1. Introduction
Noise Pollution in developing cities: Noise pollution is a menace in cities. Noise pollution is dynamic in nature and variation in traffic loads is a factor that influences the level of noise pollution in a particular area. Beeping or honking also brings a significant incremental change in noise levels. Researchers found the noise levels elevated from 2 to 13 dB [
1,
2] during honking. Beeping has become a widespread occurrence in India, particularly at congested road intersections. The congestion increases manyfold due to unlicensed camping (temporary shops at the edges of the road) on major thoroughfares [
3,
4]. Due to this congestion, the traffic levels greatly increase, and this tempts the drivers to honk more which, in turn, causes noise pollution.
The developing city intersection conforms in many ways to a hazardous area for noise pollution. In under-developed cities, often the city builds around road intersections. Intersections of national highways, state highways, or two-state highways are the prime areas around which a city builds up. The roads and crossroads (intersections) are associated with less control over traffic and multiple vehicles of varying speeds frequently run at similar times on the roads. The congestions at city intersections keep on increasing with growing establishments around the intersections. The authorities of several such unplanned cities also allow the establishment of schools, health centers, shops, sports centers, roadside open markets, etc. around the intersections, which results congestion. People living near such intersections experience loud noise throughout the day which affects people’s health. Very few studies have highlighted the importance of relating decibel levels to human health deterioration, so it becomes an important topic to study. Apart from suggesting a model to map the noise level in a particular area, the authors have tried to relate it to levels of noise exposure and determine the associated health hazards [
5].
Health Hazards of Noise Pollution: Severe noise pollution can cause detrimental effects on one’s health as it influences everyday activities (sleep, speech, and work). Further, irritation, cognitive disability, dementia, cardiovascular disorders, as well as respiratory problems have been found to be caused by ambient noise exposure [
6]. Numerous studies have been undertaken in the past to examine noise emissions as well as their influence on human wellbeing [
7]. According to recent research by the World Health Organization (WHO), 466 million individuals worldwide suffer from debilitating hearing loss, with the number expected to rise to 900 million by 2050, or one in every ten people [
8]. The cumulative disability-adjusted life years lost in European nations owing to ambient noise is anticipated to somehow be 1.6 million [
9]. Traffic noise is the main source of the noise-induced health problems. As per the literature, [
10] greater than 85 dB noise levels for 8 h in a working environment leads to noise-induced health problems. Noise exposure can cause two kinds of health effects: non-auditory and auditory. The former type includes stress, physiological and behavioral effects, and safety concerns, while auditory effects include hearing impairment resulting from excessive noise exposure. Noise-induced hearing loss (NIHL) is the main concern related to occupational noise exposure [
11]. Non-auditory effects include sleep disturbances. Uninterrupted sleep is known to be a prerequisite for good physiological and mental functioning of healthy people [
7]. It is reported that 80-90% of cases of sleep disturbance originate from noisy environments [
12]. Cardiovascular diseases, stress, and annoyance are some examples of non-auditory effects. (Further literature on health hazards is included in the
Supplementary Materials).
Noise Characterization and Noise Exposure Mapping: The authors realized the need for accurate characterization of noise levels around noisy road intersections. Noise levels vary over various locations with time [
13,
14]. Generally, noise levels are monitored at various locations and at different times [
15]. It is not possible to monitor the noise levels for every location and throughout all the days. The sampled noise levels of these monitored points are used to predict the noise levels of neighboring locations (for which no noise data are recorded) using the technique of noise modeling [
16,
17]. The predicted levels are then showcased over a map (termed a noise map), which indicates spatial variation in noise levels. These maps can indicate the hotspots for noise and areas of less noise or silence, and thus, these can be used for noise management planning. Based on a survey of the literature, authors found that there are many works related to noise mapping based on GIS mapping [
18,
19]. Regional road traffic noise management strategy was coined under RRTNMS [
20]. Under this plan, the road was graded based on the expected sound pressure level. Further, various control methods were adopted based on the ranking’s results [
21,
22]. After the publication of the Environmental Noise Directive-END2002, new initiatives were taken over computing models for efficient noise mapping and frequent upgradation of maps [
5,
23]. Software such as RLS-90, FHWA, as well as CoRTN in Sound PLAN is used frequently for noise mapping purposes. The Sound Propagation Retardation technique was used for traffic noise prediction modeling by some researchers [
2]. Scores of countries have established vehicle noise emissions standards and regulations to reduce road traffic noise levels [
24,
25]. The Ministry of Environment, Forest, as well as Climate Change of the Government of India released the noise pollution (regulation as well as control) laws in the year 2000. The Ministry directs the state governments to designate the region into silent, domestic, commercial, and industrialized regions, as per this guideline. The ‘A’ weighted noise limitations in the industrial zone are 75/70 dB (A) throughout the day/night, followed by 65/55 dB (A) in commercial, 55/45 dB (A) in residential, and 50/40 dB (A) in the quiet zone. Noise mapping can help when designing urban highways and assessing current traffic noise levels [
26,
27] The noise indicators such as L
eq, L10, L50, L90, are the other indicators sometimes used to indicate noise levels. Different governments use these indicators to anticipate sound intensity [
20,
28,
29] at important road intersections [
28]. All this necessitates the generation of noise maps for city intersections to indicate noise levels, hotspots, and estimate the primary locations of noise sources [
30,
31].
