Front. Environ. Sci. Eng. 2012, 6(4): 531–539 DOI 10.1007/s11783-010-0287-x

RESEARCH ARTICLE

Analysis of rainfall runoff characteristics from a subtropical urban lawn catchment in South-east China Jinliang HUANG (✉)1, Zhenshun TU3, Pengfei DU2, Qingsheng LI1, Jie LIN1 1 Environmental Science Research Center, Xiamen University, Xiamen 361005, China 2 Department of Environmental Science and Engineering, Tsinghua University, Beijing 100084, China 3 Fujian Institute of Oceanography, Xiamen 361012, China

© Higher Education Press and Springer-Verlag Berlin Heidelberg 2011

Abstract Characteristics of rainfall runoff from a 3.26 hm2 urban catchment with predominant land-use as lawn in Xiamen City, South-east China were investigated and analyzed. Water quality and quantity measurements of rainfall runoff were conducted for ten rainfall events over the period March, 2008 to April, 2009. The results indicated that chemical oxygen demand (COD) and total phosphorus (TP) were the major pollutants with event mean concentrations of 56.09 and 0.44 mg$L–1. From hydrograph and pollutograph analysis of two typical rainfall events, it was clear that the peak rainfall preceded the peak flowrate by about 15–20 min. Meanwhile, concentrations of major pollutants showed multiple peaks and these peaks usually preceded peak flowrate. There were no distinctive first-flush effects except for the rainfall events with the longest rainfall duration and largest runoff volume, which was verified by the fact that the first 30% runoff volume (FF30) carried 39.36% of the total suspended solids (TSS) load, 35.17% of the COD load, 28.13% of the TP load and 39.03% of the nitrate nitrogen load. Multivariate regression analysis further demonstrated that the total runoff volume had a positive correlation with the FF30 of TSS and COD. Keywords rainfall runoff, first flush, pollution characteristics, urban lawn catchment

1

Introduction

Rainfall runoff from urban areas is one of the leading causes of water quality degradation in surface waters [1]. In cities with separate storm and sanitary sewerage Received February 23, 2010; accepted October 25, 2010 E-mail: [email protected]

systems, rainfall runoff causes the primary degradation of streams [2–4]. With urbanization, the land use development to support population increases and movement requires changes to the hydrologic cycle. An increased impervious area results in greater volumes of runoff, higher flow velocities and increased pollutant fluxes to local waterways [5,6]. Urban rainfall runoff pollution sources are diverse and associated with both natural and human activities. This pollution includes precipitation, soil erosion, accumulation and atmospheric wash-off into storm sewers [7]. Therefore, in a small urban watershed with same climate and geology, land use characteristics (such as storage and trade centers in commercial and industrial areas, airport, agricultural property and public parks) reflect anthropogenic activities and could be a major driving force for stream health [4,8]. Recently, many local studies focused on different land use such as residential, industrial and undeveloped watershed [3,9–12], and the different urban surface types including road, roof and parking lots [13–17]. In China, many recent studies have focused on investigating urban stormwater runoff from different surface types such as roads and roofs [18–23]. However, characterizing rainfall runoff from the lawn catchment was seldom reported until now. As the important pervious area of surface, urban lawn suffers from intensive human activities including fertilizer application, parking and recreation. Consequently, rainfall runoff from pervious areas, including lawn, poses a potential threat for receiving water bodies in urban areas [19,20,24,25]. The objective of this study was to investigate the characteristics of rainfall runoff from an urban catchment with predominant land-use as lawn in Xiamen City, South-east China, based on overall runoff quality analysis, hydrograph and pollutograph analysis, first-flush effects analysis and multivariate regression analysis, so as to gain pointers for lawn management in urban areas of China.

