Carbon Impacts Tracer scales a single fossil fuel project's emissions into eight kinds of damage

The warming step is simple enough to check by hand. Browse's 1,597 megatonnes of CO2 ^[3] is 0.0016 trillion tonnes, and multiplying by the IPCC coefficient gives 0.00072 degrees C, the figure the tool reports ^[1]. ExxonMobil's projected 5,245 megatonnes gives 0.0024 degrees C the same way ^[2]. Run on Scarborough's 876 megatonnes, the same rule implies about 0.0004 degrees C. That warming figure is my calculation, not one the developers quote ^[3].

The "drop in the ocean" defence makes two claims: a project's emissions are tiny next to the global total, and the damage from any one project is almost impossible to track ^[2]. The tool concedes the first. Browse's lifetime emissions are about 0.07% of the roughly 2.4 trillion tonnes the world had emitted by 2019 ^[5][4]. It contests the second. Because warming rises almost linearly with cumulative emissions ^[4], a small share of emissions yields a small but countable share of harm, and the tool reports it in people and in square kilometres of glacier ^[1][13]. "This defense won't work anymore," the developers wrote ^[15].

The damage step is where the design choices sit. For each of its eight areas, the tool draws only on published studies strong enough to quantify how a rise in global temperature produces a specific harm, and the developers say that bar is why the list stops at eight ^[7]. I think that is the right bar for a number meant to be argued over. It also shapes the totals. The 484 heat-related deaths by 2100 in the Scarborough study are deaths in Europe, attributed to a gas project in Western Australia ^[11].

Event attribution, the field the developers say the tool builds on, asks how climate change altered the odds of a particular disaster or bout of extreme weather ^[16]. The Tracer answers a different question. It takes a known increment of warming and asks how much it adds to harms that published studies have already measured against temperature ^[6][7].

The thing this doesn't tell you is how wide the error bars are. The developers' account gives single figures, and the IPCC coefficient underneath them is itself an approximate value ^[4]. The input is a project's publicly documented emissions ^[6]. Every output scales directly with that input, so a disputed lifetime-emissions estimate becomes a disputed damage estimate.

The account comes from the tool's builders. Three of them worked on the Scarborough study, which researchers at six universities in Australia and the United Kingdom published last year ^[9].