CELLO poster

My newest plume video:

This model setup uses the MITgcm tutorial_deep_convection test case with the following modifications:

• Resolution: $2$ m $\times$ $2$ m $\times$ $2$ m
• Domain: $198$ $\times$ $198$ $\times$ $198$ grid cells
• Initialisation: Temperature and salt profile from mooring time series (two-day mean around September 4, 2021) with a normalised WOA September climatology used to fill at each depth level (temperature is prevented from freezing and static instabilities are homogenised to ensure $N^2 ≥ 0$). This results in a staircase and other sharp points in the potential density profile which are smoothed out using a short integration with very high diffusivities.
• Initial velocities are $0$ m s$^{-1}$ in all directions
• Timestep: $2$ s
• Equation of state: TEOS-10
• Coriolis: $f_0=-1.358 \times 10^{-4}$, $\beta=8.194 \times 10^{-12}$
• Boundary conditions: no_slip_bottom and no_slip_sides are both .FALSE., with no tracer relaxation sponges
• Surface forcing: Constant salt flux of $0.03$ g m$^{-2}$ s$^{-1}$ and heat flux of -$200$ W m$^{-2}$
• Surface temperature limitation (i.e., allowFreezing=.TRUE.): Tfreezing=-1.85 (i.e., a wee bit of supercooling)
• Non-hydrostatic solver: cg3dMaxIters=40
• Advection scheme: tempAdvScheme and saltAdvScheme are both $33$ (and diffusivities are all $0$)
• Viscosity: therein lies the question… for this run, viscC2Smag=3.1 and viscAz=1.E-4, but this may need to change in future iterations

There are still lots of improvements to be made, but this is my current best effort!