Effects of tropical cyclone on air pollution in Hong Kong

Effects of Tropical Cyclone on Air Pollution in Hong Kong by

Tuonan Li

B.Sc., University of Victoria, 2017

A Thesis Submitted in Partial Fulfillment of the Requirements for the Degree of

MASTER OF SCIENCE In the Department of Geography

©Tuonan Li, 2020 University of Victoria

All rights reserved. This thesis may not be reproduced in whole or in part, by photocopy or other means, without the permission of the author.

We acknowledge with respect the Lekwungen peoples on whose traditional territory the university stands and the Songhees, Esquimalt and WSÁNEĆ peoples whose historical

ii

Effects of Tropical Cyclone on Air Pollution in Hong Kong by

Tuonan Li

B.Sc., University of Victoria, 2017

Supervisory Committee

Dr. David Atkinson, Supervisor Department of Geography

Dr. Chris Bone, Departmental Member Department of Geography

iii Abstract

Climate and weather play a significant role in patterns of air pollution occurrence and severity. An analysis of the effect of weather on pollution parameters in Hong Kong was performed. Hong Kong is one of the world's most densely populated regions and air pollution can be problematic, which is a serious public health concern. Hong Kong is impacted by Tropical Cyclones which strongly affect weather patterns. In this research, a twelve-year record (2007-2018) of tropical cyclone (TC) and pollutant concentrations (carbon monoxide, ground-level ozone, nitrogen dioxide, sulfur dioxide, and particulate matter) were analyzed to investigate the effects of TC on air quality. It is found that the occurrences of TC are strongly related to days with elevated particulate matter, sulfur dioxide and carbon monoxide concentrations (above 90th percentile), and low concentrations (below the 10th percentile) for nitrogen dioxide. In particular, the spatial location of TC with respect to Hong Kong is found to be clearly associated with high or low pollutant concentrations. When the TC is located to the North/Northeast of Hong Kong, the air quality tends to be poor because polluted air from mainland China is advected over the city. Conversely, TC located to the West resulted in good air quality by ventilating the city with relatively clean air from the ocean.

iv Table of Contents Supervisory Committee ... ii Abstract ... iii Table of Contents ... iv List of Tables ... v List of Figures ... vi 1 Introduction ... 1 1.1 Air Pollution ... 1 1.2 Previous work ... 3 1.3 Study Area ... 4 1.4 Objectives ... 6 2 Methods... 7

2.1 Pollutant concentration data ... 7

2.2 The TC data ... 10

2.3 Air pollution emission trend ... 11

3 Results ... 13

3.1 Mean air pollution pattern ... 13

3.2 TCD vs. NTCD ... 17

3.3 Spatial distribution of TC ... 19

3.4 Pollutant emission trends ... 21

4 Discussion ... 23 4.1 Tropical cyclone ... 23 4.2 Pollutant emission ... 25 5 Conclusion ... 27 6 Reference ... 28 Appendix A ... 32 Appendix B ... 33

v List of Tables

Table 1. The data availability and classification of air quality monitoring stations in Hong Kong. ... 7 Table 2. The percentile classification of pollutant concentration data in Hong Kong. ... 9 Table 3. The average of annual pollutant concentration (μg/m³) collected from different types of stations in Hong Kong (↑=highest, ↓=lowest). ... 15 Table 4. The result of two-sample assuming unequal variances t-test (α=0.05) on pollutant

concentration (μg/m³) during TCD (sample size=236) and NTCD (sample size=1604). ... 17 Table 5. The percentile classification of pollutant concentration with the amount of TCD out of total days and its percentage. ... 18

vi List of Figures

Figure 1. The location of Hong Kong. ... 5 Figure 2. The location of 16 air pollution monitoring stations in Hong Kong. ... 7 Figure 3. The surrounding land use of the four types of air quality monitoring stations in Hong Kong (photo provided by EPD). ... 9 Figure 4. The daily location of TC events within the 800-kilometre buffer around Hong Kong from 2007 to 2018. Taiwan is the island to the east of the region and Hainan Island is located to the west of the region. ... 11 Figure 5. The mean concentration (μg/m³) of six major air pollutants in Hong Kong, 2007-2018. These plots integrate data from all 14 stations used in this study. ... 14 Figure 6. The mean monthly pollutant concentration (μg/m³) averaged by 4 station types from 2007-2018 in Hong Kong. ... 15 Figure 7. The histograms of daily pollutant concentration (μg/m³) from May to September in rural station in Hong Kong, 2007-2018. Histogram categories reflect the maximum values in each pollutant dataset. ... 17 Figure 8. The spatial distribution of TCs (yellow dot) around 800km distance (include Hainan on the west and Taiwan on the east) from Hong Kong (black dot). Columns 1-5 are arranged by percentile classifications (1= extremely low, 2=low, 3=medium, 4=high, 5=extremely high). Rows A-F are arranged by pollutant type (a= PM₂.₅ , b= PM₁₀, c= NO₂, d= SO₂, e=CO, f= O₃). ... 21 Figure 9. The Hong Kong air pollutant emission (tonnes) inventory from 2001 to 2017 (without the emission of hill fire). ... 21 Figure 10. The proportion of pollutant emission from different sources in 2017 (1: major

contributing sources include non-road mobile machineries operating in construction sites and container terminals. 2: the major sources are paved road dust, cooking fume, construction dust and quarry production for PM₂.₅ and PM₁₀ emission; paints and associated solvents, consumer products and printing for VOC emission (Environmental Protection Department, n.d.)). ... 22

1 Introduction

Tropical Cyclones (TC), also known as typhoons in the northwestern Pacific Ocean, are low-pressure systems formed in the barotropic ocean environments of the lower latitudes. They are characterized by low central pressures, corresponding powerful winds, and moderate horizontal extent (<1000km). TCs bring strong wind and precipitation along their track, which slowly disperses when they encounter land. To date, many studies have investigated the

influence of TC on the local air quality at various locations, because TCs not only create extreme weather within its wind field, but also strongly affect the atmosphere circulation in its vicinity, which can then affect pollutant concentrations (Huang et al., 2005; Lam et al., 2005; Wei et al., 2007; Wang et al., 2009; Yang et al., 2012; Chow et al., 2018). The Hong Kong Special

Administrative Region (Hong Kong), located on the south China coast at Pearl River Delta, has an annual TC season. This region also has been having air pollution issues for decades – because of its high population density and urbanization, the citizens in Hong Kong have suffered high exposure to air pollution for many years (Deng et al., 2008; Tang et al., 2018).

