Front. Environ. Sci. Engin.

China 2008, 2(2): 150–156

DOI 10.1007/s11783-008-0014-z
RESEARCH ARTICLE

An integrated assessment method of urban drainage

system: A case study in Shenzhen City, China
Xin DONG, Siyu ZENG ( ), Jining CHEN, Dongquan ZHAO
Department of Environmental Science & Engineering, Tsinghua University, Beijing 100084, China

© Higher Education Press and Springer-Verlag 2008

Abstract In recent years, the urban drainage system in [1] and protect the urban water environment [2]. Yet in China,
China is facing the dual pressure of renovation and construc- due to the rapid urbanization and inexperience in the urban
tion. This requires that the integrated assessment for the drainage system design and construction, pollution from the
planning and operation of the urban drainage system is oblig- discharge of the urban drainage system has been one of the
atory. To evaluate the urban drainage system, an integrated major factors for the deterioration of water environment.
assessment methodology based on the analytic hierarchy pro- The reasons for that lie in many aspects including insufficient
cess (AHP), integrated simulation, and fuzzy assessment is capacity of wastewater treatment plants (WWTPs), a sub-
established. This method is a multi-criteria decision adding stantial number of false connections between sewage and
approach to the assessment of the urban drainage system rainwater drainage pipelines, design of too centralized yet
comprehensively. Through the integration of the Storm Water vulnerable a drainage system as well as lack of an integrated
Management Model (SWMM), a simple wastewater treat- coordination between river systems and WWTPs. Con-
ment plant model, and a surface water quality model, an inte-
sequently, many cities have to start to renovate their mal-
grated modelling system for the urban drainage system is
functioned drainage systems. This also leads to a widespread
developed and applied as a key tool for assessment. Using the
debate regarding how to incorporate sustainable design con-
established method, a case study in Shenzhen City has been
cepts into the renovation plan, rather than simply follow the
implemented to evaluate and compare two urban drainage
system renovation plans, the distributed plan and the central- traditional design prototype, which has been developed more
ized plan. Because of the particularity of this case study, the than 100 years. Different designers may have different under-
established method is not applied entirely. Considering the standings of sustainability and the water problem the city is
water environmental impact, ecological impact, technologi- facing. Practically, for any renovation plan, a large number of
cal feasibility, and economic cost, the integrated performance options could then be proposed to the municipality and there
of the distributed plan is better. As shown in this case study, is thus a need to develop a methodology, by which different
the proposed method is found to be both effective and options could be systematically examined and relevant
practical. information could be fairly provided for decision makers.
In this paper, an integrated assessment methodology based
Keywords analytic hierarchy process (AHP), centralized on the analytic hierarchy process (AHP), urban drainage
plan, distributed plan, fuzzy assessment, integrated system simulation, and fuzzy assessment has been developed
assessment, integrated simulation, urban drainage system for the planning and renovation of the urban drainage system.
Using this integrated method, two renovation plans of
the urban drainage system, in Shenzhen River Catchment,
1 Introduction are compared. And the results of the assessment offer the
important support for decision-making.
The urban drainage system is an important infrastructure
in removing both wastewater and rainwater from city to
prevent unhygienic conditions, avoid damage from flooding 2 Methodology

