International Journal of Engineering and Science Applications

ISSN: 2406-9833 @2015 PPs-UNHAS

IJEScA

Factors Affecting Traffic Noise

Based On Road Environment Aspects
Irwan Lakawa1*, Lawalenna Samang2, Mary Selintung2, and Muralia Hustim3
1
Doctoral Student at Civil Engineering Dept., Hasanuddin University, Indonesia
2
Professor, Civil Engineering Dept., Hasanuddin University, Indonesia
3
Lecturer, Civil Engineering Dept., Hasanuddin University, Indonesia
*
email: ironelakawa@gmail.com

ABSTRACT
The purpose of this study is to analyze factors affecting traffic noise on arterial and collector roads
in Kendari. The factors are road environment aspects, including building density, tree density, leaf
lush volume, types of road environment ground surface, as well as temperature. Analytical
approach with simple and multiple regression methods was used with SPSS software ver. 22.0. The
results show that based on the determination coefficient, partially, tree density has the highest
effect by 19%, then leaf lush volume 9.7%, temperature 9.1%, building density 8.7%, and types of
road environment ground surface 5.8%. However in this study, the factors significantly affecting
the noise are only tree density, leaf lush volume, and temperature. It is based on the value of sig-
< 0.05. Simultaneously, the variability of noise levels can be explained by the variables of tree
density, leaf lush volume, and the air temperature by 31.1%, while 68.9% is explained by other
factors not examined in this study
Keywords: factors; effects; road environment; noise
Article history: Received 20 September 2015, last received in revised 11 Ocktober 2015

1. INTRODUCTION order to reduce traffic noise are performed.

TRAFFIC noise is a problem that To provide a comprehensive solution, it is
cannot be avoided due to interacting first necessary to understand the factors that
transportation system. The problem currently can affect the level of noise.
happens in big cities and medium-sized cities Several studies have examined the
around the world, including Indonesia. At factors affecting the level of noise from the
certain intensity, the traffic noise can be bad traffic aspects, among others; prediction and
for humans because it can disrupt the health simulation models of traffic noise based on
and comfort of people in the surrounding groups of vehicles divided into 8 groups: bus,
settlements and road users. The disruption can car, double cabins taxi, jeeps, trucks,
be psychological disorder that tends to lead to motorcycles, three-wheeled vehicles, and
stress and hearing loss. vans. However, this model is only used for a
Therefore, in any activities in the field steady flow of traffic condition with the
of traffic and road transport, it should be assumption that the movement of traffic is
performed prevention and control of free without any obstruction such as zebra
environmental pollution to comply with crossing or traffic light [2]. Traffic noise
environmental quality standards [1]. The prediction model can be used effectively
question now is how the control measures in under the conditions of the cities in Iran. The

IJEScA vol. 2, 2, November 2015 117

International Journal of Engineering and Science Applications
ISSN: 2406-9833 @2015 PPs-UNHAS
IJEScA

variables used include the total volume of environment parameters, weather parameters,
vehicles, speed, percentage of truck vehicles, and residence parameters [8]. The study of
length and width of road, Gradients, and factors affecting the noise in a region is not
observation distance. The measuring distance only seen from the aspects of traffic, but it is
of noise from the road side is 3m [3]. also required a study of road environment
Simple and multiple regression model aspect. Therefore, later it can be used as
of local traffic effects on noise levels for information in the handling of noise either in
arterial roads in Bali with the locus of Ngurah source or propagation path of the noise.
Rai By-Pass and Sunset Street used 2. MATERIALS AND METHODS
independent variable of traffic volume (MC, A. Research Location
LV, HV), speed (MC, LV, HV) , and distance The research was conducted in
from observation point to the nearest and Kendari, Southeast Sulawesi (Indonesia) with
furthest center line [4]. From the results of the locus of arterial and collector roads
another study, effects of the green system on becoming a major center for the movement
the noise level in three landscape conditions: of traffic including public transport
tree plants, shrubs, and trees-shrubs distribution channels. The selection of
combination, a combination of trees and Kendari as the research location because this
shrubs with average leaf lush of 75% evenly city was one of the medium-sized cities with
distributed from ground level to a height of 5 the characteristics of heterogeneous traffic, in
meters can reduce the noise by 8.2 dBA, or which the growth rate of motor vehicle was
about 25% [5]. The composition of vegetation quite high, at an average of 19% per year.
consisting of trees, shrubs and bushes can
reduce noise up to 12.25% and decrease the
temperature to 8.18% [6].
Noise characteristics of a road should
be different with the characteristics of noise
in another way. This difference occurs
because the traffic noise level is determined
by many factors, such as motor vehicles
passing through (number of vehicles, types of
vehicles, average speed of vehicles),
characteristics of the road, and other
conditions in the surrounding road [7]. Traffic Figure 1. Location and Point of Research
noise is influenced by a number of parameters Sample
that do not depend on the vehicle itself, such
as; traffic parameters, road parameters,

