Risks of phosphorus runoff losses from five Chinese paddy soils under conventional management practices

Agriculture, Ecosystems and Environment 245 (2017) 112–123

Contents lists available at ScienceDirect

Agriculture, Ecosystems and Environment journal homepage: www.elsevier.com/locate/agee

Risks of phosphorus runoff losses from five Chinese paddy soils under conventional management practices

MARK

⁎

Lingling Huaa, Jian Liub, Limei Zhaia, , Bin Xic, Fulin Zhangd, Hongyuan Wanga, Hongbin Liua, Anqiang Chene, Bin Fue a Key Laboratory of Nonpoint Source Pollution Control, Ministry of Agriculture/Institute of Agricultural Resources and Regional Planning, Chinese Academy of Agricultural Sciences, Beijing, 100081, China b Department of Plant Science, Pennsylvania State University, University Park, PA 16802, USA c Rural Energy and Environment Agency, Ministry of Agriculture, Beijing, 100081, China d Institute of Plant Protection, Soil and Fertilizer Sciences, Hubei Academy of Agricultural Sciences, Wuhan, Hubei 430064, China e Agricultural Resources and Environment Institute, Yunnan Academy of Agricultural Sciences, Kunming, Yunnan 650205, China

A R T I C L E I N F O

A B S T R A C T

Keywords: Field ponding water Nutrient management Phosphorus Rice Runoff Water quality

Phosphorus (P) runoff from arable land is a major cause for eutrophication of many surface waters. However, relatively little research has been conducted on managing P in rice (Oryza sativa L.) production systems, where farming practices differ from those of upland cropping systems due to water ponding on the soil surface (field ponding water; FPW). Because FPW is a direct source of surface runoff, identifying the main source of P and the critical period of high P concentrations in the FPW provide important insights to mitigating P runoff losses. In this study, field monitoring and laboratory incubation experiments were combined to evaluate how soil P content and conventional P fertilizer application affected FPW P concentrations in rice–wheat (Triticum aestivum L.) rotation systems of five Chinese rice producing regions. All soils had Olsen-P concentrations (10.1–20.5 mg kg−1) well below the critical levels (30–172 mg kg−1) for promoted risks of P loss. However, conventional P application rate significantly elevated FPW P concentrations compared to no P application, and P fertilizer contributed 47–92% of total P (TP) and 59–97% of total dissolved P (TDP) in the FPW. Temporarily, both TP and TDP concentrations peaked one day after P application (0.15–8.90 mg TP L−1 and 0.16–4.49 mg TDP L−1), then decreased rapidly and stabilized five days later. We conclude that fertilizer is the major source of P loss in Chinese rice production systems, and that P fertilizer rate should be optimized to reduce P concentrations in the effluent water in the first week following P application.

1. Introduction Today, eutrophication of surface water has become a worldwide environmental problem. In most of the freshwater ecosystems limited by phosphorus (P), agricultural sources of P have been identified as one primary contributor (King et al., 2015; Sharpley et al., 2015). In China, agriculture is estimated to contribute over 60% of the annual gross P loads to surface waters (Chen, 2007). This proportion of P load is predicted to even increase with the continuing intensification of agricultural production as driven by the national food security. In particular, national concerns have arisen over unreasonable use of P in agricultural production (Li et al., 2015), stressing the great need of evaluating the impacts of agricultural P management strategies on water quality (Sharpley et al., 2016). Surface runoff plays a predominant role in P loss from most of the

⁎

Corresponding author. E-mail addresses: [email protected], [email protected] (L. Zhai).

http://dx.doi.org/10.1016/j.agee.2017.05.015 Received 15 February 2017; Received in revised form 6 April 2017; Accepted 15 May 2017 0167-8809/ © 2017 Elsevier B.V. All rights reserved.

upland soils (Schroeder et al., 2004; Smith et al., 2007; Wallace et al., 2013) and the flooded soils (Liu et al., 2016). Commonly, P in surface runoff consists of both fertilizer P recently added to the soil and the soil P in the established pools (Withers et al., 2003; Liu et al., 2012). The P recently applied becomes instantly mobile after interaction with rainfall, and it constitutes a short-term source of P loss (Withers et al., 2003; Susumu et al., 2016). Depending on the type of P compounds and the presence of sorptive materials (e.g., edges of pedogenic phyllosilicates or sesquioxides that constitute the majority of pH-dependent charges in soils mineral components), water solubility of P fertilizers and potential of P loss may differ greatly. In a paired catchment study, for instance, McDowell et al. (2010) found that application of reactive phosphate rock reduced filterable reactive P by 58% and total P by 38% compared to application of superphosphate. When fertilizer is overused, the surplus of P exceeding crop needs is

Agriculture, Ecosystems and Environment 245 (2017) 112–123

L. Hua et al.

sorbed by soil materials (such as soil aggregate, calcium (Ca2+), iron (Fe3+), and aluminum (Al3+) cations), but it builds up soil P pools and becomes a long-term source of P loss (Cao and Zhang, 2004; Zhang et al., 2008). Hesketh and Brookes (2000) observed a significant correlation between plant available P (soil Olsen P) and soil solution P extracted by calcium chloride (CaCl2) solution (CaCl2 extractable P). Based on this relationship, they identified a ‘change point’ of soil OlsenP level, above which concentrations of soil CaCl2-P increased rapidly. Soil solution P can be transferred to runoff by water movement on the soil surface or to drainage by percolating water in the soil profile, and the change point concept has been used to assess potential risks of P loss from soils (Kleinman et al., 2007). Rice (Oryza sativa L.) is a staple food for roughly one third of the world’s population (Kazunori et al., 2016). In China, rice is planted on 27% of the total arable land area and it accounts for roughly 38% of the national gross grain production (Zhu et al., 2013). Paddy rice is widely grown on flat, flooded fields in regions with extensive water networks. Owing to their close or even direct interaction with water networks, paddy systems are presumably critical sources of nutrient losses to water (Morteza et al., 2016). Several previous studies identified P losses from paddy production systems as an important cause of eutrophication in the local, enclosed lakes in China (e.g., Lee et al., 2007; Zhang et al., 2007a,b). However, there is lack of understanding whether fertilizer or soil P pool is the major source of P in runoff from conventionally managed paddy fields in China. Another question is if the source of P in runoff vary with regions with different climates and soils. Also, questions remain on when and where to target management strategies to combat P runoff from paddy fields. Differing from upland cropping systems, paddy rice systems are usually established upon soils with deep, water-impermeable plough pans, and where field berms are constructed to pond water on the soil surface (the so-called field ponding water; FPW). During the rice growing season, runoff is generated when the volume of rainfall plus the volume of the FPW exceeds the capacity of field berms to enclose water (Xu and Wang, 2008; Si et al., 2000). Liu et al. (2016) found that concentrations of both total P and dissolved P (< 0.45 μm) in surface runoff were significantly correlated with the respective P forms in the FPW. Thus, monitoring of the FPW provides an efficient way of indicating risks of P runoff loss from paddy rice systems. One objective of the present study was to evaluate the contributions of P fertilizer and soil P to the P concentrations and dynamics in the FPW, under conventional P management practices in rice and wheat (Triticum aestivum L.) rotation systems in different Chinese rice producing regions. Another objective was to identify critical soil P level that can be used to indicate elevated risks of P loss from different types of paddy soils. Moreover, the study was to identify critical periods of high P concentrations in the FPW. The results will contribute to improving assessment of risks of P loss from rice producing systems.

