Characterization of surface runoff from a subtropics urban catchment.

Abstract

Characteristics of surface runoff from a 0.14-km2 urban catchment with separated sewer in Macau was investigated. Water quality measurements of surface runoff were carried out on five rainfall events during the period of August to November, 2005. Water quality parameters such as pH, turbidity, TSS, COD, TN, Zn, Pb, and Cu were analyzed. The results show that TN and COD are the major pollutants from surface runoff with mean concentration of 8.5 and 201.4 mg/L, both over 4-fold higher compared to the Class V surface water quality standard developed by China SEPA. Event mean concentration (EMC) for major pollutants showed considerable variations between rainfall events. The largest rainfall event with the longest length of antecedent dry weather period (ADWP) produced the highest EMC of TN, TSS and COD. From the pollutographs analysis, the peak concentration of TN precedes the peak runoff flow rate for all three rainfall events. The tendency of the concentration of TSS, turbidity and COD changing with runoff flow varies between rainfall events. The relationship between TSS and other parameters were analyzed to evaluate the efficiency of the physical treatment process to control the surface runoff in the urban catchment. Based on the correlation of parameters with TSS, high treatment efficiency of TSS, TN and COD was expected. The most significant event in term of first flush is the one with the strongest rainfall intensity and longest length of ADWP. TN always showed first flush phenomenon in all three rainfall events, which suggested that the surface runoff in the early stage of surface runoff should be dealt with for controlling TN losses during rainfall events.

Full text

Page 1

Journal of Environmental Sciences 19(2007) 148–152
Characterization of surface runoff from a subtropics urban catchment
HUANG Jin-liang1, DU Peng-fei1,∗, AO Chi-tan2, LEI Mui-heong3,
ZHAO Dong-quan1, HO Man-him1,3, WANG Zhi-shi2
1. Department of Environmental Science and Engineering, Tsinghua University, Beijing 100084, China.
E-mail: jlhuang@tsinghua.edu.cn
2. Faculty of Science and Technology, University of Macau, Macau, China
3. The Official Provisional Municipal Council of Macau, Macau, China
Received 7 March 2006; revised 26 May 2006; accepted 9 June 2006
Abstract
Characteristics of surface runoff from a 0.14-km2urban catchment with separated sewer in Macau was investigated. Water quality
measurements of surface runoff were carried out on five rainfall events during the period of August to November, 2005. Water quality
parameters such as pH, turbidity, TSS, COD, TN, Zn, Pb, and Cu were analyzed. The results show that TN and COD are the major
pollutants from surface runoff with mean concentration of 8.5 and 201.4 mg/L, both over 4-fold higher compared to the Class V
surface water quality standard developed by China SEPA. Event mean concentration (EMC) for major pollutants showed considerable
variations between rainfall events. The largest rainfall event with the longest length of antecedent dry weather period (ADWP) produced
the highest EMC of TN, TSS and COD. From the pollutographs analysis, the peak concentration of TN precedes the peak runoff flow
rate for all three rainfall events. The tendency of the concentration of TSS, turbidity and COD changing with runoff flow varies between
rainfall events. The relationship between TSS and other parameters were analyzed to evaluate the efficiency of the physical treatment
process to control the surface runoff in the urban catchment. Based on the correlation of parameters with TSS, high treatment efficiency
of TSS, TN and COD was expected. The most significant event in term of first flush is the one with the strongest rainfall intensity and
longest length of ADWP. TN always showed first flush phenomenon in all three rainfall events, which suggested that the surface runoff
in the early stage of surface runoff should be dealt with for controlling TN losses during rainfall events.
Key words: surface runoff; event mean concentration (EMC); first flush; Macau
Introduction
Researchers have found that the load and concentration
of suspended solid, nutrients, heavy metals and organic
pollutants discharge from urban surface runoff are higher
than that in untreated polluted and rural areas (Sartor and
Boyd, 1972; Field and Pitt, 1990). These contaminants
will detrimentally impact aquatic organisms and alter
the characteristics of the ecosystem, which may lead to
deterioration of water quality and degradation of stream
habitats in urban areas.
Since the 1960s, international society has come to
realize the urban surface runoff is the major pollution
problem in urban areas. Developed countries such as USA
and European Union paid more attentions to investigate the
characterization of urban surface runoff before developing
surface runoff management planning (Zabel et al., 2001).
In fact, before any planning is done or any practical
steps are taken to control the quality of urban runoff, it
is necessary to first specify the characteristics of urban
surface runoff (Taebi and Droste, 2004). Many researches
Project supported by the Hi-Tech Research and Development Program
(863) of China (No. 2003AA601080). *Corresponding author.
E-mail: dupf@tsinghua.edu.cn.
have been carried out to characterize the surface runoff
pollutants discharged from small catchments of differing
urban surface type, including roofs, highway, and different
land-use catchments (Yaziz et al., 1989; Chang et al.,
2004; Choe et al., 2002; Gnecco et al., 2005; Chebbo
and Gromaire, 2004; Yusop et al., 2005; Deletic, 1998;
Goonetilleke et al., 2005). In recent years, the study on
characterization of highway runoff (Zhao et al., 2001) and
roof runoff (Che et al., 2001; Wang et al., 2005) have
been reported in China. But till now, there are few reports
about the characterization of surface runoff in an urban
catchment environment in China including Macau.
The objective of this study was to investigate the char-
acteristic of surface runoff from an urban catchment with
separated sewer system in Macau, based on the analysis
of the major pollutants from surface runoff, pollutographs
analysis, correlation analysis between TSS and other pa-
rameters, and first flush phenomenon analysis so as to
develop strategy for urban surface runoff management in
Macau.
1 Methods
1.1 Description of study catchment
The YLF catchment is located in the centre of Macau

