https://doi.org/10.14311/APP.2022.33.0631
Acta Polytechnica CTU Proceedings 33:631–637, 2022 © 2022 The Author(s). Licensed under a CC-BY 4.0 licence

Published by the Czech Technical University in Prague

QUANTIFYING UNCERTAINTIES IN THE SUSTAINABILITY
EVALUATION OF CONCRETE MATERIALS CONSIDERING

REGIONAL CHARACTERISTICS IN JAPAN

Nicole Alexis Kwan Viosa, ∗, Michael Ward Henryb,
Joel Galupo Oponc

a Hokkaido University, Graduate School of Engineering, Kita 13 Nishi 8, Kita-ku, Sapporo 060-8628 Japan
b Shibaura Institute of Technology, Department of Civil Engineering, Toyosu 3-7-5, Koto-ku, Tokyo 135-8548

Japan
c MSU-Iligan Institute of Technology, Department of Civil Engineering, Faculty Member, Iligan City 9200

Philippines
∗ corresponding author: nicolealexisv@gmail.com

Abstract.
In the evaluation of concrete sustainability, what constitutes "sustainable" to one region may vary

from another. This often leads to methodological forms of uncertainties that makes the evaluation
process more complex. As such, this paper aims to quantify the effect of uncertainties in the regional
context on the sustainability evaluation of concrete materials. This is carried out by quantifying the
regional context through establishing a weighting scheme and then integrating the obtained weights
into the sustainability analysis of concrete materials in tandem with uncertainty analysis. Japan is
used as a case study because although it relatively appears as a homogeneous country, its prefectures
possess unique characteristics that may make the sustainability evaluation of concrete materials vary
across prefectures. Cluster analysis is carried out in the 47 prefectures of Japan using a set of regional
context indicators. Five clusters are identified with varying characteristics and these are translated
into different weighting schemes. The established weights are used in the sustainability evaluation
of concrete materials using multi-criteria decision-making analysis. The results showed that one mix
is the most sustainable for four of the clusters and a different mix is the most sustainable for the
remaining cluster. When uncertainty analysis is conducted, the effect of the weights in the sustain-
ability evaluation is explained by examining the average scores of the concrete mixes and the variance
of the scores across the five clusters. This investigation facilitated the understanding of how regional
differences and the uncertainties associated with it impact the evaluation of concrete sustainability.

Keywords: Regional characteristics, sustainable concrete, uncertainty analysis.

1. Introduction
Sustainable development, commonly defined as "the
development that meets the needs of the present with-
out compromising the ability of future generations
to meet their own needs" [1], has its implementation
dependent on individual governments. Each country
has varying context on this concept making the means
of attaining sustainability different among countries
[2]. With the growing awareness of sustainable de-
velopment, the concrete industry is taking efforts to
participate in the practice of sustainability. Similar
to sustainable development, the implementation of
sustainability in concrete also varies by region be-
cause concrete is a regional context material itself.
Henry and Opon [3] introduced a means for quanti-
tatively analyzing the regional context of sustainabil-
ity and linking this regionality to the sustainability
evaluation of concrete materials. However, method-
ological forms of uncertainties exist in this evaluation
which make the evaluation process more complex. As
such, it is the goal of this paper to quantify the effect

of uncertainties in the regional context on the sus-
tainability evaluation of concrete materials. Japan
was used as a case study for this analysis because al-
though Japan appears to be relatively homogeneous
as a country, it has 47 prefectures having unique char-
acteristics which may make the sustainability evalu-
ation of concrete materials vary across prefectures.

2. Methodology
The analytical flow adopted in this study is summa-
rized in Figure 1. Initially, a set of quantitative re-
gional context indicators was constructed, gathered
for all prefectures and processed accordingly. Then,
an agglomerative hierarchical clustering method was
used to explain the regional context by identifying
groups of prefectures exhibiting similar character-
istics. These clusters form the basis of weighting
schemes which represent regionality and the obtained
weights were used as inputs in the concrete sustain-
ability evaluation of a set of concrete mix utilizing
multicriteria decision-making analysis. Finally, the

631

https://doi.org/10.14311/APP.2022.33.0631
https://creativecommons.org/licenses/by/4.0/
https://www.cvut.cz/en


N. A. Vios, M. W. Henry, J. G. Opon Acta Polytechnica CTU Proceedings

Figure 1. Analytical flow.

