Are the noise levels acceptable in a built environment like Hong Kong?

Governments all over the world have enacted environmental noise directives and noise control ordinances/acts to protect tranquility in residential areas. However, there is a lack of literature on the evaluation of whether the Acceptable Noise Levels (ANLs) stipulated in the directive/ordinance/act are actually achievable. The study aimed at measuring outdoor environmental noise levels in Hong Kong and identifying whether the measured noise levels are lower than the stipulated ANLs at 20 categories of residential areas. Data were gathered from a territory-wide noise survey. Outdoor noise measurements were conducted at 203 residential premises in urban areas, low-density residential areas, rural areas, and other areas. In total, 366 daytime hourly Leq outdoor noise levels, 362 nighttime hourly Leq outdoor noise levels, and 20 sets of daily, that is, 24 L(eq,1-)h outdoor noise levels were recorded. The mean daytime L(eq,1-h) values ranged 54.4-70.8 dBA, while the mean nighttime L(eq,1-h) values ranged 52.6-67.9 dBA. When the measured noise levels were compared with the stipulated ANLs, only three out of the 20 categories of areas had outdoor noise levels below ANLs during daytime. All other areas (and all areas during nighttime) were found to have outdoor noise levels at or above ANLs.

Introduction

According to a study by the United Nations Population Division, the human population will increase from 7.2 billion in 2013 to 9.6 billion in 2050, while the world's urban population will increase from 3.6 billion to 6.3 billion over the same period.[1] As the urban areas of the world will absorb nearly all the population growth expected in the next four decades and will continue to attract some of the rural population,[1] cities will get more crowded and noisier. In Asia, populations in cities have increased rapidly in the past four decades. For example, Hong Kong's population increased from 4.00 million in 1970 to 7.17 million in 2012.[2] With the rapid growth of populations, the consumption of material and energy resources and the pressure on the environment increase. Greenhouse gas emissions, other air pollutants, water pollution, noise pollution, and municipal solid waste cause grave concerns for citizens.[345678910111213141516]

Among all pollution problems, noise is considered as an important environmental issue because it is the source of many environmental complaints.[171819] The European Union (EU) recognizes noise as an environmental stressor that adversely affects public health.[202122] Research shows that environmental noise causes numerous psychological effects such as annoyance[23242526] and negative emotions including anger, disappointment, anxiety, and even depression.[25272829] Noise is also linked to cardiovascular disease.[303132] Noise was found to have a significant negative impact on children's blood pressure[3334] and mental health.[35] Specifically, Chang et al.[34] showed that people who are exposed to high road traffic noise at around Leq,8-h 82 dBA have significantly higher prevalence of hypertension than those who are exposed to road traffic noise at around Leq,8-h 77 dBA. Hence, the EU passed Directive 2002/49/EC, also known as the Environmental Noise Directive.[36] The Directive requires all member states to produce strategic noise mapping, estimate population exposure to noise, develop noise action plans, and disseminate noise information to the public.[20]

Within Asia, Hong Kong is one of the noisiest cities in the region.[1037] To et al.[10] reported that the measured urban traffic noise levels ranged 73.4-91.4 dBA in L10,1-h, while the corresponding Leq,1-h values ranged 69.4-88.8 dBA. To et al.,[10] Mak et al.,[13] and Mak and Leung[16] indicated that the predicted noise levels based on the United Kingdom (UK)'s Calculation of Road Traffic Noise methodology (the de facto standard method used in Hong Kong) had a tendency to overestimate noise levels by 2-6 dBA in comparison with the measured data. It was not unexpected, as noise emissions from different types of vehicles have been reduced by at least several decibels in the past four decades,[38] but with the increase in road traffic, urban areas are noisier than in the past. Besides, there are many unusual building designs in Hong Kong because of the lack of usable areas. For example, high-rise buildings and schools are located in the vicinity of major roads[101638] and roundabouts,[9] and are even surrounded by two or more major roads. Some residential buildings are right next to industrial areas. These situations pose a great challenge to citizens and government departments because there are numerous concerns about noise, air pollution, safety, accessibility, and convenience, among others. This paper focuses on:

Quantifying outdoor environmental noise in Hong Kong,

Determining whether the measured noise levels are lower than the stipulated Acceptable Noise Levels (ANLs) using hourly Leq as an indicator at 20 categories of areas, and

Identifying the likely noise impact on residents, based on the World Health Organization (WHO) guidelines.[3940]