Page 2

No. 2Characterization of surface runoff from a subtropics urban catchment 149
Fig. 1 Map of the YLF (Yalianfang) urban catchment.
(Fig.1). It covers 0.14 km2and has a impermeability
of 60%. The study catchment is a densely populat-
ed residential/commercial area with 24000 inhabitants.
The catchment mainly consists of residential/commercial
(55.2%) and lawn/park (34.8%) (Fig.1 and Table 1). The
sewer network is separated.
Table 1 Description of study catchment
CharacterizationValue
Area (km2)
Land-use (%)
Factory
School
Residential and commercial
Residential
Lawn/park
Sewer type
Slope (%)
Percent impervious (%)
Stream lengths (m)
Vehicle (d-1)
Population (person)
0.14
0.52
9.45
24.54
30.66
34.83
Separated sewer
7.75
60
868
20000
24000
1.2 Sample collection and testing
Automatic monitoring station was established at the
outlet of study catchment. An automatic event sampler
(ISCO 6712) and rain gauge was equipped to grab samples
and obtain rainfall depth during storms. Sampling was
done at 5–10 min intervals in the first 60 min of storm
events and then 30 min intervals for receding flow stage.
Samplinginthestudycatchmentwascarriedoutduringthe
period from August to November in 2005. At the time of
all samplings, the runoff flow rate was also measured using
ISCO 750 area velocity module excepting for two rainfall
events on 9 August, and on 16 August. Unfortunately, the
flow meter did not work appropriately during these two
rainfall events.
The samples were collected and analyzed according to
APHA standard methods (APHA, 1992). Water quality
parameters included pH, turbidity, Total suspended solid
(TSS), COD, total nitrogen (TN), and heavy metals includ-
ing Zn, Pb, and Cu.
1.3 Characteristic of rainfall events monitored
Fiverainfalleventsweresampledduringtheperiodfrom
August to November, 2005 (Table 2).
Table 2 Characteristic of rainfall events and antecede dry weather
period condition in YLF (Yalianfang) catchment
Rainfall dateDepth
(mm)
Duration
(min)
Average intensity
(mm/h)
ADWP
(h)
2005/08/09
2005/08/16
2005/08/21
2005/08/24
2005/11/14
7.2
15.2
3.4
2.4
11.4
58
81
122
80
121
7.4
11.3
1.7
1.8
5.7
85.4
16.4
9.2
56.7
1072.5
ADWP: antecede dry weather period.
1.4 Calculation of event mean concentration (EMC)
Typically, the concentration of pollutant in surface
runoff is represented as event mean concentration (EMC)
due to the large fluctuation during the rainfall. An EMC
of a pollutant in a specific catchment is obtained from
the division of the total pollutant mass by the total runoff
volume in that event and catchment (Eq. (1)).
EMC =
N ?
N ?
i=1
¯Ci¯ Qi∆ti
i=1
¯ Qi∆ti
=
N ?
N ?
i=1
∆Mi
i=1
∆Vi
(1)
In which N,¯ Qi, Ci, ∆Mi, ∆Viand ∆ti, are number of
samples, average runoff flow rate, average runoff pollutant
concentration, runoff pollutant mass, and runoff volume in
the time interval, respectively.
2 Results and discussion
2.1 Urban surface runoff quality
A statistical summary of urban surface runoff quality for
the catchment study is given in Table 3.
Based on the data in Table 3, the following can be
discerned.
Water quality parameters whose arithmetic mean con-
centration exceeded the Class V surface water standard
developed by China SEPA were identified as the major
pollutants. Arithmetic mean concentration of TN and COD
is 8.5 mg/L and 201.4 mg/L, respectively, all exceed
the Class V surface quality standard by more than 4
factors. The maximum concentration of TN and COD from
surface runoff is 52.0 mg/L and 1274.0 mg/L whereas
their minimum concentration is 0.9 mg/L and 7.3 mg/L,
respectively, which indicates that serious nitrogen and
organic pollutant losses may happen during rainfall events
in YLF catchment. However, the concentration of heavy
metals i.e. Cu, Zn, Pb presents at a low level.
By comparison with the surface water standards for

