Problem with bounding_box parameter

[riesgutierrez]

I am looking for the product "GLDAS Noah Land Surface Model L4 monthly 0.25 x 0.25 degree V2.1 (GLDAS_NOAH025_M)" for a polygon containing Colombia (bounding_box=(-80.947, -5.872, -65.479, 14.518)), using the function "earthaccess.search_data". However, when downloading with the function "earthaccess.download", the product is downloaded for the whole planet and not for the region of Colombia.

Thank you for your cooperation.

This is the code I am using:

import earthaccess
import os
import xarray as xr
from pyproj import CRS

auth = earthaccess.login()

#username: xxxxx
#password: xxxxx


#Searching for data
results = earthaccess.search_data(
    #short_name="GLDAS Noah Land Surface Model L4 monthly 0.25 x 0.25 degree V2.1 (GLDAS_NOAH025_M)",
    doi="10.5067/SXAVCZFAQLNO",
    cloud_hosted=True,
    bounding_box=(-80.947, -5.872, -65.479, 14.518),
    temporal=("2009-01-01", "2009-12-31"),
    )

data_links = [granule.data_links(access="external") for granule in results]

files = earthaccess.download(results, "D:/CURSO_R_PYTHON/series_tiempo_ml/descargas")

[betolink]

Hi @riesgutierrez, the issue here is that earthaccess doesn't support subsetting services yet. This is a global dataset so any bounding box will match any of the files in the collection. We are in the process of integrating subsetting services like OPeNDAP, NASA's Harmony etc to have that capability in the library.

In the meantime we could use OPeNDAP directly, I think @battistowx has some examples of how. Or we could also open the files and subset them using xarray(slow unless you are in AWS). Here is an example with another global dataset: https://notebooksharing.space/view/0c9b943bfe8ec599df7ecf667c0efc8450f4d3a4c55ce359acc6949a3cc832ee#displayOptions=

Lastly, we could download and subset locally (although not optimal), I think we should have subsetting services this year!

[riesgutierrez]

@betolink
Thank you very much for the information, I will proceed to review the link you share with me.

Regards

[mfisher87]

Hope you all don't mind, converting this issue to a discussion for better visibility than a closed issue would have :)