ArticlePDF Available

Abstract and Figures

Despite growing concerns about overexploitation of sharks, lack of accurate, species-specific harvest data often hampers quantitative stock assessment. In such cases, trade studies can provide insights into exploitation unavailable from traditional monitoring. We applied Bayesian statistical methods to trade data in combination with genetic identification to estimate by species, the annual number of globally traded shark fins, the most commercially valuable product from a group of species often unrecorded in harvest statistics. Our results provide the first fishery-independent estimate of the scale of shark catches worldwide and indicate that shark biomass in the fin trade is three to four times higher than shark catch figures reported in the only global data base. Comparison of our estimates to approximated stock assessment reference points for one of the most commonly traded species, blue shark, suggests that current trade volumes in numbers of sharks are close to or possibly exceeding the maximum sustainable yield levels.
Content may be subject to copyright.
LETTER
Global estimates of shark catches using trade records
from commercial markets
Shelley C. Clarke,
1
* Murdoch K.
McAllister,
2
E. J. Milner-Gulland,
3
G. P. Kirkwood,
2
Catherine G. J.
Michielsens,
4
David J. Agnew,
2
Ellen K. Pikitch,
5
Hideki Nakano
6
and Mahmood S. Shivji
7
Abstract
Despite growing concerns about overexploitation of sharks, lack of accurate, species-
specific harvest data often hampers quantitative stock assessment. In such cases, trade
studies can provide insights into exploitation unavailable from traditional monitoring.
We applied Bayesian statistical methods to trade data in combination with genetic
identification to estimate by species, the annual number of globally traded shark fins, the
most commercially valuable product from a group of species often unrecorded in harvest
statistics. Our results provide the first fishery-independent estimate of the scale of shark
catches worldwide and indicate that shark biomass in the fin trade is three to four times
higher than shark catch figures reported in the only global data base. Comparison of our
estimates to approximated stock assessment reference points for one of the most
commonly traded species, blue shark, suggests that current trade volumes in numbers of
sharks are close to or possibly exceeding the maximum sustainable yield levels.
Keywords
Bayesian, extrapolation, fin, fishery, sampling, species, sustainable yield, wildlife.
Ecology Letters (2006) 9: 1115–1126
INTRODUCTION
Quantitative assessment of the status of exploited animal
populations usually requires accurate counts of individuals
harvested over time (Schaefer 1954; Lande et al. 2003).
However, when a substantial portion of the population off-
take is illegal, unregulated or unreported, actual harvested
numbers are highly uncertain, and intervention for conser-
vation purposes may be hampered by calls for collection of
better data. The Convention on International Trade in
Endangered Species of Wild Fauna and Flora (CITES) has
recognized that trade monitoring is critical in protecting
biodiversity, but most trade data remain decoupled from
population models used to estimate sustainable yields for
management. There is a growing body of research aimed at
determining the origin of animal products suspected to
derive from illegal harvests (Birstein et al. 1998; Roman &
Bowen 2000; Wasser et al. 2004; Shivji et al. 2005) but only a
few studies (Baker et al. 2000; Dalebout et al. 2002) have
obtained sufficient market access to allow conclusions
regarding population impacts.
Reported declines in shark populations (Baum et al. 2003;
Ward & Myers 2005) and a shark fin trade driven by rapid
economic growth in China (Clarke 2004a) make sharks an
especially relevant example of the need for trade studies.
Despite their vulnerability to overfishing (Walker 1998;
Castro et al. 1999), and the listing by CITES of some shark
species, i.e. basking (Cetorhinus maximus), whale (Rhincodon
1
Joint Institute for Marine and Atmospheric Research,
University of Hawaii and National Research Institute of
Far Seas Fisheries, 5-7-1 Shimizu-Orido, Shizuoka 424-8633,
Japan
2
Division of Biology, Faculty of Life Sciences, Imperial College
London, Prince Consort Road, South Kensington Campus,
London SW7 2AZ, UK
3
Division of Biology, Imperial College London, Silwood Park
Campus, Manor House, Buckhurst Road, Ascot, Berkshire SL5
7PY, UK
4
Finnish Game and Fisheries Research Institute, Viikinkaari 4,
PO Box 2, FIN-00791 Helsinki, Finland
5
Pew Institute for Ocean Science, University of Miami, 4600
Rickenbacker Causeway, Miami, FL 33149, USA
6
Fisheries Agency of Japan, 1-2-1 Kasumigaseki, Chiyoda-ku,
Tokyo 100-8907, Japan
7
Guy Harvey Research Institute, Nova Southeastern University
Oceanographic Center, 8000 North Ocean Drive, Dania Beach,
FL 33004, USA
*Correspondence: E-mail: scclarke@biznetvigator.com
Ecology Letters, (2006) 9: 1115–1126 doi: 10.1111/j.1461-0248.2006.00968.x
Ó2006 Blackwell Publishing Ltd/CNRS
typus) and great white (Carcharodon carcharias), sharks are
generally ignored or given a low priority in most fisheries
management organizations due to their characterization as
bycatch and low value per unit weight (Shotton 1999;
Fowler et al. 2005). The resulting lack of catch or landings
data for sharks severely inhibits the types of assessments
conducted for more data-rich fishery species. However,
shark fins are a highly valued commodity and are being
sourced globally through market channels concentrated in a
handful of Asian trading centres (Clarke 2004b). Notwith-
standing the secretive and wary nature of the shark fin trade,
monitoring these markets may currently be the best option
for determining shark exploitation levels and species
pressures worldwide.
This study extends our previous work to assess the
impact of the shark fin trade on shark species using data
collected in Hong Kong, the world’s largest shark fin
entrepo
ˆt. In an earlier study we estimated the total traded fin
weights in Hong Kong auctions for 11 common Chinese
trade name categories (comprised of fins from one or more
species), each classified according to four different fin
positions: dorsal (first only), pectoral, caudal (lower lobe
only), and other (Clarke et al. 2004). Our empirical data
included records from at least one auction held by each
trader during the period October 1999 to March 2001 and
overall comprised 29% of the total number of auctions. We
used Bayesian imputation methods (Little & Rubin 1987;
Rubin 1996) to probabilistically simulate missing data while
taking into account uncertainty in model formulation and
parameter values. Once the total weight of auctioned fins
over the 18-month period was estimated, we found that fins
in the 44 studied product type-fin position categories jointly
represented c. 46% of the total (Clarke et al. 2004).
In a subsequent study (Clarke et al. 2006), we produced
taxon-specific fin weight estimates, and described species
composition, in the Hong Kong auctions. This was
accomplished by combining market sampling and genetic
methods (Shivji et al. 2002) to derive concordances between
Chinese trade names and taxa. These species-specific
adjustments reduced traded weight quantities in each
category by 0–35%, but typically resulted in < 1% change
in the percentage distribution of fin weights among species-
specific categories for the entire market.
Building upon these previous studies, we now convert
species-specific fin weights in the Hong Kong market to
worldwide estimates of shark catch in numbers and biomass.
Our Bayesian modelling framework accounts for joint
uncertainties arising from each step and the resulting
probability intervals (PIs) for estimated quantities thus
summarize the total uncertainty in the estimates, providing
an objective means of judging the overall reliability of the
estimation. Our results represent the first fishery-independ-
ent estimate of the global shark catch for the shark fin trade,
and allow us to contrast this estimate with existing shark
catch records and assess the potential impacts on one
commonly traded shark species, the blue shark. In addition
to its application to sharks, the combined trade-based
quantitative and genetic forensic approach used here to
assess global exploitation levels provides a framework
applicable to other heavily traded, but largely unregulated
animal products (Vincent 1996; Thorbjarnarson 1999;
Milner-Gulland & Bennett 2003).
METHODS
Estimation of the number of sharks represented
To relate estimates of auctioned fin weights (Clarke et al.
2004) to whole shark equivalents, a series of conversion
factors was applied in an integrated Bayesian estimation
framework using WinBUGS software (Anon. 2004). To
produce estimates of the number of sharks present, we
estimated the mean weight of a single fin in each species-fin
position category (e.g. blue shark-dorsal fin) and used it as
the divisor for the total weight of fins in that category
(Clarke et al. 2004). The resulting estimate of the number of
fins present in each category was assumed to be equivalent
to the number of sharks for either dorsal or caudal fins, or
twice the number of sharks for pectoral fins.
Three data sets were required to estimate the mean weight
of single fins in a species-fin position category. First, data on
fin sizes provided on each auction sheet in qualitative
categories were translated into six arbitrary, numerical
classes (c). Auction records, consisting of numbers and
often weights of bags of fins by size class, were sorted by
species-fin position (s,p) category (11 species by three fin
positions) and the proportion of fins (by number of bags,
see Clarke et al. 2004) in each of the six size classes was
obtained. Second, each size class in each species-fin position
category was assigned a fin length midpoint and range based
on observations of each category-class at 17 Hong Kong
auctions over a 4-month period (n¼179). Informal
interview information from cooperative traders was also
used to assign the fin length midpoints for each class size.
The assigned midpoint lengths of dorsal, pectoral and caudal
fins for each species were cross-checked using ratios (each
fin’s length as a proportion of pre-caudal length) from
taxonomically accurate drawings and observed whole length
ranges (Compagno 1984) to ensure they were realistic.
