Journal of Dynamic Article Links < C

Environmental
Monitoring
Cite this: J. Environ. Monit., 2012, 14, 3164
www.rsc.org/jem PAPER
Trend analysis of a tropical urban river water quality in Malaysia
Faridah Othman,*a Alaa Eldin M. E.b and Ibrahim Mohamedc
Received 15th August 2012, Accepted 12th October 2012
DOI: 10.1039/c2em30676j

Rivers play a significant role in providing water resources for human and ecosystem survival and
health. Hence, river water quality is an important parameter that must be preserved and monitored. As
the state of Selangor and the city of Kuala Lumpur, Malaysia, are undergoing tremendous
development, the river is subjected to pollution from point and non-point sources. The water quality of
the Klang River basin, one of the most densely populated areas within the region, is significantly
degraded due to human activities as well as urbanization. Evaluation of the overall river water quality
status is normally represented by a water quality index (WQI), which consists of six parameters, namely
dissolved oxygen, biochemical oxygen demand, chemical oxygen demand, suspended solids,
ammoniacal nitrogen and pH. The objectives of this study are to assess the water quality status for this
tropical, urban river and to establish the WQI trend. Using monthly WQI data from 1997 to 2007, time
series were plotted and trend analysis was performed by employing the first-order autocorrelated trend
model on the moving average values for every station. The initial and final values of either the moving
average or the trend model were used as the estimates of the initial and final WQI at the stations. It was
found that Klang River water quality has shown some improvement between 1997 and 2007. Water
quality remains good in the upper stream area, which provides vital water sources for water treatment
plants in the Klang valley. Meanwhile, the water quality has also improved in other stations. Results of
the current study suggest that the present policy on managing river quality in the Klang River has
produced encouraging results; the policy should, however, be further improved alongside more
vigorous monitoring of pollution discharge from various point sources such as industrial wastewater,
municipal sewers, wet markets, sand mining and landfills, as well as non-point sources such as
agricultural or urban runoff and commercial activity.

1. Introduction result of increasing water demand from agriculture, industry,

Water is the basic unit of life. Fresh water is a limited resource as
important for industry and agriculture as for human existence.1,2
There is presently an emphasis on diffuse rather than point
sources of water pollution. These adverse effects on water are a
a
Civil Engineering Department, Faculty of Engineering, University of
Malaya, Kuala Lumpur, Malaysia. E-mail: faridahothman@um.edu.my
b
Civil Engineering Department, Faculty of Engineering, University of
Malaya, Kuala Lumpur, Malaysia. E-mail: alaa_elnoor@yahoo.com
c
Institute of Mathematical Sciences, Faculty of Science, University of
Malaya, Kuala Lumpur, Malaysia. E-mail: imohamed@um.edu.my
hydroelectric generation, and continued pollution. The conse-
quences are further exacerbated by population growth, rapid
urbanization and climate change.3,4 From an economic
perspective, water resources are over-extracted and not effi-
ciently allocated. This is due in part to the existence of market
and government failure at the local, national and international
levels. Private costs and benefits diverge from social costs and
benefits, leading to social welfare losses.3,5
The monitoring of environmental parameters is one of the
highest priorities in the evaluation of environmental status of
water resources and in environmental protection policy. The

Environmental impact
The degradation of water quality in surface water and rivers has become an important issue in Malaysia. Most of the urban rivers are
deteriorating in water quality due to the impacts of tremendous development that Malaysia is going through. This paper presents an
evaluation of the water quality for an urban tropical river using the Water Quality Index (WQI). A trend of the WQI has also been
established using a statistical approach. Our results provide an insight into the trend of the WQI and it can be used as an analytical
tool to predict the future conditions of the river water quality.

