Many recipes, and in particular those ported from AutoAssess, will expect:
- one or more "test" simulations: the new simulation(s) to evaluate, and
- a single "reference": a baseline simulation against which to compare the test(s)
Initial development of CMEW has standardised a single model development run as the experiment, and used CMIP simulations as the control.
- As a model developer,
- When I specify two model runs in CMEW
- I should be required to identify that one of them is the reference, and
- CMEW should ensure that the reference run is used as the reference by any ESMValTool recipes that use this concept
When there are only two model runs, this is addressed by #142. This user story is more general, for when more than two runs are included.
Many recipes, and in particular those ported from AutoAssess, will expect:
Initial development of CMEW has standardised a single model development run as the experiment, and used CMIP simulations as the control.
When there are only two model runs, this is addressed by #142. This user story is more general, for when more than two runs are included.