Linking noise exposure levels to health hazards of intersection: In a developing city intersection, the characteristics of noise levels can be very different. Congested unorganized intersections experience intense noise levels throughout the day. Poor shopkeepers and temporary workers often need to spend around 10 h daily in the intersection for their living. These dwellers are exposed to intense levels of traffic noises as they work almost at the edge of the road. It was required to identify the extent of noise exposure of the dwellers in the noisy corridor. It was also required to relate the noise exposure to the health hazards of the dwellers. The types of health hazards and the extent of impacts were required to be determined. Previous studies on noise mapping were limited to the identification of the hotspots of noise pollution in a city environment. The successful identification of hotspots can only indicate the alarming zones in a city where, if people are working or living for a prolonged duration, their health can be harmed. It is thus required to estimate the characteristics of noise exposure to dwellers working near intersections. The noise characterization requires sampling of noise levels and the generation of a noise map using noise modeling. Once noise levels and exposure durations are estimated accurately, they can be utilized to predict the noise exposure map for the intersection. The information on the health parameters of dwellers working near the intersection can then be related to the impact of noise on health.
The entire objectives were attempted to be addressed in three stages, that is, (a) developing of a noise map to estimate the levels of noises generated at the intersection, (b) estimating the levels of noise exposure at the intersection for the dwellers, and (c) estimating the hazards concerning the health of the dwellers, which includes the estimation of hearing impairment as well as psychological and physiological damage.
The authors have chosen a typical congested developing intersection for the study. A methodical approach was applied for sampling of noise levels at the intersection. The detailed noise maps were generated using a noise mapping technique and integrated with the duration of noise exposure to derive the noise exposure maps. Finally, noise level and noise exposure level mapping were related to the health hazard study for the noisy intersection. Assessment of health hazards for noise exposure can be made using three approaches, that is, through an audiometric test which tests any physical damage to hearing organs, physiological tests which estimates physiological damage to individuals exposed to high noise levels, and a questionnaire, which tests the general health hazards of individuals exposed to high noise level through a series of questions targeting the estimation of the relationship between lifestyle, detailed nature of noise exposure, and hazards to health (
Figure 1).
It is assumed prolonged intense exposure to noises will cause auditory losses. It can also lead to non-auditory damage leading to hypertension, cardiovascular diseases, stress, annoyance, and sleep disturbances. Considering that non-auditory damages can also be caused by lifestyle diseases, dietary patterns, air pollution, etc., these causative factors were attempted to be largely ruled out for the study site. This was accomplished by conducting a noise exposure and health hazard study for an alternative study site that was devoid of noise pollution. It is understood that if the alternative nearby study site that had a very low noise exposure level exhibited lower health hazards, in particular significantly lesser auditory damages, then the high noise exposure levels of the intersection may directly be related to the severe health hazards recorded there.
Research Gap for Noise Hazards Determination of Developing City Intersection: There is a need to estimate the hazards of traffic noise around a city intersection. The determined technique will have the following characteristics:
- ○
Characterization of detailed noise exposure to dwellers of a developing city intersection.
- ○
Determination of health hazards of noise exposure at the developing intersection.
- ○
The technique developed should be able to estimate the noise levels at a different time in the day and at different locations around the intersection.
- ○
It should be sensitive to the noise sampling period.
- ○
It should provide a reasonably accurate prediction as a function of time and space.
- ○
It should be simple to adapt.
- ○
The technique should be able to generate a noise map indicating noisy and less noisy places of the intersection.
- ○
The technique should be able to relate the estimated noise levels with the noise exposure levels of people working at and around the intersection over 8–10 h daily.
- ○
The technique should determine the noise exposure for the visitors of the intersection during their time of visit (assumed 1 h and 30 min in a day).
- ○
The technique should relate the noise exposure to the hazards to health.
