Air pollution modeling in São Paulo using bottom-up vehicular emissions inventories

Abstract

In this work, the impact of different vehicle emission inventory scenarios on air quality in the metropolitan areas of São Paulo, Baixada Santista, Vale do Paraíba, Sorocaba and Campinas is investigated. The construction of bottom-up vehicular emissions inventories is complex, being necessary to aggregate diverse information, such as the composition of the vehicle fleet (with the distribution of age, types of vehicles, types of fuel) and the emitting processes (emission factors for cold starts, exhaust and evaporative emissions). In addition, for air quality models, these emissions must still be distributed in time and space. The scenarios were constructed using different data sources, highlighting traffic simulations and vehicle counting of CET and SPTRANS, and records of vehicular displacements through GPS for the spatial-temporal distribution of vehicular flow. For the calculation of emissions, an open source software called VEIN (Vehicular Emissions Inventories, available at https://github.com/ibarraespinosa/vein) was developed. The simulated emissions for the metropolitan regions of São Paulo are larger than the emissions estimated by CETESB for all pollutants. From these scenarios, air quality simulations were performed with the WRF-Chem model. The results vary for different pollutants. In general, the daily variation of the pollutants is well simulated, showing that emissions are consistent. Despite the higher emission values found in this work, the simulated concentrations of the primary pollutants were, on average, lower than the observed concentrations. This is probably due to the fact that the simulated winds are stronger than the observed winds. This work shows new methods to develop emission inventories with different data providing a new approach to understanding air quality problems.