Geostatistical analyses of spatial distribution and origin of soil nutrients in long-term wastewater-irrigated area in Beijing, China

Abstract

The purpose of this study was to investigate the current status of soil fertility in the long-term wastewater-irrigated agricultural areas of Beiyechang District in the suburb of Beijing. A total of 103 soil samples from the top 20 cm of soil layer were collected and analyzed for macro- and micronutrient elements. The mean values of these elements were: total nitrogen (TN), 0.08 ± 0.02%; available nitrogen (AN), 60.34 ± 10.08 mg/kg; available phosphorus (AP), 19.59 ± 9.75 mg/kg; available potassium (AK), 84.22 ± 19.66 mg/kg; and availabilities of trace elements (B, Cu, Fe, Mn, Mo, and Zn), 0.29 ± 0.12, 1.44 ± 0.34, 8.97 ± 2.05, 5.44 ± 0.95, 0.27 ± 0.16, 1.16 ± 0.42 mg/kg, respectively. Compared with other similar areas, the overall soil nutrient content in the study area was still at a low level after nearly 30 years of wastewater irrigation. The organic matter content of wastewater-irrigated soils has significant impact on the accumulation of nutrient elements in these soils. Geostatistical analyses suggested that the spatial distribution of TN and AP was on account of agricultural practices, such as wastewater irrigation and fertilizer use. The availability of P, Mo, Cu and Zn may be due to the joint effect of soil parent material and wastewater irrigation. The availability of K, Fe and Mn was attributed to soil parent material. A significant availability of some elements occurred around sewage outfalls in the irrigation area and water canals, where the wastewater could be easily lifted. The knowledge of spatial distribution and sources of soil nutrients in such areas is the basis for undertaking appropriate farming and irrigation activities with rationalized utilization of treated wastewater.

Keywords:

• soil
• wastewater
• nutrient
• geostatistics
• spatial distribution
• irrigation

Introduction

Water scarcity is one of the major crises of the world today (Pereira et al. 2011), especially in arid and semi-arid regions (Maria et al. 2012). In this scenario, wastewater recycling is an important measure to reduce the dependence of freshwater resources (Weber et al. 2006), and wastewater use for irrigation has been a widespread practice with a long tradition around the world (Anikwe & Nwobodo 2002). Compared with freshwater, the wastewater possesses higher concentrations of macro- (N, P, K, S, Ca, Mg) and micronutrients (B, Cu, Fe, Mn, Mo, Zn), which are essential for plant growth (Ramirez-Fuentes et al. 2002; Rattan et al. 2005; Sophocleous et al. 2009; Jaiswal & Elliott 2011). Thus, wastewater irrigation could input some necessary nutrients to soil–plant systems (Toze 2006). A long-term use of wastewater irrigation can increase organic matter (OM) content and improve soil fertility (Horswell et al. 2003; Xu et al. 2010; Singh et al. 2012), thereby increasing crop biomasses and yields (Hussain et al. 1996; Pollice et al. 2004; Ma et al. 2008; Wu et al. 2010; Xu et al. 2010).

Several studies have investigated the characteristics of migration and enrichment of macronutrients (N, P and K) and trace elements (Cu, Fe and Zn) in soil–plant and groundwater systems as a result of long-term sewage irrigation (Yadav et al. 2002; Xu et al. 2010; Singh et al. 2012), but few studies have focused on the spatial accumulation of soil nutrients (major and trace elements) under long-term wastewater irrigation. In wastewater-irrigated areas, nutrient elements contained in the wastewater show differences in their abilities to migrate in soils, thereby resulting in higher local concentrations (Wang et al. 1997; Wu et al. 2013). Similarly, factors such as irrigation methods, soil textures and topography cause variations in the spatial accumulation of these nutrients. Under long-term wastewater irrigation, trace elements such as Cu and B can excessively accumulate in soils, thereby degrading the quality of soils and posing certain toxic effects on crops (Pereira et al. 2011). Therefore, it is necessary to understand the spatial distribution of soil nutrients in areas with a long history of wastewater irrigation to promote sustainable development and rationalized utilization of wastewater resources.

