BACKGROUND AND AIMS: The co-exposure to both traffic-related noise and air pollution may entail a mutual confounding effect of these environmental stressors on common health endpoints such as cardiovascular disease. To disentangle their effects we evaluated the correlation between measured nitrogen dioxide (NO2) concentrations and modelled noise levels (LAeq,24h) and between 15-minute ultrafine particle counts (UFP) and 15-minute noise measurements (LAeq) in the city of Girona.
METHODS: Long-term averages of NO2 were measured with passive samplers at 83 outdoor residential locations. These locations were assigned an LAeq,24h from a validated traffic noise model (NMPB-routes 96). Besides, we performed 15-minute simultaneous measurements of UFP and traffic-related noise at 31 sidewalk locations during non-rush traffic hours. These measurements were performed with a P-trak and a CESVA SC30 sonometre, respectively. Finally, we compared these results with other studies.
RESULTS: The NO2-noise correlation was 0.62 (Pearson). However, it was lower at sites with higher traffic density and in the downtown. Building density, distance to the street and traffic density were main determinants of the NO2 and noise variability. Besides, we could evaluate the relationship between measurements of UFP and traffic-related noise (spearman rank=0.55). In Chicago and Riverside, the Pearson correlation between 2-week NO2 and 5-minute noise measurements were 0.41 and 0.62, respectively, and between 5-min UFP and noise measurements it was 0.22 to 0.31 and 0.41, respectively (Allen et al, 2009). Other studies reported higher and lower correlations using modelled estimates and different air pollutants.
CONCLUSIONS: The specific characteristics of each city and each pollutant require evaluating this correlation in each study area prior to analysing the long-term effects of traffic-related noise and air pollution on potential common endpoints. Further studies should evaluate local traffic-related pollutants such as UFP and also compare modelled versus measured data and distinguish between low and high correlation locations.
