The Ocean Climate Fund's open playbook for designing, governing, financing and running marine protected areas, informed by the Azores and other leading MPAs, for countries, communities and the BBNJ high-seas process.
Designated MPAs covered 9.8% of the ocean by April 2026, but only 3.5% is effectively protected. The gap is not science: it is governance, money and durability.
We know which MPAs recover: those that are no-take or highly protected, well enforced, old enough, large enough, isolated, and backed by communities that benefit. Gökova Bay, a few hours west of Antalya, shows it: community-enforced no-take zones produced fish biomass far above unprotected sites, and a fishing cooperative reported revenue up more than 180%.
Paper parks lack staff, enforcement and finance. Strong parks get rolled back when revenue drops or politics shifts, as Palau and the US Pacific monuments show. The BBNJ Agreement, in force since 17 January 2026 with its first Conference of the Parties due by January 2027, opens the high seas to MPAs but leaves management structures undecided.
The Fund's contribution is a package that treats governance as portable infrastructure: a standard set of governance characteristics, encoded in Propel, connected to open data (the Ocean Data Initiative, Gaia and the Bermuda Standard) and to finance that outlives political cycles.
Phase 1 ships the framework for COP31; phase 2 deepens it for the first BBNJ Conference of the Parties; phase 3 turns it into tools and pilot sites.
| # | Module | What it delivers | Phase |
|---|---|---|---|
| 1 | Evidence base | What makes MPAs work, protection levels, standards | COP31 |
| 2 | Governance characteristics | The exportable governance spec (Propel schema) | COP31 |
| 3 | Implementation pathway | Step-by-step from scoping to enforcement, with gates | COP31 |
| 4 | Case library and ranking | Success cases and cautionary tales, scored on one template | COP31 |
| 5 | Success metrics | Ecological, social, economic and governance indicators | COP31 |
| 6 | Roadblocks register | Common failure points and proven mitigations | COP31 |
| 7 | Marine spatial planning | How MPAs sit inside a whole-ocean plan | BBNJ COP1 |
| 8 | Finance playbook | Start-up and recurring finance models | BBNJ COP1 |
| 9 | Monitoring and enforcement | Surveillance stack, ranger models, remote sensing | BBNJ COP1 |
| 10 | Adaptive management | Review cycles, triggers, course-correction rules | BBNJ COP1 |
| 11 | High-seas MPAs (BBNJ) | Proposal route and governance options beyond national waters | BBNJ COP1 |
| 12 | Data resources | Curated datasets wired into the Ocean Data Initiative and Gaia | BBNJ COP1 |
| 13 | Rights, equity and communities | Consent, benefit-sharing, grievance mechanisms | BBNJ COP1 |
| 14 | Toolkit and pilots | Templates, Propel module, scorecard, two or three pilot sites | 2027 |
The governance spec, the success scorecard and the data layer are the three modules that become software.
MPAs recover strongly when they are No-take, well Enforced, Old (10+ years), Large (100+ km²) and Isolated. Sites with four or five of these features vastly outperform those with one or two, and most of the world's MPAs have one or two.
Every MPA is classed by stage (proposed, designated, implemented, actively managed) and by protection level (fully, highly, lightly or minimally protected). Only implemented MPAs at full or high protection count as effective.
The certification standard for well-run protected areas, assessed on good governance, sound design and planning, effective management, and successful conservation outcomes.
The outcomes are large where these conditions hold. At Cabo Pulmo in Mexico, a community-supported no-take reserve, total fish biomass rose 463% in 14 years, with top predators up elevenfold. Recovery tracked enforcement, social cohesion and community leadership, not just the line on the map.
Design rule for this package: every MPA the Fund helps with is scored on NEOLI, classified with the MPA Guide, and benchmarked against the Green List from day one.