Third, a cooperative trader allowed length–weight measure-
ments to be taken for 10–20 fin samples from each species-
fin position category (n¼397) thus providing the basis for
conversion from fin length to fin weight.
Prior probability distribution functions were assigned for
the average number of fins per bag and fin lengths. A log-
linear relationship between fin length and fin weight was
1116 S. C. Clarke et al. Letter
Ó2006 Blackwell Publishing Ltd/CNRS
used to obtain mean fin weight for each species-fin position
category (see Appendix A). In the final step, for each
species-fin position category the annual total auctioned
weight (Wtot
s;p) was divided by the mean fin weight
Ws;pto
produce an estimate of the number of fins and correspond-
ing number of sharks (N
s,p)
:
Ns;p¼Wtot
s;p
Ws;p
:ð1Þ
Results for dorsal, pectoral and caudal fins provided
independent estimates of the number of sharks represented.
Estimation of shark biomass represented
Estimates of shark biomass were developed using an
algorithm which converted the mean dry fin length estimate
used in the previous algorithm to a mean wet fin length,
then related the mean wet fin length to the mean whole
length of the shark and then to the shark’s mean whole
biomass. In the final step, mean whole shark biomass for
each species-fin position (
Bf
s;p) was multiplied by the number
of sharks estimated for that species-fin position combina-
tion (N
s,p
) to produce an estimate of the total biomass of
sharks supplying the fin trade (Btot
s;p). Although this algorithm
requires a greater number of steps, and thus incorporates a
greater potential for uncertainty, estimates in units of
biomass facilitate comparison with landed weights reported
in fisheries statistics. The model equations for the steps
outlined here are given in Appendix B.
In order to obtain shark biomass estimates, three
additional conversion factor data sets were required. First,
published data were sought on the relationship between dry
fin length and wet fin length as rehydration experiments
using Hong Kong fins would have been prohibitively
expensive. The only available data were based on 28 blacktip
shark (Carcharhinus limbatus) individuals (Fong 1999), but as
the purpose of this conversion was merely to account for
the change in fin length due to moisture loss during drying,
results were not expected to vary considerably across
species.
Second, data on the relationship between fin length and
body length were compiled at the port of Su’Ao in eastern
Taiwan. Data collection was constrained by the species
composition of landings, occurrences of partial processing
(e.g. heads removed and tails cut), and difficulties in
accessing large sharks for measurement. Seven of the 11
species groups were sampled (n¼79) for fork length and
length of all three fin positions. The remaining four species
were assigned the parameters of the most appropriate
sampled species, usually within the same genus, based on
similarities in fin morphology. Dusky (Carcharhinus obscurus)
and bull (Carcharhinus leucas) sharks were assigned the
sandbar shark (Carcharhinus plumbeus) parameters, and the
great hammerhead (Sphyrna mokarran) was assigned the
jointly estimated parameters from smooth and scalloped
hammerheads (Sphyrna zygaena and Sphyrna lewini). Tiger
sharks (Galeocerdo cuvier) are arguably most similar in fin
morphology to sandbar or silky (Carcharhinus falciformis)
sharks, but given the small size of silky shark individuals in
the sample, and the relatively large size of tiger sharks, the
sandbar shark parameters were believed to provide a better
representation.
Third, published length–weight conversion factors, based
on measurements of 5065 sharks from the western North
Atlantic, were applied using fork length as the measurement
standard (Kohler et al. 1995). All Sphyrna spp. conversions
applied the published factors for Sphyrna lewini and all Alopias
spp. conversions applied the published factors for Alopias
superciliosus. Species-specific conversion factors for bull shark
and oceanic whitetip (Carcharhinus longimanus), which were not
included in Kohler et al. (1995), were drawn from Bonfil et al.
(1990) and Lessa et al. (1999) respectively. As these latter
length–weight relationships were based on total length, to
convert from fork length, specific-specific conversion factors
from Branstetter & Stiles (1987) for bull shark, and from
Compagno (1984) for oceanic whitetip, were applied. Because
there was insufficient information in these references to
quantify error in the total length–fork length conversion
factors, we assumed that they were precisely estimated.
Extrapolation to global trade volumes
Estimates of the annual number and biomass of sharks
represented in the global shark fin trade (G) were obtained
as follows:
G¼Np1
k1
i1
w;ð2Þ
where N
p
is the sum of estimated fins over all species for a
given fin position (or for biomass substitute Btot
p, the bio-
mass of sharks over all species for a given fin position); kis
the proportion of auctioned fins included in the studied
species categories (i.e. accounts for ÔotherÕfins of unknown
type which could not be fully modelled due to morpholo-
gical uncertainties); iis the proportion of Hong Kong im-
ports that were auctioned; and wis the proportion of global
shark fin imports that passed through Hong Kong. When
computing species-specific global totals, we assumed that all
individuals of the given species were identified in the auc-
tion records as the market category corresponding to that
species, and thus set the value of kto 1.
Due to the lack of information on product types not
examined in this study, extrapolation by weight from studied
to unstudied product types (k) assumed that the fin sizes
and shapes represented in the 11 studied product types were
representative of the unstudied product types. Based on
Letter Shark catches from trade records 1117
Ó2006 Blackwell Publishing Ltd/CNRS
model-estimated proportions of unstudied fins (treated as a
single, undifferentiated class), the median and 95% PI of the
percentage of studied fins by weight were 45.9 and
43.7–48.2 respectively (Clarke et al. 2004). The percentage
of auctioned fins within the total quantity of Hong Kong
imports, i(median 17.4, 95% PI 16.3–18.6) was estimated
using the posterior distribution for the total weight of
auctioned fins per year (Clarke et al. 2004) and a uniform
distribution representing the total quantity of fin imports to
Hong Kong as recorded in government customs statistics
for 2000 adjusted for water content of frozen fins. The
range of this uniform distribution was specified by allowing
the water content of frozen fins to vary between 70% and
80% (Clarke 2004b). Parameters kand iwere formulated as
random variables (RVs) using appropriate distributions
based on these observed mean and variance values. The
final extrapolation from Hong Kong quantities to the global
trade, w, was based on ratios of the quantity of shark fins
traded through Hong Kong to the sum of quantities traded
through the major markets of Hong Kong, Mainland China,
Singapore, Taiwan and Japan over the years 1996–2000
(Table 1; sensu Clarke 2004b). This analysis showed that
Hong Kong’s share of global imports varied between 44%
and 59% with a mean of 52%, and thus values of wwere
drawn from a uniform distribution limited between 0.44 and
0.59. For deterministic calculations, the three extrapolation
factors were combined into a base scenario formed from the
midpoint of the range of each factor.
When combining results across fin positions to estimate
totals in numbers and biomass, a mixture distribution was
computed with the density function for each fin position
weighted proportional to its precision. The use of a mixture
distribution accurately accounts for uncertainty as it
presumes that the actual value for abundance or biomass
may come from any of the three fin positions with
credibility proportional to the precision obtained for each
fin position.
RESULTS AND DISCUSSION
Estimates of shark numbers and biomass
Model results were checked at two critical steps in the
conversion factor algorithms to verify the reliability of the
results for total numbers and biomass. As estimates of
the number of sharks represented were strongly influenced
by the estimates of mean single fin weights, estimated values
of
Ws;pwere checked against deterministic estimates and
empirical observations (Table 2). In seven of 33 instances,
the deterministic values fell outside of the 95% PIs; in six of
these cases, the deterministic values were higher than the
95th percentile. The largest outlier was observed for bull
shark dorsal fins where the average deterministic fin weight
value was 83 g higher than the 95th percentile. This was
primarily because 76% of all bull shark dorsal fins were
observed in the largest size class and unlike the Bayesian
estimation, the deterministic computation ignored sampling
error. All other outlying deterministic values were within
± 25 g beyond the PI.
Another critical step in the algorithm was the conver-
sion from fin length to whole length and weight. Estimates
of whole length and weight from the model were thus
assessed against maximum and minimum values observed
in nature to gauge the reliability of the algorithms
(Table 3). This comparison revealed that caudal fin
estimators usually produced the lowest 2.5th percentiles
of whole length (in 91% of 22 length and weight
predictions) and dorsal and pectoral fin estimators usually
produced the highest 97.5th percentiles (in 45% and 55%,
respectively, of 22 predictions). None of the model
predictions exceeded the observed length or weight in
nature. Comparisons between 2.5th percentiles and empir-
ical values are not particularly useful as modelled length
values below observed minimums may be accurate due to
observed trading of fins from unborn sharks (S.C. Clarke,
personal observation).
Table 1 Estimates of the traded quantity of shark fins (tonnes) based on national customs statistics and adjusted for observed under-
reporting of Mainland China, Singapore, Taiwan and Japan relative to Hong Kong
Country 1996 1997 1998 1999 2000
Share of Mainland China, Singapore, Taiwan and Japan corrected for
double-counting among these countries*
4690 4601 4399 4311 4637
Above share inflated by 1.3158 to adjust for 24% under-reporting* 6171 6054 5788 5672 6101
Hong Kong share corrected for double-counting* 4061 4414 4086 4489 5501
Estimate of total global trade 10 232 10 468 9874 10 161 11 602
Hong Kong imports without correction for double-counting* 4513 4868 5196 5824 6788
Hong Kong percentage of total (%) 44 47 53 57 59
As the purpose of the calculation is to determine the proportion of the total trade passing through Hong Kong, the total quantity traded
through Hong Kong (regardless of whether these fins were enumerated in other countries; fifth row) is used as the numerator, and an
accurate estimate of the total global trade (free of double-counting bias; fourth row) is used as the denominator.
