Temporal patterns of instrumentation failures in atmospheric boundary layer observational systems

his study examines characteristic patterns associated with equipment failures in time series data from urban environmental monitoring systems. While prolonged data gaps represent the most apparent failure indicator, equipment malfunctions manifest through more complex signatures that can compromise forecasting accuracy. Analysis of six-year datasets from two distributed air quality monitoring networks in Krasnoyarsk (comprising 30 monitoring stations) reveals multiple failure-induced patterns, including both localized and sequential outliers. The non-stationary nature of environmental time series necessitates careful discrimination between genuine equipment failures and natural anomalies, as not all deviations require data exclusion. Current methodologies for failure pattern detection are reviewed, with particular attention to their limitations when applied to real-world non-stationary measurement series. The findings highlight critical considerations for maintaining data quality in operational environmental monitoring systems.