ACCESS-OM3 Evaluation
This website is the starting point for evaluating ACCESS-OM3 experiments. It brings together quick-look diagnostics of runs that are still going, a way to compare how two experiments are configured, and links to the full, permanent evaluation of finished experiments.
It has three parts:
- OM3 Monitor follows selected simulations while they are running, so problems can be spotted early and runs compared as they progress. It shows figures from a small set of evaluation notebooks, re-run on the latest output of each monitored experiment. Each experiment also has a status page built from the run's own payu records: recent runs, the model date reached, throughput (simulated years per day) and the latest model and PBS logs. The OM3 Monitor pages are regenerated nightly and replaced each time; nothing is archived. When an experiment stops being monitored, its pages disappear.
- Config Diff compares how two experiments are set up. Choose any two
experiments of the Experiments table and it lists the MOM6 parameters (from
MOM_input, with anyMOM_overrideapplied) that differ between them. It also shows how each experiment's settings changed during its runs: every edit to the model build, input files and settings, from the git history of its branch in access-om3-experiments. The data is read from those branches whenever this site is rebuilt. - Evaluation Summary is the permanent evaluation record. It lists the key ACCESS-OM3 experiments, one per row. To see an experiment's evaluation, click Evaluation in its row: that opens the fixed, versioned evaluation website for that run, with all its figures and notebooks. The table is taken directly from the Experiments page of the access-om3-configs documentation, and is refreshed whenever this site is rebuilt.
Spotted something wrong, or have an idea? See Contributing and feedback.