Runoff characteristics and nutrient loss mechanism from plain farmland under simulated rainfall conditions

Science of the Total Environment 468–469 (2014) 1069–1077

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

Runoff characteristics and nutrient loss mechanism from plain farmland under simulated rainfall conditions Ruimin Liu ⁎, Jiawei Wang, Jianhan Shi, Yaxin Chen, Chengchun Sun, Peipei Zhang, Zhenyao Shen State Key Laboratory of Water Environment Simulation, School of Environment, Beijing Normal University, No. 19, Xinjiekouwai Street, Beijing 100875, China

H I G H L I G H T S • • • • •

We examined the effect of rainfall events on runoff production and nutrient loss. We examined N and P loss characteristics. N and P losses are more sensitive to the runoff caused by high intensity than by medium rainfall intensity. The high vegetation fraction prevents runoff production and nutrient loss. The high rainfall intensity and soil moisture content contributed to runoff production and nutrient loss.

a r t i c l e

i n f o

Article history: Received 22 May 2013 Received in revised form 12 September 2013 Accepted 13 September 2013 Available online 4 October 2013 Editor: Eddy Y. Zeng Keywords: Non-point source pollution Plain farmland Soil nutrients Loss mechanism Rainfall simulation

a b s t r a c t In recent years, nonpoint source (NPS) pollution has become the main contributor to water quality problems. Research on nitrogen (N) and phosphorus (P) losses from farmland and the factors that influence these losses is very meaningful both for increasing the crop yield and for improving environmental water quality. To explore the mechanism by which N and P are lost from farmland in the North China Plain (NCP), 16 simulated rainfalls were conducted in 14 experimental fields (each of which had different conditions) in the NCP from July to August in 2010. The results showed that the rainfall intensity, the antecedent soil moisture content, and the vegetation cover status were the main factors that affected the surface runoff in the NCP. The runoff volume increased with the increasing rainfall intensity and the increasing soil moisture content, and decreased with the increasing vegetation cover. These factors also significantly affected the losses of P and N. The losses of P and N were positively correlated with the rainfall intensity and the antecedent soil moisture content, and negatively correlated with the vegetation cover. A longer and more intense rainfall resulted in a higher loss of N and P. Dissolved nitrogen was the predominant form of N loss. For phosphorous, the predominant loss form was greatly influenced by the rainfall intensity, the vegetation cover, and the antecedent soil moisture content. Most of phosphorus existed as dissolved phosphorus in Baizhuang (BZ) and as particulate phosphorus in Tangcheng (TC) and Fentai (FT). The minimum requirements for runoff occurrence in experimental regions were a rainfall depth of 5.1 mm, a rainfall intensity of 50 mm/h, and an antecedent soil moisture of approximately 29.6%. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Based on the discharge characteristics in the watershed, water pollution can be divided into two types: point source (PS) pollution and nonpoint source (NPS) pollution (Choi et al., 2011). Compared with PS, NPS is more intermittent and complicated and thus more difficult to control (Dou et al., 2006). Due to the large variety of sources, NPS pollution can occur at any time (Carpenter et al., 1998). It has long been known that agricultural NPS pollution accounts for a large proportion of NPS pollution and plays an important role in the water quality problems (Shortle et al., 2012; Emili and Greene, 2013). In recent years, the presence of ⁎ Corresponding author. Tel./fax: +86 10 5880 0829. E-mail address: [email protected] (R. Liu). 0048-9697/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.scitotenv.2013.09.035

NPS pollution due to agricultural activities in China has been of concern to the public (Wu et al., 2012). Many studies have indicated that the typical NPS pollution occurs in the process of rainfall–runoff (Zheng and Wang, 2002) and that more than half of the contaminants enter the water through this process (G.H. Zhang et al., 2011; H. Zhang et al., 2011). The premise of NPS pollution control has been the use of models to accurately calculate the load and describe the distribution; therefore, the identification of reasonable model parameters is crucial to obtain an accurate calculation. Many NPS models require a large number of input data and their applications are restricted by the different regional parameters. However, in some case when the detailed parameters are in shortage, parameter estimation methods or parameter simplification methods can be used to ensure that models are running well (Zhang

1070

R. Liu et al. / Science of the Total Environment 468–469 (2014) 1069–1077

Fig. 1. The location of the study area.

et al., 2007; Shen et al., 2012; R.M. Liu et al., 2013; W.B. Liu et al., 2013). However, due to the large amount of parameters contained in most NPS models and the fact that various parameters are sensitive to weather, soil types, and land use in different regions, different parameters may result in significant changes in the predicted results (Gong et al., 2011; Shen et al., 2012; Strauch et al., 2012). To obtain reasonable model parameters, it is best to go conduct a large number of field experiments to discover the loss mechanism (Shen et al., 2012). Field experiments can clearly simulate the natural situation and are widely used in the study of environmental problems. NPS pollution exhibits a significant relationship with rainfall, runoff, soil types, and soil cultivation. Thus, using field experiments to research NPS pollution is meaningful and effective (Liao et al., 2011). Rainfall is the key parameter in most field experiments because it is the main driving force of surface runoff (Dan et al., 2001). Field experiments usually involve a long experiment period, and natural rainfall involves many uncontrollable factors. Thus, it is difficult to achieve the ideal result. In contrast, a rainfall simulation experiment could be used to simulate different intensities of natural rainfall, to shorten the experiment period, and to control the experimental conditions for easy observation of the runoff occurrence and evolution (Zheng and Zhao, 2004; Cheng et al., 2008). At present, the hot spot in NPS pollution involves rainfall simulations to research the mechanism of NPS pollution and provide accurate parameters for NPS models (Jin et al., 2009; Y. Chen et al., 2012; Y.X. Chen et al., 2012; Jacinthe et al., 2002).

