NCER Working Paper Series

2017

• #117
  Download full text

  Keywords:

  Air quality, Particulate matter, Dynamic multiple equations

  A Dynamic Multiple Equation Approach for Forecasting PM2.5 Pollution in Santiago, Chile

  Stella Moisan, Rodrigo Herrera and Adam Clements

  A methodology based on a system of dynamic multiple linear equations is proposed that incorporates hourly, daily and annual seasonal characteristics to predict hourly pm2.5 pollution concentrations for 11 meteorological stations in Santiago, Chile. It is demonstrated that the proposed model has the potential to match or even surpass the accuracy of other linear and nonlinear forecasting models in terms of fit and predictive ability. In addition, the model is successful in predicting various categories of high concentration events, up to 76% of mid-range and 100% of extreme-range events as an average across all stations. This forecasting model is considered a useful tool for government authorities to anticipate critical episodes of air quality so as to avoid the detrimental impacts economic and health impacts of extreme pollution levels.