Each characteristic becomes a field in a Propel MPA governance module, so any jurisdiction, community or BBNJ proposal can start from the same, auditable structure.
| # | Characteristic | What good looks like | Proposed Propel fields |
|---|---|---|---|
| 1 | Legal mandate | Designated in statute, not only by decree; clear legal basis | legal_basis instrument_type |
| 2 | Durability (anti-rollback) | Downgrade or degazettement needs science review and a supermajority; protection can ratchet up, not down | ratchet_rule amendment_threshold |
| 3 | Objectives | Measurable conservation and social goals with baselines | objectives[] baseline_ref |
| 4 | Boundaries and zoning | Mapped zones, each with an MPA Guide protection level | zones[] protection_level |
| 5 | Activity rules | Matrix of allowed, regulated and banned activities per zone | activity_matrix |
| 6 | Decision rights | Who decides what, at what level, with what quorum | authority_map quorum_rules |
| 7 | Participation and consent | Stakeholder bodies; free, prior and informed consent where Indigenous or customary rights apply | stakeholder_bodies[] fpic_status |
| 8 | Rights and tenure | Recognised access rights for small-scale and customary users | access_rights[] |
| 9 | Enforcement | Named enforcement bodies, community rangers, penalties, surveillance | enforcement_bodies[] penalty_schedule |
| 10 | Dispute and grievance | Accessible grievance route with response times and an escalation path | grievance_process jurisdiction_of_resolution |
| 11 | Finance and fiduciary | Ring-fenced, multi-year funding; independent trust with a mixed board | finance_sources[] trustee board_composition |
| 12 | Benefit sharing | Rules for how fees, credits and fund income reach communities | benefit_rules |
| 13 | Transparency and data | Open boundaries, rules, budgets and monitoring data | data_endpoints[] disclosure_level |
| 14 | Monitoring and review | Fixed review cycle; indicators and triggers for change | review_cycle indicators[] triggers[] |
| 15 | Accountability and audit | Independent audit of finances and outcomes; public scorecard | audit_body scorecard_ref |
| 16 | Nesting | Links to the marine spatial plan, regional bodies and BBNJ | parent_plans[] bbnj_status |
Governance models supported: state-managed; co-managed by government with communities or fishers (as at Gökova); community-led and locally managed marine areas; delegated management through a public-private partnership; Indigenous-led; transboundary; and high-seas MPAs under BBNJ.
How it maps to Propel: each MPA is a governance object. Its zones and rules are the constitution, its stakeholder bodies are the voting members, and a Hub can act as the jurisdiction for contract and grievance resolution. Because the spec is a schema rather than a document, it can be exported to any jurisdiction, versioned, compared across sites and audited against outcomes.
Typical timelines run three to six years to designation (Seychelles took six, Blue Azores five) and a decade or more to full ecological recovery.
A marine spatial plan allocates the whole sea (fishing, shipping, energy, cables, aquaculture, tourism, protection) so trade-offs are settled once, in one public process, and protection is not re-argued site by site. The proven sequence is: finance commitment → whole-EEZ spatial plan → MPAs designated as its protection zones → trust fund pays for management.
The package provides a planning checklist, guidance on matching planning zones to MPA Guide protection levels, and Gaia map layers so draft plans can be viewed against Ocean Data Initiative data. Protection is most durable where boundaries are uncontested; areas beyond national jurisdiction go through BBNJ.
Tubbataha Reefs in the Philippines ranks first. Every site in the top ten combines strict no-take protection, real enforcement and decades of monitored results.
Each MPA is scored 1 to 5 per criterion; scores are weighted and summed to 100. The criteria follow the MPA Guide, NEOLI and the Green List, so the ranking uses the same scorecard as the rest of this package.