Table 1 Soil characteristics in BZ, TC and FT. Soil texture

BZ TC FT

Tidal soil Meadow cinnamon soil Calcareous cinnamon soil

Total N (g/kg)

Total P (g/kg)

Total K (g/kg)

Total C (g/kg)

Available P (mg/kg)

Available K (mg/kg)

Organic matter (g/kg)

0.48 0.35

1.30 1.71

18.48 18.23

16.06 11.89

6.92 24.02

196.41 143.14

9.59 7.08

0.24

1.02

17.62

6.17

44.37

156.61

4.72

The North China Plain (NCP) is one of the most important agricultural production regions in China. With a dominant winter wheat–summer maize double cropping system, this region provides more than 50% of the nation's wheat and approximately 33% of its maize production (Wang, Guo, et al., 2012; Wang, Wang, et al., 2012). Extensive studies have shown that an excessive use of fertilizer was very common in the NCP, particularly in regions with a high yield and high population densities, such as the plain area of the Baiyangdian watershed (R.M. Liu et al., 2013; W.B. Liu et al., 2013). The agricultural NPS pollution in the NCP was the main pollution source (Sun et al., 2010; Li et al., 2012). However, at present, it is difficult to estimate the NPS pollution load in the NCP using an NPS model because the rainfall–runoff characteristics are deeply affected by human activities. Because most of the lands are used as agricultural lands and are managed in village-scale units, the mechanism and load of the villagescale NPS pollution in the NCP are necessary prerequisites to improve the water quality in the plain areas (Cui, 2011). The existing studies that used watershed model simulations or the hydrological segmentation method to explore the village-scale NPS pollution characteristics were only performed in some southern key river watersheds in China (such as the Taihu watershed, the Dianchi watershed, and the Three Gorges Reservoir Area) and only estimated the pollution loads (Long et al., 2010; Zeng et al., 2010; Sun et al., 2012). Thus, the research on the characteristics of NPS pollution at the village-scale in the NCP is significant. In this study, 16 different artificial rainfall conditions in the NCP were simulated to analyze the loss characteristics of nitrogen and phosphorus under different rainfall conditions and the field management measures from July to August of 2010. Based on the mechanism of nutrients loss, accurate simulation parameters could be provided for the NPS models, which will prove helpful for accurate simulations and effective management. 2. Materials, methods, and experimental design 2.1. Experimental site The North China Plain, which is China's second largest plain, extends in latitude from approximately 32°N to 40°N and in longitude from

R. Liu et al. / Science of the Total Environment 468–469 (2014) 1069–1077

4

1071

Zone 4m

3m

The Experimental Zone

(c) Rainfall simulator

(a) Field experimental design

Zone

Runoff collecting tank 10cm Soil

Collection tube

Bedrock Collection barrel

(b) Runoff collector Fig. 2. The schematic diagram of the field experiments, including (a) field experimental design, (b) runoff collector, and (c) rainfall simulator.

approximately 114°E to 121°E. According to historical weather records of the past 50 years, the mean annual temperature is between 11 °C and 15 °C, and the mean annual precipitation is 500–1000 mm (Long et al., 2010; Zeng et al., 2010). The field experiments were conducted near the villages of Baizhuang (BZ), Tangcheng (TC), and Fentai (FT) in the plain area of the Baiyangdian watershed, which is located in the center of the NCP (Fig. 1). In general, soil properties can affect the NPS pollution, which includes the soil texture, the saturated soil moisture content, the organic matter, and the particle-size distribution (Dreelin et al., 2006; Ouyang et al., 2010). The soil in the southeast plain was mostly formed by alluvial flood, including tidal soil, meadow cinnamon soil, and calcareous cinnamon soil, and they all had high soil nutrition (Yan and Sun, 2008). The

soil characteristics of the three experimental sites are provided in Table 1. The typical cropping system in the region was rotational winter wheat and summer maize cultivation. After winter wheat is harvested in early June, local farmers bury the crushed straw into the surface soil, and then in late September to early October they plant summer maize. 2.2. Methods and experimental design 16 simulated rainfalls were conducted in 14 experimental plots in the NCP from July to August in 2010. A lot of indexes were used to research the rainfall–runoff process, including the runoff coefficient, the duration of the rainfall, the runoff generation time, the runoff duration, the initial loss of rainfalls, the runoff volume, the vegetation fraction and

Table 2 Characteristic of rainfall–runoff in BZ, TC and FT.

