number of tidal constituents

[Luana F. Bueno]

Hi Alan, 

Hope you are fine!

I am running an experiment with tides for the Irish Sea. 
The results are very good, the currents velocities fit pretty well to the data, since the dynamics are mains driven to the tides. However, the  tides amplitude from the ssh filed are overestimating some points and underestimating others. 

I have already ran 2 different experiments with HYCOM version 2.3.01, one with and without Self Attraction variation for each constituent  and the results are pretty much the same. 

Now I am considering running with more than 8 constituents, but I was looking the mod_tides.F90 and I can change the maximum value for ncon, but I am not sure if the others scripts are able to consider more than 8 constituents. Is it possible to run with more than 8 constituents? 

Besides that, do you think that I can get better results considering also the drag coefficient? Any other idea how I could get better results for the tide amplitude?

Many many thanks Alan. 

[Alan Wallcraft]

Regional models generally get accurate tides through their open boundaries.  Astronomical forcing and SAL are often not significant and may not be necessary.

HYCOM is hard-wired for 8 constituents.  So ncon should always be 8.  It would be possible to add more constituents, but only with non-trivial source code updates.

Tides in relatively shallow regions can be tuned by varying the quadratic drag coefficient.  If cb is negative in blkdat.input, it is replaced by a spatially varying  array read from cb.a.  The GOMb0.08 expt_01.3 example case reads cb.a, generated by cb_07_10mm.csh.  This is for a fixed roughness (z0), but see Bottom friction optimization for a description of one method of tuning cb.  Note that this is using a related but different formula for cb than the one used by hycom_botfric.  If you actually have a maps of z0, you could try defining cb.a from the map, although it isn't necessarily the case that the most accurate tides come from the observed z0.

Alan.

[Alan Wallcraft]

Here is what cb_07_10mm looks like.  It is stronger in very shallow water.

[Luana F. Bueno]

Hi Alan, 

I ran the model with the CB file build with the same range of yours. However, when analysing the results (taking into consideration SAL and/or CB file), the changes were pretty small in the amplitude, but it did affected the current fields a lot (which were pretty great when comparing to in situ data), making them a lot smaller comparing to the simulation with fixed SAL (0.06) and CB (2e-3).

I did not imagine that this factor would change the velocity that much, but it does make sense, doesn't it?

Many thanks, Alan. 

