According to the QC workshop discussion (12), it was decided to adapt the MEDATLAS 97 climatology to make it usable for the quality checks. The result should be 1x1 degree mean profiles of temperature and salinity and corresponding standard deviations.
The maximum value of the ¼ square degree standard deviation found in the 1 square degree box has been taken as the resulting standard deviation. However the results were not satisfactory because obviously too high or too low values were found in some sub-regions. It was therefore necessary to apply some QC on the standard deviations before the selection. As a result not all the means were associated with standard deviation estimates and interpolations/extrapolations have been necessary.