3164 | J. Environ. Monit., 2012, 14, 3164–3173 This journal is ª The Royal Society of Chemistry 2012
main objectives are to understand and evaluate the water avail-
ability and quality, to control and minimize the incidence of
pollutant-oriented problems, and to provide water of appro-
priate quality to various water uses like urban water supply,
irrigation water, etc.6,7 The quality of surface water and
groundwater is identified in terms of physical, chemical, and
biological parameters. A particular problem with water quality
monitoring is the large number of measured variables. The
datasets contain rich information regarding water resource
behavior. Monitoring data classification, modeling and inter-
pretation are the most important steps in the assessment of water
quality.8
River water is essential for residential use, industries, irrigation
and numerous other activities.9,10 Several reports on river water
quality assessment using physico-chemical and biological
parameters have been published elsewhere.11,12 The different
water quality parameters are expressed in different units. For
example, temperature is expressed in degree Celsius, coliforms in
numbers and most chemicals in milligrams per liter, etc. In other
words, various parameters occurring at different ranges are
expressed in different units and behave in terms of a concentra-
tion–impact relationship. Before an index can be formulated, all
this has to be transformed into a single scale usually beginning
with zero and ending with 1. Some index scales use a 0–100 range.
Computer models are employed extensively for water-quality
management of rivers and streams. These models must typically
be calibrated by adjusting a large number of parameters to attain
optimal agreement between model output and field measure-
ments.13 Mathematical–computational modeling of river water
quality is possible but requires extensive validation. Besides, it
necessitates previous knowledge of hydraulics and hydrody-
namics. To overcome these difficulties, a water quality index
(WQI) was developed.14 WQI is a mathematical instrument used
to transform large quantities of water quality data into a single
number.15 The water quality index contains unit less, single-
dimensional numbers between 0 and 100, with a higher index
value representing good water quality. A numerical index is
utilized as a management tool in water quality assessment.16,17
Water quality can be judged either by individual parameters
for any specific interest or by a few, selected, important param-
eters to judge the overall water quality. Many countries make use
of the WQI method to assess overall river status.18,19 These
indices differ from country to country but share a similar
concept, where a few important parameters are selected and
compounded to numerical rating for the evaluation of river water
quality.20,21
According to the US Atomic Energy Commission standards,22
raw water sources for domestic supply can be categorized as
Group I (excellent, requiring only disinfection), Group II (good,
requiring usual treatment such as filtration and disinfection),
Group III (poor, requiring special or auxiliary treatment and
disinfection) and would, respectively, require BOD ¼ 0.75–1.5
mg L1, 1.5–2.5 mg L1, more than 2.5 mg L1; DO ¼ 4.0–7.5
mg L1, 4.0–6.5 mg L1, 4.0 mg L1; coliform MPN ¼ 50–100/
100 mL, 50–5000/100 mL, more than 5000/100 mL; and
chloride ¼ 50 mg L1 or less, 50–250 mg L1, and more than 250
mg L1. Therefore, a river may be placed in Group I with respect
to chloride, in Group II with respect to coliform MPN, in Group
III with respect to BOD and in no group at all with respect to
This journal is ª The Royal Society of Chemistry 2012
DO. Hence, one can classify the river either in a lower class –
underestimating the river’s utility potential – or in a higher class –
making safety or a higher class value questionable. Otherwise,
the classification would arbitrarily be made based on judgment,
importance of each variable to a use, reliability of monitoring,
etc. To avoid placement in the questionable grade, the river
would tend to be placed in the lower grade, which may not result
in full potential river exploitation.
Malaysia also follows compound WQI to evaluate overall
river water quality. The existing WQI equations are proposed by
the Department of Environment Malaysia (DOE).
Spatial characteristics of land use, vegetative cover, soil, topog-
raphy, and precipitation of the concerned catchment require a tool
that can effectively manage spatial data.23,24 Spatial analysis is
considered one of the main approaches used for the river basin, and
it investigates the different land use types that affect water
quality.25,26 Spatial modeling with Geographic Information System
(GIS) is a proven method, well-documented in many previous
studies.27 The GIS is defined as an organized collection of computer
hardware, software, geographic data and personnel designed to
efficiently capture, store, update, manipulate, analyze and display
all forms of geographically referenced information. In other words,
it is a common purpose technique that handles the geographic data
in a digital format.28,29 There are two types of data that GIS
provides, raster and vector data. In this research the Digital
Elevation Model (DEM) of the study area has been generated and
used as raster data. Generally, ‘‘the Digital Elevation Models are
raster data that composed a matrix of grid cells or pixels, and each
cell or pixel represents the height or the elevation’’. DEMs are
typically used to delineate different terrain parameters such as
watershed boundaries, drainage network, contours, etc., in order to
obtain the elevations and slopes.30,31
Malaysia is blessed with abundant natural resources and a
climate conducive to commercial cultivation of crops such as
rubber and palm oil. In a relatively short period of time,
Malaysia has become a major producer of primary commodities
and the country holds a dominant world position in rubber, palm
oil and cocoa. Concomitant with the emphasis on agricultural
development in the 1960’s and 1970’s, pollution from the agro-
based industries accounted for approximately 90% of the
industrial pollution load and was the largest source of water
pollution during a period when there were inadequate provisions
for regulating the discharge of effluents.32
In Malaysia, about 95% of water comes from the inland river
systems. As the country progresses towards realizing Vision 2020
(i.e., becoming a developed nation) through the implementation
of its policy agenda for heavy industrialization, infrastructures,
and urban-expansions, water demand has increased steeply and
greater pressure is on preserving the current water resources as
well as finding alternative courses of action to improve water
quality.33 Rapid industrialization, although relatively well-plan-
ned and regulated, has been putting increased pressure on urban
areas, especially in the Klang River basin – the most densely
populated area in the country.34
Realizing the seriousness of the threat potentially posed to
river water, Malaysia has made significant efforts to improve the
river water quality in recent years. Environmental management
in Malaysia really began with the enactment of the Environ-
mental Quality Act of 1974 and the subsequent formation of the
J. Environ. Monit., 2012, 14, 3164–3173 | 3165
Department of Environment in 1976. To achieve a more effective
system of environmental management, it was realized that a