Geostatistical methods are suitable for describing the spatial distribution and spatial autocorrelation characteristics of regional variables (Webster & Oliver 2001; Li et al. 2003). The application of geostatistical methods in the field of soil science has been increasingly recognized (Guo et al. 2000; Navarro-Pedreño et al. 2007; Donato et al. 2010). Geostatistical methods can help collect information about the spatial variations of soil quality at unsampled sites (Larka & Ferguson 2004; Navarro-Pedreño et al. 2007) to predict the spatial distribution of physicochemical parameters of soils in the region (Juan et al. 2011). Several investigators have used geostatistical methods to study the spatial uncertainty of elements, including heavy metals in long-term wastewater-irrigated areas, to assess the risks of pollution (Zhang et al. 2003; Yang et al. 2008).

Therefore, the present study was conducted to analyze the soil fertility index in the agricultural soils of the Beiyechang District, a long-term wastewater-irrigated area in the Beijing suburb, using geostatistical methods. The objectives of the study were to (1) assess the concentrations and spatial distribution of soil nutrients (macro- and micronutrients), (2) determine the causes for their spatial patterns and (3) identify the possible sources.

Methods and materials

Study area

The study area is located in the Beiyechang Irrigation District in Daxing District (Figure 1) in the southeast central of Beijing (longitude 115°25ʹ to 117°30ʹ E and latitude 39°28ʹ to 41°05ʹ N). The area has a warm temperate semi-humid continental monsoon climate, with annual average rainfall of 516.4 mm and annual average evaporation of 979 mm. The rainfall is unevenly distributed throughout the year, with more than 70% of the rainfall in July, August and September.

Figure 1. Map of soil sampling site of Beiyechang Irrigation District.

The irrigation area was created in 1960 with a cultivable area of 14.8 km². Wastewater irrigation was introduced in this area in 1970. By 2007, this area had a history of over 30 years of wastewater irrigation. The wastewater was primarily from two sources: one was the city’s drainage waterway, the Liangshui River, by the diversion canal system north of the irrigation area, and the other was the industrial and domestic wastewater from the Shijingshan District in the western part of the city and the Daxing District where the irrigation area is located. The wastewater is conveyed to the entire irrigation area through main canals, branch canals and rivers in Beiyechang, and it enters the canal system of the irrigation area after being lifted by pumping stations. The soil parent material of the area is alluvium from the Yongding River, and the soil for cultivation is mainly alluvial soil. The soil texture is silty clay loam. The main crops in this area are winter wheat and summer corn.

Sampling and analysis

A total of 103 topsoil samples (0–20 cm depth) were collected from the agricultural areas in Beiyechang District in October 2007 (Figure 1). The sampling design was based on the distribution of canals and agricultural land use. The sampling density was approximately 1 sample per 0.14 km². Five individual sub-samples were taken along a single diagonal line from a 10 × 10 m grid by using a stainless steel spade. Sub-samples were mixed to obtain a composite sample for each site. The coordinates for each sampling location were recorded using a global position system receiver.

About 1.0 kg soil sample was taken from the thoroughly mixed sample of the site. Soil samples were air-dried indoor and passed through a 2.0-mm sieve. Soil pH, OM, total nitrogen (TN), available nitrogen (AN), available phosphorus (AP), available potassium (AK) and availabilities of trace elements (B, Cu, Fe, Mn, Mo and Zn) were determined. In particular, OM was determined using potassium dichromate oxidation/external heating (oil bath; Liu & Jiang 1996), and the semi-micro Kjeldahl method was used to determine TN. Ammonium acetate extraction/flame photometry was used to measure AK, and sodium bicarbonate extraction/molybdenum antimony colorimetric method (Olsen) was used to determine AP. The potentiometric method (soil/water ratio of 1:2.5; CO₂-free distilled water was used) was used to measure pH (Lu 1999). After extraction using diethylene triamine pentoacetic acid (DTPA) (1:2), available Cu (ACu), available Zn (AZn), available Fe (AFe) and available Mn (AMn) were determined using plasma spectroscopy and mass spectroscopy. After extraction using oxalic acid and ammonium oxalate (1:10), available Mo (AMo) was determined using plasma spectrometry or polarography. After extraction using boiling water (1:2), available B (AB) was determined using plasma spectrometry and a colorimetric method (Liu & Jiang 1996).

Geostatistical methods

The frequency distribution of soil nutrients were investigated by calculating skewness and kurtosis coefficients. For kriging estimation, the semivariogram, r(h), was used to quantify the spatial dependence of soil nutrient concentrations. Then, based on the spatial structure of r(h), kriging estimation was used to obtain heavy metal concentrations at unsampled locations. For this study, experimental semivariogram models were obtained using the software package GS+, while kriging estimation was performed using the software of ArcGIS 9.3 for Windows (Luo et al. 2007).

