Noise Mapp.
2017; 4:45–56
Research Article
Kuenda Laze*
Findings from measurements of noise levels in
indoor and outdoor environments in an expanding
urban area: a case of Tirana
DOI 10.1515/noise-2017-0003
Received Oct 10, 2016; accepted Mar 13, 2017
Abstract: The increasing human population and the in-
creasing number of vehicles in Tirana, Albania, emerges
the need for extensive measurements of noise levels. The
831 measurements of noise levels were taken in Tirana
in November 2015 for education purposes. Measurements
were collected in eight locations in indoor like a classroom
in a school building environment, a library and a health-
care facility and in outdoor environments like five cross-
roads. Noise levels were then compared with domestic and
World Health Organization (WHO) standards. The loga-
rithmic average of noise levels and the maximum value
of noise level measured for the period of measurement
(LAFmax) were interpolated using Interpolated Noise Lev-
els for Observer Points in ArcGIS producing noise level
maps for crossroads. The logarithmic average of noise lev-
els and LAFmax measurements were respectively above
domestic and WHO standards in 96% and 100% of out-
door environment and indoor environment locations. In-
terpolated values of logarithmic average of noise levels
and LAFmax for five crossroads were above domestic and
WHO standards indicating noise levels in road traffic could
remain high. A yearly acoustic measurement for vehicles
should be implemented. Participatory measurements of
noise levels in quite indoor environments can be used to
increase the awareness of inhabitants in Tirana.
Keywords: crossroads, dominant sounds, indoor environ-
ments, measurements, outdoor environments, prediction,
standards
*Corresponding Author: Kuenda Laze: Polytechnic University of
Tirana, Faculty of Civil Engineering, Department of Environmental
Engineering, Rr. “M. Gjollesha”, No. 54, 1023 Tirana, Albania; Email:
m1.laze@outlook.com
© 2017 K. Laze, published by De Gruyter Open.
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivs 3.0 License.
Open Access
1 Introduction
Noise is a global environmental problem [1] in cities [2].
Noise is originated from human activities like transport,
industry [3] and neighborhood areas like restaurants, bar-
café, parks, parking [4]. According to European Environ-
mental Agency (EEA) [3], noise levels increased due to in-
creasing traffic volume, and as result population exposed
to higher noise levels increased in Europe. There are ev-
idences that high noise levels above 55dBa risk human
health. People exposed to high noise levels are likely to
get cardio, hypertension illness [4], to have a low concen-
tration in task performing e.g. in school [5]. High levels of
noise could cause annoyance of population [6, 7] inhab-
iting urban areas. Indeed, high level noise causes annoy-
ance in European cities affecting the population’s percep-
tion of life quality [8].
Scientific and consultant organizations have dealt
with high noise issue [5]. Advanced methods for assess-
ment of common noise were prepared in Europe [9] to help
countries elaborate their action plans to reduce noise pol-
lution [9]. Today, the final aim of all methods, strategies,
mitigation measurement are to protect population against
high level noise i.e., above 55dBa [3, 10, 11] that are pro-
duced in restaurants, schools, by airplane and road traffic.
Road traffic noise is widely studied. Noise levels of
roads in urban places are evaluated, modelled [in urban
areas see e.g. 12, 13]. Yet, different existing models are used
in different countries. These models, however, require var-
ious input data, or uses different short and large refer-
ence times [see 14], as well as techniques. Besides noise
level prediction, effects of traffic composition are studied
on road noise [15]. Noise mapping system is developed to
detect and show in real time the sound impact caused by
roads [16]. Annual equivalent noise level under traffic ur-
ban condition is estimated using sampling as an technique
for its estimation [17]. Yet, results presented in [18] con-
stitute a relevant step forward in the iterative cycle of re-
search and design for developing tools to apply the con-
46 | K. Laze
cept of citizen observatories [19] and to empower citizens
in the assessment of acoustic comfort in public places.
Noise is an environmental problem in Albania. There
are domestic noise standards in Albania [20], which were
formulated in the year 2007, following strictly the stan-
dards of World Health Organization (WHO) [4]. Regular
measurements of noise have been collected in crossroads
for some urban areas including Tirana as part of an-
nual environmental reports in the year 2011 and 2014 as
well as there was a report dedicated to noise manage-
ment [21–23]. Measured noise levels were compared to do-
mestic, the WHO and European standards in these reports.
Yet, to author’s knowledge considering literature on noise
level measurements in Tirana and Albania, there were no
data on noise levels collected in indoor environments like
schools, library, healthcare facility from 1990 to 2015. Also,
there were no regular noise level measurements and no
monitoring activities conducted from any party for all ex-
isting and emerging industry, transport and other busi-
nesses in neighborhood areas from 2014 to 2015.
Noise levels were measured in the urban area of
Tirana in the year 2015 for education purposes. Domi-
nated sounds were observed in neighborhoods specifically
in five crossroads, in a library, a healthcare facility, and
in a classroom. This helped to better distinguish acous-
tic sound sources produced in selected measurement loca-
tions. All measurements were respectively compared to do-
mestic and WHO standards of 30dBa (healthcare facility),
35dBa (classroom) for indoor environments and 55dBa [5]
for outdoor environments.
The objectives were to 1- collect data on noise by mea-
suring noise levels using a device, 2- potentially use these
data for data analysis, 3- investigate the dominant sound
sources produced in eight locations, and 4- compare the
measurements with existing standards [24] for indoor and
outdoor environments.
There were selected five measurement locations in
five different crossroads. Crossroads were selected because
there were identified as “noise pollution hot spots” [23] by
governmental bodies that monitor noise levels in Tirana.
Indoor environments were one classroom in a school
building, a library and a healthcare facility.
Vehicles as cars, buses, motorcycles and human
sounds were respectively expected to be dominant sound
sources for outdoor and indoor environments. Data col-
lected were then used to identify the most explanatory
variable for noise levels. Ordinary least squares (OLS)
method was used to investigate the linear relationship be-
tween 1- the logarithmic average of noise level, or 2- maxi-
mum value of noise level measured for the period of mea-
surement; LAFmax (dependent variable) and explanatory
variables. The explanatory variables were the estimated
number of vehicles, the number of store of buildings and
the projected number of population. The explanatory vari-
able that explained most of high values of noise levels
at crossroads was identified. The variables of vehicle and
road type were used for calculating basic noise level for
roads following the approach of Farcaş [25] that applied
Nordic Noise Prediction method. The logarithmic average
of noise levels and LAFmax for crossroads were interpo-
lated using Interpolated Noise Levels for Observer Points
in Geographic Information System (GIS) [26] to estimate
noise levels in crossroads producing noise level maps.
2 Methodology
2.1 Criteria for measurements and location
selection
Indoor environments included one classroom, one li-
brary and one healthcare facility presenting indoor pub-
lic spaces in this study. Public spaces are not much in-
vestigated for their quality of acoustic environments like
libraries [see e.g. 27]. School building and classroom are
investigated for they are subject to noise pollution of the
surrounding area and building maintenance [5]. The class-
room, library and healthcare facility were considered be-
cause all teachers, students, readers, medical personnel
and patients need a quiet place to work and stay. Noise in
these indoor environments can be negatively affected by
the quality of inner construction, which cannot absorb ef-
ficiently noise, and by outdoor noise like road traffic. All
these noises affect negatively e.g. the student and teach-
ers performance [for more information see 5]. Dominant
sources were interesting to investigate in these indoor en-
vironments because there were no noise levels monitored
and reported by government bodies in Tirana.
Measurement locations in crossroads, which pre-
sented outdoor environments in this study, were selected
based on perception of students on noise and on noise
monitoring reports of environmental governmental bod-
ies [22, 23] in Tirana. These locations were selected in five
crossroads of the urban area of Tirana.
All measurement locations were placed along roads.
Roads had different number of road lanes and traffic com-
positions. Indoor environments were in less heavy traffic
roads (fewer roads; fewer road lanes) compared to outdoor
environments. All measurement locations in outdoor and
indoor environments are shown in map (b, c) in Figure 1.
 Measurements and dominant sounds in indoor and outdoor environments | 47