Figure 2. Sustainability evaluation framework using MCA.

Range of
z-score z≤-3 -3<z≤-2 -2<z≤-1 -1<z≤-0.5 -0.5<z<0.5 0.5≤z<1 1≤z<2 2≤z<3 3≤z

Relative
intensity
(−)

1/9 1/5 1/3 1/2 1 2 3 5 9

Relative
intensity
(+)

9 5 3 2 1 1/2 1/3 1/5 1/9

Table 1. Conversion of in-cluster mean z-scores to relative intensities for calculation of weights.

effect of methodological uncertainties in the regional
context of concrete sustainability evaluation was in-
vestigated using uncertainty analysis to understand
how these uncertainties influence concrete sustain-
ability evaluation results.

2.1. Cluster analysis methodology and
extraction of weights

In accordance with the guidelines provided by Kas-
sambara [4] for conducting cluster analysis, the
agglomerative hierarchical clustering method was
adopted using the open source statistical analysis
software R [5]. In this approach, the prefectural data
was first standardized to a mean of zero and a stan-
dard deviation of one converting them to z-scores.
After which, the distance matrix which summarizes
the distance between any two prefectures in an n di-
mensional space (where n is the number of indicators)
was calculated using the Euclidean distance method.
A hierarchical tree was created using a linkage func-
tion to group prefectures based on their distance in-
formation. Finally, a dendrogram, which is a visual-
ization of the cluster analysis results, was produced
and examined for analyzing the regionality of con-
crete sustainability. The clusters were the basis in
the extraction of weights.

Weighting of indicators is an important area in sus-
tainability evaluation as they reflect the significance
of different dimensions in their contributions to the
sustainability performance of a system [6]. A weight-
ing scheme based on the Analytic Hierarchy Process
(AHP) approach was used to reflect regionality in
the evaluation of concrete sustainability similar to
the method taken by Henry and Opon [3]. AHP is
a multi-criteria decision-making tool that structures
the decision-making process as a hierarchy composed
of quantifiable criteria and their relations and alterna-
tives towards a goal [7]. Weight determination is de-
pendent upon pairwise comparison of indicators and
since pairwise comparison is not applicable to nega-
tive or zero values, the z-scores were converted to rel-
ative intensities following the scale shown in Table 1
depending on the loading direction of the indicator
as shown in Table 5.

2.2. Concrete sustainability evaluation
methodology

After obtaining the different weights, sustainability
evaluation of the concrete mixes was conducted using
Multi-Criteria Analysis (MCA) shown in Figure 2.
Each step in the framework is explained in the fol-
lowing sections.

632



vol. 33/2022 Sustainability Evaluation of Concrete Materials in Japan

Label Data source Description (units)
Res1 a Total cement consumption in concrete per capita (100kg/capita)
Res2 a Total aggregate consumption in concrete per capita (100kg/capita)
Res3 b Total water withdrawal per capita (m3/person)
Res4 c Usage rate of slag blended cements (0% of total cement consumption)
Res5 c Usage rate of fly ash blended cements (0% of total cement consumption)
Wst1 d Total concrete construction waste generated per capita (100kg/person)
Eco1 d Total construction investment per capita (1000yen/person)
Env1 e Total CO2 emissions per capita (tons/person)
Env2 e Total energy consumption per capita (GJ/person)

Data source: a: Ministry of Economy, Trade and Industry
b: e-Stat
c: Japan Cement Association
d: Ministry of Land, Infrastructure and Transportation
e: Ministry of the Environment

Table 2. Indicators for regional context analysis data from 2010.

2.2.1. Indicator selection
Indicators are selected based on their relevance to
what is being measured. They are the fundamental
structural basis on which the sustainability of a con-
crete material can be judged [8].

2.2.2. Normalization
The indicators are expressed in different units. In or-
der for them to work together in an aggregated com-
posite, normalization is necessary. Several methods
are available for this purpose but for this paper stan-
dardization using z-scores was used.