farmers with conventional farming practices for many years, and they represent well the paddy fields in the respective provinces. Paddy soils are a group of soils formed on river sediments and interfered by groundwater movement and farming activities, and they are mainly distributed in flooded river alluvial plain, delta, and low terrace. The present study consisted of four different types of paddy soils, representative of the respective study regions. In China, these soils are widely referred as hydragric paddy soil in Hubei, purple clay soil in Zhejiang, aquic soil in Jiangsu, and purple soil in Sichuan and Yunnan (Cooperative research group on Chinese Soil Taxonomy, 2003). According to the FAO soil classification system, the soils were Cumulic Anthrosols in Hubei, Plinthic Alisols in Zhejiang, Eutric Gleysols in Jiangsu, and Calcaric Regosols in Sichuan and Yunan (ISRIC, 2014). The hydragric paddy soil is characterized by obvious deposition of manganese and iron in the profile. The purple clay soil is a heavy clay soil with low base cation saturation. The aquic soil is developed on a poorly drained landscape. The purple soil is characterized by its uniform purple or purple-red color of the entire soil profile, developed under subtropical climatic conditions. Despite both soils in Sichuan and Yunnan are a purple soil, the Sichuan soil had much higher soil bulk density and soil pH value than the Yunnan soil. Indeed, the Sichuan purple soil was alkaline (pH 8.1), while all other soils were acidic (pH 5.9–6.9). Plant available P in soil was determined according to Olsen et al., 1954, a method being widely used for determining P in Chinese paddy soils (e.g.,Cao et al., 2004; Tian et al., 2006; Wang et al., 2012; Liu et al., 2016). The Olsen P content varied from 10.1 mg P kg−1 in Hubei hydragric paddy soil to 20.5 mg P kg−1 in Yunnan purple soil. The Hubei soil also had the lowest degree of P saturation (DPS; 21%) among all the soils, while the Jiangsu soil had the highest DPS value (53.1%) owing to low iron and aluminum contents in the soil. Detailed location characteristics and soil physical and chemical properties at the start of the field experiments are presented in Table 1. 2.2. Field experiments In each field experiment, rice grew from June to October, and winter wheat grew from November to May of the next year. The cultivars of rice were Guangliangyou-476 in Hubei, Shaojing-18 in Zhejiang, 9998-3 in Jiangsu, Chuangxiangyou-9838 in Sichuan, and Chujing-28 in Yunnan. The cultivars of wheat were Zhengmai-9023 in Hubei, Yangmai in Zhejiang, Yangmai-11 in Jiangsu, Chuanmai-42 in Sichuan and Yunmai-47 in Yunnan, respectively. During the rice growing season, rice seedlings were transplanted between late May and early June, about 2 days after application of basal fertilizers. The fields were waterlogged (up to a depth of 5 to 15 cm) throughout the growing season except for a 7-day summer drainage period in the middle of July that was carried out for soil aeration. However, the fields were intermittently drained for the promotion of seedling establishment (in June) and for herbicide application if necessary. A complete randomized design with two treatments, each replicated in three field plots (20–40 m2), was used in the field. The treatments were conventional rate of P fertilizer in accordance to farmers’ practices in each province, and no P addition as the control. All sites had the same conventional P rate for rice (32 kg P ha−1), while the P rates for wheat ranged from 20 kg P ha−1 at the Jiangsu site to 42 kg P ha−1 at the Sichuan site. The P fertilizer used was single superphosphate, consisting of mainly monocalcium phosphate (Ca(H2PO4)2) and gypsum (CaSO4), and also a small amount of phosphoric acid. In term of nutrient content, the fertilizer contains 7–9% of P, 18–21% of calcium, and 11–12% of sulfur, and it is highly water soluble (http://www.ipni. net/specifics). The same, conventional rates of nitrogen (urea) and potassium (potassium sulfate) were applied in all treatments to ensure supply of nitrogen and potassium is sufficient for the crops. All P and potassium, and 60% of nitrogen were applied as basal fertilizers before transplanting of rice or at seeding of wheat, while the rest 40% of nitrogen fertilizer was top-dressed to rice at the time of heading stage,

2. Materials and methods 2.1. Experimental sites and soil properties This study was conducted on rice-wheat double cropping, one major rice production system in China. This system is mainly distributed in the Yangtze River Basin Area and the Southeast Coastal Area (Fig. 1), which annually produces over 150 million Mg of rice grains, or 80% of the national gross rice production (2013 County/City Agriculture Statistics Data). A total of 10 site-year field experiments were conducted from 2012 to 2013 at five experimental sites located in five major rice producing provinces; from west to east, namely, Dali of Yunnan Province, Ziyang of Sichuan Province, Qianjiang of Hubei Province, Changshu of Jiangsu Province, and Shaoxing of Zhejiang Province, respectively. The sites represented different climatic conditions with annual precipitation ranging from 732 mm in Yunnan to 1461 mm in Zhejiang. All experimental fields had been managed by 113

Agriculture, Ecosystems and Environment 245 (2017) 112–123

L. Hua et al.

Fig. 1. Location of the five experimental sites selected to represent rice-wheat double cropping systems in China. The background image illustrates distribution of main rice production regions in China: the Northeast Rice Producing Area represented by rice monoculture, and the Yangtze River Basin Rice Producing Area and Southeast Coastal Rice Producing Area represented by double cropping of flooded rice and upland crop (Rice Regional Layout Planning 2008–2015).

risk levels of dissolved P loss from the five experimental soils, relationships between soil Olsen-P concentrations and soil CaCl2-P concentrations were established through a series of laboratory incubation experiments. The incubation experiments included nine treatments (by three replicates), with addition of different rates of K2HPO4 to each soil sampled at the start of the field experiments in 2012 and prepared as described in Section 2.4. Specifically, 20 g soil was placed in a 50 mL incubation vial, where soil moisture content was maintained at 50% of field capacity for 7 days at 25 °C, and then K2HPO4 solution

and to wheat at the time of elongation stage. During rice growing seasons in both 2012 and 2013, FPW was sampled on 1, 3, 5, 7 and 9 days after application of P fertilizer at Hubei, Zhejiang and Jiangsu sites, and on 1, 3, 5, 9 and 15 days after P application at Sichuan and Yunnan sites, respectively.