*Data from Clarke (2004b).
1118 S. C. Clarke et al. Letter
Ó2006 Blackwell Publishing Ltd/CNRS
Estimates of the total number of sharks traded annually
worldwide, based on all fin positions combined, ranged
from 26 to 73 million year
)1
(95% PI), with an overall
median of 38 million year
)1
. Pectoral- and dorsal-based
estimates were relatively similar with median values of
29–38 million year
)1
(95% PIs of 25–36 and 30–47
Table 2 Posterior median and 95% probability intervals for weight of single fin (g) for each species-fin position combination,
Ws;p
Product type Predominant species or genus
Dorsal (n¼125; min ¼22;
max ¼305)
Pectoral (n¼127; min ¼22;
max ¼643)
Caudal (n¼148; min ¼14;
max ¼354)
Probabilistic Deterministic Probabilistic Deterministic Probabilistic Deterministic
Ya Jian Prionace glauca (blue) 41 (37–45) 43 138 (124–153) 114* 30 (24–36) 37*
Qing Lian Isurus oxyrinchus (shortfin mako) 159 (118–210) 149 127 (109–151) 140 170 (142–200) 167
Wu Yang Carcharhinus falciformis (silky) 190 (159–228) 202 139 (124–155) 135 54 (45–62) 59
Hai Hu Carcharhinus obscurus (dusky) 198 (153–232) 229 174 (153–199) 172 63 (45–75) 96*
Bai Qing Carcharhinus plumbeus (sandbar) 197 (173–224) 217 139 (115–165) 139 51 (40–61) 56
Ruan Sha Galeocerdo cuvier (tiger) 189 (113–253) 230 117 (93–153) 122 30 (13–50) 44
Chun Chi Sphyrna spp. (except Sphyrna
mokarran) (hammerheads)
79 (62–98) 88 93 (80–107) 125* 60 (52–67) 62
Gu Pian Sphyrna mokarran
(great hammerhead)
166 (103–239) 137 138 (111–166) 160 59 (36–76) 76
Wu Gu Alopias spp. (threshers) 117 (92–142) 134 194 (165–228) 190 22 (17–27) 19
Sha Qing Carcharhinus leucas (bull) 245 (198–281) 364* 220 (185–258) 196 66 (41–82) 85*
Liu Qiu Carcharhinus longimanus
(oceanic whitetip)
189 (163–213) 172 84 (75–94) 119* 48 (39–58) 52
Weights deterministically calculated using the fin length midpoints (l
Ls;pjc), the mean slopes (l
a1
s,p
) and intercepts (l
b1
s,p
), and the observed
proportions in each size class (b(c)
s,p
), are also shown. Deterministic values lying outside the predicted 95% probability intervals are marked
with an asterisk.
Table 3 Comparison between estimated whole shark lengths,
Ts;p(cm, fork length) and weights
Bs;p(kg), and observed maximum and
minimum shark lengths and weights
Product
type
Results from model
Predominant species or genus
Observed
Minimum
length
Maximum
length
Minimum
weight
Maximum
weight
Length
at birth
Maximum
length
Maximum
weight
Ya Jian 141 (C) 201 (P) 17 (C) 52 (P) Prionace glauca (blue) 30–38 334 206
Qing Lian 160 (P) 219 (D) 43 (P) 119 (D) Isurus oxyrinchus (shortfin mako) 54–63 370 506
Wu Yang 124 (C) 247 (D) 20 (C) 158 (D) Carcharhinus falciformis (silky) 45–70 274* 205*
Hai Hu 138 (C) 181 (P) 30 (C) 63 (P) Carcharhinus obscurus (dusky) 56–81 350 347
Bai Qing 132 (C) 169 (D) 26 (C) 57 (D) Carcharhinus plumbeus (sandbar) 48–64 207 118
Ruan Sha 100 (C) 175 (D) 9 (C) 51 (D) Galeocerdo cuvier (tiger) 31–78 468*
Chun Chi 127 (C) 174 (P) 23 (C) 57 (P) Sphyrna spp. (except S. mokarran)
(hammerheads)
36–48 420*
Gu Pian 111 (C) 224 (P) 16 (C) 128 (P) Sphyrna mokarran
(great hammerhead)
45–53 458 450
Wu Gu 100 (C) 215 (D) 13 (C) 133 (D) Alopias spp. (threshers) 50–89 423 364
Sha Qing 107 (C) 192 (D) 17 (C) 86 (D) Carcharhinus leucas (bull) 49 287 317
Liu Qiu 128 (C) 173 (P) 25 (C) 56 (P) Carcharhinus longimanus
(oceanic whitetip)
49–53 326 167
Model results columns list the 2.5th percentile (minimum) or 97.5th percentile (maximum) values of the lowest or highest estimate among the
three fin positions for the given product type; figures in parentheses indicate whether the estimate is for dorsal (D), pectoral (P) or caudal (C)
fins. Most observed values were taken from Froese & Pauly (2006) and converted from total length using factors given in that data base for
each species. Observed values marked with an asterisk were taken from Compagno et al. (2005) and converted to weight using factors from
Kohler et al. (1995), except in cases where the maximum observed length greatly exceeded the range of lengths observed by Kohler et al.
(1995) in which case no conversion was performed. No minimum weights or weight at birth data were available.
Letter Shark catches from trade records 1119
Ó2006 Blackwell Publishing Ltd/CNRS
million year
)1
, for pectoral and dorsal respectively) but
caudal estimates indicated a considerably higher median of
62 million year
)1
(95% PI of 50–79 million year
)1
). Appli-
cation of point estimate (e.g. mean) conversion factors from
the same data sets produced estimates of the numbers of
sharks represented in the fin trade of 30–52 million year
)1
over the three fin positions for the base extrapolation
scenario. The shark biomass represented by the global fin
trade, based on all fin positions combined, is estimated to lie
between 1.21 and 2.29 million tonnes year
)1
(95% PI) with
a median of 1.70 million tonnes year
)1
. Estimates by
individual fin positions produced a similar PI with medians
ranging from 1.37 to 1.91 million tonnes year
)1
and 95%
PIs ranging from 1.13 to 2.38 million tonnes year
)1
.
Calculations based on point estimates from the conversion
factor data bases resulted in estimates of 1.39–1.73 million
tonnes year
)1
. Species-specific estimates by number and
biomass are shown in Fig. 1.
The deterministic estimates for the total number and
biomass of sharks do not show marked differences from the
Bayesian posterior medians but, as expected, results by
individual species showed some larger deviations. This is
because the additional information included in the Bayesian
analysis provides the largest relative gains in accuracy at the
species level, whereas the total number and biomass
estimates show the averaging effect of combining results
across species. Where differences do occur, they are likely to
arise from the sparse data available for some conversion
steps. Deterministic methods rely heavily on the sample
mean to represent the best parameter estimate, but this can
be problematic when the sample size is low or the sample is
otherwise unrepresentative, and offer no reliable approach
to assessing the uncertainty in the estimates. In contrast, the
Bayesian methodology incorporates the same data but also
sampling theory and probabilistic auxiliary information,
e.g. on the mean and variance in the number of fins per bag,
to estimate the mean fin length traded by species and fin
position. The Bayesian method also utilizes a hierarchical
model structure (Gelman et al. 1995) to combine data from a
set of sampled shark species to estimate the variance in key
parameters (i.e. fin to body length conversion factors) across
the species, and provides reliable estimates of uncertainty in
the estimated quantities, as well as considerably improved
parameter estimates.
Comparison with other estimates of global shark catches
Our trade-derived figures provide a basis for evaluating the
quality of chondrichthyan (sharks, skates, rays and chimaeras)
capture production data compiled by the Food and Agricul-
ture Organization (FAO; Anon. 2005a), currently the only
data base attempting to encapsulate global catches. The data
base indicates that in 2000 the capture production for
chondrichthyans totaled 869 544 tonnes. However, of this
amount 386 547 tonnes is reported in the undifferentiated
Ôsharks, rays, skates, etc. not elsewhere indicatedÕcategory,
and thus may contain rays, skates and chimaeras that do not
contribute to the shark fin trade. Of the FAO data that are
differentiated, 218 080 tonnes (45%) are types of chon-
drichthyans used or potentially used in the shark fin trade, i.e.
shark species or guitarfish or sawfish (Rose 1996). This figure
(0.22 million tonnes) may be assumed to represent a low-end
estimate. Applying the percentage (45%) to the undifferen-
tiated capture production suggests that 174 531 tonnes of the
undifferentiated capture production is used in the shark fin
trade. Therefore, a reasonable, mid-range estimate of the total
FAO capture production supporting the shark fin trade is c.
0.39 million tonnes. If we assume the shark fin trade utilizes
all undifferentiated capture production, the estimate is c.
0.60 million tonnes (Fig. 2a).