preventive strategy was required, with numerous campaigns
launched since 1990. In addition, various studies have been
previously performed on the Klang River basin. However, to our
knowledge, no trend analysis has been carried out on this river
basin.35
In view of this circumstance, the objectives of the present study
are set as: (1) to use a suitable statistical method for trend
analysis purposes on WQI data and (2) to estimate the WQI and
water quality class at any specific station and time.
2. Materials and methods
2.1 Study area
Malaysia is a country with a tropical climate, located in South-
east Asia, with a population of 27 million. It is a fast developing
country, going through tremendous development in a vision of
becoming a developed country by 2020. The study area, i.e. the
Klang River basin, is located within the states of Selangor and
Kuala Lumpur, Malaysia. As the state of Selangor and the city of
Kuala Lumpur are undergoing remarkable development, the
river is subjected to pollution from point and non-point sources.
The water quality of the Klang River basin, one of the most
densely populated areas within the region, is significantly
degrading due to human activities as well as urbanization. The
Klang River drains an area of 1288 km2 from the steep mountain
rainforests of the main central range along Peninsular Malaysia
Fig. 1 Klang River catchment.
3166 | J. Environ. Monit., 2012, 14, 3164–3173
to the river mouth in Port Klang, spanning a distance of 120 km.
Fig. 1 shows the Klang River catchment within the Peninsular
Malaysia map. The basin encompasses the Federal territory of
Kuala Lumpur (Federal territory), parts of Gombak, Hulu
Langat, Klang, Petaling districts in the Selangor State, and the
municipal areas of Ampang Jaya, Petaling Jaya, and Shah Alam.
The Klang River originates in a mountainous area about 25
kilometers (km) northeast of Kuala Lumpur. It is joined by 11
major tributaries while passing through the Federal territory and
the area downstream of Kuala Lumpur, before amalgamating
with the Strait of Malacca at Port Klang.36
2.2 Data collections
Data collection for setting up the model can be roughly divided
into three parts: GIS data, hydrology and hydraulic data, and
water quality data.
2.2.1 GIS data. Spatial characteristics of land use, vegetative
cover, soil, topography, and precipitation of the respective
catchment require a tool that can effectively manage spatial data.
Spatial modeling with GIS is a proven method, well documented
in several previous studies.
The integration of water quality models with the geographical
information system (GIS) enables researchers to carry out
quantitative analysis of spatially distributed pollution
processes. Digital elevation models (DEMs) have become a
main source of topographic data for water quality research. So,
This journal is ª The Royal Society of Chemistry 2012
generating an accurate DEM is very important to validate the
water quality model.
From the contour line map, the digital elevation model for the
Klang River was created using a GIS technique. A GIS is used to
sub-divide the study area into small sub-catchment units based
on DEM information. A series of operations are performed for
pit removal, flow direction assignment, flow-accumulation
determination, and stream-network delineation.
2.2.2 Hydraulics and hydrology data. A hydrographic survey
was done via gauging stations installed along the main river and
its tributaries, along the existing rainfall and stream flow stations
to satisfy the hydraulics and hydrology data requirements. Fig. 2
illustrates the locations of the gauging stations.
2.2.3 Water quality data. Water quality monitoring is an
essential component of water resource conservation, manage-
ment and treatment.37–39 The location selection for monitoring is
a crucial aspect in the network construction for water quality
monitoring. If the location, for example, is not representative of
the river basin, data interpretation and presentation become
inconsequential.40–42 Parameters measured at this site include
pH, dissolved oxygen (DO), conductivity, turbidity, salinity and
temperature. Laboratory analyses were performed to determine
the water quality parameters, i.e. biochemical oxygen demand
(BOD), chemical oxygen demand (COD), total suspended solids
(TSS), nutrients, heavy metals and microbial levels of the water
samples. Fig. 2 depicts the 20 selected station sites. The stations
are labeled as St. 02–18, St. 23, 24 and St. 26.
Stations St. 02–10 are located along the main Klang River with
St. 02 nearest to the strait of Malacca, stations St. 11–13 along
Fig. 2 The locations of the gauging and water quality stations.
This journal is ª The Royal Society of Chemistry 2012
Sg. Damansara, stations St. 14 and 15 along Sg. Penchala and Sg.
Rekah, respectively, St. 16 and 26 along Sg. Kerayong, stations
St. 17, 18 and 24 along Sg. Gombak and station St. 23 along Sg.
Ampang.
2.3 WQI calculation
Six parameters, namely dissolved oxygen (DO), biochemical
oxygen demand (BOD), chemical oxygen demand (COD), sus-
pended solids (SS), ammoniacal nitrogen (AN) and pH, are
considered for the evaluation of overall status of river waters.
DO is defined as the amount of oxygen that dissolved in water
which occurs when microscopic bubbles of gaseous oxygen are
mixed in water. It is considered as one of the most essential water
quality parameters, as it is used for measuring the oxygen
amount that is available for biochemical activity in water.43 BOD
is defined as a measure of the total amount of oxygen removed
from water biologically or chemically in a specified time and at a
specific temperature. It gives an indication of the total DO
concentration which is required throughout some organic matter
degradation and oxidation. The COD is a measure of the
equivalent of the organic matter susceptible to oxidation by
strong chemical oxidants.44 Most applications of COD determine
the amount of organic pollutants found in surface water or
wastewater, making COD a useful measure of water quality. SS
often consist of a large content of suspended organic matter.
Their decomposition also consumes oxygen. AN is defined as the
amount of ammonia and ammonium compounds.45 These
compounds are transferred into the environment out of the
different sources such as waste incineration, sewage treatment,
cattle excrement and car exhausts.46 The pH of surface waters is
J. Environ. Monit., 2012, 14, 3164–3173 | 3167
specified for protection of fish life and to control undesirable
chemical reactions.
The WQI is calculated based on the above six parameters, as
shown in eqn (1). Among them, DO carries a maximum
weightage of 0.22 and pH carries a minimum weightage of 0.12 in
the WQI equation. The WQI equation eventually consists of the
sub-indices calculated according to the best-fit relationships
given in Table 1. The formula used for calculating WQI is:
WQI ¼ 0.22SIDO + 0.19SIBOD + 0.16SICOD + 0.16SISS +
0.15SIAN + 0.12SIpH (1)
In determining the trend of river water quality, intervention
modeling may be employed when different trends are observed at
different intervals. With the main interest in finding the trend, we
used the moving average of the dataset instead of the original
data yt. In doing so, we were also able to estimate the water
quality index at any time point using the moving average values.
Here, the moving average was considered to be of the order 15,
MA(15), and is denoted by y_ t. The trend model for describing a
time series y_ t is given by:
y_ t ¼ TRt + 3t (2)