2.3. Laboratory incubation experiments To obtain a change point of soil Olsen-P that can be used to assess

Table 1 Site characteristics and soil (0–20 cm) physical and chemical properties determined at start of the field experiment in 2012. Characteristics

Hubei

Zhejiang

Jiangsu

Sichuan

Yunnan

Latitude Longitude Annual precipitation (mm) Annual mean temperature (°C) Soil classification in China Soil classification in FAO Soil texture Soil bulk density (g cm−3) pH (2.5, water/soil) Organic matter (g kg−1) Total P (g P kg−1) Olsen-P (mg P kg−1) Mehlich–3 P (mg P kg−1) Mehlich-3 Fe (mg Fe kg−1) Mehlich-3 Al (mg Al kg−1) DPS (%)

30°23′N 112°37′E 1220 16.3 Hydragric paddy soil Cumulic Anthrosols Loam 1.07 6.90 22.8 0.78 10.1 272 337 989 20.5

30°02′N 121°42′E 1461 16.5 Purple clay soil Plinthic Alisols Clay 0.95 5.90 23.1 0.93 15.5 351 412 1135 22.7

31°29′N 120°05′E 1066 15.8 Aquic soil Eutric Gleysols Loam 1.02 6.70 31.0 0.53 17.5 281 230 299 53.1

30°02′N 103°10′E 976 17.4 Purple soil Calcaric Regosols Clay 1.32 8.10 21.0 0.93 16.4 484 343 1197 31.4

25°26′N 102°28′E 732 13.9 Purple soil Calcaric Regosols Clay 1.03 6.50 21.5 0.64 20.5 356 260 876 31.3

Note: DPS, degree of phosphorus saturation.

114

Agriculture, Ecosystems and Environment 245 (2017) 112–123

L. Hua et al.

The significance of difference between treatments was determined by one-way analysis of variance (ANOVA) as proposed by Djodjic et al. (2004) and Hanrahan et al. (2009). The differences in mean values were tested using Duncan’s multiple-range method, and considered as statistically significant at P < 0.05 (Djodjic et al., 2004; Hanrahan et al., 2009). The relationship between soil CaC12-P and soil Olsen-P was modelled by the two segment linear model of R-console (Girard, 2013; Xi et al., 2016):

was added at rates of 0, 30, 40, 60, 100, 160, 200, 240, 300 and 400 mg P kg−1 soil, respectively. The soils were incubated for 7 days under room temperature (25 °C) and thereafter air-dried to complete a cycle of incubation, which was repeated for four times until P concentration in soils reached equilibrium. After incubation is completed, the soils were air-dried and ground to pass through a 2-mm sieve before analysis of the concentrations of Olsen-P and CaCl2-P. 2.4. Sampling and analysis of soil, crop and field ponding water At each site, soils were collected from the plow layer (0 to 20 cm depth) using a 5 cm internal diameter auger, at the start of the field experiment in 2012 and at the end of the experiment in 2014, respectively. Five soil samples were randomly taken from each field plot and mixed together in a plastic bag to produce a composite sample. All soil samples were air-dried at 50 °C and ground to pass through a 2 mm sieve prior to use. At harvest of rice and wheat, plant samples were collected from a representative area of 1 m2 in each plot. Grains and straws were separated, air-dried at 50 °C, and ground to pass through a 0.15 mm sieve. Plant samples were weighted both before and after being air-dried to determine moisture content in the materials (%). Grain yield and straw biomass were determined based on harvest of the whole area of each plot, and calculated after taking the moisture content into account as [biomass = harvested weight x (1-moisture content%)]. Crop removal of P in grains and straws were calculated as [P uptake (grain or straw) = (P concentration in grain or straw) × biomass]. Concentration of total P in soil, grain and straw were determined using the colorimetric molybdate-ascorbic acid method (Murphy and Riley, 1962), after oxidative digestion of the plant samples with concentrated H2SO4eH2O2 (Tang et al., 2008) and digestion of soil samples with H2SO4eHClO4 (Liu et al., 2010). Plant available P in the soil was measured according to Olsen et al. (1954), with soil extracted with 0.5 M NaHCO3. The P in the soil solution was determined after shaking 10 g of the air-dried soil with 50 mL 0.01 M CaCl2 solution for 30 min (McDowell and Sharpley, 2001). The concentrations of Olsen-P and CaCl2-P in the extracts were determined using the colorimetric molybdate–ascorbic acid method (Murphy and Riley, 1962). Moreover, soil was extracted with Mehlich-3 solution (Mehlich, 1953) to determine Mehlich–3 P (Mehlich, 1984) and easily available contents of Fe and Al in soil with inductively coupled plasma optical emission spectroscopy (ICP-OES, Perkin Elmer, Wellesley, USA) (Crosland et al., 1995). These measurements were used to calculate the degree of P saturation (DPS, %) in soil as [Mehlich–3 P/(Mehlich-3 Fe + Mehlich-3 Al) × 100], with P, Fe, and Al all expressed on a mmol kg−1 basis (Breeuwsma and Silva, 1992). Soil pH was determined with a pH meter (Mettler Toledo Delta 320), with a soil (air-dried) to water (w/w) ratio of 1:2.5 (Rowell, 1994). At sampling of the FPW, a water sample was composed by five subsamples collected from random locations within each field plot. The FPW was mainly sourced from irrigation for the first two weeks after transplanting rice. Before the paddy fields were irrigated to pond water, irrigation water was sampled for each irrigation event. All water samples were stored in 500 mL polyethylene bottles in the dark at 4 °C in an ice box until analysis of P concentrations. Thereafter, concentration of total dissolved P (TDP) in water was measured on filtered samples (0.45 μm) and total P (TP) on unfiltered samples, following a method described by Murphy and Riley (1962) after a complete sulphuric/perchloric acid digestion (Olsen and Sommers, 1982). Concentration of particle-bound P (PP) was calculated as the difference between TP and TDP.

Y = a1X + b1 X ≤ T

(1)

Y = a2X + b2 X ≥ T

(2)

Where, Y is soil CaCl2-P concentration (mg kg−1); X is soil Olsen-P concentration (mg kg−1); b is interception of the regression line; a is slope; T is the critical value of soil Olsen-P (mg kg−1). The SAS software was used to examine the fitness of the data to the models suggested by Xi et al. (2016). 3. Results 3.1. Soil P balance By convention, rice was applied with 32 kg fertilizer P ha−1 at all experimental sites, but different rates of P were applied to winter wheat (Table 2). Thus, P fertilizer rate for the entire rotation varied, by sites, from 52 to 74 kg P ha−1 yr−1 (Table 3). Crop P removal by harvesting rice and wheat ranged from 36 to 49 kg P ha−1 for the control (no P) treatment, and from 43 to 59 kg P ha−1 for the conventional P treatment. The difference in the crop P removal between the two treatments was significant at Hubei and Sichuan sites (P ≤ 0.05), but not at Jiangsu, Zhejiang and Yunnan sites. Overall, 36–49 kg P ha−1 yr−1 was removed from soils in the no P treatment, while the conventional P treatment had a positive P surplus of 2–28 kg ha−1 yr−1. After two years of experiments, soil Olsen-P content in conventional P treatments was 16–169% higher than that in the control, where the difference was significant at all sites, except Hubei with the lowest initial soil Olsen-P content among all sites (Table 3). 3.2. Relationships between soil Olsen-P and soil CaCl2-P For all five paddy soils, the correlation between CaC12-P and OlsenP was well fitted by the two-segment linear model, where soil CaC12-P increased relatively slowly at lower soil Olsen-P, but rapidly after Olsen-P reached the change point (Fig. 2). The change point values of soil Olsen-P were 30 mg kg−1 at Jiangsu, 58 mg kg−1 at Zhejiang, 87 mg kg−1 at Yunnan, 94 mg kg−1 at Hubei, and 172 mg kg−1 at Sichuan, respectively. At the change points, the CaCl2-P concentrations were 1.43 mg kg−1, 1.42 mg kg−1, 2.98 mg kg−1, 0.66 mg kg−1, and 2.96 mg kg−1, respectively. At the start of the field experiments, initial soil Olsen-P contents (Table 1) were all well below their respective change point values (Fig. 2), indicating a generally low risk of P loss from the soil P pools. Notably, after two years of P application at conventional rates, soil Olsen-P content approached the change point at the Jiangsu site. 3.3. Temporal dynamics of TP and TDP in the field ponding water For all ten site-year experiments, the conventional P treatment significantly elevated TP concentrations in the FPW compared to the no P treatment (Fig. 3). In the conventional P treatment, mean TP concentrations in the FPW over the entire monitoring period ranged from 0.64 mg L−1 in the Zhejiang purple clay soil to 3.92 mg L−1 in the Hubei hydragric paddy soil. In contrast, mean TP concentrations in the FPW ranged from 0.027 mg L−1 in aquic soil to 0.51 mg L−1 in

