PACE/OCI processing

Frequently Asked Questions
Post Reply
quinten
Posts: 1021
Joined: Tue Mar 03, 2015 8:13 am

PACE/OCI processing

Post by quinten »

Hi all

Last week I pushed some changes to the GitHub that allow for processing of PACE/OCI data, with the availability of the first light data. This post will show some first results from OCI processing with ACOLITE. Note that this sensor and code are still in early operations and performance may vary. At the moment bands <345 nm are ignored. These bands will require further updates to the LUTs. If you use ACOLITE processing of OCI, please let me know if you find any issues. You should also contact me to discuss your results before distributing them! Validation will be performed in due course.

The OCI data are provided as rather large granules, and I would recommend processing for a rather small region of interest (e.g. <100x100 square km) with dsf_aot_estimate=fixed. If you want to process larger regions of interest I would suggest trying dsf_aot_estimate=tiled. I don't really recommend processing full tiles at the moment. Reflectance datasets include the detector name (blue, red, SWIR) to avoid duplicates in the overlapping region.

You need to use an up to date clone from the ACOLITE GitHub. PACE/OCI processing is not yet supported in any of the binary distribution versions, but will be included in future releases. PACE/OCI L1B data are available from NASA's EarthData. You need the L1B data in NetCDF format, that are about 1.9GB per scene. I have written some rudimentary API access to this page for search and download that I will make available in the future, likely included in the ACOLITE repository on GitHub.

The examples below have been reprojected to a UTM grid within ACOLITE (using output_projection=True), and the plotted spectra are the mean of 3x3 pixels.

The first example shows a subset over the Gironde estuary in France, processed using settings:

Code: Select all

inputfile=../PACE/Input/PACE_OCI.20240412T122654.L1B.nc
output=../PACE/Output/Gironde
limit=44.762348425684394,-1.686099592795415,45.84097732307731,-0.13903205797876694
ancillary_data=False
limit_buffer=0.1
reproject_before_ac=True
output_projection=True
PACE_OCI_2024_04_12_12_26_54_L2R_projected_rgb_rhos.png
PACE_OCI_2024_04_12_12_26_54_L2R_projected_rgb_rhos.png (71.87 KiB) Viewed 12026 times
Gironde_P1_spectra.png
Gironde_P1_spectra.png (125.06 KiB) Viewed 12026 times
The next example shows the La Plata estuary in Argentina, processed using settings:

Code: Select all

inputfile=../PACE/Input/PACE_OCI.20240418T171520.L1B.nc
output=../PACE/Output/LaPlata
limit=-36.54109787495697,-58.71242261601081,-34.1034345116905,-54.718016908115956
ancillary_data=False
limit_buffer=0.1
reproject_before_ac=True
output_projection=True
PACE_OCI_2024_04_18_17_15_20_L2R_projected_rgb_rhos.png
PACE_OCI_2024_04_18_17_15_20_L2R_projected_rgb_rhos.png (230.6 KiB) Viewed 12026 times
LaPlata_P1_spectra.png
LaPlata_P1_spectra.png (115.18 KiB) Viewed 12026 times
The final example shows Lake Taihu in China, processed using settings:

Code: Select all

inputfile=../PACE/Input/PACE_OCI.20240418T042340.L1B.nc
output=../PACE/Output/Taihu
limit=30.87803878372968,119.86213605241022,31.59138292148444,120.68267588163275
ancillary_data=False
limit_buffer=0.1
reproject_before_ac=True
output_projection=True
PACE_OCI_2024_04_18_04_23_40_L2R_projected_rgb_rhos.png
PACE_OCI_2024_04_18_04_23_40_L2R_projected_rgb_rhos.png (52.01 KiB) Viewed 12026 times
Taihu_P1_spectra.png
Taihu_P1_spectra.png (118.12 KiB) Viewed 12026 times
I hope this is useful for someone!

Quinten
Post Reply