Earth LinC station / a benchmark of common sense
Every year the world puts about 41.6 billion tonnes of CO₂ into a thin shell of air. Roughly half of it stays. This page is not a report — it is an attempt to check whether what we cover is actually what matters.
Total greenhouse gases released in 2023 — an all-time record, up 1.3% on the year before.
Share of our CO₂ that land and ocean still absorb for free. The other half accumulates.
Remaining budget for a 50% chance at 1.5 °C, from the start of 2025. About six years at today's rate.
A sustainability role inside supply chain and product is mostly a job of coverage. You draw a boundary, you count what's inside it, you improve what you can reach. And then one quiet afternoon the honest question arrives: do we cover the best of what we can cover — and even if we do, is our decarbonization still very limited?
The only way to answer is to zoom out until the company disappears and the planet's ledger is visible. Then zoom back in, and see where your hands actually are.
Writing is not just a record. Writing is thinking.— the brief for this page
Carbon is not a moral quantity, it is an accounting one. Sources in, sinks out, and whatever doesn't balance stays in the air and warms things. Here is the 2024 ledger, in gigatonnes of CO₂ per year.
Add methane, nitrous oxide and F-gases and the full basket reaches 57.1 GtCO₂e (2023). CO₂ is about three-quarters of it; methane about 17–18%; N₂O about 6%.
That residue is why concentration climbs about 2.4 ppm each year. Nature is subsidising roughly half our emissions — unpaid, unpriced, and not guaranteed.
Share of global greenhouse gases by activity. Note what this cut reveals: “energy” is not a sector, it is a verb inside every other sector. Electricity does not appear as its own line — it is folded into the industry, buildings and transport that consume it.
Sector shares: Climate Watch / WRI dataset, as presented by Our World in Data. Cut a different way — by food system rather than by smokestack — food accounts for about 34% of all emissions (Crippa et al., Nature Food, 2021). Both numbers are true. The boundary you choose decides what you see.
And on the supply side: a small number of producers dominate. The Carbon Majors database traces roughly 80% of global fossil CO₂ since 2016 to 57 producing entities. Concentration is a problem — and it is also leverage.
The half-that-disappears is the most underrated line in the whole ledger. It is not a policy, it is a favour — and the terms are quietly changing.
In 2023, record heat, drought and fire in the tropics and boreal north cut the land carbon sink to a fraction of its usual strength. A sink that behaves like weather cannot be counted on like infrastructure.
Ocean uptake has already lowered surface pH by about 0.1 — roughly 30% more acidic than pre-industrial. We are not storing carbon at sea, we are transferring it into a different problem.
A tonne burns in seconds and lingers for centuries; a fraction stays for millennia. Meanwhile engineered removal today runs at around 0.001 GtCO₂ a year — about one thousandth of one percent of what we emit.
Current policies point to roughly 3.1 °C this century. To stay on a 1.5 °C path, global emissions would need to fall about 42% by 2030. We are still setting records instead.
The system is not losing balance because one villain got worse. It is losing balance because the inflow grew for two centuries while the outflow stayed roughly the same size — and is now, in places, shrinking.
Ranked by assessed mitigation potential for 2030, drawn mainly from the IPCC AR6 Working Group III assessment, with IEA and UNEP figures where they are sharper. Percentages are against the 57 GtCO₂e we emitted in 2023 — so you can feel the size honestly. Roughly half of all this potential is available at under US$20 per tonne.
The cheapest electricity in most of the world, and it scales in months rather than decades. The bottleneck is no longer cost — it is grids, permits and storage.
Onshore wind is the other pillar of the same displacement: every terawatt-hour it makes is a terawatt-hour of coal or gas that never burns.
The single largest nature-based lever, and it costs less than restoring what we destroy. Four commodities — beef, soy, palm oil, timber — drive most tropical deforestation, and they all run through somebody's supply chain.
Reforestation, peatland and mangrove restoration, agroforestry, soil carbon. Slower, reversible, and impossible to skip.
The energy sector leaks around 120 Mt of methane a year. About 70% is technically avoidable with existing kit, and roughly 40% at no net cost, because the gas is worth money. Fastest near-term brake on warming that exists.
Heat pumps, insulation, efficient cooling, better appliances. Unglamorous, distributed across billions of decisions, and about a sixth of all emissions sits here.
