CMIP7 Data Request: v1.2.2.2 now available

On this Page

This is an old version of the Data Request. Go to the latest version here.

We are pleased to release v1.2.2.2 of the CMIP7 Data Request. Today’s patch release provides a minor updates to the six variables from the v1.2.2.1 release.

The key components of the Data Request that you can access from today are:

If you would like more information about the new structure of the Data Request for CMIP7, or some background information about how the Data Request was developed, you can find this at the bottom of this page.

This release would not have been possible without the immense amount of work contributed by the Thematic Author Team, the Data Request Task Team, and the Controlled Vocabularies Task Team.

The Data Request content

The CMIP7 Data Request specifies data requirements for CMIP7 experiments with a list of variables, with associated metadata, and mappings between scientific opportunities, groups of variables and groups of experiments. In earlier versions of the Data Request, the list of experiments was limited to the CMIP7 Assessment Fast Track. The latest version of the CMIP7 Data Request for the Assessment Fast Track (v1.2.2.2) includes additional experiments outside of the Fast Track.

The Data Request contains a subset of the variables which were requested in CMIP6, alongside a suite of new and modified variables. We have outlined some of the updates users can expect from both the CMIP6 Data Request and earlier version of the CMIP7 Request.

CMIP7 MIP tables

The CMIP7 MIP tables have been generated using the CMIP7 data request content and are available for modelling centres to start testing.

Content notes

New content notes for v1.2.2.2

What has changed?

There was an error in the root names for five Physical Parameters (areacella, areacello, areacellg, areacellr, modelCellAreai). These have been fixed in v1.2.2.2. This affects six variables.

New content notes for v1.2.2.1

Branded names

Branded variable names have been generated with version 0.12.0 of the branded variable mapper. A branded variable identifies a physical quantity and how it is sampled, and the data request additionally specifies the frequency, region, and grid type for which it is requested. For example, tas_tavg-h2m-hxy-u is the near-surface air temperature. If requested at monthly frequency, for the whole globe, on the atmospheric grid, it corresponds to the Amon.tas variable from CMIP6. A new CMIP7-era compound name has been derived to determine a unique name for each variable in the Data Request. It combines the branded name, the frequency, the realm, and the region requested. Therefore, the CMIP6 variable name Amon.tas has become atmos.tas.tavg-h2m-hxy-u.mon.GLB.

Provisional experiment names

Following discussion within ScenarioMIP community we are uncertain as to the naming scheme for scenarios. The data request has been updated with “place holder” names while we wait for confirmation of the final naming scheme. In previous versions of the data request experiments with ids starting esm-scen7 and scen7 will be replaced with a numbered entry (e.g. scenario1, scenario2). Note that this change also affects AerChemMIP and and HighResMIP experiments where the naming references a ScenarioMIP experiment.

An update to the data request will be released following confirmation of these names.

Regional variables

A number of variables are requested for the 30S-90S region. However if modellers are already providing the global version of these variables, there is no requirement to also produce the regional version.

Cell measures

Cell measures will not be variable attributes in the CMOR tables, which will be based on branded variables that are grouped by modelling realm (branded variables do not specify a variable’s grid). Grid information, along with frequency and region, is specified by the data request for each requested variable. This grid information includes the cell_measures, and modelling groups can choose to specify the cell measures in their output netcdf files (but it will not propagate automatically into the files from the CMOR tables).

Changes since v1.2.2

GitHub Issues (here and here) summarise the changes made since v1.2.2. If you would like a summary of what has changed, there is a tool within the API which can be run in the compare_variables.py command line tool. A file summarising the changes will be provided in due course.

Existing content notes

New variables for v1.2.2

The total number of variables in the v1.2.2 was slightly higher than in v1.2.1. This is due to a slight change in structure in preparation for the Variable Register the CVs Task Team is preparing. This now requires each variable in the Data Request to have its region specified. There were approximately 130 variables which were requested both globally and for the region south of 30 degrees. These variables now appear in the request separately.

Additionally, there were 6 variables identified as duplicates of each other. These duplicates have been removed.

CMIP7 Compound Names and Branded variable names

A new way of uniquely identifying variables has been proposed for CMIP7 onwards. This is known as ‘branded variable names’. The “branded variable” name is constructed from two components described below: a root name associated with the standard name defining the property, and a branding suffix, which indicates how the property should be sampled spatially and temporally. The branding suffix is composed from four components, referencing the variable’s temporal sampling, vertical sampling, horizontal sampling and area type.

