water

Article
Individual Water-Saving Response Based on
Complex Adaptive System Theory: Case Study
of Beijing City, China
Hanqing Liu, Yong Zhao * , Haihong Li *, Lizhen Wang and Qingming Wang
State Key Laboratory of Simulation and Regulation of Water Cycle in River Basin, China Institute of Water
Resources and Hydropower Research, Beijing 100038, China; hamtsing@126.com (H.L.);
wanglzh@iwhr.com (L.W.); wangqm@iwhr.com (Q.W.)
* Correspondence: zhaoyong@iwhr.com (Y.Z.); lihh@iwhr.com (H.L.); Tel.: +86-010-6878-1617 (Y.Z.);
+86-010-6878-1816 (H.L.)




Received: 27 February 2020; Accepted: 20 May 2020; Published: 22 May 2020


Abstract: It is of significance to examine the factors impacting individual water-saving behavior and
their mechanism of action for water demand management, especially for domestic water demand
management. To provide a solution to the issues of individual water-saving behavior research with
unclear influence mechanisms and subjective modeling decisions, this study provides a novel idea of
combining social survey and model study for research on domestic water conservation, introduces
complex adaptive system theory to describe the mechanism of action between individuals as well
as between individuals and external stimuli, and proposes a general framework for the model
establishment and analysis of individual water-saving response research under external stimuli.
This study uses Beijing as an empirical city. Based on the analysis results of the survey on Beijing
residents’ domestic water use combined with the problems in Beijing’s domestic water-saving efforts,
this study constructs water-saving regulation scenarios and corresponding individual water-saving
response rules; systematically simulates the change process of individual water use considering the
aspects of water price formulation, water use information, and water-saving publicity and education;
quantitatively analyzes the influence of these three types of external stimuli on individual water
saving; proposes individual water-saving promotion measures; and suggests the three aspects of
water price formulation, water use information feedback, water-saving publicity and education.

Keywords: individual water-saving response; domestic water use; water price; water use information;
water-saving publicity and education; measure analysis

1. Introduction
With the rapid economic development and continuous population growth, an increasing number
of countries are attempting to resolve the challenge of the deepening contradiction between the supply
of and demand for water resources. Accordingly, several countries are developing strategies to restrain
water demand [1]. In the past, water demand management mainly focused on agricultural and
industrial water demand. In recent years, due to a growing population and an ever-increasing standard
of living, the proportion of domestic water consumption in several countries has been increasing,
which has even become the primary growing part of water consumption. Therefore, the domestic
water demand management has been paid increasing attention. The individual is the basic unit of
domestic water use, and the final effect of domestic water saving is closely related to individual
water use behavior. Therefore, examining individual water use behavior is the micro basis of water
demand management. Understanding individual water-saving behavior and its promotion measures

Water 2020, 12, 1478; doi:10.3390/w12051478 www.mdpi.com/journal/water

Water 2020, 12, 1478 2 of 24

is essential to regulate the amount of domestic water use and alleviate the contradiction between the
supply of and demand for water resources.
Scholars in water resource management, sociology, economics, and other domains have actively
explored individual water-saving behaviors to examine the main determinants of individuals’
water-saving behaviors and ways to promote individuals’ water-saving behaviors. Early studies have
focused on the impact of individual attributes, such as gender, age, education level, and income, on
water-saving behavior [2–5]; however, these studies had certain limitations; for example, a few of
these studies revealed that women were more likely to adopt water-saving behavior than men only
through statistical data analysis but failed to explain as to why/how gender caused this difference and
determine the exact difference that this attribute caused.
With the development of a subjective and an abstract index survey method, an increasing number
of scholars have begun to examine the influence of water-saving attitudes, environmental awareness,
values, and other factors on individual water-saving behavior. Furthermore, the structural equation
model and the sociological theory method (e.g., planning behavior theory) have been used to examine
the mechanism of the factors influencing individual water-saving behavior through sociological survey
experiments [6–16].
In addition to the aforementioned factors that have been considered from an individual perspective,
numerous scholars have explored the impact of water price [1,17,18], water-saving policies [1,19],
water-saving publicity and education [20,21], water use information [14,22], water-saving technology [1],
and other external factors on individual water-saving behaviors. Furthermore, the effect of external
control measures on individual water-saving behavior has been analyzed to conduct qualitative
discussions on policy and management. These studies have directly or indirectly associated the factors
influencing individual water-saving behavior with domestic water consumption; however, they have
not explored the mechanism of factors influencing the change in an individual’s water-saving behavior
or the initiative of the individual in the process of changing water-saving behavior. Furthermore,
they have not focused on the complexity of feedback and interaction between individual and external
factors, which is not conducive to understanding the mechanism and process of individual water-saving
behavior changes affected by external factors, and how to adequately guide individuals to alter their
water use behaviors.
Complex adaptive system theory (CAS) provides an effective way to examine the behavioral
rules and their changing process in complex systems [23]. The agent-based model (ABM) of CAS,
which considers the behavior of microscopic agents as the endogenous factor of the system, generalizes
system elements as active, adaptive, and independent agents and simulates the behavior change
process of agents with the interactions among agents and environment to present the evolution process
of the system [24]. ABM has become a critical supplement to the theories and methods of water
resource management and has been effectively applied in domestic water management.
Galán et al. [25] developed an ABM comprising three social submodels for domestic water
management in Valladolid, Spain, illustrating that the combination of ABM and simulation forms a
useful research method to manage the complexity caused by multiple factors that affect the domestic
water management in metropolitan areas. Yuan et al. [26] proposed an ABM to predict urban household
water demand in Beijing, China by simulating the decision-making process of water consumption under
the joint influence of the government’s policy behavior and neighbors’ water consumption behavior,
demonstrating that water price and disposable income are essential to control household water
demand. Ma et al. [27] proposed an ABM of household water consumption simulation to test the policy
effectiveness of controlling household water consumption by simulating the water price negotiation
process between households and the government, which can provide guidance for sustainable water
resource management. Darbandsari et al. [28] introduced a new framework based on ABM to simulate
the behavioral characteristics of domestic water users and their social interactions to evaluate the
responses of domestic water users to factors affecting their water use behavior, which provided
an effective tool for testing water resource management strategies. Koutiva and Makropoulos [29]
Water 2020, 12, 1478 3 of 24

developed the urban water agents’ behavior model based on ABM to simulate the domestic water user’s
behavior in response to water demand management measures in Athens, Greece, which can effectively
calculate the evolution of domestic water demand, examine behaviors and interactions in urban water
systems, and assess the influence of various water demand management schemes. The aforementioned
research demonstrates the universality and usefulness of the theory and method applied in the research
on water demand management policy. However, in the most recent research, the rule making of water
use behavior change in the model has had strong subjectivity; that is, the parameters are mostly set
artificially, and no certain basis is provided.
In order to resolve the problem of unclear influence mechanisms in individual water-saving
behavior research as well as subjectivity and lack of basis in formulating rules and determining
parameters in ABM research, and to strengthen the relation between the model research and actual
regional situation, this study uses the method of combining a questionnaire survey with model
construction to explore the influence mechanism of individual water-saving behavior change and to
refine the model establishment and analysis process. On the basis of the results of a survey on Beijing
residents’ domestic water use and CAS, this paper constructs an individual water-saving response
model under external stimuli. By setting stimulus schemes, the change process of individuals’ water
use behavior under each scheme is systematically simulated. Moreover, the impact of external stimuli
on individual water-saving behavior is quantitatively analyzed to propose measures and suggestions
to promote individual water-saving behavior.

2. Survey on Residents’ Domestic Water Use

2.1. Survey Area

Beijing, the capital of China, a megalopolis with a population of over 20 million, has been plagued
by water scarcity over the past decades [30]. The causes of the urban water crisis can be explained via
the perspectives of water supply and water demand. From 1956 to 2000, its average annual precipitation
was 585 mm, average annual internal renewable water resources were 3.739 billion m3 , and the internal
renewable water resources per capita were less than 200 m3 since 2010 [31]. Furthermore, since
1999, Beijing has suffered several droughts continuously, the amount of surface water resources has
declined sharply, and the groundwater level has declined substantially, which has greatly threatened
the security of the water supply in Beijing. However, the urban population explosion and rapid
economic development have led to a significant increase in the rigid demand for water resources.
Consequently, Beijing has had to take unconventional measures such as interregional water transfer,
over-exploitation of groundwater, and the use of reclaimed water to support urban water demand
over the past two decades. It can be seen that the water scarcity has become a prominent bottleneck
restricting the sustainable development of Beijing.
Simultaneously, Beijing’s water use structure has changed tremendously. Due to urbanization,
industrial structure upgrading, and other reasons, agricultural and industrial water consumption has
continued to decrease over the past few decades, while domestic water consumption has continued
to increase (Figure 1). Since 2005, the domestic water consumption has surpassed agricultural water
consumption, constituting the largest usage of water in Beijing, indicating that domestic water use
may become the leading crisis factor for Beijing’s future water supply. Over the years, the increase
in domestic water consumption was partly due to the increase in population, which jumped from
14.56 million in 2003 to 21.73 million (the peak) in 2016; on the other hand, due to the economic
improvements, the lifestyles of Beijing’s residents tend to be more water intensive, with domestic
water consumption per capita increasing from 193 L/day in 2013 to 231 L/day in 2017, approximately a
20% increase in five years [32]. In recent years, Beijing’s population has been effectively controlled,
while the domestic water consumption per capita has continued to increase. Therefore, water-saving
efforts at the individual level will become the key to the Beijing’s domestic water demand management.
Water 2020, 12, 1478 4 of 24
Water 2020, 12, x FOR PEER REVIEW 4 of 25

20 Agriculture Industry Domestic

Water consumption (108 m3)

