Hi,
Dear list,
I have a question about mosaicing multidimensional datasets (netCDF,
Zarr etc). These are typically 2 or 3D grids with a time dimension
from weather or ocean models (3 or 4D in total).
Using a multidimensional VRT:
https://gdal.org/drivers/raster/vrt_multidimensional.html
it is possible to mosaic several arrays with shared dimensions along
the time or spatial dimensions without resampling.
I have a use case where I want to mosaic several datasets with
varying spatial resolution, and possibly different dimension names and
reference systems, both along the time dimension and the spatial
dimensions (onto a common spatial 'grid'). The goal is to combine
global models with local models of higher resolution and get a common
combined multidimensional array (2 or 3D + time).
This is more like a mixture of a classic 2D VRT:
https://gdal.org/drivers/raster/vrt.html and the multidimensional VRT.
I see that there is some new functionality for combining 2D parts of
multidimensional datasets, with reprojection / resampling, in a
"classic" VRT:
https://gdal.org/drivers/raster/vrt.html#arraysource
But this does not seem to exactly cover my use case. I would like
to be able to do the 2D (or even 3D) mosaicing for the spatial part
with reprojection / resampling to a common grid while also combining,
and keeping, the time dimension. Bottom line, I guess, is that I would
like to be able to use the functionality of <DerivedArray> on a
2D-part inside the <Source> element of <Array> in a multidimensional
VRT
(https://gdal.org/drivers/raster/vrt_multidimensional.html#vrt-multidimensional).
Am I right in assuming that this is not possible "out of the box"
with the current options provided by VRTDataset, so that I would need
to write some code to cover the use case myself?
yes, you are right. What you want to do isn't currently possible out of
the box. Just porting the DerivedArray in a multidimensional VRT would
not be enough as GDALMDArray::GetResampled() only supports resampling to
a common grid on the 2D space. Resampling along the time dimension would
require extra development.
Even
Best regards,
Simon Kokkendorff
Joint GEOMETOC Support Centre
_______________________________________________
gdal-dev mailing list
gdal-dev@lists.osgeo.org
https://lists.osgeo.org/mailman/listinfo/gdal-dev
--
http://www.spatialys.com
My software is free, but my time generally not.
_______________________________________________
gdal-dev mailing list
gdal-dev@lists.osgeo.org
https://lists.osgeo.org/mailman/listinfo/gdal-dev
