Introduction

Several lakes and rivers in China suffer from nitrogen and phosphate pollution. Twenty-four of 61 key lakes and pools have water quality below Class III. For the top ten rivers, 10.2% have inferior Class V water quality, 20.9% have Class IV–V quality, and 68.9% have Class I–III quality. Only 72.3% of drinking water sources in key cities attain the national water standard, and nearly 0.3 billion people in rural regions drink unsafe water1. TN, TP and COD from agricultural sources are approximately 2.7, 0.28, and 13.24 million tons, respectively, which accounts for 57.2%, 67.4% and 3.7%, respectively, of total emissions2. More than half of the groundwater in 14 investigated counties on the North China Plain have nitrate concentrations that exceed 10 mg L−13. Nitrate concentrations in 7.4% of the groundwater in Beijing suburbs exceed the national standard4.

The relationship between organic matter (OM) and soil nitrogen has always been of concern. Some researchers believe that OM can reduce nitrate loss5,6, but over-application can cause greater loss7,8,9. The inverse relationship between dissolved organic carbon (DOC) and NO3 − concentrations in Japanese streams is closely related to excess Nitrogen availability together with a carbon deficit in a soil environment10. To adjust the release of nitrogen into the soil, manure has replaced some chemical fertilizers11. To meet the N requirements of corn, poultry litter, which releases N more slowly compared to inorganic fertilizer, has been used; the use of poultry litter not only reduces N fertilizer costs for corn but also reduces the risk of nitrate N leaching into the groundwater12. Annual nitrate leaching losses in conventional plots are 4.4–5.6 times higher than in organic plots and integrated plots between conventional and organic plots, which demonstrates that organic and integrated fertilization practices more actively and efficiently support denitrifier communities, shift the balance of N emissions and nitrate losses, and cause less environmental harm13. Although only approximately 10% of the applied 15N-Iabeled fertilizer remained in the 0–30 cm layer of the control and plant manure plots, more than 25% of the applied 15N remained in the pig compost plot14. Mamo determined that nitrate leaching could be reduced by substituting compost for commercial inorganic fertilizer in vegetable fields15. A routinely cited environmental advantage of compost as a soil amendment is that it can reduce mineralization rates and decrease the potential for nitrate leaching by delaying the conversion of organic N to mobile and nitrate N16. Brinton reported potentially lower nitrate leaching from soils amended with composted rather than uncomposted manure in a maize field17. To reduce nitrate losses, organic fertilizers should not exceed 175 N kg ha−2 in nitrate vulnerable zones, but the index was 276 N kg ha−2 at the Rothamsted Experimental Station18. Bird guano resulted in greater nitrate losses when the application amount exceeded 11.2 t ha−2  19.

In the Yellow River irrigation zone of Ningxia, the concentration of ammonium in the drainage channels is 20–30 mg l−1. Only 38.3% of the sections attain a Class V water standard, so at the outer-boundary section of the Yellow River, water quality is usually at Class V during the entire year. Nitrate concentrations exceed 10 mg l−1 for nearly half of the shallow groundwater20. The SOM is usually deficient; the values range from 9.2 to 14.5 g kg−1, and the average value is 10.2 g kg−1. The application rate of fertilizer N is as high as 301 kg ha−2, which is 1.6 times the national average. The total farmland nitrogen and nitrite levels are at 61–66% and 76–81%, respectively, in surface waters. Approximately 7 billion m3 y−1 of water is drawn from the Yellow River, and 2.5 billion m3 y−1 enters the Yellow River again by percolation and filtration from upper fields. The irrigation rate is 15–45 × 103 m3 ha−2 for rice, 3.3–4.5 × 103 m3 ha−2 for wheat/corn (intercropping), and 1.5 × 103 m3 ha−2 for winter irrigation and spring irrigation21.