| Criterion | Weight | What earns a 5 |
|---|---|---|
| Ecological outcomes | 25 | Long-term, controlled monitoring shows large biomass or habitat recovery against fished sites |
| Protection level | 15 | Fully protected (no-take) core under the MPA Guide |
| Enforcement and compliance | 15 | Continuous patrols, low poaching, effective penalties |
| Durability | 15 | 20+ years protected, statute-based, no successful rollback |
| Community and economic benefit | 15 | Documented gains in catch, income or tourism for local people |
| Finance | 10 | Recurring, diversified, ring-fenced funding |
| Scale and connectivity | 5 | Large area or part of a connected network |
| Rank | MPA | Eco | Prot | Enf | Dur | Comm | Fin | Scale | Score |
|---|---|---|---|---|---|---|---|---|---|
| 1 | Tubbataha Reefs Natural ParkPhilippines · since 1988 | 5 | 5 | 5 | 5 | 4 | 4 | 4 | 94 |
| 2 | Apo Island Marine ReservePhilippines · since 1982 | 5 | 5 | 5 | 5 | 5 | 3 | 1 | 92 |
| 3 | Great Barrier Reef Marine ParkAustralia · since 1975, rezoned 2004 | 5 | 3 | 4 | 5 | 5 | 5 | 5 | 91 |
| 4 | Leigh (Cape Rodney to Okakari Point)New Zealand · since 1975 | 5 | 5 | 4 | 5 | 5 | 3 | 1 | 89 |
| 5 | Galápagos Marine ReserveEcuador · since 1998, sanctuary 2016 | 5 | 4 | 3 | 4 | 5 | 5 | 5 | 88 |
| 6 | Medes IslandsSpain · since 1983 | 5 | 5 | 4 | 5 | 4 | 4 | 1 | 88 |
| 7 | Cabo Pulmo National ParkMexico · since 1995 | 5 | 5 | 4 | 4 | 5 | 3 | 2 | 87 |
| 8 | Gökova Bay no-take zonesTürkiye · since 2010 | 5 | 5 | 4 | 4 | 5 | 2 | 2 | 85 |
| 9 | Raja Ampat MPA network (incl. Misool)Indonesia · since 2007 | 3 | 3 | 4 | 4 | 5 | 5 | 5 | 78 |
| 10 | Torre GuacetoItaly · since 1991, managed from 2001 | 4 | 3 | 4 | 4 | 5 | 3 | 1 | 75 |
| MPA | Key evidence | Main weakness |
|---|---|---|
| 1. Tubbataha Reefs Natural Park | Year-round rangers with the Navy and Coast Guard; fish biomass rising for a decade; visitor fees fund 74% of the budget. | Depends on a single fee stream; exposed to bleaching. |
| 2. Apo Island Marine Reserve | Fish density and biomass rose 12.2 and 17.3 times over 18 years of protection; catches and living standards rose with it. | Very small reserve. |
| 3. Great Barrier Reef Marine Park | No-take zones went from under 5% to over 33% in 2004; coral trout biomass averaged 81% higher inside, 2006 to 2024. | Two-thirds is not no-take; compliance gaps; heat-driven bleaching. |
| 4. Leigh (Cape Rodney to Okakari Point) | New Zealand's oldest no-take reserve; legal-size snapper 12.8 times more abundant inside than outside (2005). | Fishing at the boundary limits spillover. |
| 5. Galápagos Marine Reserve | The largest reef-fish biomass ever recorded (17.5 t/ha, mostly sharks) at Darwin and Wolf; a debt conversion funds about $17M a year. | Illegal and industrial fishing pressure at the edges. |
| 6. Medes Islands | Several predator species at carrying capacity after two decades; tourism far out-earns fishing. | Limited spillover beyond the no-take core. |
| 7. Cabo Pulmo National Park | Total fish biomass up 463% in 14 years; protection led by the local community. | Small; pressure from tourism development. |
| 8. Gökova Bay no-take zones | Biomass about four times the surrounding areas, stocks up 800%, monk seals breeding; cooperative revenue up 180%. | Grant-dependent finance; recreational angling loophole. |
| 9. Raja Ampat MPA network (incl. Misool) | Misool biomass nearly tripled in a decade; a $23M Blue Abadi trust fund; community income gains. | Fish results across the wider network are mixed. |
| 10. Torre Guaceto | Co-managed with fishers; catch per unit of effort more than twice that outside. | Partial protection over a small area. |
The Azores network (2024), Seychelles (2020) and Samoa (2025). Strong design and finance; revisited once monitoring data mature.
Papahānaumokuākea (US) would have ranked near the top on ecology but lost protection in parts in 2026; Palau's sanctuary faces pressure to reopen.
Kaş–Kekova (Türkiye), Revillagigedo (Mexico), Malpelo (Colombia), Port-Cros (France) and Bonaire.