1.1 Air Pollution

Air pollution is a subjective concept that describes the impurity of air when it contains different components produced by human activities, as compared to clean air (Brimblecombe, 2011; Stern et al., 1973). There is considerable evidence of adverse human health effects due to environmental and occupational exposures to air pollution, which have led to the development of air quality monitoring and regulations that aim to reduce emissions. In Hong Kong, the

concentrations of sulphur dioxide (SO₂), nitrogen dioxide (NO₂), fine suspended particulates (PM₂.₅), respirable suspended particulates (PM₁₀), ground level ozone (O₃) and carbon monoxide (CO) are constantly monitored across the area. All of these pollutants have been demonstrated to pose a threat to human health. For example, suspended particulates, which are

2

also known as particulate matter (PM), is a mixture of solid particles and liquid droplets with different diameters. The common pollutants, PM2.5 (diameter ≤ 2.5 μm), and PM10 (diameter ≤10 μm), were found to increase mortality over both short- and long-term, and cause damage to the respiratory and cardiovascular system (Brook et al., 2004). Ozone (O₃) can not only cause acute effects on pulmonary and cardiovascular systems, but can result in chronic effects, such as the development of asthma and atherosclerosis (Park et al., 2005; Rich et al., 2006). The pulmonary function can also be damaged by exposure to certain amounts of NO₂ and SO₂, which are toxic gases the effect of which are greater on children (Hussain et al., 2004; Anderson et al., 1997). CO, another toxic pollutant, can cause hypoxia which is harmful to high oxygen consumption organs and newborn infant (Laties & Merigan, 1979; Longo, 1977).

All of the six pollutants except O₃ are generated by human activities and natural events (hillfire); these are defined as primary pollutants. When the pollutants are modified by chemical reactions in the atmosphere, the resulting species are termed secondary pollutants. More

specifically, secondary pollutants are created by the chemical reactions between the pollutants in the atmosphere. Chemical reactions in the atmosphere are the main source for O₃, and it can represent an additional source for the other pollutants considered in this study, including particulate matter. Due to the complexity of the reactions and the mobility of pollutants in the atmosphere, pollutant concentrations can vary greatly over time and place and so are constantly being monitored and controlled to provide a healthier environment for the citizens. To provide guidance in reducing pollution effects on humans, the World Health Organization (WHO) has produced air quality guidelines for concentration standards of the “major pollutants” – SO₂, NO₂, PM₂.₅, PM₁₀, O₃ and CO – based on the current scientific evidence (WHO, 2016). WHO provides concentration targets which are implemented in Hong Kong as the standard. A

Canadian health-related pollution representation, the Air Quality Health Index (AQHI), was also adopted in Hong Kong. This approach uses a 3-hour moving average of pollutant concentration

3

data from monitoring stations to estimate health risks and provide pollution advisory information for the public (Stieb et al., 2008; Wong et al., 2012).

1.2 Previous work

Early studies of air pollution were complicated by the variety of pollutants and the number of stationary and mobile sources; this led to studies that focused more on the chemistry of air pollution (Stern et al., 1973). However, there were research efforts that identified the importance of meteorological condition on air pollution which revealed another crucial factor governing pollutant concentration—the synoptic setting (Firket, 1936; Schrenk, 1949; Wilkins, 1954; Reith, 1951). Stable and descending air concentrate the pollutants in the ground, while air that is being strongly circulated by convective (rising) activity or strong pressure gradients will transport and disperse pollutants. Therefore, understanding the nature of air motion during particular synoptic events could help to forecast the concentration of pollutants in the region. For instance, rapidly moving air, such as associated with midlatitude cyclones, was found to

effectively transport pollutants even across the ocean (Ding et al., 2009).

The importance of TC has been recognized as a major synoptic driver for most of coastal China. Despite the size and importance of Hong Kong, however, there are only a few studies focusing on the impacts of TC on air pollutant concentration on the city. Over (southern) China TC can develop a variety of wind fields based on the relative location of the TC. Other studies have found similar results for other areas; for example, the location of a TC with respect to Beijing can exert various impacts on its air quality (Wang et al., 2009). There are a number of studies investigating and modelling the relationship between TC and pollution events, especially smog events, which are mainly due to high ozone concentrations during the TC season along the south China coast ( Huang et al., 2005; Lam et al., 2005; Wei et al., 2007). Other research examined TC driven pollution episodes in the context of hourly variations on both air pollutant

4

concentrations and the atmosphere conditions to explore the reaction and transportation mechanism of the pollutants (Yang et al., 2012).

More recently, Chow et al. (2018) conducted a study of the influence of northwestern Pacific TC on air quality in Hong Kong. The study examined particulate matter and ozone from 2000 to 2015, classifying their daily concentration during the study period into different groups based on values and compared this to the pattern of TC “active days”. The spatial pattern of TC as well as the corresponding atmospheric conditions were considered. The outcome of the study indicated that TC exert an influence on the particulate matter and ozone concentrations as a function of their location with respect to Hong Kong. The classification method of Chow et al.’s study was referenced as a starting point for this study to understand the TC impacts on all six major pollutants in Hong Kong.

1.3 Study Area

Hong Kong is located on the Southeast coast of China. It is contiguous to mainland China in the North and surrounded by the ocean in the other three directions (Fig. 1). Hong Kong is classified as a humid subtropical climate, which indicates hot and humid summer with mild winter (Ackerman, 1941). The region experiences regular TC activity from the western North Pacific Ocean and South China Sea during summer and fall. On average, 30 TCs form in the West Pacific area annually; at least three of these will go on to affect Hong Kong between July and September (Hong Kong Observatory, n.d.). Combined with the summer monsoon from the Indian Ocean, atmosphere circulation in Hong Kong is fairly active, creating a moist

5 Figure 1. The location of Hong Kong.

Topographically, Hong Kong is a mountainous region with steep slopes which limits the availability of land suitable for development. As a result, the expansion of urban areas, new towns and other developments has been located mainly on reclaimed land which accounts for ~25% of Hong Kong’s total land area (Development Bureau, 2017). The contradiction between the lack of available space and high levels of urbanization lead to a high density of commercial, residential and transportation land use (Tong & Wong, 1997). With one of the highest population densities in the world, the Hong Kong citizens have relatively high exposure to locally produced air pollution (Census and Statistics Department, 2012). In response, the city introduced

regulations to limit vehicle emissions and volatile organic compounds (VOCs). Since these controls were first implemented in 2007, the air quality in the region has improved

6 1.4 Objectives

Air pollution is one of the major urban issues in modern times due to its significant damage on the health of living organisms. The dense urban landscape, active weather system and varied pollutant sources result in a complex air pollution scenario in the local area of Hong Kong. A key, episodic weather event that exerts a strong impact on pollutant concentrations in Hong Kong are tropical cyclones. In this study the impact of TC occurrence on air pollution is examined.

