[CF-metadata] representing model times in climate models

From: Steve Emmerson <steve>
Date: Fri, 30 Mar 2007 08:29:52 -0600

Jon,

The UDUNITS package is a poor tool for the handling of time in climate
model output. Whatever conventions you decide upon, programs will,
undoubtedly, have to "roll their own" code when it comes to the handling
of time.

Regards,
Steve Emmerson
UDUNITS Developer

Jon Blower wrote:
> Dear all,
>
> I'm sure this question must have been asked before but I haven't been
> able to find an answer. How should time be represented in CF-NetCDF
> files that represent model output that doesn't map to a real time -
> for example a climate model in which we are more interested in
> long-term trends of the model than in the precise time? Often the
> time axis is marked as "model years" and thus far we have got around
> this by assigning a fictive start date (e.g. the year 3000) to give a
> UDUNITS-compatible axis. What is best practice here?
>
> Also, some models use 360-day years, which I guess would confuse a lot of tools.
>
> Regards, Jon
Received on Fri Mar 30 2007 - 08:29:52 BST