Description of Studies Estimating Marine Reserve Area Requirements
~ enlarge ~
(Amended from Roberts and Hawkins, in press, and reproduced with permission)
Argues for a target of 10% of all of the marine habitats in New Zealand to be protected. The key principle at stake is that we should not fish everywhere. Some areas should be set aside as refuges from exploitation for ethical reasons. Ten percent, Ballantine says, “has a long traditional use as a figure that signifies importance without serious hurt.” It contrasts favorably with the 90% left open to exploitation and is conservative compared to the protected land area of New Zealand. However, he accepts that it represents a call to arms for conservation rather than being scientifically-based.
OBJECTIVE: RISK MINIMIZATION
Lauck et al., 1998:
Examined the combined effects of variation in stock productivity, and errors in estimating mortality and population size, on the probability of managers successfully maintaining populations above target levels. In a simple model showed that, in the face of uncertainty in fishing mortality, reserves covering between 31 and 70% of fishing grounds would be needed to maintain populations above 60% of their unexploited size (argued to be an economic optimum) over a 40
year time horizon. The area of reserve required increased with fishing intensity. Furthermore, the greater the uncertainty in fishing mortality (which is equivalent to decreasing management control), the larger the reserves required.
Recommended maintaining exploited populations at 75% of their unexploited size in order to avoid recruitment overfishing.
Guénette et al., 2000:
Used a spatially explicit model to examine whether reserves could have prevented the collapse in 1992 of the migratory northern cod (Gadus morhua) population off eastern Canada. Found that, in the absence of other management measures, reserves covering 80% of the area would have been necessary, but that with temporal closures to trawls and gill nets, reserves covering 20% of the area would have been adequate.
Looked at the use of reserves as a tool to maintain fish populations above target levels. Found that if a stock was initially heavily fished (i.e., starts at 35% of its unfished size) reserves of 20 and 30% of the management area guaranteed persistence above this level for 20 and 100 years, respectively. The greater the level of population desired, the longer the planning horizon, and the higher the degree of variability in fishing mortality (= less control over fishing), the larger reserves are required to maintain target populations. Reserves increased cumulative yields from the fishery when populations were initially heavily exploited.
Used fishery models to estimate that maintaining fish populations above 20% of their unexploited size would avoid recruitment overfishing.
Mace and Sissenwine, 1993:
For 91 fish populations (representing 27 species) in North America and Europe, calculated that the average threshold replacement stock size corresponds to a 20% spawning potential ratio (one fifth of unexploited population size). Maintaining at least a 30% spawning potential ratio would avoid recruitment overfishing for 80% of these species; therefore, a 35% spawning potential ratio would be a conservative management target. However, safe minimum population levels ranged as high as 70% for some species.
Argued that, where the nature of the relationship between population size and recruitment is unknown, a precautionary approach would be to aim to maintain populations above 40% of their unexploited size.
Used a bioeconomic model to examine effects of different reserve areas on economic yields from the Barents Sea cod fishery. Reserves reduced economic yield from the fishery but increased cod population size. The system was also modeled with an ecological shock - a ten year period of recruitment failure. Reserves supported populations through this recruitment failure and were found to be bioeconomically beneficial when there were moderate levels of movement of cod from reserves to fishing grounds (40 to 60% of cod leave reserve in a year). This allowed reserve benefits to be captured by the fishery. The largest reserves modeled, covering 70% of the management area, offered the greatest future security for stocks, but had the highest cost in terms of current yields. How large reserves should be depends on the degree to which populations are subject to external shocks, and the degree of risk managers are willing to accept. In general, reserves covering 30 to 50% of the area provided significant protection for stocks without greatly reducing current economic benefits.
Man et al., 1995:
Modeled the persistence of an exploited metapopulation distributed across a series of habitat patches. Reserves (protected patches) became highly beneficial to population persistence as the local extinction rate in patches increased (due to increasing fishing intensities). This is because reserves provided a source of offspring to replenish fished out patches. Reserves became beneficial as exploitation rates increased, reaching a maximum of 50% of the patches protected at the highest levels of fishing. However, over a wide range of fishing intensities, optimal reserve fractions ranged between 20 and 40%.