Figure 1: Map of Tirana and locations in indoor and outdoor environments: (a) location of Tirana and Albania, (b) the estimated LAFmax, (c)
the estimated logarithmic values of noise levels using Inverse Distance Weighted 3D (IDW3D). Indoor environment locations are indicated
by one, two and three in red color. Outdoor environment locations are numbered four, five, six, seven and eight in blue color. Roads are
provided by OpenStreetMap, for illustration.

All indoor environments were exposed to noise pro-
duced by road traffic. Characteristics of indoor environ-
ments (one library, one healthcare facility and one class-
room area) and outdoor environments (five crossroads) are
shown Table 1.
Parameters of vehicles flow rate capacity, number of
lanes, type of vehicles like busses, cars, taxis, motorcycles
were the same for five crossroads which were in the ring
of Tirana (Figure 1). High noise levels were measured by
environmental bodies in these crossroads. Noise level data
were presented in environmental monitoring reports in the
year 2011 and 2014 [22, 23].
2.2 Data acquisition and sampling
The Sonometer CR:306 provided readings of traditional en-
vironmental acoustic parameter of LAFmax and LAeq, t,
(A-weighted equivalent sound pressure levels). Acoustic
measurements were read in noise level (dBA). For calibra-
tion, this equipment had an equipment error of - 3dBa. The
device was held at the side of the road at a height of approx-
imately 1.3 m above the ground level.
The sound level meter was explained how to use to all
students. The 831 noise level measurements were collected
by fourteen students (henceforth “collectors”) of master
degree, profile of energy of the branch of Environmen-
tal Engineering at the Polytechnic University of Tirana in
eight locations of outdoor and indoor environments in two
days in November 2015. They practiced and learned the de-
vice to calibrate before taking noise level measurements.
Noise observations and measurements were explained (by
the author) in details to collectors.
Five groups were created. Each person participated in
data collection, sound source observations and was super-
vised (by the author). Each group was asked to spent a few
minutes to observe the place and write down about interior
design and human sounds in indoor environments and
road traffic composition flow in outdoor environments.
Then, each group received the device for measurements
of acoustic sounds. Time-history of the LAeq, T values of
the measurement was shown on the device screen. Each
group also identified the dominant sound sources. They
respectively identified acoustic sounds produced by vehi-
cles and humans as dominant sound sources in outdoor
and indoor environments. Considering types of noise sam-
pling and noise study objectives of Brown [28], measure-
ments were likely source-oriented sampling with objective
to assess (measure) the current level of noises in environ-
ments of Tirana’s urban area.
48 | K. Laze

Table 1: Characteristics of locations, noise sources and levels in dBA. Logarithmic average values of noise level, LAFmax, domestic stan-
dards and those of World Health Organization (WHO) in dBA and outdoor noise sources. Measurement locations numbered from one to
eight are in urban areas in Tirana. Photos for locations number one, two, seven and eight were taken in November 2016. Photos for loca-
tions number three, four, five and six were captured in December 2014. All photos were taken by the author.

