Designing Marine Reserve Networks

Why small, isolated protected areas are not enough

By Callum M. Roberts, Benjamin Halpern, Stephen R. Palumbi, and Robert R. Warner

For a fish, the north coast of Jamaica is a lonely place. Fishing intensities are so high that most fish are caught long before they reach maturity. Coral reefs there are festooned with traps, hooks, and nets, while spearfishers hunt all day to depths of more than fifteen meters. Diving these reefs, one is struck by the absence of fish bigger than 15 cm and by the abundance of tiny fish, mostly species of little commercial interest. What is truly amazing about Jamaica is that there are any fish at all!

With few local sources of the larger, commercially important fish remaining, Jamaican fisheries are partly supplied by fish arriving from other parts of the Caribbean where breeding stocks are in better shape—perhaps as far afield as the Turks and Caicos and Bahamas (1) (M. Watson, pers. comm.). Jamaica’s over-hunted reefs are a sink to these sources.

In the past, intensive nearshore fisheries often were supplied by populations reproducing in “natural” refuges—places that were too deep, too remote, too dangerous, or too rough to fish. The behemoth lobsters that once commanded the deep sea floor off Cape Cod are an example. But as the reach of fishing expands, much of the seas are going the way of Jamaica and there are fewer sources to rely on. Without refuges, the future for fishing is bleak.

There is a solution to this problem. We can create source populations to supply fisheries by establishing new refugesñplaces that are fully protected from all fishing. While the land and sea are different in many ways, reserves are remarkably similarñprotect an area from fishing or hunting and you get more and bigger animals in a short time. Recently, such marine reserves have been in the conservation spotlight. The findings of a new research synthesis were released at the 2001 meeting of the American Association for the Advancement of Science ( Also the National Academy of Science in the USA has just published a report on marine protected areas (2). Both groups argue that the use of fully protected reserves should be greatly expanded. Moreover, they envision reserves established throughout the oceans in networks linked by the invisible threads of larval transport.

Our present approach to creating reserves falls far short of this vision. Currently, only a few fully protected reserves exist, mostly small (< 10 km2) and highly isolated (often hundreds of kilometers apart). Together they cover less than one ten-thousandth of the sea (3). As such, the challenge we face is how to design marine reserve networks that will interact effectively and meet our goals of sustained and productive marine communities. Here we describe what such networks might look like and how to go about building them.

Tapping Into Dispersal Patterns

Good reserve networks are eminently practical. What science tells us about how to design them is completely compatible with what is feasible. We need reserves in a range of sizes (but not necessarily very large), spread over broad regions, and spaced at variable distances from one another. The key is dispersal.

Populations of some species may be entirely self-supporting, even in small reserves. Although we commonly think of marine species as having long distance dispersal, a large fraction of species appears to disperse only short distances (meters to a few kilometers). These include species as diverse as tunicates, sponges, seaweeds, corals, gastropods and amphipods, and even a few species of fish.

Yet reserves that are small enough to be practical will not encompass the full spectrum of dispersal distances of marine organisms—especially for the many nomadic species whose larvae spend weeks or months in the plankton. Measurements of advancing fronts of invading species, such as the green crab (Carcinus maenas) on the Pacific coast of North America, suggest that dispersal distances of 30-100 km are common. These distances, although substantially less than those that could be traveled through passive dispersal, imply that larvae produced in a reserve will generally settle outside its borders. Small reserves will be unable to support self-sustaining populations of all of the species within them.

If long-distance dispersers are to persist, reserves must be colonized by offspring produced elsewhere. Because most existing reserves are small and isolated, that “elsewhere” must be places that are fished. Recruitment from fishing grounds accounts for the rapid increases in the numbers of widely dispersing species seen in small reserves (Box 1). However, as fishing intensifies, some vulnerable species will be unable to persist at all outside marine reserves. That “elsewhere,” then, must lie within other marine reserves. For example, giant clams (Tridacna gigas) have become extinct throughout large areas of the Pacific. Reintroduction programs are focusing first on building populations within reserves. Many other species are heading toward extinction. For example, the once widespread “common” skate (Dipturus batis) is now confined to a few de facto refuges in European waters that are too rough to fish. The lack of reproductive stocks in unprotected areas explains the lack of recovery of species of large grouper (Serranidae) in many Caribbean reserves. We must establish reserves in networks that are sufficiently dense to exchange offspring of vulnerable species. Otherwise, we will lose them.

