Check the sensitivity of the domain_cfgs to tidal forcing.

[jdha]

Quoting Enda:
The change in the definition of the bottom may have a substantial effect on the tides,
And might be one of the places that the difference between z-s and Diego’s ME s is significant.

That might be one of the slowest parts of getting this domain right.
A simple run with tides and a straight up diff of SSH amp over a day once its spun up
might be sufficient in the first instance to observe differences in the model solutions.
Could even just be constant density ocean without any forcing, just the tides.

A short run with say just M2 say 10 days, and switching on the harmonic analysis to dump
out a M2 Amp/phase and then comparing the various set ups for the short run just in terms of sensitivity never mind accuracy in the first instant.

If they are not very different at that point can think about longer runs with multiple constituents
And comparing against obs but I have a feeling the difference could be significant so that
simple field difference of just M2 for a short run might be a good starting point and a cheap experiment.

[jdha]

Thoughts: M2, 10 days, harmonic analysis

as a very quick test perhaps go with sf_L51_r24, sf_L51_r10 (so we have an idea of the impact of rmax)

then in addition: sz_L51_r10_s26 (being quite ZPS-like) and MEs_L51_r10-07

Sound ok?

[endaodea]

Ran basic comparisons against SF12 51
constituents included:
clname(1)='M2'
clname(2)='S2'
clname(3)='O1'
clname(4)='K1'

started run 10-12-2013 ran out until March 2014. Assessed only the ssh data for JAN-FEB.

Experiments included:

• L51-sf12_0.24
• MEs_L51_r10-04
• sz_L51_r10_s21
• sz_L51_r10_s21 with CD0 = 1e-3
• sz_L51_r10_s21 with CDMAX = 0.05
• sz_L51_r10_s21 with NONLIN Bfric option instead of log layer
• sz_L51_r10_s21_with PARTIALSLIP insteasd of free slip
• sz_L75_r10_s39
• zp_L75_rXX