2.2.3. Weighting
Weighting schemes can sometimes increase or erad-
icate trade-offs between indicators, which makes
weighting a controversial issue in concrete, as cur-
rent codes are silent about it and still emphasize the
hierarchy of mechanical performance [9]. Due to this
subjectivity, clarifying how much of an impact the
weights have in the concrete sustainability evaluation
is demonstrated in this analysis.

2.2.4. Aggregation
Geometric aggregation (G) was chosen as the aggre-
gation method in this paper as it involves aggregation
by geometric mean using the product values [8] and is
appropriate when less compensability is desired [10].
Equation 1 shows how this is calculated where w is
the weight assigned to each indicator and SCI is the
indicator normalized value.

G =
n!

i=1
= (SCi)

wi (1)

Finally, the aggregated values were standardized
using z-scores followed by a transformation to t-scores
converting the results to a scale of 0 to 100 assuming
a mean of 50 and herein referred to as sustainability
score (SS).

3. Results and Discussions
3.1. Regional context indicators
For quantifying the regional context of concrete sus-
tainability, a set of 9 indicators was constructed as
summarized in Table 2 spanning four categories. The
resource consumption indicators (total cement, ag-
gregate and water) and the waste generation indica-
tor (total concrete construction waste) focus on the
flow of materials from the beginning to end of its life
cycle. The distribution of steel manufacturing and
coal power generation within Japan affects the usage
of blended cements leading to the inclusion of two ad-
ditional resource consumption indicators (usage rate
of slag blended and fly ash blended cements). Finally,
economic and environmental aspects of sustainability
were represented with three indicators measuring to-
tal construction investment, total CO2 emissions and
total energy consumption. Normalized values (e.g.
per capita) were utilized to facilitate comparison of
variables that scale with prefectural size.

3.2. Clustering analysis
With the regional context indicators, cluster analysis
was carried out to identify groups of prefectures ex-
hibiting similar characteristics. The linkage function
used was the average linkage function because it had
the highest correlation coefficient of 0.86 between the
Euclidean distance matrix of the standardized data
and the cophenetic distance matrix among the several
linkage functions tested. The dendrogram produced
by the clustering result is shown in Figure 3. The au-
thors set the number of clusters to five by cutting the
tree at the height shown to sufficiently represent re-
gionality without overly simplifying nor overly com-
plicating the analysis. Clusters are numbered from
one to five from the left.

The in-cluster mean z-scores of the five clusters for
each regional context indicator are summarized in Ta-
ble 3. Clusters 1 and 4 are single-membership clusters

633



N. A. Vios, M. W. Henry, J. G. Opon Acta Polytechnica CTU Proceedings

Figure 3. Dendrogram of agglomerative hierarchical clustering results.

Cluster Res1 Res2 Res3 Res4 Res5 Wst1 Eco1 Env1 Env2
1 0.38 0.45 -0.52 -1.44 6.04 0.31 1.98 -0.18 -0.16
2 0.38 -0.14 -0.17 0.67 -0.24 0.84 -0.39 3.17 3.19
3 -0.28 -0.26 0.04 0.02 -0.12 -0.27 -0.22 -0.19 -0.18
4 1.34 1.02 -1.58 1.29 -0.24 3.69 2.00 -0.37 -0.55
5 1.91 2.20 0.32 -0.67 -0.15 0.92 1.34 -0.42 -0.47

Table 3. In-cluster mean z-scores for regional context indicators by cluster.

containing Aomori and Shimane, respectively, that is
the in-cluster mean z-scores of these clusters are the
z-scores of the prefecture themselves. The in-cluster
mean z-scores explain the distinctive characteristics
of each cluster and these were used as inputs in cal-
culating the weights.

