Hi John,
We have used ncml to define the correct units for lat and lon axes as well
as y. Unfortunately, I don't have the ncml document with me at the moment.
I will send it if we still havent solved the problem tomorrow.
Cheers,
Adit
> the coordinate systems look ok to me. the lat and lon axis types dont
> get recognized because they use degrees as units instead of
> degrees_north and degrees_east, respectively, as CF requires. thats why
> you will get a null in your code.
>
> the grid display looks wrong, however, probably because X has "100 km"
> units, and y has nothing, so km is assumed.
>
> using the CoordSys tab of ToolsUI is one way to look for these kinds of
> issues.
>
> ads@xxxxxxxxxxxxxxxxxxxx wrote:
>> Hi John,
>>
>> The dataset is available here:
>>
>> http://topaz.nersc.no/thredds/dodsC/topaz/mersea-ipv2/arctic/tmipv2a-class1-b-be.html
>>
>> Cheers,
>>
>> Adit
>>
>>
>>
>>> Jon Blower wrote:
>>>
>>>> Hi John,
>>>>
>>>> The underlying question is, why doesn't java-netcdf recognize the axis
>>>> type? Previously (with the same dataset) the axis was recognized as
>>>> GeoX/GeoY, but now something has changed and it isn't. I think
>>>> something has changed with the remote dataset but I can't see from the
>>>> metadata (Adit provided an ncdump) what is wrong. The dataset is in
>>>> polar stereographic projection - is the metadata set up incorrectly or
>>>> is there a bug in java-netcdf?
>>>>
>>>> Cheers, Jon
>>>>
>>>> On Mon, May 18, 2009 at 5:26 PM, John Caron <caron@xxxxxxxxxxxxxxxx>
>>>> wrote:
>>>>
>>>>
>>>>> Hi Adit:
>>>>>
>>>>> The AxisType is optional, so you need to check for null.
>>>>> John
>>>>>
>>>>>
>>>>>
>>>>
>>>>
>>> ok, thats a clearer question. Adit, can you send me the actual file?
>>>
>>>
>
>
