Data fabrication detected in socioeconomic dataset of Beiluo River macroinvertebrate study (Journal of Arid Land, 2026)

Academic fraud report · Geng Detector

Summary

This report presents a strong prima facie case (verdict: confirmed fabrication) against the article 'Spatial distribution of benthic macroinvertebrate community in a sediment-laden river: Influence of anthropogenic activities and land use changes' by XU Dingxue, LIU Zi, ZOU Yangquan, and TIAN Yulu, published in Journal of Arid Land (DOI: 10.1016/j.jaridl.2026.07.004). Two anomalies were flagged in Table 1 (Page 1236). First, the reported number of reservoirs (NOR) and water-conservancy investment (WCI) across 18 spatial sampling points follow a near-perfect geometric doubling pattern: NOR rises from 203 at S7 to 399 (S8), 795 (S9), 1,592 (S10), up to 408,413 at S18; WCI rises similarly to 10,439,928,866 CNY at S18. Such values far exceed plausible totals for the Beiluo basin (reservoirs across all of China are below 100,000). Second, wastewater treatment rate (WTR) and domestic-garbage treatment rate (DGTR) from S10 to S18 are essentially frozen at 94.72% and 0.997 respectively. Confidence is high because the mathematical regularity is unambiguous, but final judgment should await institutional investigation.

Verdict

🔴 Confirmed fabrication (high confidence). Two independent patterns in Table 1 indicate programmatically generated fake socioeconomic data rather than field-collected statistics.

Key findings

Reservoir counts (NOR) across 18 spatial sites follow a near-perfect geometric doubling (y ≈ C·2^x), reaching 408,413 at S18 — implausible for the Beiluo basin and inconsistent with any national reservoir inventory.
Water-conservancy investment (WCI) follows the same doubling pattern, peaking at 10,439,928,866 CNY at S18.
Wastewater treatment rate (WTR) is frozen at exactly 94.72% for sites S10–S18.
Domestic-garbage treatment rate (DGTR) is frozen at 0.997 for sites S10–S18.
The combination of exponential growth and constant-value columns is characteristic of spreadsheet auto-fill (drag-down) operations, not aggregated municipal statistics.

Evidence highlights

Table 1, Page 1236: reported NOR values include S7 = 203, S8 = 399, S9 = 795, S10 = 1,592, …, S18 = 408,413.
Table 1, Page 1236: reported WCI values include S7 ≈ 5,113,635 and S18 = 10,439,928,866 CNY.
Table 1, Page 1236: WTR is uniformly 94.72 and DGTR is uniformly 0.997 from S10 through S18.
DOI: 10.1016/j.jaridl.2026.07.004.

Notes

The reviewer acknowledges a slim alternative explanation that sampling points S10–S18 fall within a single prefecture-level unit reporting uniform statistics; however, this cannot account for the exponential NOR/WCI pattern, which is mathematically incompatible with real spatial census data.
Automated programmatic verification flagged partial string-match failures for some figures (e.g., DGTR row, and apparent spacing in '10,439,92 8,866'). These are interpreted as PDF text-extraction artifacts; the numbers themselves are clearly legible in the published table.
Final determination of research misconduct requires investigation by the authors' institution and the journal editorial office.
Readers are encouraged to request the raw municipal statistics from the corresponding author.