where WQI ¼ water quality index; SIDO ¼ sub-index of DO; where y_ t is the moving average value in period t, TRt is the trend
SIBOD ¼ sub-index of BOD; SICOD ¼ sub-index of COD; at time period t and 3t is the error term in period t. This model
SIAN ¼ sub-index of AN; SISS ¼ sub-index of TSS; SIpH ¼ basically represents the time series y_ t by an average level, say mt,
sub-index of pH. The sub-index of all the parameters is calcu- that changes over time according to mt ¼ TRt and by the error
lated as shown in Table 1. term 3t. Here, the following is considered:
By identifying the WQI readings, water quality can then be
classified according to the following classes: TRt ¼ b0 + b1t. (3)
Class I: 93 and above.
Class II: from 78 to <93. This is referred to as the 1st order polynomial trend model. By
Class III: from 52 to <78. treating the model as a simple linear regression model,
Class IV: from 31 to <52.
Class V: from 11 to <31. y_ t ¼ b0 + b1t + 3t. (4)
Water quality in Class I is considered safe for direct drinking, ^0 and b
^1
We may obtain the estimates of b0 and b1, say b
in Class II requires treatment for drinking purposes and is safe
respectively, using the least squares estimation method. The upward
for swimming, Class III calls for intensive treatment for drinking, ^1.
(downward) trend is indicated by the positive (negative) sign of b
Class IV is only suitable for plant and domestic animal uses and
Further, similar assumptions as in the simple linear regression
Class V cannot be used for the purposes listed in Classes I–IV.
are made here; the error term 3t satisfies the constant variance,
Water quality categories are highly affected by varying charac-
independence, and normality assumptions. The assumption can
teristics in the surrounding areas.
be checked by constructing appropriate residual plots. Note that
the independence assumption of the errors is often violated when
2.4 Trend analysis time series data being considered as the time-ordered error terms
are commonly autocorrelated. The first-order autocorrelation is
Appropriate statistical tools should be used for the assessment of defined as:
the water quality trends.47–49 By treating WQI as time series data
measured monthly at the stations, we attempted to study the 3t ¼ f13t1 + at (5)
trend analysis using the data moving average values.
such that the error term in time period t, 3t, is related linearly to
the error term in time period t 1, 3t1; f1 is the correlation
coefficient between error terms separated by one time period and
Table 1 Sub-index calculations
a1,a2,. are values randomly and independently selected from a
Sub-index normal distribution having a mean zero and a variance inde-
parameter Value Conditions pendent of time. If f1 > 0, this indicates that the error terms are
positively autocorrelated. On the other hand, if f1 < 0, the error
SIDO 0 DO < 8
100 DO > 92 terms are negatively autocorrelated. In this case, the full model is:
0.395 + 0.030DO2  8 < DO < 92
0.00020DO3 y_ t ¼ b0 + b1t + f13t1 + at. (6)
SIBOD 100.4  4.23BOD BOD < 5
108e0.055BOD  0.1BOD BOD > 5 Again, the assumptions on at will be investigated by con-
SICOD 1.33COD + 99.1 COD < 20
103e0.0157COD  0.04COD COD > 20 structing appropriate residual plots and a normality test such as
SIAN 100.5  105AN AN < 0.3 the Kolmogorov–Smirnov test.
94e0.573AN  5 |AN  2| 0.3 < AN < 4 Based on the moving average values or the estimated ^ y_ t , the
0 AN > 4
SISS 97.5e0.00676SS + 0.05SS SS < 100
WQI can now be estimated at any time point in the time interval,
71e0.0016SS  0.015SS 100 < SS < 1000 and consequently their respective WQI classes.
0 SS > 1000
SIpH 17.2  17.2pH + 5.02pH2 pH < 5.5
242 + 95.5pH  6.67pH2 5.5 < pH < 7 3. Results and discussion
181 + 82.4pH  6.05pH2 7 < pH < 8.75
536  77.0pH + 2.76pH2 pH > 8.75 In this section, the status of the river water quality is presented
according to the WQI, and the WQI trend analysis is explored.