Specifically, the purpose of this research is to develop an air pollution climatology during the TC seasons using the hourly pollutant concentration data in Hong Kong and daily TC records at the western Pacific Ocean. To satisfy this mandate key objectives are identified:

1. Develop a general air pollution pattern in Hong Kong

2. Classify TC events into categories defined by the TC variables and the corresponding pollutant concentration values.

7 2 Methods

2.1 Pollutant concentration data

The Environment Protection Department (EPD) of Hong Kong operates 16 monitoring stations around Hong Kong (Fig. 2). Each monitoring station measures the hourly concentration level of SO₂, NO₂, PM₂.₅, PM₁₀ and O₃, and some include CO. The earliest stations have been collecting data since 1990, and the most recent station was established in 2017 (Table 1). The time frame for this study is from 2007-2018. As a result, the data from the most recently

established stations Tseung Kwan O and Tuen Mun are not applied in this study due to the short duration of their records. The decision to start in 2007 was made to coincide with the initiation of pollution control regulations. The hourly pollutant concentration data of the other 14 stations were obtained from the EPD website. All the data were measured in μg/m³ and stored in comma-separated values file per station annually. Then the files were combined to create datasets of hourly concentration in all 14 stations for the six pollutants.

Figure 2. The location of 16 air pollution monitoring stations in Hong Kong.

Table 1. The data availability and classification of air quality monitoring stations in Hong Kong. Station Name Start Year Station Type Station Name Start Year Station Type

Central/Western 1990 Urban Eastern 1999 Urban

8

Sham Shui Po 1990 Urban Sha Tin 1992 New Town

Tai Po 1991 New Town Tap Mun 1999 Rural

Tseung Kwan O 2017 Urban Tsuen Wan 1990 Urban

Tuen Mun 2014 New Town Tung Chung 2000 New Town

Yuen Long 1996 New Town Causeway Bay 1998 Roadside

Central 1999 Roadside Mong Kok 1992 Roadside

A preliminary examination of the mean of twelve months of pollution patterns was performed. Data from May to September were retained for examination in the context of the TC analysis. These months were chosen to specifically exclude late fall, winter and early spring because the winter months tend to always have very poor air quality, which means the impact of TC on pollution is more difficult to identify. Also, there are no TC from December to February.

Based on the type of land use surrounding the station, the monitoring stations were classified into urban, new town, rural and roadside. The urban stations are located on low urban rooftops at altitudes around 10-20 metres (Fig. 3a). The roadside stations are located at ground-level in urban settings, which are adjacent to commercial, land transportation use, and human traffic on the sidewalk (Fig. 3b). The new town stations are placed in residential areas, with altitudes around 20-30 metres (Fig. 3c). The rural station is a special “control” station: the EPD established one station in what they term a “background setting” – this is a reference to a

location that presumably feels very little impact from pollution sourced from Hong Kong itself. The rural station is located in a highly vegetated natural area at an altitude of around 10 metres (Fig. 3d). The hourly pollutant concentrations were reduced to monthly averages to indicate the generally typical annual pattern of concentration for the six air pollutants in Hong Kong. After this the monthly data from 14 stations were grouped and averaged by station type, so that each type of station would produce an air pollution pattern that could show the local air quality variations as a function of local setting. To emphasize the TC impact on air quality, the local human influences should be limited. Therefore, the station type with the lowest measurements on

9

most kinds of pollutants, which can be considered as the best air quality in the region, would be selected to represent a baseline of air quality in Hong Kong.

(a) Urban (b) Roadside

(c) New Town (d) Rural

Figure 3. The surrounding land use of the four types of air quality monitoring stations in Hong Kong (photo provided by EPD).

Pollutant data were organized into categories to simplify the analysis. The category boundaries were determined using an assessment of pollutant concentration distribution

performed using histograms. The distribution indicated by the histograms suggested categories representing the lowest 10% and highest 10% would work as the basis for analyzing TC impacts on air quality (details of this assessment are provided in the Results section). The concentration data were then classified into 5 groups based on the percentile range for further analysis (Table 2).

Table 2. The percentile classification of pollutant concentration data in Hong Kong.

10 Classification Extremely

Low

Low Medium High Extremely

High 2.2 The TC data

TC data from May to September were obtained from the Joint Typhoon Warning Centre, a division of the US National Oceanic and Atmospheric Administration (NOAA) in Hawaii. The information includes TC location and track information recorded every six hours since 1945. The comparison between TC and local air quality was performed using daily data. For this reason, the 0600 UTC position and strength data for individual TC events were selected to represent the daily location of the event because 0600 UTC is roughly midday (1400) in Hong Kong. According to the local TC warning system, an 800-kilometre distance of TC respect to Hong Kong would trigger the first level of warning (Hong Kong Observatory, 2018). This distance as the outermost point of warning initiation were adopted in the study to create an 800-kilometre spatial buffer around the region to filter only those TC within the buffer (Fig. 4) that are likely to cause impact.

In order to better isolate the signal of TC effects on air quality, the local emission signal from human activities should be minimized. Therefore, the air pollution data from the rural station Tap Mun was chosen to represent the general air condition in Hong Kong. Using the dates that were extracted from the TC events in the warning area, the daily average data from the rural station were then divided into days with TC events (TCD) or days without TC occurrences (NTCD). Out of 1841 days from May to September over 2007-2018, 236 days were classified as TCD. The average concentration of six pollutants on both TCD and NTCD were calculated to provide a general sense of the impact of TC on air pollution. A two-sample t-test for different sample sizes was applied to examine the significance of the differences between TCD and NTCD data.

11

More detailed analyses were performed on specific subsets of data. The dates of TCD in the upper and lower 10th percentile group classifications were also highlighted in order to identify the proportion of TCD in either the days with extremely low pollutant concentration or extremely high (Table 2). These number of TCD in each classification were displayed in tables. The spatial location of TC events in each group were mapped to present the spatial pattern and to identify the occurrence of clustering.

Figure 4. The daily location of TC events within the 800-kilometre buffer around Hong Kong from 2007 to 2018. Taiwan is the island to the east of the region and Hainan Island is located to the west of the region.

2.3 Air pollution emission trend

The annual emission data timeseries was assessed using a slightly longer timeframe, from 2001 to 2017, which were provided by the EPD. The data include SO₂, PM₂.₅, PM₁₀, CO, as well as nitrogen oxide (NOx) and VOCs which are the source of O₃ and NO₂ under favourable conditions (Haagen-Smit & Fox, 1954; Han et al, 2011). The level of each pollutant is the sum of pollution coming from multiple public sectors which include electricity generation, industrial

12

emission, civil aviation, road and marine transportation. Thus, considering different

contributions from the public sector could explain some of the pollutant concentration patterns. As a result, a breakdown of the specific sources for the most recent year of data (2017) was performed to explore the current emission conditions.

