Planetary Stewardship

Established Planetary stewardship is the proposition that the Earth system can be managed as a single object, against quantified limits, by institutions that do not yet exist. Its dominant expression is the planetary boundaries framework: a set of nine named Earth-system processes, each given a control variable and a numerical boundary, with the region inside all nine called a safe operating space. Established The framework is now thirty-odd years of ambition compressed into a decade and a half of publication, and it has been adopted into policy language, corporate reporting and school curricula faster than its own numbers have stabilised.

Established The single most important thing this brief has to say is that the same named boundary carries different control variables in different papers by overlapping author teams. Comparing two Nature-family papers directly: climate is radiative forcing at 1.0 W m^-2 in one and global warming at 1.0–1.5 °C in the other; biosphere integrity is Mean Species Abundance at 90% in one and intact natural ecosystem area at 50–60% of global land in the other; phosphorus is an 11 Tg P yr^-1 surplus in one and a 4.5–9.0 Tg yr^-1 cropland surplus in the other. Established Johan Rockström is an author on both. Established A count of transgressed boundaries is therefore only as stable as the definitions being counted, which is what makes the famous headline — six of nine — a claim that needs a framework attached before it means anything at all.

Established The second load-bearing number is the nitrogen boundary's published range: 61 (35–84) Tg yr^-1, a factor of 2.4 from end to end. Established It is also the boundary most often quoted as a bare integer. Frontier If the global ceiling is uncertain by that much, every national share derived from it inherits at least that uncertainty — a first-order obstacle to allocating a safe operating space between states, and one that is almost never stated when allocation is discussed.

Established A note on what this brief can stand behind. The research behind it ran under a retrieval environment in which Science and Science Advances were unreachable, as was the Stockholm Resilience Centre. Established The two canonical boundary papers are therefore bibliographic records here, not read texts: their existence, authors, venue and year are verified and nothing else is. Established Every boundary number below comes from a paper that was actually retrieved, and where a number lives in a text this brief could not obtain, the brief says so instead of supplying it.

Established The founding statement is conditional, and its own summary says so. Rockström and twenty-eight co-authors published A safe operating space for humanity in Nature 461:472–475 in 2009. Its own description reads: identifying and quantifying planetary boundaries that must not be transgressed “could help prevent human activities from causing unacceptable environmental change.” Established That is a proposal for a decision aid, phrased in the conditional. Speculative It is very widely cited as having established nine quantified planetary thresholds. Established This brief read the abstract and not the body, so it states only that the paper's own framing is conditional, and does not characterise what the body claims.

Established The most quantitatively useful retrieved source is an integrated-assessment projection, not a boundary paper. Van Vuuren and colleagues, Nature 641:910–916 (2025), project the control variables of eight of the nine boundaries forward to 2050 using the IMAGE model. Their rendering: climate change as radiative forcing, boundary 1.0 W m^-2 with the uncertainty zone running to 1.5; stratospheric ozone as less than 5% reduction from a pre-industrial 290 Dobson Units, zone 5–10%; aerosols as PM_2.5 at 10 µg m^-3, the WHO interim target, zone to 25; ocean acidification derived from 350 ppm CO₂-equivalent, zone to 450; nitrogen as cropland and pasture surplus at 62 Tg N yr^-1, zone to 82; phosphorus surplus at 11 Tg P yr^-1, zone to 15; blue-water consumption at 4,000 km³ yr^-1, zone to 6,000; land-system change as 75% of original forest area remaining, zone to 50%; and biosphere integrity as Mean Species Abundance of terrestrial plants at 90% of a natural baseline. Established Novel entities is the one boundary they do not quantify.

Established A second framework, published two years earlier and often confused with the first, uses different variables and reaches a different count. Rockström, Gupta, Qin, Lade and colleagues, Nature 619:102–111 (2023), quantify safe and just Earth system boundaries for climate, the biosphere, water and nutrient cycles and aerosols, at global and sub-global scale. Two boundary types are computed for each domain and the stricter binds. Their values: climate as global warming, safe 1.0–1.5 °C and just at or below 1.0 °C; intact natural ecosystem area 50–60% of global land, with a separate functional criterion of 20–25% seminatural habitat per square kilometre of human-modified land; surface water at 20% alteration of monthly flows; groundwater drawdown not exceeding annual recharge of roughly 16,000 km³ yr^-1; nitrogen as agricultural surplus, safe 61 (35–84) and just 57 (34–74) Tg yr^-1; cropland phosphorus surplus 4.5–9.0 Tg yr^-1; aerosols as an interhemispheric aerosol-optical-depth difference of 0.15, with the just boundary expressed instead as 15 µg m^-3 PM_2.5. Established The paper's stated result is that seven of eight globally quantified safe and just boundaries, and at least two regional ones across more than half of global land area, are already exceeded.

