Health impact pathways related to air quality changes: testing two health risk methodologies over a local traffic case study

Air pollution causes damage and imposes risks on human health, especially in cities, where the pollutant load is a major concern, although the extent of these effects is still largely unknown. Thus, taking the busiest road traffic area in Portugal as a local case study (600 m × 600 m domain, 4 m spatial resolution), the objective of this work was to investigate two health risk methodologies (linear and nonlinear), which were applied for estimating short-term health impacts related to daily variations of high-resolution ambient nitrogen dioxide (NO) concentrations modelled for winter and summer periods. Both approaches are based on the same general equation and health input metrics, differing only in the relative risk calculation. Health outcomes, translated into the total number of cases and subsequent damage costs, were compared, and their associated uncertainties and challenges for health impact modelling were addressed. Overall, for the winter and summer periods, health outcomes considering the whole simulation domain were lower using the nonlinear methodology (less 27% and 28%, respectively). Spatially, these differences are more noticeable in locations with higher NO and population values, where the highest health estimates were obtained. When the daily NO exposure was less than 6 µg.m, a fact that occurred in 95% of the domain cells and in both periods, relatively small differences between approaches were found. Analysing the seasonality effect, total health impacts derived from the linear and nonlinear applications were greater in summer (around 18% in both approaches). This happens due to the magnitude and spatial variability of NO, as the other health input metrics remained constant. This exploratory research in local scale health impact assessment (HIA) demonstrated that the use of refined input data could contribute to more accurate health estimates and that the nonlinear approach is probably the most suitable for characterising air pollution episodes, thus providing important support in HIA.