532

Front. Environ. Sci. Eng. 2012, 6(4): 531–539

2

Materials and methods

2.1

Description of study catchment

The study catchment, covering 3.26 hm2, is located on the north bank of the Yundang lagoon in Xiamen (Fig. 1). Xiamen has a subtropical monsoon climate. Annual average temperature is 21°C. Annual precipitation averages 1388 mm, of which 80% occurs between April and September, inclusively. Yundang lagoon was built by reclamation in the 1970s. As one of the most important water bodies of Xiamen, Yundang lagoon is famous for its integrated treatment project of reclamation, thus becoming one of the pilot bases for the Regional Programme on Prevention and Management of Marine Pollution in East Asian Seas sponsored by the United Nations Environment Programme. However, the water quality of the lake has become worse than the class IV Seawater Quality Standard of China in recent years, based on the Environmental Quality Bulletin of Xiamen (1996–2007) [26]. The important potential cause is that over one hundred outlets for flood discharge which are distributed along the lake

collect the rainfall runoff and then discharge it directly into Yundang lagoon. The study catchment consists of lawn, road and roof, occupying 65%, 20% and 15% of the area. According to the survey, about 70 vehicles per day park near the study catchment. The sewer network is separate from the runoff system. 2.2

Sample collection and testing

A rain gauge was set up near the study catchment to measure rainfall during rain events. The manual volumetric method and a flowmeter (SIGMA910, HACH, USA) were coupled to measure the rainfall runoff flowrate at the outlet of the catchment. Sampling was carried out using a manual grab with different time-intervals, namely, sampling at 5– 10 min intervals in the first 60 min of rainfall events and then at 30 min intervals for the receding flow stage. Ten rainfall events were monitored and sampled over the period March, 2008 to April, 2009. The general description for the 10 rainfall events monitored is given in Table 1. The samples were collected and analyzed according to

Fig. 1 Location of the study catchment [27]

Jinliang HUANG et al. Rainfall runoff characteristics from urban lawn catchment

Table 1

533

General description of the rainfall events monitored time

duration time/min

rainfall depth/mm

total runoff volume/m3

average rainfall intensity/ (mm$min–1)

antecedent dry weather period/d

No. of samples

2008-03-22

16:29–18:01

92

—

8.69

—

—

16

2008-04-12

10:27–11:32

65

5.45

2.89

0.084

8

12

2008-05-05

9:27–14:13

286

9.37

79.09

0.033

11

27

2008-05-09

15:46–17:23

97

50.7

62.01

0.052

3

13

2008-07-28

15:17–18:12

175

7.2

25.39

0.041

8

22

2008-08-06

10:24–11:53

89

4.16

58.08

0.047

6

17

2008-08-26

18:07–20:42

155

14.7

14.76

0.095

2

17

2009-03-13

18:55–20:45

110

12

85.89

0.109

3

12

2009-03-27

19:22–20:47

85

2.46

9.43

0.029

4

12

2009-04-13

10:37–11:42

65

4

42.34

0.062

2

12

date

the standard methods developed by the State Environmental Protection Agency of China [28]. Water quality parameters included total suspended solids (TSS); ammonia nitrogen (NH3-N), nitrate nitrogen (NO3-N), nitrite nitrogen (NO2-N), total phosphorus (TP) and chemical oxygen demand (COD); copper (Cu), lead (Pb) and zinc (Zn). 2.3

Methods for characterizing urban stormwater runoff

Event mean concentration (EMC) is usually used to characterize pollution concentration during a storm event since it is appropriate for evaluating the effects of stormwater runoff on receiving waters [14]. EMC is a flow-weighted average of constituent concentration and can be calculated as in Eq. (1): tn X

EMC ¼

Ct Qt Δt

t¼t1 tn X

,

(1)

Qt Δt

t¼t1

where t is the time of sample collection; Ct is the pollution concentration collected at t time (mg$L–1); Qt is the runoff flow rate measured in the t time (m3$s–1); Δt is the time interval(s); and n is the number of time intervals. The pollutant loads are always used to weight the pollution mass discharged from a certain area for a specific rainfall event. It is an index for evaluating pollutant mass level and can be calculated as in Eq. (2):