- ○
The technique must ensure that the health hazards recorded around the city intersection are indeed due to noise exposure
The authors tried to work upon the above research gap through the design of a research methodology that addresses the research gap.
Section 2 discusses the adopted Material and Methods for the study.
2. Materials
Two study sites were chosen to verify the impacts of traffic noise. The first site at Bahadurpur, Jais, in the state of Uttar Pradesh in India was representing the noisy intersection of a developing city. It was squarish in shape and had dimensions of 500 m × 500 m. The intersection of the Raebareli-Sultanpur highway and the Jagdishpur-Jais state road is centrally situated within the area. The intersection, popularly known as the Bahadurpur Chauraha, receives a high load of traffic throughout the day. The second site of Pure Ganga of a similar stretch was located within two kilometers of the first site and had a very sparse flow of traffic (
Figure 2,
Figure 3,
Figure 4 and
Figure 5).
Authors, being the residents of the area, have a comprehensive understanding of spatial and temporal variability of noise at the intersection. The chosen intersection hosted several vendors setting up their shops adjacent to roads, making the intersection more congested and crowded. Further, the dwellers work 8–10 h (9:00 a.m.–7:00 p.m.) daily near the noisy intersection, thus, were expected to be exposed to high levels of noise on a regular basis.
A standard sound pressure level meter CESVA SC-310 (a class—1 meter of range 23 dBA to 137 dBA and of a frequency range of 20 Hz to 10 kHz) was used for the collection of noise data and JUNO GPS (± 2–5 m accuracy) and Google images were used for the recording the coordinates. During the daytime, this highway is overloaded due to the trucks and local vehicles passing through the area to cities like Jais, Raebareli, and Jagdishpur. Several houses are also present near the road corridor and the residents of this area are expected to be affected by the high noise levels (
Table S1, Supplementary Material). Noise data was monitored close to road intersections (17 points) and away from the Bahadurpur intersections (17 points), strategically chosen for noise modeling. Similarly, noise samples were recorded five times a day (7:00–9:00 a.m., 9:00 a.m.–1:00 p.m., 1:00–3:00 p.m., 3:00–5:00 p.m., and 5:00–7:00 p.m.) for several days over a month. Noise samples were also taken for 10 min durations (assumed as long term), and 20 s duration (assumed as short term) at every instant. Different sampling durations were chosen to verify the impact of time-varying noise data on noise prediction. The noise data and terrain parameters between the source and predicting points were determined and integrated using a noise model. Licensed MATLAB 21 was used for the prediction of noise levels of desirous points at different times of the day. Sound pressure levels, L
eq, and time of noise exposure were used to determine noise exposure levels using an exposure computing model. The SPL, L
eq, and noise exposure levels of different points were used to generate a noise map (using licensed Arc GIS 10.2 software developed by ESRI, U.S.A.) and an exposure map for the different types of dwellers working or visiting the intersection.
The study tried to relate traffic noise exposure with the type of exposure. Noise can be short-term and impulsive or long-term at high intensity levels. Long-term noise levels as frequently found in the busy intersection of more than 75 dB can cause significant damage to the ear. Audiometry studies the physical damage to the ear. In audiometry, noises of different frequencies and intensities are played to ears to check the minimum threshold level of hearing at different frequencies. The traffic noise can induce auditory damage hearing loss, tinnitus, etc. The traffic noise can also induce non-auditory damage in the cardiovascular system causing, cardiovascular disease, hypertension, stroke, etc. It can also negatively contribute to psychoacoustic abilities in human beings, causing annoyance, sleep disturbance, cognitive dysfunction, reading ability, depression, etc. Various health hazards were evaluated for the dwellers working (living near) the developing city crossing in relation to the extent of exposure to traffic noise.
It was understood that the health hazards for the dwellers of the developing noisy city intersection can be thought to be primarily caused due to noise pollution. Thus, a nearby area that is less noisy may not cause similar health hazards to people living there. Thus, noise mapping and health hazard assessment were made for a second study site, Pure Ganga. The second study was about 2 km away from the first study site and 2 km away from any kind of noisy road traffic. An audiometric test, a medical test, and a questionnaire were conducted for study site 1 and study site 2 to record the health hazards.
