Impact Assessment of Traffic Noise in a Densely Populated Industrial City, Faisalabad Pakistan Using Geostatistical Approach and Development of Sustainable Transportation System Framework

ABSTRACT Haphazard and noisy traffic in congested and densely populated old cities has become a challenge for the environmental and transportation engineers to provide better means of transportation with raised quality of life and environment. Faisalabad is an old and major industrial city of Pakistan, which has huge influx of population due to its industrial significance. It has been facing severe traffic issues, which result in a noisy and irritating environment all around. This research work was planned to find out the health impact of noisy traffic on the residents and their social life, determining noise level variation trends with the changing density of traffic at different time intervals in a day and how variations occur during different weekdays. This study was focused to enquire the severity of noise levels due to road traffic with reference to significant selected locations throughout the city. It was aimed to develop a framework for a sustainable transportation system for Faisalabad city to cope with severe health impacts. Peak and average values were measured at significant locations, and GIS maps were also drawn spatially to demonstrate clusters of dense traffic at specific time. Furthermore, GS+ and Geostatistical module of Arc GIS 9.3 were used for evaluating sound level data. First, the data were analyzed in GS+, after doing calculation, the best possible model with the highest coefficient of determination (R2) and the minimum error of estimation (RSS) was selected, and the performance of each model was tested using cross-validation and the model which gives best result was used. Spatial prediction maps of noise level data for all the four towns of Faisalabad city were prepared by using the kriging method. Alarmingly high noise levels were observed during peak hours at specific locations. Such findings lead to the development of a framework for a sustainable transportation system to reduce the negative impact on the residents.


Introduction
Noise pollution is a machine-generated environmental threat that disturbs the activities and balance of human life.The noise being produced by the other sources may become hazardous due to its unacceptable impacts (Pettersson, 1997).Noise is a phenomenon that has physiological and psychological effects on human.It is a significant environmental problem in many urban areas.There are different kinds of noise sources.Specifically, traffic noise exerts an intense impact on residents nearby densely populated areas.Noise pollution is increasing speedily due to immense traffic load on the roads, lack of awareness, unconsciousness of vehicles impairment, and transportation.High noise levels can conduce to cardiovascular effects in humans and increasing effects of coronary artery disease.Attention on noise pollution has not been paid up to such a level as other environmental problems have gained.Faisalabad is the third most populated city of Pakistan, housing various types of industries.Population of this city is increasing progressively.Demands of vehicles for traveling requirements are also increasing day by day due to the increasing population of Faisalabad city, and thus, traffic density is increasing linearly with population density.According to a previously conducted survey, the hearing capability of the people living in this environment of such a huge traffic density is at risk (Ali & Tamura, 2003).
Noise exposure is increasingly being seen as an important environmental public health issue (Clark & Stansfeld, 2007).Persistent exposure to noise is injurious to health (Boateng & Amedofu, 2004).The damage caused by noise pollution is related to the intensity of the sound or the amount of energy it has (Bashorun & Olamiju, 2013).
Environmental noise can cause tinnitus, hearing loss, sleep disturbance, and other harmful effects on health (Mead, 2007).Exposure to noise disturbs sleep proportional to the amount of noise experienced in terms of an increased rate of changes in sleep stages and in the number of awakenings (Olokooba et al., 2010).Sleep disturbances such as difficulty to fall asleep, alterations of sleep pattern or depth, and awakenings may necessitate the need for using sleeping pills or ear plugs (Olaosun et al., 2009).
Noise can cause annoyance and aggression (Mead, 2007).Studies on traffic and aircraft noise have shown that the noise level has been associated with annoyance in a close-response relationship and both fear of the noise source and sensitivity to noise were found to strongly affect the annoyance from noise (Miedema & Vos, 1999).The degree of interference that noise causes in everyday activities precedes or leads to annoyance (Stansford & Matheson, 2003).
The non-auditory effects of noise on humans are viewed as being generally stress-related, following observations that noise exposures engender physiological reactions typical to those of stress (Ouis, 2001).Noise seems to have a negative effect on performance.It appears that the longer the exposure, the greater the effect.Children from noisy areas have been found to have heightened sympathetic arousal indicated by increased levels of stress-related hormones (Goines & Hagler, 2007).High-level noise will result in hearing loss (Donatelle, 2002).Evidence abound that constant noise exposure can damage sensitive structures in the ear.Noise-induced hearing loss, the second to ageinduced hearing loss, results from damage of the hair cells of the cochlea in the inner ear arising from continuous exposure to recreational and occupational noise (Rabinowitz, 2000).Blasts and other intense or explosive sounds can rupture the eardrum or cause immediate damage to the structures of the middle and inner ear, while hearing loss due to prolonged noise exposure is generally associated with destruction of the hair cells of the inner ear (Olaosun et al., 2009).
Unfamiliarity of vehicle impairment, industrial development, poor urban planning, conveyance, and construction activities are main factors that cause noise pollution in urban zones.Hearing loss, pressure of circulating blood, increased hostility headaches, migraine, higher fat levels, irritability insomnia, gastric ulcers, and psychological perturb are the important health risks initiated from pollution of noise (Haines et al., 2005).Another study was conducted to check the variation in traffic equivalent noise levels as the distance from the road connection increases (Abo-Qudais & Alhiary, 2004).Even the people in the zone of Katchehry Bazaar (Faisalabad), which is the heart of the town and center of commercial activities, were constantly exposed to an average noise level of 93.5 dB for about 10 to 12 hours daily (Hussain et al., 2008).
In Pakistan, due to unawareness about traffic rules, lack of sophisticated transport facilities, and adverse road conditions, most of the major cities are facing severe traffic problems.It leads to road accidents and increased noise levels.Such adverse situations result in mental stress, fatigue, and anxiety, which may develop a chronic impact on the health of residents who come across this condition for longer durations.This research work was conducted in Faisalabad, the third highly populated city of Pakistan and a wellestablished industrial hub.There are severe problems faced on a daily basis regarding poor transportation systems and intensely high noise levels on roads.It is resulting in an ill impact on the health of the residents, which are suffering from the diseases that are associated with mental stress and heart-related issues.This research was aimed to work out spatial distribution of different noise level generation zones under different conditions of time and traffic load; furthermore, it was also planned to develop a framework for traffic management to reduce the impact of pollution on the residents of Faisalabad city.