OBJECTIVES: RISK MINIMIZATION AND BYCATCH AVOIDANCE
Soh et al., 1998:
Modeled the effects of closing hotspot areas for catches of two species of rockfish in the Gulf of Alaska. The fishery for these species is unselective and currently there are high levels of discards of over-quota fish, ranging from 15 to over 60% of catches. Three areas of reserves were simulated, covering approximately 4, 9 and 16% of the trawlable shelf area of the region. Because reserves
allowed all catches to be landed, rather than fishers having to discard fish, none of the reserve areas resulted in reduced catches. Reserves played a key role in increasing biomass of both species over a 20 year time horizon, whereas without reserves, biomass declined. The authors concluded that placing reserves in hotspots of adult fish biomass would enable even the smallest areas simulated to significantly improve on current management.
OBJECTIVES: RISK MINIMIZATION AND YIELD MAXIMIZATION
Foran and Fujita, 1999:
Modeled the value of reserves on rebuilding egg output by stocks of Pacific Ocean perch (Sebastes alutus), and catches, under optimistic and pessimistic assumptions of recruitment. Found that the benefits of reserves were sensitive to levels of recruitment. For example, a 10% reserve system would decrease long-term catches by 8% if recruitment were good, while the same reserve would increase catches by 15% if recruitment were poor. As the fraction protected increased, so fishing rates outside reserves had to be increased to maintain yields. The maximum long-term catch was from a reserve area of 25% and a moderately heavy level of fishing outside. The highest catch levels can be maintained using a range of reserve sizes provided fishing effort outside can be adjusted to appropriate levels. However, reserves increased the resilience of the stock to higher levels of fishing and therefore provide a risk averse management approach.
Guénette and Pitcher, 1999:
Used a dynamic model, which included weight-fecundity and stock-recruitment relationships to examine the effects of reserves on cod (Gadus morhua). Found that reserves do not increase yields until cod are exploited at higher levels than necessary to achieve maximum sustainable yield. At higher fishing intensities, reserves prevented collapse in catch, with 30% reserves maintaining the highest yields of the four reserve areas modeled (10, 30, 50 and 70%). Larger reserves (> 30% protected) provided more robust biomass of spawning fish and reduced the number of years with poor recruitment compared to a no reserve regime. Increasing transfer rates of fish from reserves to fishing grounds decreased the benefits from reserves. However, even for highly mobile fish, reserves should be able to maintain higher spawning stocks than without them.
OBJECTIVE: YIELD MAXIMIZATION
Pezzey et al., in press:
In a bioeconomic model showed that the reserve area that maximized catches in coral reef fisheries varied between 0 and 50% of the total area, depending on the intensity of fishing outside reserves. As fishing intensity increases, so greater fractions of the fishing grounds must be protected to sustain catches. They calculated that reserves covering 21%, 36% and 40% would be required to sustain yields in the fisheries of Belize, St. Lucia and Jamaica, representing a gradient from moderate to intensive exploitation.
Sladek Nowlis and Roberts, 1997, 1999:
Using a single-species model, applied to four different species, showed that the fraction of a management area required in reserves depends on intensity of exploitation. Reserves were only effective in increasing catches when species were overfished. As fishing intensity increases, larger and larger reserves are required to sustain catches. In the most intensively exploited areas of the Caribbean, reserves covering 75-80% would be needed to maximize catches. However, at more moderate fishing intensities, reserves covering 40% of the management area would offer major benefits to yields.
Sladek Nowlis, 2000:
Modeled the effects of reserves on catches of the Caribbean white grunt (Haemulon plumieri). At moderate fishing intensities (20% of fishery recruited individuals removed per year) catches peaked with reserves covering 30% of the management area.
Sladek Nowlis and Yoklavich, 1998:
Used a population model to examine the potential for reserves to enhance catches of a Pacific rockfish, the Boccacio (Sebastes paucispinis). They found that reserves could produce moderate to great enhancements in catch depending on how overfished the species was to begin with. Optimal reserve areas, those producing the greatest long term catches, ranged from roughly 20 to 27% of the management area as fishing intensities grew.
Holland and Brazee, 1996:
Simulated the effects of reserves on catches from the red snapper (Lutjanus campechanus) fishery in the Gulf of Mexico. Found that reserves would not
benefit catches until the species was overfished. For a range of heavy exploitation rates, optimal reserve areas (those that maximized catches) increased from 15 to 29% of the area as fishing pressure increased. However, in economic terms of net present value, optimal reserve areas were reduced from these values as the rate of discounting of the future increased (in other words as the relative value afforded to present compared to future catches grows).