Measurement Indoor/outdoor Photo of location Noise Domestic and Logarithmic LAFmax

location number environments sources WHO standard, average of noise
dBA level, dBA

1 Library Main 30 58.0 64.5

road

2 Healthcare Main 30 61.9 72.02

facility road

3 Classroom Main 35 73.2 78.6

road

4 Crossroad Main 55 71.3 69.9

roads

Continued on next page

 Measurements and dominant sounds in indoor and outdoor environments | 49

Table 1: ... continued

Measurement Indoor/outdoor Photo of location Noise Domestic and Logarithmic LAFmax

location number environments sources WHO standard, average of noise
dBA level, dBA

5 Crossroad Main 55 80.5 99

roads

6 Crossroad Main 55 75.5 88.8

roads

7 Crossroad Main 55 66.5 77.1

roads

8 Crossroad Main 55 61.0 77.7

roads

Concluded
50 | K. Laze
Every group of collectors (four groups of three persons
and one group of two persons) collected at least 100 val-
ues per location during the day (8:30 – 13:30) with a sam-
pling time of 1 second for each value readings (LAeq,1sec).
They wrote down observations about sound sources in the
event, the source of the highest values of noise level pro-
duced, location, interior design (indoor environments),
the highest values of noise level (LAeq,1sec) caused by
speed, acceleration and deceleration of vehicles at traf-
fic lights (in outdoor environments), the type of vehicles
(car, bus, motorcycle), the value of LAFmax for the mea-
surement period (which varied from 5 to 11-minute noise
level measurement for every group) as well as the stan-
dard value of noise level for indoor and outdoor environ-
mental location. Then, every group took a photo of the
measurement location and wrote down the start-time and
end-time of measurements, type of device used (Sonome-
ter CR:306), date and name of location. Geographical co-
ordinates for locations were derived from Google map. All
collectors were asked about their perception on noise level
and dominant source of noise produced in measurement
locations.
All values of measurements were written down in a
notebook. For each location, the logarithmic average of
noise level values of all measurements was calculated.
The values of measurement and their logarithmic average
value were compared with standards [4] of noise level al-
lowed in every indoor and outdoor environments. The log-
arithmic average of noise level measurements and LAFmax
were used in data analysis.
2.3 Variable selection
Independent variables were selected based on assumption
that these variables could explain noise levels measured in
five locations in crossroads. Crossroads were selected due
to their high noise levels and heavy traffic. These cross-
roads were reported as high noise level areas [23] by moni-
toring governmental bodies in Tirana [21, 29]. Accordingly,
vehicles could be the mayor causes of these high noise lev-
els in crossroads in Tirana.
Tirana is an expanding urban area since the collapse
of socialism, which occurred in the year 1990. The number
of vehicles and the number of population increased con-
siderably since 1990. Population migrated from other re-
gions towards Tirana. Population change in the urban area
of Tirana was between 12% and 60% from 1989 to 2011 be-
cause Tirana was a key industrial, administrative and cul-
tural center in Albania [30] (see also literature on migra-
tion of Albanians [31, 32]). The increase of vehicle numbers
could have contributed to the change in the traffic flow rate
and to noise patterns [see e.g. 33]. According to Zhu [14],
noise emitted from vehicles is a very important variable to
establish noise prediction models. The number of stores
of the individual building was selected as an explanatory
variable in OLS analysis because building constructions
could affect noise calculations [14].
Population census data of Tirana for the year 2011 and
neighborhood population was respectively published by
the Institute of Statistics [34] and by the Municipality of
Tirana [35], and were used in OLS analysis (see 2.4 Or-
dinary least squares (OLS) method). Population had in-
creased from 2001 to 2011 [30], thus it was assumed a pro-
portional increase for every neighborhood. There was no
population data for either crossroads or neighborhoods in
2015, thus, neighborhood populations were projected in
2011 for the year 2015 as an approximate population for
crossroads. Every crossroad belonged to a neighborhood
area, thus there were selected five crossroads for five dif-
ferent neighborhoods in Tirana.
Data of vehicles, which was published by the Insti-
tute of Statistics for Tirana (www.instat.gov.al), was used
to calculate the number of vehicles in five crossroads as-
suming the number of vehicles increased by the same rate
as the number of population increased in each crossroad.
For example, if a neighborhood population increased by
1.2 from 2011 to 2015, then the number of vehicles includ-
ing the number of cars also increased by 1.2 from 2011 to
2015 in this neighborhood. The number of vehicles was ex-
pected to increase from 2011 to 2015 because there were
evidences that the number of cars had continuously been
augmented from 2001 to 2013 [36]. The number of build-
ing stores at each crossroads was given by referring to per-
sonal observations. In fact, the number of building stores
changed for one crossroad since 2014.
2.4 Ordinary least squares (OLS) method
OLS analysis was carried out in R [37] including Pearson
test (non-correlated variables should have a Person coeffi-
cient r < 0.70). OLS was used to check the linear relation-
ship between dependent and explanatory variables [38].