13 3 Results

3.1 Mean air pollution pattern

The 12-year monthly average pollutant concentrations for PM₂.₅ and PM₁₀ showed much lower concentrations in the middle of the year, approximately half of the pollution level of the highest months (Fig. 5a & b). The general trend of NO₂, O₃ and CO also indicated a lower concentration from May to September (Fig. 5c, e & f). Only SO₂ showed a relatively consistent concentration level throughout the year (Fig. 5d). As a result, May to September was selected as the study period so that the influence of pollution contributors other than TC was minimized.

0 10 20 30 40 50 1 2 3 4 5 6 7 8 9 10 11 12 Co n ce n tr ati o n (a) PM₂.₅ 0 10 20 30 40 50 60 70 1 2 3 4 5 6 7 8 9 10 11 12 (b) PM₁₀ 0 10 20 30 40 50 60 70 1 2 3 4 5 6 7 8 9 10 11 12 C o n ce n tr atio n (c) NO₂ 0 2 4 6 8 10 12 14 16 1 2 3 4 5 6 7 8 9 10 11 12 (d) SO₂ 0 10 20 30 40 50 60 70 80 1 2 3 4 5 6 7 8 9 10 11 12 C o n ce n tr atio n Month (e) O₃ 0 200 400 600 800 1000 1200 1 2 3 4 5 6 7 8 9 10 11 12 Month (f) CO

14

Figure 5. The mean concentration (μg/m³) of six major air pollutants in Hong Kong, 2007-2018. These plots integrate data from all 14 stations used in this study.

The lower summer concentrations of all six pollutants except SO₂ was apparent for all station types (Fig. 6). The roadside stations constantly measured the highest values of PM₂.₅ and PM₁₀ concentration, while the new town, urban and rural stations showed similar pollution levels (Fig. 6a & b). The NO₂ chart showed the largest variation (100 μg/m³) between the roadside stations (highest concentration) and the rural station (lowest concentration) (Fig. 6c). Stations with the higher NO₂ concentrations overall also showed the greatest range between summer and winter, while the rural station showed little variation through the year (Fig. 6c). The urban stations experienced the highest SO₂ from February to October with higher pollution levels in the summer months, while the rural station shifted from the lowest SO₂ concentration to the highest from October to January (Fig. 6d). The O₃ chart also showed the less polluted

summer trend in each station type (Fig 6e). The pollution pattern of O₃ was opposite to that of NO₂, with the highest concentrations measured at the rural stations, and the lowest recorded at the roadside station (Fig 6e). Similar to PM₂.₅ and PM₁₀, the highest CO concentrated was found at the roadside stations, and the data from other stations were mostly at the same level; but the declining trend at the middle of the year was not as significant as the measurements of PM₂.₅ and PM₁₀ (Fig. 6f). Overall, 5 out of 6 pollutants (expect O₃) observed the lowest concentration in the summer months at the rural station (Table 3). As a result, data from the rural station was selected to run the further analysis with TC.

15

Figure 6. The mean monthly pollutant concentration (μg/m³) averaged by 4 station types from 2007-2018 in Hong Kong.

Table 3. The average of annual pollutant concentration (μg/m³) collected from different types of stations in Hong Kong (↑=highest, ↓=lowest).

Station Type PM₂.₅ PM₁₀ NO₂ SO₂ O₃ CO

Urban 29 43 57 13↑ 38 710 New Town 29 43 45 11 44 689 0 10 20 30 40 50 60 70 80 1 2 3 4 5 6 7 8 9 10 11 12 (b) PM₁₀ 0 20 40 60 80 100 120 140 1 2 3 4 5 6 7 8 9 10 11 12 Co n ce n tratio n (c) NO₂ 0 2 4 6 8 10 12 14 16 18 1 2 3 4 5 6 7 8 9 10 11 12 (d) SO₂ 0 20 40 60 80 100 120 1 2 3 4 5 6 7 8 9 10 11 12 Co n ce n tratio n Month (e) O₃ 0 200 400 600 800 1000 1200 1400 1 2 3 4 5 6 7 8 9 10 11 12 Month (f) CO 0 10 20 30 40 50 1 2 3 4 5 6 7 8 9 10 11 12 Co n ce n tratio n (a) PM₂.₅ Urban New Town Roadside Rural

16

Roadside 34↑ 54↑ 103↑ 12 19↓ 916↑

Rural 25↓ 42↓ 12↓ 11↓ 71↑ 665↓

For better comparison between air quality and TC, the distribution of pollution data at the rural station was investigated. All six histograms showed a long-tailed distribution (Fig. 7). This result identified the occurrence of extremely high values in the dataset, with some of the

extremes of PM₂.₅, PM₁₀ and SO₂ exceeding the WHO pollutant daily average standard (Appendix A). 0 200 400 600 800 0 10 20 30 40 50 60 70 80 90 100 (a) PM₂.₅ 0 200 400 600 800 0 10 20 30 40 50 60 70 80 90 100 (b) PM₁₀ Frequency 0 200 400 600 800 0 5 10 15 20 25 30 35 40 45 50 (c) NO₂ 0 200 400 600 800 1000 1200 0 8 16 24 32 40 48 56 64 72 (d) SO₂

17

Figure 7. The histograms of daily pollutant concentration (μg/m³) from May to September in rural station in Hong Kong, 2007-2018. Histogram categories reflect the maximum values in each pollutant dataset.

3.2 TCD vs. NTCD

TC passage can have a noticeable short-term effect on the levels of various pollutants in the city (Appendix B). As a result, a two-sample t-test was applied to evaluate whether

differences in pollutant concentrations between TCD and NTCD data were significant (Table 4). The result showed significant differences of PM₂.₅, PM₁₀, SO₂ and O₃ data between TCD and NTCD; in all cases, the pollution levels were greater during TCD. This indicates that, in general, TC activity acts to worsen air quality for these pollutant types. However, the concentration of NO₂ and CO showed no significant differences between TCD and NTCD.

Table 4. The result of two-sample assuming unequal variances t-test (α=0.05) on pollutant concentration (μg/m³) during TCD (sample size=236) and NTCD (sample size=1604).

