The Numbers

What the Math Requires

Full Technical Validation — Version 12

All figures are first-order estimates. Peer review invited. The direction of honest correction is always toward longer timescales and greater complexity. These are the most honest numbers we have. Each claim is tagged with a confidence level: Established Science, Strong Evidence, Reasonable Inference, Informed Speculation, or Frontier Territory.

Part One: The Answer

This document validates every number behind the Trust Collective framework. It begins with the conclusion, because the conclusion is what matters.

The carbon math requires that approximately 80–90% of Earth's habitable land be returned to a natural, functioning state to bring atmospheric CO₂ from its projected peak back to the pre-industrial baseline of 280 parts per million (ppm). The absolute mathematical floor is 79%. The realistic requirement, accounting for compound stress and unidentified uncertainties, is 85–90%.

That leaves approximately 10–20% of habitable land — roughly 10 to 21 million km² — available for all of human civilization.

That sounds like a constraint. It is not.

The Trust Collective (TC) settlement and food system models were designed to give every person on Earth a quality of life that would be considered good by any standard — secure housing, abundant food, healthcare, education, culture, community, and freedom. The design constraint was not austerity. It was generosity within the smallest possible footprint.

The result: the entire TC model — cities woven through with green space and food forest, underground agriculture producing enough food for the projected peak population of approximately 10.3 billion people with a 30% surplus for resilience, underground maglev transportation, parks, villages, and cultural spaces — fits within approximately 10% of habitable land. Roughly 10.4 million km². That is slightly larger than the area of the United States or China. This footprint would not be concentrated in one place. It would be distributed across every continent, incorporating and transforming many of the places people already live. All of civilization, living well, within that footprint.

The carbon math requires a minimum of 79% for nature. The TC model needs 10% for people. That leaves a maximum gap of approximately 11% — roughly 11.4 million km² — that is neither claimed by the climate math nor needed by the settlement model. That maximum exists only at the most favorable conditions: rapid transition and high Solar Radiation Management (SRM) effectiveness. At realistic mid-range estimates, the room beyond TC is approximately 3–9%. At the high end of restoration need, it narrows to less than 1–2%.

That room is real. It means the TC model fits comfortably at every viable scenario. There is space for more generous communities, cultural landscapes, heritage sites, and places that matter to people for reasons that have nothing to do with carbon. The room is not unlimited, and its size depends on speed. But even at the tightest realistic estimate, the 10% TC model provides a quality of life most people would consider generous.

How much room remains depends on how quickly the transition happens. Every year of delay adds approximately 20 gigatonnes of CO₂ (GtCO₂) to the atmosphere, which increases how much land must be returned to nature and reduces what remains for people. Under an accelerated transition, the room beyond TC ranges from 6–11%. Under a gradual transition, it shrinks to less than 1–5%. Speed directly expands the space available for human life.

At every point in this range, the land-use transformation required is incompatible with a growth-based economy. The resource-based economy is not an ideological preference. It is what the carbon math requires.

Three independent lines of analysis — carbon math, food science, and settlement engineering — converge on this range. The convergence is the argument.

The rest of this document shows where every number comes from.

Part Two: Energy

Total Demand

TC sustained demand: approximately 20–27 TW. Peak demand during the construction era (years 0–200): approximately 28–35 TW. Derived from bottom-up sectoral analysis. Current global primary energy is approximately 18 TW (IEA 2025), but 60–65% is wasted as heat in fossil combustion (RMI 2024). An all-electric civilization serves the same useful work far more efficiently.

Energy Budget by Sector

The TC civilization requires roughly one-seventh the energy per unit of useful work compared to current fossil-based systems. Energy supply is not the constraint — geothermal alone provides approximately 50 TW at moderate depths (IEA 2024), with hundreds of TW accessible at greater depths.

Primary Source: Geothermal

Accessible geothermal with advanced drilling: approximately 50 TW baseline, expandable. IEA (2024) The Future of Geothermal Energy found 42 TW Enhanced Geothermal Systems (EGS) potential at less than 5 km depth, and approximately 600 TW at less than 8 km. Clean Air Task Force (CATF) 2024 found that 1% of superhot rock potential equals 63 TW.

Geothermal is the backbone — unlimited baseload, available everywhere, with existing pipeline infrastructure repurposed. Wells require redrilling every 20–30 years as a known engineering cycle. The energy source is permanent. Wells are maintained infrastructure. Pipeline workers have the skills and are honored and employed in the transition.

Confidence: Strong Evidence.

Deep City Geothermal Integration

City infrastructure extends 1–2 km below surface. At 2 km depth, ambient rock temperature is approximately 60–80°C (geothermal gradient of approximately 25–30°C/km; Glassley 2014). Geothermal drilling from the base requires 1–2 km less depth to reach superhot conditions. City and energy system are integrated by design.

Supplemental Sources

Surface solar: dual-use only — agrivoltaics over agricultural buffer zones and building integration. Standalone solar fields are not used. That land grows plants.

Wind, tidal, and wave energy contribute during the transition period but are phased down over time in favor of geothermal and solar, which have no moving parts exposed to weather and require less maintenance — consistent with the thousand-year infrastructure philosophy.

