TIME SERIES ANALYSIS OF NOX, O3, AND HYDROCARBON IN AMBIENT AIR AT DKI JAKARTA

• Main Author: Jiwandono, Kandu
• Format: Theses
• Language: Indonesia
• Subjects: Teknik saniter dan perkotaan; teknik perlindungan lingkungan
• Online Access: https://digilib.itb.ac.id/gdl/view/37223
• Institution: Institut Teknologi Bandung

DKI Jakarta is one of the big cities in Indonesia with the largest number of vehicles. With the number of vehicles predicted to continue to increase, there are various kinds of problems, e.g. air pollution caused by O3. Highly concentrated O3 can causes respiratory irritation and even death. UDARA (Urban hybriD models for AiR pollution exposure Assessment) was conducted to determine the relationship between air pollution and health quality in DKI Jakarta. One of the challenges in UDARA is the lack of coverage of air pollution data in DKI Jakarta even though there have been 5 automatic air quality monitoring stations. One solution offered is to use a time series model which can be expected to predict concentration at a certain time. Time series analysis and analysis of spatial variations were conducted on O3, NOX, and hydrocarbons based on 30 minutes average data from the five stations. The main objective of this study was to analyze the pollutant characters which is O3 and its precursors at each monitoring station. The analysis in this study are analysis of spatial variation with ANOVA and ANOVA post-hoc tests, trend analysis with Mann-Kendall test and visual observation, ARIMA time series calculation (Autoregressive Integrated Moving Average), and concentration forecasting in the following year based on the calculated ARIMA model. The ANOVA test shows that there are significant mean differences of all pollutant concentration throughout the monitoring stations. In general, there are a negative trend for NO2, NO, and NOX concentrations and a positive trend for O3 concentrations at each monitoring station. However, the Mann-Kendall test indicates a negative trend of O3, NO2, NO, NOX, and hydrocarbons in each monitoring station. The differences in trend analysis results is caused by the Mann-Kendall test cannot detect any trend fluctuations. The forecasting results shows that the maximum monthly average value of the 8 hour daily O3 concentration is 151.69 ?g/m3. This value exceeds the air quality guideline value issued by WHO. The accuracy of the ARIMA model was analyzed by its Root Mean Squared Error (RMSE) and Mean Absolute Percentage Error (MAPE) value. The biggest MAPE value of the O3 parameter is 18.1%. This value is good enough compared and with other similar studies with more data availability. This predicted value can be used as a reference in policy making by the stakeholder.