Methods

Study area: Hong Kong

Hong Kong is a Special Administrative Region of the People's Republic of China. It has a land area of 1,104 km2 and a population of 7.17 million. Hong Kong's living density ranges from 498 (in outlying islands) to 55,000 (in highly populated districts) people per square kilometer.[41] As one of the world's financial and tourism centers, Hong Kong welcomes 54 million visitors each year. Hence, Hong Kong is under huge environmental pressure. The Hong Kong Government established the Hong Kong Environmental Protection Department in 1986. In 1988, a comprehensive legislation on the control of the environment noise, namely the Noise Control Ordinance (NCO), was enacted.[11] The NCO covers noise from construction sites, industrial and commercial premises, residential premises, public places, and newly registered motor vehicles. In addition, noise problems are also prevented through the Environmental Impact Assessment Ordinance (EIAO). The noise standards at different areas are stipulated in the Technical Memoranda (TM) of the NCO and EIAO. The noise standards indicate that Hong Kong can be divided into urban areas, low-density areas, rural areas, and other areas. Each of the areas may be directly influenced, indirectly influenced, or not influenced by key influencing factors such as major roads or industrial areas. A major road is defined as a road that has a heavy and continuous flow of traffic exceeding 30,000 vehicles per day. Each category of these areas will have a limit of noise level during daytime, that is, from 7 AM to 11 PM and during nighttime, that is, from 11 PM to 7 AM. Table 1 shows the 20 categories of areas covered by the NCO and their corresponding ANLs. It should be noted that the ANLs would be applicable to all floors of the buildings in the same category. According to the TM of the NCO, noise measurement and assessment shall be made in terms of Leq,30-min (or more), preferably using hourly Leq, that is, Leq,1-h.

Table 1

Categorization of residential premises in accordance with the Noise Control Ordinance (NCO)

Degree to which the residential premises is affected by influencing factor	Not affected	Indirectly affected		Directly Affected
		By major road	By industrial area	By major road	By industrial area
Type of area
Rural area, including country parks or village-type developments	ASR1 A	ASR2 B	ASR3 B	ASR4 B	ASR5 B
Low-density residential area: Low-rise or isolated high-rise development	ASR6 A	ASR7 B	ASR8 B	ASR9 C	ASR10 C
Urban area	ASR11 B	ASR12 C	ASR13 C	ASR14 C	ASR15 C
Area other than those above	ASR16 B	ASR17 B	ASR18 B	ASR19 C	ASR20 C

ASR stands for Area Sensitivity Rating

A,B, and C are the ANLs stipulated in the TM of the NCO where the values are 60 dBA, 65 dBA, and 70 dBA between 7 AM and 11 PM; and 50 dBA, 55 dBA, and 60 dBA between 11 PM and 7 AM

Explanations:

 The ANLs are evaluated based on hourly Leq as stipulated in the TM of the NCO, independent from the height above ground, that is, the number of floors

 According to the NCO, “directly affected” means that the premises is at such a location that noise generated by the major road/industrial area is readily noticeable at the premise and is a dominant feature of the noise climate of the premise. A government’s plan (that is, map) shall be utilized to identify the target premises, followed by the confirmation from a site visit

 According to the NCO, “indirectly affected” means that the premises is at such a location that noise generated by the major road/industrial area, while noticeable at the premise, is not a dominant feature of the noise climate of the premise. A government’s plan (that is, map) shall be utilized to identify the target premises, followed by the confirmation from a site visit

 According to the NCO, “not affected” means that the premises is at such a location that noise generated by the road/industrial area is not noticeable at the noise sensitive receiver (NSR). A government’s plan (that is, map) shall be utilized to identify the target premises, followed by the confirmation from a site visit

Hong Kong's ANLs are at relatively high levels when compared to environmental noise criteria in other countries. For example, the City of Boston (MA, USA) indicates that the ANLs for residential areas are 60 dBA for daytime and 50 dBA for nightime, while those for the mixed residential and industrial areas are 65 dBA for daytime and 55 dBA for nighttime.[42] The WHO Regional Office for Europe conducted a study to assess night noise exposure across European countries. Their conclusion was that an Lnight,outside of 40 dBA should be set as the target in the night noise guidelines to protect the public in Europe. Meanwhile, the Lnight,outside value of 55 dBA was recommended as an interim target for the countries where the 40 dBA target cannot be achieved in the short term for various reasons.[39] In Japan, the ANLs range 50-60 dBA during daytime, that is, 6 AM to 10 PM, while the ANLs range 40-50 dBA during nighttime, that is, 10 PM to 6 AM.[43] When residential areas are facing major roads, 5-10 dBA will be added to the stated ANLs. The environmental noise standards adopted in mainland China are 50-60 dBA during daytime and 40-50 dBA during nighttime in residential areas.[44] However, even at relatively higher values, the following two research questions have yet to be properly answered. They are:

Whether outdoor noise levels are lower than the ANLs stipulated in the TM in Hong Kong, and

To what extent Hong Kong people can be exposed to noise higher than the ANLs.