Statistical analysis

To study the relationship among nutrient elements in soils and their possible sources, Pearson’s correlation coefficient analysis and coefficient of variation (C_v) were performed using SPSS version 13.0 for Windows. The correlation coefficient measures the strength of inter-relationship between two fertility parameters of wastewater-irrigated soils.

Results and discussion

Soil nutrient concentrations

Table 1 shows the descriptive statistics of soil nutrient concentrations of the study area. By comparison, soil OM of the present study area, ranging from 0.73% to 1.92% (average 1.52%), is similar to that in Dunhua wastewater-irrigated area (20 years; range 1.27% to 2.18%; average 1.69%) in Taiyuan (Liang et al. 2010), but higher than that of the topsoils (0.8% to 1.4%) after 20 years of wastewater irrigation in a study by Xu et al. (2010). The average value of TN was0.08%, and those of AN, AP and AK were 60.34, 19.59 and 84.22 mg/kg, respectively. Except for the higher AK level in studied soils, the concentration levels of TN, AN and AP compare with the mean concentration levels (0.11%, 62.85 mg/kg, 21.65 mg/kg) of the surface layer (0–20 cm) in the Zhangshi area of Shenyang after 30 years of wastewater irrigation (Zhang et al. 2008). In particular, the mean concentration levels of ACu (1.80–115; average 28.06), AZn (1.03–23.3; average 5.50), AMn (0.52–39.7; average 4.14) and AFe (0.57–36.7 mg/kg; average 5.6) in the study area were lower than those of the Delhi irrigation area in India, with a 20-year history of wastewater irrigation (Rattan et al. 2005), and the concentration of AMo was lower (2.21 mg/kg) than that of the Mezquital wastewater-irrigated area in Mexico, with a 30-year history of wastewater irrigation (Ramirez-Fuentes et al. 2002). Overall, compared with other long-term wastewater-irrigated areas, the concentrations of various nutrients in the soils of the study area were still at a relatively low level.

Table 1. Descriptive statistics of fertility parameters of studied soils.

Indicator	Maximum	Minimum	Mean	Standard deviation	CV
pH	8.70	8.00	8.28	0.14	0.02
OM (%)	1.92	0.73	1.52	0.24	0.16
TN (%)	0.13	0.05	0.08	0.02	0.20
AN (mg/kg)	81.60	27.20	60.34	10.08	0.17
AP (mg/kg)	59.80	2.20	19.59	9.75	0.50
AK (mg/kg)	146.00	46.40	84.22	19.66	0.23
AB (mg/kg)	0.78	0.05	0.29	0.12	0.43
ACu (mg/kg)	3.09	0.71	1.44	0.34	0.24
AFe (mg/kg)	15.00	4.31	8.97	2.05	0.23
AMn (mg/kg)	8.90	3.77	5.44	0.95	0.18
AMo (mg/kg)	0.86	0.04	0.27	0.16	0.58
AZn (mg/kg)	3.12	0.38	1.16	0.42	0.36

OM, organic matter; TN, total nitrogen; AP, available phosphorus; AN, available nitrogen; AK, available potassium; AB, available B; ACu, available Cu; AFe, available Fe; AMn, available Mn; AZn, available Zn.

The C_v value reflects the degree of variability of fertility: C_v < 0.10 represents weak variability, 0.10 < C_v < 1 represents moderate variability, and C_v > 1 represents high variability (Wang et al. 2001). In this study, the values of C_v ranged from 0.16 to 0.58, demonstrating moderate variability among the various studied nutrients in the following order:

AMo > AP > AB > AZn > ACu > AFe = AK > TN > AMn > AN > OM. In particular, the C_v values of pH, OM, AN and AMn were 0.02, 0.16, 0.17 and 0.18, respectively, with variability being as low as 0.20. The C_v values of AP, AMo and AB were higher than 0.40, with relatively high variability. It was assumed that elements (AN, AMn) having smaller values of C_v are likely dominated by natural sources, while those (AP, AMo, AB) with larger values of C_v are more likely to be affected by anthropogenic sources.

