Degree: Doctoral Thesis
Authors: Lei Man Tat
Supervisors: Prof. David Gonçalves, University of Saint Joseph; Prof. Francisco Cardoso Ferreira, Universidade NOVA de Lisboa, Portugal
A combination of assessment, operational forecast, and future perspective was thoroughly explored to provide an overview of the existing air quality problems in Macao. The levels of air pollution in Macao often exceed those recommended by the World Health Organization (WHO). In order for the population to take precautionary measures and avoid further health risks during high pollution episodes, it is important to develop a reliable air quality forecast. Statistical models based on linear multiple regression (MLR) and classification and regression trees (CART) analysis were successfully developed for Macao, to predict the next day concentrations of NO2, PM10, PM2.5, and O3.
Meteorological variables were selected from an extensive list of possible variables, including geopotential height, relative humidity, atmospheric stability, and air temperature at different vertical levels. Air quality variables translate the resilience of the recent past concentrations of each pollutant and usually are maximum and/or the average of latest 24-hour levels. The models were applied in forecasting the next day average daily concentrations for NO2 and PM and maximum hourly O3 levels for five air quality monitoring stations. The results are expected to support an operational air quality forecast for Macao.
The work involved two phases. On a first phase, the models utilized meteorological and air quality variables based on five years of historical data, from 2013 to 2017. Data from 2013 to 2016 were used to develop the statistical models and data from 2017 was used for validation purposes. All the developed models were statistically significantly valid with a 95% confidence level with high coefficients of determination (from 0.78 to 0.93) for all pollutants. On a second phase, these models were used with 2019 validation data, while a new set of models based on a more extended historical data series, from 2013 to 2018, were also validated with 2019 data. There were no significant differences in the coefficients of determination (R2) and minor improvements in root mean square errors (RMSE), mean absolute errors (MAE) and biases (BIAS) between the 2013 to 2016 and the 2013 to 2018 data models. In addition, for one air quality monitoring station (Taipa Ambient), the 2013 to 2018 model was applied for two days ahead (D2) forecast and the coefficient of determination (R2) was considerably less accurate to the one day ahead (D1) forecast, but still able to provide a reliable air quality forecast for Macao.
To understand if the prediction model was robust to extreme variations in pollutants concentration, a test was performed under the circumstances of a high pollution episode for PM2.5 and O3 during 2019, and a low pollution episode during 2020. Regarding the high pollution episode, the period of the Chinese National Holiday of 2019 was selected, in which high concentration levels were identified for PM2.5 and O3, with peaks of daily concentration for PM2.5 levels exceeding 55 μg/m3 and the maximum hourly concentration for O3 levels exceeding 400 μg/m3. For the low pollution episode, the 2020 period of implementation of the preventive measures for COVID-19 pandemic was selected, with a low record of daily concentration for PM2.5 levels at 2 μg/m3 and maximum hourly concentration for O3 levels at 50 μg/m3.
The 2013 to 2018 model successfully predicted the high pollution episode with high coefficients of determination (0.92 for PM2.5 and 0.82 for O3). Likewise, the low pollution episode was also correctly predicted with high coefficients of determination (0.86 and 0.84 for PM2.5 and O3, respectively). Overall, the results demonstrate that the statistical forecast model is robust and able to correctly reproduce extreme air pollution events of both high and low concentration levels.
Machine learning methods maybe adopted to provide significant improvements in combination of multiple linear regression (MLR) and classification and regression tree (CART) to further improve the accuracy of the statistical forecast. The developed air pollution forecasting model may be combined with other measures to mitigate the impact of air pollution in Macao. These may include the establishment of low emission zones (LEZ), as enforced in some European cities, license plate restrictions and lottery policy, as used in some Asian, tax exemptions on electric vehicles (EVs) and exclusive corridors for public transportations.
Keywords: Air pollution; Particulate Matter; Ozone; Macao; Statistical air quality forecast; Pollution episodes; Chinese national holiday; COVID-19