Climate Migration Planning

Established Climate migration planning is the practice of moving people before a hazard arrives rather than after it. It is a small practice: the best global census of the wider category counts twenty-seven cases of managed retreat from natural-hazard risk, and the subset that was climate-attributed and documented well enough to say whether the movers ended up better off is a handful of Pacific villages. Established It sits inside a far larger phenomenon, disaster displacement, which is overwhelmingly internal, short-duration, and counted in a way most people quoting the counts do not understand.

Established Two of the most-repeated claims in this subject are inversions of their own sources. The figure that several hundred million will be displaced by sea-level rise by 2050 comes from a paper whose authors state that their results “should therefore not be taken as projected impacts,” and which does not model the coastal defences the same authors say protect “tens or hundreds of millions” of those counted. Speculative The Teitiota communication to the UN Human Rights Committee, universally reported as the decision that recognised climate refugees, is a case the applicant lost. Established And even a win would not have produced protection, because the doctrine at issue bars removal without conferring any status at all.

Established The gap in this subject is not knowledge. It is instrument design. There is no legal category of climate refugee, no rule allocating responsibility for admitting people who cannot go home, and no agency in any wealthy country with authority to fund and execute a community relocation end to end. Frontier Each is, in principle, draftable. Established None has been drafted.

Established A note on labelling. The research behind this brief ran under a retrieval environment that blocked every primary legal instrument, every UN agency site and every displacement dataset: no treaty text was read, and no article number is printed that this brief cannot source. Speculative Where a widely reported fact could not be verified it is flagged speculative and said to be unverified — the flag describes what this brief can stand behind, not what the world is probably like.

Established Start with what the displacement data measures, because almost every downstream argument in this field misuses it. The Internal Displacement Monitoring Centre's annual reporting counts new displacements: movement events in a calendar year, not people. Established A person evacuated twice is counted twice, and a household that leaves for four days ahead of a cyclone and returns counts on identical footing with one that never goes back — which makes the totals a measure of the reach of evacuation systems as much as of harm. Speculative The figures in general circulation were not obtainable in this pass and none is stated here. Established A displacement number without its year and its counting rule is unusable, and most of the ones in circulation carry neither.

Speculative Four qualitative statements about that data are robust enough to build on, and none was verifiable in this pass against a primary dataset. New displacements associated with disasters have substantially exceeded those associated with conflict in most recent years, commonly by two to three times; the overwhelming majority is internal; the majority is weather-related, dominated by floods and storms; and most is short-duration evacuation. Established Those four together dismantle the border framing: if the phenomenon is mostly internal, mostly weather and mostly temporary, an international protection instrument addresses a small tail of it and the effective policy surface is domestic.

Established Exposure counts are the other data object, and they are not displacement projections. Kulp and Strauss, applying a neural-network correction to satellite elevation data, found roughly 110 million people occupying land below the current high tide line and 250 million below current annual flood levels. Established Under a moderate pathway with a stable Antarctica, 150 million occupy land below the projected 2050 high tide line; on a high pathway with Antarctic instability, 340 million are threatened by mid-century flooding. Established More than seventy per cent of the exposed population is in eight Asian countries. Every one of those is a count of people standing on low ground, computed by combining future water levels with contemporary population densities. Established The authors state the results should not be taken as projected impacts, note country-scale elevation error reaching plus or minus forty-nine per cent, and say that existing defences protecting “tens or hundreds of millions” of the people counted are not modelled at all. Established Exposure is an upper bound the authors flagged as an upper bound.

Established Now the legal position, stated only as far as it can be sourced. The 1951 Refugee Convention, as universally summarised, defines a refugee by a well-founded fear of persecution on grounds of race, religion, nationality, membership of a particular social group, or political opinion, held by a person outside their country of nationality; the 1967 Protocol removed the original temporal and geographic limits. Established This brief did not read the text and prints no article number for it. Climate change is not among the grounds, a rising sea is not a persecutor, and the status additionally requires crossing an international border. Established “Climate refugee” is therefore not a legal status but a rhetorical one, and the distinction determines whether anything is owed to anyone.

Speculative The nearest thing to a binding provision is regional and internal-only. The African Union's Kampala Convention is generally described as obliging States Parties to protect and assist people internally displaced by natural or human-made disasters including climate change; if that description is accurate, it is the only binding treaty provision anywhere expressly covering climate displacement, and it does not cross borders. Speculative This brief could not obtain the instrument and prints neither its article number nor its wording. Around it sits a dense layer of soft law — the Nansen Initiative Protection Agenda, the Global Compact for Migration, and a UNFCCC track from the Cancun Adaptation Framework through the Warsaw International Mechanism to a Task Force on Displacement — none of which creates status. Established A thick layer of guidance on a hard-law vacuum: useful for how to run a relocation, irrelevant to whether anyone must admit anybody.