Maximal value of noise level for the period of measure-
ments, LAFmax, was the dependent variable in a OLS data
analysis. Explanatory variables were the projected number
of inhabitants, the estimated number of vehicles and the
number of building stores along the crossroads.
Logarithmic average of noise levels, which were cal-
culated from all measurements for five crossroads, was
the dependent variable in data analysis. Independent vari-
 Measurements and dominant sounds in indoor and outdoor environments
ables to explain measured noise level were the projected
number of inhabitants, the estimated number of vehicles
and the number of building stores along the crossroads.
These explanatory variables were expected to show linear
relationships with the dependent variable of logarithmic
average of noise levels. The estimated number of vehicles
was assumed to explain most of the noise levels (either
for the logarithmic average of noise level or maximal noise
level values, LAFmax).
OLS assumed a linear relationship [see 39] between
e.g. the logarithmic average of noise level dependent vari-
able and the projected number of inhabitants, the esti-
mated number of vehicles, and the number of building
stores, following the Equation (1):
y = β0 + β1 x1 + β2 x2 + ε i (1)
where y is the dependent variable, β0 is the constant, β i
is the estimated coefficients of OLS, x is the explanatory
variable, i is the observation, and ε is the error.
Akaike Information Criterion (AIC) [40] and an ad-
justed R-squared were respectively calculated as a model
selection estimator and as a goodness-of-fit of the regres-
sion.
2.5 Prediction of noise level for crossroads
Prediction of noise level, simulation, and models are
widely used and developed in many countries [see e.g.
7, 14, 16]. A common framework (CNOSSOS-EU) has been
developed for member states of European Union [11], to as-
sess noise in urban areas.
Geographic information system (GIS) is a tool that can
be used for calculating noise predictions and for mapping
these predictions. GIS plays an important role in noise
mapping [41], while noise maps can be useful for investi-
gating noise in urban environments [42]. Geographic In-
formation System ArcGIS 10.3 [26] was used to estimate
noise levels in crossroads following the approach of “Tool
1 – Noise Calculation for Roads” [25] developed after the
Nordic Prediction Method. In this method, LAeq is calcu-
lated for each road (or part of a road), using Equation (2)
as follows:
(︃ n
)︃
∑︁ L /10
L Aeq = 10 lg 10 Aeqi
(2)
i=1
To apply this approach, vehicle number and road type
data were employed. Data were collected for dwelling,
well-kept and national asphalt roads. The estimated num-
ber of vehicles (as volume traffic) and speed allowed
| 51
for vehicles were calculated referring to literature, per-
sonal observations during data collection and to the ex-
isting legislation in Albania. A speed between 30 km/h
and 40 km/h, between 50 km/h and 60 km/h and from
70 km/h to 90 km/h was respectively used as speed lim-
its [25] for vehicles in dwelling roads, well-kept roads,
and national asphalt roads. These speed values were very
close to the highest speeds allowed for vehicles in urban
roads (40 km/h) and national roads (between 80 km/h and
90 km/h) in Albania [43]. There were no data available for
heavy vehicles, thus a priory a number 10% (personal ob-
servations) of total estimated number of vehicles was as-
sumed for heavy vehicles.
Noise levels were firstly calculated for dwelling, well-
kept and national road type. LAeq was a function of the
number of light and heavy vehicles and their speeds [25].
The logarithmic average values of noise level were respec-
tively calculated for dwelling roads, well-kept roads and
national asphalt roads. Then, a logarithmic average value
of noise level was calculated for all roads.
The estimated average value of noise levels and esti-
mated LAFmax were calculated using the algorithm of In-
terpolated Noise Levels for Observer Points (Inverse Dis-
tance Weighted 3D, IDW3D in ArcGIS) producing two maps
as follows: one map with estimated average values of noise
levels and one map with estimated LAFmax. IDW3D func-
tion uses a linear-weighted combination of observation
points (measurement locations; logarithmic average value
of noise level; LAFmax) and the weight is a function of the
distance of an observation point from any measurement
location around the observation point (see Figure 1) [44].
Then, a value (obtained by Extract values to points) of an
observation (crossroad) point was respectively calculated
using either a surface (a raster) of the logarithmic average
of noise level or a surface of LAFmax obtained by IDW3D.
3 Results
3.1 Noise levels, dominant sounds and
annotations
The 831 measurements and their logarithmic average and
LAFmax values of noise levels for indoor and outdoor en-
vironment were compared to domestic and the WHO stan-
dard [24] for indoor (30dBa and 35dBa) and outdoor envi-
ronments (55dBA). All noise level measurements obtained
in the classroom (in a break time), in the library and in the
waiting hall of the healthcare facility had higher sound
source levels than 30dBa and 35dBa. The largest differ-
52 | K. Laze

Table 2: Annotations and sound sources observed in indoor and outdoor environment locations. There were three noise level measurement
locations in indoor environments as follows: one location in a library, one location in a healthcare facility and one location in a classroom.
There were five noise level measurement locations in five crossroads, outdoor environments.