16

12

Figure 1. Water consumption by sector from 2003 to 2018. Based on data from [33].
Figure 1. Water consumption by sector from 2003 to 2018. Based on data from [33].
2.2. Survey Research
2.2. Survey Research
To understand the situation of individuals’ water use and water-saving behaviors in Beijing,
we usedTo understand the situation
a questionnaire to procureof individuals’
original data water to use and water-saving
further behaviors infactors
explore the influencing Beijing,andwe
used a questionnaire to procure original data to further explore
mechanisms of individual water-saving behaviors and measures to promote individual water-saving the influencing factors and
mechanisms
behavior. Theofcontent
individual water-saving
of the questionnaire behaviors
mainly and measures
includes thetobasic
promote
personalindividual
informationwater-saving
of the
behavior. The content of the questionnaire mainly includes the basic personal
respondents, the situation of individual water use and water saving, and their potential influencing information of the
respondents,
factors. the situation was
The questionnaire of individual
administered watertouse 523and water saving,
individuals and their
who were potential
randomly influencing
interviewed in
factors. The questionnaire was
public places with a large flow of people. administered to 523 individuals who were randomly interviewed in
public places with a large flow of people.
2.2.1. Findings
2.2.1. Findings
We summarized the noteworthy survey results as follows:
We summarized the noteworthy survey results as follows:
1. The survey results demonstrated that the influence of water price on an individual’s water
1. useThe behavior
survey results
is notdemonstrated
significant asthat the influence
a whole. Of the of water price 74%
respondents, on anstated
individual’s
that water water use
price
behavior
has little is
ornot
no significant
influence on as their
a whole. Of the respondents,
water-saving behavior, and 74% 26%
stated that water
stated that the price has little
increase in
or no influence on their water-saving behavior, and 26% stated that the
water price significantly influences their water-saving behavior. Notably, based on this finding, increase in water price
significantly
an inappropriate influences
conclusiontheir water-saving
would behavior.
be that regulating Notably,
water based
price has littleon
effectthisonfinding,
promoting an
inappropriate conclusion would be that regulating water price has
an individual’s water-saving behavior. This is because many respondents who believed that little effect on promoting an
individual’s water-saving behavior. This is because many respondents
the increase in water price has little effect on the individual’s water-saving behavior stated who believed that the
increase
that in water price
the household has charges
water little effect
andonthe theproportion
individual’s ofwater-saving behaviorexpenditure
the total household stated that the or
household water charges and the proportion of the total household expenditure
income are small and that the increase in the water price in Beijing in recent years is insufficient or income are
small
to and thattheir
stimulate the increase in the water
water-saving behavior.price Inin Beijing
addition,in recent
a systemyearsofisthree-tiered
insufficient to stimulate
pricing for
their water-saving
household behavior. In has
water consumption addition, a system of three-tiered
been implemented pricing
in Beijing (See thefor household
website water
for details:
consumption has been implemented in Beijing (See theAt present,
http://swj.beijing.gov.cn/bmxx/sjxx/201912/t20191219_1311050.html). website more for than details:
90%
http://swj.beijing.gov.cn/bmxx/sjxx/201912/t20191219_1311050.html).
of Beijing residents’ household water consumption does not exceed At present,
the firstmore tier of than 90%
water
of Beijing residents’ household water consumption does not exceed
consumption, which means that less than 10% of residents’ water use behavior is affected by the the first tier of water
consumption,
second- which means
and third-tier waterthat lessthe
price; thanper10%
capitaof residents’ water use constitutes
water expenditure behavior is affected
less thanby the
0.3%
second-
of and third-tier
their income, water
and this price; theisper
proportion toocapita
low to water
affectexpenditure
the residents’ constitutes
water use less than 0.3%
behavior of
[34].
their income, and this proportion is too low to affect the residents’ water
These findings demonstrate that the current water pricing system in Beijing cannot take full use behavior [34]. These
findings demonstrate
advantage that theleverage
of the economic current water
of thepricing
tiered system in Beijing
water price cannot take
in promoting thefull advantage
individual’s
of the economic
water-saving leverage of the tiered water price in promoting the individual’s water-saving
behavior.
2. behavior.
Many respondents were unclear about the water price and household water charges: 20.7%
2. knew
Many the respondents were unclear
current domestic about
water price forthe water price
residents and 53.9%
in Beijing, household
knewwater charges: water
the household 20.7%
knew the current domestic water price for residents in Beijing, 53.9%
charges, 73.5% of the respondents responsible for the payment of household water bills did knew the household water
not
charges, 73.5% of the respondents responsible for the payment of household water bills did not
know the domestic water price for residents in Beijing, and 31.1% did not know the household
Water 2020, 12, 1478 5 of 24

know the domestic water price for residents in Beijing, and 31.1% did not know the household
water charges. Failure to know water prices occurs because most of the current household water
charge is directly deducted from wages or automatically deducted every month after a one-time
recharge of a water card or mobile phone app. These payment methods do not automatically
provide the payer with water price and water fee information and the payer generally does not
actively query the water use information, which indicates the absence of a satisfactory feedback
mechanism of water use information in water management. Furthermore, the survey results
showed that 81.2% of the individuals who knew about household water charges thought that
increasing water prices would affect individual water-saving, while 68.9% of the individuals who
did not know about household water charges thought that increasing water prices would affect
individual water-saving. Thus, increases in water price have a greater impact on the water use
behavior of individuals who know about household water charges, suggesting that improving
individuals’ knowledge of water use information can enhance the promoting effect of water price
on individuals’ water-saving behavior [34].
3. The provision of water-saving publicity and education to the public is poor; thus, respondents’
understanding of the current situation of water resources and water-saving policies in China and
their understanding and mastering of water-saving methods and skills have been poor. Up to
63.5% of the respondents reported that they received little water-saving publicity and education,
19.9% reported that they received them once per year, and less than 17% reported that they
received them several times per year. However, 41.1% of respondents believed that water-saving
publicity and education would considerably impact individuals’ water-saving behavior, and 56.3%
believed that improving an individual’s mastery of water-saving methods and skills through
water-saving publicity and education would considerably impact individuals’ water-saving
behavior. Notably, young adults and men of all ages reported that their insufficient understanding
and mastering of water-saving methods and skills limited their water-saving behavior.

Based on the results of the questionnaire, we observed that unreasonable water pricing, untimely
feedback of water use information, and insufficient water-saving publicity and education are three
major limitations to achieving the target levels of domestic water conservation in Beijing. Therefore,
research on the influence of water price, water use information, and water-saving publicity and
education on individuals’ water-saving behavior would advance the management of domestic water
demand and the promotion of domestic water conservation in Beijing.

2.2.2. Statistical Analysis

Change in individuals’ water use behavior is often a complex process affected by multiple factors.
This section analyzes the interaction of various factors that influence individuals’ water-saving behavior.
Social psychologist Lewin [35] indicated that human behavior is the product of the interaction of
internal factors (e.g., internal conditions and characteristics of the individual) and external factors, and
individual behavior is affected and restricted by these two types of factors. Accordingly, when external
conditions change, individuals with different internal conditions or characteristics have different
degrees of responses to the change.
Therefore, we divide the influencing factors of individuals’ water-saving behavior into two
categories: individual attributes and external stimulus. According to each respondent’s responses to
the questionnaire, we assigned values to the degree that their water-saving behavior was influenced by
external stimulus (values between 1 and 5, 1 representing no influence and 5 representing significant
influence), used the chi-square test to screen out attribute categories that have significant differences
in the impact of an external stimulus on individuals’ water-saving behaviors, and finally applied
cross-statistical analysis to assess the degree of influence of individuals with different attributes in
these attribute categories on their water-saving behavior when receiving external stimuli, with the
values between 1 and 5.
 Water 2020, 12, 1478 6 of 24
Water 2020, 12, x FOR PEER REVIEW 6 of 25

Water 2020, 12, x FOR PEER REVIEW 6 of 25

Water 2020,The
the content 12, of
x FOR PEERattributes
individual REVIEW
water-saving publicity
includedand education,
gender, water-saving
age, education publicity
level, income, and education
and water-charge 6 of 25
stimulus
payment
are content
the divided
frequencyinto
of two categories:
as surveyed
water-saving teachand
knowledge
in the questionnaire.
publicity External(e.g.,
education, theincluded
stimuli currentpublicity
water-saving situation
water price of water
stimulus,
and resources
water use and
education stimulus
the content
water of
information water-saving
stimulus,
policy)into
and teach publicity
and water-saving
water-saving and education,
publicity
methods andand water-saving
education
skills. publicity
stimulus.results
Thecurrent
analysis and
According
are education instimulus
to the content
presented of
Figures
are
are divided
divided into two
two categories:
categories: teach
teach knowledge
knowledge (e.g.,
(e.g., the
the current situation
situation of
of water
water resources
resources and
and
2–5. water-saving publicity and education, water-saving publicity and education stimulus are divided into
water
water policy)
policy) and
and teach
teach water-saving methods and skills.
skills. The analysis results are presented in Figures
two categories: teach water-saving methods
knowledge (e.g., andsituation
the current The analysis
of water resultsand
resources arewater
presented
policy)inand
Figures
2–5.
3.0
2–5.teach water-saving methods and skills. The analysis results are presented in Figures 2–5.
<18 <¥3000 Never
3.0 18–30 ¥3000–¥8000 Occasionally
3.0 <18 <¥3000 Never
<18
31–44 <¥3000
¥8000–¥20,000 Never
18–30 ¥3000–¥8000 Often
18–30
45–59 ¥3000–¥8000
¥20,000–¥50,000 Occasionally
Occasionally
31–44 ¥8000–¥20,000 Often
31–44
≥60 ¥8000–¥20,000
>¥50,000 Often
2.5 45–59 ¥20,000–¥50,000
45–59 ¥20,000–¥50,000
≥60 >¥50,000
2.5 ≥60 >¥50,000
2.5

2.0
Age Monthly income Water-charge
2.0 payment frequency
2.0 Water-charge
Age Monthly income Water-charge
Age
Figure 2. Degree of influence of waterMonthly incomeon individuals
price stimulus payment
with frequency
payment different attributes.
frequency
Figure 2. Degree of of
influence of water price stimuluson
on individuals with different attributes.
3.0 Figure
Figure 2. Degree
2. Degree influenceofofwater
of influence
<18
water price
price stimulus
stimulus onindividuals with
individuals different
with attributes.
different attributes.
<¥3000
3.0 18–30 ¥3000–¥8000 Never
3.0 <18 <¥3000
Occasionally
2.5 <18
31–44 <¥3000
¥8000–¥20,000 Never
18–30 ¥3000–¥8000 Never
18–30
45–59 ¥3000–¥8000
¥20,000–¥50,000 Often
2.5 31–44 ¥8000–¥20,000 Occasionally
2.5 31–44
≥60 ¥8000–¥20,000
>¥50,000 Occasionally
2.0 45–59 ¥20,000–¥50,000 Often
45–59 ¥20,000–¥50,000 Often
≥60 >¥50,000
2.0 ≥60 >¥50,000
2.0
1.5
1.5
1.5
1.0
Water-charge
Age Monthly income
1.0 payment frequency
1.0 Water-charge
Age Monthly income Water-charge
Figure 3. Degree Age of influence of water Monthly income
use information stimulus onpayment frequency
individuals with different
payment frequency
attributes.
Figure
Figure 3. 3. Degreeof
Degree of influence
influenceof of
water use information
water stimulusstimulus
use information on individuals with different with
on individuals attributes.
different
Figure 3. Degree of influence of water use information stimulus on individuals with different
attributes.
2.5
attributes. <18
Primary
2.5 18–30
<18 Male
2.5 <18 Secondary
31–44 Primary
18–30 Primary Female
Male
2.0 18–30 Male
45–59 Tertiary
Secondary
31–44 Secondary Female
2.0 31–44
≥60 Female
2.0 45–59 Tertiary
45–59 Tertiary
1.5 ≥60
≥60
1.5
1.5
1.0
Age Education level Gender
1.0
1.0
Age
Figure 4. Degree of influence Education
of water-saving publicity education stimulus ofGender
andlevel teaching knowledge
Figure 4. Degree of influence Education
Age of water-saving publicity level
and education stimulus Gender
of teaching knowledge
on individuals with different attributes.
on individuals with different attributes.
Figure 4. Degree of influence of water-saving publicity and education stimulus of teaching knowledge
Figure 4. Degree of influence of water-saving publicity and education stimulus of teaching knowledge
on individuals with different attributes.
on individuals with different attributes.
 Water 2020, 12, 1478 7 of 24
Water 2020, 12, x FOR PEER REVIEW 7 of 25