BZ

TC FC

Plot

Antecedent soil moisture content/%

Plant height (maize)/m

Vegetation fraction/%

Fertilizer levels (N)/kg/ha−1

Rainfall intensity/mm/h

Duration of rainfall/min

Runoff generated time/min

Runoff duration time (after rain)

Initial loss of rainfall

Runoff coefficient %

1 3 4 5 6 7-1 7-2 8 9 10 11 12 TC-1 TC-2 XS-1 XS-2

43.6 39.4 37.2 38.6 35.2 38.9 45 33.4 35.4 39.1 39 36.7 28.9 36.5 27.9 31.9

2.1 2.6 1.8 2.6 2.4 2.7 2.7 2.8 2.9 3.0 3.1 3.2 0 0 0 0

85 85 45 65 65 75 75 85 85 90 90 90 0 0 0 0

0 99 165 99 0 165 165 99 0 165 99 0 0 0 0 0

30 68.6 50.5 89.4 90.3 65.4 84.7 59.4 60.5 92.1 88.3 89.8 52.2 29.6 60 30

180 120 180 90 90 66 45 120 120 90 90 60 240 360 180 270

\ 78.5 \ 32.7 20.8 36.5 5.2 63.5 72.9 17.5 20.6 16.9 21.5 30.3 30.7 10.2

\ 1.0 \ 5.8 4.9 0.8 4.0 4.7 5.0 4.6 2.7 5.3 12.0 39 \ \

90.0 89.7 151.5 48.7 31.3 39.8 7.3 89.6 72.2 25.8 30.8 14.7 18.7 14.9 30.7 5.1

0.0 0.0 0.0 0.7 4.8 2.5 21.1 0.0 0.0 7.3 1.1 8.7 4.1 39.8 2.2 17.8

1072

R. Liu et al. / Science of the Total Environment 468–469 (2014) 1069–1077 3

80

Time(min)

The data were statistically analyzed using the Microsoft Excel 2003 and SPSS 17.0 (SPSS, 2007) programs. Multi-factor analysis of variance was performed with the SPSS 17.0 for Windows software package.

9

70

8

60 50

3. Results

40

7-1

30 20

5

FT-1 TC-1

10 50

11

10

12

60

70

80

3.1. Characteristics of the experiment plots

6

90

100

Rainfall intensity(mm/h) Fig. 3. The runoff generated time under the different rainfall intensities in a single rainfall.

the fertilizer level. In addition, the loss characteristics of N and P were also analyzed. Each plot was 4-meters long and 3-meters wide. In three sides of the plots, diaphragm plates were inserted into ground. All of the diaphragm plates were 40 cm high and 10 cm of them is aboveground (Fig. 2a). The collection tubes remained horizontal to the ground so that the surface runoff could flow into these tubes. In addition, a collection barrel was placed at the end of the collection tubes to collect the surface runoff (Fig. 2b). The rainfall simulator (TSJY simulator, 2010), which was made in the Nanjing Forestry University, consisted of long sprinkler booms 4 m above the ground, and the nozzles were fixed in a triangular configuration (Fig. 2c). The rainfall simulator was set in the center of the plot. Its effective rainfall area was 12 m2, and its homogeneous degree was higher than 80%. Through the control of the nozzles, the intensities of the artificially simulated rainfall were adjusted from 0 mm/h to 120 mm/h. Based on the weather records in the past 50 years in the Baiyangdian watershed, the maximum rainfall was 962.1 mm (1956 year) and the minimum rainfall was 292.5 mm (1965 year). In 2008, the rainfall and the rainfall intensity were ranged 44.0–137.0 mm and 25.0–67.3 mm/h, respectively. Therefore, three rainfall intensities were used in the experiment: 30 mm/h (low), 60 mm/h (medium), and 90 mm/h (high). Based on the survey conducted in BZ, TC, and FT, the field experiments used three fertilizer (CO(NH2)2) levels: 0, 99 kg(N)/ha, and 165 kg(N)/ha. The runoff samples were collected 2, 4, 6, 8, 10, 15, 20, 25, 30, 40, 50, and 60 min after the runoff occurred. Then, the rest of the runoff was collected in a barrel and used to calculate the total runoff volume.

2.3. Experimental analysis The total nitrogen (TN) was determined using the Kjeldahl method, and the total phosphorus (TP) was determined by ammonium molybdate spectrophotometry. The ammonium nitrogen (NH4-N) content was measured through colorimetry using Nessler's reagent. The nitrite nitrogen (NO2-N) content was determined by diazonium coupled spectrophotometry, and the nitrate nitrogen (NO3-N) concentration was determined by ultraviolet spectrophotometry. The phosphate (PO4-P) content was measured using the molybdate blue method (GB, 2006).

Table 3 Multivariate analysis of variance of influencing factors. Rainfall intensity

Runoff coefficient Runoff generated time Initial loss of rainfall a b

F value. Significance.

Antecedent soil moisture content

Vegetation fraction

Fa

Sig.b

F

Sig.

F

Sig.