3168 | J. Environ. Monit., 2012, 14, 3164–3173 This journal is ª The Royal Society of Chemistry 2012
3.1 The trend analysis

Monthly water quality data for 20 stations from April 1998 until
November 2007 are considered. The second column of Table 2
gives the time intervals where trend analysis is performed for
each station. For instance, we consider the time series plot for
station St. 02 as shown in Fig. 3. A distinct increasing trend is
clearly observed from March 2003 onward. Hence, we may fit
either model (4) or model (6) to the moving average of
MA(15) values of the data at this interval shown by a green line
in Fig. 3–5.
It is observed that stations St. 02–05 located at the lower
stream of the main river have begun to show a positive trend
from the end of 2002 to early 2003. However, stations St. 06–10
situated at the main river’s upper stream show some trend almost
throughout the time period considered, as illustrated in Fig. 4 for Fig. 4 Time series plot of the St. 06 station.
station St. 06. As for the stations located along the tributaries
(St. 11–24), the trend is initially observed from 2001–2002 as
illustrated in Fig. 5 for St. 16. For station St. 26, available data

Table 2 Results of trend analysis for each station

Station Period with trend Trend value

St. 02 Mar 2003–Nov 2007 0.276

St. 03 Oct 2002–Nov 2007 0.182
St. 04 Feb 2003–Nov 2007 0.211
St. 05 Aug 2002–Nov 2007 0.394
St. 06 Oct 1998–Nov 2007 0.102
St. 07 Aug 1998–Nov 2007 0.060
St. 08 Aug 1998–Nov 2007 0.185
St. 09 July 1998–Nov 2007 0.143
St. 10 Nov 1998–Nov 2007 0.029
St. 11 Sept 2002–Nov 2007 0.157
St. 12 Sept 2002–Nov 2007 0.226 Fig. 5 Time series plot of the St. 16 station.
St. 13 Oct 2002–Nov 2007 0.103
St. 14 Nov 2002–Nov 2007 0.305
St. 15 Dec 2001–Nov 2007 0.217
St. 16 Apr 2002–Nov 2007 0.123
St. 17 Sept 2002–Nov 2007 0.014
St. 18 Sept 2002–Nov 2007 0.026
St. 23 Sept 2002–Nov 2007 0.074
St. 24 Dec 2002–Nov 2007 0.030
St. 26 May 2005–Nov 2007 0.175

Fig. 6 Time series plot of the St. 26 station.

are from May 2005 until November 2007, as plotted in Fig. 6.

Fig. 3 Time series plot of the St. 02 station.
Here, the data are rather short but stationary, so model (2) was
fitted on the original observations, instead of the MA(15) values.
The choice of model (4) or (6) is made based on whether the
assumptions on the residuals are satisfied. For example, we
consider fitting model (4) on MA(15) values from March 2003
until November 2007 for station St. 02, giving a residual plot as

This journal is ª The Royal Society of Chemistry 2012 J. Environ. Monit., 2012, 14, 3164–3173 | 3169
 model the autocorrelated error structure of the time series. The
positive estimate sign of b1 (b^1 ¼ 0.28, p-value ¼ 0) indicates a
positive trend, such that the river water quality at this station has
improved. A similar analysis was carried out for each station in
this study.
The results of trend analysis for each station are given in the
third column of Table 2. The values refer to the estimated slopes
^1 of the fitted trend models. Note that the p-values for all models
b
are very close to zero, suggesting that the trend value is signifi-
cant and not equal to zero. These values are all positive indi-
cating that the quality of river water in the rivers has generally
improved in the specific time period.

