Rebuilding Global Fisheries
Boris Worm,
1
*
Ray Hilborn,
2
*
Julia K. Baum,
3
Trevor A. Branch,
2
Jeremy S. Collie,
4
Christopher Costello,
5
Michael J. Fogarty,
6
Elizabeth A. Fulton,
7
Jeffrey A. Hutchings,
1
Simon Jennings,
8,9
Olaf P. Jensen,
2
Heike K. Lotze,
1
Pamela M. Mace,
10
Tim R. McClanahan,
11
Cóilín Minto,
1
Stephen R. Palumbi,
12
Ana M. Parma,
13
Daniel Ricard,
1
Andrew A. Rosenberg,
14
Reg Watson,
15
Dirk Zeller
15
After a long history of overexploitation, increasing efforts to restore marine ecosystems and rebuildfisheries are under way. Here, we analyze current trends from a fisheries and conservationperspective. In 5 of 10 well-studied ecosystems, the average exploitation rate has recently declinedand is now at or below the rate predicted to achieve maximum sustainable yield for seven systems.Yet 63% of assessed fish stocks worldwide still require rebuilding, and even lower exploitationrates are needed to reverse the collapse of vulnerable species. Combined fisheries and conservationobjectives can be achieved by merging diverse management actions, including catch restrictions,gear modification, and closed areas, depending on local context. Impacts of international fleetsand the lack of alternatives to fishing complicate prospects for rebuilding fisheries in many poorerregions, highlighting the need for a global perspective on rebuilding marine resources.
O
verfishing has long been recognized asa leading environmental and socioeco-nomic problem in the marine realm andhasreducedbiodiversityandmodifiedecosystemfunctioning (
1
–
3
). Yet, current trends as well asfuture prospects for global fisheries remain con-troversial (
3
–
5
). Similarly, the solutions that hold promise for restoring marine fisheries and theecosystemsinwhichtheyareembeddedarehotlydebated (
4
–
6
). Such controversies date back morethan a hundred years to the famous remarks of Thomas Huxley on the inexhaustible nature of sea fisheries (
7
) and various replies documentingtheirongoing exhaustion.Althoughmanagement authorities have since set goals for sustainableuse, progress toward curbing overfishing has beenhindered by an unwillingness or inability to bear the short-term social and economic costs of re-ducing fishing (
8
). However, recent commitmentsto adopting an ecosystem approach to fisheriesmayfurtherinfluenceprogressbecausetheyhaveled to a reevaluation of management targets for fisheries and the role of managers in meeting broader conservation objectives for the marineenvironment (
9
).In light of this debate, we strive here to join previously diverging perspectives and to provideanintegratedassessmentofthestatus,trends,andsolutions in marine fisheries. We explore the prospects for rebuilding depleted marine fish populations (stocks) and for restoring the eco-systems of which they are part. In an attempt tounify our understanding of the global fisheriessituation, we compiled and analyzed all availabledata types, namely global catch data (Fig. 1A),scientific stock assessments, and research trawlsurveys (Fig. 1B), as well as data on small-scalefisheries (
10
). We further used published eco-systemmodels(Fig.1B)toevaluatetheeffectsof exploitation on marine communities. Availabledata sources are organized hierarchically like a Russian doll: Stock assessments provide thefinest resolution but represent only a subset of species included in research surveys, which inturn represent only a small subset of speciescaught globally. These sources need to be inter- preted further in light of historical fisheries be-fore data collection and illegal or unreportedfisheries operating today (
11
). We focus on twoleading questions: (i) how do changes in ex- ploitation rates impact fish populations, com-munities, and yields, and (ii) which solutionshave proven successful in rebuilding exploitedmarine ecosystems?
Models.