Integrated assessment of the urban drainage system is a

Received July 13, 2007; accepted September 13, 2007 multi-criteria decision problem, because the performance
E-mail: [email protected] of an urban drainage system involves many dimensions, for
An integrated assessment method of urban drainage system 151
instance, the discharged water environmental quality, and SWMM. Alternatives should be conceptualized as simulation
economic costs [3]. Therefore, a multi-criteria decision scenarios.
adding approach is applied herein, the components and (2) Parameter identification and verification. This is the
assessment procedure of which are shown in Fig. 1. key step in the simulation procedure. It is the guarantee for
The proposed integrated assessment starts from identify- the veracity of the simulation and the rationality of the assess-
ing the system. In this step, all the components of a specific ment. Several important parameters in the integrated model-
drainage system should be taken into account, including ling system, for example, the parameters of the build-up and
the sewer system, the storm water system, the sewer inter- wash-off process in SWMM, the pollution degradation rate in
ception system, the WWTP, and the receiving water body. LRNM, are required for calibration.
Next, the assessment hierarchical model is established (3) Model running and output integration. Based on the
systematically, which is the core part of the assessment. steps (1) and (2), alternatives can be simulated continuously.
Establishing the hierarchical model is to decompose the deci- The model output should be integrated as the correlative
sion problem into elements according to their common char- quantitative attributes.
acteristics and formation of a model with different levels [4].
For those qualitative attributes, for instance, decision
In general, for an urban drainage system, the framework
makers’ preferences, fuzzy numbers are applied to fit into the
should be sufficiently comprehensive to include environ-
numerical framework of the proposed methodology [8]. For
mental, ecological, economic, and technological dimensions
one qualitative attribute, the value of the alternative that
[5]. According to the characteristics of the studied urban
represents best is 1 and the value of the alternative that repre-
drainage system, mostly through a wide consultation, each
dimension could then be decomposed into different criteria sents worst is 0, the values of the rest alternatives are between
and each criterion may consist of a number of different attri- 0 and 1 according to their performance.
butes that are considered important to the sustainable design Based on stakeholders’ preferences as given in graded
of the urban drainage system. numbers (see Table 1), the next step is to calculate the
During this assessment procedure, the most complicated weighting coefficients to different dimensions, criteria, and
step is to calculate the attributes by either numerical simu- attributes by an AHP method [9].
lation or fuzzy assessment.
In this step, an integrated model of the urban drainage Table 1 Graded number for pairwise comparison
system has been developed, which consists of a runoff-sewer comparative judgment graded number
sub-model, a WWTP sub-model, and a surface water quality
sub-model to value quantitative attributes in the hierarchical Ti and Tj are equally important 1
Ti is moderately more important than Tj 3
model [6]. In the integrated model, a runoff and sewer model Ti is strongly more important than Tj 5
SWMM (the Storm Water Management Model) by USEPA Ti is very strongly more important than Tj 7
(U.S. Environment Protection Agency), a simple WWTP Ti is extremely more important than Tj 9
model, and a surface water quality model LRNM (Looped both Ti and Tj represent the indicators the scale 2,4,6 and 8 are used
River Network Model) [7] are used in series to simulate to represent compromises
among the scale 1,3,5,7 and 9
the whole system continually and dynamically under dry
and wet weather conditions. The whole simulation procedure
can be divided into three key steps as illustrated in Fig. 2 as For the integrated assessment of the urban drainage
follows. system, an Integrated Performance Index (I.P.I) is defined.
(1) Input data editing and preparation. With the GIS tools, Based on the weighting coefficients and normalized value of
the information about catchments can be abstracted. Typical attributes, the I.P.I of each alternative urban drainage system
rainfall data should be collected and used as the input of can be calculated by the linear weighting model. The value of

Fig. 1 Procedure of an integrated assessment of urban drainage system

152 Xin DONG, et al.

Fig. 2 Integrated modeling system for urban drainage system

the I.P.I denotes the performance of the alternative. The WWTP will treat most of the wastewater in the catchment
higher I.P.I means the better performance. after it has been enlarged and upgraded to a secondary plant.
In order to convey the wastewater to the Nanshan WWTP, a
sewage tunnel with the largest diameter of 5 000 mm will be
3 Studied area constructed. In 2020, it is predicted that about 1.14x106 m3/d
wastewater will be collected by this tunnel and transported
The Shenzhen River Catchment is a central part of Shenzhen into the Nanshan WWTP. A sewer interception system
City, with 2.7x106 inhabitants in the area of 348.56 km2, will also be added into this plan with the same idea in the
a high urbanization level and a subtropical marine climate. distributed plan [11].
There is a national mangrove nature reserve located in the
Shenzhen Bay. In the catchment, there are one separated
sewer system and four WWTPs. Although a drainage system
has been constructed, the water environment of this area is
still deteriorating gradually. Most of the rivers have water
quality worse than Class V prescribed in the Surface Water
Quality Standard of China (GB3838-2002) [10]. Aiming to
the improvement of urban water environment and ecology,
two renovation plans were put forward in 2005 for this area,
the distributed plan and the centralized plan.
In the distributed plan, apart from those four existing
WWTPs which will be enlarged and upgraded, five new ones
will be constructed (Fig. 3). By 2020, the total wastewater
treatment capacity will have increased to 2.56x106 m3/d
with the wastewater treatment ratio reaching 95%. In order Fig. 3 The distributed plan
to carry out the water reuse program, all the WWTPs but
Nanshan will adopt the advanced treatment processes. For
solving the problem of false pipe connection, an additional 4 Results and discussion
sewer interception system will be built to collect the waste-
water from storm water pipes before it goes into rivers and 4.1 Hierarchy model for case study
convey it to the adjacent WWTP. Compared with the distri-
buted plan, only four WWTPs are involved in the centralized In this case study, a hierarchy model is established based on
plan (Fig. 4). With the capacity of 2.0x106 m3/d, Nanshan the aforementioned methodology. As shown in Fig. 5, there
An integrated assessment method of urban drainage system 153
4.2 Attribute calculation