Measurement of outdoor noise levels

Residential premises in Hong Kong are broadly classified into 20 categories, as shown in Table 1. Based on the Outline Zoning Plans, Outline Development Plans, and Layout Plans (hereafter called the Plans) published by the Hong Kong Planning Department, we identified 20 measurement sites for each category. We evaluated the suitability of the sites based on the recommendations in the TM of the NCO. For example, a building with apartments facing a major road in an urban area (from maps of the Plans) was categorized as a target site of Area Sensitivity Rating (ASR)14, as shown in Table 1. As the ANLs would be applicable to all floors of the buildings in the same category, we selected measurement locations at different levels, including the ground floor, 15th floor, and 30th floor in order to enhance the representativeness of the measured data for buildings in urban areas and low-density residential areas. It should be noted that noise levels due to traffic will decrease at higher floor levels in a building.[1316] Nevertheless, all the measured noise levels at different floors should not exceed the ANLs, as suggested in Table 1. For categories ASR4, ASR9, ASR14, and ASR19, we selected those units facing major roads, while for categories ASR5, ASR10, ASR15, and ASR 20, we selected those units facing industrial buildings. The list of target measurement sites was independently evaluated by a group of researchers who have performed noise monitoring and assessment work in Hong Kong.[454647] They confirmed the representativeness of the selected target sites.[46] As stated earlier, the type of each area, that is, urban, low-density residential, rural, or other, was determined based on the Plans, while whether the area was directly, indirectly, or not affected by the influencing factor including a major road or an industrial area was again determined based on the Plans [additionally, see the explanations in Table 1]. The horizontal distances between the measurement locations and the sources (such as a major road and an industrial area) ranged 3-10 m, and the common high-rise building configurations in Hong Kong were single slab, cruciform, twin tower, trident, and mixed, as reported by Lam et al.[47] We approached the owners of residential premises to seek their agreement to our performing noise measurements. In total, 203 out of the selected 400 residential owners allowed us to perform outdoor noise measurements. One-hour continuous noise measurements were carried out at 183 locations, while 24-h continuous noise measurements were carried out at 20 locations.

Noise measurements were conducted using B&K 2236 integrated sound level meters, which conformed with the International Electrotechnical Commission (IEC) specifications 60651 (Type I) and IEC 60805 (Type I). Each of the sound level meters was calibrated by a B&K 4230 acoustical calibrator before and after each set of noise measurement. The calibrations were all within ±1 dBA. The noise measurement was conducted at a position 1 m away from the window or facade at each of the selected residential premises, as indicated by the TM of the NCO. When the premises was located at the ground floor, the sound level meter was mounted at a position 1.5 m above ground level. We performed two sets of hourly Leq outdoor noise measurements at each of the selected premises during daytime between 1 PM and 3 PM on weekdays. We also collected two sets of hourly Leq outdoor noise measurements at each of the premises at night between 2 AM and 4 AM on weekdays. The horizontal distance between the measurement locations and the sidewalk of a major road or an industrial building ranged 3-10 m, while the vertical distance between the measurement locations and the major road ranged 1.5-90 m under the “directly affected” condition. The ranges were higher for those noise measurements under the “indirectly affected” and “not affected” conditions. In total, 366 daytime hourly Leq outdoor noise levels and 362 (4 discarded due to rain disturbance) nighttime hourly Leq outdoor noise levels were obtained. Other statistical noise indicators such as hourly L10, L90, and noise climate (L10-L90) were also recorded. Noise measurements were conducted between March and August while there was no fog and no rain, with the wind speed less than 5 m/s. The wind speed was monitored using a Testo 100 mm diameter vane anemometer that has a measurement range of 0-20 m/s with an accuracy of ±0.1 m/s. The other meteorological conditions, rain and fog, were checked using the readings from the automatic rain gauge networks (150 stations were set up in Hong Kong) and weather reports provided by the Hong Kong Observatory. Twenty owners of the selected residential premises (one from each of the 20 categories of residential areas) allowed us to measure daily, that is, 24 Leq,1-h outdoor noise levels on weekdays.