A range of models is available toanalyze the effects of changes in exploitation rateon fish populations, communities, and ecosys-tems. Exploitation rate (
u
t
) is defined as the pro- portion of biomass that is removed per year, i.e.,
u
t
¼
C
t
=
B
t
where
C
is the catch (or yield) and
B
is the available biomass in year
t
. Single-speciesmodels are often used to determine the exploita-tion rate
u
MSY
that provides the maximumsustainable yield (MSY) for a particular stock.Fishing for MSY results in a stock biomass,
B
MSY
, that is substantially (typically 50 to 75%)lower than the unfished biomass (
B
0
). It has beena traditional fisheries objective to achieve single-species MSY, and most management regimeshave been built around this framework. Recentlythis focus has expanded toward assessing theeffects of exploitation on communities and eco-systems (
9
).Multispecies models can be used to predict the effects of exploitation on species composi-tion,sizestructure,biomass,andotherecosystem properties. They range from simpler communitymodelstomore-complexecosystemmodels(
12
).Figure 2 displays equilibrium solutions from a size-based community model, which assumesthat fishing pressure is spread across speciesaccordingtotheirsizeandthatasubsetofspeciesremains unfished (
13
). Results of more-complexecosystem models across 31 ecosystems and a range of different fishing scenarios were remark-ably similar (fig. S1 and table S1). With increas-ing exploitation rate, total fish catch is predictedto increase toward the multispecies maximumsustainable yield (MMSY) and decrease there-after. In this example, the corresponding exploi-tation rate that gives maximum yield
u
MMSY
is~0.45, and total community biomass
B
MMSY
equilibrates at ~35% of unfished biomass (Fig. 2).Overfishingoccurswhen
u
exceeds
u
MMSY
,whereasrebuilding requires reducing exploitation below
u
MMSY
. An increasing exploitation rate causes a monotonic decline in total biomass and average body size, and an increasing proportion of spe-ciesispredictedtocollapse(Fig.2).Weused10%of unfished biomass as a definition for collapse.At such low abundance, recruitment may beseverely limited, and species may cease to play a substantial ecological role. This model suggeststhat a wide range of exploitation rates (0.25 <
u
<0.6) yield
≥
90% of maximumcatch but with verydifferent ecosystem consequences: whereas at
u
= 0.6 almost half of the species are predicted tocollapse,reducingexploitationratesto
u
=0.25is predicted to rebuild total biomass, increase aver-age body size, and strongly reduce species col-lapses with little loss in long-term yield (Fig. 2).In addition to reconciling fishery and conserva-tion objectives, setting exploitation rate below
u
MMSY
reduces the cost of fishing and increases profit margins over the long term (
14
). This sim- plemodeldoesnotincorporatefishingselectivity;however, in practice the proportion of collapsedspecies could be reduced further by increasingselectivity through improved gear technology(
15
), by closing areas frequented by vulnerablespecies, or through offering incentives to improvetargeting practices (
16
). Such strategies allowfor protection of vulnerable or collapsed species,while allowing for more intense exploitation of others.
RESEARCH
ARTICLES
1
Biology Department, Dalhousie University, Halifax, NS B3H4J1, Canada.
2
School of Aquatic and Fishery Sciences, Uni-versity of Washington, Seattle, WA 98195
–
5020, USA.
3
ScrippsInstitution of Oceanography, University of California-San Diego,La Jolla, CA 92093
–
0202, USA.
4
Graduate School of Ocean-ography, University of Rhode Island, Narragansett, RI 02882,USA.
5
Donald Bren School of Environmental Science andManagement, University of California, Santa Barbara, CA93106
–
5131, USA.
6
National Marine Fisheries Service, Na-tional Oceanic and Atmospheric Administration, Woods Hole,MA 02543, USA.
7
Commonwealth Scientific and IndustrialResearch Organisation (CSIRO) Marine and AtmosphericResearch, General Post Office Box 1538, Hobart, TAS 7001,Australia.
8
Centre for Environment, Fisheries and AquacultureScience, Lowestoft NR33 0HT, UK.
9
School of EnvironmentalSciences, University of East Anglia, Norwich NR4 7TJ, UK.
10
Ministry of Fisheries, Post Office Box 1020, Wellington, NewZealand.
11
Wildlife Conservation Society Marine Programs,Post Office Box 99470, Mombasa, Kenya.
12
Hopkins MarineStation, Stanford University, Pacific Grove, CA 93950, USA.
13
Centro Nacional Patagónico, 9120 Puerto Madryn, Argentina.
14
Institute for the Study of Earth, Oceans, and Space, Univer-sity of New Hampshire, Durham, NH 03824
–
3525, USA.
15
Fisheries Centre, University of British Columbia, Vancouver,BC V6T 1Z4, Canada.*To whom correspondence should be addressed. E-mail:bworm@dal.ca (B.W.); rayh@u.washington.edu (R.H.)
