So you want to work with climate model data
As a newcomer to climate model outputs, I needed to figure out where to access climate model data. And for something as vital to the future of the world as the climate, finding these data has been astonishingly hard. Here’s the problem: there are hundreds of models purporting to measure or account for different effects. Some of them are just to see whether we’re able to model some parts of the earth system. You’re lost in a maze of huge files and websites written for insiders. You want a big flashing neon sign that says USE THESE DATA and DOWNLOAD HERE. The purpose of this article is to be your neon sign.
Where is the data?
The world’s gold standard climate models form the basis of the IPCC reports. Naturally, there is more than one choice for such a model and scientists compare them in the Climate Model Intercomparison Project (CMIP). The most recent CMIP is CMIP6. You should use CMIP6 models because they are the best-studied models and they represent scientific consensus.
CMIP6 models are hosted in multiple locations, but you need only one, which is good, because there’s only one location that is possible to be accessed by the general public. You should use the version hosted by Google at the Google Data Marketplace. Here’s the link. The data are stored on Google Cloud at this direct link.
The data are organized into ‘model intercomparison projects’, with names like HighResMIP and Scenario-MIP.
Which MIP should I use?
CMIP6 encompasses several model intercomparison project[s]’ which share the same or similar data and scenarios. Here’s a list of them. You’ll want to use Scenario-MIP, which refers to the high-level scenarios that the IPCC wrote about. Not all MIPs model all of these scenarios. If you’re a specialist, you might want other data, but then why are you reading this guide? This guide is for people who need a flashing neon sign.
And which scenario?
There are eight standard scenarios in the IPCC report. They reference the five standard scenarios from the previous IPCC report, which are designated Shared Socioeconomic Pathway 1 through 5 (SSP1-SSP5). They are suffixed with the radiative forcing in W/m^-2 that the pathway uses. The full set of pathways in use in CMIP6 are variations on these five:
- SSP1-1.9: This one is already impossible
- SSP1-2.6: A conservative pathway which represents sharp emissions cuts.
- SSP2-4.5: An intermediate pathway. Use this pathway unless you have a particular reason not to.
- SSP3-LowNTCF: An intermediate pathway in which we specifically mitigate ‘near-term climate forcing’ – like SOx and methane – according to SSP1
- SSP4-6.0: A more aggressive intermediate pathway.
- SSP3-7.0: The standard ‘we do nothing’ pathway from CMIP5
- SSP5-3.4-Overshoot: We do SSP5-8.5 through 2040, realize our mistakes, and then take aggressive climate action.
- SSP5-8.5: A more aggressive pathway in which we do nothing
They are described in full in this paper.
And where is the data?
If you click through to e.g. scenario ssp126, you’ll find about seven
directories with names like Amon and SImon. These refer to monthly (mon)
variables. The standard, human-interpretable variables are in Amon/, like
ta (ambient temperature). There’s a full list
here. ctrl+f for
Amon. There are a few other suffixes which may be available for other runs -
*day, *3hr, and so forth. Note that most weather variables are reported at
multiple levels in the atmosphere, but surface variables are suffixed with s:
tas is temperature at the surface, but ta is temperature at all levels.
Other good variables to know:
va: northward windua: eastward windhus: specific humidityhur: relative humidityts: surface temperaturesfcWind: near-surface wind speed