Biomass reproject + update links

[HarshiniGirish]

No description provided.

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[wildintellect]

I'm redirecting review to @hrodmn and @smk0033 who are working the OGC App Pack example.

[HarshiniGirish]

@smk0033 @hrodmn need a review on this notebook please!

[hrodmn]

@HarshiniGirish I'm sorry this PR slipped past me a few months ago!

The STAC item bbox approach is clever but we don't need to do that! I am not sure why these files have a null crs but they are georeferenced. Instead of a crs they have gcps defined (gcps = Ground Control Points).

Here is how you can use it to do the reprojection without guessing using the STAC metadata:

rio_env = rasterio.Env(GDAL_HTTP_HEADERS=f"Authorization: Bearer {token}")
with rio_env:
    with rasterio.open(asset_href) as ds:
        arr = ds.read(1)
        gcps_list, gcp_crs = ds.gcps
        valid = arr[np.isfinite(arr)]
        print("Shape:", arr.shape)
        print("CRS:", ds.crs)
        print("GCPS list:", gcps_list)
        print("GCP CRS:", gcp_crs)
        print("Transform:", ds.transform)
        print("Min/Max:", float(valid.min()), float(valid.max()))

src_arr = arr.copy()
src_height, src_width = src_arr.shape

# Reproject to EPSG:4326 on a clean output grid
dst_crs = "EPSG:4326"  # don't use EPSG:4326 for raster data if possible!

dst_transform, dst_width, dst_height = calculate_default_transform(
    src_crs=gcp_crs,
    dst_crs=dst_crs,
    width=src_width,
    height=src_height,
    gcps=gcps_list
)

reprojected = np.empty((dst_height, dst_width), dtype=src_arr.dtype)

reproject(
    source=src_arr,
    destination=reprojected,
    gcps=gcps_list,
    src_crs=gcp_crs,
    dst_transform=dst_transform,
    dst_crs=dst_crs,
    resampling=Resampling.nearest
)

print("Destination CRS:", dst_crs)
print("Reprojected shape:", reprojected.shape)
print("Destination transform:", dst_transform)

I think we should encourage users to use a different CRS than epsg:4326 for array data like this, so you could figure out roughly where the sample file is then pick a projected CRS that fits the area.

[HarshiniGirish]

thanks for the feedback @hrodmn. It's up for review again!

[review-notebook-app]

View / edit / reply to this conversation on ReviewNB

hrodmn commented on 2026-06-04T14:03:26Z
----------------------------------------------------------------

I would just drop the "Because the file already contains ..." sentence because we don't need to reference the STAC item for anything.

[review-notebook-app]

View / edit / reply to this conversation on ReviewNB

hrodmn commented on 2026-06-04T14:03:27Z
----------------------------------------------------------------

Line #4.    

we can drop references to STAC item bbox everywhere!

[review-notebook-app]

View / edit / reply to this conversation on ReviewNB

hrodmn commented on 2026-06-04T14:03:28Z
----------------------------------------------------------------

Here's another STAC item bbox reference to remove.

[hrodmn]

Thanks @HarshiniGirish! I left a few more suggestions - let's remove all references to the STAC item bbox approach since we don't need to plant that seed in any reader's mind 😆

The plot at the end suggests that nodata values are not getting masked out so you should take another look at your approach.

[HarshiniGirish]

Thank you @hrodmn . I apologise this took a while to get back to.

Updated — I removed the remaining STAC item bbox reference and revised the reprojection/plotting workflow so nodata is explicitly masked before reprojection.

[review-notebook-app]

View / edit / reply to this conversation on ReviewNB

hrodmn commented on 2026-06-16T01:19:03Z
----------------------------------------------------------------

The eps value is being used to fill nodata pixels - can you figure out some way to keep those as nan ?

[review-notebook-app]

View / edit / reply to this conversation on ReviewNB

hrodmn commented on 2026-06-16T01:19:03Z
----------------------------------------------------------------

@HarshiniGirish please remove this whole cell

[hrodmn]

@HarshiniGirish I assume Claude left behind the new markdown cell with a note about not using the STAC item extent to do the reprojection 😆 - please read the whole thing and make sure all of the STAC item extent references are really gone before tagging me to review again!

I also think the plot of the reprojected data deserves another look. The valid mask from the raw data should be applied to the output and the invalid areas should have nan values instead of 20 * log10(1e-6)

[HarshiniGirish]

My bad @hrodmn : ) . Turns out I uploaded the wrong document while setting up the PR. This one looks fine. Thanks for the review