Established Set the two tables side by side and the framework's most serious internal problem is visible without any critique from outside. Climate: radiative forcing in one, degrees of warming in the other. Biosphere integrity: a species-abundance index in one, an area-of-intact-ecosystem fraction plus a per-square-kilometre habitat criterion in the other. Phosphorus: 11 Tg total surplus in one, 4.5–9.0 Tg cropland surplus in the other. Nitrogen: 62 with a zone to 82 in one, 61 with a published interval of 35–84 in the other. Aerosols: a concentration in one, a hemispheric optical-depth asymmetry in the other. Established These are not rival schools. They are the same research programme, with overlapping authorship, reporting the same named boundaries in different units with different numerical values two years apart. Frontier Physical thresholds do not usually change units between papers. Established The direct consequence is that transgression counts are not comparable across the two frameworks, and any allocation exercise must state which control variable it is allocating before it means anything.

Established The published uncertainty on these numbers is large, and it is published. Nitrogen safe boundary 61 with the interval 35–84, a factor of 2.4; nitrogen just boundary 57 (34–74), a factor of 2.2; the planetary-boundaries-lineage nitrogen value 62 with its zone running to 82, plus 32%; land-system change 75% of original forest with the far end of the zone at 50%, a span equal to a quarter of the world's original forest cover; aerosols 10 µg m^-3 with the zone to 25, a factor of 2.5; freshwater 4,000 km³ yr^-1 with the zone to 6,000; climate 1.0 W m^-2 with the zone to 1.5; phosphorus 11 Tg with the zone to 15. Established Every one of these intervals appears in the source alongside its central value. Frontier Every one of them is routinely dropped in retelling, which is the failure mode that matters most here: the science reports its uncertainty honestly and the communication layer deletes it.

Established No modelled scenario returns humanity to the safe operating space by 2050. Van Vuuren and colleagues find that under current trends and policies the situation is projected to worsen to 2050 for all planetary boundaries except ozone depletion. Established Targeted interventions — Paris implementation, dietary shift, food, water and nutrient efficiency — can, in their words, effectively reduce the degree of transgression; but even in that scenario climate change, biogeochemical flows and biodiversity remain transgressed in 2050, partly as a result of inertia. Frontier That is a stronger and more useful statement than any transgression count, and it sits uncomfortably beside the phrase “safe operating space,” which implies a returnability the framework's own best pathway analysis does not support on policy timescales.

Established Ozone is the one boundary that recovers, and the reason is not scientific. In the same projections, stratospheric ozone depletion is the only boundary that improves and the only one that remains inside the safe operating space through 2050. Established It is also the only boundary with a binding international treaty, a small number of controlled substances, and a substitutable industrial input. Frontier Read as a natural experiment, this is the most policy-relevant single fact in the subject: where the control variable is quantified, the emitters are few, and a substitute exists, an Earth-system process has in fact been governed. Speculative Whether that template generalises to nitrogen surplus, with hundreds of millions of emitters, or to biosphere integrity, which has neither a settled control variable nor a substitutable input, is the open question the rest of this brief circles.

Established One of the newly proposed boundaries has already been formally rejected in print, by someone inside the programme. Cuthbert, Gleeson, Bierkens, Ferguson and Taylor published a Matters Arising in Nature 635:E4–E5 (2024) calling the proposed groundwater Earth system boundary “flawed, unsafe and unjust.” Their four grounds: conceptually, it rests on an incorrect understanding of groundwater dynamics, which respond over timescales from months to thousands of years; methodologically, a criterion of no net decrease in storage during 2003–2016 ignores that historic pumping may already have reached equilibrium while causing ecosystem harm and land subsidence, and cannot separate natural variability from anthropogenic impact; on justice, the sub-global boundary excludes some of the poorest people from developing undeveloped aquifers while letting wealthy nations that depleted aquifers before the reference period escape accountability; and on data, GRACE satellite gravimetry was analysed at 0.25° resolution where guidance recommends 1° for basins exceeding roughly 200,000 km². Established Tom Gleeson is a co-author of this critique and a co-author of the 2023 nine-boundary update. Frontier The disagreement is internal to the field, which makes it far more informative than an external critique would be: it demonstrates that individual boundaries are contestable, and are being contested, by the people who build them.

