Since prolonged exposure to elevated ozone (O3) concentrations is known to be harmful to human health, appropriate control strategies for ozone are needed for the non-attainment area such as Seoul, Korea. The goal of this research is to assess factors linked with the 1-hour ozone exceedance through a decision tree model. Since ozone is a secondary pollutant, lag times between ozone and explanatory variables for ozone formation are taken into account in the model to improve the accuracy of the simulation. Results show that while ozone concentrations of the previous day and NO2 concentrations in the morning are major drivers for ozone exceedances in the early afternoon, meteorology plays more important role for ozone exceedances in the late afternoon. Results also show that a selection of lag times between ozone and explanatory variables affect the accuracy of predicting 1-hour ozone exceedances. The result analyzed in this study can be used for developing control strategies of ozone in Seoul, Korea.

1. Introduction1
 2. Materials and Methods
 2.1. Data
 2.2. Decision tree algorithm
 3. Results and Discussion
  3.1. Spatial and temporal variations of air pollutionand meteorology
  3.2. Correlation between ozone and its potentiallyformative variables
  3.3. Lag time between ozone and its formativevariables
 4. Conclusions
 Acknowledgement
 REFERENCES

• Sun-Kyoung Park(Hanyang Cyber University) Corresponding author