Application of the Air Quality Forecasting Analysis Using NARX Models

Abstract

Air quality management and forecasting play a crucially important role in environmental problems. It is known that air quality problem is directly related to the quality of life and human health. In order to solve this problem, there are some conventional forecasting methods used in the literature. This paper presents a new non-linear autoregressive exogenous model method. In this method, all air quality parameters are entered into the system for four different locations. These are Çanakkale Central and the districts of Çan, Lapseki and Biga. This created model provides obtaining and extracting of some unmeasured environmental pollutant parameters for other air quality stations such as Nitric oxide (NO), Nitrogen oxide (NO2), Nitrogen oxides (NOX) and Ozone (O3). Within these stations, the Çanakkale Central air quality monitoring station measures only Particulate matter (PM10) and Sulfur dioxide (SO2) parameters while others measure the parameters of PM10, PM2.5, SO2, NO, NO2, NOX and O3. Presented numerical model results are verified with measurement results and extracted acceleration error. These numerical results are realized for Çanakkale Central. Obtained results show that the forecasted parameter values are very successful and error acceleration is very low. The success of the learning process is over 90%.