Established What this brief does not have. Richardson, Steffen, Lucht and twenty-six co-authors published Earth beyond six of nine planetary boundaries in Science Advances 9(37):eadh2458 in 2023; the record is verified and the text was not obtained. Established The famous count therefore enters this brief from a title, and which six, on what control variables, at what values, is not established here. Established The same applies to the leading disciplinary critique: Montoya, Donohue and Pimm published Planetary Boundaries for Biodiversity: Implausible Science, Pernicious Policies in Trends in Ecology & Evolution 33(2):71–73 in 2018, and the title alone establishes that three senior ecologists made such an argument in a major journal. Speculative Their reasoning is not characterised here, because it was not read.

Frontier The productive question is not whether planetary boundaries is science or a policy device, but which boundaries are which. The two positions are both stateable from retrieved evidence and both are strong. Established For the policy-device reading: the founding paper frames the exercise conditionally; the Earth Commission's own justice paper states that boundaries “may need to be adjusted” to reduce harm and increase access; for climate and aerosols the just boundary binds rather than the safe one, so the operative number is normative by construction; and the same boundary changes units between papers. Established For the science reading: the boundaries are operationalised as model-projectable control variables, and eight of nine were carried through an integrated assessment model to 2050 with scenario-differentiated results, which a pure rhetorical device is not usually tractable to; they are falsifiable in practice, as the groundwater rejection shows; and the framework retrodicts correctly on its one success case, ozone, where mechanism and treaty both exist. Frontier The honest synthesis is that it is a hybrid and the framework's own authors say so — and that the hybrid character is answerable boundary by boundary rather than in general.

Frontier Biosphere integrity is the weakest link and everything downstream inherits its weakness. Three renderings, three control variables: Mean Species Abundance at 90%; intact natural ecosystem area at 50–60% of land plus 20–25% seminatural per square kilometre; and whatever the 2023 nine-boundary update uses, which this brief did not read. Established The TREE critique is aimed at exactly this boundary. Frontier The dependency runs outward: Half-Earth arguments in Biodiversity Restoration lean on the 50–60% figure as an independent Earth-system route to roughly the Half-Earth number, arrived at without invoking Wilson — which means a contested boundary is load-bearing for a conservation target covering a third of the planet's surface.

Frontier Novel entities may be a category error rather than an unmeasured boundary. It is the one boundary excluded from the only integrated-assessment treatment retrieved. Speculative For it to be a boundary in the same sense as the others, a single control variable would have to aggregate an open-ended and growing set of synthetic chemicals; no retrieved source proposes one. Handwave It is possible that the correct move is to demote novel entities from a boundary to a watchlist, and that the reason nobody does so is that nine is a better number than eight.

Speculative Boundary coupling may make boundary-by-boundary policy unsafe. A 2016 Earth System Dynamics paper by Heck, Donges and Lucht is titled Collateral transgression of planetary boundaries due to climate engineering; the record is verified and the text was not read, so this brief takes only what the title asserts — that solving one boundary can breach others. Frontier If coupling is strong, the framework's additive presentation is misleading and the object to be managed is a joint state, not nine dials. Speculative This is the point of contact with Climate Engineering and Geoengineering Governance, where the same trade is argued from the intervention side.

Speculative The control-theoretic reformulation is the most interesting live research programme and almost nobody outside a small group is running it. Heitzig, Kittel, Donges and Molkenthin published Topology of sustainable management of dynamical systems with desirable states in Earth System Dynamics 7(1):21–50 in 2016, addressing which state-space regions can be reached from which others. Established The record is verified; the text was not obtained, because the Copernicus domain was gated. Speculative The reframing it implies is strictly more useful than a transgression count: the policy question becomes which futures remain reachable rather than how many boundaries are currently crossed. Handwave Whether reachability analysis is tractable on models of Earth-system complexity is entirely unsettled, and the honest answer is that nobody has shown it.

Handwave Planetary antifragility is the furthest-out serious proposal in this subject. López-Corona, Kolb, Ramírez-Carrillo and Lovett proposed it in Earth System Dynamics 13(3):1145–1155 (2022) as an additional dimension of the safe operating space — the system's capacity to gain from volatility, which boundary framing cannot represent. Established Record verified, text not read. Handwave For it to become an objective rather than a metaphor, someone would have to define a measurable antifragility statistic for a specific Earth-system variable and then show that maximising it differs from staying inside the boundaries. Speculative Neither step has been taken in anything this brief could retrieve, which is exactly why it is worth carrying.

Frontier The field now has enough rival frameworks to need comparison papers, which is itself a finding. A 2026 preprint compares dynamic carrying capacity, planetary boundaries and safe-and-just Earth system boundaries quantitatively. Speculative A 2026 environmental-education paper treats Earth system science within and beyond planetary boundaries as curriculum. Frontier Consolidation by teaching is running ahead of consolidation by evidence, and the taught version generally does not carry the intervals.

