Bug fix in MOM_open_boundary

[c2xu]

Fixed the inconsistency for defining the reference time of tides in MOM_tidal_forcing and MOM_open_boundary.

[Hallberg-NOAA]

This change looks correct to me, and it should correctly handle the case where the tidal reference date has not been set. However, I think that we should wait to formally approve this PR until after it has undergone the testing with CEFI regional configurations.

[Hallberg-NOAA]

These changes look correct and reasonable to me, and based on my inspection of the FMS time_manager code, I am confident that all cases that used the default reference date and did not include this fix would have failed. Although this does constitute a defacto change in the default behavior, I think that this is a reasonable change and one that will not break any existing configurations.

[Hallberg-NOAA]

This PR has passed pipeline testing at https://gitlab.gfdl.noaa.gov/ogrp/mom6ci/MOM6/-/pipelines/28888.