Dispersal is critical to the success of reserves, both as conservation and fishery enhancement tools, but it is not necessary to know the dispersal behavior of every marine species to design effective reserve networks. There is a common misconception that, because exchange of offspring or propagules is necessary to sustain populations of some species in reserves, we must understand their dispersal patterns in order to put reserves in the right places. According to this view, invisible dispersal corridors pattern the ocean like highways, and reserves, like towns, should lie on these routes. Most people assume that ocean currents constitute these invisible corridors, moving eggs and larvae from place to place (4). There is little doubt that many species do make use of currents as vectors of dispersal, but most species probably do not ride them passively (5). Instead they behave in ways that interact with prevailing currents to enhance their probability of future survival.

Physical processes and behavior are combined in as many different ways as there are species, producing an exquisite variety of dispersal patterns. It is as if every species has its own road atlas, and variation in current patterns over time means that every year these atlases have to be revised and reprinted. To build reserve networks on a foundation of knowledge of the dispersal of any single species would be foolhardy.

We barely have this information for any species in any single location. Even if we did, it would be disastrous to design an entire network around such specific knowledge. The resulting plan would be ideal only for that species and others like it, failing the majority. A network of reserves must accommodate the dispersal characteristics of a huge range of species, not just a few. It must do so year after year against a background of variable ocean conditions. It must continue to perform even if those currents change as the climate warms.

How then do we create effective reserve networks? Not by detailed study of the minutiae of dispersal or by constructing multispecies models that would challenge today’s fastest supercom-puters. The answer is much simpler: bet hedging.

What we know about the range of dispersal distances indicates that we should place reserves within 10 to 50 km of one another if we want them to exchange offspring effectively. Because we know that ocean conditions change from year to year, we must place reserves on what we believe to be highways of ocean dispersal, such as the Gulf Stream current, but also on the byways, backroads, and cul-de-sacs.

We must cover the map, even the places that seem little-visited right now. As we add reserves to a network, the number of pathways among them grows faster than the number of reserves (6)(Figure 1). Some of these routes will be heavily used, others less so, but the existence of many potential connections creates breadth (all species gain some benefit) and resilience (their long-term needs can be met).

An Ecological-Economic Win-Win

What gives marine conservation planners an advantage over their terrestrial counterparts is that marine reserves can supply industry without compromising conservation objectives. The same mechanisms that transport offspring from reserve to reserve allow them to replenish exploited stocks in adjacent fishing grounds. In addition, as population densities build up in reserves, they can also begin to export juveniles and adults to fishing grounds as spillover. Fishing spots close to the Mombasa Marine National Park in Kenya, for example, have become so lucrative that fishers have agreed that only the most senior among them can fish there!

These export functions make marine reserves an attractive option and offer the possibility of an ecological-economic win-win. The economic value of reserves to fisheries, coupled with the fact that areas of the sea do not have to be purchased in the way that land does (although there are still costs), means that we have much greater opportunity for conservation.

For a reserve to benefit fisheries, it must support population build-up to levels higher than those in the surrounding fishing grounds. The greater the difference, the stronger the benefit. Marine reserves, like those on land, suffer from edge effects that diminish their effectiveness simply by emptying their contents at too high a rate. Spillover of juveniles and adults is edge-dependent. The more edge a reserve has, the faster it will export relative to its area. In addition, more dispersive species will spill over more quickly. Spillover is good for fisheries, exporting products to fishing grounds, and agile fishers tend to concentrate fishing effort close to reserve boundaries (7). Because this places species whose home ranges straddle reserve and fishing grounds at risk of capture, the more mobile species face higher risk of capture and thus need larger reserves.

Large reserves can protect a wider spectrum of species than can small reserves and will be less prone to erosion of benefits from the edges. Moreover, small reserves are hard to enforce. The smallest reserve we know of, which covered just 2.6 hectares of reef off the Caribbean island of St. Lucia (before it was incorporated into a larger protected area), lost a third of its fish biomass to illegal fishing in just a few months (3).