Page 3

150 HUANG Jin-liang et al.Vol. 19
Table 3 Summary statistics of surface quality of YLF catchment in Macau
ParameterspH Zn (mg/L)Pb (mg/L)Cu (mg/L)TSS (mg/L)TN (mg/L) COD (mg/L)
Maximum
Minimum
Arithmetic mean
Standard deviation
CEPA standard
7.6
6.7
7.2
0.2
6–9
0.185
0.008
0.055
0.039
?2
0.0153
0.0011
0.0032
0.0030
?0.1
0.0248
0.0014
0.0049
0.0040
?1
2600.0
10.0
318.6
525.3
-
52.0
0.9
8.5
10.5
?2
1274.0
7.3
201.4
292.2
?40
Sample numbers for pH, TSS, TN, Cu, Zn, Pb, and COD are 57, 45, 48, 44, 43, 40 and 47, respectively.
China, results showed that surface runoff is badly polluted
especially for TSS, COD, and TN. Yusop et al. (2005)
had similar observation during their research in a tropical
urban catchment with a land-use compose of about 27%
agriculture, 37% residential and commercial and 28%
lawn/parks/bushes. TN mainly sourced from soil loss of
lawn and garden chemical such as fertilizer (Whipple et
al., 1983). There is a park at the upstream of the YLF
catchment (Fig.1). When strong intensity rainfall event
occurred, with soil losses and nitrogen losses happened,
it may make the surface runoff quality high concentration
level of TSS and TN.
EMC for various parameters showed considerable vari-
ations between events (Table 4). The rainfall event on 14
November registered the highest EMC value for major
pollutants. The next highest is for event 24 August and the
lowest for the rainfall event on 21 August. EMC of TN and
COD for event on 14 November is 15.14 mg/L and 464.5
mg/L, are over 7–11 times higher than that of event on 21
August.
EMC from urban surface runoff depends largely on both
build-up and wash-up processes (Butcher, 2003). In this
study, EMCs for TN, TSS and COD have a close relation
with the length of ADWP and the rainfall intensity. The
largest rainfall event on 14 November with the longest
length of ADWP produced the highest TN, TSS and
COD concentration (Table 4). The other investigator also
observed that the number of dry days preceding a rainfall
event significantly affects the quality of runoff water from
the catchment systems (Yaziz et al., 1989).
Table 4 Event mean concentration of 3 rainfall events in YLF
catchment
ParameterEvent mean concentration (EMC)
2005/8/212005/8/24 2005/11/14
Turbidity (NTU)
TSS (mg/L)
TN (mg/L)
COD (mg/L)
17.7
22.34
2.22
41.14
54.52
/
1.96
169.17
60.14
788.34
15.14
464.50
2.2 Pollutographs analysis of surface runoff
Combining runoff quality and flow data at YLF catch-
ments produces pollutographs. The pollutographs of major
pollutants from three rainfall events are showed in Fig.2.
The peak concentration of TN precedes the peak runoff
flow rate in all three rainfall events. TN concentration
shows higher level at the early stage of surface runoff. The
peak concentration of TN appears at the third, first and
fourthsampleforthreerainfallevents,respectively.Ittends
Fig. 2 Pollutographs for 2005/8/21 (a), 2005/8/24 (b) and 2005/11/14 (c) rainfall event in YLF catchment.