2.5
<18
Primary
18–30
31–44 Secondary
2.0
45–59 Tertiary
≥60
1.5

1.0
Age Education level
Figure 5. Degree of influence of water-saving publicity and education stimulus of teaching skills on
Figure 5. Degree of influence of water-saving publicity and education stimulus of teaching skills on
individuals with different attributes.
individuals with different attributes.
The results demonstrated that the degree of influence of water price stimulus on individuals’
The results
water-saving demonstrated
behavior that the
was generally degree of different
significantly influencedueof water price stimulus
to an individual’s age,on individuals’
income,
and water-charge
water-saving payment
behavior wasfrequency.
generally Furthermore,
significantlythe water-saving
different due tobehaviors of individuals
an individual’s aged
age, income, and
18–44 years, middle-income individuals, and individuals who often paid their water
water-charge payment frequency. Furthermore, the water-saving behaviors of individuals aged 18– charges were the
44most
years,affected by water price
middle-income factors. The
individuals, effect
and of a waterwho
individuals use information
often paid stimulus
their wateron individuals’
charges were the
water-saving behavior was inversely proportional to the age of the individual and the payment
most affected by water price factors. The effect of a water use information stimulus on individuals’
frequency of water charges. Moreover, the effect of water use information stimulus on low-income
water-saving behavior was inversely proportional to the age of the individual and the payment
and high-income individuals was greater than that on middle-income individuals. Water-saving
frequency of water charges. Moreover, the effect of water use information stimulus on low-income
publicity and education stimulus of teaching knowledge considerably impacted the water-saving
and high-income
behavior of youngindividuals
adults aged belowwas greater
30 years,than that on
individuals middle-income
with a lower education individuals. Water-saving
level, and women,
publicity and education stimulus of teaching knowledge considerably impacted
while water-saving publicity and education stimulus of teaching skills considerably impacted the water-saving
the
behavior of young
water-saving adults
behavior ofaged below 30
individuals years,
aged individuals
18–30 years andwiththosea lower
with aeducation level, and
lower education women,
level.
while water-saving
In summary, based onpublicity and attributes
individuals’ educationand stimulus
externalof teaching
stimuli, theirskills considerably
water-saving impacted
behavior was the
affected by different
water-saving behaviordegrees.
of individuals aged 18–30 years and those with a lower education level. In
summary, based on individuals’ attributes and external stimuli, their water-saving behavior was
3. Modeling the System
affected by different degrees.
3.1. Framework
3. Modeling the System
Based on the survey and analysis of individuals’ domestic water use and water-saving behavior
in Beijing, we introduced CAS to construct an individual water-saving response model under external
3.1. Framework
stimuli. In the model, each individual with their specific attributes is regarded as the agent with an
initiative
Basedand learning
on the survey mechanism.
and analysisThe ofagent follows thedomestic
individuals’ general stimulus–response model in their
water use and water-saving behavior
interaction with the external environment; that is, after receiving the external
in Beijing, we introduced CAS to construct an individual water-saving response model under stimulus (measures to
promotestimuli.
external individuals’
In the water-saving
model, each behavior), agents
individual generate
with a corresponding
their specific water-saving
attributes behavior
is regarded as the agent
response by sensing the change in the system environment or other agents and further adjust their
with an initiative and learning mechanism. The agent follows the general stimulus–response model
behavior according to the effect on water-saving behavior. This study systematically simulates the
in their interaction with the external environment; that is, after receiving the external stimulus
change process of individuals’ water use behavior under external stimulus through the model and
(measures to promote individuals’ water-saving behavior), agents generate a corresponding water-
reflects the overall effect on individuals’ water-saving behavior under various stimulus schemes to
saving behavior
analyze individuals’ response by sensing
water-saving the change
promotion measures. in the system environment or other agents and
furtherTheadjust their
specific stepsbehavior
of modelingaccording to the
and analysis are effect on water-saving
summarized as follows: (1)behavior. This study
build a system
systematically
comprising severalsimulates the change
individuals process
(i.e., agents) on theof basis
individuals’ water
of the basic use behavior
situation under
of the research areaexternal
stimulus through
and set the initial the
statemodel and
for each agentreflects
in thethe overall
system; effect on
(2) design individuals’
stimulus schemeswater-saving
and set differentbehavior
regulatory
under various scenarios
stimulus for each external
schemes stimulus;individuals’
to analyze (3) accordingwater-saving
to the sensitivity of agents with
promotion different
measures.
attributes to various
The specific stepsexternal stimuli, formulate
of modeling and analysisthe agent response rules as
are summarized corresponding
follows: (1)tobuild
various
a system
attribute states of the agent under different stimulus schemes. This is so that agents
comprising several individuals (i.e., agents) on the basis of the basic situation of the research area and in an initial
state can generate a corresponding water-saving response according to response rules under different
set the initial state for each agent in the system; (2) design stimulus schemes and set different
regulatory scenarios for each external stimulus; (3) according to the sensitivity of agents with
different attributes to various external stimuli, formulate the agent response rules corresponding to
various attribute states of the agent under different stimulus schemes. This is so that agents in an
initial state can generate a corresponding water-saving response according to response rules under
Water 2020, 12, 1478 8 of 24

external stimulus schemes, and the response results (e.g., the change in water consumption or water
charges)
Water 2020,as
12,feedback
x FOR PEERinformation
REVIEW further affect the subsequent water-saving response of agents. This is
8 of 25
to simulate the dynamic process of individual water use changes with time under different external
different schemes
stimulus external stimulus
and reflect schemes and reflect
the dynamic theofdynamic
change water use change
and theof overall
water use and the overall
water-saving effect
water-saving
of effect ofsimultaneously;
the whole system the whole system andsimultaneously;
(4) based on theand (4) basedofon
regulation the regulation
external ofcompare
stimulus, external
stimulus,
and analyzecompare and analyzeeffects
the water-saving the water-saving effects ofschemes,
of various stimulus various stimulus
summarize schemes, summarize
the stimulus schemesthe
stimulus
with betterschemes with better
water-saving effects,water-saving
and propose effects,
measuresand propose individuals’
to promote measures to water-saving
promote individuals’
behavior.
water-saving
The framework behavior.
of modelThe frameworkand
construction of model construction
analysis is presentedand analysis
in Figure 6. is presented in Figure 6.

Step 1: set the initial state for each agent Step 3: formulate the agent response rules

External stimulus 1 Agent 1 Response result 1

Water-saving Overall
External stimulus 2 Act on Agent 2 behavior response Response result 2 water-
External stimulus 3 Agent 3 Feedback Response result 3 saving
… … … effect

Step 2: design stimulus schemes Step 4: propose measures to promote water-saving

Figure 6. Framework of model construction and analysis.

3.2. Individual Water-Saving

3.2. Individual Water-Saving Response
Response Model
Model
3.2.1. Agent Setting
3.2.1. Agent Setting
According to the analysis results in Section 2.2.2, each external stimulus has a different degree
According to the analysis results in Section 2.2.2, each external stimulus has a different degree
of influence on the water-saving behavior of individuals with different attributes. Therefore, we set
of influence on the water-saving behavior of individuals with different attributes. Therefore, we set
some attributes for each agent in the model to establish the individual response rules under external
some attributes for each agent in the model to establish the individual response rules under external
stimuli. In this model, the agent attributes included gender, age, education level, monthly income,
stimuli. In this model, the agent attributes included gender, age, education level, monthly income,
and water-charge payment frequency. We also set the initial monthly total water consumption for each
and water-charge payment frequency. We also set the initial monthly total water consumption for
agent, including the domestic water consumption of the agent in the family and in public places.
each agent, including the domestic water consumption of the agent in the family and in public places.
Based on the questionnaire samples, this model constructed a system with 500 agents, set attributes
Based on the questionnaire samples, this model constructed a system with 500 agents, set
for each agent so that the composition of agents in the system conforms to the demographic
attributes for each agent so that the composition of agents in the system conforms to the demographic
characteristics of Beijing, and set the initial monthly total water consumption for the agent according
characteristics of Beijing, and set the initial monthly total water consumption for the agent according
to the distribution law of per capita domestic water consumption in 2015 [36].
to the distribution law of per capita domestic water consumption in 2015 [36].
3.2.2. Stimulus Schemes
3.2.2. Stimulus Schemes
The external stimuli in this model comprised a water price stimulus, water use information
The external stimuli in this model comprised a water price stimulus, water use information
stimulus, and water-saving publicity and education stimulus. Each external stimulus has multiple
stimulus, and water-saving publicity and education stimulus. Each external stimulus has multiple
regulatory scenarios. Figure 7 presents the content considered in the stimulus schemes.
regulatory scenarios. Figure 7 presents the content considered in the stimulus schemes.
Water price stimulus refers to adjusting the water pricing system to regulate individual water
Water price stimulus refers to adjusting the water pricing system to regulate individual water
charges so as to stimulate individuals’ water-saving behavior. Based on the problems in the current
charges so as to stimulate individuals’ water-saving behavior. Based on the problems in the current
three-tiered water pricing system in Beijing (e.g., the excessive setting of the first-tier water consumption
three-tiered water pricing system in Beijing (e.g., the excessive setting of the first-tier water
and the unreasonable setting of the tiered water prices), this model designs three regulatory scenarios
consumption and the unreasonable setting of the tiered water prices), this model designs three
to examine the impact of water price stimulus on individuals’ water-saving behavior: (1) reduce
regulatory scenarios to examine the impact of water price stimulus on individuals’ water-saving
the scope of water consumption in the first tier to reduce the first-tier water price coverage rate (i.e.,
behavior: (1) reduce the scope of water consumption in the first tier to reduce the first-tier water price
the proportion of the agent whose water consumption is within the scope of water consumption of
coverage rate (i.e., the proportion of the agent whose water consumption is within the scope of water
the first tier to all agents); (2) increase the water price of the first tier or the second tier (because the
consumption of the first tier to all agents); (2) increase the water price of the first tier or the second
coverage rate of the third tier in real life is low; additionally, to simplify the model scenario setting,
tier (because the coverage rate of the third tier in real life is low; additionally, to simplify the model
the tiered water price in this model only considers the two-tier water price); (3) change the metering
scenario setting, the tiered water price in this model only considers the two-tier water price); (3)
change the metering cycle of water consumption, that is, change the mode of division for tiered water
consumption from annual cumulative water consumption (Mode I) to monthly cumulative water
consumption (Mode II) and quarterly cumulative water consumption (Mode III).
charges (information I), and (2) regularly inform information I and the average water consumption of
individuals in the system (information II).
Water-saving publicity and education stimulus includes two categories: one is to teach
knowledge, the other is to teach water-saving skills. The former affects individuals’ water-saving
behavior
Water by1478
2020, 12, strengthening individual water-saving awareness, and the latter promotes individuals’ 9 of 24
water-saving behavior by improving individuals’ water-saving skills and changing water use habits.
This model designs three schemes for the stimulus: (1) conduct only water-saving publicity and
cycle of water
education consumption,
stimulus that
of teaching is, change(WSPE
knowledge the mode of conduct
I); (2) divisiononly
for tiered water consumption
the water-saving publicityfrom
and
annual cumulative water consumption (Mode I) to monthly cumulative water consumption
education stimulus of teaching skills (WSPE II); (3) conduct water-saving publicity and education (Mode II)
and quarterly
stimulus cumulative
of teaching water consumption
knowledge (Mode
and skills (WSPE III).
III).