2.5. Data analysis All statistical analyses were performed using the SAS statistical software (Version 9.2). Data are presented as mean ± standard error. 115

Agriculture, Ecosystems and Environment 245 (2017) 112–123

L. Hua et al.

Table 2 Rates of P, N and K applied to rice and wheat at each experimental site. Sites

Hubei Zhejiang Jiangsu Sichuan Yunnan

Sites

Hubei Zhejiang Jiangsu Sichuan Yunnan

Treatments

Rice

No P Conventional No P Conventional No P Conventional No P Conventional No P Conventional

P P P P P

Treatments

Wheat

P (kg ha−1)

N (kg ha−1)

K (kg ha−1)

P (kg ha−1)

N (kg ha−1)

K (kg ha−1)

0 32 0 32 0 32 0 32 0 32

210 210 200 200 225 225 150 150 180 180

75 75 125 125 87 87 62 62 75 75

0 36 0 30 0 20 0 42 0 36

165 165 150 150 180 180 135 135 180 180

100 100 125 125 75 75 87 87 75 75

Rice

No P Conventional No P Conventional No P Conventional No P Conventional No P Conventional

P P P P P

Wheat

P (kg ha−1)

N (kg ha−1)

K (kg ha−1)

P (kg ha−1)

N (kg ha−1)

K (kg ha−1)

0 32 0 32 0 32 0 32 0 32

210 210 200 200 225 225 150 150 180 180

75 75 125 125 87 87 62 62 75 75

0 36 0 30 0 20 0 42 0 36

165 165 150 150 180 180 135 135 180 180

100 100 125 125 75 75 87 87 75 75

Temporal dynamics of TDP concentrations in the FPW followed similar trends to TP concentrations (Fig. 4). For all experiments, maximum TDP concentrations were observed one day after P application, despite they varied, by sites, from 0.03 to 0.56 mg L−1 in the no P treatment and from 0.16 to 4.49 mg L−1 in the conventional P treatment. The concentrations abruptly declined with time, and stabilized on day 3 in the no P treatment and five days after P applications in the conventional P treatment. Within one week after P application, TDP concentrations in the conventional P treatment were all significantly (P ≤ 0.05) higher than those in the no P treatment (Fig. 4). With regard to sites, the Hubei site (mean: 2.17 mg L−1) and the Sichuan site (mean: 1.75 mg L−1) had higher TDP concentrations than the other sites (mean: 0.43–1.48 mg L−1), all of the same conventional P treatment. In the irrigation water, which was the major source of the FPW during our monitoring period, TP concentrations ranged from 0.05 to 0.2 mg L−1, and TDP concentrations ranged from 0.02 to 0.15 mg L−1. In general, the P concentrations were similar to those measured in the FPW of the unfertilized treatments, but were much smaller than those in the fertilized treatments. This comparison additionally proves that the high P concentrations in the FPW of the fertilized treatments were derived from the fertilizer applied.

hydragric paddy soil in the no P treatment. Unexpectedly, there was a weak relationship between TP concentrations in the FPW and P content in soils. Specifically, differences in initial TP concentrations in the FPW between soils in the no P treatment (Fig. 3) could not be attributed to either soil Olsen-P or Mehlich–3 P content (Table 1). It is likely that other factors overshadowed the effects of soil P content on FPW TP concentrations. One possible reason is that crops developed better in high P soils and resulted in less P loss, compared to that in low P soils. Total P concentrations in the FPW decreased with time in both treatments, but the decrease was more abrupt in the conventional P treatment with much higher initial FPW TP concentrations (8.91–1.47 mg L−1) than in the no P treatment (1.11–0.027 mg L−1). In the conventional P treatment, TP concentrations in the FPW commonly stabilized five days after P application, with a decreasing rate of 50.6–79.9% from day 1 to day 5. Notably, TP concentrations on day 9 in the conventional P treatment were still significantly higher than the concentrations on day 1 in the no P treatment for most experimental sites and years. In the no P treatment, FPW TP concentrations fluctuated within a relatively small range of 1.11–0.02 mg L−1, despite significantly higher TP concentrations on day 1 than on day 9 in four of the five soils.

Table 3 Mean annual soil P balance and soil Olsen-P content (mean ± s.e.) in different treatments at five experimental sites. Soil Olsen-P content was measured at the end of the two-year experiment in 2014. Sites

Treatments

Hubei

No P Conventional No P Conventional No P Conventional No P Conventional No P Conventional

Zhejiang Jiangsu Sichuan Yunnan

P P P P P

Total P input (kg P ha−1 yr−1)

Crop removal (kg P ha−1 yr−1)

P surplus (kg P ha−1 yr−1)

Olsen-P (mg kg−1)

0 68 0 62 0 52 0 74 0 68

38 43 46 51 49 50 36 46 45 54

−38 25 −46 11 −49 2 −36 28 −45 14

8.70 10.1 15.5 29.4 17.5 26.6 12.5 33.6 20.5 38.2

± ± ± ± ± ± ± ± ± ±

1.5 0.3 1.9 5.1 1.5 4.2 1.0 3.7 2.7 8.4

b a a a a a b a a a

Means (n = 3) between treatments were compared for each site, and noted with different letters when the difference was significant (P ≤ 0.05).

116

± ± ± ± ± ± ± ± ± ±

0.54 1.14 0.77 2.79 0.21 1.33 2.66 3.53 4.04 1.56

a a b a b a b a b a

Agriculture, Ecosystems and Environment 245 (2017) 112–123

L. Hua et al.

Fig. 2. Relationships between soil P extracted by 0.01 M CaCl2 and by the Olsen method for the five experimental soils. (A) Hubei, Jianghan Plain; (B) Zhejiang, Taihu Plain; (C) Jiangsu, Taihu Plain; (D) Sichuan, Sichuan Basin; and (E) Yunnan, Yungui Plateau. Note: The solid point indicates the change point of the curve. R2 is the correlation coefficient of the two segment linear model. **denoted significance of correlation at P ≤ 0.01 and * denoted significance of correlation at P ≤ 0.05.