!0 Ct Qt dt tr

L¼

A n–1 X Ci þ Ciþ1

¼

i¼1

2

 A

Qi þ Qiþ1  Δti 2

,

(2)

where tr is the time of rainfall runoff duration time; n is the number of the samples collection; Ci, Qi are the pollution concentration (mg$L–1) and flow rate (m3$s–1) in the ith sample collection; Δt is the time interval(s); and A is the study area (hm2). The initial period of stormwater runoff during which the concentration of pollutants is substantially higher than that during later periods is called the first flush phenomenon [29, 30]. In general, the term “first flush” is used to indicate a disproportionately high delivery of either concentration or mass of a constituent during the initial portions of a rainfall runoff event [31]. Based on the definition of first flush with an M(V) curve [32], which is the curve of the cumulative pollutant load versus the cumulative runoff flow rate of the event (Fig. 2), the 45° line represents the event when the concentration of pollutants remain constant throughout the rainfall runoff (as curve 1 in Fig. 2). Conversely, dilution is assumed to occur when the data are below the 45° line (as curve 3 in Fig. 2). The deviation of the cumulative pollutant mass curve from the diagonal is used as a measure of the strength of the first flush [14,33]. Bertrand-Krajewskin et al. define the first flush with a quantificational criterion: The first flush phenomenon occurs when 80% of the pollution mass load is discharged by at least 30% runoff flow in a rainfall event (FF30) [34]. FF30 can quantitatively measure the first flush phenomenon easily, and it can be calculated as in Eqs. (3) and (4) [5]: T  ð30% – A0 Þ=ðA1 – A0 Þ ¼ TX ,

(3)

where T is the time interval of two sample collections which are next to the 30% flow; A0 is the cumulative flow next before the 30% flow (%); A1 is the cumulative flow next after the 30% flow; and TX is the time interval between the cumulative flow next before the 30% flow and the 30% flow. ½ðB1 – B0 Þ  TX =ðT1 – T0 Þ þ B0 ¼ B30% :

(4)

534

Front. Environ. Sci. Eng. 2012, 6(4): 531–539

in Table 2, overall runoff quality analysis was conducted. According to the Class V surface water standard developed by the China SEPA, water quality parameters of urban rainfall runoff whose mean concentration was beyond the standard were identified as the major pollutants. As shown in Table 2, mean EMCs for COD and TP were 56.09–1 and 0.44 mg$L–1, both beyond the standard. Compared with the study by Davis and McCuen from the other urban surface types [5], rainfall runoff from the study lawn catchment exhibited high concentration of TP, NH3-N, NO3+2-N and COD, which suggested that rainfall runoff pollution from urban lawn catchment should not be neglected. Interestingly, mean EMCs for heavy metals (including Pb, Zn, Cu) of rainfall runoff from the study catchment were all at very low levels. Additionally, the coefficient of variation (CV) for COD, TP, NH3-N and NO3+2-N from the study catchment was great, reflecting significant uncertainties of stormwater quality. This phenomenon was also found in Macau. To some extent, this makes urban stormwater quality control more difficult [20]. Pollution loads could be considered as a measure of the pollution mass emission per unit area. The major pollutant loads for TSS, COD, TP and NO3+2-N in the study catchment were 0.326, 0.671, 0.0047 and 0.025 kg $hm–2, respectively. Recent studies of rainfall runoff from urban green space have mostly focused on its functions for removing pollution [37–40] and water retention [41]. Obviously, characterizing rainfall runoff pollution from lawn catchment also should be considered. Similar to our study, Ren et al. [24] and Huang et al. [19] find that rainfall runoff from a lawn or park catchment in Beijing and Macau both show high COD and TP concentration. It is postulated that there would be high usage of fertilizer for the lawn. Nutrient losses might happen when storms occurs which results in the storm water quality having a high concentration level of nutrients. Whipple et al. also find that TN and TP are mainly sourced from soil losses and leaching of lawn and garden chemicals such as fertilizer application

