OCO-3 SAMs (JPL) for US GHG Center

[siddharth0248]

I am looking into OCO-3 SAMs data. I had a few questions:
https://staging.openveda.cloud/api/stac/collections/oco3-co2-sam-cogs/items/

1. Did the metadata of the raw file had any information regarding sites or locations? Was each snapshot associated to a location and type of site as shown here. By associated to a location i mean do we have multiple images across time period for a fix location A.
   e.g Location A has images for 01/01, 20/01, 05/02 ....

2. We are planning to use GOES custom interface.

• On loading the dashboard user sees all the locations, color coded by type of sites
• On selecting one location e.g. Location A
• Card gallery on the right hand side will list all the snapshots for the selected location and select/click the one which needs to be visualized

3. Instead of Animation we will have a timeline with markers for dates on which snapshots were captured

[abarciauskas-bgse]

In response to question 1, the metadata of the raw file does have information on locations/sites, which I think are identified as TARGET_ID and TARGET_NAME in the gdal output:

$ gdalinfo /vsis3/veda-data-store-staging/oco3-co2-sam-cogs_v2025.04.16/oco3-co2_volcano0010_2025-03-30T232216Z_unfiltered_xco2.tif | grep TARGET_
  TARGET_BBOX=POLYGON ((152.83 -6.55, 152.83 -3.55, 149.83 -3.55, 149.83 -6.55, 152.83 -6.55))
  TARGET_ID=volcano0010
  TARGET_NAME=v252120_Ulawun

However, TARGET_ID and TARGET_NAME are not currently standalone fields in the STAC item metadata. The TARGET_ID is a part of the file naming convention: oco3-co2_volcano0010_2025-03-30T232216Z_unfiltered_xco2.tif is associated with target volcano0010 (see also https://github.com/EarthDigitalTwin/OCO3-data-transformer/blob/master/sam_extract/writers/CoGWriter.py#L165-L166).

I think to subset by target id we can either

1. Create an explicit property in the STAC metadata that include the TARGET ID and query directly for that property. This would be more work, but more explicit. @ividito @anayeaye is currently any way to add this type of arbitrary property or the veda-dataset-pipeline stac workflow?
2. We can query for the items associated with the volcano0010 site using the "like" operator on the id field, as follows:

import requests
search_url = 'https://staging.openveda.cloud/api/stac/search'

collections_filter = {
    "op": "=",
    "args": [{"property": "collection"}, 'oco3-co2-sam-cogs'],
}

id_filter = {"op": "like", "args": [{"property": "id"}, "%volcano0010%"]}

search_body = {
    "filter-lang": "cql2-json",
    "limit": 100,
    "context": "on",  # add context for a summary of matched results
    "filter": {
        "op": "and",
        "args": [collections_filter, id_filter],
    },
}

stac_items_response = requests.post(
    search_url,
    json=search_body,
).json()

len(stac_items_response['features'])
# returns 51 items

I think using such query logic will work well with the GOES Plume interface. The UI would either have a pre-existing list of TARGET IDs and associated latitude/longitude locations for placing markers or this list could be dynamically generarted from the STAC API and cached. When a user selects a given location, a search query to the STAC API would be made like the example above.

Let me know if that helps.