Established The target here is an allocated and enforced safe operating space: quantified boundaries, divided between states by an agreed rule, with a consequence for exceedance. It exists for exactly one boundary. The chain below is ordered, and the binding link is not where most commentary puts it.

Established L1 — one operational definition per boundary. Today biosphere integrity has at least two and probably three control variables in the current literature, and climate, phosphorus, nitrogen and aerosols each have two. Frontier The fix is cheap: an assessment body fixes one control variable per boundary and versions it, so a 2027 number is comparable to a 2023 one. Established It is prerequisite to everything below, because a quantity that changes units between papers cannot be allocated. Speculative Nobody has done it.

Established L2 — the intervals are carried, not dropped. Nitrogen is 61 (35–84) in one framework and 62 with a zone to 82 in the other. Frontier Every published boundary value should carry its interval and every derived allocation should propagate it. Frontier The consequence is uncomfortable and is the reason this link matters: a global ceiling uncertain by a factor of 2.4 makes national shares uncertain by at least that much, which may make allocation politically impossible well before it is technically difficult.

Established L3 — boundaries survive adversarial review. The groundwater boundary did not. Frontier A formal Matters Arising process per boundary, as already happened once, is the cheapest quality instrument available. Established This link is working, and it is the healthiest part of the field.

Established L4 — a sharing principle is chosen. This is the binding link. Comparison literature on sharing principles exists; agreement does not, and the existence of a paper systematically comparing equal-per-capita, historical-responsibility, capability and grandfathering rules is itself evidence that none is settled. Established The groundwater critique shows the sharpest version: a boundary defined against a 2003–2016 reference period excludes some of the poorest people from development while letting earlier depleters escape accountability. Frontier An apparently technical choice — a reference period — is a distributive choice wearing a methodological costume.

Established L5 — downscaling machinery. Methods reviews, national-to-provincial nexus downscaling and at least one national-exceedance computation for a forest-specific boundary are already published. Frontier This is further along than L4, which is precisely why the blockage is visibly political rather than technical.

Established L6 — monitoring at allocation resolution. The GRACE objection is exactly this problem: analysis at 0.25° where guidance recommends 1° for basins over roughly 200,000 km². Frontier Different boundaries have wildly different observability; ozone and climate are well observed, groundwater and biosphere integrity are not.

Established L7 — a compliance instrument. Precedent exists once. Frontier It needs the ozone preconditions: few substances, a substitutable industrial input, a quantified control variable. Speculative Nitrogen surplus has the third and not the first two; biosphere integrity has none of them. Frontier A brief that presents allocation as awaiting better Earth-system data mis-locates the problem. It is an institutional blocker downstream of an unresolved normative question, and L5 being largely built is the proof.

Established This subject depends on results other fields are not currently producing. The first is a mechanism linking each control variable to a genuine Earth-system nonlinearity. Frontier For climate and ozone the mechanistic case is strong; for nitrogen, phosphorus and land-system change the boundary values read as aggregated risk thresholds rather than as identified tipping points, and no retrieved source demonstrates otherwise.

Frontier The second is reconciliation of the biosphere-integrity control variables. Mean Species Abundance and intact-ecosystem-area are not two estimators of one quantity; they measure different things and would be transgressed at different times under the same world. Speculative Whether they can be reconciled, or whether one must be abandoned, is a scientific question with no visible programme behind it.

Established The third is the assessment literature's own hygiene. The research behind this brief observed its bibliographic tooling return a result set about organ cryopreservation in response to a biodiversity-target query — silent substitution of one document for another, with no error and no low-confidence signal. Frontier That is a real fact about how citations get corrupted in machine-assembled scholarship, and this subject is governed by exactly the kind of large multi-author assessment that is assembled that way. Established Every citation in this brief was checked for internal consistency of title, venue, authors and year.

Established The fourth is retrieval itself. The nine-boundary table this brief most wanted sits behind a domain that could not be reached, and three open-access Earth System Dynamics papers on antifragility, reachability and boundary coupling were gated. Frontier A subject whose primary literature is partly unreachable to an ordinary reader is a subject whose public understanding is set by summaries, and the summaries are where the intervals go to die. Speculative That is a dependency on publishing economics rather than on science, and it is not obviously the easier of the two to move.

Established The experiments this subject needs are mostly conventions and audits, not instruments — which is why they are cheap and undone.