Pollutant TC Mean P-value (two-tail)

PM₂.₅ TCD 21 7.30E-06 (<0.05) NTCD 16 PM₁₀ TCD 36 4.26E-11 (<0.05) NTCD 27 NO₂ TCD 11 6.01E-02 NTCD 10 0 100 200 300 400 500 600 700 0 20 40 60 80 100 120 140 160 180 Concentration (e) O₃ 0 100 200 300 400 500 600 700 Concentration (f) CO

18 SO₂ TCD 10 1.34E-03 (<0.05) NTCD 8 O₃ TCD 68 2.09E-09 (<0.05) NTCD 56 CO TCD 603 2.55E-01 NTCD 586

The proportion of total TCD in the study period was 13%, while the percentile classification results indicated the percentage of all pollutants in the “extremely high” concentration category were higher than 13% (Table 5). The results also showed the highest proportion of TCD were associated with the “extremely high” classification for PM₂.₅, PM₁₀, SO₂ and CO. In the case of PM₂.₅, and PM₁₀ this accounted for more than 25% of the TCD days, and for SO₂, 23%. It was clear that O₃ was also elevated during TCD days but was mostly classified into the “medium” and “high” groups. Although CO was a little elevated in the

“extremely high” category, it was not very different from the proportion in other categories, suggesting TCs have little effect on CO concentration. It was also worth noting that more than 20% of the days in both extremely high and low NO₂ concentration group experienced TC event during the day, which indicated that TCs can bring a cleaning, as well as a polluting effect. Overall, the TCD tends to have poor air quality due to a higher pollutant concentration.

Table 5. The percentile classification of pollutant concentration with the amount of TCD out of total days and its percentage.

Pollutant Percentile Classification

Extremely Low

Low Medium High Extremely

High PM₂.₅ ¹ TCD (#) 26 / 182 36 / 274 74 / 914 46 / 274 46 / 182

19 PM₁₀ ¹ TCD (#) 6 / 182 24 / 275 100 / 911 54 / 274 48 / 182 TCD (%) 3.3 8.7 11.0 19.7 26.4 NO₂ ² TCD (#) 48 / 185 37 / 275 77 / 912 33 / 273 40 / 182 TCD (%) 25.9 13.5 8.4 12.1 22.0 SO₂ ² TCD (#) 20 / 184 42 / 274 98 / 921 34 / 269 42 / 183 TCD (%) 10.9 15.3 10.6 12.6 23.0 O₃ ² TCD (#) 7 / 183 10 / 275 141 / 916 52 / 274 26 / 183 TCD (%) 3.8 3.6 15.4 19.0 14.2 CO ² TCD (#) 28 / 184 34 / 276 105 / 275 38 / 275 31 / 182 TCD (%) 15.2 12.3 11.5 13.8 17.0

1: 98% of data were collected during TCD, 99% were recorded during study period. 2: 100% of data were collected during TCD, 99% were recorded during study period.

3.3 Spatial distribution of TC

The daily TC locations were plotted based on the percentile classification (Fig. 8). The broad pattern indicated that TCs associated with the “extremely low” classification were located well to the west side of Hong Kong, with some showing clustering around the west island of Hainan (Fig. 8 a1-f1), but the total amount of TC events were low (Table 5). Moving to the “low” classification, more TC points were found on the east side of Hainan but still to the west of Hong Kong (Fig. 8 a2-f2). The “medium” classification showed greater longitudinal spread without specific clusters of TCs (Fig. 8 a3-f3). In the “high” concentration group, the TC location were mostly located in the South China Sea, between Hong Kong and the east island of Taiwan (Fig. 8 a4-f4). Compared to the maps of “high” concentration, the position of TC with “extremely high” pollutant concentration showed clustering around the northeastern region of Taiwan, with more TCs located over the land of mainland China instead of in the ocean (Fig. 8 a5-f5). There were also other clusters displayed in the extremely high concentration group, one was located south of Taiwan; the other, south of Hainan (Fig. 8 a5, b5 & f5). In summary, there was a shift of clusters from west to east and northeast of Hong Kong, and these clusters were

20

clearly related to pollution concentration categories. However, some clusters remain in specific locations even under different classifications, which implied the other factors that might

contribute to the air quality in Hong Kong.

(a1) (a2) (a3) (a4) (a5)

PM₂.₅ (b1) (b2) (b3) (b4) (b5) PM₁₀ (c1) (c2) (c3) (c4) (c5) NO₂ (d1) (d2) (d3) (d4) (d5) SO₂

(e1) (e2) (e3) (e4) (e5)

CO

21 O₃

Figure 8. The spatial distribution of TCs (yellow dot) around 800km distance (include Hainan on the west and Taiwan on the east) from Hong Kong (black dot). Columns 1-5 are arranged by percentile classifications (1= extremely low, 2=low, 3=medium, 4=high, 5=extremely high). Rows A-F are arranged by pollutant type (a= PM₂.₅ , b= PM₁₀, c= NO₂, d= SO₂, e=CO, f= O₃). 3.4 Pollutant emission trends

Over the years there has been a general decline in pollution concentrations for all

pollutant types (Fig. 9). The most significant emission reduction has been observed in SO₂ which was mostly contributed by shipping and public electricity generation in 2017(Fig. 10a). It was worth noticing that shipping has also been the major contributor of PM₁₀, PM₂.₅, and NOx which was the source of NO₂ and O₃ (Fig 10b-d). In another words, 5 out of 6 hourly measured pollutants were highly affected by shipping emissions in Hong Kong. The VOC emission, major source of O₃, was dominated by non-combustion sources which included paints and associated solvents, consumer products and printing (Fig. 10e). As for the CO emission, more than 50% of the pollution was emitted by road transportation (Fig. 10f).

Figure 9. The Hong Kong air pollutant emission (tonnes) inventory from 2001 to 2017 (without the emission of hill fire).

0 20,000 40,000 60,000 80,000 100,000 120,000 140,000 160,000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 Emi ssi on Year CO NOx SO₂ PM₁₀ PM₂.₅ VOC

22

(a) SO₂ (b) NOx (c) PM₁₀

(d) PM₂.₅ (e) VOC (f) CO

Figure 10. The proportion of pollutant emission from different sources in 2017 (1: major contributing sources include non-road mobile machineries operating in construction sites and container terminals. 2: the major sources are paved road dust, cooking fume, construction dust and quarry production for PM₂.₅ and PM₁₀ emission; paints and associated solvents, consumer products and printing for VOC emission (Environmental Protection Department, n.d.)).

Public Electricity Generation Road Transport Navigation

23 4 Discussion

The best air quality in Hong Kong was found in the summer months, especially in June and July when all the pollutants were at the lowest concentration (Fig. 5). The least polluted rural station, however, still measured concentrations of particulate matters and SO₂ that were higher than the WHO daily standard during the less polluted months (Appendix A). The possible driving factors of these abnormal highs include variations in meteorology and pollutant emissions. These are discussed below.

