Inexact restarts

[ElizabethYankovsky]

We have run sets of cases where one simulation stops and restarts at 30 days, whereas the other simulation is run continuously. After 30 days, every set of simulations we tested has a different flow field between the restart and continuous case. The simulations we tested included cases with/without alkalinity forcing and with/without Orlanski boundary conditions. There is an additional problem when NOT using Orlanski BCs where the restarted simulations do not match up. We compared a forced, no Orlanski case with an unfroced, no Orlanski case that were both restarted at 30 days and their flow fields diverge. This issue may be saved for later as we consider the Orlanski case first. Again, for the Orlanski case our goal is to have the restarted and continuous cases have identical velocity fields post restart.

This error arose when running simulations with/without alkalinity forcing for the purpose of Ocean Alkalinity Enhancement testing in the East Pacific (25km resolution). We need to be able to have identical flow fields between the control and forced cases so that we can assess the influence of our alkalinity forcing intervention. Attached are a few slides detailing the tests. Restart_issue.pdf

[nmolem]

See my reply to issue 4.

[TomNicholas]

(@nmolem if you type hash then the issue number (i.e. #4) when you refer to another issue github will automatically link them together, which is very useful, like this #4)