Implementation of different calendar function without leap years

[TobiasKAndersen]

A scenario simulation with GOTM could be done by choosing forcing from one or several years as a period to repeat several times, thereby creating a longer scenario period built upon actual forcing data. But if the repeat period is not 4 years long, the scenario simulation will run into a problem with leaps years.
Therefore, the option of GOTM being able to run on a calendar without leap years would enable to run scenario simulations build upon repeating actual forcing data much more smoothly.

[TobiasKAndersen]

As I am not a contributor I cannot add labels, but would propose the label to be #enhancement

[bolding]

NetCDF has support for different calendars through the CFconventions - http://cfconventions.org/Data/cf-conventions/cf-conventions-1.6/build/cf-conventions.html#calendar.

So the main thing is changes to the GOTM time module - and support in flexible output to create files with the calendar attribute.

If we go in this direction we might consider introducing time zones as well.

The entire changes to the code is not neglible :-) And how do we treat observations with real date/time information in the case of alternative calendars.