Im confused of DSTART, TIDE_START, TIME_REF in Time Stamps section.
I am using tides from OTPS which I believe has a time origin of the year zero AD. Your TIME_REF can be zero, but you need to know how much time is between that time and the tidal origin time. There's a terse description here
. The ROMS clock at some time origin TIME_REF. The tidal model also has a time origin TIDE_START. Those time origins are not necessarily at the same time. Here's what I use:
DSTART = 36893.0d0 ! days
TIDE_START = -693962.0d0 ! days
TIME_REF = 19000101.0d0 ! yyyymmdd.dd
So my time origin is at the start of the 20th Century and the tidal one is something like 1900 years before that. My starting time for this run is about 101 years after my time origin.
Then there's the long-period corrections. The phase and amplitude of the eight main diurnal and semi-diurnal tides can be thought of as wobbling a bit about their base phase and amplitude. These wobbles can be computed with Mike Foreman's code, which some use to provide zeta and ubar as boundary conditions, rather than as tidal constituents. My version of ROMS has a stab at doing this internally but it's not part of the ROMS trunk.