3.5. Contribution of fertilizer and soil to P concentration in the field ponding water

3.4. Proportions of different P forms in the field ponding water Proportions of TDP and PP in TP in the FPW clearly differed between the two P management strategies. In general, TP was dominated by TDP in the conventional P treatment. Within the 7 days after P application, the percentage of TDP in TP in the FPW ranged from 43% to 84%. In contrast, PP was the main form of P in the no P treatment, where it accounted for 45–88% of TP in the FPW across all experimental sites (Fig. 5). Even so, proportions of TDP in TP tended to decrease with time in both treatments.

Comparisons between the two P treatments with and without P applications allowed us to identify contributions of fertilizer and soil to the P concentrations in the FPW under conventional P application to rice-wheat double cropping systems in China. Basically, the P measured in the FPW of the control was attributed to the P contributed by soil, and the difference in the P measured in the two treatments was presumably contributed by fertilizer. Overall, fertilizer was the pre117

Agriculture, Ecosystems and Environment 245 (2017) 112–123

L. Hua et al.

Fig. 3. Temporal dynamics of TP concentrations in field ponding water with different P treatments in five paddy rice production regions of China: (A) Hubei, Jianghan Plain; (B) Zhejiang, Taihu Plain; (C) Jiangsu, Taihu Plain; (D) Sichuan, Sichuan Basin; and (E) Yunnan, Yungui Plateau. Each point shows mean TP concentration of three replicated plots, with error bars representing standard errors. Within each region and year, mean TP concentrations were compared for all treatments and days, with statistical differences identified by symbols a, b and c.

waste of nutrients, and leads to environmental problems (Andreasen et al., 2006; Lee et al., 2007; Kleinman et al., 2011), has been an enduring concern for the Chinese agriculture. In the present study, application of P fertilizers at conventional rates did not significantly improve crop P removal (Table 3) or crop yields (data not shown) at the sites of Jiangsu, Zhejiang and Yunnan, compared to the no P application. This is probably because there was already sufficient P in their soils resulting from historical P surplus. Indeed, soil Olsen-P content at the three sites fell in the range of 15.5–20.5 mg kg−1 at the start of the field experiments (Table 1). In a previous study, Lu (2000) found no improvement of rice yield when P was applied to paddy soils with > 10 mg Olsen-P kg−1. However, others found large yield reductions in the second year following no P application (Xu and Xu, 2013), the trend of which was supported by our annual yield data (data not shown). Therefore, repeated applications of P seem to be appropriate to sustain long-term soil fertility. However, the P application rates should be optimized to lower P concentrations in the FPW. Both TP and TDP concentrations in the FPW peaked one day after P application, declined rapidly within the first three days, and became relatively stable after day 5 (Figs. 3 and 4). Since five days after P application, the dominant form of P in the FPW was transformed from

dominant source of TP and TDP in the FPW at all sites. Roughly, 46–93% of TP in the FPW was derived from fertilizer, while the rest 7–54% was from soil (Fig. 6). With regard to TDP, fertilizer contributed 59–97% and soil contributed 3–41% only. Over time, relative contribution by soil to TP and TDP in the filed ponding water tended to increase but also fluctuated. 4. Discussion Currently, the primary source of dissolved P in runoff from Chinese paddy fields is the fertilizer P. In this study, conventional P applications significantly elevated P concentrations in the FPW, in particular in the form of TDP by about 20 times, compared to no P application (Fig. 4). In a paddy field plot study, Liang et al. (2005) observed that total nitrogen and TP concentrations in surface runoff respectively reached as high as 22.2 mg L−1 and 4.8 mg L−1 following rainstorms. They attributed the high nutrient concentrations to the fertilizers applied a few days prior to the rainstorms. Given heavy rainstorms, it is expected that large amounts of P in the FPW as observed in our fertilized treatments would end up in surface runoff. Overuse of fertilizers, which results in luxury consumption and 118

Agriculture, Ecosystems and Environment 245 (2017) 112–123

L. Hua et al.

Fig. 4. Temporal dynamics of TDP concentrations in field ponding water with different P treatments in five paddy rice production regions of China: (A) Hubei, Jianghan Plain; (B) Zhejiang, Taihu Plain; (C) Jiangsu, Taihu Plain; (D) Sichuan, Sichuan Basin; and (E) Yunnan, Yungui Plateau. Each point shows mean TP concentration of three replicated plots, with error bars representing standard errors. Within each region and year, mean TP concentrations were compared for all treatments and days, with statistical differences identified by symbols a, b and c.

five days after P application. During this period, TP concentrations in the FPW were over 0.06 mg L−1 for all fivesoils, and they were particularly as high as 2.6–8.9 mg L−1 at the Hubei and Sichuan sites. In a study on red soil, Qu et al. (2016) found that both the concentration magnitudes and duration of the high concentrations in the FPW were associated with P fertilizer rates. They observed that the TP concentration in the FPW peaked at 0.73 mg L−1 and dropped to 0.27 mg L−1 two days after P application at rates of 25–30 kg P ha−1, but remained as high as 0.5 mg L−1 throughout their 9-day monitoring period when 100–130 kg P ha−1 was applied. Given coincidence of heavy rainfall events with high P concentrations in the FPW, runoff P loss would be a great concern for quality of nearby waters (Cao et al., 2005; Tian et al., 2006; Wu et al., 2012; Liu et al., 2016). Elsewhere, Sharpley et al. (2007), and Sims and Kleinman (2005) also pointed out the importance of controlling runoff P loss that occurred close to application of P to upland crops. For instance, Li et al. (2017) found

TDP to PP, which is not instantly available for the algae growth. The decline in TDP concentrations may be due to precipitation of phosphate ions with cations such as iron and calcium dissolved in the water (Reddy et al., 1999), and sorption of dissolved P to suspended sediment materials that were in the largest amounts when the soil was disturbed by irrigation but gradually settled down on the soil surface (Sharpley et al., 1981). Some of the dissolved P may be biological retained by periphyton, microorganisms, and macrophytes including rice plants (Adey et al., 1993; Svendsen and Kronvang, 1993). The reduction in both TP and TDP concentrations were also partly attributed to dilution by rainfall water. Indeed, an average of 60 mm rainfall water was observed throughout the entire monitoring period at each experimental site, with a considerable proportion of rainfall occurring at the later stage of the monitoring period. The temporal dynamics of TP and TDP concentrations in the FPW point to the need of avoiding possible runoff losses of P during the first 119

Agriculture, Ecosystems and Environment 245 (2017) 112–123

L. Hua et al.

Fig. 5. Proportions of TDP and PP in TP in field ponding water at different days after conventional and no P applications in five paddy rice production regions of China: (A) Hubei, Jianghan Plain; (B) Zhejiang, Taihu Plain; (C) Jiangsu, Taihu Plain; (D) Sichuan, Sichuan Basin; and (E) Yunnan, Yungui Plateau.