The authors conducted a study of 100 people living near to and far from the source (about 2 km distant) of the primary site. Based on the similarities of both locations, the authors considered that these should be the best locations for this study. The same type of people live in both locations in terms of their lifestyle, food habits, social status, and medical facilities. The only factor that differed was their working location. Based on comparable research survey sources, the authors framed 20 questions (for details see
Supplementary Material). Nine out of the twenty questions were about basic information, such as where you work, how many hours you worked there, and how long you’ve been there. The remnant 11 questions were based on any health-related problems he/she suffered from. Hazard assessment was also conducted through an audiometric test and a physiological test. In the audiometric test, a laptop and computer program generated tonal sound and was used to determine if the person in question suffered from a hearing impairment or not. It generated sounds of different frequencies from 250 Hz to 5000 Hz, and of different intensities, and was played using headphones. The person in question was asked if he/she heard the sound or not. Sounds of different intensities and frequencies are played. And the responders were asked about the minimum intensity for hearing the sound at a particular frequency. The loudness of speech and speech reception ability were also tested using an SPL meter to indirectly relate with a person’s loss in hearing abilities. The physiological test was conducted to measure blood pressure using a sphygmomanometer (Model No. Dr Trust ICHECK PRO 116). The survey was conducted by meeting each roadside dweller and conducting the test conveniently.
5. Discussion
Prolonged exposure to noise has ill effects. Intense traffic noise can also have auditory and non-auditory health impacts. The adverse impacts on health include hearing impairment, physiological damages to the cardiovascular system, leading to hypertension, stroke, etc., and psychological damage causing sleep disturbance, annoyance, depression, etc. The extent of impact very significantly depends on the type of noise, levels of its intensity, duration of influence to noise, etc. The noises originate from road traffic around the intersections and significantly vary in space and time. In a developing city intersection, the characteristics of noise levels can be very different. The growing of roadside shops at the intersection constricts the intersection and increase the congestion. Vendors and dwellers participate in commercial activities around the intersection, which very significantly influences the characteristics of traffic noises generated from there. Poor shopkeepers and temporary workers often need to spend around 10 h daily in the intersection for their living. These dwellers are exposed to intense levels of traffic noises as they work almost at the edge of the road. It was required to identify the extent of noise exposure of the dwellers at the noisy corridor. It was also required to relate the noise exposure to the health hazards of the dwellers. The types of health hazards and the extent of impacts were required to be determined.
Previous studies on noise mapping were limited to identifying the hotspots of noise pollution in the city environment. The successful identification of hotspots can only indicate the alarming zones in a city where, if people are working or living for a prolonged duration can harm their health. The extent of exposure to noise and its hazards to health are primarily not quantified in previous studies. Further, in the cases of the developed city intersections, the land use and land cover remain very organized, thus residential places or people’s living places remain significantly away from the traffic intersections, which avoids people’s adverse exposure to noise. In the cases of developing cities, the intersections essentially become a commercial hub drawing vendors to use the space for setting up temporary shops. Setting up of temporary roadside shops, adjacent to roads constricts the intersection. It also increases the level of congestion on roads at the intersection. The vendors and shopkeepers spend 8–10 h daily at the edge of the road (in the intersection) for their living getting exposed to intense levels of noise. As the places become marketplaces, the people from the nearby areas approach the noisy intersection for commercial activities; however, their duration of exposure to intense noise becomes lesser than those who need to spend 8–10 h daily at the intersection for living. The characteristic of noise levels prevailing at different parts of the intersection is required to be determined at different times of the day. It is also required to estimate the number of dwellers living or working at the intersections, whose health can be seriously impacted. Finally, the duration of exposure to noise is required to be determined at the intersection with the associated estimation of detailed health hazards.
The entire objectives were tried to be addressed in three stages, that is, (a) developing of noise map to estimate the levels of noise generating at the intersection, (b) estimating the levels of noise exposure at the intersection for the dwellers, and (c) estimating the hazards for health for the dwellers, which includes the estimation of hearing impairment as well as psychological and physiological damages.
In order to characterize the noise levels at the intersection, noise levels are required to be monitored. The noise levels around the intersection were systematically monitored at 17 + 17 strategic points of the noisy Bahadurpur intersection. They are monitored and tried to be predicted and mapped. To characterize the variability in traffic noise levels in a day for the area in different hours, various sampling slots were chosen for field monitoring of noise levels, i.e., at 7:00–9:00 a.m., 9:00 a.m.–1:00 p.m., 1:00–3:00 p.m., 3:00–5:00 p.m. and 5:00–7:00 p.m. Further, the noise data were collected for 10 min (long-term) and 20 s (short-term) at each sampling interval, several times over a month. The adopted approach was inspired by similar studies by other researchers (Ref: STRATEGIES FOR NOISE SURVEYS by Professor J. Malchaire, Health, and Safety Executive, UK, 2011; Current Science, Biswas et al. 2012). Different sampling time periods were tried to characterize the noise levels for the intersection comprehensively. Twelve-hour equivalent noise levels (Leq) were determined using a long-term and short-term noise sampling regime. Leq noise levels for the noisy Bahadurpur intersection were primarily close to 100 dB. There were not many differences in Leq noise levels for different sampling intervals (M1, M2….M5.) using long-term or short-term noise sampling. It thus indicates that the noise characterization for the intersection can also possibly be carried out using short-term noise monitoring, providing a quick assessment of noise levels in an area.