No. Annotations Sound source Indoor/outdoor environments

1 People talking, chatting people 67% of indoor environments
2 Road noise vehicles 100% of all environments
3 Chair dragging sitting people 33% of indoor environments
4 People speaking loudly people 33% of indoor environments
5 People footsteps people 67% of indoor environments
6 Vehicle (car, bus) horn drivers 100% of outdoor environments
7 Acceleration car drivers 60% of outdoor environments
8 Deceleration car drivers 80% of outdoor environments
9 Motorcycle engine 40% of outdoor environments
10 Buses engine 20% of outdoor environments

90,0 Measurements in outdoor environments i.e., in five

Logarithmic average value of noise levels in

80,5
80,0 75,5
73,2 71,3 crossroads, were above 55dBa, Table 1. Crossroads had
70,0 66,5

60,0
58,0
61,9
55 55 55 55
61,0
55
a wide range of noise levels measured from 48dBa to
50,0 99.3dBa, where the lowest and the highest noise levels
dBA

40,0
30 30
35 were respectively produced by vehicles stopped at a traffic
30,0
light and a car horn. Differences of noise level of 44dBa,
20,0

10,0
0,0
1 2
Logarithmic average value of noise level, dBA
Figure 2: Logarithmic average value of noise levels, domestic and
WHO standard for indoor and outdoor environments. Indoor envi-
ronment locations are one, two and three. Outdoor environment
locations are four, five, six, seven and eight.
ence between the logarithmic average value of noise level
and the standard was 38.2dBa obtained in the classroom
area (location 3 in Figure 1). Two other indoor locations
had a difference of 31.9dBa (location 2; healthcare facil-
ity) and 28dBa (location 1; library in Figure 1) indicat-
ing that indoor locations have the highest divergence val-
ues obtained from all measurement locations from their
WHO and domestic standard (Figure 2). The difference be-
tween LAFmax values (Table 1) and standards (35dBa and
30dBa) were respectively 43.6dBa, 42dBa and 34.5dBa for
the classroom, the healthcare facility and the library. The
most common annotation was road noise (as written down
by collectors) in three locations in indoor environments,
Table 2. People footsteps and people talking in halls were
common annotations in the library and healthcare facility
(locations 1 and 2). Chair dragging, people chatting, and
people speaking loudly during break time were common
in the classroom (location 3), Table 2.
30.7dBa and 28.9dBa with domestic and WHO standard
(55dBA) was caused by a car horn, a bus, a car deceleration
3 4 5 6 7 8
Locations in the fifth measurement location in Figure 1. A bus horn
Domestic and WHO standard, dBA
produced a high noise level of 84.1dBA in the crossroad lo-
cation (number 7 in Figure 1). A car acceleration at traffic
light had the highest noise level of 88.8dBa measured in
crossroad location (number 6 in Figure 1). Measurements
taken in the fifth measurement location had the highest
logarithmic average of noise level value of 80.5dBa (Fig-
ure 2) from all outdoor environments. The most common
annotations in five crossroads were car horn, acceleration
and deceleration of vehicles at traffic lights, as well as mo-
torcycles and busses that produced high sound level from
their engines, Table 2.
3.2 OLS results for crossroads
Correlation between the estimated number of cars and the
projected population and between the estimated number
of cars and the number of building stores was high (r >
0.70). Correlated variables were separated into two indi-
vidual multiple regression models for logarithmic average
and LAFmax noise level, respectively. Model performance
was calculated and compared in terms of model estima-
tor of AIC. Two OLS models for LAFmax obtained values of
AIC (model 1, 42.9 and model 2, 43.2; Table 3) indicating
both OLS models can be selected. Relationship between
 Measurements and dominant sounds in indoor and outdoor environments | 53

Table 3: Summary of OLS coeflcients, standard errors and p-values for the logarithmic average value of noise level of LAeq, and maximal
noise levels, LAFmax.

Logarithmic average of noise levels AIC Adjusted Coeflcients Standard errors p-value
R-squared
Model 1 40.5 −0.71
Projected number of inhabitants 0.268 0.526 0.660
Number of building stores 0.337 2.075 0.886
Intercept 42.705 61.693 0.560
Model 2 (best model) 38.3 −0.094
Estimated number of vehicles 0.295 0.365 0.477
Intercept 38.792 39.887 0.403
LAFmax AIC Adjusted Coeflcients Standard errors p-value
R-squared
Model 1
Projected number of inhabitants 42.9 −0.219 −0.321 0.667 0.678
Number of building stores −2.811 2.630 0.397
Intercept 132.730 78.201 0.232
Model 2 43.1 −0.27
Estimated number of vehicles 0.228 0.591 0.725
Intercept 57.721 64.570 0.437

LAFmax value of noise level and explanatory variables of
the projected human population, the estimated number
of vehicles, and the number of building stores was calcu-
lated. The OLS coefficients of the estimated number of ve-
hicles had a positive sign of coefficients, while projected
human population and the number of building stores ob-
tained negative sign coefficients. This indicated that the
estimated number of vehicles can contribute to maximum
values of noise levels measured for a period, if the number
of vehicles continued to increase, Table 3.
The model 1 and model 2 of the logarithmic average of
noise levels were respectively composed of two variables
of the estimated number of building stores and projected
population (AIC=40.5) and one variable of the estimated
number of vehicles (AIC=38.3). Model 2 was the best model
for logarithmic average of noise levels. The estimated num-
ber of vehicles had a positive relationship with the loga-
rithmic average of noise level indicating that an increased
number of estimated vehicles (or an increased traffic vol-
ume) could produce a higher noise level in crossroads.
Model-fit of R-squared was higher for two best OLS mod-
els of LAFmax (−27%, −21%) than for the best OLS model
2 of logarithmic average of noise level (−9%), Table 3.
3.3 Mapping of predicted noise level for
crossroads
Basic noise level for roads resulted respectively to 79.2dBa,
55dBa, 72.6dBa and 78.1dBa for all roads, dwelling roads,
well-kept roads, and national roads. Light vehicles con-
tributed to a noise level of 55dBA in dwelling roads. Heavy
vehicles contributed respectively to an increase of noise
levels by 12.3dBa and 9.2dBA compared to light vehicles
(cars) in well-kept and national roads, (Table 4).
The estimated maximum values of noise levels,
LAFmax, and the estimated logarithmic average of noise
levels (using IDW3D in ArcGIS 10.3) changed respectively
from 69.9dBa to 98.9dBa, and from 61.0dBa to 80.1dBa,
which were above the standard of 55dBa. The estimated
LAFmax values were very close to the measured values of
LAFmax changing from -0.1dBA to 0dBA, (Table 5). The
largest difference between the estimated logarithmic aver-
age and measured noise levels in crossroads was 0.1dBA
for location 5 (Table 5), indicating that IDW3D was a useful
tool in noise level estimations. The estimated logarithmic
average of noise level was above 61dBa in crossroads and
in all surrounding areas of crossroads in Figure 1.
54 | K. Laze

