[Gdal-dev] Re: GDAL netcdf convention CF support

[Frank Warmerdam]

On 8/18/05, Woolf, A (Andrew) <A.Woolf at rl.ac.uk> wrote:
>  
> This is great news, a good step along the road we're all headed. Norman
> mentioned what is also the end game for us - the WCS *spec* supports
> arbitrary 1-,2-,3- or 4-d slices in a fully 4-d dataset, eg: 
>   
> •3-d volume (lat-lon-height): 
> 
> http://glue.badc.rl.ac.uk/cgi-
> 
> bin/mapserv?REQUEST=GetCoverage&SERVICE=WCS&Version=1.0.0&Cov
> 
> erage=temperature&BBOX=-180,-
> 
> 90,179,90,0,1000&CRS=some3dCRS&FORMAT=CF-netCDF 
> •Hovmuller (lat vs time): 
> 
> http://glue.badc.rl.ac.uk/cgi-
> 
> bin/mapserv?REQUEST=GetCoverage&SERVICE=WCS&Version=1.0.0&Cov
> 
> erage=temperature&BBOX=10,-
> 
> 90,10,90,500,500&TIME=1996/2005/1&CRS=some3dCRS&FORMAT=CF-netCDF
> •Vertical slice through atmosphere (lat vs height): 
> 
> http://glue.badc.rl.ac.uk/cgi-
> 
> bin/mapserv?REQUEST=GetCoverage&SERVICE=WCS&Version=1.0.0&Cov
> 
> erage=temperature&BBOX=10,-
> 
> 90,10,90,0,1000&CRS=some3dCRS&FORMAT=CF-netCDF 
>   
> These are all *valid WCS requests*!!! My understanding, however, is that
> fundamental 2-D limitations in GDAL will make this extremely difficult to
> achieve in Mapserver - any views to the contrary? 

Andrew,

Indeed, as things stand in GDAL and MapServer it would be hard
to achieve support for 3D and 4D datasets smoothly.   It might be
possible to "hack something up", but I'm not sure how desirable
that would be.  At best, it would be fragile and/or dependent on
alot of extra configuration information in MapServer which would
internally still need to treat results as stacks of 2D bands. 

Best regards,
-- 
---------------------------------------+--------------------------------------
I set the clouds in motion - turn up   | Frank Warmerdam, warmerdam at pobox.com
light and sound - activate the windows | http://pobox.com/~warmerdam
and watch the world go round - Rush    | Geospatial Programmer for Rent