Noise prediction and mapping were also conducted for different times of day using modeling and mapping. Here, the methods used for noise prediction were dependent on different types of attenuation calculations (e.g., distance attenuation, building attenuation, etc.) as practiced in conventional noise modeling. The predicted results for chosen points were then used for noise mapping using GIS IDW interpolation techniques. These results were also validated with some test points at the Bahadurpur crossing. A total of 90% of predicted noise levels deviated by ± 1–4 dB from the actual noise levels recorded by the noise measurement instrument (Sound Pressure Level meter used in the study). So, this method is highly acceptable to predict the noise value in any area and generate the noise map, even a short duration of noise monitoring at strategic points can also reasonably generate a noise map for the city intersection.
The authors used the available noise mapping methods. The authors tried to predict the noise levels on and around a busy and noisy developing city intersection. The special characteristic of the intersection was the presence of a significant number of working-class populations. Many vendors, hawkers, and sellers set up their temporary shops around the road intersections (within 1 m of the edge of the road, i.e., on the pavement). These sellers need to work 8–10 h daily. The authors tried to establish a technique to determine the noise exposure and health hazards for these vulnerable groups. The authors did use the available noise prediction technique to determine the noise levels at various parts of the intersection. Monitored noise data, terrain data, and prediction models were used for the prediction of noise levels. The predicted noise levels for 34 points and the surrounding locations were integrated with GIS to develop the noise map for the study area. The noise predictions were done at different times of the day. Noise exposure levels were also computed based on the levels of noise with their likely exposure durations for various locations. Noise level and noise exposure level mapping were finally related to the health hazard study for the noisy intersection. The results were compared with that of a non-noisy nearby area. The technique of hazard estimation and its comparison were used to relate hazards of noise pollution to the vulnerable population of the study area. The authors realized that many people set up their roadside shops within 1 m from fringe/pavement and they work there 8–10 h daily. These people were exposed to high noise levels throughout their working period. Their health hazards from noise exposure were required to be evaluated. Research questions were: (i) did these people suffer due to noise exposure; (ii) did they have any health concerns, i.e., physiological, or psychological; (iii) to what extent did the health concerns exist among the dwellers of the study sites (iv) to what extent were the health concerns related to the duration of noise exposure, time of the day, and/or noise level for a location. It was attempted to address these research questions for the vulnerable population of the study site.
The prolonged intense noise exposure can lead to auditory losses along with non-auditory damage. It can very significantly impact the nervous system through the inner ear, which in turn can adversely impact the cardiovascular system and human abilities for cognition. Hypertension, ailment of the heart, stroke, stress, sleep disturbance, annoyance, and tinnitus, have frequently been reported with exposure to traffic noises (literature review can be seen in
Table 1 and
Figure 12). There is primarily limited literature on the quantification of the adverse effects of traffic noises on human life. In case the noise level or noise exposure level reduces significantly, it becomes difficult to relate the noise level with any palpable health effect. For example, at 5 m, 10 m, and 20 m from a noise source of 100 dB the receiving noise levels will be 75 dB, 69 dB, and 63 dB respectively, following the inverse square law of noise modeling as explained in equations 1 and 2. In the case of the developed intersection, people live further away from the busy intersection, and the presence of noise barriers (if present) reduces the noise levels even further. In the cases of developing intersections vendors setup their shops at the edge of road where noise levels cannot attenuate any significantly, as in the present intersection of Bahadurpur, the equivalent noise levels (L
eq) become close to 100 dB and noise exposure levels become >85dB, which is alarming. A large number of shops can be identified within the 100 dB level zones of the noise map (
Figure 13). To assess the health hazards for the dwellers the audiometric test, questionnaire and physiological tests were conducted.
Auditory losses of hearing impairments are primarily caused by the prolonged exposure to high-intensity noise levels. The non-auditory damages leading to hypertension, cardiovascular diseases, stress, annoyance, and sleep disturbances can be related to exposure of high-intensity noise levels and other causes which include lifestyle diseases, dietary patterns, air pollution, etc. The non-auditory diseases if occur due to lifestyle diseases, dietary patterns, and air pollution will not cause any damage to the auditory system or hearing impairment. It was thus, required to establish that the health hazards to dwellers of the intersection (who live/work there for 8–10 h) are primarily caused by exposure to intense noise levels. The health hazards caused by long duration noise exposure at Bahadurpur intersection were thus compared with the health hazards to non-noisy nearby study sites of Pure Ganga (village) and health hazards to visitors of the Bahadurpur intersection with a limited time of exposure (e.g., 1 h and 30 min in a round trip).
