Hyperlocal Air Pollution in London Part 2: NO2 Measured with a Low-Cost Sensor Network and Mobile Monitoring

Low-cost sensors (LCS) could help extend current air pollution monitoring programs by improving their density and time resolution. We evaluate rapid and hyperlocal pollution gradients in London using a network of LCS nodes and a Mobile Air Pollution Sensing (MAPS) car, focusing on NO2. The measurements were strongly correlated; the R2-value decreased rapidly from 0.80 to 0.19 as the distance between the MAPS car and the LCS nodes increased from 0-5 m to 16-45 m. A maximum variability of 119.0 ppb/min or 181.6 ppb/100 m is seen in urban areas on the roads, with average variability of around 6 ppb/min. Case studies in parks and tunnels revealed that NO2 mixing ratios varied significantly over small distances and times. For example, the mean mixing ratio of NO2 in the urban environment was 21.2 ppb (standard deviation (SD) = 11.8 ppb), whereas it increased to 61.3 ppb (SD = 25.0 ppb) in the Rotherhithe Tunnel, while NO2 increased from a mean of 11.4 ppb (SD = 3.0 ppb) to 22.0 ppb (SD = 16.3 ppb), from the center of Regent's Park to the roads circling around it. This study shows that pollution changes rapidly in urban areas in both space and time, and that high-density networks and mobile monitoring can significantly increase the accuracy of pollution mapping.