Materials and methods
This research was conducted in Faisalabad city, which is situated about 135 kilometers towards west from provincial capital Lahore with a latitude of 31.421045°N, a longitude of 73.089294° E, and an altitude of 184 m above the mean sea level.

Site selection
Site selection in the study area is based on the estimation that each selected site should be the illustration of the spatial distribution of noise in Faisalabad (Figure 1).Faisalabad is the third largest city having a population of 3,029,547, of which almost 42% people live in city.The noise level will be measured at significant sites including Main gate of University of Agriculture Faisalabad, (UAF), Jinnah Colony, Clock Tower, Kohinoor Plaza, Bus Station, Allied Hospital, GTS Chowk, Jhaal shopping road, GC University, and DHQ hospital, Nishatabad and Ghulam Muhammad Abad at selected time intervals (Figure 2).The experimentation and data collection were conducted from February 2018 to June 2018 in different segments, after dividing the study area in different zones as described earlier.The research plan has been explained in detail in the succeeding section.

Data description
The data related to noise pollution were taken using a Sound Level Meter, and a GPS meter was used to find the latitude and longitude of different sampling locations.The data obtained were further processed and analyzed by using GIS for making maps in Arc GIS 9 (version 9.3) and using the Kriging technique, a salient feature of Arc GIS.

Experimental Methodology
Sound levels were measured at three (3) peak hours (8:00-9:00 am, 1:00-2:00 pm, and 7:00-8:00 pm) in a day.Sound levels were noted at every five (5)-minute interval during every peak hour.This measurement was performed with the sound level meter keeping at arm's length.The sound meter had also been calibrated before taking readings at every site.The peak and average values were taken for GIS mapping.A health survey questionnaire was developed so that the impact of noise on human health could be concluded.

Geostatistical approach
After collection of data, normalization of data was done.Kriging was found to be the best available technique for interpolation and normal distribution of data.The Kriging method was used in this study for spatial variation analysis.

Exploratory data analysis
Exploratory data analysis was executed to check and explore data uniformity and consistency and remove outliers, identifying statistical distribution.The normal Quantile-Quantile plot (Q-Q plots) and histogram were plotted to check the normality of the observed data.Histogram and Q-Q Plot analyses were executed for the water table depth and each water quality parameter, and it was found that all the analyzed parameters showed mostly a normal distribution by calculating the median, mean standard deviation (SD), skewness, and kurtosis for each sample.