Fig. 2 Definition of the first flush M(V) curve. Yt the mass fraction, represents the fraction of the total pollutant mass at any time t during the storm event; Xt the volume fraction, represents the fraction of the total runoff volume that has occurred by time t for an event

where T0, T1 are the time of the two sample collections which are before and after the 30% flow; and B0, B1 are the normalized cumulative pollution loads (%) of the two samples which were collected before and after the 30% flow. Multiple variable regression analysis seeks to determine a relationship between several independent variables and one dependent variable by minimizing the total error between the observed data in the proposed regression relationship [35]. As part of this study, multiple variable regression analysis was undertaken to identify the factors influencing the first-flush effect of TSS and COD using several rainfall parameters as independent variables.

3

Results and discussion

3.1

Overall analysis of stormwater runoff quality

A statistical summary of urban rainfall runoff quality for the study catchment is given in Table 2. Based on the data Table 2 pollutants

Statistical summary of rainfall runoff quality in the study catchment EMCs/(mg$L–1)

rainfall events min

max

mean

pollution loads (mean)/(kg$hm–2) CV

EMC of urban or roads [5]/(mg$L–1) min

max

Class V surface water quality standard (GB3838-2002) [35]

TSS

10

5.48

51.27

21.51

0.7824

0.326

0

283

–

COD

10

14.91

111.97

56.09

0.5684

0.671

6

130

£40

TP

7

0.14

1.17

0.44

0.8636

0.0047

0.05

0.52

£0.4

NH3-N

6

0.29

1.38

0.88

0.4659

0.0118

0.22

0.83

£2.0

NO3+2-N

6

0.36

3.79

1.46

0.8151

0.0251

0.14

2.39

–

Cu

4

0.0037

0.0093

0.0057

0.4386

0.00014

0.003

0.142

£1

Pb

4

0.0004

0.0048

0.0021

0.9048

0.00005

0.001

0.33

£0.1

Zn

4

0.0496

0.1253

0.0899

0.3637

0.0012

0.001

1.83

£2

Jinliang HUANG et al. Rainfall runoff characteristics from urban lawn catchment

[42]. COD might have been sourced mainly from fertilizer leaching with soil losses. 3.2

Pollutographs analysis

Combining rainfall, flowrate and runoff quality data in the study catchment produced hydrographs and pollutographs. The hydrographs and pollutographs of major pollutants in two typical rainfall events are presented in Fig. 3. The peak rainfall preceded the peak runoff flowrate by about 15–20 min in these two typical rainfall events. It is understandable that lawn is a permeable surface which has more storage capacity and delays the process of transferring rainfall into runoff. Cheng et al. find that the peak runoff always appears about 20 min later than peak rainfall during heavy rainfall events [43]. Additionally, there was a multiple peak concentration of COD, TP, NH3-N, NO3-N and TSS from the lawn rainfall runoff (Fig. 3). Such a phenomenon was also found by Zhao et al. [44]. Meanwhile, the peak concentration of major pollutants mostly preceded the peak flowrate, as shown in Fig. 3. This may be due to the fact that lawn surface has a strong retention capacity for water and pollutants, and so the pollutants from the stormwater were discharged and flushed discontinuously and slowly. Similar studies in the urban catchments of Wuhan and Macau also demonstrate that maximum peak concentration for major pollutants precedes the peak runoff flowrate [19,21].