3.2 Estimation of WQI

Fig. 7 Residual plot for St. 02 based on model (4). The estimated WQI at any time point can be obtained by taking
the values of MA(15) at a particular time point within the
interval considered for trend analysis. However, for St. 26, the
fitted trend model values are used as WQI estimates since the
data are too short for moving average. The initial and final
estimated WQI values for each station within the interval
considered in the modeling are given in the second and third
columns of Table 3, respectively. The results support the trend
analysis findings, in which the estimated WQI readings are
generally higher in November 2007 than the initial WQI esti-
mates in every station.
We now look at the initial WQI classes of every station. The
Klang River falls under the Class IV category, except for most
upstream stations including St. 10, which was registered as Class
III. As the river flows downstream, the water quality begins to
deteriorate and then falls under the Class IV category. Sg.
Penchala was registered as the worst class, V. The river expe-
Fig. 8 Residual plot for St. 02 based on model (6). riences deteriorating water quality mainly in the middle and
downstream reaches of the river basin. The pollution source
locations and land use categories are presented in Fig. 10. It can
shown in Fig. 7. Obviously, the residuals are positively auto- be clearly seen from this figure that the pollution sources and
correlated and thus the independence assumption of error is
violated. To overcome this problem, we may consider model (6),
giving a residual plot as shown in Fig. 8. The new residuals are
now well-distributed around zero. Furthermore, the SPlus Table 3 WQI classes for each station
output in Fig. 9 suggests that the autoregression term f13t1 Initial moving Final moving Changes of WQI classes
(denoted by AR1 in the output) is significant (t ¼ 4.77, p-value ¼ Station average WQI average WQI (initial class–final class)
0). This implies that the first-order auto-regressive term helps
St. 02 34.81 46.62 4–4
St. 03 38.76 48.36 4–4
St. 04 36.77 48.56 4–4
St. 05 38.44 50.45 4–4
St. 06 43.09 49.52 4–4
St. 07 38.92 54.84 4–3
St. 08 33.94 55.40 4–3
St. 09 43.33 58.65 4–3
St. 10 71.97 77.97 3–2
St. 11 43.77 57.90 4–3
St. 12 45.60 57.09 4–3
St. 13 43.52 54.85 4–3
St. 14 25.69 43.33 5–4
St. 15 46.88 59.37 4–3
St. 16 41.53 50.61 4–3
St. 17 44.99 55.74 4–3
St. 18 57.48 67.54 4–3
St. 23 43.75 59.66 4–3
St. 24 81.48 85.91 2–2
St. 26 39.51 42.49 4–4
Fig. 9 SPlus output based on model (4).

3170 | J. Environ. Monit., 2012, 14, 3164–3173 This journal is ª The Royal Society of Chemistry 2012
 Fig. 10 The sources of pollution with the land use for the Klang River.

urbanization contributing to river water quality decrement are It will be of interest to observe WQI improvements according to
mostly concentrated in the middle and downstream of the river the water quality classes. For the main river, Sg. Klang, the
basin. water quality remains at Class IV for stations located downstream

Fig. 11 The changes of WQI classes.