A full documentation of the branding algorithm will be uploaded here shortly.

The branding variable algorithm development has been led by Karl Taylor (PCMDI/LLNL) and was adopted by the WIP in a formal decision on 14th January 2025.

Sea-ice variables on atmosphere and ocean grids

There are a small number of ocean variables requested on either the atmospheric grid (i.e. the cell methods are areacella) or both the ocean and atmospheric grid (i.e. there are two versions of the variable, one with cell methods areacella and one with areacello). The Oceans and Sea-ice Theme have confirmed this is an intentional decision, required to compute boundary fluxes.

Experiments outside of the CMIP7 Assessment Fast Track

A small number of experiments outside the CMIP7 Assessment Fast Track have been added to allow users to request data on the Fast Track timeline. These include experiments from TipMIP, FireMIP, and HighResMIP2, among others.

Time slices/subsets

Since v1.0, there has been discussion surrounding the use of time slices/subsets in the Data Request. Multiple modelling centres highlighted that they can be extremely difficult to implement, often meaning that requests for specific time slices/subsets are ignored. However, the Task Team still understands their use is appealing to some modelling centres in order to reduce particularly high volume requests. It was also noted that the term ‘Time Slice’ is used with different meanings in different communities (e.g., to refer to model simulations run under repeated annual cycle forcings).

Therefore, three decisions were made, impacting the CMIP7 Data Request:

Time slices are renamed to time subsets from v1.1 onwards.
Only particularly high-volume Opportunities are recommended to add time subsets. These time subsets are optional for modelling centres to implement, but provide a choice for centres if there is a need to reduce the volume of data produced.
Currently there are details about how to implement the time subsets in the Opportunity Technical Notes.

A smaller list of time subsets is also being used in CMIP7 to simplify the request.

Baseline Climate Variables

The Baseline Climate Variables have been updated to match the final version, v1.4, as per the accepted manuscript in GMD: https://egusphere.copernicus.org/preprints/2024/egusphere-2024-2363/.

Changes in CMIP6 variables

While most of the variables from CMIP6 remain unchanged, there are a few which have been edited. We have automatically flagged the variables which might have had one of the following fields edited since CMIP6.

Table/physical parameter (i.e. has the variable’s compound name changed)
Cell methods
Cell measures
Dimensions
Units
Positive direction

The updates to these fields have mostly been made to fix errors in CMIP6 variables, to add extra detail in metadata, and sometimes to adjust variable definitions or sampling. Variables with changes have been flagged in the ‘Flag – variable change since CMIP6’ field. Where a variable is flagged, we recommend comparing to the CMIP6 CMOR tables to clarify the differences; however, you will find some information in the Processing Note which will guide you towards the differences.

Associating MIPs with Opportunities

Since the v1.0beta release, we have offered MIPs the chance to tag themselves to Opportunities. The options offered to MIPs are:

If an Opportunity will produce science which is essential for their MIP. This is indicated in the field ‘MIPs – high priority.
If the science produced by an Opportunity would be relevant for their MIP, but is not essential for their science goals. This is indicated in the field ‘MIPs – lower priority’.

Please note, we have only received responses from a few MIPs so this information is incomplete.

Making pressure levels more uniform

To make the Data Request easier to implement, the Atmospheric Author Team have proposed to move all daily and monthly data requested on the CMIP6 pressure levels plev8 and plev7h onto 19 pressure levels (plev19 from CMIP6). This was implemented from v1.0.

Sub-daily data requested from the downscaling and emulator communities are still requested on a smaller number of pressure levels, due to the potential data volume bloat a move to 19 levels could cause. Data on plev3, plev7c, and plev27 remain unchanged.

Accessing the Data Request

The Online Database

The full Data Request can be explored online at https://bit.ly/CMIP7-DReq-v1_2_2_2.

Many users will not have used Airtable, or navigated through the CMIP Data Request database before. While it may seem complex at first, Airtable is an extremely powerful tool which makes viewing complex data structures much more simple when you know how to use it. Therefore we encourage users to take some time getting used to the platform in order to make navigation easier for yourself in the future.

We have created some guides to help you navigate and use the CMIP7 Data Request database. Access the Airtable guides at https://bit.ly/CMIP7-DReq-guidance.

Access the CMIP7 Data Request Airtable

Access the Airtable guides

The GitHub Software Package

The Data Request Task Team’s Technical Implementation Subgroup (DR-TISG) worked to create Python code to allow users to interact with the CMIP7 Data Request. It will provide an API and scripts that can produce lists of the variables requested for each CMIP7 experiment, information about the requested variables, and in general will support different ways of querying and utilising the information in the data request.