13.2 25.5 4.3

0.004 0.001 0.061

35.4 6.5 0.39

0.000 0.031 0.690

0.26 8.4 0.49

0.623 0.028 0.508

Before the rainfall simulation, the characteristics of the experimental sites were monitored (Table 2). The fertilizer level (0, 99, and 165 kg(N)/ha), the vegetation fraction (45%–90%), and the antecedent soil moisture content (27.9%–45%) were different in every plot. Throughout all of the field experiments, there was no runoff in experimental plots 1, 2, and 4. The experimental plot designations 7-1, 7-2, TC-1, TC-2, FT-1, and FT-2 indicated that there were two continuous rains simulated in the same experimental plot. 3.2. Surface runoff characteristics and influencing factors Because of the differences in experimental conditions (plant height, vegetation fraction, antecedent soil moisture content and rainfall intensity), runoff parameters were significantly different (Table 2). The generation of the surface runoff depended largely on rainfall characteristics and antecedent soil moisture conditions. The amount of the initial rainfall loss ranged from 5.1 mm to 151.5 mm. Once rain stopped, the runoff flow decreased rapidly, and lasted only a few minutes. In addition, with the increase in the vegetation coverage and the invariability of the rainfall intensity, the time from precipitation to runoff emergence became longer. The runoff coefficient is the ratio of the total runoff to the rainfall, and it is used to analyze the effects of the rainfall intensity on the runoff-yielding ability, thus excluding the impacts of the rainfall. The result showed that the runoff coefficient (under 60 mm rainfall) was 0 to 21.1%, and under the condition of consecutive rainfall, the runoff coefficient was significantly higher than those with single rainfalls, which was because the pre-rainfall increased the soil moisture and the runoff-yielding ability. The runoffs generated in plots 3, 8, 9, TC-1, and XS-1 were obviously delayed because the rainfall intensities were lower in these plots compared with the other plots (Fig. 3). This finding indicated that the rainfall intensity was the foremost factor that influenced the runoff generation time. The rainfall intensities in plots 3, 8, 9, TC-1, and XS-1 were the same, but the antecedent soil moisture content and the vegetation fraction in TC-1 and XS-1 were lower than those in the other plots. Thus, although the rainfall intensity was essential, the antecedent soil moisture content and the vegetation fraction could not be overlooked. It could be seen that the minimum requirements for runoff occurrence in the experimental areas were the rainfall depth of 5.1 mm, the rainfall intensity of 50 mm/h, and the antecedent soil moisture of approximately 29.6%. Using the rainfall intensity, the vegetation fraction and the antecedent soil moisture content as independent variables and the runoff generation time, the initial loss of rainfalls, and the runoff coefficient as dependent variables, the multivariate analysis was carried out. The result showed that with the same rainfall, the runoff coefficient was related to the rainfall intensity and the antecedent soil moisture content; the influence of the soil moisture content (F = 35.4, p = 0.000 b 0.05) was larger than that of the rainfall intensity (F = 13.2, p = 0.004 b 0.05). The rainfall intensity, soil moisture content, and vegetation fraction all had significant impacts on the runoff generation time; the factors influencing the runoff generation time were in such an order: rainfall intensity (F = 25.5, p = 0.001 b 0.05) N vegetation fraction (F = 8.4, p = 0.028 b 0.05) N soil moisture content (F = 6.5, p = 0.031 b 0.05). The runoff generation time and the runoff coefficient were important indicators for the runoff-yielding ability. The rainfall intensity and the soil moisture content were two of the main factors that influenced the runoff-yielding ability in the NCP. The effect of the vegetation fraction on the runoff-yielding ability

R. Liu et al. / Science of the Total Environment 468–469 (2014) 1069–1077

1073

Fig. 4. The loss processes of TN, NH4-N, NO2-N and NO3-N under a single rainfall in different experimental plots.

was not as significant as that of other factors, and it was indispensable and meaningful. (Table 3). 3.3. Characteristics of nutrients under a single rainfall The results showed that the concentration of nitrogen increased significantly with the amount of fertilizer (Fig. 4). Under moderate rain, the decreasing trend of TN, NH4-N, NO3-N, and NO2-N exhibited a significant relationship with the amount of chemical fertilizer in experimental plots 3, 8, and 9 (Fig. 4a, c, e, and g). Under heavy rain, the TN and NO3-N contents (experimental plot 6) increased significantly with the amount of fertilizer at the beginning of the rainfall but decreased to levels lower than those measured in non-fertilizer soil after a period of time (Fig. 4b and f). Due to the lower vegetation fraction in plot 6, resulting in an obvious increase in the speed and volume of the runoff, the concentrations of TN, NO3-N and NH4-N in experimental plot 6 were higher than those obtained in other plots.

