According to *this note*, I have made in the past two simulations with a moderate wind (MAG=500), low dissipation (KH=10 m2/s) but one with sponge (`exp2_w6_sponge`) and one without it (`exp2_w6`).

In this previous note, I show that the simulation with a weaker wind (MAG=100) and zero dissipation with a sponge layer (`exp2_l`), we were arriving to a quasi-steady Eulerian-mean U in the middle layer (lower panels of Fig. 1 in *this note*). On the other hand, `exp2_w6` and `exp2_w6_sponge` were run only up to day 2000 and I have been wondering if 1) we arrive also at a quasi steady Eulerian-mean flow in any of the two simulations and 2) if there is a difference between the two (perhaps, we do get the quasi steady state with sponge but not without it).

Fig. 1 shows the time series of Eulerian-mean U computed over 100 days for two locations in `exp2_w6` and `exp2_w6_sponge`, respectively. We see that we have not reached a quasi steady state even after 8000 days. Maybe we will later (because the wind is stronger) but that has yet to be demonstrated. We see, however, the variability in the mean is greater without a sponge.

To give an idea of the spatial structure of the Eulerian-mean flow, I give in Figs. 2 and 3 the maps of that component for two different periods, around day 2000 and around day 8000. We see that mean flow varies greatly between the two periods and between the two simulations.