Access the CMIP7 Data Request Software repo

Flat list of the variables

The new structure of the Data Request is centred around Opportunities, which allows for important description of, and justification for, inclusion of output variables in the Data Request. Additionally, the database structure of the Data Request allows for improved consistency of technical metadata across the many variables in CMIP. However, we appreciate that many Data Request users find a flat list of the variables useful. This flat list is available in two ways:

In the Variables tab of the online Airtable database. This can be exported to CSV format by clicking the MASTER button, then selecting ‘Export to CSV’.
In the GitHub, a JSON export of variable metadata from the Airtable database will be available shortly via the GitHub Software repo. For the v1.2.2.1 release, there are two files: in one file variables are indexed by CMIP6-era compound name (e.g., “Amon.tas”) and in the other by CMIP7-era compound name (e.g., “atmos.tas.tavg-h2m-hxy-u.mon.GLB”). The actual metadata for each variable (dimensions, standard name, etc) is identical between the two files.

Providing feedback

We would like to thank the numerous thematic author team members, MIP contacts, and modelling centre representatives who have provided feedback on the Data Request at any point during the process.

Despite the incredible amount of work and review the Data Request has undergone, there will inevitably be mistakes remaining in the request. The Task Team will shortly outline their preferred feedback mechanism on v1.2.2 to allow users to report issues. It is likely a further v1.2.3 version will be released in early Autumn, containing minor technical updates required to align the Data Request with ongoing infrastructure developments.

The new CMIP7 Data Request structure and more background information

The Data Request Task Team have developed an activity that works with community representatives to devise a controlled list of high priority variables that facilitate the majority of user needs. They proposed the following structure of the next Data Request to the CMIP Panel and WCRP-ESMO Infrastructure Panel (WIP). The Panel and WIP both approved the structure.

The Data Request includes variable definitions and the mapping of variables against the justification for the request expressed in terms of the opportunities that the data will generate. This will take the form of a controlled list of high priority variables that serve both the majority of user needs and create a harmonised set of data requests which balance scientific demand for data against modelling centre and infrastructure capacity.

The new CMIP7 Data Request Structure, showing how Experiment Groups and Variable Groups are linked to Opportunities.

The thematic author teams process

Five thematic author teams have been set up to develop the controlled list of high priority variables through through an IPO-supported and Data Request Task Team coordinated paper writing process.

An open call was conducted for authors and reviewers for the thematic papers:

Impacts & adaptation (call closed on 30 November)
Ocean & sea-ice (call closed on 01 March 2024)
Atmosphere (call closed on 08 March 2024)
Earth system (call closed on 08 March 2024)
Land & land-ice (call closed on 08 March 2024)

Appointed authors can be viewed here. The Author teams will be running open meetings and other engagement initiatives with modelling centres during this process.

Variable selection

Author teams as well as the wider CMIP community have been asked to define:

What data is requested, including CF metadata,
Why it is needed and why it is a priority,
Who will make use of it
How it will be used.

Alongside discussing individual variables, the following have been asked to define:

Experiment groups: non-exclusive grouping of experiments (e.g. ‘Fast Track’, ‘DECK’, ‘Scenarios’ etc.)
Variable groups: non-exclusive grouping of variables (e.g. monthly time slices of the baseline variables).
Opportunities: intended use-case/justification for one or multiple variable groups. Opportunities are also linked to relevant experiment groups.

Creating opportunities

Identifying opportunities helps to provide a structure to map variables against requirements. Each opportunity description conveys why this combination of variables and experiments is important and how variables and experiments contribute to impact.

Each opportunity includes:

A high-level description of the science and/or societal use and impact
A time slice to specify a block of years for which data is needed.
Experiments and variables that link to the opportunity. Some experiments and variables will be linked to more than one opportunity.

v1.0 launch event

The Data Request Task Team held a virtual launch event following the Data Request v1.0 release. The event was aimed at MIPs and Modelling Centres, though anyone from the CMIP Community was invited to register.

View the slides and recordings for the launch event here.

v1.2 launch event

The Data Request Task Team held a second virtual launch event following the Data Request v1.2 release. The event was aimed at MIPs and Modelling Centres, though anyone from the CMIP Community was invited to register.

During the session, the Task Team presented an overview of the request, including descriptions of the changes in structure since CMIP6, an overview of the information and tools available to you, and an introduction to the CMIP7 Data Request software package.