A detailed questionnaire was conducted on 100 people at each of Bahadurpur and Pure Ganga study sites. Questionnaires were strategically designed in the native language (Hindi) to estimate hours of exposure to high noise levels or working adjacent to the road intersection. It also enquired about any feeling of stress or anxiety in their life or working place (in the intersection), the experience of disturbed sleep, etc. were also enquired about separately. Feedback was taken further about any complaint of cardiovascular disease and/or related family history. Age, gender, career goal, income, expectation, etc. were also enquired about to register physiological and psychological deficiencies. A separate medical examination was conducted to identify individuals with hypertension, high pulse rate, etc. Audiometric tests were conducted on every individual to determine cases of hearing impairments. Similar tests were also conducted on 100 people at the less noisy Pure Ganga site as well. Comparing the observations of the two sites (
Figure 26 and
Figure 27), the authors found very interesting results. For people who are living near the noisy Bahadurpur intersection and in the <40 age group, 75% of people were suffering from some diseases and 25% of people were not suffering. In the >40 age group 92% of people were suffering from some diseases. Further, it was found that the hearing problem was prevalent in 79% of the people in the <40 age group, and it was 95% in the >40 age group. In the second location (Pure Ganga village), it was found that 29% of people were having certain diseases in the <40 age group, which came up to 41% in the >40 age group. In the age group of <40 years, the hearing problem was found among 13% of the population, which increased to 30% for the people of >40 age group in Pure Ganga village. This survey confirmed the traffic noise to be the prime causative factor for health hazards for the people working near the noisy Bahadurpur intersection (who spend a significant time throughout the day at the intersection).
A total of 6% of people in the Bahadurpur intersection of the <40 age group complained about cardiovascular ailments which were only 1% in the case of Pure Ganga village. Although in the >40 age group both the places had a similar percentage (7–8%) of people suffering from cardiovascular diseases. 9–12% of people in two age groups in Bahadurpur were suffering from high blood pressure which was nearly half in the case of the <40 age group people of Pure Ganga. Psychological diseases of anxiety, sleep disturbance, and stress, if combined, will be 32–36% in the Bahadurpur intersection, which was much higher than the similar figure of the Pure Ganga village area (16–15%). There were cases when people complained about suffering from multiple diseases (auditory and non-auditory in nature). In terms of suffering from multiple diseases too, the Bahadurpur intersection was found to be very harmful (16–24%) compared to that of Pure Ganga (0–6%). Tinnitus, the hearing-related ailment was also reported higher at the Bahadurpur intersection (12%) compared to that of Pure Ganga of 4–8%. The hearing impairment measured through the audiometric test corroborated further the alarming state of health hazards in the Bahadurpur intersection (79–95%) compared to that of Pure Ganga (13–30%). Different ailments were primarily found to increase with age. In the case of the noisy Bahadurpur intersection, the noise exposure had adversely impacted health even in the lower age group.
The occupation of dwellers probably played a very important role in establishing a relationship between noise exposure and health hazards. A very significant number of peoples work around the noisy Bahadurpur road intersection for over 8–10 h every day. Mostly, they set up their shops with temporary setups very close to the edge of the road (often over the road pavements). This working-class population is involved in selling fruits, vegetables, granaries, utensils, and other day-to-day items. Owners of these unshaded shops are the most vulnerable population exposed to high noise exposure. Compared to these people there were populations in Pure Ganga village in the nearby area (away from the intersection) not influenced by high noise levels. These people are primarily involved in farming and animal husbandry. They had lesser stress, sleep disturbances, and cardiovascular problems. In the absence of traffic, noise, and other external hassles, they were found to be calmer, although economically may not be better off than the noisy Bahadurpur intersection people.
The Bahadurpur intersection was very noisy and the health hazards at the place were related to the duration of exposure to noise. The visitors to Bahadurpur intersection are exposed to intense noise levels but the durations of exposure is generally significantly lesser than that of the dwellers working in the intersection 8–10 h regularly. A typical case of a visit to Bahadurpur intersection from a non-noisy nearby place (about 250–500 m away) and back taking 1 h and 30 min was considered to estimate its noise exposure during the visit. It was determined that a regular round trip to the Bahadurpur intersection by a visitor for 1 h and 30 min can have exposure levels primarily less than the alarming threshold of 85 dB. However, in the busy evening hour (5:00–7:00 p.m.) it can be more than 85 dB. Thus, regular visits to Bahadurpur intersection at a busy hour can be harmful, but that generally does not happen. The authors being native to the place approach the Bahadurpur intersection from a non-noisy place 500 m away, and hardly experience any adverse health hazard as they do not visit the intersection regularly.