Traffic noise level studies at critical sampling locations in the study area
Table 1 shows the statistical data analysis of average noise levels at 12 sampling sites., while Table 2 is about the cross validation of characteristic parameters using variogram models.Similarly, Table 3 is about statistical data analysis of peak noise levels at the same 12 sampling sites and Table 4 explains the cross validation of characteristic parameters using variogram model

Average noise level data (8:00-9:00 am)
Figure 3 shows the histogram transformation technique applied to normalize data.During normalization of data, skewness, kurtosis, and standard deviation (SD) are calculated.All the values of these parameters showed that the sound level data are normalized.The histogram shows that the maximum average value of sound level peak hour 8-9 am for selected sites is 99.82 dB and the minimum value is 87.00 dB. Figure 4 demonstrates Normal Q-Q plot of average Noise level data.
For these sound level data, variograms are drawn.Figure 5 shows that the Gaussian model is the best fit model for these sound level data and demonstrates the spatial variation of data.
GIS mapping was performed to show the variation in the sound level data for all selected sites in Faisalabad city.Due to this, it is easy to assess the data variation for different timings and locations.The GIS map in Figure 6 demonstrates that the maximum value was at Bus Station, while Clock Tower demonstrates lesser sound levels during this peak  timing.It was because at the clock tower, most of the business activity started after 9:00 am, while bus stops were crowded due to office and school timings.

Average sound level data (1:00-2:00 pm)
Data were normalized by drawing histograms of sound levels during the peak hour.In Figure 7, the values of the parameters have shown that there existed consistency in data.By checking uniformity of data, variograms were drawn as shown in Figure 8, while Figure 9 shows that the Gaussian model is the model that fulfilled the conditions of the statistical variations.The mean error, root mean square error, ASE, MSE, and RMSE were calculated for this model.
During this time interval from 1:00 pm to 2:00 PM, from Figure 10, it has been observed that the peak hour maximum sound level was 99.82 dB at Allied Hospital and the minimum was 87.01 dB at the Clock Tower.

Average sound level data (7:00-8:00 pm)
Figure 11 demonstrates that the histogram transformation technique is applied to normalize data.During normalization of data, skewness, kurtosis, and standard deviation (SD) are calculated.All the values of these parameters show that the sound level data are normalized.The histogram (Figure 11) also shows that the maximum average value of sound level peak hour 7-8 pm for selected sites is 101.6 dB and the minimum value is 85.03 dB.
By checking uniformity of data, variograms were drawn as shown in Figure 12, while Figure 13 shows that the spherical model is the model that fulfills the conditions of the statistical variations.The mean error, root mean square error, ASE, MSE, and RMSE were calculated for this model.
GIS mapping was performed to show the variation in the sound level data for all selected sites in Faisalabad city.Due to this, it is easy to assess the data.The map shows that the maximum value was at Bus Station, while GC University has a less sound level during this timing.The GIS map of sound level variation at different sampling points during this time interval is shown in Figure 14.Maximum levels of sound could be observed at the Koh e noor commercial area due to peak hours of commercial activity.Similarly, high values have also been observed in the Allied hospital area.

Peak sound level data (8:00-9:00 pm)
Data were normalized by drawing histograms of sound levels during the peak hour.In Figure 15, the values of the parameters showed that data have been consistent.
By checking the uniformity of data, variograms were drawn (Figure 16). Figure 17 shows that the Gaussian model is the model that fulfills the conditions of the statistical variations.The mean error, root mean square error, ASE, MSE, and RMSE were calculated for this model.
The GIS map for peak values for all the sampling locations has shown that the maximum sound level was 113.9 dB at DHQ Hospital and the minimum was 88.90 dB at Clock Tower (Figure 18).

Peak sound level data (1:00-2:00 am)
Figure 19 shows that the histogram transformation technique is applied to normalize data.During normalization of data, skewness, kurtosis, and standard deviation (SD) have been calculated.All the values of these parameters showed that the sound level data are normalized.The histogram showed that the maximum average value of sound level at peak hour 1-2 am for selected sites is 109.4dB and the minimum value is 95.2 dB.
For this sound level, data variograms are drawn as shown in Figure 20, while Figure 21 shows that the Gaussian model is the best fit model for these sound level data and exhibited spatial variation of data.GIS mapping (Figure 22) was performed to show the variation in the sound level data for all selected sites in Faisalabad city.Due to this, it is easy to assess the data.The map showed that the maximum value was at Allied Hospital, while Clock Tower has a less sound level during this timing.3.1.6.Peak sound level data (7:00-8:00 pm) Data were normalized by drawing histograms of sound levels during the peak hour.In Figure 23, the values of the parameters showed that data have consistency.
By checking uniformity of data, variograms were drawn (Figure 24). Figure 25 shows that the Spherical model is the model that fulfills the conditions of the statistical variations.The mean error, root mean square error, ASE, MSE, and RMSE were calculated for this model.
It is obvious from GIS mapping shown in Figure 26 that during this peak hour, the maximum sound level was 108.8 dB at Bus Station and the minimum was 91.7 dB at GC University.These