Long-Term Temperature Management: Space Shades at L1

Space shade array: approximately 100,000 tonnes total at L1. Borgue and Hein (2022) confirmed this mass for ultra-thin polymeric films with SiO₂ nanotubes. 50–100 year project. Temperature management, not energy generation.

Moon-anchored spaceline: feasible with current materials (Zylon, Dyneema). Pearson et al. (2005, NASA NIAC), Penoyre and Sandford (2019), JBIS (2023). A 48-ton system deployable via single heavy-lift launch. 20–30 year development timeline.

Energy Storage

Pumped hydro (elevated reservoirs and turbines), gravity batteries (heavy masses on cables including construction waste — nothing to break, unlimited cycles), compressed air, and thermal storage.

Transport

Shared autonomous fleet: Replaces private car ownership. Approximately 80% vehicle fleet reduction. One shared autonomous vehicle replaces 4–12 private cars (Scientific Reports 2025; PMC 2022). Parking demand reduced 80–90%.

Long-distance transport: Underground maglev systems. Near-zero surface footprint. Surface roads and parking reclaimed for ecosystem restoration.

Manufacturing CO₂ saved: approximately 9,600 Mt one-time avoided emissions from eliminating approximately 1.2 billion private vehicles from the fleet.

Part Three: Food and Land

Design capacity: enough food for every person at projected peak population of approximately 10.3 billion (UN World Population Prospects 2024), with a 30% surplus for resilience, genetic conservation, and insurance against disruption. That surplus is deliberate over-engineering — it means the food system can absorb shocks, maintain biodiversity in cultivated species, and still have capacity to spare. Total caloric demand at design: 10.95 × 10¹⁵ kcal/yr.

Nutritional requirements go beyond raw calories: approximately 60–70 grams of protein per person per day (FAO/WHO), adequate fat intake of 20–35% of calories, plus essential vitamins, minerals, and fiber. The four-zone food system delivers a complete diet, not merely energy. A formal nutritional audit against WHO/FAO dietary reference intakes is flagged as a research need.

The food system operates across four integrated zones. Each zone serves a distinct function. Together they provide redundancy: if any single zone underperforms, the others compensate. This is not optimization for minimum sufficiency. It is deliberate over-engineering for civilizational resilience.

Zone 1: Underground Core (60% of Calories)

Primary calorie engine. Climate-independent. Latitude-independent. Year-round production. Every square meter of underground growing space is a square meter of surface land returned to living ecosystem. Underground facilities are thermally buffered by surrounding rock mass, reducing heating and cooling energy. At depth, geothermal energy is closer — a facility one kilometer underground needs one kilometer less drilling to reach superhot rock.

Primary calorie production: tubers (potato, sweet potato, cassava) in aeroponic tier systems. Grains (wheat, rice) in controlled-environment tiers with optimized lighting and elevated CO₂. Legumes (lentil, bean, soy) for plant-based protein. Diversity: leafy greens and herbs in vertical panel systems. Mushrooms including light-requiring gourmet species (lion's mane, maitake, shiitake) integrated via CO₂ cascade — mushrooms in upper tiers respire CO₂ that falls onto photosynthesizing greens below. Dwarf fruit trees (strawberry, citrus, apple) under specialized LED spectrum in high-ceiling tiers.

Tier configuration: 8 effective growing tiers. Thermal management via underground rock mass as initial heat sink, transitioning to geothermal heat exchange at steady state. Blended yield: approximately 105,000–208,000 kcal/m²/yr of floor space (moderate: approximately 147,000). Derived crop-by-crop in the TC Food System Model.

Zone 1 floor area: approximately 32,000–63,000 km² (moderate: approximately 45,000 km² — roughly the area of Denmark). Climate and latitude independent.

Confidence: Reasonable Inference. Modeled, not demonstrated. Staple grain underground and underground fruit remain untested at scale.

Zone 2: Surface Agricultural Buffer (24–48% of Calories)

Grains (wheat, rice, oats, barley, corn) under agrivoltaic arrays around cities and villages. Agrivoltaics generate 6–12 TW continuous at 30% panel coverage, potentially covering the entire underground food energy demand. Low-height autonomous tractors. Orchards and perennial food forests as permaculture transition.

Small-scale traditional livestock integrated throughout Zone 2. Heritage breed cattle, goats, sheep, pigs, and poultry raised free-ranging on pasture within and around settlements. These animals serve multiple roles: high-quality protein and dairy, living genetic libraries preserving breed diversity, soil health through managed grazing, and cultural continuity with food traditions that matter to people. This is not industrial agriculture. It is village-scale animal husbandry — the way most of the world kept animals for millennia before factory farming.

Zone 2 area: approximately 2–4 million km² (approximately 1.5–3% of habitable land).

Zone 3: Food Forests (1–3% of Calories)

Multi-story perennial polycultures: canopy trees (nut and fruit), understory trees (dwarf fruit, nitrogen-fixers), shrubs (berries), herbaceous layer (perennial vegetables, herbs), ground cover, vine layer, and root layer. Research from Scotland (Crawford 2017) found mature food forests can supply caloric needs for 6–10 people per hectare. Danish and Norwegian studies found similar ranges. Tropical food forests significantly exceed these figures.