• The coverage rate of the first step water price

Water price stimulus • The tiered price of water

• The metering cycle of water consumption

• Water use information of the individual

Water use information stimulus
• Water use information of the system

• Water-saving publicity and education of teaching knowledge

Water-saving publicity and
education stimulus
• Water-saving publicity and education of teaching skills

Figure 7.
Figure 7. Regulatory
Regulatory scenarios
scenarios of
of stimulus schemes.
stimulus schemes.

3.2.3.AResponse
water useRules for the Model
information stimulus increases individuals’ understanding of their or others’ water
use to affect their water use behavior. The research content of this part of the model is the response of
1. Water price stimulus
an individual’s water-saving behavior to the active display of water use information to individuals and
According to the sensitivity of agents with different attributes to water price stimulus, the
is divided into two schemes: (1) regularly inform individual water consumption and water charges
response rules corresponding to various attributes of agents under the water price stimulus are
(information I), and (2) regularly inform information I and the average water consumption of individuals
formulated. Because the water price stimulus mainly affects household water consumption, the agent
in the system (information II).
response here refers to the change in the agent in household water use, reflected through the change
Water-saving publicity and education stimulus includes two categories: one is to teach knowledge,
in household water consumption (Wh). Under the water price stimulus, the inspiration mechanism of
the other is to teach water-saving skills. The former affects individuals’ water-saving behavior by
residents’ water-saving response occurs when the water charges exceed the psychologically
strengthening individual water-saving awareness, and the latter promotes individuals’ water-saving
acceptable range of the agent’s consumption, the agent correspondingly reduces water consumption
behavior by improving individuals’ water-saving skills and changing water use habits. This model
and then continuously adjusts their water use behavior according to the feedback information of the
designs three schemes for the stimulus: (1) conduct only water-saving publicity and education stimulus
water-saving response effect.
of teaching knowledge (WSPE I); (2) conduct only the water-saving publicity and education stimulus
The specific agent response rule is as follows: household water consumption in the current
of teaching skills (WSPE II); (3) conduct water-saving publicity and education stimulus of teaching
month (Wh(i)) is charged according to the tiered water price. If the ratio R of the monthly water charges
knowledge and skills (WSPE III).

3.2.3. Response Rules for the Model

1. Water price stimulus
According to the sensitivity of agents with different attributes to water price stimulus, the response
rules corresponding to various attributes of agents under the water price stimulus are formulated.
Because the water price stimulus mainly affects household water consumption, the agent response
here refers to the change in the agent in household water use, reflected through the change in
household water consumption (Wh ). Under the water price stimulus, the inspiration mechanism of
Water 2020, 12, 1478 10 of 24

residents’ water-saving response occurs when the water charges exceed the psychologically acceptable
range of the agent’s consumption, the agent correspondingly reduces water consumption and then
continuously adjusts their water use behavior according to the feedback information of the water-saving
response
Water effect.
2020, 12, x FOR PEER REVIEW 10 of 25
The specific agent response rule is as follows: household water consumption in the current month
(Wthe
to h(i) ) is charged
income in according
the current to the tiered(Rwater
month (i)) is price.
greaterIf the
thanratio R of theratio
a certain monthly
r, thewater chargeswater
household to the
income in the current month (R ) is greater than a certain ratio r, the household
consumption in the next month(i)(Wh(i+1)) is reduced by different agents with varying degrees of is, andwater consumption in
the household
the next monthwater (Wh(i+1) ) is reducedinbythe
consumption different
month agents with
after the nextvarying
month degrees is , and the
ofadjusted
(Wh(i+2)) is household
according to
water charges
water consumptionin thein themonth.
next month Considering
after the nextthatmonththe (W
impact ) iswater
h(i+2) of adjusted
priceaccording
stimulusto onwater charges
individuals’
in the use
water nextbehavior
month. weakens
Considering overthat
time,theweimpact
assumed of water price
that the stimulus on individuals’
aforementioned adjustment wouldwater not
use
behavior
be more than weakens
three over
timestime,
(i.e., we assumed
a duration ofthat the
three aforementioned
months) adjustment
and that water wouldwould be usednot be more
according
than
to thethree times (i.e.,
last adjusted a duration
water use mode. of three months)
The specific and that water
water-saving wouldprocess
response be usedisaccording
presentedto in the last
Figure
adjusted
8. water use mode. The specific water-saving response process is presented in Figure 8.

Figure 8.
Figure Schematicdiagram
8. Schematic diagramofof individual
individual water-saving
water-saving response
response process
process under
under the
the water
water price
price
stimulus. “①”
stimulus. “O” “②”
1 “O” “③”
2 “O” represent
3 represent the
the first,
first,second,
second,and
andthird
thirdadjustment,
adjustment,respectively.
respectively.“a” meansi =
“a”means i =i
i++1.1. “b”
“b” means
means that
that the
the water
water consumption
consumption will not change
will not change after
after three
three times
times of
of water
water use
use adjustment.
adjustment.

2. Water
2. Water use information
use informationstimulus
stimulus
In the model, water use information refers
In the model, water use information refers to
to the
the amount
amount ofof water
water used
used by
by the
the agent
agent in
in the
the home
home
and the corresponding water charges. The water use information stimulus only
and the corresponding water charges. The water use information stimulus only affects household affects household
water consumption,
water consumption,and andthetheagent
agentresponse
response refers
refers to to
thethe change
change in the
in the agent
agent in household
in household water water
use
use reflected through the change in W
reflected through the change in Wh. Different h . Different water use information stimulus schemes
water use information stimulus schemes produce produce
different response
different response processes.
processes. The
The agent
agent response
response rules
rules under
under the
the two
two stimulus
stimulus schemes
schemes inin this
this study
study
are as
are as follows:
follows:
Under the first stimulus scheme, there are two reasons to stimulate individual water-saving
behavior: (1) the water charges exceed the psychologically acceptable range of the agent’s
consumption, and (2) cumulative water consumption exceeds the first-tier water consumption,
suggesting that the agent uses too much water. Thus, the agent response rule under this scheme is to
increase the judgment of whether the water price increases from the first tier to the second tier on the
Water 2020, 12, 1478 11 of 24

Under the first stimulus scheme, there are two reasons to stimulate individual water-saving
behavior: (1) the water charges exceed the psychologically acceptable range of the agent’s consumption,
and (2) cumulative water consumption exceeds the first-tier water consumption, suggesting that the
agent uses too much water. Thus, the agent response rule under this scheme is to increase the judgment
of whether the water price increases from the first tier to the second tier on the basis of the individual
Water 2020, 12, x FOR PEER REVIEW 11 of 25
water-saving response rule under the water price stimulus. The specific water-saving response process
is presented
saving responsein Figure 9, where
process ix represents
is presented the9,
in Figure degree
whereofiximpact of the
represents water
the useofinformation
degree stimulus
impact of the water
on the
use water-saving
information behavior
stimulus of different
on the agents.
water-saving behavior of different agents.

Schematic diagram
Figure 9. Schematic diagram of
of the
the individual
individual water-saving
water-saving response
response process
process under
under the water use
information stimulus.
information stimulus. “①”“O” “O”
1 “②” “O”
2 “③” represent the
3 represent the first,
first, second,
second, and
and third
third adjustment,
adjustment,respectively.
respectively.
“a” means ii =
“a” means = ii + 1. “b”
+ 1. “b” means
means that
that the
the water
water consumption
consumption willwill not
not change
change after
after three
three times
times of
of water
water
use adjustment.
use adjustment.

In the
In thesecond
secondstimulus
stimulus scheme,
scheme, in addition
in addition to two
to the the reasons
two reasons for stimulating
for stimulating individual
individual water-
water-saving in the first stimulus scheme, individuals change their water use
saving in the first stimulus scheme, individuals change their water use behavior because their behavior because their
monthly water
monthly water consumption
consumption exceeds
exceeds the
the average
average monthly
monthly water
water consumption
consumption of of individuals
individuals inin the the
system. The specific response steps under this plan are as follows: (1) according
system. The specific response steps under this plan are as follows: (1) according to the process to the process
presented in 0
presented in Figure
Figure 9,9, we
we preliminarily
preliminarilycalculated
calculatedthe
thewater
waterconsumption
consumptionininthe thenext
nextmonth
month(W( Wh(h'i(+ );
1));
i +1)
(2) according to the comparison between the Wh(i) of the agent and the average Wh(i) of the system
(2) according to the comparison between the Wh ( i ) of the agent and the average Wh ( i ) of the system (
(Wha(i) ) and the comparison between the ratio of the monthly water charges to the income in the current
W ha ( i ) ) and the comparison between the ratio of the monthly water charges to the income in the
current month (R(i)) of the agent and the average R(i) of the system (Ra(i)), we finally calculated the
Wh(i+1) of the agent by using Formula (1).

 '
Wh ( i +1)， if Wh ( i ) ≤ Wha ( i )

  W − Wha ( i ) 
Wh ( i +1) = Wh'( i +1) − Wh ( i ) ⋅  h ( i )  ， if Wh ( i ) > Wha ( i ) and Ri > Ra ( i ) (1)
  10 ⋅ Wha ( i ) 

W ' − 0.8 ⋅ W ⋅  Wh ( i ) − Wha ( i )  ， else
h(i )  
 h ( i +1)  10 ⋅ Wha ( i ) 

3. Water-saving publicity and education stimulus

Water 2020, 12, 1478 12 of 24

month (R(i) ) of the agent and the average R(i) of the system (Ra(i) ), we finally calculated the Wh(i+1) of
the agent by using Formula (1).
 0


 Wh(i+1) , i f Wh(i) ≤ Wha(i)
 
Wh(i) −Wha(i)


 W0 − W · , i f Wh(i) > Wha(i) and Ri > Ra(i)

Wh ( i + 1 ) =
 h(i+1) h(i) 10·Wha(i) (1)
 
Wh(i) −Wha(i)


 W0 − · W · , else


h(i+1)
0.8 h(i) 10·Wha(i)

3. Water-saving publicity and education stimulus

Different from the former two types of external stimuli, a water-saving publicity and education
stimulus can affect an individual’s water use behavior in the family and in public places. We assumed
that the degree of influence on the two is the same. This part reflects the individual water-saving
response process with the change in the total water consumption (W).
For the first stimulus scheme, WSPE I affects individuals’ water-saving behavior by influencing
individual water-saving attitude. Water-saving attitude, as an emotional influence factor, plays a
significant role in promoting individuals’ water-saving behavior in the short term; however, its role
decreases with time [6], that is, the degree of influence (i1 ) of WSPE I on individuals’ water-saving
behavior decreases with time. Because we have no micro data, we assumed that the change rule of i1
was similar to the individual forgetting rule. We assumed that after an individual receives a WSPE I at
the beginning of the first month, water consumption in the following months is calculated according to
the following formula:
1
W (t) = W1 · (1 − i1 · t−1 ) (2)
b
where W (t) denotes the total water consumption of the individual in the t-th month, W 1 denotes the
total water consumption of the individual in the initial state, and b denotes the attenuation coefficient
of the degree of influence.
For the second stimulus scheme, WSPE II promotes individuals’ water-saving behavior by
increasing an individual’s understanding of water-saving methods and improving individuals’
water-saving skills. The influence of acquiring water-saving methods and skills on promoting
individuals’ water-saving behavior is stable and lasting. Therefore, we assumed that the degree of
influence (i2 ) of WSPE II on individuals’ water-saving behavior will not change with time. After an
individual receives a WSPE II at the beginning of the first month, the water consumption in the
following months is calculated according to the following formula:

W ( t ) = W1 · ( 1 − i2 ) (3)

The third stimulus scheme is to combine two types of water-saving publicity and education. The
impact is to simply add the effect of two types of water-saving publicity and education. After the
individual receives the WSPE III at the beginning of the first month, the water consumption in the
following months is calculated according to the following formula:

1
W ( t ) = W1 · ( 1 − i1 · − i2 ) (4)
bt−1

3.3. Parameter and Assumption

3.3.1. Parameter
The agent response rules of the model involve the following parameters:

• r: when the R of an agent exceeds this value, the water price factor starts to promote individuals’
water-saving behavior. At present, the average R of Beijing residents is approximately 0.5%,
and the average R of the system in the initial state of this model is 0.56%, which is consistent
Water 2020, 12, 1478 13 of 24

with the actual situation of Beijing. Regarding the r value that has started to promote individuals’
water-saving behavior, literature has not yet reached a consensus. Nevertheless, the relevant
departments in China have proposed that the R should be greater than 1% according to the actual
situation in China [37]. Therefore, this model sets r as 1%.
• b: this indicates the attenuation coefficient for the impact of water-saving publicity and education
stimulus of teaching knowledge on individuals’ water-saving behavior. In this study, we referred
to the results of similar studies, and the b value was set as 2 [38].
• is , ix , i1 , i2 respectively represent the degree of influence of the water price stimulus,
water information stimulus, water-saving publicity and education stimulus of teaching knowledge,
as well as water-saving publicity and education stimulus of teaching skills on individuals’
water-saving behavior. The setting of these parameters is based on the results of the questionnaire
and analysis (shown in Section 2.2.2).

3.3.2. Assumption
Because the change in individuals’ water use behavior in real life is a complex process, to realize
the model simulation smoothly, we assumed the following to simplify the model simulation process:
(1) the time scale of the model is month; that is, the minimum unit of measuring individual water
use change is month, and daily water use change is not considered.
(2) the time range of the model is one year; that is, we only studied the change in individual water
use within one year, and the level of socioeconomic development and individual monthly income
did not change in this year.
(3) regardless of the seasonal change in water consumption, we assume that the individual uses
water evenly within the year without external stimulus.
(4) the influence of factors beyond the scope of model consideration on individual water-saving
is ignored.

4. Results

4.1. Analysis of Individual Water-Saving Response under the Water Price Stimulus
Based on the questionnaire analysis, we qualitatively understood that the current domestic water
price for residents in Beijing has little impact on individual water use as a whole, which doesn’t
give full play to the role of water price regulation in promoting individuals’ water-saving behavior.
Further analysis demonstrates that the impact of the water price stimulus on individuals’ water-saving
behavior will be significantly different among individuals belonging to different age groups, income
levels, or water-charge payment frequency.
Based on the population structure of Beijing and the sensitivity of individuals with different
attributes to the water price stimulus, the influence of water price factors on individual water use
under the current water price system in Beijing was quantitatively studied by constructing a model.
Assuming that all agents in the system do not know the monthly water charges, the agents use water
normally according to the original water use habits in the initial month. The model calculation results
demonstrated the change in the total monthly domestic water consumption of the system under the
current water price system in Beijing (Figure 10).
In Figure 10, the month with considerable water consumption reduction lies in the first three
months. This finding occurs because 8.8% of the agents in the initial state of the model settings have
an R greater than 1%, and the water consumption of these agents does not change in the fourth to
ninth months after three months of water use adjustment. The real water saving owing to the tiered
water price starts from the tenth month, that is, the cumulative water consumption of some agents
exceeds that of the first tier and the water price enters the second tier, resulting in an R greater than 1%,
which starts to stimulate individuals’ water-saving behavior. The reduction of water consumption in
this stage is small, and the water consumption in the twelfth month is only 0.19% less than that in
Water 2020, 12, 1478 14 of 24

the ninth month. Thus, based on the current water price system in Beijing, the impact of the water
price
Waterstimulus
2020, 12, x on
FORindividuals’
PEER REVIEWwater-saving behavior starts six months later, and the degree of impact
14 of 25
is small.

Figure10.
Figure 10. Change
Change in
inthe
thetotal
totalmonthly
monthlyhousehold water
household consumption
water under
consumption the current
under water
the current price
water
price system.
system.

At
Atpresent,
present,the thecoverage
coveragerate
rateofofthe
thefirst-tier
first-tierwater
waterprice
priceininBeijing
Beijingisis90%,
90%,indicating
indicatingthat
that90%
90%ofof
the
the residents’ household water charges are not affected by the implementation of the tiered price.
residents’ household water charges are not affected by the implementation of the tiered water water
Some
price.scholars have suggested
Some scholars that the coverage
have suggested rate of therate
that the coverage first-tier
of thewater pricewater
first-tier shouldprice
be reduced
should to be
30% [39] to ensure that most residents are constrained by the tiered water price and
reduced to 30% [39] to ensure that most residents are constrained by the tiered water price and to to promote more
residents
promoteto saveresidents
more water. Into
this part,
save the first-tier
water. waterthe
In this part, price coverage
first-tier rate
water is reduced
price coveragefrom 90%
rate to 30%.
is reduced
The
fromchange
90% to in 30%.
waterThe
consumption
change in under different first-tier
water consumption underwater price coverage
different rates isprice
first-tier water presented
coveragein
Water 2020,
Figure 11. 12, x FOR PEER REVIEW 15 of 25
rates is presented in Figure 11.
Through Figure 11, we observe that reducing the first-tier water price coverage rate improves
the water-saving effect of the tiered water price; however, the annual water consumption
corresponding to the coverage rate of 30% is reduced by less than 0.3% compared with that
corresponding to 90%. Clearly, reducing the first-tier water price coverage on the basis of the original
water price has little effect on improving the water-saving effect of the tiered water price. In addition,
when the first-tier water price coverage rate changes between 90% and 80%, the water consumption
barely changes. This finding demonstrates that with an unadjusted water price, only when the first-
tier water price coverage rate decreases to a certain value can the adjustment of the coverage rate play
a better role in water-saving.

Figure 11.Change
Figure11. Changein
intotal
totalannual
annualhousehold
householdwater
waterconsumption
consumptionunder
underdifferent
differentfirst-tier
first-tierwater
waterprice
price
coverage
coveragerate.
rate.

Next, we consider the situations of increasing the first- and second-tier water price, respectively.
Table 1 presents the household water consumption with the tiered water price unchanged, the first-
tier water price increasing by ¥1, and the second-tier water price increasing by ¥1–¥5 in sequence
when the first-tier water price coverage rate is 90%, 50%, and 30%, respectively. In Table 1, when the
first-tier water price coverage rate is 90%, the water-saving effect of increasing the second-tier water
Water 2020, 12, 1478 15 of 24

Through Figure 11, we observe that reducing the first-tier water price coverage rate improves the
water-saving effect of the tiered water price; however, the annual water consumption corresponding
to the coverage rate of 30% is reduced by less than 0.3% compared with that corresponding to 90%.
Clearly, reducing the first-tier water price coverage on the basis of the original water price has little
effect on improving the water-saving effect of the tiered water price. In addition, when the first-tier
water price coverage rate changes between 90% and 80%, the water consumption barely changes.
This finding demonstrates that with an unadjusted water price, only when the first-tier water price
coverage rate decreases to a certain value can the adjustment of the coverage rate play a better role
in water-saving.
Next, we consider the situations of increasing the first- and second-tier water price, respectively.
Table 1 presents the household water consumption with the tiered water price unchanged, the first-tier
water price increasing by ¥1, and the second-tier water price increasing by ¥1–¥5 in sequence when the
first-tier water price coverage rate is 90%, 50%, and 30%, respectively. In Table 1, when the first-tier
water price coverage rate is 90%, the water-saving effect of increasing the second-tier water price by ¥5
is still not as good as that of increasing the first-tier water price by ¥1. When the first-tier water price
coverage rate is 50%, the water-saving effect of increasing the second-tier water price by ¥3 is better
than that of increasing the first-tier water price by ¥1, and the per capita annual water charges when
the second-tier water price is increased by ¥3 is less than that when the first-tier water price is increased
by ¥1; when the first-tier water price coverage rate is 30%, the water-saving effect of increasing the
second-tier water price by ¥2 is better than that of increasing the first-tier water price by ¥1, and the
per capita annual water charges when the second-tier water price is increased by ¥2 is also less than
that when the first-tier water price is increased by ¥1.

Table 1. Household water consumption after increasing the water price under different first-tier water
price coverage rates.

First Tier Water Price Tiered Water Price

Coverage Rate Variables
5,7 a 6,7 b 5,8 c 5,9 c 5,10 c 5,11 c 5,12 c
Total annual
household water 22,484.9 22,409.3 22,480.4 22,480.4 22,480.4 22,477.3 22,477.3
consumption
90%
Per capital annual
226.8 269.9 227.8 228.8 229.8 230.7 231.7
water charges
Total annual
household water 22,440.6 22,377.3 22,408.1 22,388.9 22,366.7 22,345.5 22,334.0
consumption
50%
Per capital annual
240.7 276.6 248.3 256.0 263.6 271.1 278.8
water charges
Total annual
household water 22,424.8 22,364.5 22,383.2 22,361.2 22,329.9 22,304.0 22,281.4
consumption
30%
Per capital annual
246.2 279.3 256.5 267.0 277.2 287.4 297.6
water charges
Notes: “5,7” means that the first-tier water price is ¥5 and that the second-tier water price is ¥7. Other similar
expressions have the same interpretation. a It is the original tiered water price, that is, the scenario of not changing
the tiered water price. b It is the scenario of increasing the first-tier water price. c It is the scenario of increasing the
second-tier water price.

