CF Roadmap: Time

[feggleton]

Topic for discussion

Topic for Discussion

At this year's CF Workshop we divided the roadmap preparations into multiple themes. Time was suggested as something we should discuss in terms of the future direction in CF. After a talk from Chris L (UKMO), there were some questions about time and how it is represented in CF. These were:

• Should we have other ways of representing time than are provided by UDunits syntax?
• How easy is it to map from UDunits time/calendars to other standards?
• Can we represent a wider range of non-standard calendars than currently? E.g. paleoclimate, non-earth based calendars?
• Do we want to separate time units and units of measure? (I.e. handle them with different software libraries?)

Although I am not personally leading this topic in the roadmap publication, I am just starting the discussion for now.

[chris-little]

@feggleton My first thoughts in response to your questions:
I see no reason to have time units separate from other UoMs. That said, (SI) seconds are reliable, but (calendar) days, weeks, months and years are all variable in length at some precision.
Calendars and coordinate reference systems best treated as different things.
There are plenty of conventional CRSs/UoM for time that are reasonably well defined: Millions of years ago, Unix time, Julian years and Julian days, with the advantageous characteristic of only one UoM for the time axis. Though Unix time and POSIX time have some ambiguities that most people are oblivious too.

I think CF support for non-Earth calendars and CRSs should be on the roadmap. A Mars calendar with Martian years and Martian days (called Sols) has been well used. And NASA has called for a Coordinated Lunar Time (atomic).

Notation to indicate which CRS or Calendar has been used is needed, preferably not just a CF notation, but something with the ubiquitousness of RFC3339 or ISO 8601.

I still propose to register T+hh type CRSs with OGC.

[feggleton]

@japamment wrote the questions above so she might want to comment on this.