3.3

535

First flush effect analysis

To give a general depiction of the first flush phenomenon, the M(V) curve was used in this study. The curves for TSS, COD, NO3-N, and TP were produced for ten rainfall events (Fig. 4). As shown in Fig. 4, the magnitude of the first-flush effects differed between pollutants and rainfall events. With regard to rainfall events, the most dramatic first-flush effects occurred on May 5th, 2008 and March 13th, 2009, and are characterized by the M(V) curves of COD, TSS, TP, and NO3-N as curve 2 in Fig. 2. Compared to the other rainfall events, these two rainfall events showed the longest rainfall duration (May 5th, 2008) and the largest runoff volume (March 13th, 2009), as shown in Table 1. As for the pollutants, TSS seemed to have had more significant first-flush effects than COD, as can be seen in Fig. 4. It should be noted that apart from the two rainfall events mentioned above, there were different first-flush effects for TSS and COD in terms of the other rainfall events. For example, the rainfall event which occurred on April 12th, 2008 exhibited a significant first-flush effect for TSS but not for COD, whereas the rainfall event which occurred on March 27th, 2009 showed a clear and distinctive first-flush effect for COD but not for TSS according to the M(V) curve presented in Fig. 4. Additionally, as for rainfall duration, the rainfall events with long rainfall duration did not always cause a more distinctive first-flush effect. For

Fig. 3 Pollutographs for two typical rainfall events ((a) is May 5, 2008; (b) is for May 9, 2009) in the study catchment

536

Front. Environ. Sci. Eng. 2012, 6(4): 531–539

Fig. 4 M(V) curves for TSS, COD, TP and NO3-N

example, rainfall events with the second and third longest length of rainfall duration (i.e. July 28th, 2008 and August 26th, 2008) showed no first-flush effect. In this sense, there are still some uncertainties concerning the factors influencing first-flush effects. FF30 was further used in this study to measure the firstflush effects quantitatively. Table 3 shows the statistical summary of FF30 values regarding TSS, COD, TP and NO3-N, including the extreme values, the mean values, and the standard deviations (SD) of FF30 for the 10 rainfall events. The mean values of FF30 for TSS, COD, TP and NO3-N were 39.36%, 35.17%, 28.13% and 39.03%, respectively, which are all smaller than 80%, indicating that there were no significant first-flush effects in terms of overall monitored rainfall events. These results also suggested that the order of magnitude of the first-flush effect for different pollutants from lawn runoff was as follows: TSS > NO3-N > COD > TP. This result was basically consistent with the M(V) curve presented in Fig. 4.

First-flush effects analysis enables us to deepen our understanding concerning the characteristics of stormwater runoff discharge In general, the first-flush phenomenon is influenced by many parameters, such as watershed area, rainfall intensity, impervious area, and antecedent dry weather period [14,45]. In this study, multiple regression analysis was conducted so as to identify the factors influencing the first-flush effects in the study catchment. 3.4

Multivariate regression analysis of influencing factors

Statistical Program for Social Sciences (SPSS) 13.0 for Windows was used for multiple variable regression analysis and the calculation of correlation coefficients so as to measure FF30 of TSS and COD, and to identify the factors influencing them. The parameters involved were total rainfall (Tr), flow rate (Fr), total runoff volume (V), rainfall duration (D), maximum rainfall intensity (Imax) and antecedent dry weather period (ADWP) for 10 rainfall

Jinliang HUANG et al. Rainfall runoff characteristics from urban lawn catchment

Table 3

537

Statistical summary of FF30 values regarding TSS, COD, TP and NO3-N TSS

COD

TP

NO3-N

min/%

4.4

7.7

11.3

12.3

max/%

80.56

63.08

38.8

89.3

mean/%

39.36

35.17

28.13

39.03

SD/%

22.72

20.64

9.62

29.33

Table 4

Correlation coefficients between FF30 of TSS and rainfall event characteristics β

p

Tr

1.076

0.655

– 0.031

0.937

Fr

0.453

0.336

0.421

0.259

V

0.677

0.132

0.676

0.046

R

Pr

D

– 0.711

0.436

0.181

0.641

Imax

– 1.147

0.624

– 0.146

0.709

1.706

0.230

0.281

0.465

ADWP

Table 5

Correlation coefficients between FF30 of COD and rainfall events characteristics β

