Code and data for "Detecting identical series and time-stamp inconsistencies in the European air quality record with an automated screen"

This archive contains the code, derived tables and flag lists of an automated screen for silent defects in the hourly air quality record that European countries report to the European Environment Agency (EEA): missing-data codes flagged as valid (test A), identical series reported for different sampling points (test B, rolling-hash search with time-shifted controls), changes of the diurnal phase of a series (test C) and lags between pollutants measured at the same station (test D, annual and monthly modes). The screen was applied to the verified E1a data of 40 countries for 2013–2024 (PM10, PM2.5, NO2, SO2, and O3 for consistency checks) and checked against the national data portal of Türkiye, the open data of Madrid and national sources of Ireland, Germany, Sweden, Croatia and Serbia. The archive includes injection experiments, a brute-force comparison, threshold sensitivity analyses, the national values used for the checks and the manifests of the downloaded EEA files. The raw monitoring data are public and are downloaded by the included scripts; they are not redistributed. See README.md for the order of execution. Version 1.1 adds the ozone data index and solar-noon check, all injection experiments under the autocorrelation-corrected clear-profile criterion of test C, the monthly-mode non-overlapping-window rule, the review load in unique station-years, the national hourly comparisons for the identical pairs, and the scripts that assemble the manuscript and Supplemental Material submitted to the Journal of the Air & Waste Management Association. Code: MIT licence; derived tables: CC BY 4.0.