cold water creeping upwards along topography

[sonaljit.m]

Hello,
Attached, is a timelapse video of a temperature longitudinal transect from my simulation.
https://youtu.be/wELpEsKAvqc

This is the first 2000 timesteps in my simulation, with open boundary condition in the north and south walls (left and right walls in the video). The temperature data supplied to ROMS as open boundary condition, contains cold water (near 4 degrees) at the bottom, typically at depths 4 to 5 km.

What I find strange, is that within the first 2 thousand timesteps the cold water slowly creeps up along the topography and reaches the shallow regions where the depth is less than 200m. As a result I get wrong data at the shallow regions. My bottom heat forcing is 0.0, and at the top there is a diurnally varying shortwave time series. My net surface heat flux averaged over 1 year, is around 17 W/m^2 which seems reasonable.

Is anyone familiar with this problem? Any suggestions?

Thanks,
Sonaljit.

[sonaljit.m]

This is another timelapse video of a simulation with different initial condition: https://youtu.be/sFFqxAKvUHY

[kate]

Did a race condition just eat my post? I asked for details on your horizontal diffusion and suggested MIX_ISO_TS or MIX_GEO_TS.

[jpringle]

In your sections, the bathymetry has severe stair-stepping (discontinuities). Is that real?

Also, is there any along-isobath flow which could be driving a bottom Ekman transport?

Jamie