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

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

First approach to a generalised Load() with a configuration file mechanism. Documentation still to be updated, not supporting multimember observations, not supporting atmospheric variable means and should recode the disable of extreme values for some variables.

Update pdf 2.2.0

Upgrade to 2.2.0

Increased the version number according to SEMVER.
Fixes issue 11.

Merge develop into master

This is an important merge.
After this we all will be able to follow the new developing strategy.
Please @nevensf could you confirm in the discussion below that the features 'Load ice', 'Load obsonly' and 'Plot stereo map' work properly?
Please @cprodhomme could you confirm in the discussion below that the feature 'Plot ACC' works properly?
To confirm that a feature works you can write in a comment 'Feature_name 👍'

Thanks a lot!

Merge removal of unneeded library calls

This is useful so that the package checks successfully in amdahl and moore.
Checked by myself (minor modifications).

Color bar

The addiction of the new parameter cex has been tested by Acacia Pepler and myself. If @vtorralba also agree to test it it would be nice.

Documentation corrections

Some corrections in the sample data documentation

Develop acc

Replacing MR !4
Already tested by Chloé

bugfixe: reverse order of the confidence interval, lower was in position 3 and higher was in position 1