Revealing urban air-quality representativeness through measurements and modeling under uncertainty

Assessing the spatial representativeness area of air quality monitoring stations is essential to ensure that measured concentrations adequately reflect surrounding pollutant levels. However, traditional spatial representativeness area assessments often rely on deterministic thresholds and simplified geometric assumptions, without explicitly accounting for uncertainty. This study proposes an uncertainty-aware, street-scale framework for evaluating the spatial representativeness area of urban traffic air quality monitoring stations and applies it to NO₂ concentrations during a two-week period (May 13–31, 2023) in Barcelona, Spain. Spatial representativeness is assessed using observational and modeled NO₂ data, complemented by a data-fusion field that integrates both sources. Using a 15% tolerance consistent with FAIRMODE recommendations, circular spatial representativeness areas ranged between 20 and 24 m, indicating highly localized representativeness in the dense street-canyon environment of l'Eixample. Binary maps derived from modeled and fused concentration fields revealed irregular, non-circular patterns and highlighted the influence of microscale variability. To explicitly incorporate prediction uncertainty, a probabilistic SRA is derived from the Universal Kriging field, quantifying the likelihood that concentrations fall within the representativeness interval. Probability-based maps demonstrate that representativeness depends strongly on the required confidence level and decreases substantially under stricter thresholds. The proposed framework provides a nuanced and uncertainty-informed characterization of air quality monitoring stations' representativeness, supporting improved interpretation of urban monitoring data in heterogeneous environments.