Established The doctrinal point that structures everything else: non-refoulement is a negative obligation. It stops a State returning a person to a place where they face a proscribed risk, and creates no right to enter, to work, to be housed, or to acquire status — the trajectory from non-refoulement to residence and citizenship is a specialist literature precisely because it is not automatic. Established A person who cannot lawfully be removed and has no status is in limbo, sometimes for years. Winning a non-refoulement case is the beginning of the problem, not the end of it, and this is almost never present in the advocacy framing.

Speculative Ioane Teitiota v New Zealand is cited for the opposite of what it held. Established The citation is CCPR/C/127/D/2728/2016 before the UN Human Rights Committee. Teitiota, an I-Kiribati national whose New Zealand asylum claim had been rejected, invoked the right to life, arguing that return exposed him to sea-level rise producing scarcity of habitable space, environmental degradation and saltwater contamination of the fresh water. Speculative The Committee is consistently reported to have found no violation by New Zealand, accepting the domestic assessment that the risk was not sufficiently imminent, with a timeframe on the order of ten to fifteen years said to leave room for intervening protective action; there were individual dissents. Speculative What it accepted prospectively is that climate effects may, in a future case, engage non-refoulement obligations. Established Teitiota lost. This brief could not read the views and flags the holding unverified, while noting that every secondary account agrees. Both circulating dates are correct: the views were adopted on 24 October 2019 at the Committee's 127th session and released on 7 January 2020.

Speculative The more consequential decision may be the one nobody quotes. Billy et al. v Australia, the Torres Strait Eight, decided in 2022, is commonly reported as finding Australia in violation of rights concerning privacy, family and home and the rights of minorities to enjoy their culture, while finding no violation of the right to life; this brief could not verify it and prints no case number and no article number. Frontier If the pattern holds it matters more than Teitiota, because it locates the obligation in rights that bite before life is threatened — where relocation decisions are actually taken. A parallel and possibly wider route runs through the European Convention on Human Rights.

Established Turn to the measured record of planned relocation, which is thinner and better than its reputation. The two Fijian cases documented by Piggott-McKellar and colleagues are the best-evidenced relocation outcomes in the literature, and they are best-evidenced because the authors counted the failures. Established Vunidogoloa, on Vanua Levu, moved 153 people about two kilometres inland: planned from 2012, completed in January and February 2014, driven by slow-onset tidal inundation, erosion and saltwater intrusion. The village received houses, solar power, water tanks, flush toilets, fish ponds, pineapple plantations and cattle, plus better road access to schools and clinics. Established It also received 5,000 pineapple tops against 48,000 promised and four fish ponds against eight; the houses lacked their promised kitchens; drainage was left incomplete, with unused pipes lying in the village at the time of fieldwork. Established Housing was delivered; livelihoods were not.

Established Denimanu is the most valuable case in the literature because it is a relocation that increased risk. On Yadua Island, about 170 people; nineteen households, roughly half the village, moved about 500 metres upslope after Cyclone Evan struck in December 2012 — rapid-onset, in direct contrast with Vunidogoloa. Established Houses were completed in mid-to-late 2013; they leak in heavy rain; there are two septic tanks for nineteen houses; drainage is inadequate and causing soil erosion. And the receiving site is landslide-prone: the primary school was destroyed by a landslide, and residents expect to be relocated again. Established Two cases in one country and one paper, failing differently, are the closest thing here to a controlled comparison of anticipatory against reactive movement.

Established National context and the global count. Four Fijian communities had initiated or completed a move at the time of that study; the Fijian Government recognises more than eighty further communities as needing future relocation; and Fiji was the first country in the world to publish national Relocation Guidelines, in 2018. Established Globally, Hino, Field and Mach reviewed twenty-seven cases of managed retreat as a response to natural hazard risk, totalling approximately 1.3 million people resettled, organised on two axes — who benefits, and who initiates. Established Two caveats travel with that headcount and are almost always dropped: the cases are natural hazard retreat, not climate-attributed retreat, and the total is almost certainly dominated by a few very large state-directed programmes. Frontier The durable contribution is the typology, not the number.