Materials and Methods

Study area

The trial field (106°17′52″E, 38°07′26″N) is in Lingwu County in Ningxia, which is part of the upstream irrigation zone of the Yellow River (Fig. 1). The area is a temperate arid zone with low rainfall and a dry climate. The rainy season is from July to September and accounts for 70% of the total annual precipitation. Snows in the winter is rare. The average annual precipitation is only 193 mm yr−1 and the evaporation rate is 1,763 mm yr−1. There are approximately 150–164 frost free days, and the average annual temperature is 8.9 °C. The active accumulated temperature for temperatures ≥10 °C is 3,200–3,400 °C. The average elevation above sea level is 1,130 m. There are between 2,800 and 3,100 daylight hours in a year. Agriculture essentially relies on gravity flow systems. The soil type is classified as anthropogenic-alluvial with lower fertility (Table 1). The topsoil has a high salinity and a high pH. It is a single-harvest zone. The main crops are rice, corn and spring wheat, and the typical cropping system is rice-corn rotation.

Figure 1
figure 1

Location of the study area in the Yellow River irrigation area, North Ningxia. (Note: One of your co-authors created the figure with Photoshop, and had published the figure in another journal. From: Y Z Feng et al. Features and treatment of non-point source pollution in the Ningxia Yellow River area. African Journal of Agricultural Research. 6 (24), 5541–5550 (2011)).

Full size image
Table 1 Physical and chemical properties of the trial field.
Full size table

Experiment design

Each plot was insulated by brick walls with plastic film to prevent water interchange, 40 cm high above the ground and 80 cm deep beneath the soil. There were 3 treatments: traditional without manure (CK: 0 kg ha−2), swine manure application (T1: 4,500 kg ha−2) and swine manure application (T2: 9,000 kg ha−2). The plot area was 200 m2 with 3 replications. The field trials were conducted from 2009 to 2013. The crop was rice each year except for wheat in winter 2011. Swine manure was applied and plowed 0–25 cm in the topsoil before winter irrigation.

The rates of fertilizer application in the rice field were urea, 300 N kg ha−2, triple superphosphate, 105 P205 kg ha−2, and potassium chloride, 60 K20 kg ha−2. Half of the nitrogen, total phosphate and potash were applied as base fertilizers, and the other 50% of the nitrogen was applied as a top dressing at a ratio of 3:1:1 at seeding (end of May), tillering (end of June), and booting (end of July) stages. The time of first irrigation in the paddy field was approximately mid-May, and the last irrigation was early July each year. The irrigation rate was 15, 000 m3 ha−2 in the rice growing period. Rice transplanting was in mid-May, and harvesting was in late September (approximately 120 days).

The fertilizer application rates in the wheat field were urea, 225 N kg ha−2, triple superphosphate, 150 P205 kg ha−2, and potassium chloride, 90 k20 kg ha−2. Half of the nitrogen, total phosphate and potash were applied as base fertilizers. The remaining 50% of the nitrogen was applied as a top dressing at a ratio of 3:1:1 at seeding (early March), elongation (early May) and earing (early June) stages. The irrigation scheme included winter irrigation 1,350 m3 ha−2 (end of Oct.), greening, 900 m3 ha−2 (end of Mar.), elongation, 1050 m3 ha−2 (middle of May), and heading, 1,050 m3 ha−2 (early June). Sowing occurred on 4 Oct. 2010 and harvest occurred on 29 June of the following year (269 days later).

Methods

Soil nitrate losses were determined using a resin-core device (Fig. 2, national patent No. 201020282864.4, 2010). This method had been used to determine soil nitrogen mineralization22,23,24,25. The resin-core device was comprised of a stainless steel pipe (diameter 76 mm × thickness 0.82 mm), resin bag (60 mesh nylon net, length 8 cm × width 8 cm), 2 pieces of aluminum plastic plate (diameter 74 mm with 13 eyes (diameter 3 mm)), cover, handle and antiskid axis. A strongly alkaline anion exchange resin (trade mark 717#) was used, and the pretreatment corresponded to the Method for Pretreating Ion Exchange Resins (GB/T 5476-1996). There were 4 holes in the pipe body for reducing the pressure difference between the inside and the outside of the pipe. The end of the pipe was a 10-cm long wedged surface for easy insertion into the soil. A 2-cm long aluminum plastic plate was above the wedge. The antiskid axis may have been fixed to prevent the aluminum plastic plates from dropping. The cover and the handle were convenient to place and draw the pipe body in the soil. There were 5 types of resin-core devices that were 42, 72, and 102 cm long. Initially, the resin-core devices were inserted just below the soil surface. Second, the pipes were lifted out, and the soil in the wedge and 2 cm above the wedge was collected. Third, the aluminum plastic plates, resin bags (16 g of resin in a bag), aluminum plastic plates, and antiskid axis were attached in sequence. Finally, the devices were placed back into the soil, and the wedge was refilled with soil. After a period of time, the devices were retrieved, the old resin bags were removed, and new resin bags were installed for the next test stage. The old resin bags were refrigerated. On each plot, 3 types of devices were placed in line and separated by 2 m, and the 3 duplications were placed diagonally along the plot.

