SWOT raster with wrong geometry metadata in

[sylvain.biancamaria]

Dear all,

Lately, I am trying to download SWOT Raster product using the earthaccess python library using a bounding box. Some granules provided by earthaccess.search_data() function are actually not included in the bounding box, which is an issue, as it requires to manually filter these tiles (it is especially time consuming when a long time period of SWOT data is selected).

To be more specific, I followed the PO.DAAC SWOT Cookbook tutorial https://podaac.github.io/tutorials/notebooks/datasets/SWOT_Raster_Notebook_local.html#author-nicholas-tarpinian-po.daac . The search the command (after login):
results = earthaccess.search_data(
short_name = 'SWOT_L2_HR_RASTER_2.0',
bounding_box=(-115.112686,35.740939,-114.224167,36.937819),
temporal =('2024-02-01 12:00:00', '2024-02-01 23:59:59'),
granule_name = '*_100m_*',
count =200
)
finds 3 granules instead of 2 in the tutorial. The additional granule is SWOT_L2_HR_Raster_100m_UTM60V_N_x_x_x_010_222_021F_20240201T215601_20240201T215622_PIC0_01.nc, which is located in Russia (near the Bering Sea coast). Surprisingly, the geometry for this granule in the search result is:
'Geometry': {'BoundingRectangles': [{'WestBoundingCoordinate': -180, 'SouthBoundingCoordinate': -90, 'EastBoundingCoordinate': 180, 'NorthBoundingCoordinate': 90}]}

On other location than lake Mead, I also found similar granules with the same geometry issue in the product search, like SWOT_L2_HR_Raster_100m_UTM60W_N_x_x_x_030_472_018F_20250403T031818_20250403T031839_PIC2_01.nc, SWOT_L2_HR_Raster_100m_UTM01W_N_x_x_x_030_545_136F_20250405T183334_20250405T183341_PIC2_01.nc, SWOT_L2_HR_Raster_100m_UTM01C_N_x_x_x_031_287_002F_20250417T092022_20250417T092043_PIC2_01.nc,...
When I download these granules, the variables geospatial_lon_min, geospatial_lon_max, geospatial_lat_min, and geospatial_lat_max in the netcdf header have the correct values, which correspond to the real data extent and does not cover the whole Earth. Could it be due to an issue in the metadata used by EarthData database? This is just a guess and I am not an expert in geospatial databases. The same issue happens with SWOT raster version D product. I ran these tests today (September 10th, 2025)

I would be interested to know if I am not using earthaccess.search_data correctly, or if it is an issue with the earthdata database. If this is the first case, than how should I modify the search? If it is the latest, how could I get around this problem?

Thank you in advance for your feedback,
Sylvain