Introduction

Soil erosion, defined as the detachment and displacement of soil particles from the surface to another location (Govers et al. 1990; Flanagan 2002), is associated with about 85 % of land degradation in the world, causing up to 17 % reduction in crop productivity (Oldeman et al. 1990; Nyakatawa et al. 2006). In addition to land degradation, not only is soil lost by erosion, but a proportionally higher percentage of plant nutrients, organic matter, pesticides, and fine soil particles are lost than in the original soil (Rodney et al. 2013). Eroded soils carry nutrients, pesticides, and other harmful farm chemicals into rivers, streams, and ground water resources (Gallaher and Hawf 1997) which has become an issue of significant and severe societal and environmental concern (Elsen et al. 2003; Singha et al. 2006) as it is declining crop yields and causing nonpoint source pollution of surface and ground water resources by sediment, fertilizer nutrients, and pesticide residues.

Excess use of chemical nitrogen (N) and phosphorus (P) fertilizers in agricultural activities has been observed for several decades (Sharpley et al. 1993; Simard et al. 1995; Zeng et al. 2007a) to maintain crop productivity, but it can cause degradation of water quality (Smith and Cassel 1991; Conan et al. 2003; Ren et al. 2003; Udawatta et al. 2006) of the catchment. In China, the yearly consumption of chemical fertilizers has increased from 16.6 × 106 t in 1983 to 39.8 × 106 t in 1997, and to 46.4 × 106 t in 2004 (MAC 2005). Based on monitoring 130 km of 1300 main rivers of China in 2004, in terms of the national environment quality standard for surface water (GB3838-2002), 40.6 % of the river water is poorer than the class-III water quality and 22 % is poorer than class-V (Zeng et al. 2007b). In the Midwestern United States, the agricultural sector has been identified as the single largest contributor to nonpoint source nitrate (NO3-N) pollution of surface and ground waters (Hartfield et al. 1999; Jaynes et al. 1999; Burkart and Stoner 2001). With the continuously heavy fertilizer application, uptake of nutrients from the soil keeps decreasing threatening crop productivity. Bockman et al. (1990) has estimated that at a global level, 48 % of the N-P-K nutrients used by crops in 1970 were derived from the soil, 13 % from manure, and 39 % from inorganic fertilizers. By 1990, the percentages had changed to 30 % from soil, 10 % from manure, and 60 % from inorganic fertilizers. Ayoub (1999) projected that for year 2020, the percentage would be 21 % from the soil, 9 % from the organic source, and 70 % from the inorganic fertilizers. Experiments by Ma (1987) have shown that for a given season, 30–50 % of N fertilizers and 10–20 % of P fertilizers are taken up. Then, the nutrient buildup increases potential for surface and/or groundwater pollution and eutrophication (Smith et al. 1999). According to Quan and Yan (2002), in China, the paddy soils alone annually contribute 5.94 t N/km2 to the waters, accounting for 17.5 % of the total N loadings in the water bodies. The negative impact of fertilizers on the water quality is likely to grow with time on a national scale (Zeng et al. 2007a). Best management practices (BMPs) recommend the timing and amount of fertilizer to be applied to various crops that will just meet crop needs should minimize the pollution of our soil and water resources by reducing the nitrogen content of runoff and percolation. Minimizing N and P concentrations in agricultural runoff is necessary to protect receiving waters from eutrophication (Zeng et al. 2008). N and P are known to be the limiting nutrients in aquatic ecosystems (Zhang et al. 2003).

The Three Gorges Region (TGR), an important agricultural production area with an annual precipitation of about 1100 mm, covers an area of 62,640 km2, 80 % of which consists of low mountains, hills, and valleys. The annual soil loss is approximately 157 million tons equivalent to a loss of 25 t of soil per hectare, making it one of the regions with extremely serious soil erosion. Slopes in this region range between 7 and 25°, covering 52.1 % of the area, whereas 37.5 % of the area has slopes greater than 25° (Zhou 2001; Shi et al. 2009). Shi et al. (2009) has reported that 46.2 % of the soil loss comes from cultivated slope land. Those losses constitute a threat to the well-being of over 30 million people (living in the TGR). Moreover, it reduces the benefit from the Three Gorges Dam as the reservoir will be undermined due to siltation and eutrophication (Wang et al. 2010). Recent water quality tests in the TGR imply that soil P loss has become a major cause of eutrophication (Liu et al. 2004; Wu et al. 2010).