Speculative The wealthy-country cases are named here and their figures are not asserted. Newtok to Mertarvik, a Yup'ik village in Alaska of roughly 350 residents, has been relocating across the Ninglick River because of erosion and permafrost thaw, with first families reported to have moved in October 2019 and costs widely given as upwards of US$130 million — none of it confirmable here. Speculative Isle de Jean Charles, Louisiana, a Biloxi-Chitimacha-Choctaw community, was resettled under a federal disaster-resilience competition award announced in January 2016, widely reported at US$48.3 million and described as the first federally funded climate relocation in the United States; amount, household count and outcome are unverified here. Established What is instructive does not depend on the number: a fully funded relocation with a willing community took the better part of a decade, was contested over state control and eligibility, and fractured the community politically. Speculative Kiribati under President Anote Tong pursued “Migration with Dignity,” a skills-and-pathways policy, and separately bought roughly 2,200 hectares at Natoavatu Estate on Vanua Levu in 2014 — two things the world has merged into one. Speculative The Carteret Islands move to Bougainville from around 2005, origin of most “first climate refugees” reporting, was slow, under-funded and community-led.

Frontier The category of “community” is breaking down as an operational unit. Relocation policy assumes a bounded group that can be consulted, can consent, and can move together. Established Denimanu moved nineteen of roughly thirty-eight households. Frontier Half a village is not a community giving consent; it is a village splitting, with the movers and stayers left in different risk positions and different tenure positions. Bower's work on complicating community engagement is the reference point for this, and the practical question it raises has no answer yet: what is the consent mechanism when the unit of decision is smaller than the unit of policy?

Frontier Receiving-site hazard assessment is the least institutionalised step in the entire practice, and it is cheap. Established Denimanu is the demonstration: a move away from cyclone-exposed coast into a landslide zone that then took the primary school. Frontier Nobody has published a standard requiring a multi-hazard assessment of the receiving site over the design life of the settlement, covering hazards different from the one being fled. On any reasonable benefit-to-cost reckoning this is the highest-return recommendation available in the subject, and it is unglamorous enough that it does not get made.

Frontier Anticipatory and reactive relocation are turning out to be different problems with opposite failure modes. Established Vunidogoloa was planned over years and under-delivered on livelihoods. Denimanu was executed fast in a post-cyclone funding window and under-delivered on siting and sanitation. Frontier The generalisation the field is converging on — post-disaster relocation is easier to fund and worse to execute — is supported by exactly two well-documented cases in one country, which is not enough. Whether the anticipatory route is better on outcomes, rather than merely better in principle, is an open empirical question with a tiny sample.

Frontier Livelihood provision is emerging as the measurable failure mode, and it is measurable in precisely one place. Established Five thousand of forty-eight thousand pineapple tops; four of eight fish ponds. Frontier That is the only relocation in the literature whose promised inputs were counted against delivered inputs. It should be the template for how relocation outcomes are audited, and the reason it is not is structural rather than intellectual: housing is a capital line item that a ministry can complete and photograph, and livelihood restoration is a recurrent one that nobody's budget cycle wants.

Frontier National relocation guidelines are a new instrument whose effect is untested. Established Fiji published the first set in 2018. Frontier Whether guidelines change outcomes — as against changing documentation — is the obvious evaluation question and has not been answered, in part because Fiji's post-2018 relocations are too few and too recent for a before-and-after comparison to carry weight.

Frontier The historical-continuity critique is the sharpest thing said about the field by anyone inside it. McAdam's framing of relocation as “the perennial solution to danger zones” places climate relocation in a lineage running through colonial resettlement, and the record of that lineage is overwhelmingly one of failure — of movements justified by hazard that turned out to serve other purposes, and of communities that lost more in the move than the hazard would have taken. Frontier That is not an argument against relocation. It is an argument that the burden of proof sits with the mover, and current practice does not place it there.

Frontier The literature is diversifying out of the Pacific, slowly. Urban flood-risk relocation in Kampala, resettlement in Malawi, and the Ghoramara Island case in India are entering a corpus that has been atoll-dominated and Anglophone. Frontier This matters more than it sounds, because the central empirical claim of the subject — that climate mobility is mostly internal and mostly in the Global South — is being tested against evidence drawn from the places least representative of it. Frontier A justice-oriented sub-field has also consolidated, which is a genuine intellectual development rather than a rebranding: it has produced the criteria against which relocations can be judged unjust, which the operational guidance does not supply. Frontier What it has not produced is a way to apply those criteria before a move rather than after one, which is the difference between a research field and a practice.

