Re: Loading MOPITT v6 HDF5 - array too large

Hi, Rebecca!

I downloaded the files and examined the contents with HDFView.

I think it's data producer (LaRC ASDC)'s fault or an HDF-EOS5 library
bug at the time of the MOPITT data production.

In 2008/01/01 file, StructMetadata.txt has the wrong value:

  nTime = 250841130416

In 2008/12/31 file, StructMetadata.txt has the correct value:

  nTime= 2

Due to the above mistake, I think NCL could not figure out that
nTime_MOP02 = 197103 in 2008/01/01.

I'm CC-ing this message to both ASDC support and HDF-EOS5 library developer.

--
HDF: Software that Powers Science
On Thu, May 16, 2013 at 1:43 PM, Rebecca Buchholz <rb864@uowmail.edu.au> wrote:
> Dear NCL-talk users,
>
> I'm trying to load and calculate information from MOPITTv6 level 2 HDF5
> files. I am having an unusual problem in that some files load with 2.5 x
> 10^11 time array points (way more than it should have) which requires
> calculations to be performed in loops. However, the values past about 2 x
> 10^5 seem to be zero. Some files load correctly with around 2 x 10^5 time
> points.
>
> Performing a ncl_filedump of the files shows that TimeCount_MOP02 may or may
> not be the same as nTime_MOP02 (examples below). TimeCount is the correct
> value, but the arrays are all conformed to nTime_MOP02.
> --------------------------------------------------------------------------
> Examples:
> 1) 20080101 - incorrect
> filename: MOP02J-20080101-L2V16.2.3.he5
> path: MOP02J-20080101-L2V16.2.3.he5.he5
>    file global attributes:
>       StartDateTime_MOP02 : 2008-01-01T00:00:09.936000Z
>       Year_MOP02 : 2008
>       Day_MOP02 : 1
>       Month_MOP02 : 1
>       missing_nodata_MOP02 : -9999
>       TimeCount_MOP02 : 197103
>    dimensions:
>       Unlim_MOP02 = -1
>       nTime_MOP02 = 250841130416
>       nPrs_MOP02 = 10
>       nPrs2_MOP02 = 10
>       nTwo_MOP02 = 2
>       nThree_MOP02 = 3
>       nRad_MOP02 = 12
>       nCldd_MOP02 = 10
> --------------------------------------------------------------------------
> 2) 20080101 - correct (TimeCount_MOP02 = nTime_MOP02)
> filename: MOP02J-20081201-L2V16.2.3
> path: MOP02J-20081201-L2V16.2.3.he5
>    file global attributes:
>       StartDateTime_MOP02 : 2008-12-01T00:00:49.387000Z
>       Year_MOP02 : 2008
>       Day_MOP02 : 1
>       Month_MOP02 : 12
>       missing_nodata_MOP02 : -9999
>       TimeCount_MOP02 : 188411
>    dimensions:
>       Unlim_MOP02 = -1
>       nTime_MOP02 = 188411
>       nPrs_MOP02 = 10
>       nPrs2_MOP02 = 10
>       nTwo_MOP02 = 2
>       nThree_MOP02 = 3
>       nRad_MOP02 = 12
>       nCldd_MOP02 = 10
> --------------------------------------------------------------------------
>
> Does this have something to do with NCL translating HDF5 files? I'm not sure
> why some MOPITT v6 files work and others do not. Any advice would be much
> appreciated!
>
> Files can be accessed via
> http://l4eil01.larc.nasa.gov:22000/WebAccess/drill;jsessionid=76F2189CB04FA77E3EE5C02F61337508?attrib=esdt&esdt=MOP02J.6&group=MOPITT
>
> I have attached the ncl script I am using to extract a subset of data. I
> have added an exit command after determining size of the time array. Using
> NCL v6.1.2 on my MacBook Air, with MacOSX 10.6.8.
>
> Kind Regards,
> Rebecca
>
>
> Rebecca Buchholz - PhD Candidate
>
> BSc/BA (Hons.)
>
> Centre for Atmospheric Chemistry
>
> Faculty of Science
>
> University of Wollongong NSW 2522
>
> T + 61 2 4221 3196
>
> W  http://www.uow.edu.au/science/research/cac
>
>
> _______________________________________________
> 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