“…In addition, mesoscale and submesoscale oceanic variability, which must also be parameterized in a 1-D model, can significantly impact the mean and seasonal cycle of the MLD in some regions, like the Subantarctic Zone of the Southern Ocean (du Plessis et al, 2017;DuVivier et al, 2018;Fox-Kemper et al, 2011;Lee et al, 2011;Li and Lee, 2017;Swart et al, 2015). Indeed, high-resolution simulations that resolve some subseasonal atmospheric variability and ocean mesoscale variability show that (among other properties; (e.g., Delworth et al, 2012;Griffies et al, 2015;Kirtman et al, 2012;Maltrud and McClean, 2005;Maltrud et al, 2008Maltrud et al, , 2010McClean et al, 2011;Small et al, 2014;Winton et al, 2014) the mean MLD depends on the horizontal grid spacing of the atmosphere and ocean model at resolutions between 2 • and 0.1 Harrison et al, 2018;Lee et al, 2011;Sein et al, 2018), which suggests that both atmospheric and oceanic mesoscale processes (which vary on subseasonal time scales) have a substantial impact on the annual mean MLD. Hence, high-resolution three-dimensional (3-D) ocean circulation models are more suitable than 1-D models for simulating the dynamics of the climatological mean, seasonal cycle, and subseasonal variability of the MLD on a global scale.…”