In this case study, there are two kinds of attributes for assess-
ment, which are quantitative and qualitative. All the attributes
of pollution load, urban water environmental quality, impact
on base flow and urban drainage system cost are quantitative,
and they are valued by the above-mentioned integrated
modelling and the calculation of engineering economy. The
rest of the attributes are qualitative, and they are valued by
fuzzy assessment.
Attributes of water environmental impacts Compar-
ing the pollution load produced in the distributed plan and the
centralized plan (Fig. 6), it is obvious that the former is better.
In the centralized plan, most of the municipal wastewater
Fig. 4 The centralized plan collected in the research area will be treated in the Nanshan
WWTP which is a secondary plant and the huge design capa-
city has restricted the application of the advanced processes.
are eight criteria to describe the four dimensions. For the That is the most important reason for the difference of pol-
water environmental impact dimension, two obligatory crite- lution load emission. The distributed plan has less impact
ria are involved, i.e. pollution load and urban water environ- on the exterior of the Shenzhen River Catchment. Based on
mental quality. The Shenzhen River Catchment is a littoral the simulation results, in both plans, WWTP effluents are the
and there is a national mangrove nature reserve. Considering most important pollution source for urban water environment
the ecological security and sensitive area in the catchment, and the overflow is effectively avoided in the drainage system
the dimension about ecological impact is defined as impact during the wet season. However, the wastewater tunnel will
on mangrove and impact on base flow. Based on the sustain- cause higher probability for overflow in the wet season due
able urban water system principle defined by Hellström et al. to the huge transfer amount which makes the distributed plan
[12], the rationalities of the collection system, treatment as a better option.
system, and reuse system compose the technological feasi- After the water quality simulation, the temporal change of
bility dimension. On the economic cost dimension, the urban water quality on each monitoring section can be examined.
drainage system construction and operation costs are selected Five sections have been chosen to evaluate the state of river
as the criteria. To describe each criterion in detail, there is an environment, which are the Dasha River (D), Fengtang River
attribute system that consists of 18 attributes. (F), upstream of the Shenzhen River (US), middle part of the

Fig. 5 Hierarchy model for Shenzhen case study

154 Xin DONG, et al.

Fig. 6 Total pollution load emission under two plans

Shenzhen River (MS), and downstream of the Shenzhen Comparing the river water quality under the two plans, the
River (DS). The pollutant concentration distributions of median concentrations of chemical oxygen demand (COD)
the five sections under the two renovation plans are shown and total phosphorus (TP) of almost all the rivers are lower in
in Fig. 7. The quartiles of the distribution can be used to eval- the distributed plan. Nevertheless, the rivers in the centralized
uate the river water quality during the whole year. Apart from plan have lower median concentration of total nitrogen (TN).
that, another index called the rate of reaching the standard The same trends can be obtained from RRS. The results are
(RRS) listed in Table 2 can also be used to compare the two ascribed to the different main pollution sources of rivers in
plans. The RRS means the ratio of the days on which the the different plans. As illustrated in Fig. 7, on most of the
water quality is compliant to the Class V in the Surface Water sections, the water quality distributions under the centralized
Quality Standard (GB3838-2002) during the whole year. plan spread more widely than under the distributed plan,
which means the fluctuating of the water quality is stronger
Table 2 Reaching standard rate for rivers under two plans under this plan. This will bring on the worse water environ-
river plan corresponding rate of reaching the
ment quality in the centralized plan. During the rainy day, the
section standard (RRS)/% anti-shock capability of the centralized plan is much lower.
And thus the river system quality is more unstable than that
COD TN TP
in the distributed plan. In the distributed plan, there is more
D distributed 26.6 6.6 18.6 water supplement to the urban rivers; therefore, the rivers
centralized 22.7 12.6 18.6 have larger environmental capacity and anti-shock capacity
F distributed 97.3 4.1 18.4 for pollution.
centralized 94.5 5.2 12.6
US distributed 100.0 0.3 38.9 Attributes of ecological impacts Considering the man-
centralized 99.2 1.4 22.7 grove ecosystem and base flow of rivers together, the distri-
MS distributed 100.0 1.9 30.4 buted plan has more ecological advantages according to the
centralized 97.0 12.6 25.5 quantitative and qualitative assessment (Table 3). According
DS distributed 100.0 24.9 57.0 to the values of the attributes of water environmental impact,
centralized 97.5 45.2 55.9
the distributed plan can maintain better circumstances in

