Thread by @hausfath

Climate models are complex physics-based simulations run on the worlds fastest supercomputers.

But there is a quick hack to get results similar to climate models that you can calculate in Excel. Its called the transient climate response to cumulative carbon emissions (TCRE). 1/9

TCRE is the amount of warming as a function of cumulative CO2 emissions. It turns out that warming over time is – to a first order approximation – linearly proportional to cumulative CO2 emissions, as this figure (SPM.10) from the recent IPCC AR6 demonstrates: 2/

As the AR6 notes, "Each 1000 GtCO2 of cumulative CO2 emissions is assessed to likely cause a 0.27C to 0.63C increase in global surface temperature with a best estimate of 0.45C."

Essentially, if you add up CO2 over time and multiply it by 0.45C, you get warming. That simple. 3/

You can use this for lots of simple analyses without needing any fancy supercomputers. For example, here is how flat, rising, and falling emissions translates in temperatures (you can download the excel model here: docs.google.com/spreadsheets/d…). 4/

Or this analysis showing how the pathway to net-zero influences the total warming experienced (excel model here: docs.google.com/spreadsheets/d…). You can also use the "likely" TCRE range to estimate temperature response uncertainties. 5/

Now, as with any simple model this comes with some important caveats. The biggest one is that TCRE only works for CO2; non-CO2 forcings like other GHGs (CH4, N20), aerosols, etc cannot be as easily modeled. So a TCRE-based analysis will only reflect changes to CO2. 6/

There are some simple ways around this: you can, for example, look at the difference between cumulative CO2 through a particular date in your analysis and a SSP scenario (say, SSP2-4.5), and calculate the implied difference in warming outcomes. 7/

This effectively uses assumed temperature changes from non-CO2 components of underlying scenario (e.g. SSP2-4.5). For example, given flat emissions from 2021-2100 TCRE alone says we should have 2.6C warming in 2100. If you use the difference method from SSP2-4.5 you get 2.9C. 8/

Its limitations notwithstanding, TCRE is a great tool for doing your own climate analyses. You can get global temperature change estimates broadly consistent with the results of supercomputer models in under a second on your laptop, and use them for all kinds of analyses. 9/9