Integrity review report: Retrieval of hourly aerosol single scattering albedo over land using geostationary satellite data

Academic fraud report · Geng Detector

Summary

Verdict: Questionable (Yellow). The reviewer identified two clear typographical/LaTeX errors in the manuscript but found no substantive evidence of data fabrication, image manipulation, or statistical malpractice. The errors involve (1) a formula descriptor on page 5 where the parameter H_geo is misrendered as H_iso-, producing a nonsensical expression 'H_iso-1.377622', and (2) an undefined dependent variable 'SSS_AERONET' appearing in regression equations in the abstract and results where 'SSA_AERONET' is clearly intended. All statistical metrics cited in the review (R=0.7284, R=0.8627, R=0.0969, n=1071, n=1766) were verified against the original text. Table 3 standard deviation distributions and satellite/data-access timelines were internally consistent. Image-level forensic analysis was not performed because only extracted text was available. The recommended action is to request an Erratum; no retraction is warranted.

Verdict

Questionable (Yellow). No evidence of research fraud was found. Two typographical/LaTeX errors were confirmed in the text. Statistical reporting, data provenance, and timelines are internally consistent.

Key findings

Formula description typo (medium concern): On page 5, below equation (8), the BRDF kernel coefficients are described. The original text reads: "Where H_iso=1, H_vol=0.189184, and H_iso-1.377622." The third coefficient should be H_geo = -1.377622, following the Ross-Li BRDF kernel convention. The mis-typed form creates a non-existent parameter.
Undefined variable name in regression equation (medium concern): The abstract and Results sections report "SSA_ASL=0.60*SSS_AERONET+0.38". The independent variable on the right-hand side is written as SSS_AERONET, which is never defined anywhere in the manuscript. The intended variable is SSA_AERONET. This is a clear LaTeX/typing error rather than a data issue.
Statistical metrics consistent (no concern): AHI/AERONET collocation sample sizes (1071 for ASL; 1766 for JMA) and correlation coefficients (R=0.7284 overall for ASL; R=0.8627 for the AOD>0.4 subset; R=0.0969 for JMA at low aerosol loading) are internally coherent and consistent with the cited scatter plot descriptions.
Standard deviation distribution consistent (no concern): Table 3 TOA standard deviations range from 1.11E-05 to 0.0075 and increase non-monotonically with AOD, including a sharp rise at 2.5 AOD over bright surfaces. This pattern is consistent with real instrument noise rather than synthetic fabrication.
Timeline coherent (no concern): Himawari-8 was launched in 2014 by JMA and operational from mid-2015. The study uses January–September 2020 data, last accessed 30 November 2023, which is fully consistent with publication in 2024.

Evidence highlights

DOI: 10.1038/s41612-024-00690-6
Journal: npj Climate and Atmospheric Science (2024) 7:157
Confirmed erroneous passages (verbatim from manuscript, as transcribed in the source report):
Page 5: "H_iso-1.377622" — should be "H_geo = -1.377622".
Abstract/Results regression: "SSA_ASL=0.60*SSS_AERONET+0.38" — SSS undefined.
Verified numeric values from the original text: R=0.7284 (ASL, n=1071); R=0.8627 (AOD>0.4 subset); R=0.0969 (JMA, low aerosol); Table 3 TOA SD range 1.11E-05 to 0.0075.
Data timeline: Himawari-8 launch 2014; study window Jan–Sep 2020; data last accessed 30 November 2023.

Notes

Limitation: only extracted text was available to the reviewer; pixel-, noise-, and metadata-level image forensics (Checks 1 and 3 of the standard Geng protocol) were not executed.
No image duplications, no p-hacking patterns, no signs of fabricated statistics were detected within the scope of text-based review.
The author team should be encouraged to publish an Erratum correcting the H_geo descriptor and the SSA/SSS variable inconsistency. A formal misconduct investigation is not warranted on the basis of current evidence.