Nevertheless, some small reserves have performed surprisingly well. Much of what we have learned about marine reserves has come from small, pilot projects. For example, the Leigh Marine Reserve in New Zealand covers just over 5 square kilometers. Twenty years after it was created, densities of a valuable snapper (Pagrus auratus) were nearly 40 times higher in the reserve than in the surrounding fishing grounds, and spiny lobsters (Jasus edwardsii) had increased by 5-11 percent per year (8,9).

So what is the optimal size for a marine reserve? It turns out there is none. Make reserves too large and spillover to fisheries will be staunched. Make them too small and nothing will benefit. What is important is not to make reserves all the same size. We need some large reserves to provide security for more mobilespecies. In most places, human concerns constrain the upper size limit of reserves. Population pressure, the location of ports, shipping lanes, dumping sites, and oil fields, among many other factors, limit our options for creating large reserves. These constraints will generally keep reserves within the proper size range (a few kilometers to a few tens of kilometers across) to optimize spillover to fisheries.

Marine reserves should typically increase in size moving from nearshore to offshore. Small reserves will be harder to identify in offshore areas, harder for fishers to comply with, and thus harder to enforce. Offshore species also tend to be more mobile and require larger areas. Spreading reserves out, rather than creating a few large ones, is particularly critical to small-scale coastal fishers. They have much more limited access to fishing grounds than fleets of larger boats that operate far offshore. Reserves displace fishers, and if the reserves are too large they will create winners (those fishing near reserve boundaries) and losers (those who fished near the middle). Networking the area to be protected spreads benefits.

Migratory species, like those that form the mainstay of most temperate fisheries, can also be served well by networked reserves. Reserves work best when they can offer permanent protection to resident adults. This has led some people to dismiss them as a management tool for animals that migrate long distances. However, most migratory species undergo migration bottlenecks or use places that are critical to particular life stages. Reserves, although often seasonal, are already used to protect migratory species at their most vulnerable. In the U.S. Virgin Islands, spawning aggregation sites for red hind groupers (Epinephelus guttatus) are now permanently protected. It is too late though for the Nassau grouper (E. striatus), whose massive Virgin Islands’ aggregation was wiped out by fishing in the 1970s.

Fishery managers have long used closures of nursery grounds to protect juvenile fish and their habitats from by-catch and damage. Closures offer a simple means of increasing yields by allowing young animals to reach certain sizes before being caught. For example, horseshoe crab (Limulus polyphemus) nursery grounds are protected seasonally in the Delaware Bay. By incorporating into reserve networks sites that constitute migration bottlenecks or that support critical life stages, we can offer important refuges to migratory species.

Representation and Replication

We can use the same network designs that benefit fisheries to benefit conservation. On the basis of two criteria alone—representation and replication—we come close to achieving ideal networks. Maps of candidate reserve networks designed to include replicated representatives of all habitats look like archipelagoes, spreading islands of protection across broad regions.

The simplest way to protect biodiversity is to incorporate into reserves representatives of all habitats in all biogeographic regions (6). Habitats can act as a surrogate for species in reserve planning, simplifying the task of deciding what to protect. This approach reduces the need for detailed, species-level mapping and population estimates.

New computer tools help simplify the task of designing candidate networks, making it easier to juggle multiple selection criteria and the options they generate. What those tools reveal is just how wide a range of choice there is. For any given set of goals, for example, protecting 20 percent of each different habitat type in a region, there are literally thousands of possible network designs. This means that we can be flexible in locating reserves and that we can accommodate most people’s concerns. The use of such tools has been helpful in the process of choosing sites for fully protected zones in the California Channel Islands National Marine Sanctuary.

To safeguard biodiversity over the long term, reserve networks must be disaster-proof. Images of oil-drenched animals in the Galapagos Marine Reserve prove that reserve designation cannot guarantee protection. Catastrophes can strike anywhere, but we can insure against them by protecting several representatives of every habitat type in different reserves. We can also build resilience into networks by protecting habitat in proportion to the prevailing frequency and severity of natural or human disasters (10). Places regularly subject to severe disturbance need greater relative protection. For example, the Galapagos lie in the heart of a region affected by El Niños. Strong El Niños occur about once a decade and have wrought havoc with marine life. As such, the protection bar for the islands should be raised to give life a chance to spring back more quickly when favorable conditions return.

