OBS for Lab Sea MLD are not consistent with what we assess in the model

[oceandie]

The OBS for the Lab Sea MLD that we use are mean and std (see MXL_lab_mean_obs.txt) while in the models we assess the time-evolution MLD in March - so the models will be always to deep than the obs value.

Possible solutions:

1. use MXL_lab_max_obs.txt instead -> however, since this is from clim, we won't have the std

2. compute the March MLD from EN4 monthly Dataset -> we will have both mean and std and potentially using the same averaging area of the models, but requires more computations!

3. use March MLD from different climatologies (WOA, EN4, ifremer) -> both mean and std, maybe less work than option 2, although probably the std will be close to 0.