This journal is ª The Royal Society of Chemistry 2012 J. Environ. Monit., 2012, 14, 3164–3173 | 3171
(St. 2–6). The next three stations further upstream (St. 7, 8, 9) see an
improvement from Class IV to Class III, and the furthest station
(St. 10) moved WQI class towards better, from III to II. For the
tributaries, WQI readings at all DOE stations along Sg. Dam-
ansara (St. 11, 12, and 13), Sg. Rekah (St. 16) and Sg. Ampang (St.
23) improved from Class IV to Class III, while Sg. Penchala (St. 14)
improved from Class V to Class IV. Regarding Sg. Gombak, the
reading at St. 24 (upper stream) remained unchanged at Class II,
while that at St. 17 and 18 (downstream) improved from Class IV
to Class III. Fig. 10 illustrates the sources of pollution with land use
of the Klang River, while in Fig. 11 the changes of WQI classes can
be observed.
3.3 Environmental management
Findings from this study portray that the Klang River water
quality is experiencing an improvement trend since 1998 for the
middle and upstream stretches, and again slightly later in 2002–
2003 for the downstream stretches. The tributaries have started
to show a positive trend in 2001–2003. Water quality enhance-
ment is highly dependent on improved wastewater and sewerage
treatments, other improvement efforts and policies under the
government initiatives, as well as people’s awareness of the need
to preserve our vital water resources. Numerous campaigns and
awareness programs are being promoted by the government
since the 1990’s to recover river water quality. The ‘Love Our
River Campaign’ was launched in 1992 to improve the water-
front and river.50 By far, this is considered the most compre-
hensive long-term awareness program on water resources in
Malaysia. The campaign aimed to promote public awareness of
the importance of rivers and highlight the critical state of river
pollution.51 In addition, six main activities were carried out
during this program, namely River Adoption, River Watch,
River Expeditions, Education (talks, visits, briefings, etc.), River
Beautification and Love Our Rivers Month. The following 7th
Malaysia Plan (1996–2000) not only focuses upon ensuring
more sustainable development through environment preserva-
tion, but also emphasizes the integration of environmental
considerations with the economic and social development
process. Education on environmental awareness and campaigns
was also intensified to ensure active participation from all
sections of society.52
4. Conclusions
This study presents the results of river water quality assessment
using the WQI, and attempts to establish a trend analysis of the
WQI via the statistical approach. The applied methods detected
a statistically increasing trend of the WQI in the downstream of
Klang River’s main stem in the period between August 2002
and November 2007, while the trend started in 1998 for the
upstream reaches. It is interesting to note that the middle rea-
ches improved from Class IV to Class III and some tributaries
also exhibited some improvements in terms of WQI. The
reasons for the progress within this catchment, especially the
tributaries, are most likely related to the proactive policies and
efforts promoted by the Malaysian government. It should be
noted that the downstream remained a Class IV, probably due
to the amount of pollution loading as well as urbanization and
3172 | J. Environ. Monit., 2012, 14, 3164–3173
land development within the area. More efforts need to be made
to improve the river water quality.
Acknowledgements
We would like to express our thanks to the Department of Irri-
gation and Drainage, and the Department of Environment,
Malaysia for their co-operation in performing this study. We would
also like to express our gratitude to the University of Malaya
Research Grant (UMRG) for supporting this study. We are most
grateful and would like to thank the reviewers for their valuable
suggestions, which led to a substantial improvement of the article.
References
1 B. Khalil and T. B. M. J. Ouarda, J. Environ. Monit., 2009, 11, 1915–
1929.
2 J. Bartram and R. Ballance, Water Quality Monitoring: A Practical
Guide to the Design and Implementation of Freshwater Quality
Studies and Monitoring Programmes, Published on behalf of
UNESCO, WHO and UNEP by E&FN Spon, 1996.
3 E. Birol, K. Karousakis and P. Koundouri, Sci. Total Environ., 2006,
365, 105–122.
4 UNEP, Global environmental outlook 2000, Earthscan Publications, 2000.
5 J. L. R. Proops, Ecological Economics, 1991, vol. 3, pp. 263–265.
6 A. Loukas, Desalination, 2010, 250, 266–273.
7 B. H€ulya, Water SA, 2006, 32(3), 389–393.
8 S. Karen, CCME Water Quality Index 1.0 User’s Manual, 2001.