Used a bioeconomic model to examine effects of reserves on spawning stock size, catches and costs of fishing for a mobile species like the cod (Gadus morhua). Assumed open access fishing outside reserves and found that reserves would have to be very large (70-80% of the management area) in order to produce catches and spawning stock levels equivalent to those of an optimally controlled fishery (one where stock size is held at 60% of the unexploited level). However, optimal control is an unrealizable economic abstraction and, compared to open access, reserves fared well. When covering between 50 and 80% of the area they produced increases in spawning stocks of 40-130%. Catches were greater than open access over a range of 10-80% of the area protected. The area that needs to be protected reduces when controls on fishing are implemented in remaining fishing grounds. However, reserves increased the costs of fishing and tended to promote overcapacity. The model ignored possible increases in catch from increased reproduction by the stock.
Used a yield per recruit model for Georges Bank cod (Gadus morhua) to examine reserve effects on spawning stock biomass and yield in relation to reserve area, fishing pressure and rate of movement of fish from reserves to fishing grounds. Reserves were very effective at increasing spawning stock biomass. However, they decreased catches unless there were moderate rates of movement of fish from reserves to fishing grounds (although the model did not consider possible enhancements in catch that might be provided by increased reproduction by protected stocks). Reserves became more effective as fishing intensities increased, and the area of reserve needed to increase catch grew as the mobility of the fish increased. For transfer rates from reserve to fishing grounds of 50% of the population per year, reserve areas of between 10 and 40% of the fishing grounds increased catches, the area needed rising over this range as fishing intensities increased.
Used a yield per recruit model to examine effects of reserves on catches of fish on Pacific coral reefs. Reserves substantially increased spawning stock biomass for three model fish species with differing levels of mobility. Spawning stock increases were greatest for the least mobile species, and reserves became more beneficial as fishing intensities increased. However, reserves almost always decreased yield per recruit. Nevertheless, increases in spawning stock biomass reduce risk of over-exploitation, and reserves ranging from 20 to 50% of the management area would offer significant levels of insurance against overfishing, although at increasing cost to present catches. Like Polacheck (1990), DeMartini ignored the possible benefits from increased reproduction by protected stocks. If included, reserves could potentially have increased catches (see Sladek Nowlis and Roberts 1997, 1999).
Hastings and Botsford, 1999:
Found that, for a wide range of biological conditions, marine reserves could offer equivalent yields to conventional fishery management tools. For species that reproduce over long lifespans, the fraction of area that needs to be protected as reserves is smaller than the fraction of the adult population that needs to be protected under conventional management. This is because animals can reproduce over longer periods in reserves than fishing grounds. For example, maintaining reproductive output at 35% of the unexploited level might require less than 35% of the area in reserves.
Botsford et al., 1999:
Modeled the effects of reserves on catches of California red sea urchins (Strongylocentrotus franciscanus). They showed that reserves would benefit catches where the slope of the stock versus recruitment curve is shallow (i.e., the species is vulnerable to recruitment overfishing). By contrast, if the slope is steep, and the species is therefore resilient to recruitment overfishing, reserves would reduce catches (although still increasing spawning stocks). However, the value of the slope is uncertain for most fished species, including this urchin. They found that, over the range of vulnerability where reserves would increase catches, the catch-maximizing fraction of the management area in reserves varied from 8 to 33%. For the most probable level of vulnerability for the sea urchin, they concluded that reserves covering 17% of the coast could increase long-term equilibrium catches by 18%.
Attwood and Bennett, 1995:
Modeled the effects of reserves on catches of three species of surf zone fish that are targeted by recreational anglers. Reserves would increase catches for two of
the species, while reducing risk of recruitment overfishing of the third by increasing spawning stocks. Modeled catches of Galjoen (Dichistius capensis) peaked at 65% of the fishing grounds in reserves, while those for blacktail (Diplodus sargus) peaked at around 25-30% of the coast protected. The results suggested a combined management strategy would be successful for the three species, with one third of the area protected distributed into reserves between 7 and 22km long across the coast of South Africa.