Figure 2
figure 2

Resin core device used to determine soil nitrate loss.

Full size image

The nitrate in the resin bags was extracted in Potassium chloride solution (1 mol/L KCl) and determined using ultraviolet spectrophotometry. The nitrate loss is calculated using the following equation:

$$nitrate\,loss\,(kg\,h{a}^{-2})=\frac{nitrate\,amount\,in\,resin\,(kg)}{cross\,sectional\,area\,of\,pipe\,({m}^{2})}\times 10,000\,({m}^{2})$$

The data were analyzed using SPSS 19 and Excel 2010, and the probability was determined to test the significance (α = 0.05) in different treatments using the one-way analysis of variance (ANOVA).

Results

Nitrate leaching loss

Table 2 shows nitrate leaching losses. Nitrate leaching losses in the 30-cm layer were 16.95–17.50 kg ha−2 in the rice field and 13.91 kg ha−2 in the winter wheat field; there were significant differences (p < 0.05) between the rice and winter wheat fields. However, the treatments and CK were not significantly different (p > 0.05) in the 30-cm layer. In the 60-cm layer, T2 and CK were significantly different (p < 0.05) in 2010 and 2013. T1 and CK were significantly different (p > 0.05) only in 2013. T1 and T2 were significantly different (p < 0.05) only in 2010. In the 90-cm layer, T2 and CK, and T1 and CK were significantly different in 2010, 2012 and 2013. T2 and T1 were significantly different only in 2012. In the 90-cm layer, T2 and CK, and T1 and CK were all significantly different (p < 0.05) in 2010, 2012 and 2013. T2 and T1 were significantly different (p < 0.05) only in 2012. There were significant differences (p < 0.05) between the rice and winter wheat fields. The 90-cm layer and 30-cm layer, and the 6- cm layer and 30-cm layer were significantly different, but there were no significant differences between T1 and T2. Figure 3 shows the nitrate leaching trends. Nitrate leaching losses declined with increasing soil depth in all years, and the losses in the winter wheat field were less than in the rice field. Nitrate leaching losses in the rice field were very similar over 3 years, but the losses in the last year were less than in the first and third years.

Table 2 Nitrate leaching losses at different soil layers and treatments in the trial period.
Full size table
Figure 3
figure 3

Nitrate leaching trends for 4 years.

Full size image

The process of soil nitrate leaching loss during the entire growth period

In Fig. 4, the ratios of nitrate leaching losses are shown for different periods. In the rice field, the key period of soil nitrate leaching was before the end of June (45–50 days/130–140 growing days) (Fig. 4a); nearly 70% of nitrate had leached at the 30-cm layer, and almost 68–75% of nitrate had leached under the 30-cm layer. For winter wheat, there was a nearly 70–85% (Fig. 4b) nitrate leaching loss from each soil layer before the end of June. The key periods of soil nitrate leaching losses were at the early stages of rice and winter wheat; for rice it was at end of June (tillering phase), and for winter wheat it was at the end of June (earing phase).

Figure 4
figure 4

Ratio of nitrate leaching loss during the entire growth period.