the Sichuan soil than the Yunnan soil, even though both are purple soil (Calcaric Regosols). The Sichuan soil had a pH of 8.1, which indicates stronger P fixation by calcium than other soils with a pH of 5.9–6.9 (Busman et al., 2002). However, this was not observed in the present study, probably because the effect of pH on P solubility was overshadowed by the P application rate. Moreover, Zhai et al. (2015) found that binding of P by Ca and Mg in alkaline soils is often weaker than binding by Fe and Al in acidic soils. Our results did reveal an increase in soil available P content and thus an elevated P loss potential after two years continuously conventional rates of P application to all soils. Specifically, soil Olsen-P content increased by 0.7 to 10 mg kg−1 annually. Soils may become saturated with P over time. For example, the Jiangsu aquatic soil had an Olsen-P change point of 29.5 mg kg−1, while it had an initial Olsen-P content of 17.5 mg kg−1, and an annual accumulation rate of 4.5 mg Olsen-P kg−1 under conventional P treatment. Presumably, its Olsen-P content will exceed the change point in 3 years. Previous studies demonstrated that the potential risk of P dissolving from plow-layer soil to water substantially increased when the content of soil available P exceeded the change point (Hesketh and Brookes, 2000; Kleinman et al., 2007). Thus, our results demonstrate the need of reducing conventional P application rates for the long-term benefits of water quality in rice

that TP concentrations in the agricultural runoff can reach as high as 1.2 mg L−1 in storm events. In China, most rice is produced in rotation with upland crops such as winter wheat and oilseed rape. Therefore, immediate P loss after fertilizer application needs to be controlled in the entire paddy production systems. All five experimental soils had a relatively low risk of P dissolution. Soil available P contents (10–40 mg kg−1) were lower than the threshold values for an abrupt increase in P loss potential (29–172 mg kg−1). As a result, P concentrations in the FPW did not correlate well with content of initial soil available P. Therefore, the varying TP and TDP concentrations in the FPW with experimental sites may be attributed to different amounts of P inputs, and soil sorption characteristics (Zhang et al., 2015). The highest TP and TDP concentrations were observed in the conventional P treatment at Hubei and Sichuan sites. The P rates were the same during rice growing seasons at the five study sites, but Hubei site (36 kg ha−1) and Sichuan site (42 kg ha−1) received the largest amount of P fertilizer during wheat growing season among all sites, resulting in an annual P surplus of 25 and 28 kg P ha−1, respectively. Song et al. (2007) reported that frequent drought and flooding events in paddy soils could lead to significant P release by desorption of P that had previously adsorbed or precipitated to soil particles. The higher P rates may also explain the greater solubility of 120

Agriculture, Ecosystems and Environment 245 (2017) 112–123

L. Hua et al.

Fig. 6. Contribution of fertilizer and soil to TP and TDP in field ponding water under conventional P management in five paddy rice production regions of China: (A) Hubei, Jianghan Plain; (B) Zhejiang, Taihu Plain; (C) Jiangsu, Taihu Plain; (D) Sichuan, Sichuan Basin; and (E) Yunnan, Yungui Plateau.

robust P recommendation programs for paddy systems.

production regions. Apart from the reduction of P fertilizer rate, the type of P fertilizer and timing of fertilization may also need to be considered, such as in Hubei and Sichuan. Nonetheless, P needs to be supplemented to replenish the P removed by crops in the systems with no P applications. While several regional studies in China showed that most of the paddy soils in these studies had available P contents ranging from 13 to 25 mg L−1, a considerable proportion of soils contain higher or lower level of P (Wang et al., 2014a,b, 2012; Xie et al., 2001). Such a diverse range of P contents brings difficulties in P recommendations for rice. Moreover, P management during other crop seasons in rotation with rice such as winter wheat also needs attention, as P rates applied to these crops, which are often in excess of crop needs, can also largely contribute to build-up of P pools in paddy soils. To minimize P loss, P application rate should be based on crop needs and modified by the P amounts already in the soil (Sharpley, 2016). For instance, farmers are not allowed to apply more P than crop removal rate to the fields where soil test P concentration is already high (Olsen-P > 26 mg kg−1) in Great Britain (Defra, 2010). In China, currently established nutrient recommendations are mainly focused on crop production, with relatively less consideration of water quality. Work remains to establish

5. Conclusions Sustainable development requires improved water quality in production systems (Zhao et al., 2006; Wang et al., 2014a,b). In China, where paddy rice is widely grown, management of P runoff plays an important role in achieving agricultural sustainability. Under conventional P applications representing regional P management scenarios, all five experimental soils had plant available P contents below their respective threshed values for promoted P loss potential. As a result, soil P contributed only 7–54% of P concentrations in the FPW. In contrast, P application at conventional fertilizer rates accounted for 46–93% of P concentrations in the FPW. After P application, total P concentrations in the FPW remained significantly higher than those in the control treatment throughout the entire monitoring period. Total P concentrations were the highest within the five days after P application in all the experiments, indicating the need of best management practices for mitigating P loss from this time period. Moreover, the excessive P rate applied to other crops in rotation with rice should be reduced to avoid build-up of soil P pools, which is particularly 121