Correlation between nutrients and among parameters

Table 2 presents the Pearson correlation coefficients of various elements in the soils. Soil pH showed highly significant positive correlation (p < 0.01) with OM, macronutrients (AN, AP, AK) and micronutrients (AFe, AMn, ACu) in the soils. OM showed a highly significant positive correlation (p < 0.01) with nutrients in the soils (except AB and AZn) and a significant positive correlation with AZn (p < 0.05). Using correlation analysis, Maria et al. (2012) showed that soil OM significantly influences the accumulation of nutrients such as TN and AP. Wang et al. (1997) showed that soils with high levels of clay particles and OM due to wastewater irrigation demonstrated strong adsorption and holding effects on trace elements, and both trace heavy metals (Cu and Zn) and OM accumulate in the soils through the process of wastewater irrigation (Hu et al. 2008).

Table 2. Pearson correlation coefficients of fertility parameters of studied soils.

Indicator	pH	OM	TN	AN	AP	AK	AB	ACu	AFe	AMn	AMo	AZn
pH	1
OM	−0.39**	1
TN	−0.08	0.37**	1
AN	−0.36**	0.67**	0.33**	1
AP	−0.31**	0.30**	0.1	0.44**	1
AK	−0.22**	0.42**	0.27**	0.35**	0.28**	1
AB	−0.06	−0.06	0.001	0.15	0.18	−0.004	1
ACu	−0.30**	0.33**	0.11	0.28**	0.28**	0.38	−0.03	1
AFe	−0.39**	0.40**	0.25**	0.37**	0.32**	0.38**	0.04	0.44**	1
AMn	−0.34**	0.43**	0.32**	0.37**	0.13	0.32**	−0.03	0.26**	0.48**	1
AMo	−0.17	0.32**	−0.05	0.08	−0.01	−0.02	−0.01	0.03	0.08	0.24*	1
AZn	−0.17	0.21*	−0.14	0.13	0.35**	0.12	0.01	0.33**	−0.03	0.1	0.14	1

OM, organic matter; TN, total nitrogen; AP, available phosphorus; AN, available nitrogen; AK, available potassium; AB, available B; ACu, available Cu; AFe, available Fe; AMn, available Mn; AZn, available Zn.

* Correlation is significant at the 0.05 level (2-tailed).

** Correlation is significant at the 0.01 level (2-tailed).

TN showed a highly significant positive correlation (p < 0.01) with AN, AK, AFe and AMn. Moreover, AN, AK and AP were highly significantly correlated with each other (p < 0.01). Among various trace elements, AFe was highly significantly correlated with AMn and ACu (p < 0.01). ACu was highly significantly correlated with AZn (p < 0.01), and AMo was significantly correlated with AMn (p < 0.05). AN and AK had a significant impact on the spatial distribution of AFe, AMn and ACu (p < 0.01). Furthermore, AP was highly significantly correlated with AFe, ACu and AZn (p < 0.01), and AB was weakly correlated with all other elements. Thus, accumulation and coexistence of various elements in wastewater-irrigated soils are highly correlated (Hu et al. 2008).

Analysis of spatial structures of elements

Soil nutrients (major or trace elements) are not only affected by human activities but are also related to soil textures and structure conditions, as well as they show certain trends in spatial distribution (Huang et al. 2007). To verify whether the data followed a normal distribution, skewness and kurtosis were used: if skewness was close to 0 and kurtosis close to 3, then the distribution was normal. Table 3 shows the results of normality tests of soil nutrients and the analysis of overall trends. Soil pH, AFe and AMn followed a normal distribution and showed a first-order dominant trend, so the universal kriging (UK) interpolation model was suitable. Soil OM, TN, AN and AK followed a normal distribution without an overall trend, so the ordinary kriging (OK) interpolation model was suitable. Soil AP, ACu, AZn and AB were not normally distributed and showed no overall trends, so the disjunctive kriging (DK) interpolation model was suitable. AMo followed a log-normal distribution, so the logistic normal kriging (log-OK) model was suitable.

Table 3. Spatial distribution of fertility parameters of studied soils.

	No transformation		Logarithmic transformation
Indicator	Skewness	Kurtosis	Skewness	Kurtosis	Distribution of the model	Overall trend	Suitable interpolation model
pH	0.62	3.49			Normal	First-order dominant trend	UK
OM	−0.69	3.46			Normal	None	OK
TN	0.44	3.13			Normal	None	OK
AN	−0.48	3.63			Normal	None	OK
AP	1.10	5.38	−0.91	4.84	Non-normal	None	DK
AK	0.64	3.26			Normal	None	OK
AFe	0.35	3.43			Normal	First-order dominant trend	UK
AMn	0.38	3.37			Normal	First-order dominant trend	UK
AB	0.56	4.31	−1.18	4.82	Non-normal	None	DK
ACu	1.22	7.57	−0.05	4.05	Non-normal	None	DK
AZn	1.90	8.44	0.31	4.49	Non-normal	None	DK
AMo	1.25	4.84	−0.48	3.49	Log-normal	None	Log-OK

OK, Ordinary Kriging; Log-OK, Logistic Normal Kriging; DK, Disjunctive Kriging; UK, Universal Kriging; OM, organic matter; TN, total nitrogen; AP, available phosphorus; AN, available nitrogen; AK, available potassium; AB, available B; ACu, available Cu; AFe, available Fe; AMn, available Mn; AZn, available Zn.

