clippig netcdf using rioxarray

[gkb999]

Hello team,

The following is my NC input bounds as follows:
image

My guess: it includes global values.

What I'm looking at: Clipping for the following bounds (in lat-lon)
min_lon = 140 & min_lat = -90 & max_lon = 220 &max_lat = -65 (Southern hemisphere)

Adding to this, my projections are in linear units: meters,
My input nc's projection is as below:

I need to clip for the said lat-lon.
As I don't have idea what these values mean in meters, I used utm.from_latlon(-65, 140): to get a particular value in meters and then tried clipping as da.rio.clip_box(minx= 2035.06, miny=500000.00, maxx=2791172.45, maxy= 547155.12,)

NOW, THE CLIPPED ARRAY HAS:

Does this mean 3 longitudes and 112 latitudes???

Then, this is undoubtedly the wrong way.
Can someone please suggest me the right way?

What I also tried:

1)Converting XY to lat-lon and then select from data array: DID NOT WORK
2) Transforming the projection of the NC data using TRANSFORMER.TRANSFORM: NOT SUCCESSFUL AGAIN
3)Last method, getting lat-lo values into meters, then feed as input for the bounding box - DIDNT WORK AGAIN.
Please provide me with the appropriate way.
I'm sure I have messed up using multiple ways and am half-lost already.

[snowman2]

Have you tried this?

https://pyproj4.github.io/pyproj/stable/api/transformer.html#pyproj.transformer.Transformer.transform_bounds

transformer = Transformer.from_crs("EPSG:4326", data.rio.crs, always_xy=True)
data.rio.clip_box(*transformer.transform_bounds(140, -90, 220, -65))

[gkb999]

wow, after adding inplace = true

It seems it worked, need to test if it is on the right spot.
Thanks a ton!!!

[snowman2]

The raster CRS need to be set to: EPSG:3412

data.rio.write_crs("EPSG:3412", inplace=True)

[gkb999]

Thanks for the correction,
Yet, that doesn't seem to change any.
Please refer to the following

Also, why is spatial_ref null : ()

[snowman2]

why is spatial_ref null : ()

See: https://corteva.github.io/rioxarray/stable/getting_started/crs_management.html

[gkb999]

Thanks a lot, that makes sense now