Table 4: Basic noise level for roads. Road types, vehicle speed in km/h, (the least) number of road lane, number of light and heavy vehicles.
Time was 11 minutes. No heavy vehicles were assumed for dwelling roads.

Estimated basic noise Road type Vehicle Number of Number of light Number of heavy vehi-
level for roads, dBA speed, km/h road lane vehicles cles
55 Dwelling 30-40 1 1 vehicle per second -
72 Well-kept 50-60 2 2 vehicles per second 1 vehicle per every five
seconds
78 National 70-90 3 5 vehicles per second 1 vehicle per every five
seconds

Table 5: Estimated noise levels in crossroads. Measurement location number is the same as in Table 2.

Measurement location Estimated average value of Logarithmic average value Estimated LAFmax,
number, crossroads noise levels, dBA of noise levels, dBA LAFmax, dBA dBA
4 71.31 71.30 69.90 69.9
5 80.59 80.49 98.99 99.0
6 75.54 75.49 88.80 88.8
7 66.51 66.50 77.10 77.1
8 61.02 61.00 77.70 77.7

noise levels. These estimations suggested that noise levels

4 Conclusions can be predicted using road types and number of heavy
and light vehicles, and IDW can be used to interpolate
Vehicles and human sounds were respectively dominated
noise level. Interpolated values of LAFmax and logarith-
sounds in outdoor and indoor environments. People talk-
mic average of noise level in Figure 1 showed that people
ing, chatting were the most common annotations in class-
reading in library and waiting in healthcare facility were
room, healthcare facility, and library indoor environ-
exposed to high levels of noise from above 61dBa. This in-
ments. A car or bus horn was the most common annota-
dicated that road traffic could contribute to an increase of
tions in all crossroads.
the noise levels in library and healthcare facility, which
Logarithmic average values of noise level were above
were (measured) approximately twice as high as the stan-
the standard of 30dBA, 35dBA for indoor and over the stan-
dard of 30dBA.
dard of 55dBA for outdoor environments. This indicated
The methodological approach used in this study is
people were exposed to high noise levels and to health
promising in estimating noise levels in an expanding and
risk [24] as well as to poor task performance in school [5]
growing urban area of Tirana. The present study find-
and in library.
ings provided a basis for improving predictions by using
OLS models of the estimated LAFmax of noise level
yearly noise data, vehicles data, building stores and in-
fitted moderately well with observations of LAFmax (R-
habitants in neighborhoods as well as by collecting exten-
squared was −27% to −21%). The independent variable
sively new data on pavement type, road gradient, noise di-
of the estimated number of vehicles explained most of
rectivity [road traffic modelling see e.g. 14], road conditions
LAFmax and logarithmic average of noise level. The es-
and building constructions. There is an opportunity to col-
timated number of vehicles cannot only be an important
lect data e.g. of building façade to predict noise levels [see
explanatory variable for noise level increase (as shown
e.g. 25]. These data can change predictions of noise levels
by OLS analysis in an expanding urban area like Tirana),
in Tirana.
but also for noise modelling. Zhu [14] has found that their
This study showed that noise measurements relied on
model accuracy increased with number of vehicles in-
participatory approach of collectors (students) for educa-
cluded in their model.
tion purposes. Similarly, much can be done on inhabi-
Basic noise level calculations for roads and estima-
tant participatory on measuring noise levels [7] in restau-
tions of LAFmax and logarithmic average of noise level car-
rants [45, 46], food courts [47] to increase population and
ried out in IDW3D-GIS, were comparable with measured
businesses awareness. This could help business reduce
 Measurements and dominant sounds in indoor and outdoor environments
noise levels in their environments. For example, in indoor
environments, much can be done by the designers to care-
fully consider acoustic details of the interior by shaping
sound sources to be positively perceived with attention to
vocal comfort, ambient levels [46] in public spaces. De-
signers could also update the needs of people to acoustic
indoor environment using public spaces because needs of
people for acoustic requirements of space could be higher
for their key activity e.g. for ‘reading and thinking’ than
for ‘interacting and communicating’ in library [27]. Par-
ticipatory measurements of noise levels can be a useful
approach to involve inhabitants, business people, school
people to increase their awareness about noise and noise
reduction measures in Tirana.