The other indispensable factor that significantly influences the loss of nitrogen was found to be the rainfall intensity. The runoff volume under heavy rain conditions was higher than that measured under moderate rainfall conditions. In addition, the content of N during heavy rain was higher than that obtained during moderate rain. However, due to dilution, the concentrations of TN, NH4-N, NO2-N, and NO3-N expressed a downward trend over time as a result of a single rain. The concentrations of both TP and PO4-P fluctuated under moderate rain conditions likely due to the rainfall intensity and the vegetation fraction (Fig. 5). Despite a high vegetation fraction in experimental zones 3, 8, and 9, the concentrations of both TP and PO4-P still fluctuated significantly. With the exception of the vegetation cover, the loss of TP and PO4-P also exhibited a strong relationship with the rainfall intensity. The runoff volume under heavy rain conditions was higher than that measured under moderate rainfall conditions. In addition, the content of P under heavy rain conditions was higher than that measured under moderate rain conditions.

1074

R. Liu et al. / Science of the Total Environment 468–469 (2014) 1069–1077

Fig. 5. The loss processes of TP and PO4-P under a single rainfall in different experimental plots.

The outputs of P and N were both influenced by the rainfall intensity and the vegetation cover, but because only nitrogen fertilizer was used in the field experiments, the loss of N was more than that of P. 3.4. Characteristics of nutrients under consecutive rainfall Since the soil moisture content was increased after the first rainfall, which was conducive to the formation of surface runoff and the output of nutrients, the outputs of TN, TP, and DP during the second rainfall were higher than those measured during the first rainfall, and the runoff was generated earlier in the second rainfall compared with the first rainfall (Fig. 6). The average concentrations of TN, TP, and DP in the second rainfall were lower than those measured during the first rainfall, whereas the contents of NH4-N and NO3-N during the second rainfall were higher than those measured during the first one. The reason for this difference might be the improvement in the soil moisture content and the runoff volume after the first rainfall, and the decrease in the average nutrient concentration is likely due to dilution. 3.5. The pollution assessment of runoff Event Mean Concentration (EMC) is often used to evaluate the pollution degree in runoff. The EMC values of TN, NH4-N, NO3-N, TP, phosphate and DP in different experimental plots are shown in Table 4. The EMC of TN in BZ (6.30 ± 1.57 mg/L) was significantly higher than that in TC (3.10 ± 0.18 mg/L) and FT (2.97 ± 0.56 mg/L). The reason might be that the nitrogen fertilizer was only applied in the field of BZ. Moreover, most of the nitrogen in surface runoff existed as dissolved nitrogen, NO3-N accounted for 41.85%–81.32%, NH4-N accounted for 0.31%–54.7%, and NO2-N accounted for 0%–5.3%. The EMC values of TP in BZ, TC and FT were 0.49 mg/L, 0.72 mg/L and 1.06 mg/L, respectively. In BZ, most of the phosphorus in surface runoff existed as dissolved phosphorus, and phosphate accounted for 63.41%. While in TC and FT, the phosphorus in surface runoff existed as particulate phosphorus, and the dissolved phosphorus in TC and FT accounted for 10.51% and 31.77%, respectively. In BZ, TC and FT, the content of TN in runoff all had surpassed the worst level of the national groundwater environmental quality (GB, 2002). The loss of phosphorus was less than that of nitrogen, but for most plots, the content of TP in runoff also had surpassed the worst level of the national groundwater environmental quality (GB, 2002).

4. Discussion The physical characteristics of the soil affect the loss of N and P, which includes the soil type, the saturated soil moisture content, the organic matter, and the particle-size distribution (Dreelin et al., 2006; Ouyang et al., 2010; L.P. Zhang et al., 2012; W.W. Zhang et al., 2012). In a sandy dry field, particulate nitrogen and phosphorus carried by runoff sands were the primary forms of nitrogen and phosphorous losses under typical rainfall conditions (Qian et al., 2010). However, these were different from the loss characteristics in the NCP. In addition to soil characteristics, the terrain condition also affects the loss of N and P (Zhao et al., 2011). In the NCP, the total phosphorous and nitrogen losses due to runoff in agricultural land were higher in sloped farmland compared with flat farmland. However, in the hilly area of China, the concentrations of ammonium nitrogen and phosphate in the runoff are always maintained at low concentrations during the rainfall (Jiang et al., 2008). The rainfall intensity, vegetation fraction and soil moisture content were three factors that influenced the runoff volume in the NCP. Some researchers have demonstrated that the vegetation fraction is the major factor that influences the watershed NPS nutrient movement and some studies have indicated that the low vegetation fraction results in an obvious increase in the speed and volume of the runoff (Bakhsh et al., 2005; Udawatta et al., 2006; Zuo et al., 2008; G.H. Zhang et al., 2011; H. Zhang et al., 2011; Liang et al., 2012; L.P. Zhang et al., 2012; W.W. Zhang et al., 2012). The rainfall intensity and the soil moisture content influenced largely the runoff-yielding ability (Bouldin et al., 2004; Liang et al., 2012; Rivera-Ruiz et al., 2012). Beyond that, the three factors also had a significant influence on the loss of N and P. For the vegetation fraction, some researchers have shown that there are large fluctuations in the nutrient concentrations in bare soil because of the runoff that is formed before the soil internal voids are saturated and because of the lag effect of the extraction process between the surface and the internal nutrient concentrations (Sharpley, 1997; Dan et al., 2001; Shigaki et al., 2007; Hahn et al., 2012). For the rainfall intensity and the soil moisture content, the higher rainfall intensity and the higher soil moisture content are conducive to the formation of surface runoff and the output of nutrients (Vadas et al., 2004; Udawatta et al., 2006; Shigaki et al., 2007; Zuo et al., 2008; Y. Chen et al., 2012; Y.X. Chen et al., 2012; Hahn et al., 2012). In addition, fertilization also affects the loss of N and P (Wang, Guo, et al., 2012; Wang, Wang, et al., 2012). In the NCP, the use of fertilizer can obviously increase the output of N and P. In South China, the loss

