Nonlinear Models to Forecast Ozone Peaks
In this paper, the problem of forecasting tropospheric ozone concentration is approached by means of nonlinear black-box modeling techniques1. In particular Neuro-Fuzzy, Auto Regressive Ciclostationary and Nonlinear Set Membership models are identified and tested in the one day ahead forecast of daily maximum ozone concentration in Brescia, a highly populated and industrialized area in the Po Valley (Northern Italy). The model performances are assessed by means of indexes and statistical indicators suggested by the European Environment Agency.
KeywordsOzone Concentration European Environment Agency Auto Regressive Hourly Concentration Ozone Episode
Unable to display preview. Download preview PDF.