p

R

pr

Tr

1.907

0.689

0.102

0.793

Fr

– 0.605

0.489

– 0.057

0.884

V

0.479

0.472

0.570

0.109

D

– 0.449

0.789

0.132

0.735

Imax

– 2.127

0.647

– 0.062

0.875

ADWP

– 0.261

0.831

– 0.305

0.424

events. The analytical results including the beta coefficients β and their corresponding probability value p; the Pearson partial correlation coefficient R and the probability value Pr are presented in Tables 4 and 5. As shown in Table 4, the partial correlation coefficient R of V for the FF30 of TSS was 0.676 and the probability value Pr 0.046 was smaller than 0.05, indicating that the FF30 of TSS had a significantly positive correlation with V. It can be further seen that the rainfall events with the largest runoff volume were most likely to have a higher FF30 of TSS. This result can be viewed as further proof for the first-flush phenomenon presented by M(V) and FF30 in the above sections. It should be noted that this result is different from those of Deletic [33] and Taebi and Droste [46]. They conclude that the magnitude of the first flush effects is positively and strongly correlated with the maximum rainfall intensity. Our finding is also different from research in Wuhan [21]. That research finds that the magnitude of the first flush of TSS is correlated to ADWP. According to the Pr and R values presented in Table 5, the FF30 of COD also had a slightly positive correlation with V, which is the same as the FF30 of TSS.

4

Conclusions

Chemical oxygen demand (COD) and total phosphorus (TP) were the major pollutants with event mean concentrations of 56.09 and 0.44 mg$L–1. From hydrograph and pollutograph analysis of two typical rainfall events, it is clear that the peak rainfall preceded the peak flowrate by about 15–20 min. Meanwhile, concentrations of major pollutants showed multiple peaks and mostly preceded peak flowrate. There were no distinctive first-flush effects except for the rainfall events with the longest length of rainfall duration and largest runoff volume, which was verified by the fact that the first 30% runoff volume (FF30) carried 39.36% of the total suspended solids (TSS) load, 35.17% of the COD load, 28.13% of the TP load and 39.03% of the nitrate nitrogen load. Multivariate regression analysis further demonstrated that the total runoff volume had a positive correlation with the FF30 of TSS and COD. Acknowledgements This research was supported by the National Natural

538

Front. Environ. Sci. Eng. 2012, 6(4): 531–539

Science Foundation of China (Grant No. 50778098) and the Youth Project of Fujian Provincial Department of Science & Technology (No. 2007F3093).

References 1. US Environmental Protection Agency. National water quality inventory: 1998 Report to Congress. 2000. Available via DIALOG. http://www.epa.gov/305b/98report 2. Walsh C J. Urban impacts on the ecology of receiving waters: a framework for assessment, conservation and restoration. Hydrobiologia, 2000, 431(2–3): 107–114 3. Yusop Z, Tan L W, Ujang Z, Mohamed M, Nasir K A, Tan L W. Runoff quality and pollution loadings from a tropical urban catchment. Water Science and Technology, 2005, 52(9): 125–132 4. He L M, He Z L. Water quality prediction of marine recreational beaches receiving watershed baseflow and stormwater runoff in southern California, USA. Water Research, 2008, 42(10-11): 2563– 2573 5. Davis A P, McCuen R H. Stormwater Management for Smart Growth. New York: Springer Science, 2005, 138–140 6. Grimm N B, Faeth S H, Golubiewski N E, Redman C L, Wu J, Bai X, Briggs J M. Global change and the ecology of cities. Science, 2008, 319(5864): 756–760 7. Novotny V. Water Quality: Diffuse Pollution and Watershed Management. New York: John Wiley and Sons, 2003 8. Zushi Y, Masunaga S. Identifying the nonpoint source of perfluorinated compounds using a geographic information system based approach. Environmental Toxicology and Chemistry, 2009, 28(4): 691–700 9. Lee J H, Bang K W. Characterization of urban stormwater runoff. Water Research, 2000, 34(6): 1773–1780 10. Shah V G, Dunstan R H, Geary P M, Coombes P, Roberts T K, Rothkirch T. Comparisons of water quality parameters from diverse catchments during dry periods and following rain events. Water Research, 2007, 41(16): 3655–3666 11. Kang J H, Lee Y S, Ki S J, Lee Y G, Cha S M, Cho K H, Kim J H. Characteristics of wet and dry weather heavy metal discharges in the Yeongsan Watershed, Korea. Science of the Total Environment, 2009, 407(11): 3482–3493 12. Weston D P, Holmes R W, Lydy M J. Residential runoff as a source of pyrethroid pesticides to urban creeks. Environmental Pollution, 2009, 157(1): 287–294 13. Gromaire M C, Garnaud S, Saad M, Chebbo G. Contribution of different sources to the pollution of wet weather flows in combined sewers. Water Research, 2001, 35(2): 521–533 14. Lee J H, Bang K W, Ketchum L H Jr, Choe J S, Yu M J. First flush analysis of urban storm runoff. Science of the Total Environment, 2002, 293(1-3): 163–175 15. Chang M, McBroom M W, Scott Beasley R. Roofing as a source of nonpoint water pollution. Journal of Environmental Management, 2004, 73(4): 307–315 16. Kim L H, Kayhanian M, Zoh K D, Stenstrom M K. Modeling of highway stormwater runoff. Science of the Total Environment, 2005, 348(1-3): 1–18