The technique has been able to predict the levels of noise exposure for the noise corridor, the place of working for the vulnerable group. The attempt used monitoring of noise levels and to predict noise level and generate the exposure maps. More number of data points and a greater number of time intervals could have been chosen to generate higher resolutions of exposure maps (in space and time). However, it would have involved a larger cost and time for data collection, processing, and modeling which might improve the prediction accuracies somewhat. On the other hand, it may be more desirable to devise a technique that involves a lesser number of noise data sampling but still can generate a noise map and reasonably predict exposure and health effects for the vulnerable pockets in the city intersection. Keeping this in mind the authors had devised short-term noise monitoring for 20 s instead of 10 min and came up with similar noise maps which deviate little and still indicate high noise level adjacent to the intersection. It could have also been used for noise exposure mapping and for relating to health hazards. In the future work, attempts can be made to reduce the sampling further and still relate to noise hazards for a place conservatively.
The noise exposure levels at the roadside open-air shops near the noisy urban intersection of Bahadurpur were found to be very high (close to 100 dB). The results of physiological study and audiometric studies also indicated the marks of serious damage to hearing, cardio-vascular and psychological abilities (as explained in the
Section 4 and
Section 5,
Figure 15 and
Table 1). The questionnaire-based survey conducted on the 100 people living at the Bahadurpur study site also confirmed serious adverse impacts (e.g., hearing impairment, tinnitus, sleep disturbance, cardiovascular diseases, hypertension, etc.). The serious noise pollution at Bahadurpur was considered to be the prime cause of the adverse health impacts on the population. Auditory, cardiovascular, and psychological diseases were significantly less among the population of the nearby Pure Ganga village having no significant traffic noise problem (
Table 1 and
Table 2). Further, the authors found a parallel in terms of noise-induced adverse health impacts of Bahadurpur intersection with that of Quito Ecuador (published by Virginia et al in Int. J. Environ. Res. Public Health 2022) [
61]. Their study indicated the community response to noise in the hotspot at the major Road of Quito (Ecuador). The authors found that 13.11% of participants were not annoyed while 25% were slightly annoyed, 35.66% were moderately annoyed and 26.23% participants were highly annoyed; thus, in total 85.89% of participants were annoyed with light to high intensity of annoyance. Similarly, 32.38% were slightly sleep-disturbed, 31.15% were moderately sleep-disturbed, and 19.67% were highly disturbed of sleep; thus, in total, 83.2% of participants were sleep disturbed with low to high intensity. Similar to the Bahadurpur intersection, the study site of Mariscal Sucre Avenue at Quito Ecuador was underdeveloped and people there used to live in temporary roadside accommodation, experiencing very intense traffic noise levels, causing serious health damages.
6. Conclusions
Noise pollution has been a menace for the developing Bahadurpur city intersections. In the unplanned urban areas, the city was built around the road intersections. Traffic of different types, speed, and direction crowds the city intersections. Roadside shops, markets, businesses, and public establishments pull many people close to the roadway intersection adding congestion. There have been studies to predict the noise levels in different parts of the city due to road traffic. However, determination of noise exposure levels, especially the health hazards (of noise pollution) for people working close to road intersections are generally not attempted. A typical underdeveloped city intersection in Bahadurpur, UP, India was chosen. A technique was established to determine the noise exposure levels relating to health hazards. The novel technique suggested a way to first characterize the noise levels at places in terms of noise levels. The characterized noise levels were then related to noise exposure levels. During exposure prediction, it used an approach practiced in industrial exposure calculation. Finally, the noise exposure maps were related to ground truths of health parameters for the dwellers working at the noisy intersection. The tinnitus, sleep disturbance, annoyance, hypertension, cardiovascular ailment, etc., near the noisy Bahadurpur intersection were found to be high. A total of 79% of the dwellers (of <40 Age group) working close to the noisy Bahadurpur intersection were found to experience severe health hazards, which increased to 95% with the elderly population. The statistics decreases to 13% (of <40 age group) and 30% (of >40 age group) for the less noisy Pure Ganga study area. Similarly, the hearing impairments were also found to be high among the dwellers of the noisy intersection area. For example, over 12% of the responders in the noisy Bahadurpur intersection indicated cases of tinnitus, which become 8% for the quieter Pure Ganga study area. The health matrix of the noisy intersection area was compared to the same as that of the non-noisy area and was found to be very significantly less harmful. The statistical comparison with ANOVA confirmed that the noisy and non-noisy areas are not equally impacted. Further, the high health hazards observed amongst people of the noisy area was primarily due to high exposure to traffic noise.
