The climate numbers behind the plans keep moving

Removal timing flips the ranking

When should carbon dioxide be removed from the atmosphere — early or late? In a preprint posted this week, researchers show the answer depends on the metric you use. The ranking of equal-total carbon dioxide removal schedules is reversed depending on whether the metric is endpoint concentration or time-integrated forcing. The same tonnes, the same total — and the schedule that wins changes when the yardstick changes. m:…2418

The trade-off is large. Moving 2,014 Gt of CO2 removal earlier in a schedule adds 59 to 85 W m⁻² yr of integrated forcing benefit. But later removal of 2,014 Gt of CO2 results in 18 to 30 ppm more CO2 benefit at the endpoint compared to earlier removal. Earlier removal wins on cumulative forcing; later removal wins on the final concentration. Neither schedule is better in general — only better on a given metric. m:…2418

A collapse that delays a collapse

Another preprint finds a similar reversal of intuition in the ocean and ice. A collapsed Atlantic Meridional Overturning Circulation — AMOC — state delays the tipping of the Greenland Ice Sheet by 160 ppm CO2. The cooling that follows an AMOC collapse slows Greenland's melt, so the failure of a tipping element pushes another's threshold further away. Cooling elsewhere can slow melting in another place, and the system does not respect the neat boundaries that plans are drawn around. m:…1047

The authors quantify the stabilising effect as equivalent to 0.9°C of global warming. That does not make AMOC collapse good news — it is a major disruption in its own right — but it means Greenland's fate cannot be read off a global temperature number alone. The thresholds interact, and a plan that treats them as independent will misjudge both. m:…1047

The methane inventories are low

Another preprint turns from models to measurement. Using satellites to characterise urban methane emissions, researchers find that landfills contribute 7% to 53% of total urban methane emissions across the ten studied cities — a wide range that reflects how different cities' waste systems are. High-resolution Bayesian inversion using TROPOMI and GHGSat data improves spatial attribution of urban methane sources, pinning emissions to specific sites rather than spreading them across a grid. m:…2100

The standard inventories come out low by comparison. EDGAR v8 systematically underestimates grid-level landfill emissions compared to GHGSat observations. Cities planning methane cuts from inventory numbers may be aiming at the wrong total — and the satellite data suggests the landfill share is both larger and more variable than the books say. An inventory that undercounts its largest source is not a small error; it is a plan aimed at the wrong target. m:…2100

What this means for plans

The papers share a warning for anyone writing a climate plan: the number you plan against depends on choices that are rarely stated — which metric, which threshold, which inventory. A removal schedule that looks best on cumulative forcing looks worse on endpoint concentration. A Greenland threshold that assumes a stable AMOC is wrong if the AMOC collapses. A methane plan built on EDGAR v8 may undercount its biggest urban source. m:…2418 m:…1047 m:…2100

None of the papers says planning is impossible — only that a single headline number is not a plan. Report the metric alongside the number, test the threshold against the system it sits in, and check the inventory against what satellites actually see. The numbers are not wrong; they are incomplete without the choices behind them. m:…2418 m:…2100