Associated with the rapid economic and social development in Hong Kong and the neighboring Pearl River Delta (PRD) region, visibility impairment has become increasingly serious in recent years. The number of low visibility hours at Hong Kong increased steadily from 899 hours in 2000 to 2267 hours in 2007. Fine particulate, also called PM_2.5 (particulate matter less than 2.5 μm in diameter), is the known major culprit for visibility degradation. It is generally observed that the degree of visibility degradation increases with concentration level of PM_2.5. The annual average mass concentration of PM_2.5 in Hong Kong increased from 31.8 μg/m³ in 2000 to 46.9 μg/m³ in 2004, and then decreased to 38.6 μg/m³ in 2007. The increase in the number of poor visibility hours from 2004 to 2007 is unex...[ Read more ]

Associated with the rapid economic and social development in Hong Kong and the neighboring Pearl River Delta (PRD) region, visibility impairment has become increasingly serious in recent years. The number of low visibility hours at Hong Kong increased steadily from 899 hours in 2000 to 2267 hours in 2007. Fine particulate, also called PM_2.5 (particulate matter less than 2.5 μm in diameter), is the known major culprit for visibility degradation. It is generally observed that the degree of visibility degradation increases with concentration level of PM_2.5. The annual average mass concentration of PM_2.5 in Hong Kong increased from 31.8 μg/m³ in 2000 to 46.9 μg/m³ in 2004, and then decreased to 38.6 μg/m³ in 2007. The increase in the number of poor visibility hours from 2004 to 2007 is unexpected in light of the decrease in PM_2.5 during this period. This discrepancy has indicated that knowledge of PM_2.5 mass alone is not sufficient in assessing impact of PM on visibility degradation. This thesis work investigates potential reasons beyond PM mass, such as PM size and chemical composition, in impacting the visibility degradation. In addition, apportionment of visibility degradation to major aerosol source categories has been carried out and empirical formulas linking light extinction with aerosol constituents are established for the PRD region.

Source apportionment of light extinction at Tsuen Wan, an urban site in Hong Kong, to seven aerosol source categories was carried out by applying Positive Matrix Factorization (PMF) to PM_2.5 chemical composition data and visibility data in the periods of November 2000 - October 2001 and November 2004 - October 2005. The seven sources and their respective contributions to light extinction are: secondary sulfate (31±9%), mixture of biomass burning / industrial combustion emissions (BBIC) (23±6%), automobile exhaust (22±8%), secondary nitrate (10±5%), sea salt (10±7%), and residual oil combustion (3±4%). Between the two campaign years (2000/01 vs. 2004/05), the contributions to light extinction by the top three sources have changed. The contribution of BBIC dropped from 27±7% (79.1 Mm^-1) in 2000/2001 to 19±5% (65.7 Mm^-1) in 2004/2005; the contribution of automobile exhaust decreased from 26±9% (74.9 Mm^-1) in 2000/2001 to 15±7% (51.9 Mm^-1) in 2004/2005; however, the contribution of secondary sulfate increased from 25±8% (73.1 Mm^-1) in 2000/2001 to 36±10% (121.4 Mm^-1) in 2004/2005. The source apportionment results indicate that the control strategies on vehicular emissions and industrial combustion emissions have taken effect and that the formulation and implement of control strategies on SO₂ emissions need to be strengthened in order to lower the ambient PM_2.5 levels and improve visibility in Hong Kong.

Examination of high PM hours under conditions of similar PM_2.5 mass and RH reveals significant differences in light scattering coefficient (B_sc) between winter and summer. Similar observations were found in the daily samples, for which detailed chemical composition are available. A case study using two daily samples 20041103 and 20050812 was conducted. The ISORROPIA-Mie model was used to examine how aerosol size distributions and the mixing state of elemental carbon (EC) could significantly impact light extinction by aerosols. The modeling results show that size variation in the droplet-mode size of aerosols (in the range of 0.5-1.2 μm, near visible wavelength) may cause a change in B_sc larger than 50% for aerosols having the same bulk PM_2.5 mass and chemical composition. The different mixing state of EC (external versus internal mixture) could also contribute to the relative difference of B_sc as large as 34%. Weather chart, wind field data and back trajectory analysis as well as chemical tracers show that the summer and winter samples examined here were influenced by air masses of different origins. The winter samples were influenced by air masses from the northeast of Hong Kong. The summer samples were influenced by air masses from the northeast of Hong Kong, where residual oil combustion emissions related to container ports were more prominent. The different origins of air masses provide physical basis for possibility of different size and mixing state of aerosols between the summer and winter samples.

The long-term record of PM and visibility at the urban site of Tsuen Wan and the suburban site of Tung Chung from 2000 to 2007 were analyzed using the ISORROPIA-Mie model. The aim is to explain the non-linear relationship between annual PM concentration and the number of visibility degradation days. The increase of B_ext (272 to 333 Mm^-1) from 2000 to 2004 at Tung Chung was mainly caused by the increase in the PM mass concentration (44.9 to 62.1 μg/m³). From 2004 to 2007, although the mass concentration decreased (62.1 to 54.1 μg/m³), the annual B_ext stayed at a high level (~ 325 Mm^-1) due to the increase contributions from sulfate (16% to 23%) and organic matter (25% to 31%). At the Tsuen Wan site, the increase of mass concentration (49.8 to 62.7 μg/m³) was largely responsible for the increase of annual B_ext from 2000 to 2004 (280 to 334 Mm^-1). The slightly decrease of annual B_ext from 2004 to 2007 (334 to 304 Mm^-1) was more related to the variation of aerosol size, which likely occurred due to reduction in local vehicular emissions and increase in source types at the urban site.

The IMPROVE formulas developed in the US was found to significantly overestimate B_ext at Guangzhou and Nansha, two sites in the PRD. A new empirical formula to link light extinction with bulk aerosol chemical constituents is established for the PRD region. Gauss-Seidel iteration and Multiple Linear Regression were used to construct the new empirical formula. It was found that the formula with exponent added on OM term can provide the best estimate for B_ext, with R² of 0.85 and 0.89 and the slope values of 0.99 and 0.70 for Guangzhou and Nansha data, respectively. The applicability of the new formula over the PRD area was tested on PM chemical speciation data in Hong Kong. Compared with the IMPROVE formulas, the new formula developed in this thesis work can more accurately estimate B_ext, evidenced by the mean fractional bias <= 10% for the lowest 75% B_ext. The new formula underestimates B_ext by 30% for the largest 25% B_ext.