4.1 Tropical cyclone

The most significant meteorological event that could change the local wind field is TC. During the study period, 13% of the days experienced TC. Therefore, the ratio of TCD in each percentile classification group was assumed to be close to 13% and higher in the lower

concentration groups as the TC was hypothesised to have a cleaning effect on the local air qualiy. However, the result showed a higher percentage of TCD in the extremely high

concentration group, indicating the polluting effects of TC. With all of the “extremely high” and “high” classifications that accounted for most of the TCD, the NO₂ data acted the opposite as its maximum percentage was in the extremely low classification (Table 5). This could due to the reversible chemical reaction that converts NO₂ to O₃ or vice versa, keeping a balance between the amount of NO₂ and O₃ in the troposphere. During the formation of O₃, solar radiation, which is plentiful during summer, especially in the rural area without the building blocking, is required (Haagen-Smit & Fox, 1956). This could explain how the NO₂ was consumed, and the reason for the rural station to have the highest O₃ concentration through the year (Fig. 6e). Moreover, this photochemical reaction can decompose O₃ and produce NO₂ which theoretically keeps an equilibrium between the 2 pollutants; but the emission of VOCs, which are able to travel a long distance to the downwind area, can act as a source of additional VOC supply into a region which disrupts the balance, changing the relationship between O₃ and NO₂ from negative liner to

24

nonlinear where the formation of O₃ would first enhanced and then reduced after the peak as the concentration of NO₂ kept increasing (Jhun et al., 2015; Wei et al., 2018; Sillman, 1999). Because the TC event changes the local wind direction, the amount of the pollutants that were transported and the extent of their chemical reaction varied, creating favourable conditions for either consuming or producing NO₂ which could explain the higher TCD percentage in both extreme high and low classifications (Table 5). Moreover, other factors include cloud cover and relative humidity which can be affected by TC could also impact the efficiency of solar radiation in the chemical reaction so that the concentration of NO₂ varied (Kleinman, 1991). This shows that TC events are able to influence concentrations of most of the pollutant types in Hong Kong.

As for the spatial distribution of TCs, all the observed events followed a similar spatial pattern –the shift from the west of Hong Kong with extremely low concentration to the east with the highs (Fig. 8). With the knowledge of TC and its impact on pressure changes, one can predict the large-scale wind direction based on the TC locations. For instance, most of the TC that were labelled in the “extremely high” and “high” classification, were located around the Taiwan, which would generate northwesterly wind in Hong Kong. TC located around Hainan would enhance the southwesterly wind that already existed as the summer monsoon (Kim et al, 2013). The major difference between the northwesterly and southwesterly wind is source region over which the wind is blowing: air advected from over the ocean is assumed to be clean, while air advected from over the continental area of mainland China contains pollutants that are emitted from human activities. Furthermore, due to the friction with the land surface, the speed of

north/northwest winds would be reduced (Chow et al, 2018). The ocean wind, on the other hand, would have higher speed, due to its combination with the monsoon, and less land friction on its track. This would give the wind greater ability to ventilate the pollution from Hong Kong. Also, the weather conditions including wind speed and air stability are related to the distance from TC, because there is a zone of descending air at the TC’s peripheral circulations in the earliest stages

25

of the approach, this creating warmer and dryer atmosphere condition near the surface; results in a more stable atmosphere with the inversion layer in Hong Kong that could reduce the

dispersions of pollutants and enhance the accumulation (Munksgaard et al, 2015; Feng et al., 2007). This could explains the location of TCs in the 3 less-extreme classifications showed they were closer to Hong Kong. However, without knowing the intensity and the traveling speed of TC, the exact distance to Hong Kong was unknown. This could contribute to the locational spread of TC in the “medium” classification (Fig.8 a3-f3). More variables would need to be included in an analysis to better isolate the locations of the TCs in the “medium” category.

One potential limitation to applicability was the statistically based TC classification because the air pollution product is more applicable when it is more directly related to human health. For this study, the high classification might not necessarily represent a high health risk; rather, only a high measurement within the dataset. For instance, the health standard for the daily averaged concentration of CO is 7000 μg/m³ (Appendix A) which is much higher than the data that were utilized in this study (Fig. 7f). As a result, the health-related data is necessary to be included in the pollution study, such as the WHO air quality guidelines and the AQHI, so that a standard could be applied as the absolute threshold to determine the health-related pollution effects by TC. The AQHI was launched and reported since 2013 in Hong Kong, but its algorithm was only released recently by the EPD. Thus, the AQHI data were not available for this study as the value before 2013 was unknown. Future studies could add a direct analysis of AQHI to obtain a better understanding of the direct impacts of TC on human health risk associated with air pollution.

4.2 Pollutant emission

Local emission is another contributing component to air quality. Accordingly, the extent of study period should be thoughtfully selected due to the significant decline of the emission in Hong Kong (Fig. 9). Chow et al. (2018) showed a 57% of TCD in the “extremely high”

26

concentration of PM₁₀ in Hong Kong which was double the result from this study. One of the reasons was the study period of previous research was from 2000 to 2015; at this time emissions were high because there was no regulation control. When combined the high historical

concentration data with the lower current data, the high percentile classification would be dominant by the historical data including all the TC that occurred during the early twenty first century. In this study, the pollution data since 2007 were obtained based on the implementation year of the regulations. However, there was still a significant drop in the emission of SO₂, and some decreases in NOx and CO data which could lead to a biased result. Similarly, the TC season was defined as July to October by Chow et al. (2018) when the air quality in October was worse on average, the TC impacts could be less significant as the contribution of local emission was higher. Overall, more work is required for data testing to make sure the TCD classification is not affected by the emission and its long-term trend.

As for the recent emission, the shipping sector becomes the highest contributor of multiple pollutants (Fig. 10). With the rank of fifth busiest container port in the world in 2017, the shipping industry in Hong Kong was heavily polluted with the major emitter being ocean going vessels (Hong Kong Maritime and Port Board, 2020; Environmental Protection

Department, n.d.). One can assume that major sources of the shipping pollutants are located at the coast. When the wind blows from the ocean to Hong Kong, it can transport the pollutants emitted by the shipping industry if the upwind area was aliened with the ports, result in the downwind area being polluted by the emission. Under this situation, the southerly TC wind which was defined as clean marine air from this study could be incorrect due to the marine emission. As a result, more variables on local sectors should be included to reduce the variation of TC impact caused by the local emission.

