Thanks @sbrus89! I'll test this out for the case I was using this for earlier today and report back on what I find.

@pwolfram, just so you know, I haven't even complied this yet. I just wanted get something started.

Thanks @sbrus89, looking forward to testing it out now!

Previously, tidal signal was not being picked up when ssh_gradient was used. This fixes that issue and tidal forcing is now activated when ssh_gradient is used.

As discussed with @pwolfram, it may make sense to merge mpas_ocn_tidal_potential_forcing.F into mpas_ocn_vel_tidal_potential.F.

Also, this address #458.

Testing with new USDEQU240at120cr20rr5 case

• no wetting /drying
• config_pressure_gradient_type = 'ssh_gradient'
• config_dt = '00:01:00'
• config_time_integrator = 'RK4'
• config_mom_del4 = 4.0e8
• config_self_attraction_and_loading_beta = 0.09
• config_hurricane_min_depth = 5.0

Results are reasonable with the exception of the weird resonance near 10-23

@xylar and @mark-petersen, this is ready for your review, especially related to @sbrus89 's question about

Merge mpas_ocn_tidal_potential_forcing.F into mpas_ocn_vel_tidal_potential.F.?

My take is it is probably best to do this but would be good to get additional feedback on this issue.

I think it makes sense to do this, but definitely want to see what @xylar and @mark-petersen think. Just for context, right now:

mpas_ocn_tidal_potential_forcing.F contains all the routines to initialize the astronomical factors and compute the equilibrium potential forcing.

mpas_ocn_vel_tidal_potential.F computes the gradient of the equilibrium potential and does the tendency update.

Both have separate init routines, which is redundant.

Same as above but with config_pressure_gradient_type = 'pressure_and_zmid'

@pwolfram, are these with or without wetting/drying?

I don't see any reason you can't just combine those two files and avoid some redundancy. I have done something similar with land ice fluxes because momentum and tracer fluxes are interrelated.

@sbrus89, no wetting / drying. Need to keep the different concerns separated.

@sbrus89, can you also make ssh_gradient the default pressure for the hurricane cases? Please let me know when you are ready for me to review. I think it safe to merge the two files before @mark-petersen, @xylar and myself do the formal review prior to merge.

@mark-petersen and @xylar, @sbrus89 is running some final tests but this is now a good time for you both to do a code review prior to our merge.

Comparison between split explict and RK4 with ssh_gradient

Comparison between previous split explicit (with pressure_and_zmid), current split explicit (with pressure_and_zmid), and current split explicit (with ssh_gradient)

@pwolfram, I think everything checks out here. The SE and RK4 results with ssh_gradient are nearly identical, and the SE with pressure_and_zmid matches the previous results.

Interestingly for this case, it's a mixed bag on whether pressure_and_zmid or ssh_gradient gives better results. Some stations are improved, others are worse.

Thank you all for the hard work!