Forecasting of Daily PM10 Concentrations in Brno and Graz by Different Regression Approaches

In Brno and Graz, winter PM10 concentrations frequently surpass the EU limit of 50 µg/m³ due to poor air dispersion conditions. This study demonstrates that reliable daily PM10 forecasts can be generated using multiple linear regression and generalized linear models, incorporating both current measurements and next-day meteorological data. Both approaches yield similarly accurate predictions, as assessed by multiple quality measures. These models could inform future regulatory decisions, such as traffic restrictions, by providing actionable short-term air quality insights.

Stadlober, E., Hübnerová, Z., Michálek, J., & Kolář, M. (2016). Forecasting of Daily PM10 Concentrations in Brno and Graz by Different Regression Approaches. Austrian Journal of Statistics, 41(4). https://doi.org/10.17713/ajs.v41i4.169