What is the workflow for reprojecting PACE OCI data?

[rodbager3579]

Hello, good day.

I'm processing L1B data from PACE OCI to L2 using the SFREFL suite. My problem comes when trying to reproject the bands through command lines, as it doesn't run L2bin or L3mapgen. I've tried other methods with Python and libraries, but since I'm not using a regular grid, I can't get a proper reprojection.
Can anyone advise me or give me suggestions on how to reproject and mask PACE OCI data? My goal is to automate the entire process.


The image on the right shows the reprojection using SNAP, the one on the left shows one of my attempts to do the reprojection with Python.

Thank you very much.

[OB ScienceSW - jlindo1]

Hi there!

My first question is over your reasoning for processing L1B into L2 using SREFL. Earth Data already has L2 SREFL files, and you could download them through obdaac_download as well. After downloading an SREFL file, you could run mapgen on the command line with a --projection flag to indicate which projection you'd like.

I'm happy to answer more questions!

Best,
Jakob

[sacaplan]

Hello,

If you're looking for the preprocessed L2 SFREFL files, Jakob's answer would work well.

If you want to process the data on your own, though, rioxarray package handles PACE's irregular L2 grids well, and is what I'd suggest if you're looking to reproject your processed data! We will have a tutorial out on the Help Hub soon (with more explanation for each step and why they're necessary), but I've outlined the basic process that should work for you in the image below. This method is relatively newly implemented, so please make sure your Python environment has the most updated versions of rioxarray, xarray, libgdal-core, and rasterio.

Cheers,
Skye