Quinn et al., 1993:
Used a population model to explore the role of reserves in managing the fishery for the red sea urchin (Stronglyocentrotus franciscanus) in California. This species is subject to strong Allee effects at reproduction and at recruitment. They require high adult densities for successful fertilization of eggs, and juveniles recruit to areas of high adult density and survive best under an adult ‘spine canopy.' The authors simulated the effects of reserves on population sizes and catch rates for no reserves and three reserve areas: 17, 33 and 50% of the coast. Population sizes and sustained catches were greatest with 50% of the coast protected for all except the lightest level of fishing examined. This result was partly due to the spacing of reserves in relation to dispersal distance of the sea urchins. At the lowest fraction of the coast protected, reserves were too far apart for offspring to disperse from one to another.
Simulated the effects of creating reserves in the North Sea on the fishing mortality of cod (Gadus morhua). Found that creating reserves covering 10% of the area would lead to reduction of mortality of only 5% at the lowest transfer rate of cod from reserves to fishing grounds. Protecting 25% could reduce mortality by 10-14%. However, cod were assumed to be homogeneously distributed across the region as was fishing effort. A more realistic simulation would probably have found greater benefits from protecting the same fractions of the area but in places where cod are more aggregated and catch higher.
OBJECTIVE: BIODIVERSITY REPRESENTATION
Turpie et al., in press:
Divided the South African coast into fifty-two 50km sections to explore designs for systems of marine reserves that would represent all species of marine fish present, and all biogeographic areas. Analyses of complementarity were used to design the most space-efficient systems of reserves. A system covering 10% of the coast could be designed that would represent 97.5% of the species. However,
this would not represent 15 species endemic to South Africa. A reserve system covering 29% of the coast would represent all of the species. Representing all species in the core regions of their ranges, a commonly stated conservation goal, would require 36% of the coast to be protected.
Bustamante et al., 1999:
Developed a design for a representative system of fully-protected zones for coastal habitats in the Galapagos Marine Reserve. This reserve covers the entire archipelago. Their objectives were to protect all of the ‘visiting sites' in the archipelago, areas of high biological importance, and to represent all the different coastal habitat types in each of the five biogeographic zones encompassed by the islands. To achieve this, they calculated it would be necessary to protect 36% of the coastline from fishing.
Halfpenny and Roberts, in review:
Designed a reserve system for the continental shelf seas of north-western Europe with the aim to represent all habitats and biogeographic regions present, and to replicate them in different reserves. Two systems covering 10% of the region were designed and were successful in achieving sufficient replication for most, but not all of the biogeographic regions and habitats.
OBJECTIVE: MAINTENANCE OF GENETIC VARIATION
Trexler and Travis, 2000:
Modeled the effects of fully-protected reserves to prevent or reverse undesirable selective effects of fishing, and promote genetic diversity. Found that, under the most likely selective regimes, a reserve covering just 1% of the management area would have marked conservation benefits. Benefits increased rapidly with the proportion of the area in reserves. A 10% reserve decreased directional selection by 60%, while a 20% reserve would eliminate the selective effects of fishing from the population entirely.
OBJECTIVE: INCREASE CONNECTIVITY AMONG RESERVES
Roberts, in review a:
Used a simple model in which reserve size and the fraction of the management area covered by reserves were varied to explore connectivity among reserves. Connectivity rapidly increased (= decreasing inter-reserve distances) as the proportion protected increased. For any given reserve proportion, connectivity also
increased as the size of individual reserves was decreased. Connectivity increases were asymptotic, with the greatest decreases in inter-reserve distance manifested over the range of 5-30% of the management area protected, with reserves getting 76% closer to each other over this range of protection. He also examined connectivity as the ‘target size' of reserves for dispersing propagules, expressed as the number of degrees of horizon covered by reserves. Target size increased steeply as the proportion of the management area protected grows, and was four times greater at 30% of the area in reserves compared to 5%.
OBJECTIVE: MAINTENANCE OF UNDISTURBED HABITAT
Allison et al., in review:
Looked at the effect of natural and human catastrophes on coastal ecosystems. Calculated that, if our aim is to protect a certain proportion of habitats in an undisturbed state, we must protect a larger fraction of the area. How much larger depends on the spatial extent of disturbance events, their frequency and rate of recovery of habitats. The more frequent a disturbance, and the longer the recovery time, the larger the fraction of a management area that must be protected in order to meet conservation targets.