Harnessing Opportunity and Necessity

Most protected areas are built upon opportunity or necessity. Rarely are they created as part of some grand network design. Certainly, it is now possible to design networks from scratch, but most attempts to do so have been met with vigorous resistance from those who would be affected.

A much better approach is to continue to use the powerful engines of opportunity and necessity to drive network construction. Several of the United States’ National Marine Sanctuaries were created only when oil and gas exploration was imminent. Some habitats are so fragile or so threatened that they need urgent protection. For example, we are discovering rich deep-water coral communities just as new fishing technology is opening their domains to exploitation. Unless we act decisively, habitats will be destroyed faster than we can describe them.

Likewise, the will to protect fish stocks sometimes only develops after their collapse. We can use these opportunities as they arise, especially in the early stages of network development. Network designs can evolve as each new reserve is added, according to what remains unprotected. It is only later in the process, when networks near completion, that choice of locations for the final reserves becomes more constrained (Box 2).

The virtual absence of reserves in the sea is unfortunate given the pressing need for better management. But it also means that we have the opportunity to do a much better job of conservation in the sea than we have done on land. Furthermore, the fact that few marine species have become extinct in recent times, at least to our knowledge, gives us the chance to act before it is too late. Because marine reserves can help sustain industry as well as conserve biodiversity, it is possible to conceive of them covering large fractions of the seas. The near absence of property ownership in the sea allows us to put reserves in the right places to the benefit of many.

Literature Cited

1. Roberts, C.M. 1997. Connectivity and management of Caribbean coral reefs. Science 278:1454-1457.

2. N.R.C. (National Research Council). 2001. Marine Protected Areas: Tools for Sustaining Ocean Ecosystems. National Academy Press, Washington, D.C.

3. Roberts, C.M. and J.P. Hawkins. 2000. Fully protected marine reserves: A guide. Endangered Seas Campaign, WWF-US, Washington, DC, and University of York, UK. Available to download in English, and Spanish from:

4. Roberts, C.M. 1998. Sources, sinks, and the design of marine reserve networks. Fisheries 23:16-19.

5. Barber, P.H., S.R. Palumbi, M.V. Erdmann and M. Kasim Moosa. 2000. A marine Wallace’s line? Nature 406:692-693.

6. Ballantine, W.J. 1997. ‘No-take’ marine reserve networks support fisheries. In D.A. Hancock, D.C. Smith, A. Grant and J.P. Beumer (eds). Developing and Sustaining World Fisheries Resources: The State and Management. Second World Fisheries Congress, Brisbane, Australia. Pp. 702-706.

7. McClanahan, T.R. and B. Kaunda-Arara. 1996. Fishery recovery in a coral reef marine park and its effects on the adjacent fishery. Conservation Biology 10:1187-1199.

8. Kelly, S., D. Scott, A. B. MacDiarmid and R. C. Babcock. 2000. Spiny lobster, Jasus edwardsii, recovery in New Zealand marine reserves. Biological Conservation 92: 359-369.

9. Willis, T.J., R.B. Millar and R.C. Babcock. 2000. Detection of spatial variability in relative density of fishes: Comparison of visual census, angling, and baited underwater video. Marine Ecology Progress Series 198:249-260.

10. Allison, G.W., S.D. Gaines, J. Lubchenco and H.P. Possingham. Ensuring persistence of marine reserves: Catastrophes require adopting an insurance factor. Ecological Applications. In Press.

Suggested Reading

Findings from the Marine Reserves Working Group at the National Center for Ecological Analysis and Synthesis will be published in a special issue of Ecological Applications in late 2001.

Two books cited above describe fully protected marine reserves in detail for managers, decision makers, and stakeholders: Roberts and Hawkins (2000) and NRC (2001).

Keeping oceans wild: How marine reserves protect our living seas, produced by the Natural Resources Defense Council, summarizes several experiences with marine reserves in the USA.

It is available to download from:


This work is a contribution of the Developing the Theory of Marine Reserves Working Group, supported by the National Center for Ecological Analysis and Synthesis, a center funded by the National Science Foundation (Grant # DEB9421535), the University of California at Santa Barbara, the California Resources Agency, and the California Environmental Protection Agency. This work was also supported by a fellowship to CR from The Pew Charitable Trusts and by a grant from the U.S. World Wildlife Fund.

Illustration ©Guy Billout