JiscMail Logo
Email discussion lists for the UK Education and Research communities

Help for DEVORAC Archives


DEVORAC Archives

DEVORAC Archives


DEVORAC@JISCMAIL.AC.UK


View:

Message:

[

First

|

Previous

|

Next

|

Last

]

By Topic:

[

First

|

Previous

|

Next

|

Last

]

By Author:

[

First

|

Previous

|

Next

|

Last

]

Font:

Proportional Font

LISTSERV Archives

LISTSERV Archives

DEVORAC Home

DEVORAC Home

DEVORAC  June 2012

DEVORAC June 2012

Options

Subscribe or Unsubscribe

Subscribe or Unsubscribe

Log In

Log In

Get Password

Get Password

Subject:

Re: AW: Time information

From:

Gareth Thomas <[log in to unmask]>

Reply-To:

Gareth Thomas <[log in to unmask]>

Date:

Thu, 14 Jun 2012 15:45:20 +0100

Content-Type:

text/plain

Parts/Attachments:

Parts/Attachments

text/plain (112 lines)

FYI, AATSR L1B uses a day count from 2000/01/01 00:00:00 UTC (i.e. 
essentially a Julian day). I assume ATSR-2 (in AATSR format) does 
something similar, but I don't know what the reference date is.

I agree with Richard that time should be stored as time elapsed since 
some defined date, and using a the earliest start date from the input 
data makes sense. My own preference would be to use days (stored as 
double precision floats) rather than seconds as the unit (I think it is 
a more human-friendly), but I am not fussed.

Regarding conversion between different formats, I guess we need two things:
- A conversion from/to UTC to/from a Julian day (i.e. time elapsed since 
some date).
- A similar conversion from TAI to/from Julian day. Like Richard, I am 
not familiar with TAI either, btw.
David Simpson from Goddard has F90 routines that do the former (as far 
as I can tell): http://caps.gsfc.nasa.gov/simpson/calendars/index.html
There is also CALPAK: 
http://people.sc.fsu.edu/~jburkardt/f_src/calpak/calpak.html

If I understand correctly TAI time will be 35 seconds ahead of UTC as of 
the end of June, but I can't find software that will convert between it 
and UTC. Essentially, you'd need a list of the times at which leap 
seconds were added to do the conversion.

If you are short of something to do, there is a document which describes 
Envisat time keeping here: 
http://eop-cfi.esa.int/CFI/ENV_CFI_DOCS/mcd2.0.pdf

Gareth

On 14/06/12 15:15, Jerg Matthias wrote:
> I agree that a simple integer number which gives the elapsed seconds is much more handy if you want to collocate later etc. I am also struggling a little bit to find conversion tools for those different formats. Does anybody have experience with those things or knows what to look for exactly? Ideally it should be something in fortran which could be easily linked in the preprocessing so that everything after that step is uniform wrt time.
>
> Matthias
>
> -----Ursprüngliche Nachricht-----
> Von: ORAC Developers [mailto:[log in to unmask]] Im Auftrag von Richard Siddans
> Gesendet: Donnerstag, 14. Juni 2012 15:31
> An: [log in to unmask]
> Betreff: Re: Time information
>
> I don't know about the difference between UTC and TAI but I would definitely favour storing time+date as a elapsed seconds since a given date (double precision value).
> 1/1/70 as per AVHRR seems a sensible start date for this application.
>
> I don't think its a good idea to store times as just an HHMMSS string.
> It might be easy to read by eye but is very messy to handle computationally.
>
> Richard
> -----------------
> Richard Siddans ([log in to unmask])
> Rutherford Appleton Laboratory, Chilton OX11 0QX, U.K.
> Tel: (44) (0)1235 446428
> Fax: (44) (0)1235 445848
>
>
>> -----Original Message-----
>> From: ORAC Developers [mailto:[log in to unmask]] On Behalf Of
>> Caroline Poulsen
>> Sent: 14 June 2012 13:58
>> To: [log in to unmask]
>> Subject: Re: Time information
>>
>> Hi Matthias,
>> just looked at the data definition document and cut and pasted the
>> below
>>
>>
>> The identifying time for this data set in UTC.
>>
>> Format is [HH[MM[SS]]] where HH is the two digit hour from 00 to 23,
>> MM is the two digit minute from 00 to 59 and SS is the two digit
>> second from 00 to 59.
>>
>>
>> -----Original Message-----
>> From: ORAC Developers [mailto:[log in to unmask]] On Behalf Of
>> Jerg Matthias
>> Sent: 14 June 2012 13:37
>> To: [log in to unmask]
>> Subject: Time information
>>
>> Hello everybody,
>>
>> I included the reading and "computation" of time information for MODIS
>> and AVHRR for each pixel (line). However, it is not clear to me what
>> time standard that is. E.g. MODIS uses TAI (international atmosc time)
>> with reference 1.1.1993, AVHRR seems to use 1.1.1970 but I do not know
>> the time standard, e.g. UTC or TAI. TAI and UTC seem to differe in the
>> sense that UTC has leap seconds included, TAI doesn't ave that. I also
>> would not presently know how to transform those standards into one
>> another. It is also not clear to me what we are supposed to use or
>> what AATSR uses.
>>
>> Anyway, the additional code has just been added to the repository and
>> is online now.
>>
>> Cheers,
>>
>> Matthias

-- 
Address: Dr G E Thomas
          Atmospheric, Oceanic & Planetary Physics
          Clarendon Laboratory
          Parks Road
          Oxford OX1 3PU
          England

Email: [log in to unmask]
Telephone: +44/0 1865 272 894
Fax:       +44/0 1865 272 923

Top of Message | Previous Page | Permalink

JISCMail Tools


RSS Feeds and Sharing


Advanced Options


Archives

April 2014
March 2014
February 2014
January 2014
December 2013
November 2013
October 2013
September 2013
August 2013
July 2013
June 2013
May 2013
April 2013
October 2012
September 2012
August 2012
July 2012
June 2012
May 2012
April 2012
March 2012
February 2012
January 2012
December 2011


WWW.JISCMAIL.AC.UK

Secured by F-Secure Anti-Virus CataList Email List Search Powered by the LISTSERV Email List Manager