3.3. Extraction of weights
The calculated weights are shown in Table 4 and it
can be deduced from these results that different clus-
ters put different weights on the indicators. As these
values were normalized to have a total sum of 1, it is
easier to compare than the in-cluster mean z-scores.
Cluster 1 consisting Aomori puts a very high weight
on Res5 derived from the high value of the prefec-
ture’s fly ash usage rate. Cluster 2 composed of Oita,
Okayama and Yamaguchi puts the same emphasis on
Env1 and Env2 attributed to the high value of its
CO2 emissions and energy consumption. Cluster 3
containing 38 clusters have equal weighting for all the
indicators and can be said to represent the average
conditions in Japan. Cluster 4 with only Shimane
has its greatest weight on Wst1 associated with its
high concrete construction waste. Finally, cluster 5
comprising Okinawa, Toyama, Fukui and Yamanashi

has a number of indicators with high weight includ-
ing Res1, Eco1 and especially Res2 related to its high
total aggregate consumption. These weights are par-
allel with the in-cluster mean z-scores of the clusters.

3.4. Concrete sustainability evaluation
A set of 9 indicators for concrete sustainability eval-
uation was constructed as shown in Table 5. Each
concrete sustainability indicator has its correspond-
ing regional context indicator. The table shows the
loading direction of each indicator which indicates the
behaviour that contributes to increased sustainability
and is employed in the calculation of weights. For ex-
ample, decreasing the consumption of raw materials
improves the sustainability of a concrete mix.

A test set of three concrete mixes was drawn from
Noguchi et.al. [11] and the mix proportions are shown
in Table 6. To render consistency in the analysis,
mixes with 28th day compressive strengths of ap-
proximately 30MPa which are generally used for civil
engineering infrastructures were chosen. Three new
alternatives were created containing high-grade recy-
cled aggregates are also shown in the table. High-
grade recycled aggregates were assumed to have the
exact same properties and cost as normal aggregates

634



vol. 33/2022 Sustainability Evaluation of Concrete Materials in Japan

Cluster Weight (w) Res1 Res2 Res3 Res4 Res5 Wst1 Eco1 Env1 Env2
1 w1 0.05 0.05 0.02 0.15 0.44 0.05 0.15 0.05 0.05
2 w2 0.04 0.04 0.04 0.07 0.04 0.07 0.04 0.33 0.33
3 w3 0.11 0.11 0.11 0.11 0.11 0.11 0.11 0.11 0.11
4 w4 0.12 0.12 0.01 0.12 0.04 0.35 0.19 0.04 0.02
5 w5 0.16 0.26 0.05 0.11 0.05 0.11 0.16 0.05 0.05

Table 4. Indicator weights for each cluster.

Label Indicator Units Loading Related regionalcontext indicator
Cem Cement content kg/m3 − Res1

Agg Natural aggregatecontent kg/m
3 − Res2

Wat Water content kg/m3 − Res3

Bfs Blast furnaceslag content kg/m
3 + Res4

Fa Fly ash content kg/m3 + Res5

Rca Recycled aggregatecontent kg/m
3 + Wst1

Cst Construction materialscost yen/m
3 − Eco1

CO2 CO2 emissions kg−CO2/m3 − Env1
Ener Input energy MJ/m3 − Env2

Table 5. Indicators for concrete sustainability evaluation.

Mix
28th day

f ′c
W/B W C BS FA S NA RA

(MPa) (−) (kg) (kg) (kg) (kg) (kg) (kg) (kg)
Ref 30.57 0.57 184 325 0 0 783 1063 0
A 30.57 0.57 184 325 0 0 783 0 1063
B1 30.44 0.58 192 200 133 0 806.2 965.4 0
B2 30.44 0.58 192 200 133 0 806.2 0 965.4
F1 30.23 0.56 186.7 238.1 0 94.1 847 949.9 0
F2 30.23 0.56 186.7 238.1 0 94.1 847 0 949.9

Table 6. Concrete mix proportions per m3.

for simplicity. Thus, concrete with high-grade recy-
cled aggregates have the same properties as mixes
with normal aggregates, but the inventory data shows
that they require more energy to process to go back
to the original state which is reflected by higher CO2
emissions and energy consumption [12].

Table 7 shows the calculated indicator values for
each mix wherein the six constituent material indica-
tors were all determined directly from the mix pro-
portions. Due to the lack of data on the regional
variation of the costs of constituent materials, this
analysis uses a single value per mix obtained from
Henry and Opon [3] and calculated to get the unit
cost per mix. Lastly, the two environmental impact
indicators were computed using Japanese inventory
data [13].