Method: ecological scores reward controlled, long-term comparisons, which favours older, well-studied sites. Scale carries only 5 points by design, because the evidence shows enforcement and protection level matter more than size. Scores are the Fund's provisional assessment from published evidence and will be reviewed with partners and re-run each year.
French Polynesia's Tainui Atea, about 4.8 million km² since 2025, is more than twice the size of the next largest. Size and strictness are different things: only a few of the ten are mostly fully protected, and two have been partly reopened to fishing.
| Rank | MPA | Area (km²) | Strictly protected | Status |
|---|---|---|---|---|
| 1 | Tainui AteaFrench Polynesia (France) · 2025 | 4,800,000 |
About 1.09 million km² highly or fully protected | Announced at UNOC 2025; strict zones being expanded |
| 2 | Marae MoanaCook Islands · 2017 | 1,981,965 |
Small share | Mostly multiple-use, whole EEZ |
| 3 | French Southern Territories National Nature ReserveFrance · 2016, extended 2022 | 1,600,000 |
Includes about 255,000 km² highly protected around Saint Paul and Amsterdam | Implemented |
| 4 | Ross Sea Region MPAHigh seas (CCAMLR) · 2016 | 1,550,000 |
1.12 million km² fully protected (72%) | Implemented; 35-year term |
| 5 | Papahānaumokuākea Marine National MonumentUnited States · 2006, expanded 2016 | 1,508,846 |
Formerly all no-take | Parts reopened to commercial fishing, 2026 |
| 6 | Natural Park of the Coral SeaNew Caledonia (France) · 2014 | 1,292,967 |
Small share | Criticised as under-enforced |
| 7 | Pacific Islands Heritage Marine National MonumentUnited States · 2009 | 1,270,480 |
Reduced | Commercial fishing reopened 50 to 200 nautical miles out, 2025 |
| 8 | South Georgia and South Sandwich Islands MPAUnited Kingdom · 2012 | 1,240,000 |
Partial no-take | Implemented |
| 9 | Coral Sea Marine ParkAustralia · 2018 | 989,836 |
Partial | Multiple-use zoning |
| 10 | Pitcairn Islands Marine ReserveUnited Kingdom · 2016 | 834,334 |
Fully protected | Implemented; borders the Gambier strict zone |
Total area Strictly protected, where a source states it. Areas follow the World Database on Protected Areas, updated with French Polynesia's 2025 designation; sources differ by a few percent. 89% of global MPA area sits in the 100 largest MPAs (MPAtlas), so this list largely decides the global 30x30 figure. Download the data (JSON).
The strongest cases pair real protection with local benefit and locked-in finance; the cautionary ones lost protection when revenue or politics shifted.
| Case | Model | Result | Lesson |
|---|---|---|---|
| Gökova Bay, Türkiye | Co-managed no-take zones (since 2010), community marine rangers | No-take biomass about 25 times unprotected sites; Akyaka cooperative revenue up more than 180% in three years; monk seals breeding | Community enforcement plus visible income gains wins fishers over |
| Kaş–Kekova, Antalya | Special Environmental Protection Area with no-fishing zones (2020) and rangers (2019) | Early signs of recovery in predatory fish; biomass, seal and temperature monitoring | The COP31 host region's own model, extending the Gökova approach along the coast |
| Azores, Portugal | Regional law (2024) through the Blue Azores partnership | 287,000 km², 30% of the sea; half fully and half highly protected; €10M implementation agreement (2025) | Science plus participation, with partnership finance from day one |
| Seychelles | Debt swap, then a whole-EEZ marine spatial plan, then an independent trust (SeyCCAT) | 30% of the EEZ designated in 2020; trust grants of about US$700k a year | Finance and spatial planning designed together; trust independent of government |
| Galápagos and Hermandad, Ecuador | Largest debt conversion (2023), with a guaranteed blue bond | About $17M a year to 2040, plus a permanent endowment | Guarantees make the economics work; transparency later questioned by civil society |
| Belize | Blue bond (2021) with milestones | About $180M for conservation over 20 years; 30% ocean target | Milestone-linked finance forces delivery |
| Samoa | Nine fully protected MPAs (2025) | Effective protection of 25.6% | Small island states can lead on quality, not just area |
| Palau (cautionary) | 80% no-take sanctuary funded by a visitor fee | Pressure to reopen when COVID cut tourism revenue; public support remains above 70% | A single revenue source is a single point of failure |
| US Pacific monuments (cautionary) | Executive designation under the Antiquities Act | Commercial fishing reopened across several monuments, 2025 to 2026 | What one executive creates, another can undo; statute and ratchet rules matter |
Every indicator needs a baseline before designation and a fixed reporting cycle.
