<blockquote type="cite"><p><span>Dear MITgcm Support Community,</span></p><p><span>I am running coupled atmosphere–land–ocean experiments based on the </span><em><span>cpl_aim+ocn</span></em><span> configuration.</span><br><span>My goal is to validate the model setup using a fully coupled PI-type experiment </span><strong><span>without SST/SSS relaxation or prescribed surface fluxes</span></strong><span>, before proceeding to paleoclimate simulations. The experiment is conducted with </span><code><span>aim_fixed_pCO2 = 285e-6</span></code><span>, following a PI control configuration.</span></p><p><span>As
 the simulation progresses (after about 500 years), some parts of the 
ocean (e.g., the Southern Ocean and the Mediterranean Sea) become 
unrealistically fresh, as shown in the attached figure, and North 
Atlantic Deep Water (NADW) formation appears to cease.</span></p><p><span>I would appreciate guidance on the following points:</span></p><ul data-spread="false"><li><p><span>Are there known issues or common pitfalls in </span><em><span>cpl_aim+ocn</span></em><span> that could lead to excessive ocean freshening?</span></p></li><li><p><span>In
 PI-control fully coupled experiments, is some form of SST/SSS 
relaxation generally required to maintain stable NADW formation?</span></p></li><li><p><span>When SST/SSS relaxation is turned off, are there specific parameters in </span><code><span>data.land</span></code><span> and/or </span><code><span>data.aimphys</span></code><span> that should be adjusted to avoid such behavior?</span></p></li></ul><p><span>Any suggestions on likely causes, useful diagnostics, or where to focus debugging efforts would be greatly appreciated.</span></p><p><span>Thank you very much for your time and support.</span></p><p><span>Best regards,</span><br><span>Fisher</span></p>
        </blockquote><br>