Air quality evaluation in selected areas of Sydney and Kigali and the proposal of pollutants prediction models

Abstract

This thesis provides an evaluation of Air Quality in selected areas of Sydney and Kigali using historical and measured data, respectively. In addition to the evaluation, the obtained data is used to develop mathematical models for the prediction of dependent air pollutants using Multi Linear and Principal Component Regression. Historical data from monitoring stations for four pollutants, PM2.5, PM10, NO2 (nitrogen dioxide), and SO2 (sulphur dioxide), in North Parramatta was obtained from the Department of Planning, Industry and Environment. A comparative analysis was conducted to evaluate the levels of these pollutants with respect to Australia and World Health Organisation Air Quality Standards in the periods before (2018 and 2019), during (2020 and 2021), and after (2022, 2023 and 2024) the COVID-19 pandemic. The comparative analysis showed that in the pre Covid-19 period, PM2.5 and PM10 exceeded the recommended levels in 438 (2018 and 2019) recordings. This level decreased to 164 times during COVID-19 times (2020 and 2021) and continued to decrease post Covid-19 (2022 and 2023) to 103 times.

Although the data shows a trend of decrease in number of times the pollutants exceeded the recommended values as well as the recorded spikes reduction in pollution patterns compared to pre and during COVI-19 times, Parramatta North Air Quality still experiences significant pollutant spikes as evidenced by recorded data for the period under research. During the pandemic times (2020 and 2021) the pollution frequency reduced but the spikes persisted. In 2020, PM2.5 and PM10 recorded maximum values of 170.9 μg/m³ and 803.7 μg/m³ which are 6.8 times and 16 times the recommended values, respectively. In 2021 and 2022, these maximum values of PM2.5 and PM10 decreased to 101.4 μg/m³ and 185.2 μg/m³, respectively. Nevertheless, these values are still 4 times and 7.4 times the recommended values, respectively.

Finally, the acquired data for North Parramatta and Kigali were used to develop mathematical models to predict the levels of f PM10 as a dependent pollutant from levels of PM2.5, SO2, NO2, and PM1 as independent pollutants. For Parramatta North, a Principal Component Regression model was used to predict PM10 against PM2.5, SO2, and NO2 while for Kigali, a Multi Linear Regression model was used to predict PM10 from levels of PM1 and PM2.5. Results of both models show a positive correlation between the dependent and independent pollutants.

Date of Award	2024
Original language	English
Awarding Institution	Western Sydney University
Supervisor	Robert Salama (Supervisor)