The concrete sustainability evaluation results us-
ing MCA are shown in Figure 4. Each cluster’s re-

sults represent one implementation of MCA. Looking
at each cluster, varying weighting schemes lead to
varying sustainability scores for the different mixes.
For clusters 1, 3, 4 and 5, the most sustainable con-
crete material is only one mix (F2). While for cluster
2 the most sustainable concrete material is a differ-
ent mix (B1) whose selection is heavily supported by
the high weight placed on CO2 and Ener indicators
by this cluster and among the concrete mixes, mix B1
has the lowest CO2 emission and energy consumption
values. Despite the same most sustainable concrete
material for four of the clusters, the ranks of the other
mixes change. But, a high level of confidence can be
presumed on mix F2 considering a general choice for
the most sustainable concrete material for the entire
country.

635



N. A. Vios, M. W. Henry, J. G. Opon Acta Polytechnica CTU Proceedings

Mix Cem Agg Wat Bfs Fa Rca Cst CO2 Ener
Ref 325 1063 184 0 0 0 5,775.04 255.12 1213
A 325 0 184 0 0 1063 5,775.04 270.86 1563.8
B1 200 965.4 192 133 0 0 5,679.39 162.63 861.84
B2 200 0 192 133 0 965.4 5,679.39 176.92 1180.4
F1 238.1 949.9 186.7 0 94.1 0 5,265.96 190.22 956.62
F2 238.1 0 186.7 0 94.1 949.9 5,265.96 204.28 1270.1

Table 7. Concrete mix characteristics (units are shown in Table 5).

Figure 4. Sustainability scores of concrete mixes by cluster.

Mix Average SS RankAverageSS Variance RankVariance
Ref 36.00 6 7.27 1
A 41.08 5 29.11 5
B1 50.55 4 32.62 6
B2 57.09 2 18.46 2
F1 54.41 3 27.18 4
F2 60.87 1 21.61 3

Table 8. Descriptive statistics of uncertainty analysis for each concrete mix.

3.5. Uncertainty analysis
The subject of uncertainty propagation in concrete
material sustainability evaluation that arises from
weighting schemes is discussed in this section (i.e.
how much the weights impact the sustainability eval-
uation results). The method undertaken for this anal-
ysis is where the SS of each mix for all of the clusters
were aggregated and the corresponding average and
variance together with their individual ranks were
computed as shown in Table 8.

When choosing a sustainable concrete material in
a national level but at the same time being aware of
the existence of the regional context, saying that mix
F2 is the most sustainable concrete mix because it is
ranked 1st by four out of five clusters has uncertain-
ties associated with it. The uncertainty analysis is a
way of quantitatively examining how much consider-
ation of the regional context will affect the selection
of a sustainable concrete mix. As an example, Mix
F2 was ranked 1st based on the average sustainability
score as it showed to be the most sustainable concrete
material for four of the clusters, but considering the
variance of this selection, it ranked 3rd which can rea-

sonably be identified as sustainable. However, when
looking at mix B2 which ranked 2nd based on the
average sustainability score, it showed a lower vari-
ance. This then is where the choice between choosing
a moderately scored material with low variance ver-
sus a highly scored material with moderate variance
comes in. Generally, a mix with higher average sus-
tainability score and lower variability is desired as
this entails a more confident result. Based on this
idea and referring to Figure 5, the bottom right mix is
the most desirable decision and is identified to be mix
F2. These results cannot be an output by conduct-
ing concrete sustainability evaluation alone. Viewing
the mixes in the perspective of uncertainty analysis
gives evidence that even though considering regional
context in concrete sustainability evaluation as repre-
sented by the weights, the most sustainable concrete
material for a national level nonetheless shows to be
the same material that displayed the most sustainable
character in a cluster level.

636



vol. 33/2022 Sustainability Evaluation of Concrete Materials in Japan

Figure 5. Scatter plot of Average SS vs Variance.