31 JULY 2009 VOL 325
SCIENCE
www.sciencemag.org
578
These results suggest that there is a range of exploitation rates that achieve high yields andmaintain most species. To test whether current fisheries fall within this range, we evaluatedtrends in 10 large marine ecosystems for which both ecosystem models and stock assessmentswere available (
10
). Figure 3A shows exploita-tion rate and biomass trajectories derived from 4to 20 assessed fish or invertebrate stocks per ecosystem. These stocks typically represent most of the catch, and we assumed that trends in their exploitation rates represent the community as a whole. Ecosystem models were used to calculate
u
MMSY
(light blue bars) and the exploitation rateat which less than 10% of the fished species are predictedtobecollapsed(
u
conserve
,darkbluebars).Across the 10 examined ecosystems, MMSY was predicted at multispecies exploitation rates of
u
MMSY
= 0.05 to 0.28 (mean of 0.16), whereasavoiding10%collapseratesrequiredmuchlower exploitationratesof
u
conserve
=0.02to0.05(meanof 0.04).Up to the 1990s, assessed species in 6 of the10 ecosystems had exploitation rates substantial-lyhigherthanthosepredictedtoproduceMMSY(Fig. 3A). Only the eastern Bering Sea has beenconsistentlymanagedbelowthatthreshold.Sincethe 1990s, Iceland, Newfoundland-Labrador, the Northeast U.S. Shelf, the Southeast AustralianShelf, and California Current ecosystems haveshown substantial declines in fishing pressuresuch that they are now at or below the modeled
u
MMSY
. However, only in the California Cur-rent and in New Zealand are current exploita-tion rates predicted to achieve a conservationtarget of less than 10% of stocks collapsed (Fig.3A). Declining exploitation rates have contrib-uted to the rebuilding of some depleted stocks,whereas others remain at low abundance. Aver-aged across all assessed species, biomass is stillwell below
B
MSY
in most regions. However, biomass has recently been increasing above thelong-term average in Iceland, the Northeast U.S. Shelf, and the California Current, whileremaining relatively stable or decreasing else-where (Fig. 3A).
Scientific stock assessments.
Stock assess-ments quantify the population status (abundance,length, and age structure) of targeted fish or invertebratestocks.Weexploredthestatusof166stocks worldwide for which we were able toobtain estimates of current biomass and exploi-tation rate (Fig. 3B). For about two-thirds of theexaminedstocks(63%),biomass(
B
)hasdropped below the traditional single-species management target of MSY, that is,
B
<
B
MSY
. About half of those stocks (28% of total) have exploitationratesthatwouldallowforrebuildingto
B
MSY
,that is,
u
<
u
MSY
,whereasoverfishingcontinuesintheremainder (
u > u
MSY
in 35% of all stocks).Another 37% of assessed stocks have either not fallen below
B
MSY
or have recovered from previous depletion; most stocks in this category(77%) are in the Pacific. The weight of theevidence, as shown by the kernel density plot inFig. 3B, indicates that most assessed stocks have
Fig. 1.
Data sourcesused to evaluate globalfisheries.(
A
)Globalcatchdata; colors refer to thenatural logarithm of theaverage reported catch(metrictonkm
−
2
year
−
1
)from 1950to 2004). (
B
)Otherdata:Stockassess-ments quantify the statusof exploited populations;researchtrawlsurveysareusedtoestimatefishcom-munity trends; ecosystemmodelsareusedtoassessresponses to fishing. Eco-systems that were ana-lyzed in some detail arehighlightedingreen(notoverfished), yellow (lowexploitationrate,biomassrebuilding from overfish-ing),orange(lowtomod-erate exploitation rate,notyetrebuilding),orred(high exploitation rate).
< -5< -4< -3< -1< 0> 1> 2> 4> 6> 8
Ln(Catch)
< -6
A
=Research surveys=Stock assessments=Ecosystem models
B
www.sciencemag.org
SCIENCE
VOL 325 31 JULY 2009
579
RESEARCH ARTICLES
fallenbelowthe biomassthatsupportsmaximumyield (
B
<
B
MSY
) but have the potential torecover, where low exploitation rates (
u
<
u
MSY
)aremaintained.Notethatmoststockassessmentscome from intensely managed fisheries indeveloped countries, and therefore our resultsmay not apply to stocks in many developingcountries,whichareoftennotassessedbutfishedat high exploitation rates and low biomass. Fullresults are provided in table S2.When we combined the biomass estimatesof stocks assessed since 1977 (
n
= 144, Fig.4A), we observed an 11% decline in total bio-mass. This trend is mostly driven by declines in pelagic (mid-water) species, whereas large de-clinesindemersal(bottom-associated)fishstocksin the North Atlantic were offset by an increasein demersal biomass in the North Pacific after 1977. This shows how a global average canmask considerable regional variation. Althoughsome ecosystems showed relative stability (e.g.,theeasternBeringSea,Fig.4B),someexperienceda collapse of biomass (e.g., eastern Canada,Fig. 4C), whereas others indicated rebuildingof some dominant target species (e.g., Northeast U.S. Shelf, Fig. 4D). These regional examplesillustrate different stages of exploitation andrebuilding.