IJEScA vol. 2, 2, November 2015 118

International Journal of Engineering and Science Applications
ISSN: 2406-9833 @2015 PPs-UNHAS
IJEScA

B. Data Collection performed on each observation segment

The number of observations sample along 25m. Percentage of building density
were 49 points on arterial and collector roads. was calculated by the equation:
Criteria for determining the sample point
used incidental sampling technique by
Dbg = x 100%
considering the function of road and
(2)
environmental conditions of the road.
Surveys of noise and temperature were
Where, Dbg is building density (%), Ldi is
conducted for 10 minutes at each observation
width of buildings walls (m), Ps is
sample point. SLM microphone was placed
observation segment length (m). Tree density
at a distance of 6m from the edge of
was calculated by the equation:
pavement. To obtain the value of equivalent
continuous noise level (Leq), it was calculated DPh = x 100%

by the following equation [7] [8]: (3)

Leq = L50 + 0,43 (L1 – L50) Where, Dph is tree density (%), Pki is width of
(1) leaf canopy (m), Ps is observation segment
length (m). Leaf volume calculation used the
Where, Leq is equivalent continuous noise equations in Table 1 [9].
(dBA), L50 is noise exceeded 50%, and L1 is
noise exceeded 1%.
Table 1. Leaf Lush Volume Based on
Canopy Shape

Figure 2. SLM Placement Scheme

Types of environmental surfaces Leaf volume in Table 1 applied to the

between road side and buildings (yard) were leaf lush percentage of 100%. If less than
identified according to the types of surface 100%, then the leaf volume was multiplied
namely; (1) grass, (2) land, (3) paving-block, by the percent value of the leaf lush in Figure
(4) concrete rebates. The identification was 3.

IJEScA vol. 2, 2, November 2015 119

International Journal of Engineering and Science Applications
ISSN: 2406-9833 @2015 PPs-UNHAS
IJEScA

value, it turns out that the building density

and type of ground surface road environment
are not significant because p-value is > 0.05.
Most relationship between predictor variable
and response are non-linear. Multiple-
regression model used several predictor
variables and one response variable. The goal
Figure 3. Estimated Leaf Lush Percentage
was to determine the effects of predictor
variable simultaneously. Based on the output
3. RESULT AND DISCUSSION
of Table 2, the variables used were the
Variables becoming the predictor
variable having significant value (p-value) <
consisted of; building density (Dbg), trees
0.05, they were; tree density (Dph), leaf lush
density (Dph), leaf lush volume (Vda), types of
volume (Vda), and temperature (Su).
road environment ground surface (Pbu),
temperature (Su). These were independent Table 3. Multiple Regression Model
variables constituting component forming the
model. The response variable was the noise
level (NL). Analytical approach in
determining the factors affecting noise used
statistical method of simple and multiple-
Table 3 showed the alternative models
regression with SPSS ver. 22.0.
of multiple regression analysis.Therefore, the
Simple regression analysis is a method
reflection effect of the building wall is almost
to determine the partial effects of each
non-existent, similarly in the type of ground
predictor variable on the response variable.
surface, on which the number of observation
The effects of the variables were seen from
samples that do not have obstacles such as
the probability of significance and coefficient
trees are only 18% of the total sample. It is
of determination (R²). The confidence level
possible that this is the cause of the
used was 95%, meaning the significance
relationship between ground surface type and
value (p-value) less than or equal to 5%, and
noise not significant.
had a coefficient close to 1 (R² ≈ 1).
The analysis results in Table 2 show Table 2. Model and Effects of Partial
Variables
that the tree density has the highest effect,
that is by 19%, then leaf lush volume 9.7%,
temperature 9.1%, building density 8.7%, and
type of ground surface road environment 5.8
%. However, if seen from the significance

IJEScA vol. 2, 2, November 2015 120

International Journal of Engineering and Science Applications
ISSN: 2406-9833 @2015 PPs-UNHAS
IJEScA

Based on the examination of statistical assumption tests are, among others:

parameters in Table 3, the selected model is normality, heteroscedasticity,
number 2 because it meets all the test multicollinearity, and autocorrelation.
parameters, namely; Fcount 3.15 > Ftable 2.81, In Table 4 it can be seen that all the
standard error of the estimate 0.00 < standard variables have the value of asymp. sig-KS >
deviation of 0.14, sig (p-value) 0.023 < 0.05, 0.05, meaning that the data are relatively
and determination coefficient of 31.1%. equal to the average, so it can be said that the
Although the building density and the type of residual meets the normality assumption.
road environment ground surface does not Detection of the presence of multi-collinearity
have significant effect in this study, based on symptoms can be seen from VIF value. It is
the study of theory, the two variables affect said there are no multi-collinearity symptoms
the noise level. This is possible because in if VIF < 10. In Table 5, the VIF value of Dph,
this study, the distance between front building Vda, and Su variables are smaller than 10,
wall and sound level meter (SLM) varies, of meaning that there are no multicollinearity
which approximately 70% have the distance symptoms. Independency test aims at
between 5-20 m. detecting the presence of autocorrelation
Therefore, the reflection effect of the symptoms by seeing the value of Durbin-
building wall is almost non-existent, similarly Watson statistics.
in the type of ground surface, on which the The requirement: du < DWcount < 4-du
number of observation samples that do not (du: upper limit of Durbin-Watson). The
have obstacles such as trees are only 18% of value of DWcount = 1.849 (output model
the total sample. It is possible that this is the summary). The value is in the required range,
cause of the relationship between ground so there are no autocorrelation symptoms.
surface type and noise not significant. Heteroscedasticity symptoms can be
Good or bad multiple linear regression determined by scatter plot diagram between
model has requirements that must be met, that NL predictive variable (Fits) and residual
is classical assumption. The classical variable.

Table 4. One-Sample Kolmogorov-Smirnov Test

IJEScA vol. 2, 2, November 2015 121

International Journal of Engineering and Science Applications
ISSN: 2406-9833 @2015 PPs-UNHAS
IJEScA

From Figure 4, it can be concluded that ground surface 5.8 %. However in this study,
there is no heteroscedasticity symptoms the factors significantly affecting the noise are
because the plots are evenly spread above only tree density, leaf lush volume, and
and below axis 0 without forming a specific temperature. It is based on the value of sig-α
pattern. < 0.05.
Tree obstruction is a factor that makes
it difficult to identify in detail the effects of
ground surface types on noise, because the
sound of vehicle received by SLM has been
absorbed by the trees. Therefore, to analyze
the effects of ground surface, it needs to
perform measurement in a place without
obstacles. Likewise, of the effects of building
density, the placement of SLM should be in
the range of 1-5m reflective distance from
building wall and without obstruction from
the source of noise.
Figure 4. Scatter Plot Diagram
5. REFERENCES
[1] Undang-Undang Republik Indonesia
Table 5. Coefficients
Nomor 22 Tahun 2009 Tentang Lalu
Lintas dan Angkutan jalan.
[2] Sooriyaarachchi, R.T., & Sonnadara,
D.U.J., (2008). “Modelling Free Flowing
Vehicular Traffic Noise”. Engineer.
Volume. 40(2), pp: 43-47.
[3] Golmohammadi, R., Abbaspour, M.,
Nassiri, P., Mahjub, H., (2009). A
4. CONCLUSIONS
compact model for predicting road traffic
Simultaneously, the variability of noise noise. Journal Environ Health Sci. Eng,
level can be explained by the variables of tree Vol. 6(3), pp: 181-186.
[4] Wedagama, D.M.P. (2012). “The
density, leaf lush volume and temperature by
Influence of Local Traffic on Noise
31.1%, while 68.9% is explained by other Level (Case Studi: By Pass Ngurah Rai
factors not examined in this study. and Sunset Road, Bali)”. Bumi Lestari
Based on the determination coefficient, Journal of Environment. ISSN 1411-
9668. Vol. 12(1), pp: 24-31.
partially the tree density has the highest [5] Umiati, S., (2011). “Pengaruh Tata Hijau
effects by 19%, then leaf lush volume by Terhadap Tingkat Kebisingan Pada
9.7%, temperature by 9.1%, building density Perumahan Jalan Ratulangi Makassar”.
Jurnal Teknika, Vol. 2, pp: 12-19.
by 8.7%, and type of road environment

IJEScA vol. 2, 2, November 2015 122

International Journal of Engineering and Science Applications
ISSN: 2406-9833 @2015 PPs-UNHAS
IJEScA

[6] Pudjowati, U.R., Yanuwiadi, B., [8] Handayani, D., (2007). “Pengkajian
Sulistiono, R., Suyadi., (2013). “Effect of Faktor - Faktor Tingkat Kebisingan Jalan
Vegetation Composition on Noise and Perkotaan”. Jurnal Puslitbang Jalan dan
Temperature in Waru-Sidoarjo Highway, Jembatan. Vol. 24(2).
East Java, Indonesia”. International [9] Balitbang PU., (2005). Pedoman Mitigasi
Journal of Conservation Science. ISSN dampak kebisingan akibat lalu lintas
2067-533X. Vol. 4(4), pp: 459-466. jalan, Pd T-16-2005-B. Jakarta:
[7] Mediastika, C.E., (2005). Akustika Departemen Pekerjaan Umum Republik
Bangunan: Prinsip‐Prinsip dan Indonesia.
Penerapannya di Indonesia, Erlangga,
Jakarta.

IJEScA vol. 2, 2, November 2015 123

International Journal of Engineering and Science Applications
ISSN: 2406-9833 @2015 PPs-UNHAS
IJEScA

IJEScA vol. 2, 2, November 2015 124