R. Liu et al. / Science of the Total Environment 468–469 (2014) 1069–1077

1075

Fig. 6. The loss processes of TN, NH4-N and NO3-N, TP and DP under consecutive rainfall in different experimental plots.

of nitrate nitrogen and phosphorous is obviously influenced by fertilization in the following order: organic manure N fertilizer N organic manure (Liang et al., 2004). Moreover, some researchers have found that the return of straw to the field effectively reduces the surface runoff and the nutrient loss from farmlands under natural rainfall conditions. However, straw mulch has an insignificant influence on the TN and TP concentration lost as a result of runoff. Whereas dissolved nitrogen is the predominant form of nitrogen loss, particle phosphorous is the main form of phosphorus loss (Wang et al., 2010). In the NCP, dissolved nitrogen was the predominant form of N loss, but for phosphorous, because P was less mobile than N and less soluble,

the predominant loss form was influenced by the rainfall intensity, the vegetation cover, and the antecedent soil moisture content (De Jager and Houser, 2012). In addition, due to dilution, the concentrations of particulate nitrogen, particulate phosphorus, and total phosphorous were much higher in the early stage of the rainfall–runoff process compared with those in the later stage (Black, 1899; Rivett et al., 2008). 5. Conclusions The roles of rainfall, vegetation fraction and antecedent soil moisture content in runoff and nutrient loss were studied in field plots using

1076

R. Liu et al. / Science of the Total Environment 468–469 (2014) 1069–1077

Fig. 6 (continued).

rainfall simulators. The field experiment produced very useful information. In summary, the results are as shown: (1) It was proven that the rainfall intensity, the antecedent soil moisture content and the vegetation fraction were the main factors that affected the runoff-yielding ability. (2) It emerged clearly that the vegetation fraction was an effective method of preventing runoff and nutrient loss. The runoff production and nutrient export increased with the increasing rainfall intensity and the increasing soil moisture content. (3) It can be concluded that the minimum requirements for runoff occurrence in the experimental areas were the rainfall depth of 5.1 mm, the rainfall intensity of 50 mm/h, and the antecedent soil moisture of approximately 29.6%.

Table 4 Event Mean Concentration (EMC) of the pollutants. Site

Plot

TN (mg/L)

NH4-N (mg/L)

NO3-N (mg/L)

NO2-N (mg/L)

TP (mg/L)

Phosphate (mg/L)

DP (mg/L)

BZ

3 5 6 8 9 10 11 12 7-1 7-2 TC-1 TC-2 FT-1 FT-2

7.08 4.88 7.78 5.49 4.54 6.57 6.05 4.12 7.55 8.98 2.97 3.22 3.37 2.58

2.38 1.51 2.33 0.87 0.48 0.47 0.79 0.75 4.15 3.05 0.02 0.01 0.03 0.03

4.15 3 5.14 3.02 2.86 4.99 4.92 2.85 3.16 5.36 1.68 1.61 1.73 1.53

0.25 0.26 0.27 0.06 0.05 0.05 / / 0.2 0.21 / / / /

0.18 0.35 0.68 0.2 0.21 0.51 0.4 0.61 0.59 1.21 0.47 0.97 1.25 0.87

0.14 0.31 0.58 0.09 0.12 0.15 0.11 0.25 0.53 1.12 / / / /

/ / / / / / / / / / 0.06 0.08 0.32 0.33

TC FT

(4) N and P concentrations in runoff depended on the rainfall intensity and the manure application. In this context, N and P losses are more sensitive to the runoff caused by high intensity than for runoff caused by medium rainfall intensity. Indeed, as expected there were more TN and NO3-N but less TP, NH4-N, NO2-N and PO4-P in runoff. Overall, further research is needed to use watershed model simulations to explore the NPS pollution characteristics in the NCP. Acknowledgments The research was funded by the National Natural Science Foundation of China (Grant No. 41001352), the Fundamental Research Funds for the Central Universities, the Open Research Foundation of Pearl River Hydraulic Research Institute, PRWRI (No. [2010] KJ01) and the Water Specific Project of China (2008ZX07209-007). References Bakhsh A, Kanwar RS, Karlen DL. Effects of liquid swine manure applications on NO3-N leaching losses to subsurface drainage water from loamy soils in Iowa. Agric Ecosyst Environ 2005;109(1–2):118–28. Black PE. The science of watershed hydrology. USA: Ann Arbor Press; 1899. Bouldin JL, Farris JL, Moore MT, et al. Vegetative and structural characteristics of agricultural drainages in the Mississippi Delta landscapes. Environ Pollut 2004;132(3): 403–11. Carpenter SR, Caraco NF, Correll DL, et al. Nonpoint pollution of surface waters with phosphorus and nitrogen. Ecol Appl 1998;8(3):559–68. Chen Y, Wang M, Wang G, et al. Nitrogen runoff under simulated rainfall from a sewage-amended lateritic red soil in Fujian, China. Soil Tillage Res 2012a;123:35–42. Chen YX, Liu RM, Sun CC, et al. Spatial and temporal variations in nitrogen and phosphorous nutrients in the Yangtze River Estuary. Mar Pollut Bull 2012b;64(10):2083–9. Cheng F, Xu XZ, Gao JH, et al. Advance of research on the rainfall simulators for soil erosion. Sci Soil Water Conserv 2008;6(2):107–12.