Frontier One: publish a versioned control-variable registry. For each of the nine boundaries, one control variable, one definition, one value, one interval, one version number, with a documented change log when it moves. Established The measurable outcome is that two papers two years apart become comparable, which today they are not. Speculative The obstacle is that fixing a definition freezes a scientific argument that some authors would rather keep open.

Frontier Two: recompute one national allocation under four sharing rules and publish the spread. Take nitrogen, whose global ceiling is published as 61 (35–84) Tg yr^-1, and allocate it to a set of states under equal-per-capita, historical-responsibility, capability and grandfathering rules, propagating the ceiling's interval through each. Frontier The predicted result is that the between-rule spread exceeds the within-rule uncertainty for most states, and that for some the two are comparable — which would establish, quantitatively, that the argument about allocation is normative rather than empirical. Established No retrieved source has run this.

Frontier Three: re-run the groundwater boundary at the recommended resolution. The GRACE objection is specific and testable: 0.25° where guidance says 1° for basins over roughly 200,000 km². Frontier Recomputing the sub-global boundary at 1° and reporting how the exceedance map changes is a well-defined experiment that would either rehabilitate the boundary or retire it.

Speculative Four: a retrodiction test on ozone. Run the boundary framework backwards over the Montreal period and ask whether it would have identified the ozone boundary as the tractable one ex ante, on the criteria of emitter count, substitutability and control-variable quality. Frontier A framework that scores its own boundaries for governability, rather than only for transgression, would be a different and more useful instrument.

Handwave Five: compute a reachability set for one coupled boundary pair. Even a toy demonstration — which future states of a two-boundary system remain attainable from today under bounded intervention — would test whether the control-theoretic reformulation is a research programme or a metaphor.

Established The whole edifice rests on observing systems, and their resolution is a scientific parameter, not a technical detail. The clearest published instance is the GRACE dispute: the same satellite gravimetry data analysed at 0.25° supports a groundwater boundary that its critics say collapses when analysed at the 1° guidance resolution for large basins. Frontier That is an engineering constraint determining a policy number.

Established Observability differs by an order of magnitude across the nine. Ozone column and radiative forcing have long, calibrated, redundant observing records. Frontier Groundwater storage is inferred from a gravity mission whose footprint is coarser than most aquifers. Frontier Biosphere integrity, on either of its two control variables, is assembled from modelled abundance indices or from land-cover classifications, neither of which is a direct measurement of the quantity the boundary names. Speculative A framework that reports nine boundaries as comparable objects is implicitly claiming a uniformity of observation that does not exist.

Established There is a measured analogue for what happens when a monitoring framework is built above the scale at which the data exist. The Global Urban Monitoring Framework comprises 77 indicators across five domains and was endorsed by the UN Statistical Commission in March 2022; a 2026 analysis in npj Urban Sustainability reports that fewer than 20% of all indicators are, on average, reported at city level, with most reported nationally. Frontier An indicator framework whose evidence base is more than 80% at the wrong scale is a measurement constraint that becomes an institutional one, and boundary downscaling to national allocations faces the same failure in reverse.

Frontier The unglamorous requirement is continuity. Earth-system control variables are trends, and a trend needs an instrument record longer than the political cycle that funds it. Speculative No retrieved source in this pack quantifies mission-continuity risk for the specific instruments the boundary framework depends on, and this brief does not supply a figure it could not obtain — but the structural point stands: the boundary values are only as durable as the observing programmes behind them, and those are appropriated annually.

Established The nearest neighbour is conservation policy, and the dependency runs in the direction people expect least. The biosphere-integrity boundary supplies the Earth-system warrant for area targets; the delivery literature then supplies the reality check. Established Global restoration prioritisation finds that restoring 15% of converted lands in priority areas could avoid 60% of expected extinctions while sequestering 299 gigatonnes of CO₂, with cost-effectiveness varying up to thirteen-fold with spatial allocation. Established Protected-area workforce accounting finds current terrestrial personnel at roughly 555,000, one per 37 km², against approximately 3 million needed for a 30% estate. Frontier A boundary is a target; these are the delivery functions, and they are covered in Biodiversity Restoration.

Established Cities are where several boundaries are actually breached, and where the local version of biosphere integrity has been measured. A 2026 Nature Cities study of stream communities across the São Paulo metropolitan region — a megacity of over 22 million — found food-web complexity collapsing from networks of more than 150 trophic links to as few as three, with a 1,446-fold reduction in energy flow. Frontier That is what a transgressed biosphere boundary looks like at the scale where people live; see Sustainable Megacities.