Statistical analysis

The measured noise levels were statistically analyzed. The range, mean, and standard deviation (SD) of the measured noise levels at each ASR category (daytime as well as nighttime noise measurements) were determined. The Kruskal-Wallis tests were conducted to compare daytime and nighttime noise exposures in terms of Leq,1-h, L10,1-h, and L90,1-h at different ASRs and to identify whether there was significant difference among those ASR sites. Normality of the measured noise levels at each ASR category was checked using the Kolmogorov-Smirnov test, in which a P value greater than 0.05 implies the measured values to be normally distributed, while a P value equal to or less than 0.05 implies the measured values to be not normally distributed. When the statistical distribution was normally distributed, the percentage of the noise levels that was equal to or greater than ANL was determined based on the normal probability function. When the statistical distribution was not normally distributed, the percentage of the noise levels that was equal to or greater than ANL was determined using the number of exceedances divided by the number of the measured noise levels.

Results and Discussion

Daytime outdoor noise levels

Table 2 presents the measured noise levels in terms of Leq,1-h, L10,1-h, L90,1-h, and noise climate (that is, L10-L90) at each category of the areas during daytime. It was found that the mean daytime Leq,1-h values ranged 54.4-70.8 dBA, while the mean daytime L10,1-h values ranged 58.9-72.9 dBA. The Kruskal-Wallis test was conducted to compare daytime noise exposures in terms of Leq,1-h at different ASRs. The results showed that there was significant difference (c2 = 181.24, df = 19, P < 0.000) among daytime Leq,1-h levels at different ASRs. In general, outdoor noise levels at different categories of urban areas were higher than that of low-density residential areas and rural areas.

Table 2

The number of measurement sites, the total number of measured noise levels, Leq,1-h, L10,1-h, L90,1-h, and noise climate (NC) for daytime noise measurements

Noise parameters	ASR1	ASR2	ASR3	ASR4	ASR5
No. of sites	11	11	10	7	8
No. of noise measurements	22	22	20	14	16
Leq,1-h in dBA
Range	45.6-64.3	50.9-70.4	49.6-65.1	57.9-68.6	53.7-61.1
Mean (SD)	56.3 (6.1)	56.3 (5.3)	57.0 (4.0)	63.0 (3.5)	58.1 (2.2)
L10,1-h in dBA
Range	48.5-69.8	53.6-72.4	53.0-68.9	59.4-71.2	57.2-65.5
Mean (SD)	58.9 (6.5)	57.5 (3.7)	60.9 (4.7)	65.3 (3.6)	62.2 (2.4)
L90,1-h in dBA
Range	42.4-57.2	42.6-54.4	41.7-56.5	51.1-62.1	48.8-56.8
Mean (SD)	49.8 (4.0)	49.6 (3.2)	49.6 (4.5)	56.9 (3.4)	53.3 (2.9)
NC in dBA
Range	2.1-17.6	3.0-12.0	0.9-18.7	2.8-16.7	1.3-15.5
Mean (SD)	9.1 (4.2)	7.9 (2.7)	11.3 (5.1)	8.5 (4.5)	8.9 (4.4)
	ASR6	ASR7	ASR8	ASR9	ASR10
No. of sites	14	12	8	8	7
No. of Leq,1h measurements	28	24	16	16	14
Leq,1-h in dBA
Range	49.7-67.5	50.4-70.0	50.2-60.2	55.9-75.1	58.4-74.0
Mean (SD)	59.1 (4.6)	59.9 (5.4)	54.4 (3.3)	64.8 (5.2)	64.9 (4.3)
L10,1-h in dBA
Range	50.9-69.2	52.9-70.6	52.1-70.3	61.9-78.6	64.1-74.4
Mean (SD)	61.6 (5.0)	63.2 (5.7)	59.0 (5.6)	67.4 (4.8)	68.5 (3.3)
L90,1-h in dBA
Range	46.5-66.8	46.2-64.0	45.7-57.1	52.4-63.0	51.9-63.7
Mean (SD)	55.0 (6.1)	54.2 (5.3)	50.9 (3.6)	57.9 (3.3)	57.9 (3.3)
NC in dBA
Range	1.3-12.6	3.6-17.9	4.2-13.2	4.0-19.1	6.9-17.0
Mean (SD)	6.6 (3.3)	9.1 (4.7)	8.1 (2.9)	9.5 (4.9)	10.6 (3.1)
	ASR11	ASR12	ASR13	ASR14	ASR15
No. of sites	11	10	5	12	5
No. of Leq,1h measurements	22	20	10	24	10
Leq,1-h in dBA
Range	60.1-72.6	59.6-70.0	57.7-66.8	65.2-77.8	62.7-71.7
Mean (SD)	63.8 (3.4)	64.0 (3.6)	61.7 (3.5)	70.8 (3.8)	65.9 (3.1)
L10,1-h in dBA
Range	61.6-74.9	60.8-73.6	59.2-70.9	65.9-80.1	63.7-71.9
Mean (SD)	65.8 (4.1)	65.7 (3.7)	64.1 (4.4)	72.9 (4.3)	68.0 (3.0)
L90,1-h in dBA
Range	56.9-68.9	52.7-64.6	54.7-60.1	62.5-74.9	61.5-70.1
Mean (SD)	60.4 (3.6)	59.8 (2.9)	57.0 (1.7)	68.5 (3.4)	64.6 (3.5)
NC in dBA
Range	2.9-7.6	2.0-13.3	2.9-13.8	1.4-8.6	1.8-6.5
Mean (SD)	5.5 (1.4)	5.9 (2.8)	7.1 (3.7)	4.3 (2.2)	3.4 (1.7)
	ASR16	ASR17	ASR18	ASR19	ASR20
No. of sites	13	10	10	8	4
No. of Leq,1h measurements	26	20	20	16	8
Leq,1-h in dBA
Range	55.4-66.6	57.5-69.1	57.6-74.2	54.7-77.9	60.9-70.3
Mean (SD)	61.5 (3.6)	63.9 (3.5)	65.1 (4.5)	67.4 (7.3)	65.0 (3.9)
L10,1-h in dBA
Range	56.2-69.3	59.2-70.8	58.6-78.2	56.6-79.0	62.4-71.1
Mean (SD)	64.1 (4.0)	65.9 (3.5)	67.7 (5.4)	68.7 (7.9)	65.9 (3.2)
L90,1-h in dBA
Range	54.3-62.2	54.3-66.4	55.7-66.5	51.2-72.4	57.2-64.4
Mean (SD)	58.7 (2.3)	60.8 (4.5)	60.8 (3.8)	63.6 (8.0)	61.0 (3.1)
NC in dBA
Range	1.2-10.0	1.6-9.8	2.9-11.9	1.4-11.1	1.0-6.1
Mean (SD)	5.5 (2.4)	5.1 (2.0)	7.0 (2.5)	5.1 (2.9)	4.0 (2.2)