27 5 Conclusion

Tropical Cyclones of the western Pacific can affect the local air quality in Hong Kong depending on the direction from which they approach the city. Although the mechanism of pollutant formation varied, the general spatial pattern of TC with certain concentration level of PM₂.₅, PM₁₀, SO₂, NO₂, O₃ and CO were found to be similar. The relationship between TC approach/positioning in the east or west and local air quality is an important consideration for the summer air quality forecast, because it is likely that citizens would be exposed to unhealthy air when a TC approaching from the east moves into the nearby water body. However, the

prediction could likely be improved by considering more meteorological variables and local variations. Some of the variables that were mentioned in the discussion section could be helpful to improve the prediction. This study explores the impact of TC locations on the six major air pollutants in Hong Kong, offers a potential prediction method on summer air quality in the region.

28 6 Reference

Ackerman, E. A. (1941). The köppen classification of climates in North America. Geographical

Review, 31(1), 105-111. doi:10.2307/210420

Anderson, H., Spix, C., Medina, S., Schouten, J., Castellsague, J., Rossi, G., . . . Katsouyanni, K. (1997). Air pollution and daily admissions for chronic obstructive pulmonary disease in 6 european cities: Results from the APHEA project. European Respiratory Journal, 10(5), 1064-1071. 10.1183/09031936.97.10051064

Brimblecombe, P. (2011). Air Pollution. Environmental Encyclopedia (pp. 38-41). 4th ed., Vol. 1. Detroit: Gale.

Brook, R., Franklin, B., Cascio, W., Hong, Y., Howard, G., Lipsett, M., . . . Expert Panel on Population and Prevention Science of the American Heart Association. (2004). Air pollution and cardiovascular disease - A statement for healthcare professionals from the expert panel on population and prevention science of the american heart association. Circulation, 109(21), 2655-2671. 10.1161/01.CIR.0000128587.30041.C8

Census and Statistics Department. (2012). 2011 Population Census: Summary Results. Hong Kong: Census and Statistics Department.

Chow, E. C. H., Li, R. C. Y., & Zhou, W. (2018). Influence of tropical cyclones on hong kong air quality. Advances in Atmospheric Sciences, 35(9), 1177-1188. doi:10.1007/s00376-018-7225-4

Deng, X., Tie, X., Zhou, X., Wu, D., Zhong, L., Tan, H., . . . Deng, T. (2008). Effects of southeast asia biomass burning on aerosols and ozone concentrations over the pearl river delta (PRD) region. Atmospheric Environment, 42(36), 8493-8501.

doi:10.1016/j.atmosenv.2008.08.013

Development Bureau. (2017). Reclamation Outside Victoria Harbour (Report No. 07/2017). Retrieved from https://www.devb.gov.hk/filemanager/en/content_1054/Paper_07_2017.pdf Ding, A., Wang, T., Xue, L., Gao, J., Stohl, A., Lei, H., . . . Zhang, X. (2009). Transport of north

china air pollution by midlatitude cyclones: Case study of aircraft measurements in summer 2007. Journal of Geophysical Research - Atmospheres, 114(D8), D08304.

doi:10.1029/2008JD011023

Environmental Protection Department. (2007). Air pollution control (volatile organic compounds) regulation (Report No. EP351/V5/1 Pt.4). Retrieved from

https://www.epd.gov.hk/epd/sites/default/files/epd/english/environmentinhk/air/prob_solutio ns/files/Notification_letter.pdf

Environmental Protection Department. (n.d.). Hong Kong Air Pollutant Emission Inventory. Retrieved from

https://www.epd.gov.hk/epd/english/environmentinhk/air/data/emission_inve.html

Feng, Y., Wang, A., Wu, D., & Xu, X. (2007). The influence of tropical cyclone melor on PM10 concentrations during an aerosol episode over the pearl river delta region of china:

Numerical modeling versus observational analysis. Atmospheric Environment, 41(21), 4349-4365. doi:10.1016/j.atmosenv.2007.01.055

29

Firket, J. (1936). Fog along the meuse valley. Transactions of the Faraday Society, 32, 1192. doi:10.1039/tf9363201192

Geddes, J. A., Murphy, J. G., & Wang, D. K. (2009). Long term changes in nitrogen oxides and volatile organic compounds in Toronto and the challenges facing local ozone control.

Atmospheric Environment, 43(21), 3407-3415. doi:10.1016/j.atmosenv.2009.03.053

Haagen-Smit, A. J., & Fox, M. M. (1956). Ozone formation in photochemical oxidation of organic substances. Industrial & Engineering Chemistry, 48(9), 1484-1487.

doi:10.1021/ie51400a033

Han, S., Bian, H., Feng, Y., Liu, A., Li, X., Zeng, F., & Zhang, X. (2011). Analysis of the relationship between O3, NO and NO2 in tianjin, china. Aerosol and Air Quality Research, 11(2), 128-139. doi:10.4209/aaqr.2010.07.0055

Hong Kong Maritime and Port Board. (2020). Summary Statistics on Port Traffic of Hong Kong. Retrieved from https://www.hkmpb.gov.hk/document/summary_statistics.pdf

Hong Kong Observatory. (2018). Hong Kong’s tropical cyclone warning signals. Retrieved from https://www.hko.gov.hk/publica/gen_pub/tcws.pdf

Hong Kong Observatory. (n.d.). Climate of Hong Kong. Retrieved from https://www.hko.gov.hk/en/cis/climahk.htm

Huang, J., Jimmy C. H. Fung, Alexis K. H. Lau, & Qin, Y. (2005). Numerical simulation and process analysis of typhoon-related ozone episodes in hong kong. Journal of Geophysical

Research - Atmospheres, 110(D5), D05301-n/a. doi:10.1029/2004JD004914

Hussain, I., Jain, V. V., O’Shaughnessy, P., Businga, T. R., & Kline, J. (2004). Effect of nitrogen dioxide exposure on allergic asthma in a murine model. Chest, 126(1), 198-204.

10.1378/chest.126.1.198

Jhun, I., Coull, B. A., Zanobetti, A., & Koutrakis, P. (2015). The impact of nitrogen oxides concentration decreases on ozone trends in the USA. Air Quality, Atmosphere & Health, 8(3), 283-292. doi:10.1007/s11869-014-0279-2

Kim, J., Zhou, W., Cheung, H. N., & Chow, C. H. (2013). Variability and risk analysis of hong kong air quality based on monsoon and el niño conditions. Advances in Atmospheric

Sciences, 30(2), 280-290. doi:10.1007/s00376-012-2074-z

Kleinman, L. I. (1991). Seasonal dependence of boundary layer peroxide concentration: The low and high NOx regimes. Journal of Geophysical Research: Atmospheres, 96(D11), 20721-20733. doi:10.1029/91JD02040

Lam, K. S., Wang, T. J., Wu, C. L., & Li, Y. S. (2005). Study on an ozone episode in hot season in hong kong and transboundary air pollution over pearl river delta region of china.