Accordingly, it is concluded that when the first-tier water price coverage rate is high,
the water-saving effect caused by increasing the second-tier water price is not obvious, and reducing
the first-tier water price coverage rate promotes the water-saving effect of improving the second-tier
water price. When increasing the second-tier water price and increasing the first-tier water price
achieve the same water-saving effect, the increase in water charges in the former case is less than
Water 2020, 12, 1478 16 of 24

that in the latter case because increasing the first-tier water price increases the water charges of all
agents, including agents with lower water consumption, while increasing the second-tier water price
only increases the water charges of agents with excessive water consumption. This demonstrates that
in the case of the low coverage rate of the first-tier water price, increasing the second tier does not
increase the water charges of basic water consumption in life but effectively controls unnecessary water
consumption, that is, it can achieve a better water-saving effect and will not increase the individual
economic burden too much.
The results of this comprehensive study on the impact of reducing the first-tier water price
coverage rate and increasing the tiered water price on household water use are presented in Figure 12.
We draw the following conclusions from Figure 12:

(1) All curves in Figure 12 demonstrate a downward trend from left to right, indicating that reducing
the first-tier water price coverage rate under any water price promotes individuals’ water-saving
behavior. When the coverage rate is greater than 60%, the curve decline range is smaller; when the
coverage rate is less than 60%, the curve decline range is large; and when the coverage rate is
reduced from 55% to 50%, the curve decline range is the largest, indicating that when the first-tier
water price coverage rate is less than 60%, adjusting the first-tier water price coverage rate can
better indicate the water-saving effect of the tiered water price.
(2) The solid lines in Figure 12 are dense on the left and sparse on the right, which demonstrates
that when the first-tier water price coverage rate is high, increasing the second-tier water price
has no obvious effect on promoting individuals’ water-saving behavior. When the first-tier
Water 2020, 12, x FOR PEER REVIEW 17 of 25
water price coverage rate is reduced to 50%, increasing the second-tier water price significantly
increases the reduction in household water consumption. Especially when the second-tier water
price is increased to ¥13, the water-saving effect of increasing the second-tier water price is
price is increased to ¥13, the water-saving effect of increasing the second-tier water price is
particularly significant.
particularly significant.
(3) Comparing the dotted lines with the solid lines in the figure, we observe that when the first-tier
(3) Comparing the dotted lines with the solid lines in the figure, we observe that when the first-tier
water price coverage rate is greater than 75%, the water-saving effect of the second-tier water
water price coverage rate is greater than 75%, the water-saving effect of the second-tier water price
price increased to ¥20 is not as good as that of the first-tier water price increased by ¥1.
increased to ¥20 is not as good as that of the first-tier water price increased by ¥1. Additionally,
Additionally, the dotted line intersects with all the solid lines representing the second-tier water
the dotted line intersects with all the solid lines representing the second-tier water price of more
price of more than ¥8. This indicates that increasing the first-tier water price only has an obvious
than ¥8. This indicates that increasing the first-tier water price only has an obvious advantage in
advantage in terms of the water-saving effect when the first-tier water price coverage rate is high
terms of the water-saving effect when the first-tier water price coverage rate is high and loses its
and loses its advantage in water-saving effects when the first-tier water price coverage rate is
advantage in water-saving effects when the first-tier water price coverage rate is low.
low.

22,500
Total annual household water consumption (m3)

5,7
22,400 5,8
5,9
22,300 5,10
5,11
5,12
22,200 5,13
5,14
22,100 5,15
5,16
5,17
22,000 5,18
5,19
21,900 5,20
6,7

21,800
90% 85% 80% 75% 70% 65% 60% 55% 50% 45% 40% 35% 30%

The first-tier water price coverage rate

Figure 12. Change in the total

total annual
annual household
household water consumption under the change in the first-tier
coverage rate
water price coverage rate and
and tiered
tiered water
water price.
price.

In addition, the current tiered water pricing system divides the tiered water consumption on the
basis of annual cumulative water consumption. In Figure 10, the reduction in water consumption
because of the water price entering the second tier starts from the tenth month, that is, individual
water use behavior starts to be constrained by the tiered water price at a later period under Mode Ⅰ,
 Water 2020, 12, 1478 17 of 24

In addition, the current tiered water pricing system divides the tiered water consumption on the
basis of annual cumulative water consumption. In Figure 10, the reduction in water consumption
because of the water price entering the second tier starts from the tenth month, that is, individual
water use behavior starts to be constrained by the tiered water price at a later period under Mode I,
and the water consumption in the previous months is not affected by the tiered water price. Therefore,
we analyzed the impact of two modes of the tiered water pricing system that changes the metering
cycle of water consumption (Mode II, Mode III) on individual water consumption. The results of the
model demonstrated that under any first-tier water price coverage rate and any tiered water price,
the effect of these three modes on promoting individuals’ water-saving behavior is approximately
Mode III > Mode I > Mode II.
We further analyzed the changes in the monthly household water consumption for the year under
the three modes. For example, we considered a case where the first-tier water price coverage rate is
50% and the second-tier water price is increased to ¥13. In Figure 13, under Mode II, some agents with
large water consumption are charged with the second-tier water price in the middle and later period of
the first month. When the R of the first month of the agent is greater than 1%, water consumption
is reduced in the second month. After three months of adjustment, water consumption of the agent
enters a stable state. Therefore, Mode II can enhance the role of the water-saving effect of tiered water
price in the first month and let the agent with large water consumption enter the water-saving state
as soon as possible. Under Mode I, only after the sixth month can the agent pay water charges with
Water the
2020,second-tier water
12, x FOR PEER price, and their water consumption begins to be constrained by the tiered water
REVIEW 18 of 25
price. However, only part of the monthly water consumption is charged based on the second-tier water
price
agents’ R when Mode 1%.
exceeding II is That
adopted, and whole-month
is, Mode I can exert thewater consumption
water-saving mayof
effect bethe
charged
tieredbased
water onprice
the on
second-tier water price when Mode I is adopted; thus, Mode I leads to more agents’
more agents. Accordingly, Mode I has a better water-saving effect than Mode II. Mode Ⅲ makes up for R exceeding 1%.
That is, Mode I can
the shortcomings exert the
of those twowater-saving
modes. Mode effect
Ⅲofcanthe give
tieredfull
water price
play to on
themore agents. Accordingly,
water-saving effect of the
Mode I has a better water-saving effect than Mode II. Mode III makes up for the shortcomings of those
tiered water price earlier and let the water-saving effect of the tiered water price play on more agents;
two modes. Mode III can give full play to the water-saving effect of the tiered water price earlier and
thus, it is the best of the three modes. Considering the seasonal variation in individual water
let the water-saving effect of the tiered water price play on more agents; thus, it is the best of the
consumption
three modes. in real life, if thethe
Considering tiered water
seasonal pricingindivides
variation the water
individual tieredconsumption
water consumption inifterms
in real life, the of
quarterly
tieredcumulative
water pricing water consumption,
divides different
the tiered water divisions
consumption inof tiered
terms water consumption
of quarterly cumulative can waterbe set
for different
consumption, different divisions of tiered water consumption can be set for different seasons such thatmore
seasons such that the water-saving effect of the tiered water pricing can be
scientifically and rationally
the water-saving effect of exerted.
the tiered water pricing can be more scientifically and rationally exerted.

1950
Total monthly household water

Mode Ⅰ Mode Ⅱ Mode Ⅲ

1910
consumption (m3)

1870

1830

1790

1750
1 2 3 4 5 6 7 8 9 10 11 12
Time (month)
Figure 13. Change
Figure in the
13. Change total
in the monthly
total monthlyhousehold
household water consumptionunder
water consumption under three
three modes
modes (when(when
the the
first-tier water coverage rate is 50%, the first-tier water price is ¥5 and the second-tier water price
first-tier water coverage rate is 50%, the first-tier water price is ¥5 and the second-tier water price is is ¥13).
¥13).

4.2. Analysis of Individual Water-Saving Response under the Water Use Information Stimulus
Based on the results of the questionnaire, we know that the amount of good information
feedback is insufficient between Beijing residents and their water use information, and individuals
do not know much about household water consumption and water charges. The results
Water 2020, 12, 1478 18 of 24

4.2. Analysis of Individual Water-Saving Response under the Water Use Information Stimulus
Based on the results of the questionnaire, we know that the amount of good information feedback
is insufficient between Beijing residents and their water use information, and individuals do not
know much about household water consumption and water charges. The results demonstrated that
individuals who know more about water use information are generally more sensitive to water price
factors, and the impact of water price on their water use behavior is greater than that on individuals
who do not know water use information, which is one of the reasons why other studies have posited
that water use information can significantly positively affect water-saving behavior [14]. According to
the questionnaire analysis, we also know that different types of individuals have significant differences
in their attention to and understanding of water use information. The effect of water use information
stimulus on water-saving behavior may be more prominent for individuals who do not focus much on
or know about water use information. Therefore, the effect of promoting water saving differs when
different types of individuals are stimulated by water use information.
Based on this analysis, we then quantitatively compared the implementation of two types of water
use information stimulus (Information I, Information II) with the absence of the water use information
stimulus (No Information) through model simulation. The change in the total monthly household
water consumption in three cases is presented in Figure 14. In Figure 14, the implementation of the
water use information stimulus can promote individual water saving. Compared with No Information,
the implementation of Information I can reduce the total household water consumption by approximately
1.4% over the whole year. The implementation of Information II has a more obvious effect on promoting
individuals’ water-saving behavior, reducing the total annual household water consumption by
approximately 8.5%.
Water 2020, 12, x FOR PEERThis finding occurs because Information I mainly affects the agents with19large
REVIEW of 25
water consumption and with little attention to water use information originally, whereas Information II
thataffect
can exceeds theagents
all the average water
with consumption
water consumptionofthat theexceeds
system,the which results
average in consumption
water the average ofwater
the
consumption
system, which of the system
results decreasing
in the average waterwith the change
consumption in system
of the the agent’s water-saving
decreasing behavior.
with the change in
Therefore,
the agent’s the implementation
water-saving of Information
behavior. II can
Therefore, the form a virtuous
implementation circle of interaction
of Information II can formbetween the
a virtuous
agent’s
circle of water-saving behavior
interaction between theand the water-saving
agent’s system’s water-saving
behavior andlevelthe
tosystem’s
improvewater-saving
the system’slevel
water-to
saving level
improve to a greater
the system’s extent andlevel
water-saving finally
to astabilize the whole
greater extent and system in the state
finally stabilize of a high
the whole water-
system in
saving
the statelevel.
of a high water-saving level.

2000
No Information Information Ⅰ Information Ⅱ
Total monthly household water

1900
consumption (m3 )

1800

1700

1600

1500

1400
1 2 3 4 5 6 7 8 9 10 11 12
Time (month)
Figure 14. Change in the total monthly household water consumption in the three cases.

4.3. Analysis of
4.3. Analysis of Individual
Individual Water-Saving
Water-SavingResponse
Responseunder
underthe
theWater-Saving
Water-SavingPublicity
Publicityand
andEducation
EducationStimulus
Stimulus
The results of the questionnaire demonstrated that Beijing residents are poorly educated about
waterThe
saving, their
results ofunderstanding of the
the questionnaire current situation
demonstrated thatofBeijing
water resources
residents and
are water
poorlypolicies
educatedin Beijing
about
water saving, their understanding of the current situation of water resources and water policies in
Beijing is unsatisfactory, and they have not mastered water-saving methods and skills. Many
respondents in the survey emphasized that strengthening water-saving publicity and education
should be the fundamental measure to promote universal water conservation. Many scholars have
stated that water-saving publicity and education play crucial roles in promoting individuals’ water-
Water 2020, 12, 1478 19 of 24

is unsatisfactory, and they have not mastered water-saving methods and skills. Many respondents
in the survey emphasized that strengthening water-saving publicity and education should be the
fundamental measure to promote universal water conservation. Many scholars have stated that
water-saving publicity and education play crucial roles in promoting individuals’ water-saving
behavior [19,20,40], but elaboration is difficult on the impact process of water-saving publicity and
education on individuals’ water-saving behavior, and few studies have discussed it separately according
to the content of water-saving publicity and education.
Our study compared and analyzed the change in the total monthly water consumption in four cases
of no water-saving publicity and education stimulus (No WSPE) and three types of water-saving publicity
and education stimulus (WSPE I, WSPE II, WSPE III). The model results (Figure 15) demonstrate that
the water consumption reduction in the first month after the implementation of WSPE I is significantly
more than that of WSPE II, the water-saving effect of WSPE I is slightly worse than that of WSPE II
in the second month, and the effect of WSPE I almost disappears seven months later. If WSPE III is
implemented, the water-saving effect in the early period is obvious; that is, the water consumption
reduction in the first month is approximately 2.67 times that in the implementation of WSPE II, and the
water-saving effect in the later period basically depends on the role of WSPE II, which can reduce the
water consumption
Water 2020, approximately 4% compared with No WSPE.
by REVIEW
12, x FOR PEER 20 of 25

3000
Total monthly water consumption

2800

2600
(m3 )

No WSPE
2400
WSPE Ⅰ

2200 WSPE Ⅱ
WSPE Ⅲ

2000
1 2 3 4 5 6 7 8 9 10 11 12
Time (month)
Figure 15. Change in the total monthly water consumption in four cases.