Food forests are the transition gradient between human life and wild ecosystem. A person walks from their apartment through a neighborhood food forest, past fruiting trees and berry hedges, and the forest gradually becomes wilder — less maintained, more diverse, more dense — until it is indistinguishable from the restored ecosystem beyond. There is no hard boundary. The city fades into food forest, and food forest fades into wilderness.

Food forests reach full productivity in 8–15 years. This aligns with the TC transition timeline: planted in the first decade, productive by the second. Their primary value is nutritional diversity (fruits, nuts, berries, herbs that are difficult to produce at scale underground), community-scale food sovereignty, genetic conservation of thousands of cultivars, and the ecological gradient from settlement to restoration. Area: 200,000–500,000 km².

Zone 4: Wild Harvest (Growing Over Centuries)

In a world with 80–90% restored ecosystem, wild harvest is not marginal. It is a substantial and growing food source. This includes large and small game (deer, elk, moose, pronghorn, wild turkey, waterfowl, and restored bison herds), managed marine harvest (wild-caught fish, shellfish, seaweed), freshwater fish from restored rivers and lakes, wild-harvested plants (berries, nuts, tubers, greens), and the systematic culling of invasive species such as feral pigs.

AI-assisted drone herding manages wild populations humanely. Behavioral algorithms direct animals through sorting infrastructure for selective harvest. Females, juveniles, and breeding stock are released. Animals selected for harvest are directed to processing. This replaces industrial feedlots with ecosystem-integrated protein production that simultaneously maintains grassland health through grazing pressure — exactly as bison maintained the Great Plains for millennia.

Wild harvest grows over centuries as ecosystems mature. The food system does not degrade over time. It improves. Each century of restoration produces more abundant wildlife, more productive fisheries, more diverse wild plant communities. By century three or four, wild harvest may contribute significantly more than the first-order estimates used here.

Confidence: Informed Speculation for early estimates. Strong Evidence that restored ecosystems produce increasing wild food over time.

The Protein Architecture

The TC protein system is integrated across all four zones, providing a complete, culturally familiar portfolio.

Underground aquaponics forms a closed-loop system. Black soldier fly (BSF) larvae convert food waste into high-protein feed. Fish — tilapia, catfish, trout, and other species suited to controlled environments — eat the BSF larvae. Humans eat the fish. The same aquaponic systems produce greens and herbs from nutrient-rich water. One integrated cycle turns waste into protein and vegetables with near-zero external input.

Legumes (lentils, beans, soybeans, chickpeas) are grown underground in dedicated tiers, providing plant-based protein at scale. Mushrooms provide additional protein along with essential micronutrients.

Surface livestock in Zone 2 provides dairy, eggs, and meat from heritage breeds raised humanely on pasture. Wild game from Zone 4 provides lean, nutrient-dense protein that increases as ecosystems recover. Sustainable marine harvest resumes once ocean populations are healthy.

No single source carries the system. The redundancy is deliberate. If aquaponics underperforms, legumes and surface livestock absorb the difference. If marine harvest is slower to recover, terrestrial sources fill the gap. The protein portfolio is diverse by design.

Total Agricultural Footprint

Total: approximately 2.5–4.5 million km², or approximately 2–3.5% of habitable land.

Agricultural land freed: approximately 44–46 million km² returned to native ecosystem (approximately 90–95% of current agricultural land).

Total system at moderate estimates: 100% of projected peak population fed with 30% surplus. Conservative: 86%. Optimistic: 113%. If Zone 1 underperforms, Zone 2 absorbs the difference — requiring more surface land and reducing the restoration percentage. Food system performance and restoration percentage are coupled variables.

The 10% Breakdown

The TC model requires approximately 10% of habitable land — roughly 10.4 million km². For scale: that is slightly larger than the area of the United States or China. Roughly a third of Africa, or about twice the area of the European Union.

This footprint is not one contiguous block. It is thousands of communities distributed across every continent, woven into the landscape wherever people already live and wherever the geography is suitable. Many existing cities and towns are incorporated into the design — retrofitted, deepened, and connected. New communities are built where optimal geology, climate, and existing population density make them most useful. The footprint is global and distributed by design.

That 10% breaks down as follows: settlements and urban infrastructure occupy approximately 5–7% of habitable land, including cities, villages, parks, cultural spaces, and all surface infrastructure. Zone 2 surface agriculture occupies approximately 1.5–3%. Zone 3 food forests occupy approximately 0.2–0.4%. Underground facilities (Zone 1 food production, maglev transportation, utilities) add negligible surface footprint because they are beneath the surface.

Ten percent is not arbitrary. It is the ceiling of engineered component calculations, each derived from specific area data, with margin built in for more generous communities and design flexibility. The settlement footprint is determined by population density in nature-integrated cities. The agricultural footprint is determined by the four-zone food model. Together they fit within a space that most people can visualize — one large country — and provide a quality of life that exceeds what most of the world experiences today.

Transition-Period Food Security

During the 30–60 year construction period, existing agriculture continues in each region until TC food systems come online. No person goes hungry because of the transition. Underground food production is built and brought to capacity in each adopting region before the corresponding surface agricultural land is released for restoration. The food system transitions in sequence, not all at once.