Our median biomass estimate for the global shark fin
trade based on all fin positions combined (1.70 million
tonnes year
)1
) is more than four times higher than the mid-
range FAO-based figure (0.39 million tonnes), and nearly
three times higher than the high FAO estimate (0.60 million
tonnes year
)1
). Independent estimates for the three fin
positions also indicate median values three to five times
higher than the FAO-based figures (Fig. 2b). Differences
between our estimates and the FAO figures may be
attributable to factors suppressing FAO landings data such
as unrecorded shark landings, shark biomass recorded in
non-chondrichthyan-specific categories, and/or a high
frequency of shark finning and carcass disposal at sea.
Shark finning is prohibited by national bans in several
countries including the USA, the European Union, South
Africa, Brazil and Costa Rica (Fowler et al. 2005), and
regulated through administrative measures in other coun-
tries including Australia and Canada. The practice of finning
is also contrary to recommendations or resolutions agreed
by the International Commission for the Conservation of
Atlantic Tunas, the Inter-American Tropical Tuna Com-
mission, the Indian Ocean Tuna Commission and the
Northwest Atlantic Fisheries Organization. Despite several
successful prosecutions for violations in the USA (Anon.
2005b), global enforcement of finning restrictions remains
minimal and finning undoubtedly continues.
In addition, our trade-based biomass calculations may
underestimate global shark catches. For example, due to the
lack of data on domestic production and consumption of
shark fins by major Asian fishing entities such as in Taiwan
and Japan, unless exported for processing and then re-impor-
ted, these fins are not accounted for within our methodology
(Clarke 2004b). Furthermore, shark mortality which does not
produce shark fins for market, e.g. fishing mortality where the
entire carcass is discarded, is also not included. These dis-
crepancies suggest that world shark catches are considerably
1120 S. C. Clarke et al. Letter
Ó2006 Blackwell Publishing Ltd/CNRS
)
se
n
no
td
na
s
u
oht
(
s
sa
mo
iB
C
P
D
A
(b)
(a)
00
70
0
6005004003002
Blue
00109080
706050
4
d
aehremmaH
0
8
070605040
3
C
P
D
A
y
k
liS
C
P
D
A
0
6
050
40
30
2
01
0
pitetihwcinaecO
00
1
08
0
604
0
2
rehserhT
070
60
50
4
030
2
01
C
P
D
A
rabdnaS
C
P
D
A
0
6
0504030201
okamniftrohS
0504030201
daehremmahtaerG
040302010
C
P
D
A
lluB
C
P
D
A
0302010
yksuD
543210
C
P
D
A
regiT
)snoillim(s
k
rahsforebmuN
C
P
D
A
02510150
eulB
3
210
daehremmaH
3210
C
P
D
A
ykliS
C
P
D
A
5.115.00
p
i
tetihwcinaec
O
420
rehserh
T
25.115.00
C
P
D
A
rabdnaS
C
P
D
A
5.
11
5
.
00
okamnif
t
ro
hS
5.11
5
.00
da
eh
remm
a
h
ta
er
G
2
5
.1
1
5.00
C
P
D
A
l
luB
C
P
D
A
15.
0
0
yksuD
3
.
0
2
.
01.
00
C
P
D
A
regiT
Figure 1 Estimates by species of the num-
ber and biomass of sharks utilized per year
in the shark fin trade worldwide. Medians
(circles) and 95% probability intervals (lines)
are shown. Fin positions are abbreviated as:
dorsal (D), pectoral (P), caudal (C) and all fin
positions from a mixture distribution (A).
Letter Shark catches from trade records 1121
Ó2006 Blackwell Publishing Ltd/CNRS
higher than reported, and thus shark stocks are facing much
heavier fishing pressures than previously indicated.
Comparison with species-specific reference points
Despite concerns regarding the potential overexploitation of
sharks (Anon. 1999), only a small number of studies have
produced estimates of maximum sustainable yield (MSY) or
other reference points for sharks by species. These studies
include stock assessments of blue sharks in the North
Pacific (Kleiber et al. 2001), blue sharks and shortfin mako
sharks in the North and South Atlantic (Anon. 2005c), and
large coastal shark populations in the western Atlantic and
Gulf Mexico with species-specific estimates for sandbar and
blacktip sharks (McAllister et al. 2001; Corte´s et al. 2002). As
each study was conducted for an individual ocean basin,
their reference points require extrapolation to a global value
for comparison with global fin trade-based estimates.
We chose a wide ranging and continuously distributed
species, the blue shark Prionace glauca (Compagno 1984;
Nakano & Seki 2003), to illustrate our method. We made
the simplifying assumption that although reproductive rates,
natural mortality and fishing selectivity may vary among blue
shark populations, MSY per unit area can be represented by
a single value over the entire range of this speciesÕhabitat.
Blue shark habitat was defined as the area between 50°N
and 50°S latitude worldwide and extending to the coastline
in each ocean basin (Compagno 1984; Nakano & Seki 2003).
Using an equal area projection in a geographical information
system, the North Pacific blue shark habitat was calculated
at 75.35 million km
2
, the Atlantic habitat at 72.16 mil-
lion km
2
, and the area of global habitat at 287.84 mil-
lion km
2
. The ratios used to extrapolate the regional MSY
values were therefore set at 1 : 3.82 for the North
Pacific : global extrapolation and 1 : 3.99 for the Atlan-
tic : global extrapolation.
Based on the extrapolated North Pacific MSY estimate in
number of sharks (Kleiber et al. 2001), the blue shark global
MSY is 7.26–12.60 million sharks year
)1
. The median global
trade-based estimate for the number of blue sharks utilized
each year based on all fin positions combined (10.74 mil-
lion year
)1
) is very similar to the MSY estimate, but the 95%
PI (4.64–15.76 million year
)1
) spans a broader range than
the MSY estimate. Based on summing the minimum and
maximum MSY biomass estimates for the North and
South Atlantic basins (Anon. 2005c), the global MSY is
0.73–1.09 million tonnes year
)1
, which exceeds the trade-
based median catch biomass estimate for all fin positions
(0.36 million tonnes year
)1
)byc. 200–300% and also lies
completely above the trade-based 95% PI of 0.20–0.62
million tonnes year
)1
.
Acknowledging the margins of error, and the likely
downward bias of trade-based estimates, our evaluation,
using a Pacific numbers-based reference, suggests that blue
sharks globally are being harvested at levels close to or
possibly exceeding MSY. In contrast, our comparison with
an Atlantic, biomass-based MSY reference point suggests
catch levels may be less problematic. Given that we have no
population estimate, we are not able to evaluate the actual
sustainability of our estimated catch levels. The MSY
reference point is the highest possible catch that could
theoretically be sustainable, and thus any catch that
approaches or exceeds this level is of concern.
CONCLUSIONS
As a result of the global nature of our assessment we cannot
evaluate the exploitation status of individual populations.
Furthermore, the blue shark is one of the most prolific and
resilient of shark species (Smith et al. 1998; Corte´s 2002) and
thus our blue shark results cannot be used to make
: d e t a i t n e r e f f i D s k r a h s
: d e t a i t n e
r e f f i
D s y a r , s e t a k s s a r e a m i h c d n a
:
d
e t a
i
t
n
e r e f f i d n
U s k
r
a h s t
o
n d
e
m u s s a
:d e t a i t n e r
e
f f i d n U s k r a h s d e m u s s a
0 8 0 8
1 2
1
3 5 4 7 1
7 1 9 4 6 2
6 1 0 2 1 2
) s e n n
o t d n
a
s u o
h t ( s s a m o i B
O A
F
l a d u a C : d e s a b e d a r T
l
a r
o
t
c
e
P :
d
e s a b
e d a r T
l a s r
o
D : d e s a b e d a r T
s n i f l l A : d e s a
b
e
d a r T
0 0 5 2 0 0 0 2 0 0 5 1 0 0 0 1
0 0 5
0
(a)
(b)
Figure 2 (a) FAO Capture Production for 2000 (Anon. 2005a)
showing quantity (tonnes) of chondrichthyans reported in undif-
ferentiated and differentiated categories; (b) adjusted FAO Capture
Production data compared with estimates for shark biomass
represented in the global shark fin trade by fin position [median
(circle) and 95% probability interval (line) for fin estimates or range
for FAO data].
1122 S. C. Clarke et al. Letter
Ó2006 Blackwell Publishing Ltd/CNRS
inferences about other shark species. Conclusions regarding
the sustainable or unsustainable use of other species, and
thus the shark fin trade as a whole, will require more detailed
species-based stock assessment reference points. However,
given the lower productivity of the other species common in
the fin trade (Smith et al. 1998; Corte´s 2002), the large
difference between trade-derived estimates of exploitation
and the catch estimates reported to the FAO adds to
growing concerns about the overexploitation of sharks.
Direct measures of catches will continue to be desirable
for managing harvests of fish and wildlife populations.
However, there are many situations for which the quality
and quantity of harvest data are so poor that trade data
provide equal or better opportunities for understanding
whether species are threatened by exploitation. This is
particularly the case for fisheries catching sharks, in which
shark monitoring and management systems are only
beginning to be implemented, and for which handling
practices, such as finning, distort landings data. This study
provides a quantitative characterization of the shark fin
trade and a probabilistic linkage between trade volumes and
sustainability reference points. To improve the accuracy of
the algorithm, future research should focus on improving
the estimates for those quantities that had the greatest
uncertainty in this analysis due primarily to low sample sizes
and possibly non-representative sampling (e.g. the length of
fins in each auction size class and fin length to body length
conversion factors).