The choice of semivariogram was mainly based on the principle of least mean prediction error. As shown in Table 4, exponential functions had the least mean prediction error for soil pH, OM, TN and AN, so the optimal semivariograms were exponential functions. Gaussian functions had smaller prediction errors for soil AP, AFe, AMn, AB and AMo, so the optimal semivariograms were Gaussian functions. The mean prediction errors were the least when using spherical functions to predict soil AK, ACu and AZn, so the optimal semivariograms were spherical functions.

Table 4. Suitable variograms for fertility parameters of studied soils.

	Mean error under different variograms
Factor	Spherical function	Exponential function	Gaussian function	Suitable semivariogram
pH	0.1461	0.1438	0.1477	Exponential
OM	−0.0031	−0.0024	−0.0041	Exponential
TN	−0.0003	−0.0003	−0.0003	Exponential
AN	−0.2738	−0.2339	−0.3354	Exponential
AP	0.2090	0.2400	0.1978	Gaussian
AK	−0.0514	0.0838	−0.0663	Spherical
AB	−0.0003	−0.0014	−0.0001	Gaussian
ACu	−0.0006	0.0009	−0.0032	Spherical
AFe	0.0116	0.0121	0.0097	Gaussian
AMn	−0.0061	−0.0067	−0.0057	Gaussian
AMo	0.0005	0.0005	0.0002	Gaussian
AZn	−0.0041	−0.0053	−0.0043	Spherical

OM, organic matter; TN, total nitrogen; AP, available phosphorus; AN, available nitrogen; AK, available potassium; AB, available B; ACu, available Cu; AFe, available Fe; AMn, available Mn; AZn, available Zn.

Semivariograms and correlating index were calculated, and the results are presented in Table 5, including a nugget value (C₀), sill (C₀ + C₁) and ratio of nugget to sill. The sill and nugget (C₀) describe spatial heterogeneity (Cambardella et al. 1994). The sill expresses the attributes of the system or the maximum variation of regional variables; the higher the sill, the larger the degree of total spatial heterogeneity. However, the sill is not effective to compare regional variables because of the larger influence of the sill itself and the measurement unit. The nugget expresses the spatial heterogeneity of the stochastic component. A large nugget variance shows that there is an assignable process of small scale; it cannot be used to compare the differences of the stochastic components of different variables. However, the ratio of nugget to sill (C₀/C₀ + C₁) reflects the total spatial heterogeneity in nugget variance and hence is very useful (Li et al. 2013).

Table 5. Spatial variation and accuracy of fertility parameters.

Indicator	Range/m	Nugget (C0)	Partial sill (C1)	Sill (C0 + C1)	C0/C0 + C1
pH	7854.92	0.0076	0.0198	0.2440	0.57
TN	7854.92	0.0002	0.00004	0.0003	0.86
OM	7854.92	0.0313	0.0390	0.0703	0.45
AN	7854.92	66.1160	55.3970	121.5130	0.54
AP	3622.68	0.8231	0.2393	1.0624	0.77
AK	1282.51	159.75	258.7100	418.4600	0.38
AB	3900.20	0.7802	0.4526	1.2329	0.63
ACu	7854.92	0.6247	0.6881	1.3128	0.48
AFe	5014.76	2.9255	2.2055	5.1310	0.57
AMn	3585.75	0.6442	0.3666	1.0108	0.64
AMo	7854.92	0.3082	0.1479	0.4561	0.68
AZn	1333.52	0.4623	0.4844	0.9467	0.49

OM, organic matter; TN, total nitrogen; AP, available phosphorus; AN, available nitrogen; AK, available potassium; AB, available B; ACu, available Cu; AFe, available Fe; AMn, available Mn; AZn, available Zn.