Nutrient runoff losses from agricultural systems have been studied extensively over the past few decades (Sims et al. 1998; Simard et al. 2000), with most agricultural runoff studies focusing on N and P losses (Zhang et al. 2003). As large losses of nutrients in runoff continue to occur from cropped watersheds, studies emphasis on N and P losses from purple soil intercropped by winter wheat-summer maize under different fertilizer applications rates are rarely reported in TGR. However, management of the sloping land is a key factor in determining ecological sustainability in this region.

In this study, a 5-year plot experiment with intercropping winter wheat-summer maize was conducted. It is aimed to investigate the impact of fertilizer applications, i.e., control set without fertilizer (CK), combined manure with chemical fertilizer (T1), single chemical fertilizer (T2), and chemical fertilizer with increasing fertilization (T3), on plot runoff, sediment yield, and N and P concentrations in runoff water. The expectations are that our results will contribute to the BMPs recommendations minimizing N and P concentrations in agricultural runoff from purple soil in TGR.

Materials and methods

Site description

The Three Gorges Region is located in the subtropical zone of China characterized by a temperate climate with mean annual temperature of 18 °C, mean summer temperature of 28 °C, and mean winter temperature of 6 °C. Rich soils support abundant agricultural and forestry resources, but the terrain shows marked changes in elevation. The study was conducted from 2010 to 2014 at the Southwest University experimental field (106° 24′ 20′′ E, 29° 48′ 42′′ N) on a 15 °C slope in Beibei district. The area is characterized by an average annual precipitation of 1100 mm, a mean annual temperature of 18.3 °C, an average annual sunshine of 1270 h and an average annual frost-free period of 317 days. The topography is dominated by mountains and hills. The main soil types are purple soil and the basics characteristics are showed in Tables 1 and 2.

Table 1 Basics physical characteristics of the purple surface soil in the study area
Full size table
Table 2 Basics properties on nutrient contents of top soil in the study area
Full size table

Experiment design and treatments

Twelve experimental plots were established on west facing hill-slopes, on which rain-fed sweet potatoes were grown before the plots were established, with intercropping winter wheat-summer maize. To reduce the effects of position, all plots were established on the same 15° slope and adjacent to each other. The size of each experimental plot was 32 m2 (8 × 4 m). In order to prevent water and nutrients exchange between adjacent plots, a concrete cement border, 30-cm deep by 30-cm width extended 20 cm above ground, was built. At the base, a tank with vertical scale was placed to collect runoff and sediments.

From 2010 to 2014, the following four fertilizer treatments (Table 3) were set up: (1) manure with fertilizer as applied by local farmers (T1), (2) optimum fertilizer as recommended by research (T2), (3) optimum fertilizer with increasing fertilization (T3), and (4) control set with no fertilizer (CK). Chemical fertilizer was made with urea, calcium superphosphate, and potassium chloride and rated according to each treatment as showed in Table 3. The proportion of N, P, and K was estimated at 46.4 % of N material in urea, 12 % of P2O5 within calcium superphosphate, and 60 % of K2O within potassium chloride respectively. Within farmyard manure, those proportions were estimated at 0.2 % of N, 0.2 % of P2O5, and 0.2 % of K2O. Each treatment had three duplicates. All treatments were oriented parallel to the slope. Field management was similar to that used by local farmers. The seeds of winter wheat were sown in November and harvested manually in May of following year, and those for summer maize were sown in April and harvested manually in August.