Established The workback plan to a working protection status for cross-border disaster displacement has ten links, and the field routinely stops at link three. Speculative Link 1, measuring the cross-border fraction: weak, because internal displacement is counted and cross-border is not. Frontier Link 2, an administrable trigger definition that does not require individual-level climate attribution: not drafted anywhere. Speculative Link 3, non-refoulement extended in principle to climate risk: achieved, if Teitiota holds as reported. Frontier Link 4, lower-threshold rights engaged: emerging, via the Torres Strait pattern and the European Convention route. Established Link 5, status conferred rather than removal barred — residence, work rights, duration, family unity, a path to settlement, in statute: not achieved anywhere on earth. Link 6, a rule allocating responsibility among States for who admits: never solved, for any caseload, in the history of refugee protection. Frontier Links 7 through 10 — funding distinct from adaptation finance, a regional instrument, domestic legislation converting case law into a category, and operation at scale — are all downstream of 5 and 6.

Established Link 6 binds, with link 5 as the near-term practical blocker, and the asymmetry between them is the finding. Link 5 can be solved unilaterally: any State that wishes to may legislate a complementary-protection category tomorrow, and that is historically how complementary protection was built — national legislatures converting a stream of successful non-removal claims into a status. Established Link 6 cannot be solved by anyone acting alone, and no protection regime has ever solved it. Frontier The tractable path is national legislation; the global instrument is not the tractable path, which means the thing to watch is domestic immigration statutes, not COP texts.

Established The parallel workback, to relocation that leaves people better off, has a different binding link. Frontier Receiving-site multi-hazard assessment over the settlement's design life: cheap, rarely done, Denimanu is the counterexample. Sub-community consent mechanism: weak. Established Housing delivered: generally achieved. Services delivered to specification: failed at Denimanu, at two septic tanks for nineteen houses. Livelihoods restored, budgeted and audited: failed at Vunidogoloa, and this is the link that binds. Frontier Land tenure secured at the receiving site: underexamined, and in Fiji it interacts with customary tenure in ways nobody has modelled. Outcomes monitored for ten years or more: almost never done, anywhere. The two workbacks bind at different links for the same underlying reason: each turns on a recurrent obligation that somebody must own permanently — a State's duty to admit, or a ministry's duty to keep a plantation alive after the ribbon is cut — and recurrent obligations are what neither international law nor a capital budget is built to carry.

Frontier This subject waits on four results that other fields would have to produce, and none of them is anyone's priority. The first is attribution at the household level: a method that can say whether a given move was climate-influenced without leading the respondent, given that migrants themselves attribute their decisions to land, work, education and family. Frontier Without it, climate migration cannot be counted, and what cannot be counted cannot be funded.

Established The second is a global inventory of coastal defences. Kulp and Strauss state that defences protecting tens or hundreds of millions of the people in their exposure counts are not modelled, because no such inventory exists. Frontier Until it does, every exposure-to-displacement conversion is unbounded on the side that matters most.

Frontier The third is longitudinal household data linking climate stress to mobility outcomes by wealth quintile, which is the only way to settle whether severe stress produces movement or immobility, and which requires panel surveys running a decade or more in exactly the places where survey infrastructure is weakest. Frontier The fourth is mundane and immediately actionable: the case composition behind the 1.3-million retreat figure, which would tell the field whether managed retreat has a track record or a handful of state-directed outliers with a large denominator. Established That one is a supplementary-material problem, not a research programme.

Frontier There is also a dependency running the other way, and it is the reason this brief is organised around instruments. Nothing in the physical science of sea-level rise, cyclone intensity or heat exposure is now the limiting input to a decision about whether to move a village; the hazard signal has been sufficient for a decade. Frontier What is missing is legal drafting, statutory authority and a budget line, none of which any research programme produces. Speculative A brief that treats this subject as awaiting better projections has mistaken which link binds.

Frontier The measurements that would move this field are unusually cheap, which is a signal about why they are not being made. One: audit delivery against contract. For every planned relocation, publish the promised schedule of houses, services and livelihood inputs and the delivered schedule against it, on the Vunidogoloa model — pineapple tops promised, pineapple tops delivered. Frontier This requires no new method and produces the only outcome data the field has ever generated.

Frontier Two: ten-year outcome monitoring on at least one cohort. Income, health, school attendance, tenure security and hazard exposure at the receiving site, measured at intervals, with a comparison group of households that did not move. Frontier Nothing of the kind exists for any climate relocation anywhere, which is why the question “does relocation help?” has no evidence-based answer.

Frontier Three: a receiving-site multi-hazard protocol, trialled prospectively. Assess candidate sites for landslide, flood, seismic, fire and water-supply hazard over the settlement design life, publish the assessment before the move, and revisit it after. Established Denimanu is the counterfactual that shows what its absence costs.