C₀/(C₀ + C₁) < 0.25 indicates a strong spatial correlation among various factors within their variable ranges, with the structural factors being the major factors affecting the spatial variation; 0.25≤C₀/(C₀ + C₁)≤ 0.75 indicates a medium spatial correlation, and C₀/(C₀ + C₁) > 0.75 indicates a weak spatial correlation (Guo et al. 2000). Structural factors (parent material, topography and soil types) lead to increased soil spatial correlation, and random factors (irrigation, fertilizer use and tillage measures) weaken spatial correlations. In terms of the effect of ratio of nugget and sill of various elements, the order was as follows: TN > AP > AMo > AMn > AB > AFe = pH > AN > AZn > ACu > OM > AK. In particular, the nugget to sill ratio of TN and AP was 0.86 and 0.77, respectively (both larger than 0.75), indicating significant impact on their spatial distribution by wastewater irrigation, fertilizer use and other human activities. The nugget to sill ratio of other indicators was between 0.38 and 0.68, denoting medium spatial correlation, indicating that their spatial distribution was affected by the joint effect of structural factors and random factors (Huang et al. 2007).

Spatial distribution of various nutrients

Figure 2a–2l shows the spatial distribution maps of various soil nutrients plotted using ArcGIS according to the derived semivariograms shown in Table 4. As shown in Figure 2a and 2b, the pH values of wastewater-irrigated soils show an overall increasing trend from the north to the south in the range of 8.0 to 8.7, and the distribution of OM is high in the north and low in the south (Figure 2d). AN is high in the north, and significant enrichment occurred along the main canals (at the inlets of branch canals) and is significantly correlated with the distribution of OM (0.60; p < 0.01). Significant enrichment of TN and AP occurred around the diversion branch canals (Figure 2c and 2f), and the nugget to sill ratio of these was also relatively high. Anthropogenic processes (long-term wastewater irrigation and fertilizer use) were the major factors that affected the accumulative distribution of TN and AP. AN and AK had highly significant correlation (p < 0.01) with AFe and AMn. Most of Fe and Mn came from the soil parent material. Thus, AN may be affected by the soil parent material and anthropogenic processes. There was a significant AK-enriched region in the northeast part of the study area (at the ends of branch canals; Figure 2e). As shown in Figure 2g–2l, the spatial distribution of various trace elements in the study area showed certain correlations among them and was significantly affected by soil OM. Significant enrichment of AFe occurred in areas controlled by the second and third branch canals, and significant enrichment of AMo occurred at the heads of some branch canals. At the same time, AMo was high in the southern central region, and AB and AMn were high in the central part of the irrigation area (a drainage river is located in that area). The characteristics of ACu and AZn enrichment were similar (highly significant correlation; p < 0.01) and showed a trend of accumulation in regions near the main water canal, and there was good accumulation of AZn in the northeastern part of the study area. According to the analysis of correlation coefficients and geostatistical analysis, AMo, ACu and AZn may have been subject to the joint effect of soil parent material and wastewater irrigation.

Figure 2. Spatial distribution maps of fertility parameters of studied soils.

Therefore, macro elements, namely TN, AN and AP, and micro elements, namely ACu, AZn and AMo, were higher at wastewater outfalls or along the main canals and branch canals. This distribution trend might be related to the source condition of wastewater irrigation, as follows: being located closer to the main canal led to higher irrigation probabilities, larger water diversion, easier soil nutrient enrichments and higher soil fertilities (Wang et al. 1997; Khan et al. 2007). According to the analysis of correlation coefficients, the ratio of nugget and sill, the characteristics of spatial distribution, and the parent material had a largereffect on AFe, AMn, AB and AK distribution in the wastewater-irrigated soils of study area.

In sum, compared with other similar areas, the overall nutrient content in the wastewater-irrigated soils of the study area was still at a low level even after over 30 years of wastewater irrigation. However, the long-term wastewater irrigation has been the primary reason for the spatial distribution of OM and some nutrients in the soils. Significant enrichment of elements such as N, P, Cu and Zn occurred in the areas around sewage outfalls in the irrigation area and water canals, from where the wastewater could be more easily lifted than from the interior areas. In view of the results of the study, more effective measures of irrigation with treated wastewater might need to be examined for this district. Moreover, continuous monitoring of B, Cu and Zn concentrations should be enforced in areas that engage in treated sewage irrigation to prevent excessive accumulation of these elements in the soils from posing a threat to crop production and human health.

Acknowledgments

This paper was supported by the National 863 project under Grant [grant number 2012AA101404-1]; a Ministry of Water Resources public project under Grant [grant number 201101051]; and projects for talents of Beijing 2010.