If your intervals are contiguous, time_bounds[i,1] should equal
time_bounds[i+1,0] (See the beginning of section 7.1 in the CF spec.) So for
January 2013, you should set the endpoint to 2013-02-01 00:00:00.
(The resulting potential for confusion is why I think it's best practice to put
the time coordinate at the midpoint of the interval rather than either end.)
There's been some discussion of time resolution on this list, but nothing
that's made it into the spec.
Cheers,
--Seth
On Fri, 15 Mar 2013 23:12:27 +0100
Andreas Hilboll <lists at hilboll.de> wrote:
>Seth, thanks for this nice walkthrough! I still have one question,
>regarding 3):
>
>for Jan 2013, I would chose 2013-01-01 00:00:00 as the start point, but
>what about the end point?
>
>- 2013-01-31 23:59:59
>- 2013-01-31 00:00:00
>- ...
>
>Is it possible to somehow set the time 'resolution'
>
>Cheers, Andreas.
>
>
>> 1) Set the time values to the midpoint of the time interval.*
>>
>> 2) Set a "cell_methods" attribute on the data variable with a
>> value of "time: mean (interval: 1 month)".
>>
>> 3) Create a time_bounds variable with dimensions (time,2)
>> whose values are the start and end points of each time interval.
>>
>> 4) Set a "bounds" attribute on the time coordinate variable
>> equal to the name of your time_bounds variable.
>>
>> *CF allows you to put the time coordinate anywhere in the
>> interval, but in my experience at the midpoint is the best option,
>> as it is the most intuitively obvious and least likely to cause
>> problems if you ever create aggregations over longer periods.
>>
>> Cheers,
>>
>> --Seth
>>
>> On Fri, 15 Mar 2013 22:45:39 +0100
>> Andreas Hilboll <lists at hilboll.de> wrote:
>>> Hi,
>>>
>>> I want to store monthly averages of gridded satellite measurements of
>>> atmospheric trace gas columns. I'm wondering how I should specify the
>>> time axis, i.e. set the time for each monthly aggregate to the first, or
>>> the central, or the last day of the month. How should I specify that the
>>> values are actually monthly averages?
>>>
>>> Thanks for your ideas!
>>> Cheers, Andreas.
>>> _______________________________________________
>>> CF-metadata mailing list
>>> CF-metadata at cgd.ucar.edu
>>> http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
>>
>
Received on Fri Mar 15 2013 - 16:56:16 GMT
