Additional features

This release includes:

• GIR carbon cycle (only for CO2), eliminating dependence on scipy.root and speeding up runtime
• Two-layer temperature model from Held et al. (2010)/Geoffroy et al. (2013), allowing for ocean heat uptake to be diagnosed directly. Shout outs to Glen Harris and Zeb Nicholls
• ScmDataFrame reader, for compatibility with iiasa-climate-assessment and openscm-runner to ingest scenarios from IPCC Working Group 3

Nothing here should break back-compatibility but there is now a minefield of configurations to run fair_scm with and the interface became unmanageable a long time ago to the uninitiated. Sensible defaults will be packaged with openscm-runner, otherwise this will be sorted out in v2.0.

Finally, python2 is no longer guaranteed to work. Python 3.5 should work but 3.6+ is recommended for future compatibility with unit handling.

Add keywords to `fair_scm` call

For allowing user to modify hardwired CMIP5/AR5 emissions to forcing relationships.

RCMIP phase 1

Code freeze for RCMIP phase 1: www.rcmip.org

Minor fixes to aerosol, ozone and contrail forcings to specify a custom baseline.

RCMIP-interim

A few interim patches included in the trunk before the target RCMIP phase 1 release:

• Fixed Stevenson aerosol forcing in call from fair_scm
• API keyword to turn off scaling F2x in call to fair_scm (we don't want this by default now, as AR6 F2x may not be the same as AR5)
• Add AR6 gases molecular weights and radiative efficiencies (not included in multigas run yet)
• ValueError for attempting to infill CFCs before 1765

Inverse FaIR

This new release of FaIR allows diagnosis of CO2 emissions from concentrations. See the fair.inverse module for details (and also the Examples page at https://fair.readthedocs.io/en/latest/examples.html#running-in-concentration-driven-mode).

Restart consistency

The behaviour of restart_out and restart_in has been formalised. This allows the carbon cycle and ocean thermal response state from one run to be saved and used as boundary conditions for a second run.

One internal book-keeping change: inside FaIR, CO2 concentrations are now an absolute value rather than a difference from pre-industrial.

FaIR rename; repo restructure; GWP tool

FAIR has been renamed FaIR: the acronym has changed to reduce confusion with an existing environmental policy decision model called FAIR.

Repository has been updated to include code coverage reports, pull request / issue templates, ReadTheDocs yaml and Makefile.

A Global Warming potential calculator has also been added.

No changes to science code from earlier in the 1.3 series.

Concentration driven functionality / ensemble generator / bug fixes

Bug fixes:

• #40: some data files not included when installing
• #39 (part of): historical forcing scaling factor default time series did not have enough years

New features:

• #39: concentration driven (and forcing driven) runs now possible
• #38: ensemble generator with joint lognormal ECS/TCR distribution provided

Small updates

Here #34 (a tool to calculate steady state emissions) and #36 (consistency in RCP naming) have been addressed. Technically this should be a minor release, rather than a patch as #34 adds functionality - we'll do semantic versioning from v2.0 I promise - but unless anything is broken the v1.3 series indicates (to me, at least!) that nothing should have changed from the GMD paper.

Multigas "final" release

Default aerosol treatment has been changed since the v1.2 series:

• organic carbon, sulfate, nitrate and VOC follow their exact radiative efficiency from Aerocom
• black carbon's radiative efficiency is lowered, to account for rapid adjustments (semi-direct effects), so that the default ERFari in 2010 is -0.45 W/m².

Black carbon on snow now has a default efficacy of 3, to match what is specified in Smith et al. (2018).

The option to add a time-varying scaling factor for CO₂ forcing in CO₂-only mode has been included.

Linked documentation has also been added (not included in packaged version - hopefully...).