Speculative Intervention technologies are the framework's hardest adjacency. Solar radiation management and large-scale carbon removal are proposals to move one control variable directly, and the retrieved literature asserts that doing so causes collateral transgression of others. Frontier Climate Engineering, Geoengineering Governance and Atmospheric Management carry the intervention side; this brief carries the accounting frame that says an intervention is only a gain if the joint state improves.

Frontier The least obvious adjacency is administrative rather than physical. Every step in section 4 — fix a definition, publish an interval, choose a sharing rule, audit compliance — is a public-administration problem wearing Earth-system clothing. Speculative Existential Risk Governance asks a structurally identical question about a different class of thresholds, and runs into the same wall: nobody has authority over a global quantity, and the quantity is not defined the same way twice.

Established No retrieved source describes an operating international institution with authority over any planetary boundary except ozone. That is a strong statement and it is honestly bounded: the research behind this brief could not reach the Earth system governance literature, and its absence here is a gap in retrieval rather than proof of absence in the world. Established What is verified is that the Earth Commission has produced a justice framework, a boundary set and an emerging comparison literature, and that allocation research exists at national and provincial scale. Frontier What is not evidenced anywhere in the pack is a treaty instrument, outside ozone, with a quantified control variable and a compliance mechanism attached to it.

Frontier So what would a body with authority over an Earth-system process actually look like? The only worked example says: a named control variable with an agreed definition and interval; a small enough set of regulated flows that compliance is auditable; a substitutable industrial pathway so that compliance is affordable; a scheduled tightening; and a consequence. Established The Montreal architecture has all five. Speculative Nitrogen surplus has a quantified control variable and hundreds of millions of emitters, so auditability fails. Speculative Biosphere integrity has neither a settled control variable nor a substitutable input, so two of the five preconditions are missing before negotiation starts.

Frontier The institutional design problem is therefore not authority in the abstract but the object of authority. A body given authority over “the nitrogen boundary” would immediately have to say whether it governs 62 or 61, total surplus or agricultural surplus, and with what interval — and the answer redistributes obligations between states before any negotiation about shares begins. Speculative This is why the sequencing in section 4 puts definition before allocation: the definitional choice is a quiet allocation.

Speculative A minority institutional proposal worth carrying: version the boundaries like a standard rather than publishing them like findings. Semantic versioning, a change log, a deprecation policy for retired boundaries such as groundwater pending re-analysis, and a stated rule for when a value may move on normative rather than empirical grounds. Handwave No body has claimed this role, and the framework's authors have never proposed it in anything this brief retrieved. Frontier It is, however, the smallest institutional change that would make the numbers usable in law.

Established The framework's own authors state that boundaries may need to be adjusted on justice grounds. The Earth system justice paper — Gupta, Liverman and twenty-four co-authors, Nature Sustainability 6:630–638 (2023) — says that living within planetary limits requires attention to justice because biophysical boundaries are not inherently just, that stringent boundaries may affect just access to food, water, energy and infrastructure, and that boundaries may need to be adjusted to reduce harm and increase access. Frontier A quantity revisable on normative grounds is doing something other than reporting a biophysical threshold. Established This is the strongest internal evidence for the policy-device reading, and it is made by the framework's own authors rather than by its critics.

Established For climate and aerosols, harm to humans sets the operative limit, not Earth-system stability. The safe-and-just framework reports that justice considerations constrain the integrated boundaries more than safety considerations for those two domains; the just climate boundary at or below 1.0 °C is stricter than the safe boundary of 1.0–1.5 °C. Frontier People who quote the framework as physics are quoting an ethical judgement for those two boundaries.

Established Grandfathering is not hypothetical here; it has been demonstrated on a specific proposed boundary. A criterion of no net groundwater decrease during 2003–2016 excludes states that have not yet developed their aquifers while exempting states that depleted theirs before the window opened. Frontier Every reference period is a distributive choice, and the ethical content of an allocation is usually already fixed by the time the allocation formula is discussed.

Speculative There is a heterodox reading that the safe operating space is a labour and consumption question rather than an emissions question — a 2026 preprint explores work within planetary boundaries beyond waged labour. Established Record verified, text not read. Handwave Carried because the framework's distributive implications have been argued far more thoroughly on the production side than on the side of who has to change what they do.

Frontier If the modelling is right, the safe operating space is not somewhere this civilisation returns to on any timescale it plans over. Climate, biogeochemical flows and biodiversity remain transgressed in 2050 even under strong policy, partly through inertia. Speculative The options are then boundary revision, negative-flux technologies at scale, or acceptance of a multi-century return time — and only the third requires nothing new to be invented.

