[telemac2d][telemac3d] Fix securing HBOR to be positive (tide+TOMAWACT3D)

When modelling tide with tidal database different from TPXO (e.g. JMJ or "miscellaneous" e.g. FES, NEA or PREVIMER), there is no minimum threshold to 0 for HBTIDE (see BORD_TIDE or BORD_TIDE_MISC subroutines). So is when coupling TELEMAC-3D with TOMAWAC-T3D for HBOR (see BORD3D subroutine).

In TPXO module, HBTIDE%R(K) = MAX( 0.D0 , Z-ZF(NBOR(K)) ) is written. It is now postponed outside in TIDAL_MODEL_T2D and TIDAL_MODAL_T3D as done for other tidal databases directly for HBOR.

It will enable to do generic treatments when modelling tide independent from chosen tidal database (e.g. inverted barometric effect).