How the City Integrates with the Food System

The underground food system and the city above it form a single breathing organism. Air from homes, from biodigesters processing food waste, and from surface intakes flows downward through the underground agricultural tiers. Plants in the growing chambers absorb CO₂ and release oxygen. The air that returns to homes is scrubbed — high in oxygen, low in CO₂, potentially lower than ambient outdoor concentrations. The city's own biological processes clean the air its residents breathe.

If the underground agricultural capacity is sufficient, the system can also scrub some portion of outside air drawn in from the surface — a net atmospheric cleaning function built into the city's infrastructure. The exact concentrations require modeling, but the principle is established: photosynthesis in the growing tiers processes CO₂ from every source the city generates.

Building with Concrete: A Carbon Solution

Food waste enters anaerobic digesters, producing biogenic methane for energy and CO₂ for carbonation curing in concrete. The CO₂ is permanently mineralized into the building material itself — not stored temporarily, but locked into the crystal structure of the concrete. This is permanent sequestration measured in geological time.

Carbonation curing at 45% efficiency produces concrete with no strength loss (Fu et al. 2024). Huang et al. (2025, ESSD) found that current global cement materials already absorb 0.86 GtCO₂/yr from approximately 300 billion tonnes of existing stock. The TC builds orders of magnitude more concrete than currently exists, and every cubic meter of it is a carbon sink.

Traditional concrete uses steel rebar for reinforcement. The TC replaces steel with basalt fiber — made from crushed volcanic rock, requiring no mining of iron ore and no blast furnaces. Basalt fiber reinforcement reduces embodied CO₂ by approximately 70–90% compared to steel (Kingston University lifecycle assessment). It does not corrode, giving it a lifespan measured in centuries rather than decades. This is concrete built to last a millennium.

The production cycle is carbon-negative when done correctly: biogenic CO₂ from digesters is captured during curing, basalt fiber replaces steel, and the finished structure absorbs additional atmospheric CO₂ over its lifetime through natural carbonation. Building with concrete in the TC is not a climate cost. It is a climate solution. Infrastructure is built to last: concrete bins, not plastic trays. A growing tier installed in Year 10 is still operating in Year 1,010.

Part Four: The Carbon Sequestration Stack

Seven sources of carbon removal — actual drawdown, not avoidance of emissions. Each source draws carbon from the atmosphere and stores it in living systems, mineral structures, or ocean biomass. All layers operate simultaneously at full deployment.

Human Land Use: The Full Picture

Total habitable land surface: approximately 104 million km².

Standard statistics report approximately 48–50 million km² (46–48%) as human-used (FAO 2021; Ellis et al. 2010). This substantially understates the actual footprint.

Total human impact: 85–90% of habitable land is significantly affected. Only 10–15% remains genuinely wild and ecologically functional. Venter et al. (2016, Nature Communications) found 75% under measurable human pressure. Sanderson et al. (2002, BioScience) found 83% directly influenced. Mo et al. 2023 (Nature) found up to 80% of the world's forests degraded.

All approximately 90 million km² of economically exploited land — cropland, pasture, logged forests, overgrazed woodland, drained wetlands — forms one interconnected pool of human land use. Restoration draws from this entire pool. A logged forest is as much an economic asset as a wheat field. Restoring either one means ending its current economic function.

The Seven Sources

A. Degraded Natural Land Uplift — approximately 500 GtCO₂. Approximately 20–25 million km² of logged forests, overgrazed woodland, altered fire regimes, and drained wetlands. Mo et al. 2023 (Nature) identified 138 GtC of conservation potential in existing degraded forests alone. Restoring these lands to full ecological health does not require relocating anyone. It requires ending the economic activities that currently degrade them. Rate: approximately 50% of full cropland-to-ecosystem rate (FAO data). Peak: approximately 2–4 GtCO₂/yr.

Confidence: Reasonable Inference.

B. Pasture Restoration — approximately 1,300 GtCO₂. Approximately 29 million km² (90% of global pasture, retaining approximately 3 million km² for TC agriculture). Pasture is the least efficient food use of land and is released first. Requires TC food systems to replace grazing in each adopting region before land is transitioned. Temporal lag from adoption: 8–15 years. Peak: approximately 13 GtCO₂/yr.

Confidence: Informed Speculation.

C. Cropland Restoration — approximately 860 GtCO₂. Approximately 14.4 million km² (90% of global cropland, retaining approximately 1.5 million km² for TC Zone 2 surface agriculture). Released last — requires underground food systems fully operational. Temporal lag: 12–20 years. Peak: approximately 7.2 GtCO₂/yr.

Confidence: Informed Speculation.

D. Enhanced Weathering — approximately 250 GtCO₂. Coordinated deployment of crushed silicate minerals across adopting regions. Full rate: approximately 2 GtCO₂/yr (upper Beerling et al. 2020 range). Tu et al. 2026 projects 0.7–1.1 GtCO₂/yr by 2100 under realistic adoption. Full rate defensible under TC conditions with unlimited clean energy and no cost constraint. Phases in over decades.

Confidence: Reasonable Inference.

E. Concrete Carbonation — approximately 140 GtCO₂. Available after TC city construction begins (years 30–200). CO₂ mineralization into building materials. Carbonation curing at 45% efficiency with no strength loss (Fu et al. 2024). Huang et al. (2025, ESSD) found current cement materials absorb 0.86 GtCO₂/yr globally from approximately 300 billion tonnes of stock. Ongoing: approximately 0.5–1 GtCO₂/yr from mature infrastructure.

