Why the Social Cost of Carbon is Six Times Higher Than We Think

Plain-language explainer for doi:10.5281/zenodo.20653285 (#403)

The central idea in one sentence

Standard climate economics treats each tipping element (Arctic ice, Amazon rainforest, West Antarctic ice sheet, …) independently; this paper proves that once you account for the triangular interactions between them, the social cost of carbon jumps from ~$51/tonne to ~$316/tonne.

The problem with “bilateral” climate models

Current integrated assessment models (IAMs) compute the damage from climate change by asking: how much does each tipping element cost if it tips? They add up the answers. This is the $H^0$ (bilateral) approach.

But tipping elements interact. The Amazon tipping makes Sahel droughts more likely. Arctic ice loss accelerates Greenland melting. West Antarctic ice and permafrost feedbacks are coupled. The $H^1$ (triangular) approach asks: can the bilateral damages be assembled into a globally consistent picture, or do the triangular interactions create irresolvable contradictions?

The mathematical finding

The paper applies sheaf cohomology — a tool for asking whether locally consistent data can be assembled globally — to the network of climate tipping elements.

The $H^1$ signal turns positive at $T^* \approx 1.8°C$ of warming — between the Paris Agreement targets of 1.5°C and 2.0°C. At this temperature, the triangular couplings between tipping elements become load-bearing: you cannot price the risks bilaterally and get a consistent answer. The correction factor is approximately $6\times$:

Approach Social cost of carbon
Bilateral ($H^0$, standard IAMs) ~$51/tonne
$H^1$-corrected (triangular interactions) ~$316/tonne

Why this matters for policy

The Paris Agreement targets were set without accounting for $H^1$ effects. The analysis suggests 1.5°C is closer to the true topological tipping threshold than policymakers may realise — and that the economic case for aggressive mitigation is much stronger than bilateral models indicate.

For the full technical treatment, see doi:10.5281/zenodo.20653285