R. Liu et al. / Science of the Total Environment 468–469 (2014) 1069–1077 Choi GC, Lee JH, Yu JC, et al. Laboratory assessment of biofihn process and its microbial characteristics for treating nonpoint source pollution. Korean J Chem Eng 2011;28(5): 1207–13. Cui HM. Analysis on effect of agricultural non-point source pollution on the water environment in Baiyangdian Basin. Mod Agri Sci Technol 2011;7:298–300. Dan BQ, Yin CQ, Yu J, et al. Study on phosphorus transport in the surface layer of soil with rainfall simulation method. Acta Sci Circumst 2001;21(1):7–12. De Jager NR, Houser JN. Variation in water-mediated connectivity influences patch distributions of total N, total P, and TN, TP ratios in the Upper Mississippi River, USA. Freshwater Sci 2012;31(4):1254–72. Dou PQ, Wang XY, Wang LH. Research progress on the mechanism of nitrogen and phosphorus in non-point source pollution. J Cap Norm Univ (Nat Sci Ed) 2006;27(2):93–8. Dreelin EA, Fowler L, Carroll CR. A test of porous pavement effectiveness on clay soils during natural storm events. Water Res 2006;40(4):799–805. Emili LA, Greene RP. Modeling agricultural nonpoint source pollution using a geographic information system approach. Environ Manage 2013;51(1). GB. Environmental quality standards for surface water; 2002. GB. Methods for chemical analysis of water editors; 2006. Gong Y, Shen Z, Hong Q, et al. Parameter uncertainty analysis in watershed total phosphorus modeling using the glue methodology. Agric Ecosyst Environ 2011;142(3–4): 246–55. Hahn C, Prasuhn V, Stamm C, et al. Phosphorus losses in runoff from manured grassland of different soil P status at two rainfall intensities. Agr Ecosyst Environ 2012;153: 65–74. Jacinthe PA, Lal R, Kimble JM. Carbon dioxide evolution in runoff from simulated rainfall on long-term no-till and plowed soils in southwestern Ohio. Soil Tillage Res 2002;66(1):23–33. Jiang R, Zhu B, Tang JL, et al. Characteristics of nitrogen and phosphorus losses in typical rainfall–runoff events in a small watershed in hilly area of purple soil. J AgroEnviron Sci 2008;27:1353–8. Jin K, Cornelis WM, Schiettecatte W, et al. Effects of different soil management practices on total P and Olsen-P sediment loss: a field rainfall simulation study. Catena 2009;78(1):72–80. Li WH, Shi YL, Gao LH, et al. Occurrence of antibiotics in water, sediments, aquatic plants, and animals from Baiyangdian Lake in North China. Chemosphere 2012;89(11):1307–15. Liang T, Wang H, Kung HT, et al. Agriculture land-use effects on nutrient losses in West Tiaoxi Watershed, China. J Am Water Resour Assoc 2004;40(6):1499–510. Liang FF, Jiang XJ, Yuan JJ, et al. Main features of the loss of nitrogen and phosphorus and rainfall intensity influence in the slope farmland of the Three Gorges Reservoir Area. J Soil Water Conserv 2012;26(4):81–5. Liao YS, Zhuo MN, Li DQ, et al. Impact of fertilizer and crop growth on non-point source pollution. Ecol Environ 2011;20(5):886–91. Liu RM, Zhang PP, Wang XJ, et al. Assessment of effects of best management practices on agricultural non-point source pollution in Xiangxi River watershed. Agric Water Manag 2013a;117:9–18. Liu WB, Fu GB, Liu CM, et al. A comparison of three multi-site statistical downscaling models for daily rainfall in the North China Plain. Theor Appl Climatol 2013b;111(3–4): 585–600. Long D, Gao YC, Singh VP. Estimation of daily average net radiation from MODIS data and DEM over the Baiyangdian watershed in North China for clear sky days. J Hydrol 2010;388(3–4):217–33. Ouyang W, Skidmore AK, Toxopeus AG, et al. Long-term vegetation landscape pattern with non-point source nutrient pollution in upper stream of Yellow River basin. J Hydrol 2010;389(3–4):373–80. Qian XY, Shen GX, Huang LH, et al. Characteristics of nitrogen and phosphorus losses with rainfall–runoff from sandy dry field in Chongming Dongtan. J Soil Water Conserv 2010;24:11–4. Rivera-Ruiz P, Oropeza-Mota JL, Martinez-Menes MR, et al. The rainfall–runoff–erosion relationship in hillsides and microwatersheds. Tecnol Cienc Agua 2012;3(4):151–66. Rivett MO, Buss SR, Morgan P, et al. Nitrate attenuation in groundwater: a review of biogeochemical controlling processes. Water Res 2008;42(16):4215–32.