Frontier Four: a non-leading survey instrument for migration drivers, validated against a sample where the environmental signal is independently known — a post-cyclone district, say — to establish how much climate influence a conventional survey misses. Frontier Five: a natural experiment on guidelines. Fiji's 2018 guidelines create a before-and-after boundary; comparing relocations either side of it on delivery-against-contract is the cheapest available test of whether procedural instruments change outcomes, and it is a test that any government publishing guidelines ought to want run. Frontier Six: measure the stayers. Every documented relocation has a group that did not move — nineteen households of thirty-eight at Denimanu — and nobody has followed them, so the field has no evidence at all about what a partial relocation does to the people left behind. Speculative Seven: read the primary sources. This cluster produced two confirmed cases of a source being cited for something it does not say, one documented count drift, and one retrieval tool that served an entirely unrelated paper under a correctly addressed identifier. Frontier A systematic citation audit of the fifty most-cited claims in climate-migration policy documents is a weekend of work and would probably be the highest-value output in the subject.

Established A relocation is a small settlement build, and the engineering failures in the record are elementary. Houses at Denimanu leak in heavy rain, which rusts fittings and jams doors. Drainage was incomplete at Vunidogoloa and inadequate at Denimanu, where it is causing soil erosion on the new slope. Established Sanitation was under-provided at a ratio of two septic tanks to nineteen houses. Kitchens promised at Vunidogoloa were not built. Frontier None of these is a hard engineering problem; all of them are specification-and-supervision problems, which is a different failure and needs a different fix.

Frontier The site works are where the real engineering sits, and they are the part that gets skipped. Slope stability at an upslope receiving site, ground investigation adequate to identify landslide susceptibility, drainage designed for the actual rainfall regime rather than the coastal one the community left, and a water supply that does not depend on the aquifer that is salinising. Frontier The move at Denimanu was 500 metres. Five hundred metres is enough to change the governing hazard entirely, and the practice does not treat it that way.

Frontier The recurrent-cost engineering is the productive frontier. Established Solar power, water tanks and flush toilets were delivered in Fiji; what was not delivered was the productive base — the plantations and ponds that make the settlement an economy rather than a housing estate. Frontier Designing relocations around productive assets with a maintenance regime and a named owner, rather than around dwelling units, is the specification change with the largest expected effect and no constituency behind it. The same applies to the services: a septic tank is a capital item with a desludging schedule, and the schedule is the part that determines whether the settlement is habitable in year five. Speculative A relocation budget that does not carry a ten-year operations and maintenance line is not a relocation budget; it is a construction budget with a village attached.

Established Coastal protection economics is the adjacent field that most directly determines how much relocation there will be. Vousdoukas and colleagues found benefit-cost ratios for raising European dykes averaging 8.3 to 14.9 depending on scenario, with country values from 1.6 to 34.3 — and, in the same analysis, that protection is economically justified along only 23.7 to 32.1 per cent of Europe's coastline, that minority carrying at least eighty-three per cent of the available damage reduction. Established Read correctly, that is a quantitative statement that most coastline will not be defended, arrived at from inside a defend-oriented analysis. Frontier Retreat is not the alternative to that finding; it is its complement, and the two literatures barely speak.

Frontier Domestic social protection is the adjacent system that would actually deliver, if the phenomenon is mostly internal. Cash transfer infrastructure, urban land and housing policy, informal-settlement upgrading, and rural livelihood programmes reach internal climate migrants without anyone having to call them climate migrants. Frontier The instrument is domestic and unglamorous, and it is the one that would move numbers.

Frontier Anticipatory humanitarian action — forecast-triggered cash and evacuation, disbursed before impact rather than after — is the closest operational analogue to anticipatory relocation, works on days-to-weeks horizons, and has an evidence base that planned relocation lacks entirely. Speculative Voluntary buyout programmes for flood-prone property in the United States are the largest quantitative record of retreat inside a democracy; the pack behind this brief could not obtain the studies and states no figures from them, but that is where the numbers on democratic retreat live. Frontier Labour mobility schemes — seasonal work visas, regional free-movement arrangements, skills pathways — are how Pacific migration actually happens, and they are administered as migration policy rather than climate policy, which is exactly the reframing the subject needs. Speculative They also carry a quiet advantage over any protection status: a work visa confers the thing non-refoulement withholds, which is the right to earn.

Speculative The binding constraint in a wealthy country is authority, not money. Bronen and Chapin's argument from the Alaskan cases is that no single United States federal agency has authority to fund and execute a community relocation, so funding must be assembled from programmes designed for other purposes — erosion control, housing, disaster recovery, infrastructure — each with its own eligibility rules, none of which contemplates moving a village. Speculative The research pass could not obtain the paper and states none of its figures. Frontier The structural claim survives that gap, because it is corroborated by the shape of every documented case: relocations proceed at the speed of the slowest funding stream, and no agency owns the outcome.

