The time range for TEMPO N02 V03 via Harmony

[kodiologist]

The page https://cmr.earthdata.nasa.gov/search/concepts/C2930763263-LARC_CLOUD.html says that TEMPO N02 V03 covers 2023-08-02 to 2025-09-16. But when I try to download some data from June 2024, I get no files, and the message "WorkItem warned: nasa/harmony-opendap-subsetter:1.2.2: Temporal range request outside supported dimension range for {'/time'}". Am I doing something wrong? Here's my Python script:

Code: Select all

from datetime import date
import harmony

collection = 'C2930763263-LARC_CLOUD'

client = harmony.Client(auth = (USERNAME, PASSWORD))

job = client.submit(harmony.Request(
    collection = harmony.Collection(id = collection),
    spatial = harmony.BBox(-100.4, 18.1, -97.4, 20.6),
    temporal = {'start': date(2024, 6, 3), 'stop': date(2024, 6, 4)},
    max_results = 1))
client.wait_for_processing(job)
print(client.status(job))

[ASDC - ghayescrepps]

Thank you for reaching out with your question. The ASDC service team reviewed your code and determined that this issue comes from an unfortunate combination of request parameters. With max_result=1, the request is limited to the first granule that contains data for June 3, 2024. In this case, the first captured granule has metadata time bounds from 2024-06-02 23:49:39 to 2024-06-03 00:29:25.

However, for Level-3 granules, post-processing collapses the time information to a single timestamp representing all data in the granule. For this particular granule, that timestamp is 2024-06-02 23:49:59.

As a result, when the subsetter tries to select data starting at 2024-06-03 00:00:00, it finds no matching data within that granule and returns a “no data found” error.
If you change your request to max_result=2 or more than 2, you will capture other granules that have data within your requested time limits.

If you have additional questions, please reach out.

[kodiologist]

Thanks. When I set `max_results = 2`, I get the same warning, but I can download a file, which is what matters. The warning is still produced with the result set expanded further with `temporal = {'start': date(2024, 6, 1), 'stop': date(2024, 6, 30)}` and `max_results = 10`.

[ASDC - ghayescrepps]

I am glad to hear that you were able to get your downloaded files. With max_results=2, Harmony captures 2 files, the first file from the original problem, and the second file that contains timestamp within specified limits. Therefore, Harmony reports the issue with no data even if it's just one file out of many so the warning is the expected behavior.

[kodiologist]

I see. Forgive me if this is off-topic, but I've noticed that Harmony login, at least with the Python library, is very slow: the line `client = harmony.Client(auth = ("Kodiologist", PASSWORD))` takes 2 minutes and 17 seconds. The option `should_validate_auth = False` only seems to defer this delay to a later operation rather than save time. Is this expected? I'm running Python 3.13.7 on Ubuntu 25.10.

[ASDC - ghayescrepps]

Thank you for letting us know of this behavior. I spoke with the Harmony development team and they said that this is not the expected behavior (i.e., it shouldn't be taking that long). They determined that the extended time was due to a bug related to how the authentication request was being queued and handled. They are going to fix this bug to avoid this issue in the future.