When the Kruskal-Wallis tests were performed on L10,1-h and L90,1-h at different ASRs, the results showed that there was significant difference between daytime L10,1-h values at different ASRs (P < 0.000) as well as significant difference between daytime L90,1-h values at different ASRs (P < 0.000).

Figure 1 shows the box and whisker plots for ASR1-20. This figure indicates that at least one of the measurements taken at each site in the four residential areas in Hong Kong exceeded the ANL, except for categories ASR5 (rural area that is directly affected by an industrial area), ASR8 (low-density residential area that is indirectly affected by an industrial area), and ASR13 (urban area that is indirectly affected by an industrial area).

Figure 1

Box and whisker plots for ASR1-20 (Daytime noise exposure)

Nighttime outdoor noise levels

Table 3 presents the measured noise levels in terms of Leq,1-h, L10,1-h, L90,1-h, and noise climate (that is, L10-L90) at each category of the areas during nighttime. It was found that the mean nighttime Leq,1-h values ranged 52.6-67.9 dBA, while the mean nighttime L10,1-h values ranged 54.7-70.5 dBA. The Kruskal-Wallis test was conducted to compare nighttime noise exposures in terms of Leq,1-h at different ASRs. The results showed that there was significant difference (χ2 = 142.52, df = 19, P < 0.000) among nighttime Leq,1-h levels at different ASRs. In general, outdoor noise levels at different categories of urban areas were higher than those of low-density residential areas and rural areas.