On a broader level, this study illustrates a rigorous
Bayesian statistical methodology, applicable across species
and markets to high-value animal parts including ivory, and
species traded for traditional medicine or as luxury foods.
The methodology provides considerably more reliable
estimates than simple deterministic estimates obtained from
sample means and regression estimates of conversion factors.
The Bayesian methodology is recommended for future trade
applications as it offers more precise and reliable estimates by
combining data with auxiliary information (e.g. auction
structure), relating similar data sets (e.g. when sampling is
constrained), and serving as a framework for incorporating
new information as monitoring programmes develop.
ACKNOWLEDGEMENTS
This study was supported by Imperial College London, the
David and Lucile Packard Foundation, the Eppley Founda-
tion, the Wildlife Conservation Society, and the Florida Sea
Grant Program. The authors gratefully acknowledge the
contributions of Shou-Jen Joung and his laboratory at the
National Taiwan Ocean University, Clare Marshall of
Imperial College London, Samu Ma
¨ntyniemi of the Finnish
Game and Fisheries Research Institute, Tsutomu Nishida of
the Japan National Research Institute of Far Seas Fisheries
and Iago Mosqueira of AZTI Tecnalia to this study. R.A.
Myers, E. Corte´s and an anonymous referee provided
constructive comments which considerably improved the
manuscript.
REFERENCES
Anon. (1999). International Plan of Action for the Conservation and
Management of Sharks. Document FI:CSS/98/3. FAO, Rome.
Anon. (2004). WinBUGS Software (v. 1.4.1). Available at http://www.
mrc-bsu.cam.ac.uk/bugs/winbugs/contents.html. ( Last accessed
29 August 2006).
Anon. (2005a). FISHSTAT Plus (v. 2.30), Capture Production Data-
base 1970–2003. Available at: http://www.fao.org/fi/statist/
FISOFT/FISHPLUS.asp. (Last accessed 29 August 2006).
Anon. (2005b). 2004 Report to Congress Pursuant to the Shark Finning
Prohibition Act of 2000, United States National Marine Fisheries Ser-
vice. Available at http://www.nmfs.noaa.gov/by_catch/Shark%
20Finning%20Report.pdf. (Last accessed 29 August 2006).
Anon. (2005c). Report of the 2004 inter-sessional meeting of the
ICCAT subcommittee on by-catches: shark stock assessment.
Col. Vol. Sci. Pap. ICCAT, 58, 799–890.
Baker, C.S., Lento, G.M., Cipriano, F. & Palumbi, S.R. (2000).
Predicted decline of protected whales based on molecular
genetic monitoring of Japanese and Korean markets. Proc. R. Soc.
London, Ser. B, 267, 1191–1199.
Baum, J.K., Myers, R.A., Kehler, D.G., Worm, B., Harley, S.J. &
Doherty, P.A. (2003). Collapse and conservation of shark
populations in the northwest Atlantic. Science, 299, 389–392.
Birstein, V.J., Doukakis, P., Sorkin, B. & DeSalle, R. (1998).
Population aggregation analysis of three caviar-producing spe-
cies of sturgeons and implications for the species identification
of black caviar. Conserv. Biol., 12, 766–775.
Bonfil, R.S., David de Anda, F. & Mena, A.R. (1990). Shark fish-
eries in Mexico: the case of the Yucatan as an example. NOAA
Tech. Rep. NMFS, 90, 427–441.
Branstetter, S. & Stiles, R. (1987). Age and growth estimates of the
bull shark, Carcharhinus leucas, from the northern Gulf of Mexico.
Environ. Biol. Fishes, 20, 169–181.
Castro, J.I., Woodley, C.M. & Brudek, R.L. (1999). A Preliminary
Evaluation of Status of Shark Species. FAO Fisheries Technical Paper
380. Food and Agriculture Organization, Rome.
Clarke, S. (2004a). Shark Products Trade in Hong Kong and Mainland
China and Implementation of the CITES Shark Listings. TRAFFIC
East Asia, Hong Kong.
Clarke, S. (2004b). Understanding pressures on fishery resources
through trade statistics: a pilot study of four products in the
Chinese dried seafood market. Fish and Fisheries, 5, 53–74.
Clarke, S., McAllister, M.K. & Michielsens, C.G.J. (2004). Esti-
mates of shark species composition and numbers associated
with the shark fin trade based on Hong Kong auction data.
J. Northw. Atl. Fish., 35, 453–465.
Clarke, S., Magnussen, J.E., Abercrombie, D.L., McAllister, M. &
Shivji, M. (2006). Identification of shark species composition
and proportion in the Hong Kong shark fin market based on
molecular genetics and trade records. Conserv. Biol., 20, 201–211.
Compagno, L.J.V. (1984). Sharks of the World, Parts I (Hexanchiformes
to Lamniformes) and II (Carcharhiniformes). FAO Species Catalogue,
Vol. 4. Food and Agriculture Organization, Rome.
Letter Shark catches from trade records 1123
Ó2006 Blackwell Publishing Ltd/CNRS
Compagno, L.J.V., Dando, M. & Fowler, S. (2005). Collins Field
Guide: Sharks of the World. Harper Collins, London.
Corte´s, E. (2002). Incorporating uncertainty into demographic
modeling: application to shark populations and their conserva-
tion. Conserv. Biol., 16, 1048–1062.
Corte´s, E., Brooks, L. & Scott, G. (2002). Stock Assessment of Large
Coastal Sharks in the U.S. Atlantic and Gulf of Mexico – Final Meeting
Report of the 2002 Shark Evaluation Workshop. United States
National Marine Fisheries Service, Panama City.
Dalebout, M.L., Lento, G.M., Cipriano, F., Funahashi, N. & Baker,
C.S. (2002). How many protected minke whales are sold in Japan
and Korea? A census by microsatellite DNA profiling. Anim.
Conserv., 5, 143–152.
Fong, Q.S.W. (1999). Assessment of Hong Kong shark fin market:
implications for fishery management. Doctoral Dissertation,
University of Rhode Island, Kingston, RI.
Fowler, S.L., Camhi, M., Burgess, G.H., Cailliet, G.M., Fordham,
S.V., Cavanagh, R.D. et al. (2005). Sharks, Rays and Chimaeras: the
Status of the Chondrichthyan Fishes. International Union for Con-
servation of Nature, Gland.
Froese, R. & Pauly, D. (2006). FishBase. Available at http://
www.fishbase.org. (Last accessed 29 August 2006).
Gelman, A., Carlin, J.B., Stern, H.S. & Rubin, R.B. (1995). Bayesian
Data Analysis. Chapman and Hall, London.
Kleiber, P., Takeuchi, Y. & Nakano, H. (2001). Calculation of
Plausible Maximum Sustainable Yield (MSY) for Blue Shark (Prionace
glauca) in the North Pacific. United States National Marine
Fisheries Service, La Jolla, CA.
Kohler, N.E., Casey, J.G. & Turner, P.A. (1995). Length–weight
relationships for 13 species of sharks from the western North
Atlantic. Fish. Bull., 93, 412–418.
Lande, R., Engen, S. & Saether, B. (2003). Stochastic Population
Dynamics in Ecology and Conservation. Oxford University Press,
Oxford.
Lessa, R., Paglerani, R. & Santana, F.M. (1999). Biology and
morphometry of the oceanic whitetip shark, Carcharhinus long-
imanus (Carcharhinidae), off north-eastern Brazil. Cybium, 23,
353–368.
Little, R.J.A. & Rubin, D.B. (1987). Statistical Analysis with Missing
Data. John Wiley and Sons, New York.
McAllister, M.K., Pikitch, E.K. & Babcock, E.A. (2001). Using
demographic methods to construct Bayesian priors for the
intrinsic rate of increase in the Schaefer model and implica-
tions for stock rebuilding. Can. J. Fish. Aquat. Sci., 58, 1871–
1890.
Milner-Gulland, E.J. & Bennett, E.L. (2003). Wild meat – the
bigger picture. Trends Ecol. Evol., 18, 351–357.
Nakano, H. & Seki, P. (2003). Synopsis of biological data on the
blue shark (Prionace glauca Linnaeus). Bull. Fish. Res. Agen. Japan,6,
18–55.
Roman, J. & Bowen, B.W. (2000). Mock turtle syndrome: genetic
identification of turtle meat purchases in the southeast United
States. Anim. Conserv., 3, 61–65.
Rose, D.A. (1996). An Overview of World Trade in Sharks and Other
Cartilaginous Fishes. TRAFFIC International, Cambridge.
Rubin, D.B. (1996). Multiple imputation after 18+ years. J. Am.
Stat. Assoc., 91, 473–489.
Schaefer, M.B. (1954). Some aspects of the dynamics of popula-
tions important to the management of commercial marine
fisheries. Bull. Inter-Am. Trop. Tuna Comm., 1, 27–56.
Shivji, M., Clarke, S., Pank, M., Natanson, L., Kohler, N. &
Stanhope, M. (2002). Rapid molecular genetic identification of
pelagic shark body-parts conservation and trade-monitoring.