Agriculture, Ecosystems and Environment 245 (2017) 112–123

L. Hua et al.

Liang, X., Tian, G., Li, H., Chen, Y., Zhu, S., 2005. Study on characteristic of nitrogen and phosphorus loss from rice field by natural rainfall runoff. J. Soil Water Conserv. 19, 59–63 (In Chinese). Liu, E.K., Yan, C.R., Mei, X.R., He, W.Q., Bing, S.H., Ding, L.P., et al., 2010. Long-term effect of chemical fertilizer, straw, and manure on soil chemical and biological properties in Northwest China. Geoderma 158, 173–180. Liu, J., Aronsson, H., Ulen, B., Bergstrom, L., 2012. Potential phosphorus leaching from sandy topsoils with different fertilizer histories before and after application of pig slurry. Soil Use Manage. 28, 457–467. Liu, J., Zuo, Q., Zhai, L.M., Luo, C.Y., Liu, H.B., Wang, H.Y., et al., 2016. Phosphorus losses via surface runoff in rice-wheat cropping systems as impacted by rainfall regimes and fertilizer applications. J. Integra. Agric. 15, 667–677. Lu, R.K., 2000. The Analysis Methods of Soils and Their Agrochemistries. China Agricultural Science and Technology Press, Beijing. McDowell, R.W., Sharpley, A.N., 2001. Phosphorus losses in subsurface flow before and after manure application to intensively farmed land. Sci. Total Environ. 278, 113–125. McDowell, R.W., Littlejohn, R.P., Blennerhassett, J.D., 2010. Phosphorus fertilizer form affects phosphorus loss to waterways: a paired catchment study. Soil Use Manage. 26, 365–373. Mehlich, A., 1984. Mehlich 3 soil test extractant: a modification of Mehlich 2 extractant. Commun Soil Sci. Plant. 15, 1409–1416. Morteza, Y., Javad, S., Mahmood, S.S., 2016. On dealing with the pollution costs in agriculture: a case study of paddy fields. Sci. Total Environ. 556, 310–318. Murphy, J., Riley, J.R., 1962. A modified single solution method for the determination of phosphate in natural waters. Anal. Chem. Acta. 27, 31–36. Olsen, S.R., Sommers, L.E., et al., 1982. Phosphorus. In: Page, A.L. (Ed.), Methods of Soil Analysis, Part 2, Agron Monoger. 9, 2nd ed. ASA and SSSA, Madison, WI, pp. 403–429. Olsen, S.R., Cole, C.V., Watanabe, W.S., Dean, L.A., 1954. Estimation of Available Phosphorus in Soil by Extraction with Sodium Bicarbonate. Qu, H.P., Zhou, L.Q., Huang, M.F., Huang, J.S., Wei, Y.L., Xie, R.L., et al., 2016. Phosphorus balance in paddy soils and its environmental effect under different phosphorus application rates. J. Plant Nutri. Ferti. 22 (1), 40–47. Reddy, K.R., Kadlec, R.H., Flaig, E., Gale, P.M., 1999. Phosphorus retention in streams and wetlands: a review. Crit. Rev. Environ. Sci. Tech. 29, 83–146. Rowell, D.L., 1994. The Meaning of pH, and Its Measurement. pp. 159–161. Sharpley, A.N., Menzel, R.G., Smith, S.J., Rhoades, E.D., Olness, A.E., 1981. The sorption of solute phosphorus by soil material during transport in runoff from cropped and grassed watersheds. J. Environ. Qual. 10, 211–215. Sharpley, A.N., Herron, S., Daniel, T., 2007. Overcoming the challenges of phosphorusbased management in poultry farming. J. Soil Water Conserv. 62, 375–389. Sharpley, A.N., Bergstrom, L., Aronsson, H., Bechmann, M., Bolster, C.H., Borling, K., et al., 2015. Future agriculture with minimized phosphorus losses to waters: research needs and direction. AMBIO 44, 163–179. Sharpley, A.N., Kleinman, P., Jarvie, H., Flaten, Don., 2016. Distant views and local realities: the limits of global assessments to restore the fragmented phosphorus cycle. Agric. Environ. Lett. 1, 160024. http://dx.doi.org/10.2134/ael2016.07.0024. Sharpley, A.N., 2016. Managing agricultural phosphorus to minimize water quality impacts. Sci. Agric. 73, 1–8. Si, Y., Wang, S., Chen, H., 2000. The loss of farmland nitrogen and phosphorus and water eutrophication. Soils 188–193 (In Chinese). Sims, J.T., Kleinman, P.J.A., 2005. Managing agricultural phosphorus for environmental protection. In: Sims, J.T., Sharpley, A.N. (Eds.), Phosphorus: Agriculture and the Environment. American Society of Agronomy Madison, WI, USA (p. 1021–1068). Smith, D.R., Owens, P.R., Leytem, A.B., Warnemuende, E.A., 2007. Nutrient losses from manure and fertilizer applications as impacted by time to first runoff event. Environ. Pollution. 147, 131–137. Song, K.Y., Zoh, K.D., Kang, H., 2007. Release of phosphate in a wetland by changes in hydrological regime. Sci. Total Environ. 380 (1–3), 13–18. Susumu, S.A., Seiko, H., Takayuki, U., Takeshi, Y., Sadahiro, Y., Satoshi, Y., 2016. Excessive application of farmyard manure reduces rice yield and enhances environmental pollution risk in paddy fields. Arch. Agron. Soil Sci. 62, 1208–1221. Svendsen, L.M., Kronvang, B., 1993. Retention of nitrogen and phosphorus in a Danish Lowland: implications for the export form the watershed. Hydrobiologia 251, 123–135. Tang, X., Li, J.M., Ma, Y.B., Hao, X.Y., Li, X.Y., 2008. Phosphorus efficiency in long term (15 years) wheat-maize cropping systems with various soil and climate conditions. Field Crop Res. 108, 231–237. Tian, Y., He, F., Yin, B., Zhu, Z., 2006. Dynamic changes of nitrogen and phosphorus concentrations in surface water of paddy field. Soils 38, 727–733 (In Chinese). Wallace, C.B., Burton, M.G., Hefner, S.G., DeWitt, T.A., 2013. Effect of preceding rainfall on sediment, nutrients, and bacteria in runoff from biosolids and mineral fertilizer applied to a hayfield in a mountainous region. Agric. Water Manage. 130, 113–118. Wang, S., Zhao, X., Xing, G., Gu, Y., Shi, T., Yang, L., 2012. Phosphorus pool in paddy soil and scientific fertilization in typical areas of Taihu Lake Watershed, China. Soils 44, 158–162 (In Chinese). Wang, W.C., Xu, D.M., Chau, K.W., Lei, G.J., 2014a. Assessment of river water quality based on theory of variable fuzzy sets and fuzzy binary comparison method. Water Resour. Manage. 28, 4183. http://dx.doi.org/10.1007/s11269-014-0738-4. Wang, Y., Zhao, X., Wang, L., 2014b. Accumulation, environmental risk and control of phosphorus in rice/wheat rotation farmland in Taihu Lake watershed. J. AgroEnviron Sci. 33, 829–835 (In Chinese). Withers, J.A., Ulen, Barbro, Stamm, Christian, Bechmann, Marianne, 2003. Incidental phosphorus losses-are they significant and can they be predicted? J. Plant Nutr. Soil Sci. 166, 459–468.

