WIT Press


Ground-level Ozone Forecast Based On Machine Learning

Price

Free (open access)

Volume

74

Pages

8

Published

2004

Size

276 kb

Paper DOI

10.2495/AIR040051

Copyright

WIT Press

Author(s)

R. Zabkar, J. Zabkar & D. Cemas

Abstract

In this paper we apply methods of machine learning to the problem of groundlevel ozone forecasting, using measured data and data calculated by the numerical weather prediction model ALADIN (Aire Limitee Adaptation Dynamique developement InterNational). Our goal is to build a simple ozoneforecasting tool to predict the daily maximum of ground-level ozone concentration per day using meteorological and air quality data. Tropospheric ozone episodes in Slovenia are mainly due to the local traffic sources and the long-range transport of ozone and its precursors, generally originating from Western Europe. Ozone forecast model was developed for the purpose of issuing public alerts to avoid exposure to high ground-

Keywords