Conserv. Biol., 16, 1036–1047.
Shivji, M.S., Chapman, D.D., Pikitch, E.K. & Raymond, P.W.
(2005). Genetic profiling reveals illegal international trade in fins
of the great white shark, Carcharodon carcharias.Conserv. Genet.,6,
1035–1039.
Shotton, R. (1999). Case Studies of the Management of Elasmobranch
Fisheries. FAO Fisheries Technical Paper 378. Food and Agri-
culture Organization, Rome.
Smith, S.E., Au, D.W. & Snow, C. (1998). Intrinsic rebound
potentials of 26 species of Pacific sharks. Mar. Freshw. Res., 49,
663–678.
Spiegelhalter, D.J., Best, N.G., Carlin, B.P. & van der Linde, A.
(2002). Bayesian measures of model complexity and fit. J. R. Stat.
Soc. B, 64, 583–639.
Thorbjarnarson, J. (1999). Crocodile tears and skins: international
trade, economic constraints, and limits to the sustainable use of
crocodilians. Conserv. Biol., 13, 465–470.
Vincent, A.C.J. (1996). The International Trade in Seahorses. TRAFFIC
International, Cambridge.
Walker, T.I. (1998). Can shark resources be harvested sustainabil-
ity? A question revisited with a review of shark fisheries. Mar.
Freshw. Res., 49, 553–572.
Ward, P. & Myers, R.A. (2005). Shifts in open-ocean fish com-
munities coinciding with the commencement of commercial
fishing. Ecology, 86, 835–847.
Wasser, S. K., Shedlock, A.M., Comstock, K., Ostrander, E.A.,
Mutayoba, B. & Stephens, M. (2004). Assigning African elephant
DNA to geographic region of origin: applications to the ivory
trade. Proc. Natl Acad. Sci. USA, 101, 14847–14852.
APPENDIX A: METHODOLOGICAL DETAILS FOR
ESTIMATION OF SHARK NUMBER
For each species-fin position category s,p the chance that an
auctioned bag of fins falls in bin c,p(c)
s,p
,was assigned a
relatively uninformative Dirichlet prior probability density
function (pdf):
pðcÞs;pDirichletðaðcÞs;pÞ;ðA1Þ
where a(c)
s,p
¼1 for all c.
The observed frequency of bags in the six length bins for
each s, p category (b(c)
s,p
) was given a multinomial likelihood
function conditioned on the probabilities p(c)
s,p
:
bðcÞs;pMultinomial pðcÞs;p;X
c
bðcÞs;p
!
:ðA2Þ
The Dirichlet prior pdf is conjugate to the multinomial
likelihood function, and the Gibbs sampler in WinBUGS
utilizes this conjugacy to form a highly efficient sampling
algorithm (Spiegelhalter et al. 2002) to estimate the posterior
pdf of p(c)
s,p
and other quantities of interest.
The mean fin length
Ls;pjcin bin cwas modelled as a
normal random variable (RV), truncated at the lower (y
s,p|c
)
and upper (z
s,p|c
) bin boundaries
1124 S. C. Clarke et al. Letter
Ó2006 Blackwell Publishing Ltd/CNRS
Ls;pjcNðl
Ls;pjc;r2
Ls;pjcÞ:ðA3Þ
The mean (l
Ls;pjc) was set equal to the available fin length
midpoint while the variance (r2
Ls;pjc) took account of the
uncertainty in the number of fins falling within each bin
(F
s,p|c
) through the following equation
r2
Ls;pjc¼ðzs;pjcys;pjcÞ
12 Fs;pjc
2
where
Fs;pjc¼pðcÞs;pX
c
bðcÞs;pfs;pjc:ðA4Þ
Note that the variance of a uniformly distributed RV is
equal to the squared difference between the upper and lower
boundaries divided by 12.
The RV for the average number of fins per bag in each
size bin c,f
s,p|c
is log-normally distributed:
fs;pjclog normalðlogðlfÞ;r2
s;pjcÞðA5Þ
with
r2
s;pjc¼log 1 þvarðfs;pjcÞ
l2
f
!
;ðA6Þ
where
varðfs;pjcÞ¼ varðfbÞ
pðcÞs;pP
c
bðcÞs;p
ðA7Þ
given that varðfb) is the variance in fins per bag, and l
f
is the
average number of fins per bag.
Based on expert judgment gained through market
observations, l
f
and varðfb) were given the following prior
pdfs truncated at zero:
lfnormalð50;14:142Þ;ðA8Þ
varðfbÞnormalð500;10002Þ:ðA9Þ
Sensitivity analyses proved the results to be insensitive to
realistic changes in these priors. The RV for the probability
that a fin falls into a particular bin cwas computed as fol-
lows:
pf
s;pjc¼Fs;pjc
P6
i¼1Fs;pji
:ðA10Þ
The RV for the mean length across all bins was thus given
by:
Ls;p¼X
6
c¼1
ðpf
s;pjc
Ls;pjcÞ:ðA11Þ
The model used the fin length–weight data to estimate a
joint posterior pdf for the slope (a1
s,p
) and intercept (b1
s,p
)
of a log linear relationship between RVs fin length (
Ls;p) and
fin weight (
Ws;p) for each species-fin position category (s,p).
The RV
Ls;pwas then used in conjunction with the RVs for
these parameters to predict the RV mean fin weight:
log
Ws;p¼a1s;plogð
Ls;pÞþb1s;p:ðA12Þ
APPENDIX B: METHODOLOGICAL DETAILS FOR
ESTIMATION OF SHARK BIOMASS
Within the model, mean wet fin length (
Lwet
s;p) was computed
by estimating RVs for the slope (a2
p
) and intercept (b2
p
)ofa
linear relationship between mean wet fin length (
Lwet
p) and
mean dry fin length (
Lp) for each fin position, p, using
diffuse priors and the Fong (1999) data. As the 95%
posterior PIs of the intercepts for each fin position all
overlapped with zero, the linear relationship was simplified
to:
Lwet
s;p¼a2p
Ls;p:ðB1Þ
Slopes (a3
s,p
) and intercepts (b3
s,p
) of the wet fin length
(Lwet
s;p) to whole shark length (T
s,p
) linear relationship were
estimated for each category (s,p) using a hierarchical model
(Gelman et al. 1995) of the data collected in this study:
log Tobs
s;p¼a3s;pðlogðLwet;obs
s;pÞlogð
Lwet;obs
s;pÞÞ þ b3s;p:
ðB2Þ
Due to high variances in some of the data, lengths were
transformed to log space to avoid rare occurrences of
negative values and individual fin lengths were standardized
by the mean of all observed fins for each species – fin
position combination. The RVs a3
s,p
and b3
s,p
were pre-
sumed to be drawn from normal distributions, each with a
single global mean and variance across all species for each
fin position:
a3s;pNormðla3p;r2
a3pÞ;ðB3Þ
b3s;pNormðlb3p;r2
b3pÞ:ðB4Þ
Diffuse priors were assigned to the hyperparameters
la3p;r2
a3p;lb3p;and r2
b3p.
Random variables a3
s,p
and b3
s,p
were applied to the
logged value of
Lwet
s;p(eqn B1), standardized by the mean
of the observed fin lengths by species and fin posi-
tion (logðLwet;obs
s;pÞ), to predict the RV for mean whole shark
length (
Ts;p):
log
Ts;p¼a3s;pðlogð
Lwet
s;pÞlogðLwet;obs
s;pÞÞ þ b3s;p:
ðB5Þ
The mean biomass per shark (
Bf
s;p) was modelled as a log
normal RV:
Letter Shark catches from trade records 1125
Ó2006 Blackwell Publishing Ltd/CNRS
Bf
s;plog normalðl
Bs;p;r2
Bs;pjTs;pÞ;ðB6Þ
where l
Bs;p¼logð
Bs;p), the natural logarithm of the mean
biomass per shark by s,pwas obtained by applying:
logð
Bs;pÞ¼a4sþðb4slog
Ts;pÞ:ðB7Þ
The published values of the length–weight relationship
are the mean values for normally distributed RVs a4
s
and
b4
s
. An approximation for the variance in the natural log-
arithm of mean weight per animal given the mean length
(r2
Bs;pj
Ts;p) was derived using properties of the linear regres-
sion model as follows:
r2
Bs;pj
Ts;pffið1r2
sÞb42
sr2
Ts;p
r2
s
1
ns
þðlogð
Ts;pÞlogðTs;pÞÞ2
r2
Ts;pðns2Þ
"#
;
ðB8Þ
where b4
s
is the slope parameter, r
s
is the correlation
coefficient reported for the published values of a4
s
and
b4
s
for species s, n
s
is the number of samples used to
estimate the published values of a4
s
and b4
s
, and log (T
s,p
)
and r2
Ts;pare the mean and variance of the natural log-
arithm of input shark lengths used in the estimation of a4
s
and b4
s
.