Page 4

No. 2Characterization of surface runoff from a subtropics urban catchment151
to decrease with the runoff flow, no matter whether another
flow peak exits or not. Especially for the rainfall event on
24 August, the highest concentration of TN appears at the
first sample and then gradually decreases, showing strong
flushing effects (Fig.2b). This suggests that TN source
might be exhausted and subjected to dilution at the initial
phase of surface runoff.
The tendency of the concentration of TSS, Turbidity and
COD changing with runoff flow varies between rainfall
events. For the rainfall events on 21 August and 24 August,
the peak runoff flow rate precedes the peak concentrations
of TSS, Turbidity and COD (Figs.2a and 2b). On the
contrary, the peak runoff flow rate followed by the peak
concentrations of TSS, Turbidity and COD for the event
on 14 November (Fig.2c).
The rainfall characteristic, ADWP condition and land
cover in the study catchment influences the profile of pol-
lutographs. Lower rainfall intensity (rainfall events on 21
August and on 24 August) made rainfall slowly infiltrate,
generate the surface runoff, and wash off the pollutants
on the surface of the catchment. As mentioned before,
a park located in the upstream of YLF urban catchment
may cause longer time processes. Comparably, the rainfall
event on 14 November with more depth (11.4 mm) and
rainfall intensity (5.7 mm/h) enabled the buildup mass on
the surface to wash off quickly. In additional, it has the
longest length of ADWP, which provided a more buildup
mass in the same street-sweeping condition.
2.3 Relationship between TSS and other parameters
The relationship between TSS and other parameters
were analyzed to evaluate the efficiency of the phys-
ical treatment process and summarized in Table 5.
The degree of correlation was in the following order:
COD>TN>Zn>Pb>Cu. TSS had high correlation with
COD and TN with a high value (>0.85) of the coeffi-
cient of determination R2, which indicates that a close
relationship between soil losses and nitrogen discharge
from urban surface runoff. It could be surmised that most
of COD and TN exist in the form absorbed by soil
particles (Goonetilleke et al., 2005). From a management
perspective, structural surface improvement measures such
as detention basins or sediment traps would be effective
in removing most of the nutrient pollutants and organic
pollutants in the urban surface runoff discharge from the
Table 5 Regression coefficients between TSS and other parameters
ParametersCOD TN Zn PbCu
Regression
coefficients
Sample no.
0.8990.870 0.2170.1320.006
35 3631 3132
YLF catchments.
2.4 First flush phenomenon analysis
The phenomenon known as “the first flush of storm
runoff” usually assumed that the first part of runoff is
most polluted (Deletic, 1998). To assess the first flush,
researchers usually use curves of the cumulative fraction
of total pollutant mass vs. the fraction of total cumulative
runoff volume for the event. Geiger (1987) used the point
of maximum divergence from the 45 slopes to quantify
the first flush (Deletic, 1998). In this study, the first flush
phenomenon was identified by Geiger’s definition.
The curves of major pollutants from YLF catchment
were reproduced for three rainfall events (Fig.3). In-
sufficient data makes any statistical analysis scarcely
significant, although some interesting information can be
obtained by looking at the characteristics of the monitored
rainfall events and the related curves (Gnecco et al., 2005).
The most significant event in term of first flush is the one of
November 14, 2005. Compared to the other rainfall events,
this rainfall event owns the longest length of ADWP and
the strongest rainfall intensity, and both buildup and wash
off processes also are the strongest, which maybe lead to
the significant first-flush effects.
Additionally, TN always showed first flush phenomenon
in all three rainfall events, which suggested that the runoff
in the early stage of surface runoff should be dealt with for
controlling TN losses during rainfall events.
3 Conclusions
TN and COD are the major pollutants from surface
runoff with mean concentration of 8.5 and 201.4 mg/L,
both over 4-fold higher compared to the Class V surface
water quality standard developed by China SEPA.
Event mean concentration (EMC) for major parameters
showed considerable variations between rainfall events.
The largest rainfall event with the longest length of ADWP
produced the highest TN, TSS and COD concentration.
Fig. 3 Cumulative curves of Turbidity, TSS, TN, COD in three rainfall events. (a) 2005/8/21; (b) 2005/8/24; (c) 2005/11/14.

