Improvements in generalisation of experimental data load. Any combination of experimental datasets of any numbers of ensembles can be loaded, either 2-D or global mean variables, either stored in one file per ensemble per start date or in on file for the whole run. The experiment with greatest number of members must be the first in the list.

Merge branch 'develop-generalised-load' of https://gitlab.cfu.local/cfu/s2dverification into develop-generalised-load