Sustainable transport management framework
Ultimate goals and policy guideline of the Sustainable Transport Management Framework for Faisalabad city, having a population of about 3.0 Million at the time of study, were defined as follows; • Traffic volume reduction through a sustainable approach • Reduction of the negative impact on the environment and human health Based on the findings of the research work conducted in the study area and major observations identified, the framework for sustainable transport management was developed, considering the predefined policy and goals.Salient features of this framework have been explained in successive paragraphs, which once implemented, can have a significant impact on the environment, human health, and quality of living of the city.

Traffic control and road safety and maintenance
Traffic signals and strict observation of traffic rules are the most powerful devices for traffic control.Traffic signals must be installed at every important crossing, and it should be followed strictly.Using latest artificial  intelligence, the duration of traffic signals for peak hours should be optimized to maintain the consistency in the traffic flow and avoid the undue delay in shifting the traffic at the signals.Considering the onfield traffic control requirements, lane rules should be strictly implemented, particularly during rush hours.Fixation of road studs near crossings and diversions would facilitate its implementation.Traffic rules awareness among the drivers of all categories of vehicles must be ensured, and there should be identification of traffic signs, which must be displayed at significant places.Initially, for the purpose of awareness of drivers, unaware of traffic rules, help and assistance from well-trained college students could be acquired.Media campaigns should be initiated to influence public about the road safety and traffic noise pollution.Penalty should be enforced to those who break the traffic rules, and by adopting digital and automatic means for controlling the traffic, strict surveillance could be ensured.Modes of transport should be limited during specific hours, allowing only certain vehicles to move through crowded vicinities and important areas, like hospitals and schools, for example, animal driven carts, trucks, and intercity buses.
Improvement of quality of roads to facilitate smooth flow of traffic causes lesser noise and lesser risk of accidents.There should be a standard green belt separating the two-way traffic.This green belt must work as a buffer zone to reduce the impact of noise and at the same time, reduce the intensity of high beam lights of traffic coming from other sides.It will help in reducing the negative psychological impacts on the drivers and will facilitate safe journey.Low-power motor bikes should be encouraged within city, and there should be separate lane marked and reserved only for light weight motor bikes and cycles and other such vehicles.

Adaptation of economical traffic routing for Faisalabad city by urban planning
Instead of clustering of infrastructure and facilities, uniform distribution and dispersal of resources should be planned.It can also be expressed as decentralization of urban resources.It would lead to parallel growth of the city in all directions, resulting in the reduction of need to travel across the urban locality to reach other corners of the city to access a particular commodity/facility.For a sustainable transport system, it is important to develop cities in such a manner to facilitate non-motorized transport management for the city.Only reliance on public transport will not suffice the need, but it is also required that the urban planning should be done to optimize the travelling demands of the residents in terms of land allocation to specific purpose including hospitals, schools, markets, and residential areas.It must be planned to reduce the load on major hospitals of the city, i.e., Allied Hospital on the western side.Instead of extending the scope of this hospital, it is proposed to plan a state-of-the-art hospital with advanced treatment facilities on the eastern side of the city as well.Similarly, high schools and other teaching institutions should be properly located to equally distribute the load at peak hours of transportation.Proper urban planning for economical traffic routing is more feasible for newly developed urban communities.But for older cities like Faisalabad, the best possible approach could be "Radial Wheel Concept."It is about connecting all important locations through "ring roads," with inner and outer circles.A Ring Road project is the need of the hour for Faisalabad city to interlink the alarmingly extending boundaries of the city.Particularly, due to the development of Industrial Zones on M-3 motorway, it is required to plan a ring road across the city to reduce the overcrowded mid-town traffic.

