Geng Fraud Detection Report: "A new Einstein coefficient method for mesopause–lower thermosphere atmosphere temperature retrieval under a non-local thermal equilibrium situation" (DOI: 10.1364/OE.498765)

Academic fraud report · Geng Detector

Summary

Verdict: Cleared (no indicators of academic fraud detected). This 2023 Optics Express paper by Haotian Li et al. introduces an Einstein-coefficient-based temperature retrieval method for the mesopause–lower thermosphere under non-LTE conditions, using MIGHTI observations from the ICON satellite. Text-based audit across six dimensions found: (1) a fully self-consistent timeline, with ICON's 2019 launch, 2022 observation dates, and submission in 2023 all coherent; (2) realistic, non-smooth error profiles, including a 130 K deviation for the line-intensity method at 120–140 km and an RMS spike to 65–94 K at 130 km versus ACE-FTS, used honestly as evidence of N2 airglow contamination rather than hidden; (3) mathematically self-consistent derivations of the emission-to-line-intensity ratio (∝ ν²) from Einstein A/B coefficients; and (4) plausible statistical treatment using RMS, slope, and intercept. Limitations: only the PDF text was available, so pixel-level image reuse and splicing checks (Western-blot-type forensics) were not performed.

Verdict

Cleared. No textual, methodological, timeline, statistical, or logical indicators of academic fraud were identified. Pixel-level image forensics could not be performed because only the PDF text was supplied.

Key findings

Timeline consistency (Rule 6: references & methodology anomalies): ICON satellite launch (2019), observation dates (5 Jan, 17 Jan, 19 Apr 2022), submission (2023), and HITRAN 2020 database reference (published 2022) are internally coherent and chronologically valid.
Realistic error magnitudes (Rule 2: fabricated-data detection): Reported deviations are large and non-monotonic — Einstein coefficient method max error 45 K, line-intensity method max error 130 K in the 120–140 km N2-contaminated band; RMS vs. ACE-FTS spikes to 65 K (Einstein) and 94 K (line-intensity) at 130 km. Such "ugly" residuals argue against synthetic smoothing and are openly used to motivate the N2 contamination discussion.
Mathematical/physical derivation integrity (Rule 6): The ratio ε/S ∝ ν² follows cleanly from Einstein A and B relations (Eqs. 3–8), supporting the non-LTE inversion logic without internal contradiction.
Statistical plausibility (Rule 4): Analysis uses RMS, slope (k), and intercept (b) to define Δy; no p-hacking, no forced significance, and uncertainty is reported as altitude-dependent rather than uniformly suppressed.
No image-text contradictions: Figure captions are consistent with surrounding narrative (image pixel-level reuse and splicing checks were out of scope).

Evidence highlights

DOI: 10.1364/OE.498765 (Optics Express, Vol. 31, No. 19, 11 Sep 2023).
Observation dates: 2022-01-05, 2022-01-17, 2022-04-19 (MIGHTI/ICON).
Database: HITRAN 2020 (Ref. 28).
Quantitative residuals: 45 K and 130 K maximum deviations (Section 5.2); 65 K and 94 K RMS at 130 km vs. ACE-FTS (Section 5.3).
Core relation: ε/S ∝ ν² (Eq. 8), derived from Einstein coefficients.

Notes

Scope limitation: analysis is text-only; Rule 1 (image reuse) and Rule 3 (image splicing) were not executed at the pixel level.
Confidence: moderate-to-high for the checked dimensions; any future image-level recheck would not be expected to change the verdict unless undisclosed figure manipulation exists.
Recommendation: no action required; the paper can be cited as a normal research contribution pending standard peer review.