We diddid not
not collect
collect data
data on
on the
the specific
specific impact
impact of of water-saving
water-saving publicity
publicity and education on
individuals’ water-saving
water-savingbehavior
behaviorand andcould
couldnot
not obtain
obtain relevant
relevant datadata from
from other
other research;
research; thus,thus,
we
we could
could notnot accurately
accurately simulate
simulate thethe impactprocess
impact processofofwater-saving
water-savingpublicity
publicity and
and education
education on
individuals’
individuals’water-saving
water-saving behavior. However,
behavior. more more
However, importantly, this research
importantly, this highlights the importance
research highlights the
of conducting
importance ofscenario analyses
conducting rather
scenario than sharp
analyses simulations.
rather than sharpFurther research
simulations. could research
Further collect data on
could
the change
collect datain on
household waterinconsumption
the change household after
waterwater-saving
consumption publicity and education.publicity
after water-saving On this basis,
and
further research
education. on basis,
On this the response
further process
researchofon
individual water-saving
the response process ofunder the water-saving
individual water-savingpublicity
under
and education stimulus
the water-saving forand
publicity several times can
education analyze
stimulus forthe effecttimes
several of various water-saving
can analyze publicity
the effect and
of various
education
water-saving frequencies
publicityonandindividuals’
education water-saving
frequencies on behavior. This water-saving
individuals’ can assess which frequency
behavior. This and
can
assess which
method frequencywater-saving
of conducting and method publicity
of conducting water-saving
and education canpublicity
achieve aand education
better can achieve
water-saving effect
a better
and water-saving
guide water-savingeffect and guide
publicity and water-saving
education work publicity and education work more scientifically.
more scientifically.

4.4. Analysis of
4.4. Analysis of Individual
Individual Water-Saving
Water-Saving Promotion
Promotion Measures
Measures
According
According to
to the
the analysis
analysis of
of the
the impact
impact of
of water
water price
price stimulus
stimulus on
on individuals’
individuals’ water-saving
water-saving
behavior,
behavior, we observed that under the current water pricing system in Beijing, the impact of
we observed that under the current water pricing system in Beijing, the impact of aa water
water
price
price stimulus
stimulus on
on individuals’
individuals’ water-saving
water-saving behavior
behavior isis small. Thus, we
small. Thus, we suggest
suggest aa reduction
reduction inin the
the
first-tier water price coverage rate and an increase in the second-tier water price to take full advantage
of the promotion effect of the tiered water price on individuals’ water-saving behavior. The model
results suggested to reduce the first-tier water price coverage to less than 50%. In real life, first-tier
water consumption can be set as the water consumption necessary to fulfill the basic living needs of
residents according to the actual situation of the region. In addition, the results of the model
Water 2020, 12, 1478 20 of 24

first-tier water price coverage rate and an increase in the second-tier water price to take full advantage
of the promotion effect of the tiered water price on individuals’ water-saving behavior. The model
results suggested to reduce the first-tier water price coverage to less than 50%. In real life, first-tier water
consumption can be set as the water consumption necessary to fulfill the basic living needs of residents
according to the actual situation of the region. In addition, the results of the model demonstrated
that dividing tiered water consumption based on quarterly cumulative water consumption is more
conducive to promoting water-saving behavior. Therefore, we suggest changes in the metering cycle
of water consumption in the current tiered water pricing system from yearly to quarterly. Because of
the substantial differences in individual water consumption in different seasons in real life, we suggest
setting different divisions of tiered water consumption for each season according to the seasonal change
in water consumption.
According to the analysis of the impact of a water use information stimulus on individuals’
water-saving behavior, promoting individuals to procure water use information could effectively
enhance the water-saving effect of water price on individual water use behavior. Therefore, we suggest
an increase in the access to water use information and improvements to the feedback mechanism of
water use information. These include sending an SMS regularly, adding an app reminder function,
installing real-time monitoring and display equipment for household water use, and applying other
means to provide timely feedback on water price, water consumption, and water charges. In addition,
disclosing other individuals’ water use information and increasing individuals’ understanding of the
average water use of the system can effectively stimulate individuals to reduce their water consumption
and make the whole system work to conserve more water. Therefore, when providing feedback
on water use information to individuals, other residents’ water use information can be added for
comparison. Moreover, neighborhood committees, community building management committees,
and other organizations can compare and report the monthly, quarterly, and annual water use of the
community, and praise and reward families with better water-saving behavior or families with greater
progress to strengthen positive emotions for individuals’ water-saving behavior and to create a good
water-saving atmosphere.
According to the analysis of the impact of water-saving publicity and education stimulus on
individuals’ water-saving behavior, the effect of water-saving publicity and education stimulus of
teaching knowledge on individuals’ water-saving behavior in the early period is relatively significant,
but the effect obviously weakens with time. The effect of water-saving publicity and education stimulus
of teaching skills in the early period is not as good as that of water-saving publicity and education
stimulus of teaching knowledge, but its promoting effect on individual water-saving behavior is
stable and lasting. Combining the two types of water-saving publicity and education can improve the
significance and durability of water-saving effects. At present, the situation of water-saving publicity
and education in Beijing is not ideal. Therefore, additional efforts to promote water-saving publicity
and education are necessary. We suggest enriching the content and forms of water-saving publicity and
education and determining the frequency of various water-saving publicity and education efforts based
on the influence of various water-saving publicity and education efforts. Individuals’ water-saving
behavior changes with time, so promoting water-saving publicity and education efforts need to be
more normative and systematic.

5. Discussion
In this study, CAS is applied to study individuals’ water-saving behavior. Based on full
consideration of the interaction between individual attributes and external stimuli, we establish an
individual water-saving response model under external stimuli. By directly stimulating the micro water
use behavior under different stimulus schemes, we predict overall water consumption of the system,
evaluate the water-saving effect of different external stimulus schemes, and provide suggestions for
individual water-saving promotion measures and water demand management. Different from many
model studies, this study also conducts a social survey in the research area and integrates the analysis
Water 2020, 12, 1478 21 of 24

results of the social survey into the establishment and analysis of the model, which effectively solves
the problem of too much subjectivity and the lack of basis when determining parameters in this type
of model. Moreover, this model considers the heterogeneity and mutual influence among micro
individuals and depicts the internal mechanism that affects the change in individuals’ water-saving
behavior. Therefore, the simulation results of the model can more truly reflect the change in the
trend of the regional water-saving situation under the water-saving control measures, which can
provide a reference for research on regional water demand management and have great value in
practical applications.
There are several limitations because of the complexity of the influencing factors and processes of
individual water-saving behavior as well as the difficulty of micro data collection. First, we do not
comprehensively consider the influencing factors of individual water-saving behavior but assume
and simplify some influencing factors in the model. Thus, further research could consider other
key influencing factors (e.g., family characteristics, seasonal changes, and water-saving technology
development) that can be appropriately added according to the actual needs, reduce the assumptions,
and study the influence of each key factor on individuals’ water-saving behavior in a more in-depth
manner. Second, our study has insufficient data support. Limited by the existing basic data, the rules
of some models lack objective data support, such as research on the impact of water-saving publicity
and education on individuals’ water-saving behavior. Notably, to create a water-saving publicity and
education plan, researchers should record the change in individual water consumption with time
after each water-saving publicity and education and analyze the law of the influence of the form and
frequency of water-saving publicity and education on individuals’ water-saving behavior with time.
Furthermore, researchers must use this as the basis to determine the model parameters, improve the
accuracy of the model, and make the research more scientific, rigorous, and objective.
Although there is much room for improvement in research, any method of studying human
behavior has its inherent limitations. The human ability of exercising free will, thinking creatively, and
innovating means that human actions, especially in cases that have not been previously experienced,
are fundamentally difficult to predict accurately [41]. As far as the individual water-saving behavior
studied in this paper is concerned, its influencing factors and processes are complicated, and they will
vary in different periods and different regions. Therefore, we put forward a research idea to grasp
the main problems of individual water-saving efforts in a certain period of the study area through
social surveys, and then combined with regional characteristics to drive modeling decision-making
with the purpose of practical problem-solving, and finally, through model simulation to judge the
effectiveness of different regulatory measures in promoting individual water saving. Managers in
different regions can formulate models based on the regional characteristics and make differentiated
water-saving management policies based on the model’s results.
Further thinking about domestic water management, there is a certain limit for domestic
water-saving; water demand management and water supply management should be considered
at the same time in practical water resources management [42]. Because human beings have a basic
demand for domestic water to fulfill their basic health and hygiene needs and excessively curbing
of domestic water demand will reduce the quality of life, individual domestic water-saving should
be moderate [34,43,44]. Moreover, treating and making good use of domestic sewage can enlarge the
amount of water resources available, and then balance water supply and demand from the overall
allocation of water resources [45]. Therefore, in the future domestic water management and research,
we need emphasize that domestic water-saving should be to save unreasonable consumption of water
resources to scientifically formulate water-saving management policies to curb unreasonable domestic
water use, and also call for strengthening the treatment and reuse of domestic sewage.

6. Conclusions
The management of domestic water demand is one of the key points of water demand management
in areas with high economic development level. In recent years, an increasing number of scholars
Water 2020, 12, 1478 22 of 24

and managers studied the measures of domestic water-saving from the perspective of individual
behavior. This paper explores the feedback interaction between individuals and external factors in the
process of changing water use behavior, with the ultimate goal of proposing individual water-saving
promotion measures for regional individual water-saving problems. A research idea of combining
a social survey and model study was put forward. Furthermore, based on CAS, a general model
framework of individual water saving behavior response research was proposed, and the process of
model establishment and analysis was refined in combination with the actual situation of the region.
This study considers Beijing as the research area. Through the survey of individual domestic
water use in Beijing, we observe three notable problems in the domestic water-saving efforts in
Beijing: unreasonable water pricing, untimely feedback of water use information, and insufficient
water-saving publicity and education. Furthermore, according to the survey results, the relation
among the impact of water price stimulus, water information stimulus, water-saving publicity and
education stimulus on individual water-saving behavior and individual attributes (i.e., gender, age,
education level, income, and water-charge payment frequency) was analyzed. It was found that for
individuals with different attributes, different external stimuli were considered and their water-saving
behavior was affected in different degrees. Therefore, we fully consider the regional demographic
characteristics and the heterogeneity between micro individuals in the model, so that the model has
more regional characteristics. Additionally, aiming at the shortcomings of the current water price
system, water information feedback mechanism, and water-saving publicity and education in Beijing,
this study proposes suggestions on individual water-saving promotion measures by setting different
control schemes and analyzing the water-saving effect under each scheme.

