Priorities for large MSU experiments

Copying the google doc here so we can continue the discussion started at the COSIMA2022 workshop.

Starter questions:

  1. What are your dream model runs for next year? In 3 years? In 10 years?
  2. Paleo vs. high-resolution
  3. Longer BGC runs? e.g. RYF with BGC
  4. High-res regional vs global?
  5. Coupled high-res ACCESS-OM-01 simulations with UM atmosphere as per ESM1.5.
  6. ACCESS-OM3 - do we want to do IAF cycles as for OM2?
  7. Waves and tides?
  8. How much of the saved data is unused? Should we track this?
  9. Are there variables or time frequencies that would be really useful to have output for?

I’ll add the ideas we had as separate posts.

ACCESS-OM2-01 0.1° RYF with BGC - long spinup

1000yr ACCESS-OM2-025 BGC IAF spinup - Matt C’s accelerated adjustment?

MOM6 with COBALT biogeochemistry?

MOM6 experiments at high resolution to tweak vertical coords etc and confirm performance/realism - test in regional configs?

ERA5 forcing in ACCESS-OM2

ACCESS-OM2-01 IAF cycle with meltwater perturbation as basal melt

MOM6 Panantarctic + ice cavities at 1/20deg

Coupled high-res ACCESS-OM-01 simulations with UM atmosphere as per ESM1.5

Ryan commented: Is the (international) ocean-only community going to stick with OMIP-style IAF cycles? Or would there be a case to transition to a different protocol (e.g. Huguenin et al. 2022,

Experiments with different sea level, e.g. RYF with LGM sea level (120m lower) at high or low resolution? Or higher sea level?

here’s an example of a 0.1° paleo experiment with late-Eocene topography - check out the movie:

Do you want people to thumbs up the ones they want to indicate their preference?

Sure. Also is there a way to have threaded comments on each of these suggestions? That’s the other reason I split them up.

Nope. Discourse, by design, does not do threaded replies. See this example of it action.

I think this is a key issue that we should discuss more. When do we delete data from old experiments on the common /g/data drives (that aren’t published elsewhere as part of publications)?

Agreed this is good to think about. Also useful to see if there are diagnostics that we currently save as standard but are never or only rarely used.

We had some discussion on tracking which files are actually used here: Track how often files are accessed? · Issue #231 · COSIMA/cosima-cookbook · GitHub
I think we need something like this if we are going to have a data life cycle where data eventually gets deleted / shifted to tape.

Potentially related: @dale.roberts has configured the analytics database such that it might be suitable for this use case

OK thanks. So if we reply to (or quote) specific posts Discourse will link them.