Atmospheric Environment, 39(11), 1967-1977. doi:10.1016/j.atmosenv.2004.11.023

Laties, V. G., & Merigan, W. H. (1979). Behavioral effects of carbon monoxide on animals and man. Annual Review of Pharmacology and Toxicology, 19(1), 357-392.

10.1146/annurev.pa.19.040179.002041

Longo, L. D. (1977). The biological effects of carbon monoxide on the pregnant woman, fetus, and newborn infant. American journal of obstetrics and gynecology, 129(1), 69–103.

30

Martins, E., Nunes, A., & Correa, S. (2015). Understanding ozone concentrations during weekdays and weekends in the urban area of the city of rio de janeiro. Journal of the

Brazilian Chemical Society, 26(10), 1967-1975. doi:10.5935/0103-5053.20150175

Munksgaard, N., Zwart, C., Kurita, N., Bass, A., Nott, J., & Bird, M. (2015). Stable isotope anatomy of tropical cyclone ita, north-eastern australia, april 2014. Plos One, 10(3), e0119728. doi:10.1371/journal.pone.0119728

Park, S. K., O'Neill, M. S., Vokonas, P. S., Sparrow, D., & Schwartz, J. (2005). Effects of air pollution on heart rate variability: The VA normative aging study. Environmental Health

Perspectives, 113(3), 304-309. 10.1289/ehp.7447

Reith, J. W. (1951). Los angeles smog. Yearbook of the Association of Pacific Coast

Geographers, 13(1), 24-32. doi:10.1353/pcg.1951.0001

Rich, D. Q., Mittleman, M. A., Link, M. S., Schwartz, J., Luttmann-Gibson, H., Catalano, P. J., . . . Dockery, D. W. (2006). Increased risk of paroxysmal atrial fibrillation episodes associated with acute increases in ambient air pollution. Environmental Health Perspectives, 114(1), 120-123. 10.1289/ehp.8371

Schrenk, H. H. (1949). Air pollution in donora, Pa.: epidemiology of the unusual smog episode of October 1948 (pp. 149-162). Washington D.C.: Federal Security Agency, Public Health Service, Bureau of State Services, Division of Industrial Hygiene.

Sillman, S. (1999). The relation between ozone, NOx and hydrocarbons in urban and polluted rural environments. Atmospheric Environment, 33(12), 1821-1845. doi:10.1016/S1352-2310(98)00345-8

Stern, A. C., Wohlers, H. C., Boubel, R. W., & Lowry, W. P. (1973). Air Pollution.

Fundamentals of Air Pollution (pp. 19-33). New York and London: Academic Press.

Stieb, D. M., Burnett, R. T., Smith-Doiron, M., Brion, O., Shin, H. H., & Economou, V. (2008). A new multipollutant, no-threshold air quality health index based on short-term associations observed in daily time-series analyses. Journal of the Air & Waste Management Association, 58(3), 435. doi:10.3155/1047-3289.58.3.435

Tang, R., Tian, L., Thach, T., Tsui, T. H., Brauer, M., Lee, M., . . . Barratt, B. (2018). Integrating travel behavior with land use regression to estimate dynamic air pollution exposure in hong kong. Environment International, 113, 100-108. doi:10.1016/j.envint.2018.01.009

Tong, C. O., & Wong, S. C. (1997). The advantages of a high density, mixed land use, linear urban development. Transportation, 24(3), 295-307. doi:10.1023/A:1004987422746

Wang, X., Ho, K., Chow, J. C., Kohl, S. D., Chan, C. S., Cui, L., . . . Watson, J. G. (2018). Hong kong vehicle emission changes from 2003 to 2015 in the shing mun tunnel. Aerosol Science

and Technology: Articles in Honor of Peter McMurry, 52(10), 1085-1098.

doi:10.1080/02786826.2018.1456650

Wang, X. Q., Wand, Z. F., Yu, T., Gong, Y. B. (2009). Foreshadowing of the western pacific tropical cyclone track to PM10 air pollution episode in the Beijing area. Chinese Science

Bulletin, 54(5), 830-835. 10.1007/s11434-008-0562-4

Wei, W., Li, Y., Wang, Y., Cheng, S., & Wang, L. (2018). Characteristics of VOCs during haze and non-haze days in beijing, china: Concentration, chemical degradation and regional

31

transport impact. Atmospheric Environment, 194, 134-145. doi:10.1016/j.atmosenv.2018.09.037

Wei, X. L., Li, Y. S., Lam, K. S., Wang, A. Y., & Wang, T. J. (2007). Impact of biogenic VOC emissions on a tropical cyclone-related ozone episode in the pearl river delta region, china.

Atmospheric Environment, 41(36), 7851-7864. doi:10.1016/j.atmosenv.2007.06.012

Wong, T.W., Tam W., Lau A., Ng S., Yu I.TS, Wong, A., Yeung, D. (2012). A Study of the Air Pollution Index Reporting System.(Report No. AP 07-085). Retrieved from

https://www.aqhi.gov.hk/pdf/related_websites/APIreview_report.pdf

Yang, J. X., Lau, A. K. H., Fung, J. C. H., Zhou, W., & Wenig, M. (2012). An air pollution episode and its formation mechanism during the tropical cyclone nuri's landfall in a coastal city of south china. Atmospheric Environment, 54, 746-753.

32 Appendix A Classical Pollutants Averaging Time Time-Weighted

Average (μg/m³) Health Endpoints used for AQGs development

PM₁₀

Annual

20

Total, cardiopulmonary and lung cancer mortality 24h 50 PM₂.₅ Annual 10 24h 25 O₃ 8h daily

max 100 Daily mortality

NO₂

Annual

40 Respiratory effects in children 1h daily

max 200 Bronchial responsiveness in asthmatics

SO₂ 24h 20

All age mortality and childhood respiratory disease

10min 500 Respiratory symptoms in asthmatics

CO

24h

7000

COHb levels in non-smokers blood below 2%

8h 10000

1h 35000

Appendix A. Summary Table of WHO Air Quality Guidelines. Reprinted from WHO Expert Consultation: Available evidence for the future update of the WHO Global Air Quality Guidelines (AQGs), by World Health Organization, 2015, pp44., retrieved from

http://www.euro.who.int/__data/assets/pdf_file/0013/301720/Evidence-future-update-AQGs-mtg-report-Bonn-sept-oct-15.pdf Copyright 2016 by World Health Organization

33 Appendix B

Appendix B. The hourly air pollutant concentration (μg/m³) during the TC Molave (0906) from

July 16th to 21st, 2009. Molave moved towards Hong Kong from Southeast, reached the 800-km buffer at late night of July 17th, and passed through Hong Kong at the dawn of July 19th.