Merging data arrays WindowError: bounds and transforms are inconsistent

[nilsleh]

I am trying to merge some DataArrays based on spatial coordinates. The data I am working with is climate data on a global grid. I am running into rasterio.errors.WindowError: Bounds and transform are inconsistent when trying to use the merge_arrays() function. I recreated the issue in this google colab notebook. The real data I have does not inherently have a transform defined that could be accessed by rioxarray so I think there is one created on the fly and I get an error. I am wondering how I should handle my data to make the merge_array() work.

[RichardScottOZ]

Did write_crs definitely work?

[nilsleh]

No, I think we are still having that issue, although that PR for which we need it has gone stale for now microsoft/torchgeo#1490 but will be picked up again.

[snowman2]

I believe that this is the source of your issue:

lats = np.linspace(lat_min, lat_max, SIZE)

Your raster is upside down. If you reverse your y coordinates in the raster, the bounds will be as expected. Otherwise, you will need to change the input bounds to have miny and maxy swapped.