1999). An analysis of the possible utility of time–
area closures to minimize billfish bycatch by U.S. pelagic longlines.
Fisheries Bulletin 97, 243–55.
Longline effort and Atlantic marlin Marine and Freshwater Research 417
Goodyear, C. P. (2001). Atlantic blue marlin and yellowfin tuna: com-
parative population vulnerability to fishing mortality. American
Fisheries Society Symposium 25, 219–24.
Hinton, M. G., and Nakano, H. (1996). Standardizing catch and effort
statistics using physiological, ecological, or behavioral constraints
and environmental data, with an application to blue marlin (Makaira
nigricans) catch and effort data from the Japanese longline fish-
eries in the Pacific. Bulletin of the International Commission for
Conservation of Atlantic Tunas 21, 171–200.
Hutchings, J. A. (1995). Seasonal marine protected areas within
the context of spatial–temporal variation in the northern cod
fishery. In ‘Marine Protected Areas and Sustainable Fisheries’.
(Eds N. L. Shackell and J. H. M. Willison.) pp. 39–53. (Science
and Management of Protected Areas Association: Wolfville, Nova
Scotia.)
Hutchings, J. A. (1996). Spatial and temporal variation in the density
of northern cod and a review of hypotheses for the stock’s collapse.
Canadian Journal Fisheries andAquatic Science 53, 943–62.
http://www.publish.csiro.au/journals/mfr
Kenchington, R. A. (1990). ‘Managing Marine Environments.’ (Taylor
and Francis: NewYork.)
Lauck, T, Clark, C. W., Mangel, M and Munro, G. R. (1998). Implement-
ing the precautionary principle in fisheries management through
marine reserves. Ecological Applications 8(Suppl.), S72–8.
Murawski, S. A. (1992). The challenges of finding solutions in
multispecies fisheries. In ‘Proceedings of the National Industry
Bycatch Workshop, 4–6 February 1992, Newport, Oregon.’ (Eds
R. W. Schoning, R. W. Jacobson, D. L. Alverson, T. G. Gentle and Jan
Auyong.) pp. 35–45 (Natural Resources Consultants: Seattle, WA.)
Reynolds, R. W., and Smith, T. M. (1994). Improved global sea surface
temperature analyses. Journal of Climate 7, 929–48.
Shackell, N. L., and Willison, J. H. M. (Eds) (1995). ‘Marine Pro-
tected Areas and Sustainable Fisheries.’ (Science and Management
of Protected Areas Association: Wolfville, Nova Scotia.)
Manuscript received 5 September 2001; revised and accepted 21 March
2002.
A preview of this full-text is provided by CSIRO Publishing.
Content available from Marine and Freshwater Research
This content is subject to copyright.