4. Conclusions
In this paper, quantifying the regional context of con-
crete sustainability was demonstrated that lead to ob-
taining different weighting schemes. This then was
integrated in the evaluation of concrete sustainabil-
ity. However, uncertainty analysis in this evaluation
suggests that those weights, while they can lead to
different evaluations of concrete mixes at a mesoscale
(cluster level), the effect of the weights potentially
may be negligible in a macroscale (national level) con-
sidering that the assumptions made in the analysis
holds the same and true. This is evidenced by the
varied sustainability scores for each mix per cluster,
but the rank of the most sustainable concrete mate-
rial does not change for one mix except in one clus-
ter. With this at hand, it can be concluded that the
weights are not leading to a significant difference of
the selection of the most sustainable concrete mate-
rial. Therefore, applying a simple weighting scheme,
such as equal weighting, may be sufficient for the rep-
resentative sustainability evaluation of concrete.

References
[1] G. H. Brundtland. Our common future-Call for

action. Environmental Conservation 14(4):291-4, 1987.
https://doi.org/10.1017/S0376892900016805.

[2] M. Henry, Y. Kato. Sustainable concrete in Asia:
Approaches and barriers considering regional context.
Construction and Building Materials 67:399-404, 2014.
https:
//doi.org/10.1016/j.conbuildmat.2013.12.074.

[3] M. Henry, J. Opon. Regional context indicators for
the evaluation of concrete sustainability: an analysis at
the prefectural level in Japan. The 8th International
Conference of Asian Concrete Federation, China, 2018.

[4] A. Kassambara. Practical guide to cluster analysis in
R, Edition 1, sthda.com, 2017.

[5] R Core Team. A language and environment for
statistical computing R Foundation for Statistical

Computing Vienna, 2018.
http://www.R-project.org/.

[6] X. Gan, I. C. Fernandez, J. Guo, et al. When to use
what: Methods for weighting and aggregating
sustainability indicators. Ecological Indicators
81:491-502, 2017.
https://doi.org/10.1016/j.ecolind.2017.05.068.

[7] T. L. Saaty. Decision making for leaders RWS
Publications, RWS Publications, Pittsburgh, 2008.

[8] M. J. Burgass, B. S. Halpern, E. Nicholson, et al.
Navigating uncertainty in environmental composite
indicators. Ecological Indicators 75:268-78, 2017.
https://doi.org/10.1016/j.ecolind.2016.12.034.

[9] J. Opon, M. Henry. Understanding the propagation
of uncertainty in concrete material sustainability
evaluation The 8th International Conference of Asian
Concrete Federation, China, 2018.

[10] M. Saisana, A. Saltelli. Rankings and Ratings:
Instructions for Use. Hague Journal on the Rule of
Law 3(02):247-68, 2011.
https://doi.org/10.1017/s1876404511200058.

[11] T. Noguchi, F. Tomosawa, K. M. Nemati, et al. A
practical equation for elastic modulus of concrete. ACI
Structural Journal 106(5):690-96, 2009.
https://www.researchgate.net/profile/Kamran-Nem
ati/publication/279700549_A_practical_equation
_for_elastic_modulus_of_concrete/links/5cbd7a
30a6fdcc1d49a5ea2c/A-practical-equation-for-e
lastic-modulus-of-concrete.pdf.

[12] M. W. Henry. Formation and evaluation of
sustainable concrete based on social perspectives in the
Japanese concrete industry, 2010.

[13] T. Uomoto, C. Hashimoto. Recommendation for
Shotcreting (Draft) Part Tunnels by NATM by the
Japan Society of Civil Engineers. Shotcrete for
Underground Support X, p. 14-29, 2006.
https://doi.org/10.1061/40885(215)3.

637

https://doi.org/10.1017/S0376892900016805
https://doi.org/10.1016/j.conbuildmat.2013.12.074
http://www.R-project.org/
https://doi.org/10.1016/j.ecolind.2017.05.068
https://doi.org/10.1016/j.ecolind.2016.12.034
https://doi.org/10.1017/s1876404511200058
https://www.researchgate.net/profile/Kamran-Nemati/publication/279700549_A_practical_equation_for_elastic_modulus_of_concrete/links/5cbd7a30a6fdcc1d49a5ea2c/A-practical-equation-for-elastic-modulus-of-concrete.pdf
https://doi.org/10.1061/40885(215)3