The Carbon Bathtub — emissions, concentration, and who actually moves the needle

The level — atmospheric CO₂ (ppm)

Observed CO₂ (NOAA) If emissions had kept growing after 2013 Your future scenario Pre-industrial ≈ 280 ppm

—

ppm avoided by the post-2013 plateau

—

observed CO₂, 2025

—

2050 under your scenario

—

2050 vs pre-industrial

The faucet — emissions (GtCO₂/yr)

Actual emissions (GCB) Counterfactual growth Growth pressure, no levers Net, with levers + removal

—

your levers cut (% of global)

—

synth fuel displaces, 2050

—

net emissions, 2050

—

carbon removal in 2050

This is where your levers actually live. Each wedge bends the line down while it's rolling out — then it's spent, and the line goes back to following growth pressure. That's the compounding asymmetry: growth compounds forever, one-time cuts get eaten. The exception is substitution: a cheaper synthetic fuel takes a share of the fossil base, so when the base grows, the displacement grows with it — the one lever that compounds. The level (top chart) is the integral of this one.

The faucet's baseline

Counterfactual: kept growing at 2.2%/yr

The realized 2000–2013 trend. The red line is "what we avoided."

Growth pressure from 2026 +0%/yr

Demand growth — GDP, population, energy use — pushing emissions up if the transition stalls. It only goes up: decline isn't a dial here, because real decline is nothing but the sum of concrete interventions — the levers below. (An abstract "−X%/yr decarbonization" slider would double-count them.) Growth compounds; your levers must outrun it.

Levers phase in over 15 yrs

Nothing deploys overnight. Your behavior + systemic wedges ramp linearly from 2026 to full effect over this period.

Who changes their behavior?

Systemic levers (infrastructure, not lifestyle)

These aren't anyone's personal choice — they're how the world makes power, steel, and movement. Note the scale vs. everything above.

Substitution — synthetic fuel on solar

Oil + gas displaced by 2050 0%

The Terraform Industries play: cheap solar → direct air capture + electrolysis → pipeline-grade hydrocarbons. CO₂-neutral (it's circular, not removal), and it drops into existing pipes, furnaces, and engines — no downstream replacement. Modeled differently from every other lever on purpose: it's a share of the growing fossil base (oil + gas ≈ half of today's emissions), adopted on an accelerating cost-curve ramp. Because it's a fraction, not a fixed wedge, it's the one lever growth pressure can't eat. Caveats: doesn't touch coal, cement chemistry, or agriculture; assumes leakage stays low.

Carbon removal (a second drain)

Removal capacity by 2050 0 Gt/yr

Direct air capture, BECCS, enhanced weathering — ramps linearly from ~zero today to this rate in 2050. Reality check: novel CDR removes ~0.002 Gt/yr today; IPCC pathways lean on 3–12 Gt/yr by mid-century; past ~10 is speculative scaling. It can push net emissions below zero — watch how slowly the level responds even then.

What am I looking at? (and the honest caveats)

Real data: emissions are the Global Carbon Budget 2025 series (fossil + land-use), CO₂ is NOAA's global annual mean. The counterfactual gap uses the Joos et al. (2013) impulse-response carbon-cycle model — the same one Zeke Hausfather used — and reproduces his ~8 ppm. The future curve runs that model forward, calibrated to match observations in 2025. The pre-industrial line is the ~1750 global mean of 280 ppm — the level the Holocene climate (and all of agriculture-era humanity) ran at. Carbon removal is modeled as negative emissions on a linear ramp to your 2050 capacity; the Joos response treats removals symmetrically, which is roughly right at these scales.

Behaviour savings are developed-world per-capita figures (Wynes & Nicholas 2017). The regional availability factor is the honest, debatable part: a plant-rich diet or skipped flight saves little where people barely eat beef or fly. Numbers are illustrative — directionally real, not precise. (I left out the "have fewer kids" lever on purpose — it's contested and beside the point.)

One modeling rule keeps the books honest: decline is only ever built bottom-up from named interventions (the wedges + removal, each on a rollout ramp). The single exogenous dial is demand growth — an abstract "decarbonize −X%/yr" slider would just double-count the levers it abstracts over. Wedges and removal are additive (fixed Gt); fuel substitution is multiplicative (a share of the growing base, on a t² cost-curve ramp) — that structural difference, not a tuning choice, is why it alone keeps pace with growth.

Takeaway: PPM lags; emissions lead. And cutting is mostly about where the emissions already are — not about asking everyone, everywhere, to do the same thing.