Conditional on the RVs a2
p
,b2
p
,a3
s,p
and b3
s,p
in the
Monte Carlo Markov chain (MCMC) iteration, and presum-
ing that the sample distribution of lengths of sharks is
approximated by those whose fins were auctioned, it can be
assumed that log (T
s,p
) is a normally distributed RV with
mean logð
Ts;p) and variance r2
Ts;p=ns. This is because logð
Ts;p)
derives from the mean length of auctioned fins for the
category s,p and given the above conditions, and in the
absence of the length data used to estimate a4
s
and b4
s
,the
most suitable proxy for log (T
s,p
) is a RV with mean
logð
Ts;p) and a variance derived from sampling theory. The
variance r2
Ts;pwas obtained in each MCMC iteration by a
transformation of the variable for auctioned dry fin lengths
modelled in eqns A1–A11. Mean dry fin length per bin c
(
Ls;pjc, see eqn A11) is transformed to mean wet fin length
for bin c(
Lwet
s;pjc) using eqn B1. Mean wet fin length per bin c
is transformed to the natural logarithm of the mean whole
shark length per bin c(log
Ts;pjc) using eqn B5. The variance
r2
Ts;pis then approximated as follows:
r2
Ts;p¼X
6
c¼1
ðpf
s;pjcðlogðTs;pjcÞlogð
Ts;pÞÞ2Þ:ðB9Þ
Editor, Michael Hochberg
Manuscript received 10 April 2006
First decision made 22 May 2006
Manuscript accepted 27 July 2006
1126 S. C. Clarke et al. Letter
Ó2006 Blackwell Publishing Ltd/CNRS
... Charismatic marine megafauna that inhabit coastalscapes are mobulid rays, carcharhinid sharks and cheloniid sea turtles. They fulfill key roles as predators or grazers in coastal ecosystems in the (sub)tropics (Ferretti et al., 2010;Roff et al., 2016), yet their persistence is threatened by (over)fishing, loss of breeding habitat, pollution, habitat disturbance and pathogens (Clarke et al., 2006;Estes et al., 2011;Brooks et al., 2013;Vianna et al., 2013;Ward-Paige et al., 2013;Dwyer et al., 2020;Jatmiko and Catur Nugroho, 2020). As a result, most of these marine megafauna are now listed as vulnerable, near-threatened or endangered on the international union for conservation of nature red list of threatened species (Marshall et al., 2011;Brooks et al., 2013). ...
... The global demand for marine animal products such as shark fins (Clarke et al., 2006), swim bladders (Sadovy and Cheung, 2003;Clarke, 2004), and ray gill plates (White et al., 2006;Ward-Paige et al., 2013) is unsustainable (Berkes et al., 2006;Lenzen et al., 2012). Particularly for the slower life history species has the intense fishing exploitation that targets these demands resulted in population declines and increased risks of extinction, sometimes with synergistic effects of environmental conditions (Jennings et al., 1999;Schindler et al., 2002). ...
Article
Full-text available
Understanding why different life history strategies respond differently to changes in environmental variability is necessary to be able to predict eco-evolutionary population responses to change. Marine megafauna display unusual combinations of life history traits. For example, rays, sharks and turtles are all long-lived, characteristic of slow life histories. However, turtles also have very high reproduction rates and juvenile mortality, characteristic of fast life histories. Sharks and rays, in contrast, produce a few live-born young, which have low mortality rates, characteristic of slow life histories. This raises the question if marine megafaunal responses to environmental variability follow conventional life history patterns, including the pattern that fast life histories are more sensitive to environmental autocorrelation than slow life histories. To answer this question, we used a functional trait approach to quantify for different species of mobulid rays, cheloniid sea turtles and carcharhinid sharks – all inhabitants or visitors of (human-dominated) coastalscapes – how their life history, average size and log stochastic population growth rate, log(λs), respond to changes in environmental autocorrelation and in the frequency of favorable environmental conditions. The faster life histories were more sensitive to temporal frequency of favourable environmental conditions, but both faster and slower life histories were equally sensitive, although of opposite sign, to environmental autocorrelation. These patterns are atypical, likely following from the unusual life history traits that the megafauna display, as responses were linked to variation in mortality, growth and reproduction rates. Our findings signify the importance of understanding how life history traits and population responses to environmental change are linked. Such understanding is a basis for accurate predictions of marine megafauna population responses to environmental perturbations like (over)fishing, and to shifts in the autocorrelation of environmental variables, ultimately contributing toward bending the curve on marine biodiversity loss.
... Saldaña-Ruiz et al. (2017) reconstructed the historic landings of a shark fishery in the Gulf of California and showed a relative decline in landings for several species. Clarke et al. (2006) used trade data and genetic techniques to reconstruct point estimates of species-specific catches in the global shark fin trade. Clarke et al. (2006) found that for species which were traded, catches were three to four times higher than the official catch statistics. ...
... Clarke et al. (2006) used trade data and genetic techniques to reconstruct point estimates of species-specific catches in the global shark fin trade. Clarke et al. (2006) found that for species which were traded, catches were three to four times higher than the official catch statistics. Overall catches of sharks and rays in WA were 57% larger than those derived from official statistics. ...
Article
Reliable catch information is scarce for most sharks and rays worldwide, with almost half of the stocks considered to be Data Deficient due to limited species-specific catch statistics. Western Australia (WA) hosts a diverse number of shark and ray species, some of which are considered to be threatened with extinction at a global level. Commercial catch statistics only account for shark and ray landings. The present study used the best available information to reconstruct unaccounted and unreported catches for 47 shark and ray taxa to better understand the impact of fishing. For some species, there was good agreement between reconstructed catches and reported landings, but overall reconstructed catches were 57% higher than those derived from official statistics alone, underestimating the actual extraction level for many species. The reconstructed catch time series provide the basis for the assessment of all species of sharks and rays captured in WA, including protected species that interact with commercial and recreational fisheries.
... Furthermore, apart from providing an important perspective to interpreting functional and life-history evolution as being the sister group to all other extant jawed vertebrates (Gnathostomata) [15], they exhibit a genomic architecture that is likely closer to the ancestral vertebrate condition compared to teleosts [16]. Their commercial value, especially of their meat, fin and liver is increasing as targeted teleost fish become less accessible [17,18]. As a result, overfishing has profoundly altered shark and ray populations [19][20][21] and several species are facing a two-fold higher extinction risk compared to finfish [22]. ...
Article
Full-text available
Chondrichthyes occupy a key position in the phylogeny of vertebrates. The complete sequence of the mitochondrial genome (mitogenome) of four species of sharks and five species of rays was obtained by whole genome sequencing (DNA-seq) in the Illumina HiSeq2500 platform. The arrangement and features of the genes in the assembled mitogenomes were identical to those found in vertebrates. Both Maximum Likelihood (ML) and Bayesian Inference (BI) analyses were used to reconstruct the phylogenetic relationships among 172 species (including 163 mitogenomes retrieved from GenBank) based on the concatenated dataset of 13 individual protein coding genes. Both ML and BI analyses did not support the "Hypnosqualea" hypothesis and confirmed the monophyly of sharks and rays. The broad notion in shark phylogeny, namely the division of sharks into Galeo-morphii and Squalomorphii and the monophyly of the eight shark orders, was also supported. The phylogenetic placement of all nine species sequenced in this study produced high statistical support values. The present study expands our knowledge on the systematics, genetic differentiation, and conservation genetics of the species studied, and contributes to our understanding of the evolutionary history of Chondrichthyes.
... Secondary fins have little to no commercial value, which means fishers will not lose substantial income from participating. A combination of morphological identification, DNA testing, and regressions of fin size to body size can be used to reconstruct the fishery's species and size-composition from these fins (Clarke et al. 2006). We successfully implemented the approach in Belize, which provided novel information on their domestic shark fishery, and considered how other nations could adopt this type of program to obtain critical information necessary to manage their shark fisheries. ...
Article
Developing-world shark fisheries are typically not assessed or actively managed for sustainability; one fundamental obstacle is the lack of species and size-composition catch data. We tested and implemented a new and potentially widely applicable approach for collecting these data: mandatory submission of low-value secondary fins (anal fins) from landed sharks by fishers and use of the fins to reconstruct catch species and size. Visual and low-cost genetic identification were used to determine species composition, and linear regression was applied to total length and anal fin base length for catch-size reconstruction. We tested the feasibility of this approach in Belize, first in a local proof-of-concept study and then scaling it up to the national level for the 2017-2018 shark-fishing season (1,786 fins analyzed). Sixteen species occurred in this fishery. The most common were the Caribbean reef (Carcharhinus perezi), blacktip (C. limbatus), sharpnose (Atlantic [Rhizoprionodon terraenovae] and Caribbean [R. porosus] considered as a group), and bonnethead (Sphyrna cf. tiburo). Sharpnose and bonnethead sharks were landed primarily above size at maturity, whereas Caribbean reef and blacktip sharks were primarily landed below size at maturity. Our approach proved effective in obtaining critical data for managing the shark fishery, and we suggest the tools developed as part of this program could be exported to other nations in this region and applied almost immediately if there were means to communicate with fishers and incentivize them to provide anal fins. Outside the tropical Western Atlantic, we recommend further investigation of the feasibility of sampling of secondary fins, including considerations of time, effort, and cost of species identification from these fins, what secondary fin type to use, and the means with which to communicate with fishers and incentivize participation. This program could be a model for collecting urgently needed data for developing-world shark fisheries globally. Article impact statement: Shark fins collected from fishers yield data critical to shark fisheries management in developing nations.
