Intercomparison of air quality models in São Paulo, towards an operational ensemble forecasts system

Predicting air quality in megacities is challenging due to the diversity and variability of emission sources, as well as the specific meteorology and photochemistry occurring in the urban boundary layer.

São Paulo is by far the largest city in South America, one of the biggest megacities of the world, located near the coast and on a plateau at about 800 m above sea level, in a tropical climate. A megacity such as São Paulo is therefore a challenge for regional air quality models, which must be used at a resolution high enough to sufficiently accurately represent the processes leading to the high concentrations and high diurnal variability of the main pollutants. On the other hand, the measurement network is composed of 26 stations within the metropolitan area and another 63 within the state of São Paulo mostly in or near other cities, which constitutes an excellent support for evaluating the model outputs.

In this study, we assess the strengths and weaknesses of modeled concentrations of regulated pollutants (CO, O3, NO2, PM2.5, PM10), over three contrasting time periods in 2019. Four Chemistry-Transport models are involved in this intercomparison of high-resolution modeling results, less than 5 km. We study primary pollutants, meteorology, photochemistry as well as the performance of ozone and PM2.5 alerts when WHO air quality standards are not met. The results show that all models have good performance depending on the period and pollutants, and the performance of multi-model median is the best, as has already been shown for other regions.

In the framework of the KLIMAPOLIS project, the perspective of our study is to build an operational air quality forecasting system for the São Paulo region based on ensemble forecasts.