[MITgcm-support] RBCS extending to bottom & shelf dry points in lake setup

[Jody Klymak]

Perhaps you can clarify this a bit by adding something indicating the water Depth to your plots?   It looks to me like you have set a Gaussian temperature fill, but the topography is actually just  a straight cylinder at the edges.

If that is true, then the fact that the water extends downwards is just convection because you have dense water over lighter (assuming you are using a non-linear equation of state).

I’m also confused by your Y axis only going from -3400 to -3150, whereas you specify YC from -3400 to +3400.

In all, I suspect you have gotten your dimensions confused, and have not set the bathymetry up the way you intended.

Cheers,  Jody



From: MITgcm-support <mitgcm-support-bounces at mitgcm.org> on behalf of Fateme Sharifi <fateme.sharifi1391 at gmail.com>
Date: Sunday, August 31, 2025 at 10:49 PM
To: MITgcm-support at mitgcm.org <MITgcm-support at mitgcm.org>
Subject: [MITgcm-support] RBCS extending to bottom & shelf dry points in lake setup

You don't often get email from fateme.sharifi1391 at gmail.com. Learn why this is important<https://aka.ms/LearnAboutSenderIdentification>

Dear MITgcm users,

I am setting up a small lake experiment with Gaussian bathymetry, a dry shelf band, and a shallow “moat” region (within 400 m of shoreline, depth ≤ 5 m) where I apply RBCS restoring toward 4 °C.

My input files are generated with a MATLAB script (attached) that writes:

  *   Depth.bin (outside/shelf = 0 m),

  *   hFacC.bin with proper partial cells,

  *   rbcs_mask.bin limited to shallow moat layers only (z_mid ≤ moat_depth, hFacC > 0),

  *   rbcs_Tr1_fld.bin set to 4 °C inside the mask.

At model start (t=1), the RBCS mask behaves correctly (restricted to the top ~5 m near the boundary). However, after just 2–3 timesteps, the RBCS influence appears to extend vertically to the bottom (see attached figure with Y–Z sections at t=1 and t=3). This was unexpected because the written rbcs_mask.bin only has ones in shallow moat layers.

I also noticed some sensitivity to shelf dry points: despite setting Depth=0 and checking that hFacC=0 in the shelf region, I worry that something in RBCS or partial cells may be activating those points during runtime.

My questions:

  1.  Has anyone seen RBCS restoring “leak” into deeper layers, even if the input mask is strictly shallow?

  2.  Is there an additional runtime parameter (e.g. in data.rbcs or data) needed to prevent the restoring from being applied throughout the column?

  3.  What is the recommended way to guarantee that RBCS remains limited to the intended top layers throughout the run (not just in the input mask)?

  4.  Any best practices for handling dry shelf points with RBCS in lakes or enclosed basins?

I have attached both my MATLAB setup script (setup_gaussian_lake_rbcs_revised_v2.m) and the figure (vertical_section_temperature.jpg).

Best regards,

Fatemeh Sadat Sharifi
Ph.D. Candidate at
The Leibniz Institute of Freshwater
 Ecology and Inland Fisheries (IGB)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.mitgcm.org/pipermail/mitgcm-support/attachments/20250901/e844bc34/attachment.html>