Re: Question on the "cd_calendar" and CRUTEM4 nc data file?

This is a follow-up from a person at the CRU. The CRUTEM4 does
have an error in the time dimension units.

"The Met Office actually creates the netcdf version, I have spoken to
the person there that deals with CRUTEM and he was aware of the problem
and it was fixed last month. He was unaware that we hosted the netcdf
data on our website, so in the near future we will post the file
containing the fix (otherwise identical).

If you need the corrected version now, then please download from the MO
website."

Climatic Research Unit
School of Environmental Sciences
University of East Anglia

On 4/8/12 8:55 AM, Dennis Shea wrote:
> I believe the units of time on the file are incorrect.
>
> float t(t) ;
> t:units = "days since 1850-01-00 00:00:00" ;
> t:time_origin = "00-JAN-1850:00:00:00" ;
> ---
> Use of the 00 for the day is rather unusual.
> It would result in a date of 1849-12-31-00 which is what
> is being returned by cd_calendar. I sent an email to CRU
> regarding this issue.
>
>
> I think the units should be
> t:units = "days since 1850-01-01 00:00:00" ;
> t:time_origin = "01-JAN-1850:00:00:00" ;
>
> ---
> You can use the ncatted operator to change the units attributes
> on the file or you can do the following in NCL:
>
> t = f->t
> t@units = "days since 1850-01-01 00:00:00"
> t@time_origin = "01-JAN-1850:00:00:00"
>
> date = cd_calendar(t, -1)
>
> Good luck
>
> On 4/8/12 1:10 AM, apple wrote:
>> Hi, all
>>
>> Now I use NCL 6.0 on a macbook (X-Lion 10.7), and "cd_calendar" is a new function in NCL6.0.
>>
>> I download two NC files (HadSST2_1850on.nc and CRUTEM4.nc) from the CRU
>>
>> (http://www.cru.uea.ac.uk/cru/data/temperature/ ). When I do as follow
>>
>> " f1 = addfile (" HadSST2_1850on.nc","r")
>> f2 = addfile("CRUTEM4.nc',"r")
>>
>> t1 = f1->time
>> t2 = f2->t
>>
>> yyyymm1 = ut_calendar(t1,-1)
>> yyyymm2 = ut_calendar(t2,-1) "
>>
>> I get
>> " warning:ut_calendar: Invalid specification string. Missing values will be returned.
>> (0) HadSST2 : from 185001 to 201202
>> (0) CRUTEM4 : from -2147483647 to -2147483647
>>
>> However, When I use the " cd_calendar" to replace the " ut_calendar" , as the following
>>
>> " yyyymm1 = cd_calendar(t1,-1)
>> yyyymm2 = cd_calendar(t2,-1) "
>>
>> I get
>> (0) HadSST2 : from 185001 to 201202
>> (0) CRUTEM4 : from 184912 to 201011
>>
>> We can see one month lag at the beginning between two data , which is inconsistent with the instruction
>> on the webpage http://www.cru.uea.ac.uk/cru/data/temperature/ . According to the webpage, both data should
>> start at the same time,
>>
>>
>> Thanks!
>>
>>
>>
>> Lin
>>
>>
>>
>>
>>
>>
>>
>> _______________________________________________
>> ncl-talk mailing list
>> List instructions, subscriber options, unsubscribe:
>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
> _______________________________________________
> ncl-talk mailing list
> List instructions, subscriber options, unsubscribe:
> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
_______________________________________________
ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
Received on Mon Apr 16 09:49:45 2012