1077

Sharpley AN. Rainfall frequency and nitrogen and phosphorus runoff from soil amended with poultry litter. J Environ Qual 1997;26(4):1127–32. Shen ZY, Liao Q, Hong Q, et al. An overview of research on agricultural non-point source pollution modelling in China. Sep Purif Technol 2012;84:104–11. Shigaki F, Sharpley A, Prochnow LI. Rainfall intensity and phosphorus source effects on phosphorus transport in surface runoff from soil trays. Sci Total Environ 2007;373(1): 334–43. Shortle JS, Ribaudo M, Horan RD, et al. Reforming agricultural nonpoint pollution policy in an increasingly budget-constrained environment. Environ Sci Technol 2012;46(3): 1316–25. Strauch M, Bernhofer C, Koide S, et al. Using precipitation data ensemble for uncertainty analysis in SWAT streamflow simulation. J Hydrol 2012;414:413–24. Sun TW, Chen JJ, Shi ZT, et al. Investigation on non-point sources pollution of a typical village and pollution loads of the runoff in Baiyang Lake watershed. J Anhui Agric Sci 2010;38(30):16991–3. Sun TW, Chen JJ, Wang H, et al. Study on non-point source pollution loads in villages along the Fuhe River, Baiyangdian watershed. Res Environ Sci 2012;25(5):568–72. Udawatta RP, Motavalli PP, Garrett HE, et al. Nitrogen losses in runoff from three adjacent agricultural watersheds with claypan soils. Agric Ecosyst Environ 2006;117(1): 39–48. Vadas PA, Kleinman P, Sharpley AN. A simple method to predict dissolved phosphorus in runoff from surface-applied manures. J Environ Qual 2004;33(2):749–56. Wang J, Guo XS, Wang YQ. Effect of straw mulch on nitrogen and phosphorus loss from farmlands in Chaohu Lake Region under natural rainfall condition. Chin J Eco-Agric 2010;18:492–5. Wang J, Guo XS, Wang YQ, et al. Characteristics of nutrient runoff losses from farmland under different tillage and fertilization methods in Chaohu Lake Region. J Soil Water Conserv 2012;26:6–11. Wang J, Wang EL, Yang XG, et al. Increased yield potential of wheat–maize cropping system in the North China Plain by climate change adaptation. Clim Change 2012;113(3–4):825–40. Wu L, Long TY, Cooper WJ. Temporal and spatial simulation of adsorbed nitrogen and phosphorus nonpoint source pollution load in Xiaojiang Watershed of Three Gorges Reservoir Area, China. Environ Eng Sci 2012;29(4):238–47. Yan LZ, Sun B. China's farmland ecosystem nutrient cycling, balance and its management. Beijing: Science Press; 2008. Zeng R, Zhao YW, Yang ZF. Emergy-based health assessment of Baiyangdian watershed ecosystem in temporal and spatial scales. Procedia Environmental Sciences. Amsterdam: Elsevier Science BV; 2010. Zhang Q, Chen Y, Yu Q, et al. A review on non-point source pollution models. Appl Ecol 2007;18(8). Zhang GH, Liu GB, Wang GL, et al. Effects of vegetation cover and rainfall intensity on sediment-bound nutrient loss, size composition and volume fractal dimension of sediment particles. Pedosphere 2011a;21(5):676–84. Zhang H, Zeng FT, Fang HY, et al. Impact of consecutive rainfall on non-point source pollution in the Danshui River catchment. Acta Sci Circumst 2011b;31(5):927–34. Zhang LP, Fu XT, Wu XY. Sediment content and nitrogen and phosphorus load characteristics of surface runoff on bamboo forest slopes: a simulation test. J Appl Ecol 2012a;23:4. Zhang WW, Li H, Sun DF, et al. A statistical assessment of the impact of agricultural land use intensity on regional surface water quality at multiple scales. Int J Environ Res Public Health 2012b;9(11):4170–86. Zhao W, Yang PL, Li HS, et al. Characteristics of soil erosion, nitrogen and phosphorous losses under three grassland use patterns in Hulun Lake watershed. Trans Chin Soc Agric Eng 2011;27:220–5. Zheng Y, Wang XJ. Advances and prospects for nonpoint source pollution studies. Adv Water Sci 2002;13(1):105–10. Zheng FL, Zhao J. Artificial simulation rainfall and the rainfall simulation equipment introduction. Res Soil Water Conserv 2004;11(4):177–8. Zuo HJ, Zhang Q, Xu LG. Study on the factors influencing nitrogen leaching loss from farmland and controlling measures. Environ Pollut Control 2008;30(12):83–9.