The forcing is astronomical (the Sun and Moon) and initialized in subroutine tides_set (mod_tides.F90). The only complicated part is the nodal corrections:
! argUMENTS and ASTROL subroutines SUPPLIED by RICHARD RAY, March 1999
We often use five tidal constituents, rather than the full eight available in HYCOM, because this allows a tidal analysis from 28 days of hourly fields:
2 'tidflg' = TIDES: tidal forcing flag (0=no;1=bdy;2=body;3=bdy&body)
2 'tidein' = TIDES: tide field input flag (0=no;1=yes;2=sal)
00011111 'tidcon' = TIDES: 1 digit per (Q1K2P1N2O1K1S2M2), 0=off,1=on
The tidein=2 option indicates that we are reading in the SAL constituents, i.e. we are using the observed SAL. In order to improve the S2 tide, ESPC-D also includes atmospheric pressure forcing.
If all 8 tidal constituents, are included the analysis requires about 6 months of hourly fields.
So ESPC-D-V01 inherited its five tidal constituents from the usual HYCOM with tides setup. It should have used all eight, and that oversight was corrected in ESPC-D-V02. Both include the Ngodock et al (2016) ASEnKF correction.
Note that ESPC-D-V02 is assimilating T & S & SSH, but it is not assimilating tides. For example. the deep water global mean RMS error vs TPXO10atlas (which does assimilate tides) is 2.02 cm and its Atlantic error is 3.37 cm. These are very good results for a non-assimilative global OGCM. See
Tidal Analysis of ESPC-D-V02.
For more information, see:
B.K.
Arbic, M.H. Alford, J.K. Ansong, M.C. Buijsman, E.J. Metzger, H.E.
Ngodock, J.G. Richman, J.F. Shriver, I. Souopgui, A.J. Wallcraft, L.
Zamudio, et. al. 2018
A Primer on Global Internal Tide and Internal Gravity Wave Continuum Modeling in HYCOM and MITgcm
Alan.