Research trawl surveys.
The best sources of information to assess the state of fished commu-nities are repeated scientific surveys that include both target and nontarget species. We analyzedresearch trawl survey data from 19 ecosystemswhere such data were available (see Fig. 1B for locations and fig. S2 and table S3 for full data set). We found that community trends averagedacross all surveys (Fig. 4E) were broadly similar tothecombinedbiomasstrendsseenintherecent assessments (Fig. 4A), with similar signatures of stability (Fig. 4F), collapse (Fig. 4G), and re-covery(Fig.4H)inselectedregionalecosystems.Few of these surveys, however, reached back tothe beginning of large-scale industrial exploita-tion in the 1950s and early 1960s. Where theydid, for example, in the Gulf of Thailand and in Newfoundland, they revealed a rapid decline intotal biomass within the first 15 to 20 years of fishing (fig. S2) as predicted by ecosystemmodels (Fig. 2). These declines were typicallymost pronounced for large predators such asgadoids (codfishes) and elasmobranchs (sharksand rays). Subsequent to the initial decline, total biomass and community composition have oftenremained relatively stable (fig. S2), althoughthere may be substantial species turnover andcollapses of individual stocks (see below).Across all surveys combined (
10
), we docu-mented a 32% decline in total biomass, a 56%decline in large demersal fish biomass (species
≥
90 cm maximum length), 8% for medium-sizeddemersals (30 to 90 cm), and 1% for small de-mersals (
≤
30 cm), whereas invertebrates increased by 23% and pelagic species by 143% (Fig. 4E).Increases are likely due to prey release from de-mersal predators (
17
,
18
).The trawl surveys also revealed changes insize structure that are consistent with model predictions: average maximum size (
L
max
)declined by 22% since 1959 when all commu-nities were included (Fig. 4M). However, therewere contrasting trends among our focal regions:
L
max
changedlittleintheeasternBeringSea over the surveyed time period (Fig. 4N), droppedsharply in the southern Gulf of St. Lawrence,eastern Canada (Fig. 4O), as large demersalstocks collapsed, and increased because of re- building of large demersals (particularly haddock)on Georges Bank, Northeast U.S. Shelf (Fig. 4P).These trends included both target and nontarget species and show how changes in exploitationrates affect the broader community. PublishedanalysesoftheGulfofSt.Lawrenceandadjacent areas in eastern Canada demonstrate that thesecommunity shifts involved large changes in pre-dation regimes, leading to ecological surprisessuch as predator-prey reversals (
19
), trophiccascades (
17
), and the projected local extinctionof formerly dominant species (
20
). Research onthe Georges Bank closed area (
21
) and in marine protected areas worldwide (
22
) has shown howsome of these changes may reverse when pred-atory fish are allowed to recover. This revealstop-down interactions cascading from fishers to predators and their multiple prey species as im- portant structuring forces that affect community patterns of depletion and recovery (
18
).
Global fisheries catches.
The benefits andcostsinvolvedinrebuildingdepletedfisheriesaredemonstratedbyananalysisofcatchdata.Global
0.00.20.40.60.81.0020406080100
Exploitation rate
P e r c e n t o f m a x i m u m
MaintainingbiodiversityMaintaininghigh catchMaintaininghigh employment
Total catchTotal biomassMean LmaxCollapsed species
MMSY
RebuildingOverfishing
Fig. 2.
Effects of increasing exploitation rate on a model fish community. Exploitation rate is theproportion of available fish biomass caught in each year. Mean
L
max
refers to the averagemaximum length that species in the community can attain. Collapsed species are those for whichstock biomass has declined to less than 10% of their unfished biomass. This size-structured modelwas parameterized for 19 target and 2 nontarget species in the Georges Bank fish community (
13
).It includes size-dependent growth, maturation, predation, and fishing. Rebuilding can occur to theleft, overfishing to the right, of the point of maximum catch. Three key objectives that informcurrent management are highlighted: biodiversity is maintained at low exploitation rate, maximumcatch is maintained at intermediate exploitation rate, and high employment is often maintained atintermediate to high exploitation rate, because of the high fishing effort required.
31 JULY 2009 VOL 325
SCIENCE
www.sciencemag.org
580
RESEARCH ARTICLES
Search History:
Result 00 of 00
00 results for result for
p.
More From This User
Notes
Load more