Page 5

152HUANG Jin-liang et al. Vol. 19
TN concentration shows higher level at the early stage
of runoff. The peak concentration of TN precedes the peak
runoff flow rate for all three rainfall events. While the
tendency of the concentration TSS, turbidity and COD
changing with runoff flow varies between rainfall events.
The rainfall characteristic, ADWP condition and land
cover affects the profile of the pollutographs.
The relationship between TSS and other parameters
were analyzed to evaluate the efficiency of the physical
treatment process to control the surface runoff in the urban
catchment. The degree of correlation was in the following
order: COD>TN>Zn>Pb>Cu. Based on the correlation
of parameters with TSS, high treatment efficiency of TSS,
TN and COD was expected.
The most significant event in term of first flush is the
one with the strongest rainfall intensity and longest length
of ADWP. TN always showed first flush phenomenon in
all three rainfall events, which suggested that the runoff in
the early stage of surface runoff should be dealt with for
controlling TN losses during rainfall events.
Acknowledgements: The authors would like to extend
their appreciation to the IACM laboratory for providing
zealous support to analyze the samples.
References
APHA (American Public Health Association), American Water
Works Association, and Water Environment Federation,
1992. Standard methods for the examination of water and
wastewater[M]. Washington, DC: Am Public Health Assoc.
Butcher J B, 2003. Buildup, washoff, and event mean concentra-
tion[J]. Journal of the American Water Association, 39(6):
1521–1528.
Chang M, McBroom M W, Beasley R S, 2004. Roofing as a
source of nonpoint water pollution[J]. Journal of Environ-
mental Management, 73: 307–315.
Che W, Liu H, Wang H et al., 2001. Research on roof runoff
pollution and its utility in Beijing City[J]. China Water and
Wastewater, 17(6): 57–61.
Chebbo G, Gromaire M C, 2004. The experimental urban catch-
ment “Le Marais” in Paris: what lessons can be learn form
it?[J]. Journal of Hydrology, 299: 312–323.
Choe J S, Bang K W, Lee L H, 2002. Characterization of surface
runoff in urban areas[J]. Water Science and Technology,
45(9): 249–254.
Deletic A, 1998. The first flush load of urban surface runoff[J].
Water Research, 32(8): 2462–2470.
Field R, Pitt R E, 1990. Urban storm-induced discharge impacts:
US Environmental Protection Agency Research Program
Review[J]. Water Science and Technology, 22(10/11): 1–7.
Geiger W, 1987. Flushing effects in combined sewer systems[C].
In: Proc. 4th Int. Conf. on Urban Storm Drainage, Lau-
sanne, Switzerland. 40–46.
Gnecco I, Berretta C, Lanza L G et al., 2005. Storm water
pollution in the urban environment of Genoa, Italy[J].
Atmospheric Research, 77: 60–73.
Goonetillekea A, Thomasb E, Ginnc S et al., 2005. Under-
standing the role of land use in urban stormwater quality
management[J]. Journal of Environmental Management,
74: 31–42.
Sartor J D, Boyd G B, 1972. Water pollution aspects of street
surface contaminants[S]. United States Environmental Pro-
tection Agency, Washington, DC, EPA-R2-72-081.
Taebi A, Droste R L, 2004. Pollution loads in urban runoff and
sanitary wastewater[J]. Science of the Total Environmental,
327: 174–184.
USEPA, 1995. Controlling nonpoint source runoff pollution from
roads, highway and bridges[S]. US EPA, Report EPA-841-
F-95-008a, USA.
Wang H, Liu M, Liu H et al., 2005. Primary research on pollution
of urban roof runoff[J]. Resources and Environment in the
Yangtze Basin. 14(3): 367–371.
Whipple W, Grigg S, Gizzard T et al., 1983. Stormwater manage-
ment in urbanizing areas[M]. NJ: Prentice-Hall, Englewood
Cliffs.
Yaziz M I, Gunting H, Sapari N et al., 1989. Variations in
rainwater quality from roof catchments[J]. Water Research,
23(6): 761–765.
Yusop Z, Tan L W, Ujang Z et al., 2005. Runoff quality and
pollution loadings from a tropic urban catchment[J]. Water
Science and Technology, 52(9): 125–132.
Zabel T, Milne I, Mckay G, 2001. Approaches adopted by the
European Union and selected Member States for the control
of urban pollution[J]. Urban Water, 3: 25–32.
Zhao J, Liu S, Qiu L et al., 2001. The characteristics of express-
way runoff quality and pollutants discharge rule[J]. China
Environmental Science, 21(5): 445–448.