Confidence: Informed Speculation.

F. Ocean Recovery — approximately 175 GtCO₂. Central estimate of 100–300 GtCO₂ range. Ending industrial-scale harvest allows marine biological populations to recover. Sustainable, managed harvest resumes once populations are healthy. Rate: approximately 0.5–1 GtCO₂/yr.

Confidence: Informed Speculation.

G. Existing Forest Sink Continuation — approximately 500 GtCO₂. Currently approximately 8–10 GtCO₂/yr (Global Carbon Budget 2024). Once emissions reach zero, this becomes pure drawdown. Diminishes as atmospheric CO₂ drops and forests mature. Estimated: approximately 400–600 GtCO₂ cumulative over the first 100 years.

Confidence: Reasonable Inference.

Not counted: Forest protection (approximately 600–800 GtCO₂ stored) is avoidance, not drawdown — essential but does not reduce atmospheric CO₂. Direct Air Capture (DAC) has a legitimate role in late-stage drawdown (years 400+) and as an emergency buffer, but is not counted as a primary tool.

Total Drawdown Capacity

Several of these tools are late-stage. Concrete carbonation requires TC cities to be built. Enhanced weathering requires coordinated industrial deployment. Ocean recovery takes decades to centuries. During the early decades of the transition, ecosystem restoration and existing forest sinks do the drawdown work. SRM does the temperature work. Each tool arrives when the transition makes it available.

Biome-Disaggregated Sequestration

Total active restoration area: approximately 51 million km². Rates are landscape averages adjusted for natural disturbance regimes (fire, windthrow, storms, landslides). Disturbance is part of healthy ecosystems, not a refutation of restoration-based sequestration.

Sources: Cook-Patton 2020, Conant 2017, Loisel 2014, Günther 2020, Alongi 2014, Pan 2011.

Tropical forests contribute 40–50% of peak drawdown, making equatorial regions the mathematically correct location for first demonstration communities and early adoption.

Published comparison: Lewis et al. (2025, Nature Geoscience) found 96.9 GtC total through 2100 at 15–30% restoration. TC restores substantially more land. Super-linear effects (improved water cycling, wildlife-mediated dispersal) roughly offset diminishing returns from less productive biomes. Net approximately neutral. Scaling brackets TC central estimates.

Confidence: Informed Speculation.

Composite Sequestration Curve

Peak net drawdown: approximately 6–9 GtCO₂/yr under baseline conditions, or approximately 8–14 GtCO₂/yr with automation acceleration. With automation (human and robotic assist in years 0–15, robotic primary in years 15–40, full autonomous from years 40–100+), timeline compression of approximately 40–50% brings full restoration area to years 60–100 rather than 150.

Negative Feedbacks

Arctic methane: approximately 5 GtCO₂e/yr (range 2–8). Permafrost thaw already triggered. Persists 1–2 centuries after peak temperature (UNFCCC 2025). Over 30% of the Arctic-Boreal region is now a net source (ICCI 2025). Sources: Schuur et al. (2022), Hugelius et al. (2024), Yuan et al. (2024).

Ocean outgassing: approximately 4 GtCO₂/yr (range 2–6). Carbonate buffer chemistry resists equilibrium change. Currently oceans absorb approximately 2.9 GtC/yr (Global Carbon Budget 2024). This reverses as atmospheric CO₂ drops.

Combined feedback emissions: approximately 9–14 GtCO₂e/yr. These persist for decades to centuries and are the primary reason net drawdown is dramatically lower than gross drawdown during the critical middle period.

Dedicated Underground Algae — Eliminated

Rigorous analysis showed energy cost of underground algae at climate scale is prohibitive. Same energy produces 10 times more removal via DAC. Preserved as documented elimination.

Part Five: Solar Radiation Management

SRM is not optional. Without it, feedback emissions accelerate past the capacity window and the math does not work at any plausible Perturbation Airborne Fraction (PAF). SRM provides the thermal bridge that keeps ecosystems viable during the centuries required for restoration to draw down atmospheric CO₂. SRM and restoration always appear together — SRM without restoration is a band-aid on a wound that never heals. Restoration without SRM loses the race against feedbacks.

Marine Cloud Brightening (MCB)

Salt spray into low marine clouds. Immediate deployment. Core coral and Arctic protection. Feingold et al. (2024, Science Advances): 31-scientist consensus roadmap. Harrison et al. (2020): proof-of-concept on Great Barrier Reef. Latham et al. (2013): sub-global MCB ameliorates coral bleaching. Lower governance complexity than Stratospheric Aerosol Injection (SAI). Continues indefinitely as ecosystem support. Available for regional correction of precipitation impacts from SAI.

Stratospheric Aerosol Injection (SAI)

Global temperature bridge. Deployed years 0–2. Two deployment pathways: conservative (existing aircraft first) and accelerated (parallel platforms from year one). Both achieve meaningful cooling within 5–10 years. Duffey et al. (2025, Earth's Future): low-altitude high-latitude SAI feasible with existing aircraft. Smith (2020, Environ. Res. Lett.): approximately $18B/yr per degree Celsius avoided. Purpose-built fleet: 1–2 decades.