17. Kim L H, Ko S O, Jeong S, Yoon J. Characteristics of washed-off pollutants and dynamic EMCs in parking lots and bridges during a storm. Science of the Total Environment, 2007, 376(1-3): 178– 184 18. Zhao J Q, Sun Q Q. Quality characteristics of urban-road runoff and regulation of pollutants discharge. Journal of Chang’an University: Natural Science Edition, 2002, 22(2): 21–23 (in Chinese) 19. Huang J L, Du P F, Ao C T, Lei M H, Zhao D Q, Ho M H, Wang Z S. Characterization of surface runoff from a subtropics urban catchment. Journal of Environmental Sciences, 2007, 19(2): 148– 152 20. Huang J, Du P, Ao C, Ho M, Lei M, Zhao D, Wang Z. Multivariate analysis for stormwater quality characteristics identification from different urban surface types in macau. Bulletin of Environmental and Contaminated Toxicology, 2007, 79(6): 650–654 21. Li L Q, Yin C Q, He Q C, Kong L L. Catchment-scale pollution process and first flush of urban storm runoff in Hanyang, Wuhan City. Acta Scientiae Circumstantiae, 2006, 26(7): 1057–1061 (in Chinese) 22. Dong X, Du P F, Li Z Y, Yu Z R, Wang R, Huang J L. Hydrology and pollution characteristics of urban runoff: Beijing as a sample. Chinese Journal of Environmental Science, 2008, 29(3): 607–612 (in Chinese) 23. Ballo S, Liu M, Hou L, Chang J. Pollutants in stormwater runoff in Shanghai (China): Implications for management of urban runoff pollution. Progress in Natural Science, 2009, 19(7): 873–880 24. Ren Y F, Wang X K, Han B, Ou Z Y, Mao H. Chemical analysis on stormwater-runoff pollution of different underlying urban surfaces. Acta Ecologica Sinica, 2005, 25(12): 3225–3230 (in Chinese) 25. Zhang Z, Fukushima T, Onda Y, Mizugaki S, Gomi T, Kosugi K, Hiramatsu S, Kitahara H, Kuraji K, Terajima T, Matsushige K, Tao F. Characterisation of diffuse pollutions from forested watersheds in Japan during storm events—its association with rainfall and watershed features. Science of the Total Environment, 2008, 390 (1): 215–226 26. Xiamen E P B. Environmental Quality Status Bulletin of Xiamen (1996–2007). Available via DIALOG. http://www.xmepd.gov.cn/sj/ ContentList.aspx?CmsList = 102 (in Chinese) 27. Huang J L, Tu Z S, Du P F, Lin J, Li Q S. Uncertainties in stromwater runoff data collection from a small urban catchment, Southeast China. Journal of Environmental Sciences, 2010, 22 (11): 1703–1709 28. State Environmental Protection Administration of China. Methods of Monitoring and Analyzing for Water and Wastewater. Beijing: China Environmental Science Press, 2002, 276–279 (in Chinese) 29. Gupta K, Saul A J. Specific relationships for the first flush load in combined sewer flows. Water Research, 1996, 30(5): 1244–1252 30. Deletic A B, Maksimovic C T. Evaluation of water quality factors in storm runoff from paved areas. Journal of Environmental Engineering, 1998, 124(9): 869–879 31. Sansalone J J, Cristina C M. First flush concepts for suspended and dissolved solids in small impervious watersheds. Journal of Environmental Engineering, 2004, 130(11): 1301–1314 32. Geiger W. Flushing effects in combined sewer systems. Proceeding of 4th Int. Conf. on Urban Storm Drainage. Lausanne, Switzerland, 1987, 40–46

