Marine Protected Areas That Talk to Each Other: The Science of Networks

Here is an awkward fact about marine protected areas: most marine life doesn't live in them. It lives between them. A fish may spend its adult life behind a reserve's invisible boundary, but it arrived as a millimetre-long larva carried on a current from somewhere else — perhaps a reef fifty kilometres along the coast, perhaps the next reserve up-current, perhaps nowhere protected at all. Protect a single patch of ocean and you fence your house; you don't necessarily secure your food supply. This is why the modern science of MPAs is really the science of networks — systems of reserves sized, spaced and positioned so that what drifts out of one reserve seeds the next. Get the geometry right and reserves amplify each other; get it wrong and you own a scatter of beautiful islands that quietly empty out.

If you haven't read our primer on what MPAs protect and how, start there. This article is about the design layer above: what turns a collection of protected areas into a network that behaves like a single living system.

The larval fact at the heart of it all

Almost everything on a kelp reef — abalone, rockfish, urchins, lobsters, most kelps themselves — reproduces by releasing young life into the water column. Larvae (or kelp spores) drift for minutes or months depending on the species, then settle wherever currents deposit them. Adult abalone move metres in a lifetime; their larvae can travel tens of kilometres. A reef's population is therefore rarely self-contained: it is maintained by the arrival of other reefs' offspring, a process called connectivity that we explore in depth in Marine Connectivity and Drifters and Swimmers.

Connectivity has a polar opposite: self-recruitment, the fraction of a reef's offspring that settles back home. For most species, both happen — reefs are partially self-seeding and partially subsidised. And that mixture creates the core design problem. A reserve too small leaks: its adults reproduce, but nearly all the larvae wash out to sea into fishing grounds. A network spaced too far apart starves: each reserve is a source with no protected destinations, and disturbed fished reefs between them never get reseeded. A well-designed network threads the needle — reserves big enough that a useful share of their larvae settle at home, close enough together that the outbound larvae land in each other or in the fished reefs the network is meant to sustain.

The design rules, and where they came from

Over two decades of theoretical work and — increasingly — empirical measurement, a rough grammar of network design has been agreed upon. The numbers are guidelines, not laws, but they are remarkably consistent across studies:

• Size: no smaller than a few kilometres across. Many syntheses converge on reserves at least several kilometres long alongshore, so that short-dispersing species (abalone, some kelps, many invertebrates) self-seed meaningfully.
• Spacing: typically within 10–20 km of the next reserve. This matches the median dispersal distance of many ecologically important reef species. Species with month-long larval lives can jump much bigger gaps, so spacing targets are set by the short-dispersers — often the ones fisheries care most about.
• Replication: protect more than one example of every habitat. Storms, marine heatwaves and disease strike patchily — recall what sea star wasting did to the US west coast, or how a single heatwave can strip a whole stretch of kelp, as in Marine Heatwaves 101. Two or three spaced copies of each habitat type insure the network against local catastrophe.
• Representativeness: include the whole depth and exposure gradient. A network of only one habitat type protects one cast of characters; kelp forests, deep reefs, soft sediments, estuaries and canyons each host different species with different dispersal scales.
• Placement along current pathways. Up-current reserves disproportionately subsidise down-current reefs; in some regions a handful of "source" reefs seed huge stretches of coastline, and protecting them returns outsized value.

These principles are why serious network proposals — California's coast-wide MPA network, Australia's marine park zoning, the emerging Great Southern Reef planning — are designed in current-aware units rather than ad-hoc park-by-park.

Spillover: what a network does for the water in between

The case for networks isn't only ecological insurance; it's also fisheries economics. Inside a well-enforced reserve, fish get older, bigger and vastly more fecund — egg production scales steeply with body size. Two outflows follow. Adult spillover: fish and lobsters, denser inside, wander across the boundary and get caught, concentrating catches at the edges — the "fishing the line" skippers know well. Larval subsidy: the far larger flow, as the reserve's reproductive output rains onto fished reefs down-current. Because a big old female rockfish may produce orders of magnitude more larvae than the young fish that replaced her in the catch, even modest reserves can subsidise surprisingly large fished areas — provided the network is spaced so those larvae land somewhere fishable rather than in oceanic dead ends. This is the larval-dispersal calculus that connects MPA science to solutions for fisheries and coastal councils weighing their local marine planning, as in our work with governments.