Fig. 7 Water quality of rivers under two plans

An integrated assessment method of urban drainage system 155
the Shenzhen Bay with comparatively suitable salinity, Table 4 Values of attributes of ecological impacts
sufficient nutrients, and clean water, which are necessary for criteria attributes distributed centralized
a good habitat to botanic species and a healthy mangrove eco-
collection system construction 1 (higher) 0 (lower)
system. Enough freshwater replenishment from the drainage
rationality feasibility
system improves the capability of the river system to utilization of 1 (higher) 0 (lower)
resist seawater intrusion in the distributed plan which has a existing system
lower concentration and slighter fluctuation on river salinity treatment system utilization of 1 (higher) 0 (lower)
(Fig. 8). rationality existing system
system reliability 0.5 (equivalent) 0.5 (equivalent)
reuse system construction 1 (higher) 0 (lower)
Table 3 Values of attributes of ecological impact rationality feasibility
system reliability 0.5 (equivalent) 0.5 (equivalent)
criteria attributes distributed centralized

mangrove ecosystem salinity for growth 1 (lower) 0 (higher)

nutrient for growth 1 (more) 0 (less) 4.3 Integrated assessment
water quality 1 (better) 0 (worse)
base flow salinity change 1 (lower) 0 (higher)
Based on the value of the attributes in the hierarchy mode, the
two plans can be compared firstly. As shown in Figs. 6–8
and Tables 2–4, except the equivalent system reliability of
the treatment system and reuse system in the two plans, the
rest of the attributes have absolute predominance in the distri-
buted plan. The distributed plan has a lower pollution load, a
higher RRS, and bigger fuzzy numbers for the attributes of
ecological impacts and technological feasibility and a lower
cost.
As is shown, this case study is specific, and all the attri-
butes in the hierarchy model for the assessment represent that
the distributed plan is better. Thereby, the further steps, weight
calculation and performance index valuation, are not required
in this case study, which made the application of the inte-
grated assessment method easier. In this specific case study,
the synthesis and integrated assessment can be completed by
comparison of assessment attributes.