Frontier A competition is a terrible instrument for a right. Speculative The Isle de Jean Charles resettlement was won in a federal competitive grant round. Frontier That means communities in identical circumstances that did not apply, or applied and lost, received nothing — and it means the winning community's terms were set by the grant's design rather than by their own. Scale requires a funded, non-discretionary entitlement with a statutory basis: a threshold of hazard that triggers a right to relocation assistance, administered by an agency with the authority to buy land, build, and transfer tenure. Frontier Essentially nowhere has one. Established That is the institutional gap in one sentence.

Established The international architecture is procedurally rich and substantively empty. Speculative The Nansen Initiative and its successor Platform on Disaster Displacement produced an endorsed Protection Agenda that binds nobody; the Global Compact on Migration names climate among drivers and binds nobody; the UNFCCC track has an adaptation framework, a loss-and-damage mechanism and a task force on displacement, and none of them confers status. Frontier What that architecture can do is normalise displacement as a legitimate object of national adaptation planning and open a funding channel — which is why the useful institutional ask is not a new treaty but a line in national adaptation plans and a loss-and-damage window that pays for moving people before the loss rather than compensating it afterwards. Speculative That ask is small enough to be granted and specific enough to be audited, which is more than can be said for anything currently being negotiated.

Established A protection regime designed around movers systematically misses the worst-off. Frontier Migration requires resources: money for transport, a contact at the destination, documents, a saleable asset. The poorest households have none of these, so the severest climate stress can produce immobility rather than displacement — the trapped-populations finding. Frontier If that holds, every instrument in this subject is aimed at the people who could get out, and the people who could not are invisible to all of them.

Established The same distributional failure arrives from a second, unrelated direction. In the European coastal-protection analysis, optimal adaptation reduces expected damages by 96 to 98 per cent but reduces the number of people flooded by only 59 to 66 per cent. Established Protection optimised on asset value protects assets far better than it protects people, because that is what the objective function says. Frontier Asset-weighted adaptation under-protects people in the defence frame and in the migration frame alike, by different mechanisms, and that convergence is the strongest quantitative ethical claim available in the whole subject — and it comes from mainstream economic analysis rather than from a critique of one.

Frontier Consent is the unresolved procedural question. Established Nineteen of thirty-eight households moved at Denimanu. Frontier There is no accepted mechanism for consent when a village splits, no rule about what the stayers are owed, and no doctrine on whether a relocation that half a community refuses is voluntary. The colonial record of resettlement is the reason this matters: the historical lineage is full of moves that were procedurally consented and substantively coerced, and the field has not built the safeguards that would distinguish its own practice from that record. Frontier There is a further asymmetry inside consent itself — the people asked to agree are current residents bearing identifiable, immediate costs, while the beneficiaries are future residents and future taxpayers who cannot be asked at all. Speculative That is the worst political-economy profile any public decision can have, and it is why retreat loses referenda that protection wins.

Speculative The deepest question in the subject is whether a State survives the loss of its habitable territory. Whether a State whose land becomes uninhabitable retains legal personality, its maritime baselines, and its exclusive economic zone is open and unsettled. Frontier For an atoll State the EEZ question is not symbolic: fisheries licence revenue drawn from that zone can be the largest single item in the national budget, and it is tied to baselines drawn from land that may not be there. Fixing baselines against sea-level rise, by State practice or authoritative determination, would preserve the revenue and with it the possibility of a nation that persists as an institution and a fund while its population lives elsewhere.

Speculative “Migration with dignity” is a proposition about what a country is. Kiribati's policy under Tong was to build skills and migration pathways so that I-Kiribati could move as qualified migrants rather than as destitute claimants. Frontier Read one way that is the most realistic climate policy any small State has adopted. Handwave Read another way it is a managed dissolution, and the deliberate refusal of the refugee framing that accompanies it — because the label implies the State has failed and prejudices its sovereignty and its claims — means that a nation's most likely end is a process nobody involved is permitted to name.

Handwave Push the century out and the exotic version of this subject is a deterritorialised State. A government with a treasury, a citizenry, a diplomatic personality and a maritime zone, exercising no jurisdiction over any land — the sovereign as trust fund and passport issuer. Speculative Nothing in international law obviously forbids it and nothing in international law provides for it. Handwave It would be the first genuinely new kind of polity since the nation-state, and it would arrive not through anyone's design but because the alternative was to say out loud that a country had ended.