Urban mass transit (UMT) for Faisalabad city
Government must develop such policies to encourage public transport mode rather than individual transport.A mega urban mass transit project would be the most appropriate solution to reduce the traffic load on the roads.Frequency of urban mass transit can be optimized based on the mass transit influx requirements, which may vary time to time throughout the day.Having an urban mass transit would reduce the traffic density on the main roads by reducing the demand of traffic vehicles, particularly the private ones.The presence of more private vehicles on the roads results in unwanted space occupancy with lesser productivity.Due to wide expansion of Faisalabad city along Satiana Road, Jhang Road, Sargodha Road, Lahore Road, and Samandari Road, it is highly recommended to establish urban mass transit facility along all these roads, having interconnecting points at significant places.

Traffic sorting/uniformity in the vehicle size and category
A well-sorted transportation system would facilitate in ease with which the traffic will flow smoothly in a particular direction.But if there are a variety of vehicles of all sizes and categories, it would hinder the continuity in the traffic flow.Keeping in view the typical scenario of traffic accustomed in Faisalabad city, intrusion of all kinds of traffic vehicles caused traffic problems, particularly during peak rush hours.Therefore, it is important to limit the vehicle size and category to run on the roads, while other vehicles should be provided alternate service roads.If service roads are not possible to be developed, there should be clear separation through road studs and partitioning on the main roads to distinguish different kinds of vehicles.Three-wheel drives and animal-driven carts must be replaced with a more stable vehicle for luggage carriage.The mode of transports must be limited during specific hours, and only certain vehicles should be allowed to move through crowded vicinities and important areas, like hospitals and schools.The vehicles to be excluded from the main roads during peak hours include animal carts, three wheelers, intercity bus services, and load carrying trucks.
A concise framework for a sustainable traffic management system, ideally to be implemented in Faisalabad city, in particular, and, generally, for any other place with similar characteristics and traffic behaviour and issues is represented in Figure 27.Subjected to establishment, this framework in Faisalabad city would lead to a sustainable transport management system in the city and will reduce the air and noise pollution issues along with reducing the probability of accident occurrence due to haphazard transportation, thus ensuring maximum road softy measures.

Conclusion
This research work concludes that major spatial variations in noise levels were observed at DHQ hospital, Allied hospital, Main gate of UAF, Bus station, and GTS chowk during peak hours.During 8-9 am, the noise level at the main gate of UAF was maximum because timing of university starts from 8 am and there was a huge rush of bikes, cars and rickshaws.The maximum value of noise was 113 dB at DHQ hospital during 8-9 am.This means that there is huge rush of traffic at that time.During 8-9 am, at most of the other sites, the noise level was comparatively low due to less traffic noise, while from 1-2 pm and 7-8 pm, there is a great rush of individuals riding on bikes, cars, rickshaws, buses, and trucks.Overall results showed that Jinnah town and Lyallpur town had a great noise pollution during peak hours affecting respondents working or living at those places.The framework developed for a sustainable transport management system can lead to a tremendous shift in quality of life and health and environment conditions of residents of Faisalabad, subjected to the strict and time execution of it.This framework is truly based on real data, providing a rational solution to the existing problem of intense noise levels and improper and haphazard traffic situations prevailing in the city under investigation, which signifies the novelty of this research work.Findings of this research work are beneficial for concerned authorities of law making and implementation to adopt this framework genuinely for the uplift of urban infrastructure and raising the standard of living of the inhabitants of the city.

Figure 3 .
Figure 3. Histogram of average noise level data.

Figure 4 .
Figure 4. Normal Q-Q plot of average noise level data.

Figure 7 .
Figure 7. Histogram of average noise level data.

Figure 8 .
Figure 8. Normal Q-Q plot of average noise level data.

Figure 11 .
Figure 11.Histogram of average noise level data.

Figure 12 .
Figure 12.Normal Q-Q plot of average noise level data.

Figure 15 .
Figure 15.Histogram of peak noise level data.

Figure 16 .
Figure 16.Normal Q-Q plot of peak noise level data.

Figure 19 .
Figure 19.Histogram of peak noise level data.

Figure 20 .
Figure 20.Normal Q-Q plot of peak noise level data.

Figure 23 .
Figure 23.Histogram of peak noise level data.

Figure 24 .
Figure 24.Normal Q-Q plot of peak noise level data.

Figure 27 .
Figure27.Four roads to a sustainable transportation system in Faisalabad city.

Table 1 .
Statistical analysis of average noise levels at 12 sites.

Table 2 .
Cross-validation of L avg (characteristic parameters of variogram models).

Table 3 .
Statistical analysis of peak noise levels at 12 sampling sites.

Table 4 .
Cross-validation of L peak (characteristic parameters of variogram models).