Speculative That has an uncomfortable consequence for the framing. “Safe operating space” is a phrase about a place one can be in or out of. Frontier A framework whose own best pathway analysis shows no return within a policy horizon is describing a trajectory, not a room, and the metaphor may be doing quiet damage to the politics by implying that a sufficient effort restores a prior state.

Frontier The consolidation risk is real and measurable. The framework is entering curricula while its control variables remain inconsistent between papers by the same authors. Speculative A generation taught nine dials with six shaded red, and no units and no intervals, will find it harder rather than easier to negotiate an allocation later, because the numbers they were taught will not survive contact with a treaty draft.

Handwave The optimistic reading, stated as what it is. Ozone shows that a planetary process can be governed when the object of governance is well posed. Speculative If the definitional work in section 4 were actually done, the subject would move from a rhetoric of limits to an engineering of ceilings — and there is no scientific reason it cannot, only an institutional one. Frontier The measure of whether this civilisation ever became a steward of anything will not be how many boundaries it named; it will be whether it ever governed a second one.

Horizons for a framework whose science is faster than its institutions, and whose institutions are faster than its definitions.

• 10 yr: Frontier A versioned control-variable registry is technically trivial and would settle the comparability problem; expect instead continued parallel publication of two frameworks with overlapping authors. Frontier Downscaling to national indicators continues and improves, because it is publishable without agreement. Speculative At least one more boundary receives a formal Matters Arising, because adversarial review is the one part of this chain that is functioning. Frontier Boundary-referenced corporate disclosure spreads faster than the definitions stabilise, which means firms reporting against a nitrogen number whose published interval spans a factor of 2.4 without printing the interval.
• 25 yr: Speculative A binding instrument for a second Earth-system process is conceivable only where the ozone preconditions hold; the most plausible candidate is a nutrient-flux instrument covering a small number of large fertiliser producers rather than the emitters themselves. Speculative Allocation between states remains blocked on the sharing rule rather than on data, and the downscaling literature keeps getting better at a task nobody has authorised. Handwave A serious possibility over this horizon is that the framework quietly stops claiming nine boundaries and reports a smaller set it can actually defend.
• 50 yr: Speculative If reachability analysis becomes tractable, the policy object changes from “how many boundaries are crossed” to “which futures remain attainable,” which is a different negotiation with different winners. Handwave An enforceable allocation of a global biophysical ceiling between sovereign states has no precedent outside ozone and would be the largest voluntary transfer of sovereignty in history. Speculative The likelier route is indirect: trade instruments and supply-chain rules that price a boundary-referenced flux without anyone ever agreeing a national share.
• 100 / 250+ yr: Handwave Return to the safe operating space on any control variable other than ozone is a multi-century proposition under the retrieved modelling, and the framework does not currently say so out loud. Handwave Whether “planetary stewardship” survives as a concept, or is replaced by an antifragility or reachability formulation that treats volatility as the object rather than the enemy, is genuinely open. Handwave On this horizon the interesting artefact may be the observing record itself: a continuous, calibrated, multi-century series for a handful of control variables would be worth more than any framework built on top of it.

• Depends on A settled operational definition for each control variable — a scientific and editorial task rather than an observational one, and the cheapest undone thing in the subject. It depends further on Earth-system observation records long enough to define a trend rather than a snapshot, at resolutions matched to the basins, biomes and jurisdictions the numbers will be applied to; and on the conservation and land-use delivery literature, which is what says in staff and money what any biosphere target would actually cost to meet.
• Requires (not on this map) Three things that are not research results. First, a body with authority over an Earth-system process: the ozone regime is the only working example, and it works because the control variable is quantified, the emitters are few and the input is substitutable — none of which holds for nitrogen or biosphere integrity. Second, an allocation of the safe operating space between states: the downscaling machinery is largely built, so what blocks allocation is the choice of sharing rule, and the fact that a global ceiling published as 61 (35–84) Tg yr^-1 hands every national share a factor-of-2.4 uncertainty before anyone argues about fairness. Third, earth-observation funding that survives political cycles: control variables are trends, trends need instrument records longer than the appropriations that pay for them, and the resolution of a single gravity mission has already been decisive in whether one proposed boundary stands.
• Enables National absolute-sustainability accounting, boundary-referenced corporate and financial disclosure, and any future compliance instrument for an Earth-system process other than ozone all wait on this. It is also the accounting frame against which climate-intervention proposals have to be scored, since an intervention is only a gain if the joint state improves — and the retrieved literature asserts that solving one boundary can breach others. Every downstream use inherits whatever definitional instability sits upstream of it, which is why the registry question in section 4 comes first.
• Adjacent Sits alongside climate engineering and geoengineering governance, which propose to move a control variable directly and therefore need this accounting frame to say whether they helped; to biodiversity restoration, which supplies the delivery arithmetic behind the biosphere boundary and the reality check on any area target derived from it; to sustainable megacities, where several boundaries are locally breached and where the collapse has actually been measured; and to existential-risk governance, which asks a structurally identical question about a different class of thresholds and runs into the same problem of who has authority over a global quantity.