The noise maps and noise exposure level maps indicated that pronounced health hazards can be expected around the noisy road intersection. The vendors, sellers, hawkers, etc. working adjacent to noisy roads become the most vulnerable groups. The locations of these shops within 1–2 m from the road edge, their nature of structure (open or semi-open), the nature of congestion to the roads, and the nature of congregation of people to shops or roads increasing noise levels contribute to the health hazards.
The study showcased techniques to relate noise exposure to its likely health effects. It clearly indicates how to monitor the noise data and then use a prediction technique to develop a noise map for an area. It also showcased how the time-varying noise map is integrated with noise exposure duration to generate the exposure map for an area. The noise map and exposure map indicated vulnerable zones in the intersection where noise exposure levels can be alarming and mark the shops/people who are working or living there are going to experience severe health hazards. The study also indicated the types of diseases expected when people are exposed to high noise levels for a long duration.
The noise and health hazard mapping integrated noise and exposure levels over space and time. The use of a different sampling time for noise monitoring (i.e., long-term and short-term) and using them for noise prediction showcased that noise or exposure level maps of reasonable accuracy can be generated using lesser sampling time, or sampling points as well. Thus, noise hazard zonation mapping is possible with smart planning, and monitoring using lesser cost in a GIS environment.
The established technique not only generates the noise or exposure map for an area, but it also indicates the potential health hazards in an area and the nature of health deficiencies that are likely from noise exposure. Along with serious hearing impairments, people are likely to suffer more from cardiovascular diseases, hypertension, and psychological challenges such as anxiety, stress, sleep disturbance, etc. Elderly people exposed to intense noise levels for a long duration will have more severe damaging impacts.
The predicted noise maps/hazard maps have been able to demarcate the hazardous zone successfully. The risk or vulnerability reduces sharply with distance (e.g., 25 dB attenuation in 5 m). Thus, any idea for noise management can be planned for the dwellers, keeping their shops away from the hazardous corridor. Keeping the vendors away will reduce the crowds on the roads, which would enable a smoother flow of traffic, and less honking causing a reduction in noise. The research clearly indicated that the vendors of roadside shops at the Bahadurpur intersection are badly exposed to intense noise levels. The temporary shops at the edges of roads in the intersections are about 5 m from the road center. These shops can be shifted back further by at least 5 m, which will cause an attenuation in noise level by 25 dB as per the inverse square law used in the noise modeling (Equations (1) and (2)). Further, instead of setting up a temporary shop in open spaces, if the vendors are encouraged to construct their shops far away (from the intersection, which can be closed from all sides), the direct propagation of sounds from vehicles can be obstructed, which will very significantly attenuate the noise exposure levels. The Government’s understanding, planning, and regulation to set up separate places for commercial activities will safeguard the vendors or shopkeepers form exposure to high noise levels. Shifting the shops away from the intersection will also reduce the congestion at the intersection, traffic flow will be smoother, and honking will reduce significantly which will reduce the noise levels at the intersection very significantly.
The use of noise-absorbing materials for building the shops and nearby hutments can reduce noise exposure levels. Further, the use of absorbing noise barriers and guide walls adjacent to the fringes of the road intersections will stop vendors from occupying the roads. It will reduce the congestion and the levels of exposure to high noise levels for the vendors or shopkeepers. The measures of traffic control will improve the speed of the traffic, and will reduce the noise levels greatly (Schiavi et al. 2006).
The introduction of electric vehicles can also contribute to an improvement in the local acoustics. Noise levels from electrical vehicles can be up to 6 dB (A) quieter than those of conventional vehicles. However, only slower speeds have been proven to be useful (Grubesa and Suhanek 2021). The implementation of a “smart traffic management” policy with centralized control over traffic signals and sensors can manage traffic flow in accordance with the current condition of the roads in cities. This may lead to fewer instances of congestion and less honking overall, especially in developing nations. The authors plan to improve the noise-induced health hazard estimation technique. So that the technique can be applied over a large area and efficiently predict the likely health hazards at several intersections or around hotspots. Further, the authors plan to predict the impact of various mitigating measures, in terms of deciding the optimum location for the relocation of vendors’ shops, design of barrier and traffic guide, use of E-vehicles, and the smart traffic management plans that will reduce the noise pollution and the noise-induced health hazards.