Table 3

The number of measurement sites, the total number of measured noise levels, Leq,1-h, L10,1-h, L90,1-h, and noise climate (NC) for nighttime noise measurements

Noise parameters	ASR1	ASR2	ASR3	ASR4	ASR5
No. of sites	10	11	10	7	8
No. of noise measurements	20	22	20	14	16
Leq,1-h in dBA
Range	44.2-64.0	45.6-62.3	43.1-63.1	56.6-63.8	47.0-60.2
Mean (SD)	54.4 (5.1)	52.9 (4.8)	52.6 (5.7)	60.5 (2.2)	53.6 (5.3)
L10,1-h in dBA
Range	46.1-69.5	46.4-69.1	44.9-71.0	58.2-80.6	48.2-63.7
Mean (SD)	56.9 (6.3)	55.0 (5.2)	58.7 (6.6)	65.2 (6.1)	57.6 (5.5)
L90,1-h in dBA
Range	41.4-53.1	44.5-53.2	40.8-59.1	54.4-59.5	44.4-54.9
Mean (SD)	49.6 (3.4)	48.1 (2.3)	48.0 (5.1)	57.1 (1.8)	49.7 (3.9)
NC in dBA
Range	2.6-17.3	1.9-21.8	3.2-21.1	2.6-22.3	2.6-15.9
Mean (SD)	7.3 (4.2)	6.9 (5.2)	10.7 (5.1)	8.1 (6.0)	7.9 (3.4)
	ASR6	ASR7	ASR8	ASR9	ASR10
No. of sites	14	12	8	8	7
No. of Leq,1h measurements	28	24	16	16	14
Leq,1-h in dBA
Range	49.9-65.3	50.3-65.3	46.5-65.3	51.4-70.1	50.4-63.3
Mean (SD)	57.2 (4.5)	57.3 (4.8)	54.7 (6.7)	60.8 (5.2)	58.5 (4.4)
L10,1-h in dBA
Range	52.3-67.5	51.9-68.9	53.0-71.0	55.7-74.1	52.1-70.5
Mean	59.1 (4.7)	60.7 (5.3)	59.7 (6.7)	64.1 (5.3)	62.1 (6.4)
L90,1-h in dBA
Range	47.7-64.4	45.7-59.6	45.6-55.9	49.2-61.8	47.5-57.5
Mean	54.3 (4.7)	51.6 (3.7)	50.0 (3.9)	55.9 (4.2)	53.4 (3.2)
NC in dBA
Range	1.2-10.2	4.3-20.5	1.9-18.1	3.7-15.7	0.9-14.0
Mean	4.8 (3.0)	9.1 (4.3)	9.6 (5.0)	8.2 (4.1)	8.7 (4.7)
	ASR11	ASR12	ASR13	ASR14	ASR15
No. of sites	10	10	5	11	5
No. of Leq,1h measurements	20	20	10	22	10
Leq,1-h in dBA
Range	56.0-70.4	56.2-66.6	55.8-66.1	61.2-74.7	57.2-67.7
Mean (SD)	61.3 (3.8)	61.4 (3.4)	60.1 (4.2)	67.9 (4.1)	62.6 (3.6)
L10,1-h in dBA
Range	57.7-73.6	57.6-70.1	56.7-69.6	64.8-78.1	58.1-69.1
Mean	63.5 (4.6)	63.8 (3.9)	63.2 (5.0)	70.5 (4.4)	64.6 (4.1)
L90,1-h in dBA
Range	51.6-64.7	54.6-62.6	51.8-59.3	54.8-72.0	56.4-65.9
Mean	56.9 (4.1)	58.2 (3.1)	56.2 (2.5)	64.0 (4.2)	61.5 (3.4)
NC in dBA
Range	1.9-15.4	2.1-9.8	2.4-12.4	1.4-11.9	1.2-5.6
Mean	6.6 (3.5)	5.5 (2.3)	7.0 (4.0)	6.5 (3.2)	3.1 (1.5)
	ASR16	ASR17	ASR18	ASR19	ASR20
No. of sites	13	10	10	8	4
No. of Leq,1h measurements	26	20	20	16	8
Leq,1-h in dBA
Range	50.8-64.9	50.5-65.2	55.6-72.5	54.6-72.8	57.2-65.0
Mean (SD)	58.9 (4.0)	59.6 (4.7)	62.2 (5.2)	64.4 (6.2)	61.8 (3.2)
L10,1-h in dBA
Range	53.2-67.0	52.1-68.3	57.2-76.3	56.8-71.1	59.7-67.6
Mean (SD)	61.0 (4.6)	62.5 (5.1)	65.3 (5.9)	67.4 (7.4)	63.2 (3.5)
L90,1-h in dBA
Range	47.1-61.3	47.9-62.7	46.6-64.8	51.3-66.7	53.6-60.6
Mean (SD)	56.1 (4.0)	56.4 (4.8)	57.3 (5.0)	61.0 (5.4)	58.1 (3.3)
NC in dBA
Range	1.1-10.6	3.9-10.0	3.1-16.2	2.0-13.0	2.0-7.2
Mean (SD)	4.9 (2.7)	6.1 (2.1)	8.1 (3.6)	6.4 (3.6)	5.0 (2.4)

When the Kruskal-Wallis tests were performed on L10,1-h and L90,1-h at different ASRs, the results showed that there was significant difference between nighttime L10,1-h values at different ASRs (P < 0.000) as well as significant difference between nighttime L90,1-h values at different ASRs (P < 0.000).