Established “Six of nine planetary boundaries have been crossed” is not a settled measurement you can quote without a framework attached. In the evidence available here the count comes from a paper title, the text of which was unreachable. Established It is also not the same statement as “seven of eight globally quantified safe and just Earth system boundaries are already exceeded,” which comes from a different framework with a different denominator, a different boundary set and different normative content. Frontier Quoting both, averaging them, or reconciling them arithmetically is an error. Established Name the framework whenever a count is used.

Established The deeper problem: the same boundary carries different control variables in different papers by overlapping author teams. Climate is radiative forcing at 1.0 W m^-2 in one and warming at 1.0–1.5 °C in the other. Biosphere integrity is Mean Species Abundance at 90% in one and intact ecosystem area at 50–60% of land plus 20–25% seminatural per square kilometre in the other. Phosphorus is an 11 Tg P yr^-1 surplus in one and a 4.5–9.0 Tg yr^-1 cropland surplus in the other. Established Rockström authors both papers. Frontier So the well-formed statement is not “the nitrogen boundary is 62 Tg” but “in van Vuuren and colleagues' 2025 rendering, the nitrogen control variable is cropland and pasture surplus with a boundary at 62 Tg N yr^-1 and an uncertainty zone to 82.” Established The enthusiast form of this error is a nine-segment infographic with six wedges shaded red and no units anywhere on it.

Established “The safe operating space is a physical property of the Earth system.” For climate and aerosols the just boundary binds rather than the safe one, and the framework's own justice paper states that boundaries may need to be adjusted to reduce harm and increase access. Frontier At least part of the operative number is normative by design and by the authors' own statement. Established This is not a critique smuggled in from outside; it is in the programme's own literature.

Established “The boundary is X” — where the published interval has been silently deleted. Nitrogen 61 (35–84), a factor of 2.4. Aerosols 10 with a zone to 25, a factor of 2.5. Freshwater 4,000 with a zone to 6,000. Land-system change 75% of original forest with the far end at 50%, a span equal to a quarter of the world's original forest cover. Established All published at source. Frontier The practical rule this brief follows: carry the interval every time, and if there is no room for the interval, do not use the number.

Established “The framework is unfalsifiable.” A proposed boundary was formally challenged in Nature on conceptual, methodological, justice and data-resolution grounds, by a group including a co-author of the nine-boundary update. Frontier Whatever else is true, individual boundaries are contestable and are being contested. Speculative What has not been shown is that a rejected boundary is subsequently retired from the framework's summary graphics; this brief could not check that, and flags it as unverified rather than asserting it.

Established “Novel entities has been quantified.” It is explicitly the one boundary not projected in the only integrated-assessment treatment retrieved. Frontier A news item asserting that plastic pollution has pushed Earth past all nine boundaries circulates with a citable identifier attached; it is journalism, not a study, and it should not be cited as evidence for a boundary count.

Established “Ambitious policy returns us to the safe operating space.” The retrieved modelling says climate, biogeochemical flows and biodiversity remain transgressed in 2050 even under strong policy, partly from inertia. Frontier Optimistic framings of the framework contradict its own best pathway analysis, and the contradiction is in the same journal family.

Established “Planetary boundaries” and “safe and just Earth system boundaries” are not the same framework. They have overlapping authorship, different variables, different counts and different normative content. Frontier This is the single most common conflation in secondary writing on the subject, and it is the reason two different transgression counts circulate as though they were successive measurements of one thing.

Established A structural warning about how numbers in this field propagate. The clearest documented case in the surrounding literature is a restoration-cost paper whose title carries 310,000 EUR per hectare while its own results report 195,000 once four statistical outliers are excluded — the inflated figure is in the title, so indexing and citation carry it and the correction stays in the body. Frontier Boundary values are exposed to the same mechanism: what propagates is the headline, and the interval is in the body.

Frontier Finally, a caution about the machinery of citation itself. During the research for this brief a bibliographic query returned a result set about organ cryopreservation in response to a question about biodiversity targets — a silent substitution of one document for another, with no error signal. Speculative Assessment reports in this field are assembled at exactly the scale where such substitutions are invisible. Established The defence is boring and effective: resolve every identifier, and check that title, venue, authors and year cohere before a citation ships.