An Analysis into the Temporal Variations of Ground Level Ozone in the Arid Climate of Makkah Applying k-means Algorithms

Abstract

Analysis of temporal variations helps identify the time when an air pollutant is more likely to exceed the air quality standards. This paper characterises the temporal variations of ground level ozone (O3) in Makkah. In addition to graphical presentations, like time variation plots, this paper applies k-means algorithm to test its applicability for analysing the diurnal cycles of O3. Diurnal, weekly and seasonal variations in O3 concentrations are analysed using data for 2012. O3 data for 1997 - 2007 are used to investigate the annual cycle ofO3. The average annual cycle shows the highest concentration in September and lowest in December. O3 concentrations are lower in colder months probably due to lower solar radiation levels, whereas unexpectedly highest O3 concentrations are observed in September. O3 levels decline in the hottest months - June and July probably due to chemical and biophysical feedbacks. Furthermore, O3 diurnal cycles are clustered into 4 and 12 subgroups using k-means algorithm. The clusters are considerably different from the monthly and seasonal diurnal cycles, probably because of the anomalies found in various months or seasons that biases the average cycles. These anomalies are addressed by k-means algorithm putting them into a separate group.

Keywords

• ground level ozone
• temporal variations
• k-means algorithm
• Makkah