What these horizons track is instrument design, not climate. The physical signal is already sufficient to justify moving people; what is missing is the legal and financial machinery to do it.

• 10 yr: Frontier Expect more national relocation guidelines on the Fijian model, and the first evaluations of whether they change outcomes. Expect at least one wealthy country to legislate a hazard-triggered relocation entitlement at sub-national level. Speculative Expect the lower-threshold-rights route — privacy, family, home, culture — rather than the right to life to become the main litigation channel. Established Do not expect a protection status: nothing is in draft anywhere. Frontier Expect the first serious attempt to audit a relocation against its own delivery schedule, because it is cheap and someone will notice.
• 25 yr: Speculative A regional complementary-protection instrument covering disaster displacement is plausible, most likely following the African and Latin American regional models rather than a global one. Loss-and-damage finance reaching anticipatory relocation in more than pilot volumes is plausible on this horizon and is the single most consequential thing that could happen. Frontier The trapped-populations finding, if confirmed by longitudinal data, reorients the field away from movement entirely.
• 50 yr: Speculative Relocation at metropolitan scale in a democracy — a city district retreating on a schedule, with compensation as an entitlement rather than a grant — is the threshold test, and it has never been passed. Expect it first where a statutory delta-fund style mechanism already removes the decision from annual politics. Speculative Expect the first authoritative determination on whether maritime baselines survive the land they were drawn from, since the fisheries revenue attached to that question forces it long before the last island goes. Frontier Expect, too, that the phrase ‘managed retreat’ will have been abandoned for something less honest, because the word retreat has never survived contact with a planning inquiry.
• 100 / 250+ yr: Handwave A State that persists with no habitable territory, holding a maritime zone and a treasury on behalf of a dispersed citizenry. Or the opposite, and more likely: no new categories at all, the whole phenomenon absorbed into ordinary internal migration and ordinary urbanisation, uncounted, unfunded and unnamed, which is what has happened to every previous episode of environmentally driven movement in human history.

• Depends on This brief depends on results the field has not been asked to produce: a household-level climate-attribution method for migration decisions, a global inventory of existing coastal defences, longitudinal data linking climate stress to mobility outcomes by wealth quintile, and a published case composition behind the one global retreat headcount anybody quotes. None of these is expensive and none is anyone's priority, because each would constrain a number that is currently useful precisely for being unconstrained. This brief also depends, unusually for a technology tree, on primary-source verification: two of the most-cited claims in the subject are inversions of their own sources, and the correction requires nothing more than that somebody open the papers.
• Requires (not on this map) Three constraints sit outside anything a research programme can deliver. The first is a protection status for cross-border disaster displacement — not a bar on removal, which non-refoulement already supplies in principle, but residence, work rights, duration and family unity in statute, which exists nowhere. The second is loss-and-damage finance that can be spent on moving people before the loss rather than compensating it afterwards; anticipatory relocation currently falls between adaptation finance and disaster response and is funded by neither. The third is the integration of displacement into national adaptation plans, which is the mechanism by which an internal phenomenon acquires a domestic budget line — and since climate mobility is overwhelmingly internal, this is the constraint that governs the largest number of people.
• Enables A working protection status would enable a great deal downstream: an insurable and financeable category of pre-emptive move; a place for displacement in national adaptation planning rather than only in disaster response; a basis for allocating loss-and-damage finance to movement rather than to reconstruction alone; and an end to the position in which a person who cannot lawfully be removed also cannot lawfully work. Relocation practice that delivered livelihoods and services on the same audited footing as housing would enable something the field has never had — an honest answer to whether relocation leaves people better off — and that answer is the precondition for anyone funding it at scale.
• Adjacent Adjacent to coastal defence economics, which determines how much coastline will be defended and therefore how much retreat there must be, and which finds protection economically justified along only a quarter to a third of Europe's coast; to anticipatory humanitarian action, which already disburses cash and moves people ahead of impact on horizons of days and has the evidence base that planned relocation lacks; to domestic social protection and urban planning, which is the delivery channel for internal movement; and to labour-mobility policy, which is how most climate-influenced movement across borders actually happens and which nobody counts as climate policy at all.

Established “Three hundred million people will be displaced by 2050.” This is a corruption of Kulp and Strauss 2019, and the clearest case in this cluster of a founding paper explicitly denying the result it is cited for. Established Their figures count people living on land below a projected water level: 150 million below the projected 2050 high tide line on a moderate pathway, 340 million threatened by mid-century flooding on a high pathway with Antarctic instability. Established The authors state that the analysis combines future water levels with contemporary population densities and “should therefore not be taken as projected impacts.” Three corruptions have been layered on it: exposure turned into displacement; a scenario's number quoted without its scenario; and, worst, the omission of the authors' statement that existing defences protecting “tens or hundreds of millions” are not modelled at all. Established There is no peer-reviewed global projection of sea-level-rise displacement. Not a contested one, not a disputed one — none.