| Domain | Core indicators | Typical data source | Cadence |
|---|---|---|---|
| Ecological | Fish biomass and size inside vs outside; apex predators; habitat cover (seagrass, coral, kelp); key species; invasive species | Visual census, eDNA, baited cameras, satellite habitat maps | Annual |
| Climate | Blue carbon stocks; temperature and heatwave exposure; recovery after bleaching or heat events | Sensors, remote sensing, field cores | Annual; sensors continuous |
| Compliance | Fishing effort inside boundaries; detected incursions; patrol coverage; prosecutions | Vessel tracking, ranger logs, community reports | Monthly |
| Social | Catch and income outside boundaries (spillover); food security; public support; jobs | Landings data, household surveys | Every two years |
| Economic | Tourism revenue and fees; cost per km² managed; share of budget funded | Fee records, budgets | Annual |
| Governance | MPA Guide stage and level; NEOLI score; Green List progress; grievances opened and closed | Propel record, audits | Annual |
| Durability | Years since designation; strength of the legal instrument; years of finance runway; rollback attempts | Legal review, finance model | Annual |
Public scorecard: one page per MPA showing protection level, NEOLI score (0 to 5), biomass inside against outside, compliance rate, years of finance runway and community benefit, rendered on the Ocean Data Initiative and Gaia map layers.
| Roadblock | How it shows up | What has worked |
|---|---|---|
| Fisher opposition | Short-term catch loss; distrust of outsiders | Co-design of zones; fishers as paid rangers; publish spillover data early (Gökova) |
| No recurring finance | Designated but unstaffed: a "paper park" | Close finance before designation; independent trust; diversified revenue |
| Single-source revenue | A revenue shock triggers reopening | Three or more revenue streams; a reserve fund covering two or more years of costs (Palau lesson) |
| Weak legal instrument | Rollback by decree or executive order | Statute; ratchet clause; science review for any downgrade (US monuments lesson) |
| Enforcement gap | Illegal fishing; recreational take of juveniles | Satellite and vessel monitoring, community patrols, real penalties; close recreational loopholes |
| Contested boundaries | Neighbours dispute jurisdiction; protection becomes geopolitics | Stay within undisputed waters; use BBNJ or regional bodies beyond them |
| Data scarcity | No baseline, so no proof of results | Rapid baseline surveys; eDNA and remote sensing; open data from day one |
| Capacity and institutions | Overstretched agencies; knowledge leaves with staff | Co-management agreements; training; documented, exportable governance (Propel) |
| Climate change and invasive species | Warming, heatwaves, Lessepsian species | Climate-smart siting (refugia, connectivity); fisheries and markets for invasive species |
| Equity failures | Benefits captured by tourism operators or elites | Explicit benefit-sharing rules; grievance mechanism; consent |
Most failures come from funding the launch and not the running costs: staff, patrols, monitoring, community programmes and maintenance.
| Mechanism | How it works | Examples |
|---|---|---|
| Debt conversion | Sovereign debt refinanced at lower cost; the savings fund conservation | Galápagos (about $17M a year to 2040); Seychelles, Belize, Barbados, Gabon, Bahamas |
| Project Finance for Permanence | All funders commit at once; money is released only when the full plan and policy are agreed | Large-scale land and sea programmes |
| Conservation trust fund | Independent fund with a mixed board: endowment, sinking or revolving | SeyCCAT (government holds 3 of 9 board seats) |
| Visitor and user fees | Entry, dive, mooring, yacht and cruise fees | Palau's environmental fee; Tubbataha's conservation fee |
| Blue carbon and biodiversity credits | Credits from seagrass, mangrove and other habitats | Emerging; standards still maturing |
| Fisheries and subsidy reform | Redirect harmful subsidies; licence fees support management | Repurposing subsidies from ten countries could fund global ocean protection |
| Public budgets and aid | Domestic budget lines; GEF; development banks | GEF Small Grants funded Gökova's start |
| Corporate and private | Sponsorship, insurance, impact investment | OCF and The Reefline's ReefTile |
| Gaia ocean layer | Sale of virtual ocean squares, with proceeds to protection | OCF's planned mechanism, tied to specific MPAs |
Design rules: close finance before designation; at least three independent revenue streams; a reserve covering two or more years of running costs; transparent fund reporting; and a cost model per km² so funders can see what protection costs.