Figure 2 shows the box and whisker plots for ASR1-20 for nighttime noise measurements. This figure indicates that many of the measured noise levels exceeded the ANLs in different categories of residential areas in Hong Kong.

Figure 2

Box and whisker plots for ASR1-20 (Nighttime noise exposure)

Daytime and nighttime outdoor noise levels

Normality tests were performed using IBM SPSS 20.0. Normality was determined based on the significance level of 0.05 from the Kolmogorov-Smirnov test, that is, a P value greater than 0.05 was taken to be normally distributed and a P value equal to or less than 0.05 to be not normally distributed. Table 4 shows that 65% of daytime noise data sets and 70% of nighttime noise data sets were normally distributed. For each of those data sets identified as “normally distributed,” the percentage of the noise levels that was equal to or greater than ANL was determined based on the normal probability function. For each of those data sets identified as “not normally distributed,” the percentage of the noise levels that was equal to or greater than ANL was determined using the number of exceedances divided by the number of the measured noise levels, as shown in Table 4.

Table 4

Normality check of the measured noise levels and the percentage of the noise levels greater than ANL at different measured sites

Daytime noise levels			Nighttime noise levels
Site: Noise levels that are normally distributed	Acceptable noise level (ANL)	Percentage higher than ANL	Site: Noise levels that are normally distributed	Acceptable noise level (ANL)	Percentage higher than ANL
ASR1	60	27.2	ASR1	50	79.9
ASR4	65	29.1	ASR4	55	98.7
ASR5	65	0.0	ASR6	50	94.0
ASR8	65	0.0	ASR7	55	68.4
ASR9	70	17.0	ASR9	60	56.1
ASR11	65	36.6	ASR10	60	36.4
ASR12	70	5.8	ASR11	55	94.1
ASR13	70	0.0	ASR13	60	51.1
ASR14	70	58.0	ASR14	60	96.7
ASR15	70	11.1	ASR15	60	75.3
ASR17	65	37.6	ASR16	55	82.9
ASR18	65	51.1	ASR17	55	82.8
ASR20	70	11.9	ASR18	55	90.9
ASR2	65	9.1	ASR20	60	70.5
ASR3	65	5.0	ASR2	55	18.2
ASR6	60	57.1	ASR3	55	30.0
ASR7	65	16.7	ASR5	55	50.0
ASR10	70	7.1	ASR8	55	37.5
ASR16	65	15.4	ASR12	60	50.0
ASR19	70	50.0	ASR19	60	68.8

Table 4 shows that only three categories of areas, namely ASR5, ASR8, and ASR13, would have noise levels below ANLs during daytime. These three areas are rural areas directly affected by an industrial area (ASR5), and low-density and urban areas indirectly affected by an industrial area (ASR8 and ASR13). The results are not surprising because most of the industrial areas in Hong Kong have had very little industrial activity since the mid-1980s. Many industrial buildings are used as warehouses and offices. In urban areas that are directly affected by major roads (ASR14), it was found that 58% of the samples were exposed to noise levels greater than ANL during daytime.

There were only four sites with the percentage of exceedances being less than 50%, that is, 18.2% at ASR2 (rural areas indirectly affected by major roads), 30.0% at ASR3 (rural areas indirectly affected by industrial areas), 37.5% at ASR8 (low-density areas indirectly affected by industrial areas), and 36.4% at ASR10 (low-density areas directly affected by industrial areas) during the nights.

To gain in-depth understanding of daily noise exposure at different categories of residential areas, 20 sets of daily, that is, Leq,1-h noise measurements were obtained from the sampled locations. The results are shown in Figure 3.

Figure 3

Daily Leq,1-h noise levels measured at different categories of residential areas in Hong Kong Explanation: The red line (—) represents the ANL

Figure 3 shows that one or more of Leq,1-h values exceeded ANLs at each of the sampled sites, except those of ASR5 and ASR8. At the sample sites of ASR3 and ASR13, results show that one Leq,1-h value exceeded ANL at night. Figure 3 also shows that the measured noise levels in the urban area directly affected by a major area were almost a constant throughout the 24-h period and the measured noise levels at nighttime exceeded ANL by 5-7 dBA, at around 65 dBA.

