Differences in vertical salinity structure between v2.2 and v2.3 (nested in GOMl0.04 expt 50.1)

[Alfredo Terrazas]

Dear HYCOM community,

I am comparing two simulations performed with HYCOM versions 2.2 and 2.3. Both configurations are nested in the GOMl0.04 experiment 50.1, using the same nesting strategy and comparable external forcing (CFSR).

At a specific analysis point, I computed vertical statistical profiles of salinity (long term minimum, maximum, mean, and standard deviation over time) for both simulations. I am attaching a figure showing the comparison.

The main issue is that the vertical salinity structure obtained with v2.3 differs significantly from the one produced by v2.2. The v2.2 simulation shows a vertical structure that agrees well with in situ observations available at that location (and some other with the same issue), while v2.3 departs from that structure in a systematic way.

To facilitate troubleshooting, I am also attaching the configuration files used in both cases (blkdat.input, forcing settings, vertical grid parameters, etc.).

My objective is to identify which parameters or processes I should examine first in order to recover, in v2.3, a vertical salinity structure closer to what I obtained with v2.2 (and to observations).

From your experience, I would appreciate guidance on:

• Key parameter changes between v2.2 and v2.3 that could affect vertical stratification.

• Modifications in vertical mixing schemes or default coefficients.

• Changes in relaxation, nudging, or boundary condition handling.

• Any known differences in sigma-layer behavior between these versions.

If there are specific diagnostics you recommend computing to isolate the source of the discrepancy, I would be very interested in applying them.

Thank you in advance for your guidance.

Alfredo.

[Alan Wallcraft]

The figure does not say which is red and which is black.

There is a major surface forcing difference:

<    3      'empflg' = E-P     forcing   flag (0=none,3=net_E-P,1-2,4-6=sst-bas_E)
---
>    4     'empflg' = E-P     forcing   flag (0=none,3=net_E-P, 1,2,4=sst-bas_E)

In 2.3 you are specifying E-P (empflg=3), which means that forcing.precip must be P-E rather than just P.  If your precip file is just P, this means that there is no evaporation term in E-P for salt forcing (there still is evaporation in the heat flux though).

See what happens if you change this to 6 (same as flxflg).

The only other issue I see is with turbidity:

112c111
<    2    'jerlv0' = initial jerlov water type (1 to 5; 0 for KPAR, -1 for CHL)
---
>    1    'jerlv0' = initial jerlov water type (1 to 5; 0 for KPAR, -1 for CHL)

jervl0 =2 is less clear water than =1, and is probably more realistic.  Better still would be -1 for CHL see seawifs_mon_chl.csh.

Alan.