Table 3 Fertilizers application rates in the experiments
Full size table

Runoff and sediment collection and measurement

After each rainfall event, the runoff was collected at the base of the plot within a tank with vertical scale which records the runoff total volume. Samples from same treatment were carefully mixed throughly using plastic bottle to make a composite. The sink sediments at the bottom of the bottle were then oven dried at 105° for 24 h, weighted and regarded as sediments concentration. The amount of soil eroded was calculated by multiplying the amount of runoff with the sediment concentration. Part of composite sample was later centrifuged and filtered to analyze for NO3-N, NH4-N, total-N (TN), and total-P (TP) using routine analytical methods (Bao 2000). The TN was determined using the Kjeldahl method, and the 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 amount and the intensity of rainfall were recorded in an automatic meteorological station.

Data analysis

Experimental data were analyzed using the SPSS 20.0 system. Spearman’s rank correlation coefficients were calculated and regression analysis was performed to show the relationship between the amount of runoff and soil loss. One-way analysis of variance (ANOVA) was carried out to determine the differences of runoff, soil loss, and concentration of runoff nutrient elements among different treatments. Least significant difference (LSD) at p = 0.05 was used to elucidate any significant differences.

Results and discussion

Rainfall events and runoff volumes

From 2010 to 2014, 49 rainfall events have been recorded during the experimental period. They ranged from 16.1 to 85.1 mm in 2010, 18.9 to 124.2 mm in 2011, 19.8 to 101.2 mm in 2012, 33 to 117.5 mm in 2013, and 20.5 to 84.58 mm in 2014. Rainfall events with values over 60 mm were considered as extremes events as showed by the different peaks in the yearly rainfall curve (Fig. 1). The rainfall events have produced different amount of surface runoff. The variation in the amount of rainfall may mainly explain the difference in surface runoff as runoff events are usually the result of unique weather and soil conditions (Douglas et al. 1998). In 2010, two extreme rainfall events were recorded, 74.5 mm on June 16 and 85.1 mm on August 25, which generated runoff that accounts for 31.78 % of the 285.80 L m−2 total runoff. In 2011, two extreme events were also recorded, 115 mm on June 20 and 124 mm on June 29, with runoff that accounts for 34.88 % of the 135.79 L m−2 total runoff. Two records in 2012 were extreme: 60.7 mm on April 21 and 101.2 mm on July 9, accounting for 34.55 % of the 166.02 L m−2 total runoff. In 2013, a single extreme event accounted for 42.29 % of the 121.50 L m−2 total runoff, being 117.5 mm on June 17. Four events with over 60 mm were recorded in 2014: 64.2, 75.2, 84.58, and 75.28 mm with 33.79, 24.45, 23.93, and 27.03 L m−2 respectively, amount of total runoff. These runoffs represent 58.72 % of the 185.96 L m−2 total runoff. From those results, it can be noticed that some higher isolated rainfall events generated the amount of runoff that accounted for more than one third of total runoff or for several relatively lower rainfall events. In the literature, numbers of scientists have concluded that the runoff was mainly controlled by rainfall intensity as can be seen in Fig. 1. Then, runoff rate increased in response to the increasing rainfall intensity. Raindrops in high intensity of rainfall impact topsoil with more kinetic energy inducing some extent of soil sealing with consequences to prevent infiltration and increase the amount of runoff. According to Huang et al. (2013), higher rainfall intensities yield raindrops with more kinetic energy with consequences to increase the amount of runoff and disfavor the soil infiltration. Zhou et al. (2009) found that runoff and sediment export increased with increasing rainfall intensity in study conducted in Chongqing. The large rainfall events cause a brief initial high rate of discharge that is an important fraction of the total discharge (Black 1996). The linear regression (Fig. 1) confirmed the significant relationships between rainfall and runoff of the four treatments (R 2 = 0.402–0.692, p < 0.05, n = 49) (Fig. 2). Rainfall events account for 40 to 70 % of runoff variability which is established by an important relatively, lower rainfall events producing low runoffs.