Vessel tracking, satellite radar and night-light detection of dark vessels, through platforms such as Global Fishing Watch and Skylight.
Fixed and drop sensors, eDNA and baited cameras. The Fund's low-cost sensor work in Goa and the Caribbean plugs in here.
Community marine rangers and fisher reporting, as at Gökova and Kaş–Kekova, with the coast guard for interdiction and penalties that exceed the gain from poaching.
Mooring buoys to end anchor damage on seagrass, boundary markers, patrol boats, sensor replacement, data systems and training, all in the recurring budget.
The ratchet rule: changes that increase protection follow the normal process. Any change that reduces protection needs independent science review, a supermajority and a public record of the evidence.
The BBNJ Agreement entered into force on 17 January 2026, counts 86 Parties, and must hold its first Conference of the Parties before 17 January 2027. Only about 1.5% of the high seas is protected today.
Under the Agreement, Parties propose area-based management tools, including MPAs, based on the best available science and traditional knowledge; a Scientific and Technical Body reviews them; consultations follow; and the Conference of the Parties decides. Still to be settled is who manages high-seas MPAs (the Secretariat, the Conference, a committee or groups of states), along with the budget, the special fund for developing countries and the clearing-house for data.
What this package offers BBNJ:
The Sargasso Sea, with its Bermuda-hosted commission, is a natural first case for the Fund to profile.
The data module curates the open sources an MPA team needs and wires them into the Ocean Data Initiative and the Gaia map.
| Resource | What it holds | Use in the package |
|---|---|---|
| Protected Planet / WDPA | Official global protected-area boundaries | Base boundary layer |
| MPAtlas | MPAs classified by MPA Guide stage and protection level | Effectiveness layer; scorecard benchmark |
| ProtectedSeas | Activity regulations per MPA | Rules layer for the activity matrix |
| Global Fishing Watch | Vessel tracking and fishing effort | Compliance indicator |
| Allen Coral Atlas | Global reef habitat maps and bleaching monitoring | Habitat baseline (tropical) |
| Global Mangrove Watch | Mangrove extent and change | Blue carbon and habitat |
| OBIS and GBIF | Species occurrence records | Biodiversity baseline |
| Copernicus Marine | Ocean physics, biogeochemistry, temperature | Climate exposure |
| EMODnet | European seabed habitats and human activities | Mediterranean and Azores baselines |
| OCF: largest MPAs dataset | The ten largest MPAs with area, strict share and status (JSON) | Scale layer; second ranking |
| OCF Ocean Data Initiative | Open ocean data checked by the QEM, with Gaia map layers | Where all of the above comes together |
Each source keeps its own licence and terms.
Consent protocols, recognition of customary tenure, gender inclusion (such as the Gökova fisherwomen programme) and a standard grievance mechanism.
Siting for climate refugia, larval connectivity between sites and representativeness, so MPAs work as networks rather than islands.
Guidance on fishery closures, community areas and other measures that count toward 30x30 when they deliver outcomes.
Visitor limits, mooring plans and fee design, so tourism funds protection without degrading it.
Many MPAs also protect archaeology and living culture, such as Kekova's submerged ruins, which broadens support and finance.
Framing MPAs as investments in livelihoods and food security, with a stakeholder engagement kit.
An optional guardian or legal proxy for the ecosystem in each MPA's decision body, drawing on rights-of-nature governance models.
AI agents for monitoring triage, ranger reporting, multilingual stakeholder questions and annual scorecards, with identity and audit through Mandala and HubID.
Governments, communities and partners working on marine protection can use this plan and help build it out.