Livestock is the biggest single source of agricultural methane. Assessed potential ranges widely (0.7–8 Gt) because it depends entirely on how many people change, not on whether the physics works.
Over a billion tonnes of food is wasted each year — around a fifth of what reaches consumers — carrying an estimated 8–10% of global emissions with it. Emissions already spent, for nothing.
More than one in five cars sold worldwide in 2024 was electric. The value only lands fully when the grid behind the plug is clean — sequence matters.
Together roughly 14–16% of global CO₂, and half of cement's emissions come from the chemistry itself. Clinker substitution, scrap-based electric steel, hydrogen, capture. The hardest tenth on this list — and the one that needs demand-side buyers most.
Assessed total across all mitigation options by 2030: 31–44 GtCO₂e per year — more than half of current emissions. The constraint is not the absence of solutions. It is deployment rate, capital, and who is willing to change a specification.
Two different questions. Which industries are big? — steel, cement, chemicals, food. Which are accelerating? — a very different list, and the one a product and supply chain role should be watching, because today's small line is next decade's structural problem.
These are the “hard to abate” industries for a physical reason, not a political one: they need extreme heat, they emit from chemistry rather than combustion, their plants last 30–50 years, and their products are traded as commodities where a 10% green premium loses the tender.
Which is exactly why the lever sits with buyers. A steel mill cannot justify a new process for a market that does not exist yet. Purchase orders create that market.
Around 1.5% of world electricity in 2024, projected by the IEA to roughly double toward ~945 TWh by 2030. The load is new, concentrated, and growing faster than clean supply in most grids.
Of global emissions, and back above pre-pandemic levels. Sustainable fuels remain a fraction of a percent of jet fuel used. No electrification path for long haul this decade.
Roughly 1 Gt a year, carrying 80%+ of world trade. Ships built today burn fuel in 2050. Every product with an ocean leg owns a slice of this.
Space-cooling demand is on track to multiply by mid-century as heat and incomes rise. A warming feedback made of appliances — and of refrigerant gases hundreds of times stronger than CO₂.
The IEA has called petrochemicals the largest driver of future oil demand growth. As fuel demand peaks, feedstock demand is meant to replace it — inside packaging, textiles and components.
Estimates vary wildly by boundary — which is itself the finding. When a sector cannot agree on its own footprint within a factor of four, it cannot yet manage it.
Here is the number that reframes a supply chain job. CDP finds that for the average company, supply chain emissions are around 11.4 times its direct operational emissions. Which means a perfect, fully renewable, zero-emission headquarters can still be a rounding error.
Coverage is the real work. Not because scope 3 is fashionable, but because that is where the tonnes physically are — in the steel of a housing, the polymer of a casing, the freight lane, the farm three tiers upstream that nobody has ever visited.
Most corporate footprints are built from industry-average emission factors, not measured data. Averages cannot detect improvement — which means a lot of reported reduction is boundary movement, not physics.
Offsets do not cancel a tonne emitted; at planetary scale the removal capacity to match our emissions simply does not exist yet. Reduce first, and be precise about what the residual really is.
The reading was the easy half. These are the ones still open — offered so that you add yours, rather than accept mine.
If half of our emissions are absorbed for free by land and ocean, and that half is weakening, should every target be measured against a shrinking allowance rather than a fixed one?
Our reported footprint uses average emission factors. If a supplier genuinely halved their process emissions tomorrow, would our number move at all? If not, what are we actually managing?
Which single specification in our product — one material, one thickness, one tolerance — carries the most tonnes? Do we even know its name?
Fossil methane abatement is 40% free and among the fastest brakes on warming. Why does almost no corporate roadmap contain a methane line?
We buy compute the way we once bought freight — invisibly, by the unit. What happens to our footprint when AI workloads triple?
Are we optimising the product we sell, or the product the world will actually need? Efficiency inside a growing volume can still add tonnes.
If our scope 3 is 11 times our scope 1 and 2, why is 11 times the effort not going there?
What would we have to stop making entirely — and who inside the company is allowed to ask that out loud?
None of these need an answer today. They need an owner, a date, and one measurable next step. That is how action flows.
Every figure above traces to one of these. Where estimates differ between bodies, the range is shown rather than the flattering end. Check them — disagreeing with a number is a better outcome than believing it.
Reporting years differ by source (2016 for the sector split, 2023–2024 for totals). Percentages are rounded. Nothing here is a forecast; it is a snapshot with its edges shown.