Speculative “Teitiota means climate refugees are protected” / “the UN ruled they cannot be sent back.” Teitiota lost. The Committee found no violation by New Zealand on the facts, accepting that the risk was not sufficiently imminent; what it accepted was the prospective proposition that climate effects may engage non-refoulement in some future case. Established Reporting inverted a loss into a win at scale, and the inversion now sits in secondary academic literature as well as press coverage. The claim is wrong twice over, because even a win would not have produced protection: non-refoulement bars removal and confers no status. Established The date confusion resolves cleanly: views adopted 24 October 2019, released 7 January 2020, both correct about different events. This brief did not read the views and labels the holding unverified; it states the correction anyway, because every secondary source agrees.

Established “Climate refugee” is a legal category. It is not, anywhere. The 1951 Convention's grounds are race, religion, nationality, membership of a particular social group and political opinion; climate is not among them, a rising sea is not a persecutor, and the status requires a border crossing. Speculative The nearest binding provision appears to be the Kampala Convention's coverage of internal disaster displacement in Africa. Established This brief read no primary legal instrument and prints no article number from any of them. Frontier The usable terms are “climate-displaced person” and “disaster displacement,” said together with the fact that the status does not exist.

Frontier “It should be a legal category, though.” A serious position seriously argued: the term names a real harm and generates pressure. Established Against it: it misdirects effort to the small cross-border fraction; affected States reject the framing because it implies state failure and prejudices sovereignty; reopening the 1951 Convention is regarded across the field as suicidal, since renegotiation would narrow rather than widen; and it implies a persecutor, which breaks the doctrine, since emitting States are not what the person is fleeing. Established And a new instrument still hits the allocation problem: recognition without a rule about who admits produces no protection.

Established “Managed retreat has already moved 1.3 million people for climate reasons.” Hino, Field and Mach's twenty-seven cases and roughly 1.3 million people are natural hazard retreat, and the total is very likely dominated by a few large state-directed programmes. Frontier The headcount is the part that travels and the part that means least.

Established “Relocation moves people out of harm's way.” Denimanu moved 500 metres upslope into a landslide-prone site, lost its primary school to a landslide, and expects to move again. Frontier Relocation is a hazard-substitution decision taken as though it were a hazard-elimination decision.

Speculative “The main obstacle is money.” In Alaska the obstacle is that no agency has authority to fund and execute a move. Established In Fiji, housing was funded and delivered while livelihoods were not. Frontier Authority and recurrent-cost budgeting, not capital, are the constraints. Speculative The circulating Alaskan cost per head, if right, is high but not obviously beyond a wealthy state — a figure this brief could not verify and does not assert.

Speculative “Kiribati bought land in Fiji to move its population there.” The 2014 Natoavatu Estate purchase was framed by the Kiribati government substantially in terms of food security and investment, with President Tong resisting the “buying a new country” reading; “Migration with Dignity” was a separate skills-and-pathways policy. Frontier Two things merged into one story more dramatic than either.

Established “Climate migration will be a wave of people arriving at rich-country borders.” Speculative It is overwhelmingly internal, largely short-duration, and substantially invisible inside ordinary rural-to-urban migration; and severe stress can reduce mobility for the poorest. Frontier The border framing misallocates attention in the direction that makes the phenomenon look threatening rather than the direction that would help anyone. The corollary is the most useful reframing available: the effective instrument for most climate migration is domestic social protection and urban planning, not international refugee law.

Speculative “Two hundred million” and “one billion climate migrants by 2050.” The 200-million figure traces to work from the 1990s and 2000s described, including by its author, as an extrapolation rather than a projection; the billion figure comes from a threat-register publication and is an exposure count. Established Neither should be used, and there is nothing better to replace them with.

Established And one about the machinery of citation itself. The retrieval pass behind this brief recorded a bibliographic tool serving, twice, the full text of an unrelated paper — a mouse digit-regeneration study — under a correctly addressed identifier for a coastal-defence paper, with no error signalled. Frontier An agent that had not checked the returned title would have attributed the wrong content to the right citation. Established This subject is governed by assessment reports assembled from secondary citations at scale, and has already produced two confirmed inversions. The base rate is high enough that “everyone knows X” should be read as a prompt to open the paper, not as permission to skip it.