Extensive monitoring of noise levels and a yearly
acoustic measurement for vehicles as well noise level
and noise reduction measurements in quite indoor envi-
ronments like healthcare facility and libraries should be
strongly monitored by governmental bodies as a routine
monitoring procedure.
Acknowledgement: The author is very thankful to three
anonymous reviewers for their comments, to Prof. Dr. F.
Asdrubali for its support to this publication as well as to
students of master degree, profile of energy of the branch
of environmental engineering at the Polytechnic Univer-
sity of Tirana that collected the data for the course of
“Noise protection”, in 2015. The road data were collected
by the author in the year 2008 for her doctoral research
conducted in Germany.
References
[1] International Institute of Noise Control Engineering, Noise emis-
sions of road vehicles - effect of regulations. Final Report by
the I-INCE Working Party on the Effect of Regulations on Road
Vehicle Noise by Ulf Sandberg Convener of the I-INCE Work-
ing Party on Noise Emissions of Road Vehicles Swedish Na-
tional Road and Transport Research Institute (VTI) SE-581 95
Linköping, Sweden. 2001.
[2] International Institute of Noise Control Engineering, Guidelines
for Community Noise Impact Assessment and Mitigation. Final
Report of the I-INCE Technical Study Group on Community Noise:
Environmental Noise Impact Assessment and Mitigation (TSG
6). Prepared by the Members and Consultants of TSG 6, L.S.
Finegold, Editor. 2011.
[3] European Environmental Agency, Europe’s environment: The
Dobriš Assessment. Noise and Radiation. Editors: D. Stan-
ners and P. Bourdeau. Available at: http://www.eea.europa.eu/
publications/92-826-5409-5/page016new.html. Accessed on
13 November 2014. 1995.
| 55
[4] WHO, Guidelines for community noise. Edited by Birgitta
Berglund, Thomas Lindvall, Dietrich H Schwela. Geneva. 1995.
[5] Chetoni, M., et al., Global noise score indicator for classroom
evaluation of acoustic performances in LIFE GIOCONDA project,
in Noise Mapping. 2016. p. 157.
[6] Jakovljevic, B., K. Paunovic, and G. Belojevic, Road-traflc noise
and factors influencing noise annoyance in an urban popula-
tion. Environment International, 2009. 35(3): p. 552-556.
[7] Guillaume, G., et al., Noise mapping based on participative mea-
surements, in Noise Mapping. 2016. p. 140.
[8] Aspuru, I., et al., LIFE+2010 QUADMAP Project: a new methodol-
ogy to select, analyze and manage Quiet Urban Areas defined by
the European Directive 2002/49/EC, in Noise Mapping. 2016. p.
120.
[9] Kephalopoulos, S., et al., Advances in the development of com-
mon noise assessment methods in Europe: The CNOSSOS-EU
framework for strategic environmental noise mapping. Science
of The Total Environment, 2014. 482–483: p. 400-410.
[10] OECD, The state of the environment. Organisation for Economic
Cooperation and Development, Paris. 1991.
[11] Kephalopoulos, S., M. Paviotti, and F. Anfosso-Lédée, Com-
mon Noise Assessment Methods in Europe (CNOSSOS-EU). EUR
25379 EN. Luxembourg: Publications Oflce of the European
Union, 2012, 180 pp. European Commission Joint Research Cen-
tre Institute for Health and Consumer Protection TP 281 21027 -
Ispra (VA). Italy. Document prepared. 2012.
[12] Calixto, A., F.B. Diniz, and P.H.T. Zannin, The statistical modeling
of road traflc noise in an urban setting. Cities, 2003. 20(1): p.
23-29.
[13] Piccolo, A., D. Plutino, and G. Cannistraro, Evaluation and analy-
sis of the environmental noise of Messina, Italy. Applied Acous-
tics, 2005. 66(4): p. 447-465.
[14] Zhu, L., et al., Rapid Modeling Method for Road Traflc Noise
Sources: A case study in Beijing, in Noise Mapping. 2015.
[15] Filho, J.M.A., A. Lenzi, and P.H.T. Zannin, Effects of traflc com-
position on road noise: a case study. Transportation Research
Part D: Transport and Environment, 2004. 9(1): p. 75-80.
[16] Sevillano, X., et al., DYNAMAP – Development of low cost sen-
sors networks for real time noise mapping, in Noise Mapping.
2016. p. 172.
[17] Gaja, E., et al., Sampling techniques for the estimation of the
annual equivalent noise level under urban traflc conditions. Ap-
plied Acoustics, 2003. 64(1): p. 43-53.
[18] Aspuru, I., et al., CITI-SENSE: methods and tools for empowering
citizens to observe acoustic comfort in outdoor public spaces,
in Noise Mapping. 2016. p. 37.
[19] Liu, H.-Y., et al., A conceptual approach to a citizens’ obser-
vatory – supporting community-based environmental gover-
nance. Environmental Health, 2014. 13(1): p. 107.
[20] Ministry of Environment Forests and Water Administration, Reg-
ulation No. 8, date 27.11.2007. Concerning noise threshholds
in indoor and outdoor environments (in Albanian). Available at:
www.moe.gov.al. Accessed on 20 November 2014. 2007.