Fig 1
figure 1

Relationship between rainfall and runoff with the control (CK), manure (T1), fertilizer (T2), and increase fertilizer (T3) treatments

Full size image
Fig 2
figure 2

Characteristic of rainfall events associated with control (CK), manure (T1), fertilizer (T2), and increase fertilizer (T3) treatments runoff during the experimental period

Full size image

Runoff and sediment yield

Table 4 showed the runoff volumes and sediment yield during the experimental period. The runoff volume ranged from 9.57 to 100.36 L m−2, and the annual mean volume of each treatment significantly varied in CK > T3 > T2 > T1 order. Fertilization rates had significant impact on slope runoff. The CK treatment with no fertilizer application produced the highest runoff with an annual mean value of 1.55, 1.42, and 1.34 times higher than T1, T2, and T3 mean values respectively. Compared to CK, the T1, T2, and T3 reduced the runoff volume by 35.7, 29.6, and 16.8 %, respectively. More rainfall water was retained by treatment where combined manure with chemical fertilizer was applied. The T1 presented a comparative advantage in reducing rain splash, increasing infiltration, and reducing runoff volume. Likely, the soil conditions in T1 and T2 were favorable for vegetation growth, good coverage which reduced raindrop velocity when arriving to the soil surface and increased the rain water infiltration thus decreased runoff and sediment detachment. The degree of vegetation cover and litter production from each treatment are the main factors controlling surface runoff and soil loss. Vegetation canopy was confirmed to play a key role in protecting soil surfaces from erosion (Wei et al. 2007). It was recognized that litter production and organic matter accumulation could reduce soil-water loss (Eagleson 1982; Boer and Puigdefábregas 2005). The results of Li et al. (2007) showed that continuous vegetation cover reduced rain splash and runoff and finally caused less damage to soil aggregates and improved soil structure permanently. Year to year comparison within a single treatment also showed difference in runoff volume. For the CK treatment, it varies from 23.93 L m−2 in 2013 to 100.36 L m−2 in 2010. This is mainly due to the variance in number of recorded rainfall events that have produced runoffs. Also, the period of sampling as some months seemed to have recorded a comparatively higher rainfall values.

Table 4 Total mean runoff and sediment in different treatments
Full size table

Results on sediment yield have been collected for 3 years (2012 to 2014). They ranged from 4.06 to 13.20 t km−2 and significantly varies from treatments in CK > T3 > T1 > T2 order. The CK treatment precipitated the higher sediments yield representing 1.67, 1.98, and 1.26 times of T1, T2, and T3 mean sediments yield respectively. Compared to CK, the T1, T2, and T3 reduced the sediments detachment by 40.5, 20.9, and 49.6 %, respectively. Less detachment of soil particles occurred in T2 treatment where optimum chemical fertilizer was applied. The T2 may have set a good covering and litter production. In T3, the increasing fertilizer application with time may have affected the soil structure strength, decreasing the N-P-K nutrients uptake by plants and influenced the crop development and the soil resistance to the detachment forces.

The linear regression for the four treatments (Fig. 3) showed a significant relationship between runoff and sediment yield (R 2 = 0.572–0.832, p < 0.05, n = 49). The runoff explained for 57 to 83 % of the variation in sediment which is strongly correlated. The runoff rate and sediment concentration were also influenced by the soil surface features. The variation in R 2 coefficient indicated that plots with different ground covers may have received different rainfall which generated different runoff and sediment detachment.

Fig 3
figure 3

Relationship between runoff and sediment yield with the control (CK), manure (T1), fertilizer (T2), and increase fertilizer (T3) treatments

Full size image

Nitrogen and phosphorus losses

The total nitrogen (TN) losses via the runoff during the experimental period ranges from 1.33 to 7.51 kg hm−2, the ammoniac nitrogen (NH4 +-N) from 0.07 to 1.24 kg hm−2, the nitrate nitrogen (NO3 −-N) from 0.10 to 1.45 kg hm−2, and the total phosphorus (TP) from 0.07 to 1.24 kg hm−2 (Table 5).