Jinliang HUANG et al. Rainfall runoff characteristics from urban lawn catchment

33. Deletic A. The first flush load of urban surface runoff. Water Research, 1998, 32(8): 2462–2470 34. Bertrand-Krajewskin J L, Chebbo G, Saget A. Distribution of pollutant mass vs volume in stormwater discharges and the first flush phenomenon. Water Research, 1998, 32(8): 2341–2356 35. McLeod S M, Kells J A, Putz G J. Urban runoff quality characterization and load estimation in Saskatoon, Canada. Journal of Environmental Engineering, 2006, 132(11): 1470–1481 36. State Environmental Protection Agency (SEPA), China. Environmental quality standard for surface water quality standard (GB 3838–2002). 2002. Available at http://english.mep.gov.cn/standards_reports/standards/water_environment/quality_standard/ 37. Lind B B, Karro E. Stormwater infiltration and accumulation of heavy metals in roadside green areas in Goteborg, Sweden. Ecological Engineering, 1995, 5(4): 533–539 38. Larm T. Stormwater quantity and quality in a multiple pond-wetland system: flemingsbergsviken case study. Ecological Engineering, 2000, 15(1-2): 57–75 39. Boving T B, Neary K. Attenuation of polycyclic aromatic hydrocarbons from urban stormwater runoff by wood filters. Journal of Contaminant Hydrology, 2007, 91(1–2): 43–57 40. Read J, Wevill T, Fletcher T, Deletic A. Variation among plant

41.

42.

43.

44.

45.

46.

539

species in pollutant removal from stormwater in biofiltration systems. Water Research, 2008, 42(4–5): 893–902 van Woert N D, Rowe D B, Andresen J A, Rugh C L, Fernandez R T, Xiao L. Green roof stormwater retention: effects of roof surface, slope, and media depth. Journal of Environmental Quality, 2005, 34 (3): 1036–1044 Whipple W, Grigg N S, Grizzard T, Randall C W, Shubinski R B, Tucker L S. Stormwater Management in Urbanizing Areas, Englewood Cliffs, New Jersey: Prentice- Hall, Inc., 1983 Cheng J, Yang K, Xu Q X. Rainfall-runoff storage-infiltration effect of LUCC in highly urbanized region on a catchment’s scale: Shanghai urban green space system as an example. Acta Ecologica Sinica, 2008, 28(7): 2972–2980 (In Chinese) Zhao J W, Shan B Q, Yin C Q. Pollutant loads of surface runoff in Wuhan City Zoo, an urban tourist area. Journal of Environmental Sciences, 2007, 19(4): 464–468 Barbosa A E, Hvitved-Jacobsen T. Highway runoff and potential for removal of heavy metals in an infiltration pond in Portugal. Science of the Total Environment, 1999, 235(1–3): 151–159 Taebi A, Droste R L. First flush pollution load of urban stormwater runoff. Journal of Environmental Engineering, 2004, 3(4): 301– 309