Sentinel-3 projection

[FedericOldani]

Hi there,
I am trying to download Sentinel-3 LST images but I am stuck with the following error:
Cannot automatically compute the resolution, since asset 'lst-in' on item 0 'S3B_SL_2_LST_20230101T091518_20230101T091818_0179_074_264_2160' doesn't provide enough metadata to determine its native resolution.

This happens with any item in my tests. Should I compute the params in some way?
Thank you in advance

[TomAugspurger]

Can you share a code snippet that reproduces the probelm?

[FedericOldani]

Here the code:

import pystac_client
import planetary_computer
import stackstac

catalog = pystac_client.Client.open(
    'https://planetarycomputer.microsoft.com/api/stac/v1',
    modifier=planetary_computer.sign_inplace,
)

bbox = (7.566238975059995, 44.323814339090575, 7.568329074631334, 44.326032959203175)

search = catalog.search(
    collections=['sentinel-3-slstr-lst-l2-netcdf'],
    bbox=bbox,
    datetime='2023-04-01/2024-05-12'
)
items = search.item_collection()

data = (
    stackstac.stack(
        items,
        assets=['lst-in'],
        bounds_latlon=bbox,
        rescale=False,
        xy_coords='center',
        properties=True,
        epsg=4326
        )
)

Note that without epsg param it returns the error:
ValueError: Cannot pick a common CRS, since asset 'lst-in' of item 0 'S3A_SL_2_LST_20230401T092051_20230401T092351_0179_097_150_2160' does not have one.

[TomAugspurger]

I'm not too familiar with sentinel-3 data, but do you know if those assets have a CRS? Looking at gdalinfo on the href from https://planetarycomputer.microsoft.com/api/sas/v1/sign?href=https://sentinel3euwest.blob.core.windows.net/sentinel-3/SLSTR/SL_2_LST___/2024/05/29/S3B_SL_2_LST____20240529T213622_20240529T213922_20240530T112857_0180_093_285_5760_PS2_O_NT_004.SEN3/LST_in.nc, I don't see anything that looks too promising.

[neoporteria]

The georeference of Sentinel-3 data comes in a different file (geodetic-in and cartesian-in), so it's necessary to extract the coordinates from there and georeference them with snap or using them as gcps