Open for having the different reference values in `ea_normalise` be defined with spatial data, and not just simple staic values.