4.4 Discussion
Fig. 8 Salinity distribution of rivers
In recent years, more and more new conceptions about
Attributes of technological feasibility In the techno- the urban drainage system have emerged from the idea of
logical feasibility dimension, the two plans have been com- the sustainable urban water system. The centralized system
pared with the methods of field investigation, literature cannot occupy the predominant position any longer because
review, and expert consultation, and evaluated by the fuzzy of the huge network system. The distributed system reduces
numbers (Table 4). The distributed plan can make the best of the dependence on the network and decreases the risk of
the existing collection system and treatment system, and have construction and investment. Except the traditional functions,
higher construction feasibility of the collection system and the urban drainage system should also play some new roles,
reuse system. Additionally, there is equivalent reliability including promoting the effective utilization of water, ensur-
for the treatment system and reuse system in the two plans. ing the healthy ecology, and reducing the risk for cities. These
Generally, under the distributed plan, the efficiency of the conflicts between new conceptions and traditional ways are
urban drainage system can be enhanced and the construction obviously shown in this case study.
risk of the drainage system will decrease.
Attributes of economic cost Using the calculation of
engineering economy, the urban drainage system cost which 5 Conclusions
consists of construction cost and operation cost can be valued.
In the centralized plan, the system cost is 5.68x109 RMB, Based on the AHP, integrated simulation, fuzzy assessment,
which is much higher than that of the distributed plan which and expert and stakeholder consultation, an integrated
is 4.47x109 RMB. It is the cost of the big wastewater tunnel assessment method for the urban drainage system is estab-
in the centralized plan which contributes a lot to the higher lished which provides a comprehensive and practical way
investment. to evaluate the system in planning and operation phases. In
156 Xin DONG, et al.
this method, the urban drainage system can be examined
on several dimensions, for example, the environmental, eco-
References
logical, technological, and economic aspects and a modelling 1. Karrman E. Strategies towards sustainable wastewater
system for the urban drainage system is established by inte- management. Urban Water, 2001, 3(1–2): 63–72
grating the sub-models of the compositions of the urban 2. Erbe V, Risholt L P, Schilling W, Londong J. Integrated modeling
drainage system. This is a significant requirement for the for analysis and optimization of wastewater systems—the
sustainable urban water system. Odenthal case. Urban Water, 2002, 4(1): 63–71
3. Malmqvist P A, Heinicke G, Karrman E, Stenstrom T A, Svensson
Using this method, two urban drainage system renovation
G. Strategic Planning of Sustainable Urban Water Management.
plans in Shenzhen City, i.e. the distributed plan and the London, UK: IWA Publishing, 2006, 4–6
centralized plan which are brought forward to improve 4. Chang K F, Chiang C M, Chou P C. Adapting aspects of
urban water environment, are evaluated and compared on GBTool 2005—searching for suitability in Taiwan. Building and
the dimensions of water environmental impact, ecological Environment, 2007, 42: 310–316
impact, technological feasibility, and economic cost compre- 5. Lundin M, Morrison G M. A life cycle assessment based procedure
for development of environmental sustainability indicators for
hensively. The result of the assessment shows the integrated
urban water system. Urban Water, 2002, 4(2): 145–152
performance of the distributed plan has obvious advantages. 6. Dong X, Chen J N, Zhao D Q. Application of SWMM model in
Because of the particularity of this case study, the method urban drainage system. Water and Wastewater Engineering, 2006,
established in this paper is not applied entirely. For the general 32(5): 106–109 (in Chinese)
case, one alternative does not have absolute predominance 7. Xu Y J, Zeng S Y, Zhang T Z. Dynamic water quality model for
in all the assessment attributes; therefore, all the steps of this looping river network based on the framework of uncertainty
analysis: A case study of Wenruitang River Basin. Advances in
method should be implemented during the assessment.
Water Science, 2005, 16(4): 574–580 (in Chinese)
In China, the emergence of many environmental problems 8. Martinez L, Liu J, Yang J B. A fuzzy model for design evaluation
related to the urban drainage system has triggered the system based on multiple criteria analysis in engineering system. Inter-
renovation and reconstruction. These cities are facing the national Journal of Uncertainty, Fuzziness and Knowledge-based
same problems as Shenzhen City. In addition, the construc- Systems, 2006, 14(3): 317–336
tion of the urban drainage system in China is dramatically 9. Lamata M T. An alternative solution to the analytic hierarchy
process. International Journal of Intelligent Systems, 2006, 21:
lagging behind the rapid growth of urban population and eco-
425–441
nomic development. In this situation, the method proposed 10. Wang F, Li Y D. Approach on the sustainable utilization of water
in this paper can provide an effective and comprehensive tool resources in Shenzhen. Chongqing Environmental Science, 1999,
to evaluate alternatives and supply new design prototypes for 21(4): 12–15 (in Chinese)
the system construction. 11. Shenzhen Plan Bureau. Report on Drainage System Distributed
Plan in Shenzhen River Cacthment. Shenzhen: Shenzhen Plan
Acknowledgements This work was supported by the National Basic Bureau, 2005 (in Chinese)
Research Program of China (Grant No. 2006CB403407). The authors would 12. Hellstrom D, Hjerpe M, Moeffaert D V. Indicators to Assess Eco-
like to acknowledge Dr. Xuehua Liu and Dr. Lv Zhou for their contribution logical Sustainability in the Urban Water Sector. Report 2004:3 of
to the ecological and techno-economic impact assessment in the case study. Urban Water. Sweden: Chalmers University of Technology, 2004