Combined model for PM10 forecasting in a large city

Abstract

We present the results of a PM10 forecasting model that has been applied for air quality management in Santiago, Chile during recent years. The daily operation of this model has served to inform in advance to the population about the air quality they will find in different areas of the city and to help environmental authorities in the decision to take actions on days when concentrations are in ranges considered significantly harmful and to impose restrictions to the activity of the city in advance, when extreme episodes are foreseen. At present, national PM10 standard for 24 h average is 150 μg m -3. According to the range where the concentrations fall, five levels or classes of air quality are defined: good (A), regular (B), bad (C), Critical (D) and Emergency (E). Forecasting is based on the combination of artificial neural networks and a nearest neighbor method. Inputs to the models are concentrations measured at several monitoring stations distributed throughout the city and meteorological information in the region. Outputs are the expected maxima concentrations for the following day at the site of the same monitoring stations. Results for last three years (2009, 2010, 2011) indicate that the model may be considered as an important tool for air pollution control. © 2012 Elsevier Ltd.