Alternative SRM compounds beyond sulfur aerosols are under active investigation. Calcite (Keith et al., Harvard) neutralizes stratospheric acid rather than creating it. Titanium dioxide (TiO₂) and aluminum oxide are also candidates. Each compound has different optical properties, atmospheric chemistry interactions, and potential side effects. A comprehensive comparison of non-sulfur SRM compounds is flagged as a research need.

Each additional half-degree of cooling roughly doubles deployment because it counteracts more existing warming, not just preventing future warming.

SAI Handoff to L1 Shades

Years 100–120, with 20–40 year overlap. Spaceline operational by approximately year 25–35, lunar manufacturing by approximately year 40–60, shade deployment gradual. L1 operational phase-down rate: 0.1°C/decade optimal, 0.3°C/decade maximum safe (Parker and Irvine 2018). L1 retained as precision instrument for permanent fine-tuning.

Moderate Overshoot Cooling

The TC proposes moderate radiative forcing overshoot: SAI and L1 shades cool 0.5–1.0°C below natural trajectory during peak drawdown (years 30–200). Not cooling to pre-industrial — controlled over-correction to accelerate ocean heat release. Cooler atmosphere steepens thermal gradient, speeds ocean-to-atmosphere heat transfer.

Trade-off: Reduced photosynthetically active radiation may slow ecosystem growth. Monsoon sensitivity: SAI must avoid disrupting monsoon patterns. Underground food reduces agricultural monsoon dependency. Hemispheric symmetry reduces risk.

Confidence: Frontier Territory. No published literature addresses over-cooling during multi-century drawdown.

Termination Shock

If SRM stops suddenly, temperatures snap back. Restoration must reach self-sustaining drawdown before SRM phase-out. This is the key coupling variable between SRM and restoration.

Under a growth-based economy, SRM is moral hazard — it masks consequences while CO₂ builds. When SRM stops, termination shock is worse because CO₂ accumulated. TC SRM is categorically different: restoration heals underneath. A band-aid while the wound heals.

The Positive Feedback Loop of Cooling

Every half-degree of SRM cooling reduces feedback emissions and improves carbon dioxide removal efficiency (Henry's Law: cooler water holds more dissolved CO₂). Less outgassing means more restoration sticks. Faster drawdown enables more cooling. This is the first feedback loop in the climate crisis that runs for restoration rather than against it.

It only works if healing is happening underneath the SRM.

Part Six: How Much Room Depends on Speed

The Resistance Problem

Carbon removal is harder than carbon emission. When CO₂ is removed from the atmosphere, the oceans and soils respond by releasing some of the CO₂ they had previously absorbed. This partially counteracts the removal effort. The Earth system resists drawdown more strongly than it resists accumulation.

The Perturbation Airborne Fraction (PAF) captures this resistance. A PAF of 0.65 means that for every tonne of CO₂ removed from the atmosphere, only 0.65 tonnes stays removed. The remaining 0.35 tonnes is replaced by outgassing from oceans and soils as they rebalance. Published PAF values range from 0.4 to 0.7 over multi-century timescales (Jones et al. 2016 via NASEM 2018). Removal efficiency decreases as total removal increases (Tokarska and Zickfeld 2015). The response is asymmetric (Zickfeld et al. 2021).

This gives us the core equation: Total Removal Needed equals the Atmospheric Deficit divided by PAF.

PAF is the single most consequential variable in this analysis. A small change in PAF produces an enormous change in how much land must be restored — and therefore how much room remains for people.

PAF Under Cooled Conditions

Under SRM, surface temperatures are held lower than they would otherwise be. Cooler ocean surfaces dissolve and retain more CO₂ — a well-established physical relationship known as Henry's Law. Oceans that are cooler should outgas less during drawdown, meaning a higher proportion of each removed tonne stays removed. Working assumption: PAF under SRM improves to 0.65–0.70, compared to 0.45–0.60 without SRM.

Confidence: Frontier Territory. Supported by first-principles physics but no empirical data exists. No published literature addresses carbon dioxide removal efficiency under SRM conditions. This is the highest-priority target for external validation.

An important caveat: this improvement is not immediate. The deep ocean retains accumulated heat for centuries. SRM cools the surface, but the ocean's thermal mass means that PAF improvement phases in gradually as ocean temperatures decrease. In the early decades of drawdown, PAF may be closer to published values than to the SRM-adjusted estimates used here.

This caveat also strengthens the case for early SRM deployment. The sooner cooling begins, the sooner each tonne of removed carbon stays removed — and the more room remains available for people.

Buffer

A 30% buffer is applied to total removal needed throughout this document. It accounts for feedback emissions exceeding central estimates, sequestration rates falling below modeled values, transition timing delays, SRM deployment delays, food system underperformance requiring more surface land, and unknowns that have not yet been identified. Several of these variables are correlated — higher temperatures simultaneously worsen feedback emissions, reduce sequestration rates, and delay SRM effectiveness. All numbers include the buffer unless otherwise noted.

