Mobile monitoring for mapping spatial variation in urban air quality: Development and validation of a methodology based on an extensive dataset

Mobile monitoring is increasingly used as an additional tool to acquire air quality data at a high spatial resolution. However, given the high temporal variability of urban air quality, a limited number of mobile measurements may only represent a snapshot and not be representative. In this study, the impact of this temporal variability on the representativeness is investigated and a methodology to map urban air quality using mobile monitoring is developed and evaluated. A large set of black carbon (BC) measurements was collected in Antwerp, Belgium, using a bicycle equipped with a portable BC monitor (micro-aethalometer). The campaign consisted of 256 and 96 runs along two fixed routes (2 and 5 km long). Large gradients over short distances and differences up to a factor of 10 in mean BC concentrations aggregated at a resolution of 20 m are observed. Mapping at such a high resolution is possible, but a lot of repeated measurements are required. After computing a trimmed mean and applying background normalisation, depending on the location 24–94 repeated measurement runs (median of 41) are required to map the BC concentrations at a 50 m resolution with an uncertainty of 25%. When relaxing the uncertainty to 50%, these numbers reduce to 5–11 (median of 8) runs. We conclude that mobile monitoring is a suitable approach for mapping the urban air quality at a high spatial resolution, and can provide insight into the spatial variability that would not be possible with stationary monitors. A careful set-up is needed with a sufficient number of repetitions in relation to the desired reliability and spatial resolution. Specific data processing methods such as background normalisation and event detection have to be applied. •Mobile monitoring of BC is performed in an urban environment using a bicycle.•Mapping the local BC concentration at a high resolution of up to 20 m is possible.•A large number of repeated measurements are required to obtain representative results.•The number of runs could be reduced by background normalisation and trimmed mean.•Guidelines for mobile monitoring campaigns are proposed.