New empirical tide solutions have been computed for the global ocean, following methods going back to Schrama & Ray (1994). Global high-resolution assimilation methods have markedly advanced over the past decade, and it is now difficult for purely empirical methods to improve models such as FES2014 and TPXO.9 -- at least until several years of SWOT data are available. Nevertheless, some improvements are obtained in polar regions and for minor tides such as J1 (which in FES2014 was a purely hydrodynamic model without data assimilation).

There may be some interest in the tide prediction software package associated with GOT5, which is a complete rewrite of earlier software with enhanced capabilities. For one, the package can handle netCDF formats (as in FES2014), native OTIS binary, and the older GOT4 ascii format. It can even handle (as sledgehammer to as pin) a simple list of harmonic constants for a single location. Options at initialization can turn nodal modulations on or off. Also on option, minor constituents can be inferred or not, using either piecewise linear admittance functions or Munk-Cartwright Fourier functions. For inference, the software keeps track automatically of which tides need to be inferred and which are input. Constituents can be identified with a standard Darwinian name, or with a numerical 7-digit (extended) Doodson number which completely defines a tidal argument. The latter can thus handle any obscure compound tide of any order. Although written in Fortran-2008, some other languages (e.g., Julia) can call the software directly.