Improving vertical temperature stratification

[lcbernardo]

Hi everyone. I have been working on my ROMS simulation for a part of southwestern Japan for quite some time now, but I could not seem to properly reproduce the vertical stratification, at least in comparison to the Global HYCOM model I use for the boundary forcing (please see attached PDF file, as well as the *.h and *log files). I have tried various strategies already, such as changing the number of layers, changing the vertical mixing schemes, turning off the tides, and shifting the grid eastward so that my western boundary no longer partially lies on land, but I could not seem to get good temperature stratification, especially in the northern half of the domain.

Now I know that a challenging feature in this area is the passage of the Kuroshio Current, and hence I am thinking that perhaps nudging to a climatology may be the next strategy to try, but I'm not sure if the problem lies elsewhere in the existing setup. The rx1 criterion also is seriously violated, but I wonder if this could really be a big factor influencing the large difference with the HYCOM trends. I'd like to ask if anyone has any advice regarding the settings I am using which may help improve the simulation outputs. Any help would be greatly appreciated!