Author Contributions: Conceptualization and methodology, Y.Z., H.L. (Haihong Li) and H.L. (Hanqing Liu);
software, formal analysis, and investigation, H.L. (Hanqing Liu); resources, H.L. (Haihong Li) and L.W.;
data curation, H.L. (Hanqing Liu); writing—original draft preparation, H.L. (Hanqing Liu); writing—review and
editing, L.W., H.L. (Haihong Li), Y.Z., and Q.W.; supervision, Y.Z. and H.L. (Haihong Li); project administration,
Y.Z. and H.L. (Haihong Li); funding acquisition, Y.Z. All authors have read and agreed to the published version of
the manuscript.
Funding: This research was funded by the National Key Research and Development Program of China, grant number
2016YFC0401407; and the Basic Research Fund Project of China IWHR, grant number WR0145B622017.
Acknowledgments: Thanks to Tong Ling of China Agricultural University for her help during the research;
Qian Penghao for his help in programing; and Zhang Xin, Jiang Yasai, Yang Fubo, Chen Yanxin, Li Zhiqi,
Zhang Qiongfang, Jin Rong, Wang Jingjing, and Wang Pu for their help in the questionnaire survey.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Grafton, R.Q.; Ward, M.B.; To, H.; Kompas, T. Determinants of residential water consumption: Evidence and
analysis from a 10-country household survey. Water Resour. Res. 2011, 47, 47. [CrossRef]
2. De Oliver, M. Attitudes and Inaction. Environ. Behav. 1999, 31, 372–394. [CrossRef]
3. Zelezny, L.C.; Chua, P.-P.; Aldrich, C. New Ways of Thinking about Environmentalism: Elaborating on
Gender Differences in Environmentalism. J. Soc. Issues 2000, 56, 443–457. [CrossRef]
4. Tindall, D.B.; Davies, S.; Mauboulès, C. Activism and Conservation Behavior in an Environmental Movement:
The Contradictory Effects of Gender. Soc. Nat. Resour. 2003, 16, 909–932. [CrossRef]
5. Gregory, G.D.; Di Leo, M. Repeated Behavior and Environmental Psychology: The Role of Personal
Involvement and Habit Formation in Explaining Water Consumption1. J. Appl. Soc. Psychol. 2003, 33,
1261–1296. [CrossRef]
6. Randolph, B.; Troy, P. Attitudes to conservation and water consumption. Environ. Sci. Policy 2008, 11,
441–455. [CrossRef]
7. Willis, R.M.; Stewart, R.A.; Panuwatwanich, K.; Williams, P.R.; Hollingsworth, A.L. Quantifying the influence
of environmental and water conservation attitudes on household end use water consumption. J. Environ.
Manag. 2011, 92, 1996–2009. [CrossRef]
Water 2020, 12, 1478 23 of 24

8. Wolters, E.A. Attitude–behavior consistency in household water consumption. Soc. Sci. J. 2014, 51, 455–463.
[CrossRef]
9. Salvaggio, M.; Futrell, R.; Batson, C.D.; Brents, B.G. Water scarcity in the desert metropolis: How environmental
values, knowledge and concern affect Las Vegas residents’ support for water conservation policy. J. Environ.
Plan. Manag. 2013, 57, 588–611. [CrossRef]
10. Aprile, M.C.; Fiorillo, D. Water conservation behavior and environmental concerns: Evidence from a
representative sample of Italian individuals. J. Clean. Prod. 2017, 159, 119–129. [CrossRef]
11. Untaru, E.-N.; Ispas, A.; Candrea, A.N.; Luca, M.; Epuran, G. Predictors of individuals’ intention to conserve
water in a lodging context: The application of an extended Theory of Reasoned Action. Int. J. Hosp. Manag.
2016, 59, 50–59. [CrossRef]
12. Russell, S.V.; Fielding, K.S. Water demand management research: A psychological perspective. Water Resour.
Res. 2010, 46, 46. [CrossRef]
13. Fu, Y.; Wu, W.; Gao, J. Simulation and conservation of the end use water based on behaviour intervention
modelling. Procedia Eng. 2015, 119, 761–770. [CrossRef]
14. Perren, K.; Yang, L. Psychosocial and Behavioural Factors Associated with Intention to Save Water Around
the Home: A Greek Case Study. Procedia Eng. 2015, 119, 1447–1454. [CrossRef]
15. Ramsey, E.; Berglund, E.; Goyal, R. The Impact of Demographic Factors, Beliefs, and Social Influences on
Residential Water Consumption and Implications for Non-Price Policies in Urban India. Water 2017, 9, 844.
[CrossRef]
16. Fan, L.; Wang, F.; Liu, G.; Yang, X.; Qin, W. Public Perception of Water Consumption and Its Effects on Water
Conservation Behavior. Water 2014, 6, 1771–1784. [CrossRef]
17. Arbués, F.; Barberán, R.; Villanúa, I. Price impact on urban residential water demand: A dynamic panel data
approach. Water Resour. Res. 2004, 40, 40. [CrossRef]
18. Savenije, H.; Van Der Zaag, P. Water as an Economic Good and Demand ManagementParadigms with Pitfalls.
Water Int. 2002, 27, 98–104. [CrossRef]
19. Athanasiadis, I.N.; Mentes, A.K.; Mitkas, P.A.; Mylopoulos, Y.A. A Hybrid Agent-Based Model for Estimating
Residential Water Demand. Simul. 2005, 81, 175–187. [CrossRef]
20. Berk, R.A.; Schulman, D.; Mckeever, M.; Freeman, H.E. Measuring the impact of water conservation
campaigns in California. Clim. Chang. 1993, 24, 233–248. [CrossRef]
21. Linkola, L.; Andrews, C.J.; Schuetze, T. An Agent Based Model of Household Water Use. Water 2013, 5,
1082–1100. [CrossRef]
22. Wichman, C.J. Information provision and consumer behavior: A natural experiment in billing frequency.
J. Public Econ. 2017, 152, 13–33. [CrossRef]
23. Holland, J.H. Hidden Order: How Adaptation Builds Complexity; Addison-Wesley: Reading, MA, USA, 1995.
24. Gilbert, N. Agent-Based Models; Sage: London, UK, 2007.
25. Galán, J.M.; Del Olmo, R.; López-Paredes, A. An agent-based model for domestic water management in
Valladolid metropolitan area. Water Resour. Res. 2009, 45, 45. [CrossRef]
26. Yuan, X.-C.; Wei, Y.-M.; Pan, S.-Y.; Jin, J.-L. Urban Household Water Demand in Beijing by 2020:
An Agent-Based Model. Water Resour. Manag. 2014, 28, 2967–2980. [CrossRef]
27. Ma, Y.; Shen, Z.; Kawakami, M.; Suzuki, K.; Long, Y. An Agent-Based Approach to Support Decision-Making
of Total Amount Control for Household Water Consumption. In Advances in Geographic Information Science;
Springer Science and Business Media LLC: Berlin Heidelberg, Germany, 2012; pp. 107–128.
28. Darbandsari, P.; Kerachian, R.; Estalaki, S.M. An Agent-based behavioral simulation model for residential
water demand management: The case-study of Tehran, Iran. Simul. Model. Pr. Theory 2017, 78, 51–72.
[CrossRef]
29. Koutiva, I.; Makropoulos, C. Modelling domestic water demand: An agent based approach. Environ. Model.
Softw. 2016, 79, 35–54. [CrossRef]
30. Li, B.; Sivapalan, M.; Xu, X. An Urban Sociohydrologic Model for Exploration of Beijing’s Water Sustainability
Challenges and Solution Spaces. Water Resour. Res. 2019, 55, 5918–5940. [CrossRef]
31. Beijing Water Authority. Beijing Water Resources Bulletin 2010. Available online: http://swj.beijing.gov.cn/
zwgk/szygb/201912/P020191219479481875049.pdf (accessed on 5 February 2020). (In Chinese)
32. Beijing Municipal Bureau Statistics. Beijing Statistical Yearbook 2018. Available online: http://202.96.40.155/
nj/main/2018-tjnj/zk/indexch.htm (accessed on 5 February 2020). (In Chinese).
Water 2020, 12, 1478 24 of 24

33. Beijing Water Authority. Beijing Water Resources Bulletin 2003-2018. Available online: http://swj.beijing.gov.
cn/zwgk/szygb/ (In Chinese)
34. Sauri, D. Water Conservation: Theory and Evidence in Urban Areas of the Developed World. Annu. Rev.
Environ. Resour. 2013, 38, 227–248. [CrossRef]
35. Lewin, K.; Heider, F.; Heider, G.M. Principles of topological psychology; McGraw-Hill: New York, NY, USA,
1936.
36. Wang, H.; Zhao, Y.; Li, H.; Wang, L.; Jiang, S. Investigation and Analysis on the law of household water use
in Beijing. In Proceedings of the 2016 China (Ningxia) International Forum on efficient utilization of water
resources, Yinchuan, Ningxia, China, 30 May 2016. (In Chinese).
37. Chen, J.; Lu, J.; Wang, X.; Zhang, Y. Analysis on the current situation of water expenditure of urban residents
in the east, middle and west of China. Chin. Price 2017, 8, 29–31. (In Chinese)
38. Ding, F. A Case Study on Practical Curriculum of Public Water-saving Education. Master’s Thesis, East China
Normal University, Shanghai, China, 2015. (In Chinese).
39. Liu, X. A Novel Water Pricing Model of the Utility Analysis for Beijing Urban Residents. Master’s Thesis,
Capital University if Economics and Business, Beijing, China, 2017. (In Chinese).
40. Sharvelle, S.; Dozier, A.; Arabi, M.; Reichel, B. A geospatially-enabled web tool for urban water demand
forecasting and assessment of alternative urban water management strategies. Environ. Model. Softw. 2017,
97, 213–228. [CrossRef]
41. Garcia, M.; Portney, K.; Islam, S. A question driven socio-hydrological modeling process. Hydrol. Earth Syst.
Sci. 2016, 20, 73–92. [CrossRef]
42. Katz, D. Undermining Demand Management with Supply Management: Moral Hazard in Israeli Water
Policies. Water 2016, 8, 159. [CrossRef]
43. Rak, J.; Pietrucha-Urbanik, K. An Approach to Determine Risk Indices for Drinking Water–Study Investigation.
Sustainability 2019, 11, 3189. [CrossRef]
44. Inman, D.; Jeffrey, P. A review of residential water conservation tool performance and influences on
implementation effectiveness. Urban Water J. 2006, 3, 127–143. [CrossRef]
45. Zhao, Y.; Zhu, Y.; Wang, J.; Gao, X.; Li, H.; Zhu, Y.; Jiang, S.; Yong, Z.; Guohua, H.; Jianhua, W.; et al.
Water stress assessment integrated with virtual water trade and physical transfer water: A case study of
Beijing, China. Sci. Total Environ. 2020, 708, 134578. [CrossRef]

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).