Estimating the spatial variability of fine particles at the neighborhood scale using a distributed network of particle sensors

Small-scale heterogeneity of airborne pollutants may have implications for accurate exposure estimation in environmental health studies. However, it has been studied thus far mainly near main roads and over relatively short periods. The emergence of low-cost miniature particle sensors enables deployment of multi-sensor nodes for studying the spatial variability of ambient pollutants at fine spatial scales for extended time periods. We carried out measurements of fine ambient particles, both in terms of particle number concentrations (PNC; 0.3 < d < 3 mu m) and particle mass concentrations (PM2.5; 0.3 < d < 2.5 mu m), for more than three months (Dec 2015-Mar 2016; N = 1953 h) using a distributed network of optical particle counters. The network consisted of seven nodes that were deployed in a residential urban area, five nodes in one neighborhood (similar to 1.5 km(2)) and two nodes in neighboring neighborhoods. While collocated with a reference monitoring instrument the sensors' readings were highly correlated (Pearson's r> 0.9; RMSE similar to 5 mu g m(-3)) and the variance of the observations when the reference PM2.5 measurements were <20 mu g m(-3) (similar to 90% of the records) was very low. Significantly higher heterogeneity was observed during the sensor deployment in the neighborhood, suggesting spatial variability of airborne particles at the neighborhood scale. Studying the spatial variability during different conditions (meteorological, day of the week, time of day) revealed signatures of human activity, suggesting specific sources that possibly contribute to the observed inner-neighborhood variability.