Audit Report: 'Retrieval of Temperature Profiles in the MLT Region Using the O2(b1Σ+g) Airglow With Spectral Contamination Correction of N2 1PG Emissions' (IEEE TGRS, 2024)

Academic fraud report · Geng Detector

Summary

This audit report rates the paper as highly suspicious, primarily due to internal data inconsistencies and fundamental writing errors. The most serious finding is a direct contradiction between Section IV (Self-Consistency Analysis) and Section V (Discussion and Conclusion): the body text states the post-correction mean bias is 6 K, while the conclusion states 9 K, even though both cite the identical 80% improvement figure. A second confirmed finding identifies a clear factual error in the conclusion, which incorrectly describes the target wavelength as 726 nm instead of the 762 nm O2-A band consistently used throughout the rest of the paper. A third potential issue involves garbled LaTeX/PDF rendering of Equation (9) (RMSE formula), though this is judged as insufficient evidence of misconduct. Overall confidence in the textual inconsistencies is high because they directly contradict statements within the same document. Limitations include the absence of pixel-level image analysis since raw figures were not provided. The final determination of misconduct requires an institutional investigation.

Verdict

🟠 Highly suspicious — Based on confirmed internal contradictions and factual errors identified in the published text.

Key findings

Internal numerical contradiction between body and conclusion (Severity: 🔴, Verified ✅): Section IV-A states the mean bias after N2 spectral contamination correction is 6 K (80% reduction), while Section V states the mean bias is 9 K (80% improvement). The improvement percentage is identical, but the absolute value differs.
Core wavelength error in conclusion (Severity: 🟡, Verified ✅): Section V begins with a statement referring to the "726 nm O2-A band airglow retrieval," which contradicts the paper's title, abstract, introduction, and full derivation, all of which consistently use the 762 nm (specifically 762.6 nm) O2-A band.
Malformed RMSE formula in Equation (9) (Severity: 🟡, Unverified ⚠️): The extracted LaTeX reads $U_{rmse} = \sqrt{\frac{\sum^N_{i=1} - [X_i - F] \cdot^2}{N}}$ , containing a stray negative sign and misplaced operator. This may reflect PDF extraction artifacts or an actual LaTeX typo in the published article.

Evidence highlights

Quantitative discrepancy: Reported post-correction bias values of 6 K vs. 9 K with identical 80% improvement claim — DOI: 10.1109/TGRS.2024.3492178, Sections IV-A and V.
Wavelength inconsistency: "726 nm" appearing only in the conclusion paragraph vs. "762 nm" / "762.6 nm" used throughout title, abstract, introduction, and methodology.
Formula anomaly: Equation (9) on Page 5 contains non-standard LaTeX syntax that does not match the conventional RMSE definition sqrt(Σ(Xᵢ − F)² / N).

Notes

Authors listed: Kuijun Wu, Haotian Li, Jiarui Su, Daoqi Wang, Zhihua Wang, Weiwei He.
Publication venue: IEEE Transactions on Geoscience and Remote Sensing, Vol. 62, 2024.
Confidence: High for the 6 K vs. 9 K discrepancy and the 726 nm error, since both can be verified by direct textual comparison within the article. Low for the formula issue, pending verification of the original typeset PDF.
Limitations: No raw figures were supplied, so image-based manipulation (e.g., in airglow spectra or temperature profile plots) could not be assessed. The report is AI-assisted and does not constitute a formal finding of misconduct.
Recommended actions: Contact the corresponding author to clarify the bias discrepancy and the wavelength error; file a comment on PubPeer; request that IEEE TGRS issue an erratum clarifying the corrected bias value and target band.