Machine Learning‐Based Swath Bias Correction for OCO‐3 Snapshot Area Mapping Mode Observations

Details

• Journal Title:
  Earth and Space Science
• Personal Author:
  Mauceri, Steffen ; Nelson, Robert R. ; O'Dell, Christopher W. ; Taylor, Thomas E. ; Kiel, Matthäus ; Massie, Steven ; Payne, Vivienne H. ; Chatterjee, Abhishek
• NOAA Program & Office:
  CIRA (Cooperative Institute for Research in the Atmosphere)
• Description:
  Space‐based monitoring of urban carbon dioxide (CO 2) emissions is essential for climate policy, but systematic measurement biases can compromise emission quantification accuracy. NASA's Orbiting Carbon Observatory‐3 (OCO‐3) provides unique Snapshot Area Mapping (SAM) mode observations that can capture city‐scale CO 2 plumes in single overpasses, yet some SAMs exhibit swath‐based XCO 2 biases of 1-3 ppm that can mask or artificially enhance true emission signals. Here, we develop a machine learning framework to identify and correct these artifacts while preserving genuine atmospheric CO 2 variations. We trained a random forest classifier on 1,723 manually labeled SAMs spanning August 2019 to February 2025 to distinguish scenes requiring bias correction from those with legitimate spatial gradients in XCO 2 . The model uses five key diagnostic features identified through systematic feature selection: inter‐swath XCO 2 discontinuity, scene homogeneity, coarse‐ and fine‐mode aerosol optical depths, and spectral alignment corrections. A median‐alignment algorithm then removes detected biases by adjusting swath‐level offsets while conserving total column CO 2 . The classifier achieves a precision and recall of 79% and 63%, respectively, prioritizing the avoidance of false corrections that could otherwise remove true atmospheric signals. Applied to the complete OCO‐3 SAM record, consisting of 19,839 unique SAMs, we estimate that approximately 9% have an identifiable swath bias. This work addresses a critical OCO‐3 data quality issue that has hindered the operational use of OCO‐3 SAMs for urban fossil fuel emissions monitoring. The proposed correction framework enhances OCO‐3 data reliability for use in updating emission inventories and provides methodological insights for future greenhouse gas satellite missions.
• Source:
  Earth and Space Science, 13(7)
• DOI:
  https://doi.org/10.1029/2025ea004853
• ISSN:
  2333-5084 ; 2333-5084
• Format:
  pdf
• Publisher:
  American Geophysical Union (AGU)
• Document Type:
  Journal Article
• License:
  https://creativecommons.org/licenses/by/4.0/
• Rights Information:
  CC BY
• Compliance:
  Library
• Download URL:
  https://repository.library.noaa.gov/view/noaa/76046/noaa_76046_DS1.pdf
• File Type:
  [PDF - 10.37 MB]
• Collection(s):
  NOAA Cooperative Institutes
• Main Document Checksum:
  urn:sha-512:02ae55e1e42d45c723c4a62141d6605afd460c58aa64f0a7b0562026fe935732d2182cd5b7d789896bd382a70151266315e7d4e3d0236460909554c27953379c