Why single reserves so often disappoint

The history of marine protection is littered with small, isolated reserves that produced lovely local results — bigger fish, happier divers — yet failed to shift the trajectory of the wider coast. The network lens explains the pattern:

• Too small to self-seed. A reserve a few hundred metres across exports essentially everything it produces; when a heatwave or disease knocks it back, there is no protected neighbour to reseed it, and recovery relies on luck and distance.
• No replication of habitats. One reef type, one depth band, one exposure: a single event can erase the whole estate at once.
• Boundaries drawn by politics, not physics. Reserves placed where conflict was lowest rather than where currents deliver. An up-current source reef left unprotected is the conservation equivalent of paving the headwaters.
• No pressure relief outside the boundary. A reserve cannot offset unmanaged fishing, pollution or warming everywhere else; it is a strong node in a system, not a substitute for managing the system.

None of this is an argument against reserves — the evidence that well-designed, well-enforced protection works is overwhelming. It is an argument that the unit of design is the coast, not the cove.

What the science still argues about

Honest network science has live edges. Measuring connectivity directly is hard: larvae are tiny, abundant and mostly unmarked, so researchers triangulate with genetics (parentage analysis can literally match settlers to their parents), otolith chemistry, and high-resolution ocean models fertilised with drifter data. Dispersal kernels turn out to be fat-tailed — most larvae stay close, a lucky few go very far — and both tails matter: the nearby flow runs the fishery subsidy, while the rare long jumps knit distant reserves together and drive colonisation of recovering reefs, which is exactly the dynamic that matters for restoring degraded reefs at scale. There's also active debate about how networks should account for climate velocity — species shifting poleward as the ocean warms — which argues for protecting climatic refuges (thermal refugia) and for spacing reserves along the direction species will need to travel, not only along today's currents. And there is a growing recognition that people are part of the geometry: networks co-designed with Indigenous sea-country rangers and local fishers outperform paper parks, because enforcement and stewardship are as much social infrastructure as patrol boats. That is why Tidewright's restoration programs are built as partnerships with councils and Traditional Owners — see our community programs.

Key takeaways

• Most reef species reproduce by broadcasting drifting larvae, so no reef is an island: populations are maintained by connectivity between reefs, and conservation must be designed around it.
• Effective networks follow recurring design rules: reserves at least several kilometres long, neighbouring reserves typically within 10–20 km, every habitat replicated two or more times, and placement aligned with current pathways.
• Networks pay for the water in between via adult spillover and — far more importantly — larval subsidies from the big old, hyper-fecund animals that only fully protected areas grow.
• Single, small, politically-where-easy reserves tend to produce local ecological wins but fail to reseed after disturbance and rarely lift the wider coast.
• The frontier problems are measuring real connectivity (genetics, otoliths, models), designing for climate velocity and thermal refugia, and building networks with — not around — the communities who steward them.

FAQ

How big does a marine reserve need to be to work?

There is no single magic number, but syntheses of reserve science repeatedly converge on reserves at least a few kilometres across alongshore. Smaller reserves export nearly all the larvae they produce and can't self-seed the short-dispersing species — abalone, many urchins, some kelps — that are often the ecological keystones. Bigger is generally better for self-recruitment, provided spacing rules are still met so exports land in other protected or well-managed reefs.

What's the difference between a marine reserve and an MPA?

"MPA" is the umbrella term — any geographically defined ocean area managed for conservation, from strictly no-take sanctuaries to zones allowing some fishing or aquaculture. A "marine reserve" or "no-take zone" is the strict subset where extraction is prohibited entirely. Networks typically embed no-take cores (which rebuild age structure and fecundity) within broader multiple-use MPAs, getting both conservation and social licence.

Do protected areas actually help fisheries outside their boundaries?

For well-enforced no-take areas, the evidence for adult spillover and larval subsidy is strong, though size varies by species and configuration. Fishing-the-line effects — better catches near boundaries — are commonly reported by fishers. What reserves cannot do is compensate for chronic overfishing or habitat destruction across the rest of the seascape; they are a powerful complement to good fisheries management, not a replacement for it.

How do scientists know larvae actually travel between reserves?

Three converging lines of evidence: genetic parentage analysis (matching juveniles inside one reserve to adult parents in another), chemical signatures in fish ear bones (otoliths) that record the water they drifted through, and dispersal simulations run on measured current fields, increasingly validated against genetic estimates. All three agree on the fat-tailed pattern — most larvae stay near home, a small but crucial fraction travels far enough to link reserves.

Should a coastline have many small reserves or a few large ones?

Network theory answers: both, deployed as one system. A few large reserves provide strong self-seeding cores and resilience; a chain of smaller stepping-stone reserves maintains connectivity across long coastlines and subsidises fishing grounds between them. The design goal isn't size or count in isolation — it's that the whole coast's larvae reliably have somewhere protected to come from and somewhere viable to land.