For a fair comparison of models, the verification should only be done for
dates and locations that are common to all models. common_cases
takes
a harp_list object as input and then identifies and filters to only those
cases that are common to all of the forecast models in the harp_list object.
By default this is done with the SID, fcst_dttm and lead_time columns, but
extra columns can be added via ...
. If one of the columns is a vertical
coordinate ("p", "z", "ml" for pressure, height and model level
respectively), that column will also be included.