Improving the modeling of road dust levels for Barcelona at urban scale and street level

Road dust emission is an emerging issue in air quality due to the lack of remediation measures in contrast to vehicle exhaust emissions. The evidence of receptor modeling studies allows for quantifying impact on a few receptors, but the high cost of PM chemical speciation data and the questionable representativeness of single monitoring sites, limit considerably the development of population exposure estimates and epidemiologic studies based on georeferenced data. This study attempts to initiate and promote urban-scale dispersion modeling for road dust emissions, which will allow for a more robust estimate of population exposure and health outcomes. The TNO URBIS (URBan Information System) model was applied in the city of Barcelona, implementing a Gaussian line source and a street canyon dispersion model, together with new experimental estimates of road dust emission factors and algorithm to describe the time variability. Annual, daily and hourly road dust contributions were simulated and validated against observation of PM10, mineral dust and hourly PM2.5-10 concentrations. Results show that road dust contributed 9-15% to PM10 levels at background sites, and 23-44% at traffic sites. Highest contributions were modeled in the commercial/residential district where most of population live and work (Eixample) structured by 120 m wide square blocks, separated by roads with >10,000 vehicles per day. Street level contributions rise up to 20 μg/m3 (96% of roads) and an additional 3% of roads within 20-40 μg/m3. Hourly simulations of road dust contributions revealed to benefit from the implementation of the new emission module (Amato et al., 2012), able to describe the exponential recovery of road dust emission potential after rain events, when compared to common approach such as the use of constant emission factor or an ON/OFF approach. Correlation coefficients with observed data varied from 0.61, 0.58 and 0.43 for annual, daily and hourly means, respectively, revealing a clear improvement in terms of both spatial and temporal variability. However, more efforts need to be done in validating the model in different climatic scenarios and evaluating the seasonal variation of road dust emissions, due to droughts or Saharan dust events.