What's new in 2.3.00

Significant changes since 2.2.98:

a) Freeform Fortran 90.

b) Move the config/ directory (Make fragments) inside the HYCOM-src directory.

c) CESMBETA and NAVYESPC NUOPC cap examples. These are provided as examples of how to add a "cap" for coupling. Neither one is suitable for general use.

d) Equation of state is chosen at compile time through new Macros: EOS_SIG0 / EOS_SIG2 and EOS_7T / EOS_9T / EOS_12T / EOS_17T. See the Compilation (Make) wiki page for more details.

e) New macro for thermobaricity correction: set KAPPAF_CENTERED to use layer centered pressure in converting sigma to sigma-star instead of upper interface pressure.

f) E-P-R is now always treated as a water flux (previously a salt flux), and wtrflx is in the archive files. SSS relaxation is always a salt flux, and E-P-R is still applied as a virtual salt flux if epmass is not set true. Four new blkdat variables:
     empbal: E-P balance flag (0=none,1=offset,2=scale)
     emptgt: E-P balance target (mm/week, into ocean)
     sefold: e-folding time for SSS relaxation (days, -ve to input)
     sssbal: SSS rlx balance flag (0=none,1=offset,2=scale)
Use sefold=30.0 and sssbal=0 to recover the previous SSS relaxation bahaviour.

g) epmass now applies E-P-R freshwater flux to the top layer, so epmass=true gives a natural (mass flux) surface freshwater flux.

h) btrmas=true makes HYCOM exactly mass conserving, but this makes the model more costly to run. Normally (btrmas=false) it is approximately mass conserving. When mass conserving, epmass=true will exactly conserve salt although there is salt exchange with sea ice to account for. The archive field thknss is always dp', where sum(dp') = bathymetry, but now new archives contain oneta where dp = dp' * oneta so that dp can be calculated if needed. HYCOM now always uses dp (not dp') for layer-based region-wide sums and time means. To confuse matters, dp in the HYCOM source code is actualy dp' (and so should have been called dpp).

i) New Steric SSH option (sshflg=2) to make montg1 steric SSH. This is probably a better approach than sshflg=1, since then pbavg (bottom pressure anomaly) is non-steric SSH. As with sshflg=1, the split between steric and non-steric SSH assumes that the long term mean SSH is entitrely steric. So you need to read in the long term mean in relax.montg.[ab].

j) Addition of a 365 days no leap calendar (yrflag=4).

k) Energy loan model has bug corrections and the lwflg.eq.-1 logic.

l) The Asselin filter is now applied near the end of the time step.