... The long generation times and low intrinsic population growth rates of many sharks make them inherently susceptible to overexploitation 1,7,19 . Globally, sharks are landed for their meat, fins, gill plates and liver oil 20,21 and catches increased to an estimated peak of 63-273 million individuals in the early 2000s before declining owing to overfishing 6 . The first warnings of the dire status of sharks were based on boom-and-bust catch patterns and the increasing international trade in shark fins 22,23 . ...
Article
Full-text available
Overfishing is the primary cause of marine defaunation, yet declines in and increasing extinction risks of individual species are difficult to measure, particularly for the largest predators found in the high seas. Here we calculate two well-established indicators to track progress towards Aichi Biodiversity Targets and Sustainable Development Goals: the Living Planet Index (a measure of changes in abundance aggregated from 57 abundance time-series datasets for 18 oceanic shark and ray species) and the Red List Index (a measure of change in extinction risk calculated for all 31 oceanic species of sharks and rays). We find that, since 1970, the global abundance of oceanic sharks and rays has declined by 71% owing to an 18-fold increase in relative fishing pressure. This depletion has increased the global extinction risk to the point at which three-quarters of the species comprising this functionally important assemblage are threatened with extinction. Strict prohibitions and precautionary science-based catch limits are urgently needed to avert population collapse, avoid the disruption of ecological functions and promote species recovery.
... These aggregations can represent feeding or breeding locations when adults [16,66], and nurseries or growing grounds during early life stages [67][68][69]. Their aggregation behaviour also increases their vulnerability to exploitation [70,71], hence monitoring their distribution and behaviour during these periods is vital for appropriate management and conservation [70,[72][73][74]. For example, knowing where and when they aggregate can inform spatial protection management strategies, such as marine protected areas and fisheries management. ...
Article
Full-text available
Over the past decade, drones have become a popular tool for wildlife management and research. Drones have shown significant value for animals that were often difficult or dangerous to study using traditional survey methods. In the past five years drone technology has become commonplace for shark research with their use above, and more recently, below the water helping to minimise knowledge gaps about these cryptic species. Drones have enhanced our understanding of shark behaviour and are critically important tools, not only due to the importance and conservation of the animals in the ecosystem, but to also help minimise dangerous encounters with humans. To provide some guidance for their future use in relation to sharks, this review provides an overview of how drones are currently used with critical context for shark monitoring. We show how drones have been used to fill knowledge gaps around fundamental shark behaviours or movements, social interactions, and predation across multiple species and scenarios. We further detail the advancement in technology across sensors, automation, and artificial intelligence that are improving our abilities in data collection and analysis and opening opportunities for shark-related beach safety. An investigation of the shark-based research potential for underwater drones (ROV/AUV) is also provided. Finally, this review provides baseline observations that have been pioneered for shark research and recommendations for how drones might be used to enhance our knowledge in the future.
... More recently, however, developing markets and depleting numbers of traditionally commercial fish have made these "bycatch" sharks and rays increasingly desirable. Sharks and rays are also intentionally caught and killed because of the perceived threat they pose to humans as well as the incessant demand for shark products, including liver oil, fins, and gills (Fowler et al. 2002;Clarke et al. 2006;Lack & Sant 2009). Habitat depletion and environmental contamination also represent substantial dangers to Chondrichthyans. ...
Article
Full-text available
Cartilaginous fish include sharks, rays, skates, sawfish, and chimaeras. Their habitat ranges from shallow coastal waters to deep ocean floors, estuarine areas as well as rivers and inland waters. Overfishing is considered to be the main threat to their existence, but there are many more stressors that these species face. Pollution is an issue that concerns aquatic organisms at every level, and Chondrichthyans are no exception. Here, we looked at their IUCN Red List assessment, and noticed a lack of information regarding anthropogenic contamination for these species. Out of 1124 cartilaginous fish species assessed, only 17 Selachimorpha and 32 Batoidea species were considered to be facing a “pollution threat”; in most cases, the threat was assigned not from direct ecotoxicological studies of the specimens, but because the species inhabited areas likely to be contaminated. An update on the conservation status of these species is urgently needed. Further, there is a fundamental need to study the effects of contaminants on Chondrichthyans as they play a key role in aquatic ecosystems.
... According to FAO global catch production statistics , total landings of Isurus oxyrinchus increased by 69% from 2004-2009-2016(FAO 2018. And the proportion of I. oxyrinchus of global shark fins in international trade has declined from 2.7% (Clarke et al. 2006) to 0.2%-1.2% (Fields et al. 2018) over the past two decades, along with a historical decline (first 10 years with data vs. last 10 years) of 16.4% to 96% in different regions (CoP18 Proposal 42). ...
Article
Multiple paternity has been demonstrated in a variety of sharks with different reproductive modes (i.e., viviparous, ovoviviparous, adelphophagy, oviparous), although the number of sires per litter varies considerably among species. To date, such analyses have focused mainly on coastal and nearshore shark species due to the difficulty in sampling oceanic sharks. In the present study, we observed multiple paternity in the oceanic shark Isurus oxyrinchus from seven polymorphic microsatellite loci and three litters collected from Nanfangao Fishing Port. Paternity tests showed that an average of 4.6 sires were assigned to each litter of I. oxyrinchus using COLONY software, and that the average number of sires dropped to 2.5 when using GERUD. These findings suggest that multiple paternity could be a common reproductive strategy used by the shortfin mako shark, and that this mating system should be integrated into a demographic model to make more accurate population projections and risk analyses in the future.
Article
Full-text available
Extinctions on land are often inferred from sparse sightings over time, but this technique is ill-suited for wide-ranging species. We develop a space-for-time approach to track the spatial contraction and drivers of decline of sawfishes. These iconic and endangered shark-like rays were once found in warm, coastal waters of 90 nations and are now presumed extinct in more than half (n = 46). Using dynamic geography theory, we predict that sawfishes are gone from at least nine additional nations. Overfishing and habitat loss have reduced spatial occupancy, leading to local extinctions in 55 of the 90 nations, which equates to 58.7% of their historical distribution. Retention bans and habitat protections are urgently necessary to secure a future for sawfishes and similar species. Available at: https://advances.sciencemag.org/content/7/7/eabb6026
Preprint
Full-text available
Understanding the link between seamounts and large pelagic species (LPS) is critical for guiding management and conservation efforts in open water ecosystems. The seamounts along the Cocos Ridge in the Eastern Tropical Pacific (ETP) are thought to play a critical role for LPS moving between Cocos Island (Costa Rica) and Galapagos Islands (Ecuador). However, to date, research efforts to understand pelagic community structure beyond the borders of these oceanic Marine Protected Areas (MPAs) have been limited. This study used drifting-pelagic baited remote underwater video stations (BRUVS) to characterize the distribution and relative abundance of LPS at Cocos Ridge seamounts. Our drifting-pelagic BRUVS detected a total of 21 species including sharks, large teleosts, small teleosts, dolphins and one sea turtle, of which 4 are threatened species. Relative abundance and richness of LPS was significantly higher at shallow seamounts (<400m) compared to deeper ones (>400m) suggesting that seamount depth could be an important driver structuring LPS assemblages along the Cocos Ridge. Our cameras provided the first visual evidence of the schooling behaviour of S. lewini at two shallow seamounts outside the protection limits of Cocos and Galapagos Islands. However, further research is still needed to demonstrate a positive association between LPS and Cocos Ridge seamounts. Our findings showed that drifting pelagic BRUVS are an effective tool to survey LPS in fully pelagic ecosystems of the ETP. This study represents the first step towards the standardization of this technique throughout the region.
Article
Full-text available
The oceanic whitetip shark, Carcharhinus longimanus, represented 29% of the elasmobranch catch in 197 longline sets conducted off Brazil in the equatorial Atlantic (1°N to 9°S, 30° to 40°W) from 1992 to 1997. A total of 258 individuals were caught (121 males and 137 females) ranging from 71 cm to 250 cm (total length, TL). No significant difference was detected in the length vs eviscerated weight relationship between sexes. Significant differences between sexes were found for twelve morphometric features. Claspers 2.6-21.0 cm in length were found in males measuring 114 to 235 cm. In individuals up to 187 cm, claspers are flexible and smaller than 10 cm. A substantial increment in size of claspers occurred in 190 cm individuals. Testes ranging from 3 to 170 g were recorded in individuals 95 to 235 cm in length and epididymis width varied from 0.4 in to 2.4 cm. In females, nidamental gland width varied between 0.5 to 4.4 cm in 105 to 250 cm. Ovary weights from 5 to 180 g were recorded. Vitellogenic follicles were not observed in females smaller than 180 cm, but in larger individuals they vary from 1.0-4.4 cm. Liver weights ranged from 180 g in a 100 cm female to 7,500 g in a 250 pregnant female. First maturity class is 180-190 cm for both sexes. Three pregnant females (203, 213 and 250 cm in length) were caught with embryos of approximately 20 cm (n = 3), 10 cm (n = 4), and fertilised eggs (n = 9), respectively. Males and females were longer and heavier from July through December. Testis and liver weights in males were also significantly higher during this period. The mean weight of ovaries and follicle diameter decrease during this period which can be attributed to ovulation. A new-born shark caught in August with an unhealed umbilical scar suggested that birth takes place at about 70 cm.