9 M. Kido, Y. Yustiawati, M. Syawal, S. Sulastri, T. Hosokawa,
S. Tanaka, T. Saito, T. Iwakuma and M. Kurasaki, Environ. Monit.
Assess., 2009, 156, 317–329.
10 P. Luo, B. He, K. Takara, B. H. N. Razafindrabe, D. Nover and
Y. Yamashiki, J. Environ. Monit., 2011, 13, 2819–2829.
11 S. M. Avvannavar and S. Shrihari, Environmental Monitoring and
Assessment, 2008, 143, 279–290.
12 M. N. Madhyastha, P. D. Rekha and K. C. Shashikumar, J. Hum.
Ecol., 1999, 10(1), 29–34.
13 G. J. Pelletier, S. C. Chapra and H. Tao, Environ. Model. Software,
2005, 21, 419–425.
14 K. Prakash, L. Seockheon, L. Young-Soo, K. Sushil and K. Siddhi,
Environ. Monit. Assess., 2007, 132, 93–110.
15 S.-G. Nives, Water Res., 1999, 33, 3423–3440.
16 C. G. Cude, J. Am. Water Resour. Assoc., 2001, 37, 125–137.
17 M. Pandey and S. M. Sundaram, Int. J. Ecol. Environ. Sci., 2002, 28,
139–142.
18 D. Chapman and Hall, Water Quality Assessments, E & FN SPON
An Imprint of Chapman & Hall, London, 1st edn, 1992.
19 D. Swaroop Bhargava, Environ. Pollut., Ser. B, 1983, 6, 51–67.
20 A. Almamun and A. Idris, in Twelfth International Water Technology
Conference, IWTC12, Alexandria, Egypt, 2008.
21 P. F. Silvia and D. A. Wunderlin, Water Res., 2000, 34, 2915–2926.
22 USAEC, in Environmental Engineers’ Handbook, ed. D. H. F. Liu and
B. G. Liptak, CRC Press, 1997.
23 J. A. Griffith, Water, Air, Soil Pollut., 2002, 138, 181–197.
24 H. Juahir, S. M. Zain, A. Z. Aris, M. K. Yusoff and M. B. Mokhtar,
J. Environ. Monit., 2009, 12, 287–295.
25 O. Buck, D. K. Niyogi and C. R. Townsend, Environ. Pollut., 2004,
130, 287–299.
26 S. T. Y. Tong and W. Chen, J. Environ. Manage., 2002, 66, 377–393.
27 Y. F. Huang, X. Chen, G. H. Huang, B. Chen, G. M. Zeng, J. B. Li
and J. Xia, Hydrobiologia, 2003, 494, 127–134.
28 ESRI, Understanding GIS: The ARC/INFO Method, ESRI, Redlands,
CA, 1995.
29 M. F. Goodchild, in Environmental Modeling with GIS, ed. M. F.
Goodchild, B. O. Parks and L. T. Steyaert, Oxford University
Press, New York, 1993.
30 S. Antipolis, Int. J. Comput. Vis., 2006, 72(2), 107–132.
31 C. E. Nobel and D. T. Allen, Process Saf. Environ. Prot., 2000, 78,
295–303.
32 DOE, Progress in Malaysia Towards Environmentally Sound and
Sustainable Development 1976–1990, Department of Environment,
This journal is ª The Royal Society of Chemistry 2012
 Ministry of Science, Technology and the Environment, Malaysia,
1991.
33 S. A. Muyibi, A. R. Ambali and G. S. Eissa, Water Resour. Manag.,
2008, 22, 485–508.
34 R. Bradley, Environmentalist, 2010, 30, 228–241.
35 A. R. Abdullah, TrAC, Trends Anal. Chem., 1995, 14, 191–198.
36 A. Development-Bank, ADB Annual report 1996, ed. A. D. Bank,
1996, RRP: MAL 26009, pp. 1–3.
37 K. H. Reckhow, J. A. S. Donigian, J. R. Karr, J. Mandrup-Poulsen,
H. S. Mcdonald, V. Novotny, R. A. Smith and C. O. Yoder, Assessing
the TMDL Approach to Water Quality Management, The National
Academies Press, 2001.
38 L. H. MacDonald, A. W. Smart and R. C. Wissmar, Monitoring
Guidelines to Evaluate Effects of Forestry Activities on Streams in
the Pacific Northwest and Alaska, University of Washington Water
Center, Seattle, Washington, 1991
39 K. Mei, Y. Zhu, L. Liao, R. Dahlgren, X. Shang and M. Zhang,
J. Environ. Monit., 2011, 13, 2755–2762.
40 B. Khalil, T. B. M. J. Ouarda and A. St-Hilaire, J. Environ. Monit.,
2011, 13, 2190–2205.
41 T. G. Sanders, R. C. Ward, J. C. Loftis, T. D. Steele, D. D. Adrian
and V. Yevjevich, Design of Networks for Monitoring Water
This journal is ª The Royal Society of Chemistry 2012
Quality, Water Resources Publications, Littleton, Colorado, 1st
edn, 1983.
42 R. C. Ward, J. C. Loftis and G. B. McBride, Design of Water Quality
Monitoring Systems, Wiley, New York, USA, 1990.
43 Z. G. Ji, Hydrodynamics and Water Quality: Modeling Rivers, Lakes,
and Estuaries, Wiley-Interscience, 2008.
44 W. Viessman and M. J. Hammer, Water Supply and Pollution Control,
Pearson Prentice Hall, the University of Michigan, USA, 7th edn,
2005.
45 M. T. Jafari and T. Khayamian, Talanta, 2008, 76, 1189–1193.
46 A. Mirmohseni and A. Oladegaragoze, Sens. Actuators, B, 2003, 89,
164–172.
47 R. M. Hirsch, R. B. Alexander and R. A. Smith, Water Resour. Res.,
1991, 27, 803–813.
48 D. Jaruskova and I. Liska, J. Environ. Monit., 2011, 13, 1435–1445.
49 A. I. McLeod, K. W. Hipel and B. A. Bodo, Environmetrics, 1991, 2,
169–200.
50 N. S. Abdul-Latip, T. Heath, S. Shamsuddin, M. S. Liew and
K. Vallyutham, in International Conference on Sustainable Building
and Infrastructure (ICSBI), Kuala Lumpur, Malaysia, 2010.
51 N. Ishak and N. Sabanayagam, in New Straits Times, Malaysia, 2006.
52 NAHRIM, Awareness and Advocacy, Malaysia, 2008.
J. Environ. Monit., 2012, 14, 3164–3173 | 3173