Current State (April 2026)

CO₂: approximately 429 ppm (Met Office 2026 projected annual), 53% above pre-industrial 280 ppm. 1 ppm equals 7.82 GtCO₂ (CDIAC). Total GHG emissions: approximately 55 GtCO₂e/yr. Fossil CO₂: 38.1 GtCO₂/yr (2025, record). Temperature: 1.44°C above pre-industrial (2025 WMO). Three-year mean 2023–2025: 1.48°C. Remaining 1.5°C budget: approximately 170 GtCO₂ (approximately 4 years). Rate of CO₂ increase accelerating: 2024 saw 3.75 ppm, largest single-year jump on record.

CO₂ Target: 280 ppm

The target is restoration of atmospheric CO₂ to the pre-industrial baseline of 280 ppm. This is the concentration at which Earth's climate systems operated in stable equilibrium for millennia before the industrial era. Below-280 analysis — exploring whether temporary sub-baseline drawdown could accelerate deep ocean heat dissipation — is a separate study to be conducted in 10 ppm increments as marginal cost analysis.

Scenario 1: Gradual Cascade

First adoption approximately 2036. Phased adoption over 30 years.

Initial adoption begins in the mid-2030s. The transition spreads gradually across regions over three decades, with holdout nations continuing to emit until approximately 2066.

2026–2036 (pre-adoption): approximately 40 GtCO₂/yr. CO₂ rises approximately 27 ppm to approximately 456 ppm. 2036–2046 (partial adoption): approximately 30 GtCO₂/yr average. CO₂ rises approximately 20 ppm to approximately 476 ppm. 2046–2056 (accelerating): approximately 20 GtCO₂/yr. CO₂ rises approximately 13 ppm to approximately 489 ppm. 2056–2066 (completing): approximately 10 GtCO₂/yr from holdouts. CO₂ rises approximately 7 ppm to approximately 496 ppm. Early sequestration from first adopters partially offsets late-stage emissions, shaving approximately 10–20 ppm off the peak.

Central estimate for peak CO₂: approximately 500 ppm, around 2060–2070.

Confidence: Informed Speculation.

Atmospheric deficit: 220 ppm × 7.82 = approximately 1,720 GtCO₂.

Without SRM, the math fails. With SRM at mid PAF (0.65), the TC model fits, but the room beyond it is thin — approximately 3% of habitable land for cultural landscapes, expanded communities, or transition flexibility. At high PAF (0.70), room expands to approximately 5%. Every percentage point of room matters.

This 40-year cascade also approaches the limits of the societal stability window. Converging analyses of food system disruption, infrastructure degradation, and political fragmentation suggest that the capacity for sustained international cooperation degrades significantly within 15–30 years under continued climate pressure (Herrington 2021, Kemp et al. 2022, Nebel et al. 2024).

Scenario 2: Accelerated Cascade

A climate-vulnerable early adopter by approximately 2031. Near-universal adoption by 2040.

Climate-vulnerable nations — low-lying coastal countries, nations losing agricultural capacity, regions facing existential climate threats — have the strongest motivation for early adoption. A nation in this position could plausibly begin transition by the early 2030s, creating proof of concept that accelerates international adoption.

2026–2031 (pre-adoption): approximately 38 GtCO₂/yr. CO₂ rises approximately 13 ppm to approximately 442 ppm. 2031–2035 (early adopters): approximately 32 GtCO₂/yr. CO₂ rises approximately 10 ppm to approximately 452 ppm. 2035–2040 (rapid cascade): approximately 20 GtCO₂/yr. CO₂ rises approximately 6 ppm to approximately 458 ppm. 2040–2045 (phase-down): approximately 5 GtCO₂/yr. CO₂ rises approximately 2 ppm to approximately 460 ppm. 2045+: near-zero emissions globally. SRM holding temperature stable. Sequestration ramping up.

Central estimate for peak CO₂: approximately 462 ppm, around 2043–2045.

Confidence: Informed Speculation. Optimistic but physically plausible.

Atmospheric deficit: 182 ppm × 7.82 = approximately 1,423 GtCO₂.

Speed saves approximately 297 GtCO₂ of atmospheric burden compared to the gradual scenario. That is the quantified value of urgency.

At SRM mid PAF (0.65), approximately 9% of habitable land remains available beyond what the TC model requires — roughly 9.4 million km². At high PAF (0.70), the room reaches 11%. Under the accelerated cascade, the transition provides meaningful room for generosity beyond the core model.

Both scenarios arrive at the same structural conclusion. Speed does not change what the physics requires. It changes how much room remains for people on the other side. The faster this happens, the more generous the result.

What This Means for the Economy

Returning 79–90% of habitable land to a natural state requires ending or fundamentally transforming every major extractive industry. The scale of this transformation is the reason the climate problem has not been solved. It is also the reason a resource-based economy is structurally necessary.

A growth-based economy requires expanding resource extraction and expanding markets. This transition removes both simultaneously. Each adopting region removes resources from the extractive economy and consumers from the demand economy. This is true at every point in the 79–90% range.

The solution to climate change breaks money. That is not a political statement. It is what the carbon math requires.

Part Seven: Sea Level Rise

Permanent on any human timescale. TC floor approximately 2.0–2.2m above 2024 baseline. Coastal abandonment is rational. TC builds inland. Projections consistent with IPCC AR6 WGI, SROCC, WCRP (2022).