important for the soils with relatively low critical P values. Our results suggest that farmers’ conventional rates of P application in Chinese rice production systems can be reduced to minimize the impact of fertilization on downstream water quality while maintaining current level of grain production. Reduction of P rates across China will also help to alleviate the national or even international P insecurity, provided that global P reserves are predicted to be depleted within 200 years (Cordell et al., 2009). Acknowledgments This study was funded by the National Key Research and Development Program of China (2016YSD0800500); the Special Fund for Agro-scientific Research in the Public Interest (201003014) and the National Natural Science Foundation of China (41203072). References Adey, W., Luckett, C., Jensen, K., 1993. Phosphorus removal from natural water using controlled algal production. Restor. Ecol. 1993, 29–39. Andreasen, C., Litz, A.S., Streibig, J.C., 2006. Growth response of six weed species and spring barley (Hordeum vulgare) to increasing levels of nitrogen and phosphorus. Weed Res. 46, 503–512. Breeuwsma, A., and Silva, S., 1992. Phosphorus fertilization and environmental effects in the Netherlands and the Po region (Italy). Rep 57. Agric. Res. Dept., Soil and Water Research. Wageningen, The Netherlands Busman, L., Lamb, J., Randall, G., Rehm, G., Schmitt, M., 2002. The Nature of Phosphorus in Soils. http://www.extension.umn.edu/agriculture/nutrient-management/ phosphorus/the-nature-of-phosphorus/. Accessed 26 Dec 2016. Cao, Z.H., Zhang, H.C., 2004. Phosphorus losses to water from lowland rice fields under rice–wheat double cropping system in the Tai Lake region. Environ. Geochem. 26, 229–236. Cao, Z.H., Lin, X.G., Yang, L.Z., Hu, Z.Y., Dong, Y.H., Yin, R., 2005. Ecological function of Paddy Field Ring to urban and rural environment I . characteristics of soil P losses from paddy fields to water bodies with runoff. Acta. Pedologica Sinica. 42, 799–804 (In Chinese). Chen, M., 2007. Nutrient Balance Modelling and Policy Evaluation in China Farmingfeeding System [D]. Tsinghua University, Beijing. Cooperative research group on Chinese Soil Taxonomy,, 2003. Chinese Soil Taxonomy. Science Press, Beijing. Cordell, D., Drangerta, J.-O., White, S., 2009. The story of phosphorus: global food security and food for thought. Global Environ. Change 19 (2), 292–305. Crosland, A.R., Zhao, F.J., McGrath, S.P., Lane, P.W., 1995. Comparison of aqua regia with sodium carbonate fusion for the determination of total phosphorus in soils by inductively coupled plasma atomic emission spectroscopy. Commun. Soil Sci. Plant Anal. 26, 1357–1368. Defra, 2010. The Fertiliser Manual (RB209) Department for Environment, Food and Rural Affairs, TSO (The Stationary Office). (Retrieved, Nov 19 2014 from http:// www.defra.gov.uk/publications/2011/03/25/fertiliser-manual-rb209/). Djodjic, F., Borling, K., Bergstrom, L., 2004. Phosphorus leaching in relation to soil type and soil phosphorus content. J. Environ. Qual. 33, 678–684. Girard, R., 2013. This Class Implements Optimized List of Continuous Convex Piecewise Linear Functions. (> http://finzi.psych.upenn.edu/R/library/ConConPiWiFun/ html/cplfunctionvec.html. Accessed 1 May 2015.). Hanrahan, L.P., Jokela, W.E., Knapp, J.R., 2009. Dairy diet phosphorus and rainfall timing effects on runoff phosphorus from land-applied manure. J. Environ. Qual. 38, 212–217. Hesketh, N., Brookes, P.C., 2000. Development of an indicator for risk of phosphorus leaching. J. Environ. Qual. 29, 105–110. ISRIC, 2014. Soil and Terrain Database (SOTER) for China. ISRIC −World Soil Information and FAO. (http://www.isric.org/projects/soter-china. Accessed 26 Dec 2016). Kazunori, M., Tamon, F., Toshichika, I., Nittaya, C., Uday, P., Motoki, N., Yasushi, I., Tsuneo, K., 2016. Prediction of future methane emission from irrigated rice paddies in central Thailand under different water management practices. Sci. Total Environ. 566–567, 641–651. Kleinman, P., Sullivan, D., Wolf, A., Brandt, R., Dou, Z.X., Elliott, H., et al., 2007. Selection of a water-extractable phosphorus test for manures and biosolids as an indicator of runoff loss potential. J. Environ. Qual. 36, 1357–1367. Kleinman, P., Sharpley, A., Buda, A.R., McDowell, R.W., Allen, A.L., 2011. Soil controls of phosphorus in runoff: management barriers and opportunities. Can. J. Soil Sci. 91, 329–338. Lee, C.H., Lee, Y.B., Lee, H., Kim, P.J., 2007. Reducing phosphorus release from paddy soils by a fly ash-gypsum mixture. Bioresour. Technol. 98, 1980–1984. Li, H.G., Liu, J., Li, G.H., Shen, J.B., Bergstrom, L., Zhang, F., 2015. Past, present, and future use of phosphorus in Chinese agriculture and its influence on phosphorus losses. AMBIO 44 (2), 274–285. Li, L.G., He, Z.L., Li, Z.G., Li, S.L., Wan, Y.S., Stoffella, Peter J., 2017. Spatiotemporal change of phosphorous speciation and concentration in stormwater in the St. Lucie Estuary watershed, South Florida. Chemosphere 172, 488–495.

122

Agriculture, Ecosystems and Environment 245 (2017) 112–123

L. Hua et al.

Zhang, N., Li, C., Li, Y., 2007a. Accumulation and releasing risk of phosphorus in soils in Dianchi watershed. Soils 665–667. Zhang, Z.J., Zhang, J.Y., He, R., Wang, Z.D., Zhu, Y.M., 2007b. Phosphorus interception in floodwater of paddy field during the rice-growing season in TaiHu Lake Basin. Environ. Pollution. 145, 425–433. Zhang, F.S., Wang, J.Q., Zhang, W.F., Cui, Z.L., Ma, W.Q., Chen, X.P., et al., 2008. Nutrient use efficiencies of major cereal crops in China and measures of improvement. Acta Pedologica Sinica. 45, 915–924. Zhang, S.J., Wang, L., Ma, F., Zhang, X., Li, Z., Li, S.Y., et al., 2015. Can arbuscular mycorrhiza and fertilizer management reduce phosphorus runoff from paddy fields? J. Environ. Sci 33, 211–218. Zhao, M.Y., Cheng, C.T., Chau, K.W., Li, G., 2006. Multiple criteria data envelopment analysis for full ranking units associated to environment impact assessment. Int. J. Environ. Pollu. 288, 448–464. Zhu, D.F., Chen, H.Z., Xu, Y.C., Zhang, Y.P., 2013. Constraints and countermeasures of the mechanization of double rice production in China. China Rice. 19, 1–4 (in Chinese).

Wu, J., Fan, J., He, Y., Tu, R., Tan, B., Xu, H., et al., 2012. Dynamics of nitrogen and runoff loss in ponding water of paddy field under different fertilization practices. Ecol. Environ. Sci. 21, 1561–1566 (In Chinese). Xi, B., Zhai, L.M., Liu, J., Liu, S., Wang, H.Y., Luo, C.Y., et al., 2016. Long-term phosphorus accumulation and agronomic and environmental critical phosphorus levels in Haplic Luvisol soil, northern China. J. Integra. Agric. 15, 200–208. Xie, C., Zhou, X., Yao, L., 2001. Phosphorous situation in soil and demand prediction for P fertilizer of Guangdong Province. Soil Fertil. 13–15. Xu, A.L., Wang, P., 2008. Phosphorus losses with surface runoff from farm lands in polder area around Taihu basin. J. Agro-Environ. Sci. 27, 1106–1111 (In Chinese). Xu, X.J., Xu, J.P., Ning, Y.W., et al., 2013. Effects of nitrogen-phosphorus reduction and phosphorus application patterns on crop yields and soil nutrients in rice-wheat rotation system. Soil Fertilizer Sci. China 5, 75–79 (In Chinese). Zhai, L.M., CaiJi, Z.M., Liu, J., Wang, H.Y., Ren, T.Z., Gai, X.P., et al., 2015. Short-term effects of maize residue biochar on phosphorus availability in two soils with different phosphorus sorption capacities. Biol. Fertil. Soils. 51, 113–122.

123