Quantifying prediction and intervention measures for PM_2.5 by a PDE model

Increasingly series global air pollution has become a focus of global concern, and the high accuracy forecasting for PM_2.5 is difficult due to its large fluctuations. Thus, this paper aims to provide a forecasting model for PM_2.5 concentration, along with quantifying the effects of potential intervention measures for pollution control. To achieve this goal, a specific partial differential equation (PDE) model is proposed using network, clustering, and real daily data. The proposed model describe the combined influence of local emission, transboundary transmission, and human prevention, which helps high accuracy predictions. The results show that the prediction accuracy is well acceptable, up to average 95% and 90% based on our accuracy measures. Furthermore, the daily PM_2.5 concentrations of all-city clusters under different intensities of interventions are forecasted by performing sensitivity analysis. Simulation results verify that stricter control measures result in lower PM_2.5 concentration; joint-control of air pollution is more efficient due to the reality of transboundary pollution.