Table 5 Nitrogen and phosphorus losses in runoff in different treatments. Values are presented as mean ± SD
Full size table

The losses were significantly (p < 0.05) higher in the CK treatment. The T1, T2, and T3 treatments significantly decreased the total nitrogen loss compared to the CK treatment (p < 0.05); the reduction was 41.2, 40.4, and 26.9 % respectively. However, no significant difference in total nitrogen loss was observed between the T1 and T2 treatments. The NH4+-N loss varies in the order CK > T3 = T2 > T1. Compared with control, the T1, T2, and T3 treatments decreased the NH4 +-N loss by 35.2, 11.8, and 11.8 %, respectively. The NO3 −-N loss varies in the order CK > T3 > T2 > T1. The T1, T2, and T3 compared with controls, respectively, reduced NO3 −-N loss by 42.1, 34.1, and 31.8 %. The TP loss under fertilizations varies in CK > T3 > T2 > T1 order. The T1, T2, and T3 treatments compared to CK respectively reduced TP loss by 33.33, 31.25, and 25 %. The T1 combined manure with chemical fertilizer and the T2 optimum fertilizer applications retained much TN, NH4 +-N, NO3 −-N, and TP nutrient on the plots. The nutrients accessibility and their uptakes by the plants have been much easier in T1 and T2 than in other treatments. Douglas et al. (1998) noted that only a small percentage of the N and P applied as fertilizer would be lost in runoff when the fertilizer was applied at recommended rates and at optimum times. Then agricultural management practices have a major influence on soil nutrient losses (Baumhardt and Jones 2002; Abrisqueta et al. 2007; Wang et al. 2010). Variance analysis showed that fertilization had significant effect on nutrients losses. Comparing T1 to T2 treatment, in average, the TN, NH4 +-N, and NO3 −-N nutrient losses in T1 respectively reduced by 1.3, 26.7, and 5.9 % than T2. The combined manure with chemical fertilizer modify and improve the structure of the soil (Li et al. 2003), promote the formation of soil aggregate which increase the adsorption capacity of soil (Liu et al. 2004), and enhance the infiltration and thus reduce the N loss. Meanwhile, the mineralization of organic N in the manure treatment consumes protons, and that is why reduction in NH4+-N is much higher. Comparing T2 to T3 treatment, the TN and NO3 −-N loss in T3 treatment increased by 22.6 and 3.9 % than T2. The increase of fertilizer rate did not reduce the loss of nitrogen. Increasing fertilizer may produce higher chemical concentration in soil sediment fraction. It has been shown that the lower amount of chemicals generally lower the contents of nitrogen and phosphorus in the soils (Bu et al. 2008). When considering treatments from year to others, the variability on nitrogen and phosphorus is visible. This may be due to the difference in runoff, and sediments yield strongly correlate with rainfall events which yearly vary in number of events and total volume. Likely, year with poor rainfall characteristics (number of events, intensity, and depth) produced lower nitrogen and phosphorus losses.

Conclusion

We carried out a 5-year (2010 to 2014) study to characterize the soil erosion and nitrogen and phosphorus losses caused by rainfall under different fertilizer application levels. The results indicated significant positive effects of fertilizer treatments on surface runoff and nutrients losses compared to the plot set without fertilizer. The combined manure with chemical fertilizer treatment showed highest reduction of runoff and sediments losses. The rainfall-runoff and the runoff-sediment relationships set a significant correlation described by power function. The T1 combined manure with chemical fertilizer and the T2 optimum fertilizer applications have retained much more TN, NH4 +-N, NO3 −-N, and TP nutrient on the plots. The treatment T1 has improve the soil structure with positive effect on soil adsorption capacity, better option in reducing loss of soil total nitrogen and total phosphorus. From those results, it could be suggested for farmers to adopt combined manure with chemical fertilizer technologies to achieve the optimum soil water and nutrient conservation on purple slope cropland. Study on the effect of rating the manure combined with chemical on nutrient losses remain open and continue to remain our concern.