Part Eight: The Full Timeline

Year 0: Decision. MCB and SAI deployed. Grid decarbonization begins.

Years 0–30: Temperature still rising 0.3–0.5°C (underlying; SRM-masked surface may be held stable). Tell people now. Sea level: 3–5 mm/yr. Pastureland restoration begins.

Years 30–60: Emissions approximately 55 to 5 GtCO₂e/yr. Cities 30–50% populated. Major cropland liberation.

Years 60–120: Near-zero emissions. CO₂ at peak approximately 462–500 ppm (depends on transition speed; faster reductions yield lower peak).

Years 100–120: SAI hands off to L1 space shade array (20–40 year overlap transition).

Years 120–200: Active drawdown. CO₂ falling. Earliest restorations beginning to mature.

Years 200–400: CO₂ approaching 280 ppm pre-industrial baseline.

Years 400–700: CO₂ stabilizing near 280 ppm. Temperatures falling. Sea level plateau approximately 2m. L1 shades functioning as precision instrument. Ecosystems maturing.

Years 700–1,000+: Full planetary climate stabilization. Deep ocean equilibration. Sea level permanently elevated. Dynamic equilibrium.

Temperature Recovery

The ocean has absorbed more than 90% of excess heat (IPCC SROCC 2019; IPCC AR6 WGI 2021). Deep ocean equilibration: 500–1,000 years (Clark et al. 2016, Nature Climate Change). Surface air temperatures may not return to 1950s levels (approximately 0.2°C above pre-industrial) until roughly years 300–500. Extreme weather persists for generations. This confirms the necessity of thousand-year infrastructure.

Part Nine: Ocean Interventions

Artificial Upwelling (AU): Cold deep water pulses reduce coral bleaching. Sawall et al. (2020, Frontiers Mar. Sci.): less than 2 hrs/day significantly reduces stress. Not effective for carbon sequestration — may upwell dissolved inorganic carbon. Use: targeted coral protection only.

Whale pump: Restoring whale populations creates natural iron fertilization. Lavery et al. (2010, Proc. R. Soc. B): 50 tonnes Fe/yr from 12,000 Southern Ocean sperm whales. Net approximately 200,000 tC/yr; approximately 2M tonnes at pre-whaling levels. Ecosystem restoration strategy with carbon co-benefits — not primary carbon tool (Boor et al. 2023 caveat noted).

Iron fertilization: Mixed trial results. Supplemental, monitored, limited locations. Not primary strategy.

Part Ten: What We Do Not Yet Know

Honest gaps, in order of importance.

1. PAF under SRM. No published literature. This variable determines how much room remains for people — whether the realistic estimate is closer to 80% or 90% for nature. The PAF improvement phases in over time as ocean temperatures decrease, making the time-varying curve a critical research need.

2. Food system performance. TC Zone 1 underground yields are modeled, not demonstrated. Underperformance increases surface land requirements and reduces the room available.

3. Cascade timing. Every year of delay adds approximately 20 GtCO₂ to the atmospheric burden. The difference between the two scenarios is approximately 297 GtCO₂ — which translates directly into how much land remains for people.

4. Societal stability window. Converging collapse vectors suggest 15–30 years before cooperation capacity degrades (Herrington 2021, Kemp et al. 2022, Nebel et al. 2024). If only the accelerated scenario produces generous room, the urgency of beginning increases.

5. Holdout emissions. Non-adopting nations contribute an additional approximately 200–375 GtCO₂ under the gradual scenario.

Queued Analysis

Geographic cascade modeling with country-level land data. Time-varying PAF curve under SRM. Sub-baseline cooling scenarios. Multi-scenario SRM analysis at half-degree increments. Biome-weighted land restoration rates. Concrete stock estimation. SRM overshoot modeling. Monsoon sensitivity analysis. Short-term SRM compound alternatives (calcite, TiO₂, aluminum oxide, sea salt, and other non-sulfur approaches). Renewable energy long-term comparison.

Conclusion

The climate requires that approximately 80–90% of Earth's habitable land be returned to a natural, functioning state. The absolute floor is 79%. The realistic requirement is 85–90%.

The TC settlement and food system models provide every person on Earth with a generous quality of life within 10% of habitable land. The climate math leaves 10–20% available. That gap is room — room for more generous communities, cultural landscapes, and a transition that does not demand austerity.

SRM is non-negotiable. Without it, the math does not work at any plausible PAF. With it, the math works — and the faster SRM and restoration begin, the more room remains for people.

The faster this happens, the more generous the result. Under the accelerated scenario, 7–12% of habitable land remains beyond what the TC model needs. Under the gradual scenario, that room shrinks to 1–6%. Speed does not just improve the climate outcome. It directly expands the space available for human life.

At every point in the range, the land-use transformation required is incompatible with a growth-based economy. The resource-based economy is not an ideological preference. It is what the carbon math requires.

Three independent lines of analysis — carbon math, food science, and settlement engineering — converge on the same range. The convergence is the argument.

These numbers are first-order estimates with honest ranges and named gaps. They exist to be checked. Every serious challenge brought to them has made the framework more precise.

Version 12  |  April 2026  |  Working Document

From the Trust Collective Project

The thread grows brighter with every person who chooses it.