Summarizing the observations from Figures 1–3 and Tables 1–4, one can conclude that noise levels at most locations in Hong Kong very likely exceed the stipulated ANLs, in particular during nighttime, as shown in Figures 2 and 3, and Table 4.

When the measured noise levels at the rural area, low-density residential area, and urban area, which are indirectly or directly affected by a major road and by an industrial area were compared pair-wise, the results showed that the measured outdoor noise levels at residential premises with the influence of noise from a major road were higher than those at residential premises with the influence of noise from an industrial area [See Figures 1–3].

Strengths and limitations

This study was a territory-wide noise survey in Hong Kong. The right side of Table 4 shows that Hong Kong's nighttime noise levels very likely exceed the ANLs. It shows that the percentage of exceedances ranged from 18.2% at ASR2 (rural areas indirectly affected by major roads) to 98.7% at ASR4 (rural areas directly affected by major roads). For the areas (rural, low-density, urban, and others) directly affected by major roads, the percentage of exceedances ranged 56.1-98.7%. In addition, the percentage of exceedances ranged 50.0-96.7% in urban areas (ASR11-ASR15). These results indicate that 50% or more inhabitants of Hong Kong, that is, 3.6 million people are exposed to noise levels that are higher than ANLs at nighttime. In particular, as more than 90% of the Hong Kong population live in urban areas, 3.2 million people can be exposed to noise levels at or above 60 dBA at night. Nevertheless, it should be noted that all measured noise levels were recorded outdoors, and sound insulations due to facades and windows were not taken into consideration. A more precise noise modeling, including sound insulation of building facades, will give a more accurate prediction of noise exposure of Hong Kong's population. The findings of this study were comparable with other studies conducted in Hong Kong and other cities in the world.[364849] For example, Lam and Ma[36] estimated that about one million Hong Kong people were affected by excessive traffic noise using L10,1-h at 70 dBA as a criterion few years ago. Among other cities, Stansfeld and Crombie[48] reported that 20% of the population of London were exposed to a noise level at or above Leq,12-h 60 dBA during daytime. Lopez et al.[49] reported that the nighttime hourly Leq values recorded in the big towns of Spain were at or above 60 dBA throughout the night.

It should also be noted that the means of different sets of Leq-1h provided information on noise exposure, but these arithmetic means could underestimate the noise nuisance felt by people as they do not take into account the logarithmic nature of noise levels.

Conclusion

Environmental noise affects people. Many governments have established noise standards and ANLs with an aim to protect citizens from adverse physiological, psychological, and social effects associated with noise exposure. Unfortunately, literature on noise measurement, monitoring, and control rarely looks at whether or not the measured outdoor noise levels exceed the stipulated noise standard/ANL, and the extent of exceedances, if any. This paper reports a territory-wide noise survey in the built environment of Hong Kong. Results show that the mean daytime hourly Leq values ranged 54.4-70.8 dBA, while the mean nighttime hourly Leq values ranged 52.6-67.9 dBA. When the measured outdoor noise levels were compared with the stipulated ANLs, it was found that only three categories of areas, namely rural area directly affected by an industrial area, low-density residential area indirectly affected by an industrial area, and urban area indirectly affected by an industrial area, had outdoor noise levels below ANLs during daytime. This is because Hong Kong's industrial areas have little industrial activity and many industrial buildings have been used as warehouses and offices in recent years. All other areas (and all areas during nighttime) have outdoor noise levels at or above the ANLs.

Implications

A significant percentage of Hong Kong inhabitants in urban areas can be exposed to a noise level above ANL, that is, 60 dBA at night, a level at which adverse health effects occur according to the WHO.[3940] The WHO[39] reported that the maximum number of awakenings for an Lnight,outside of 60-65 dBA is around 300 per year and may even cause psychiatric disorders. Here, the concerned government departments shall review and revise the existing noise management for new urban development and redevelopment. Besides, they shall look at the purposes of setting ANLs in light of the findings that most ANLs are not achievable. In building design, improving sound insulation[50] and sound absorption, such as by installing green roofs,[51] can protect residents from high outdoor noise levels.

Hong Kong is one of many congested cities in China. The findings of this study imply that similar studies can be duplicated in other places to assess environmental noise exposure across China's cities.

Financial support and sponsorship

Nil.

Conflicts of interest

The authors report no conflicts of interest.