Analyses of atmospheric pollutants in Hong Kong and the Pearl River Delta by observation-based methods

Despite continuous efforts paid on pollution control by the Hong Kong (HK) environmental authorities in the past decade, the air pollution in HK has been deteriorating in recent years. In this thesis work a variety of observation-based approaches were applied to analyze the air pollutant monitoring data in HK and the Pearl River Delta (PRD) area. The two major pollutants of interest are ozone and respirable suspended particulate (RSP, or PM10), which exceed the Air Quality Objective more frequently. Receptor models serve as powerful tools for source identification, estimation of source contributions, and source localization when incorporated with wind profiles. This thesis work presents the first-ever application of two advanced receptor-models, positive matrix factorization (PMT) and Unmix, on the PM10 and VOCs speciation data in HK.; Speciated PM10 data were collected from a monitoring network in HK between July-1998 and Dec-2005. Seven and nine sources were identified by Unmix and PMF10, respectively. Overall, secondary sulfate and vehicle emissions gave the largest contribution to PM10 (27% each), followed by biomass burning/waste incineration (13%) and secondary nitrate (11%). Sources were classified as local and regional based on its seasonal and spatial variations as well as source directional analysis. Regional sources accounted for about 56% of the ambient PM10 mass on an annual basis, and even higher (67%) during winter. Regional contributions also showed an increasing trend, with their annual averaged fraction rising from 53% in 1999 to 64% in 2005. The particulate pollution in HK is therefore sensitive to the regional influence and regional air quality management strategies are crucial in reducing PM level in HK. On the other hand, many species with significant adverse health impacts were produced locally. Local control measures should be strengthened for better protection of public health.; Secondary organic carbon (SOC) could be a significant portion of OC in particles. SOC was examined by using PMF-derived source apportionment results and estimated to be sum of OC present in the secondary sources. The annual average SOC in HK was estimated to be 4.1 mugC/m3 while summertime average was 1.8 RgC/m3 and wintertime average was 6.9 ggC/m 3. In comparison with the SOC estimates by the PMF method, the method that uses elemental carbon (EC) as the tracer for primary OC to derive at SOC overestimates by 78-210% for the summer samples and by 9-49% for the winter samples. The overestimation by the EC tracer method was a result of incapability of obtaining a single OC/EC ratio that represented a mixture of primary sources varying in time and space. It was found that SOC and secondary sulfate had their seasonal variation in sync, suggesting common factors that control their formation. The close tracking of SOC and sulfate appears to suggest that in-cloud pathway is also important for SOC formation.; Speciated VOCs were obtained in four air quality monitoring stations (AQMSs) in HK from August-2002 to August-2003. Both Unmix and PMF identified five stable sources. Mixed solvents gave the largest contributions ranging from 34% at rural Tap Mun to 52% at urban Central/Western. The wind directional analysis indicates the main source location at the central PRD area. Regional transport accounts for about 19% of the total VOC, while the two local and vehicle-related sources are responsible for 27%. By weighing the abundance and reactivity of each VOC species, mixed solvent use is estimated to be the largest contributor of local ozone, with the contributions ranging from 42% at Tung Chung to 57% at Tap Mun. The next largest is the vehicle exhaust, accounting for about 28% in Yuen Long. Biogenic emission is responsible for nearly 20% of the ozone generation at Tap Mun but this figure is likely underestimated.; Distinct secondary inorganic aerosol (SIA) responses are expected to the reduction of different precursors as a result of no...