[21] Ministry of Environment Forests and Water Administration,
National plan for noise management (in Albanian). Available at:
www.mjedisi.gov.al/files/userfiles/Ajer/Plani_veprimit_per_
zhurmat.doc. Accessed on 20 November 2014. 2010.
[22] National Agency of Environment, Report of the Environmental
State in the year 2014 (in Albanian). Available at: www.akm.gov.
al. Accessed on 27 November 2016. 2014: Tirana.
56 | K. Laze
[23] Ministry of Environment Forests and Water Administration
and Agency of Environment and Forests, Report of Environ-
ment 2011: monitoring of noises in urban areas (in Albanian).
Available at: http://www.mjedisi.gov.al/files/userfiles/Public/
RAPORTI_I_GJENDJES_NE_MJEDIS_2011.pdf. Accessed on 20
November 2014. 2012.
[24] World Health Organisation, Guidelines for community noise,
Geneva. Available from: http://www.who.int/docstore/peh/
noise/guidelines2.html. [last cited on 1999]. 1995.
[25] Farcaş, F. and Å. Sivertunb, Road traflc noise: GIS tools for noise
mapping and a case study for Skåne region. The International
Archives of the Photogrammetry, Remote Sensing and Spatial
Information Sciences, Vol. 34, Part XXX. 2009.
[26] ESRI, ArcGIS 10.3. http://www.esri.com. 2016.
[27] Xiao, J. and F. Aletta, A soundscape approach to exploring de-
sign strategies for acoustic comfort in modern public libraries:
a case study of the Library of Birmingham, in Noise Mapping.
2016. p. 264.
[28] Brown, A. and K. Lam, Urban noise surveys. Applied Acoustics,
1987. 20(1): p. 23-39.
[29] Ministry of Environment Forests and Water Administration,
Inter-sectorial environmental strategy. National strategy for
development and integration (in Albanian). Available at:
http://www.mjedisi.gov.al/files/userfiles/Strategjia_Ndersekt
oriale_e_Mjedisit_2007.pdf. Accessed on 20 November 2014.
2007.
[30] Institute of Statistics, Migration in Albania. Retrived from www.
instat.gov.al on 3 October 2016. 2014, INSTAR.
[31] Calogero, C., et al., A Country on the Move: International Mi-
gration in Post-Communist Albania. International Migration Re-
view, 2006. 40(4): p. 767-785.
[32] King, R.L., Albania as a laboratory for the study of migration
and development. Journal of Southern Europe and the Balkans,
2005. 7(2): p. 133-155.
[33] Smiraglia, M., et al., Predicting Hourly Traflc Noise from Traflc
Flow Rate Model: Underlying Concepts for the DYNAMAP Project,
in Noise Mapping. 2016. p. 130.
[34] Institute of Statistics, Population and housing census 2011:
main results. Available at www.instat.gov.al. Accessed on 22
December 2015. 2012.
[35] Municipality of Tirana, Bulletin of statistics of the year 2009.
Available at www.tirana.gov.al. Retrieved on 20 December 2015.
2009.
[36] Ministry of Transport and Infrastructure, Report on Road Safety
for the Year 2013 (in Albanian). Available at www.transporti.gov.
al. Accessed on 22 December 2015. n.a.
[37] R Core Team, R: A language and environment for statistical com-
puting. R Foundation for Statistical Computing, Vienna, Austria.
URL http://www.R-project.org/. 2016.
[38] Woolridge, J.M., Woolridge, Jeffrey M. (2003). Introductory
econometrics: A modern approach. Mason, OH: Thomson
South-Western. . 2003.
[39] Greene, W.H., Econometric analysis (PDF) (5th ed.). New Jersey:
Prentice Hall. ISBN 0-13-066189-9. Retrieved from http://stat.
smmu.edu.cn on 2016-04-19. 2002.
[40] Burnham, K.P. and D.R. Anderson, Multimodel Inference. Socio-
logical Methods & Research, 2004. 33(2): p. 261-304.
[41] de Kluijver, H. and J. Stoter, Noise mapping and GIS: optimising
quality and eflciency of noise effect studies. Computers, Envi-
ronment and Urban Systems, 2003. 27(1): p. 85-102.
[42] Tsai, K.-T., M.-D. Lin, and Y.-H. Chen, Noise mapping in urban
environments: A Taiwan study. Applied Acoustics, 2009. 70(7):
p. 964-972.
[43] Botim i Qendrës së Publikimeve Zyrtare, Kodi rrugor i Repub-
likes se Shqiperise. Miratuar me ligjin nr.8378, datë 22.7.1998,
ndryshuar me ligjin nr.9808, datë 24.9.2007 dhe ligjin nr.10
488, datë 5.12.2011. Qendra e Publikimeve Zyrtare. ISBN: 978-
9928-01-025-4 (in Albanian). Available on www.qbz.gov.al. Ac-
cessed on 28 November 2016. 2012.
[44] Childs, C., Interpolating Surfaces in ArcGIS Spatial Analyst -
Esri. Available on https://www.esri.com/news/arcuser/0704/
files/interpolating.pdf. Retrived from www.esri.com on 3 Octo-
ber 2016. 2004.
[45] Lindborg, P., Psychoacoustic, physical, and perceptual features
of restaurants: A field survey in Singapore. Applied Acoustics,
2015. 92: p. 47-60.
[46] Lindborg, P., A taxonomy of sound sources in restaurants. Ap-
plied Acoustics, 2016. 110: p. 297-310.
[47] Navarro, M.P.N. and R.L. Pimentel, Speech interference in food
courts of shopping centres. Applied Acoustics, 2007. 68(3): p.
364-375.