Full size image

Discussion

Soil nitrate concentration

As shown in Table 3, the soil nitrate concentration was low in the rice field in mid-April, increased at the end May and the end of June, and then gradually decreased at the end of July and August and early September. Soil nitrate concentrations were approximately the same in the middle of April and at the end of July (p > 0.05), at the end of May and the end of June, and at the end of August and early September. The difference between T2 and CK was significant except at the end of May and June, and the difference between T1 and CK was not significant in any month. The difference between T2 and T1 was significant, except at the end of May and June. Therefore, in the early growing period, the soil nitrate concentration was usually high and low, respectively, before rice transplanting and at the end of the growing period; furthermore, the soil nitrate concentration was approximately the same in the two stages. Therefore, the contribution of soil nitrate concentration from fertilizer N was greater than from swine manure. Of course, the contribution of nitrate from swine manure should not be ignored, especially in the irrigated field. Soil nitrate concentration in the winter wheat field at the end of May was greater than during other periods (p < 0.05), and it was lowest at the end of June. The difference between T2 and CK was significant in mid-April and at the end of May, and the difference between T1 and CK was significant only in mid-April. The difference between T2 and T1 was significant only at the end of May. Soil nitrate concentration in the winter wheat field was greater than in the rice field, possibly because there is no leaching in dry farming. No samples were obtained after harvest because weeds were difficult to eradicate in the winter wheat field.

Table 3 Soil nitrate concentration in the 30 cm layer (mg kg−1).
Full size table

Soil nitrate concentration dropped and the carbon to nitrogen ratio increased with increased organic matter; therefore, nitrate leaching can be controlled at a higher carbon to nitrogen ratio because soil microbes fix mineral nitrogen26. Too great a volume and long-term manure application possibly caused more nitrate to leach27. Soil nitrate accumulation increased as manure application increased, and manure over-application caused more nitrate to accumulate below a 200 cm depth28. Soil microbes convert nitrogen into nitrate and improve the probability of nitrate leaching under a high organic carbon to nitrogen ratio29. Soil nitrate concentration increased as soil depth increased, and the peak value appeared in the 100–120-cm layer in a peach orchard30; but Zhang found that the peak value of nitrate accumulation was in the 0–60-cm layer31. Fertilizer reduction, improved methods of fertilizing, and fertilizer efficiency can decrease soil nitrogen accumulation and leaching32. Manure over-application caused the nitrate concentration to increase sharply in the soil profile in a vegetable field33.

Nitrate of soil leachate

There were 4 time samples of soil leachate in the rice field each year. Table 4 shows the soil leachate nitrate concentrations.

Table 4 Nitrate concentration of soil leachate (rice field) (mg l−1).
Full size table

In the 30-cm layer, the difference between T1 and CK reached significance in early and late June, as did the difference between T2 and T1 (p < 0.05). T1 and CK were not significantly different (p > 0.05) in any month. Nitrate concentration was not significantly different (p > 0.05) at the end of May compared with early June, but there were clear differences (p < 0.05) between early June, the end of June and the end of July. The sampling intervals were short, which led to similar nitrate concentrations at the end of May and in early June, and at the end of June, the nitrate concentration of soil leachate was not low, which was coincident with the soil nitrate concentration. At the end of July, the nitrate concentration of the soil leachate was greatly reduced. In the 60-cm layer, the difference between T2, T1 and CK reached significance at the end of May; the nitrate concentration of T1 was the largest. Only T1 and CK were significantly different in early June; T2 and CK were not significantly different, and CK was greater than T2. The treatments and CK were not significantly different at the end of June and July. The nitrate concentration differences in the 60-cm layer were greater than in the 30-cm layer between the end of May and early June. The nitrate concentration was reduced at the end of June and greatly reduced at the end of July. In contrast with the CK treatments, the differences in nitrate concentration in the 90-cm layer almost reaching significance in all periods, except T1 at the end of July, Almost all treatments were greater than CK. The change of nitrate concentration in the 90-cm layer was the same as in the 30-cm and 60-cm layers. The chicken manure application caused high soil leachate nitrate and ammonium concentrations in the paddy field34.

Soil TN and SOM

In the 0–30-cm-layer, the TN values for CK, T1 and T2 were 0.72, 0.78 and 0.88 g kg−1, respectively, and were improved by 7.72% and 22.04% (P < 0.05), respectively. This result was the same as for the other tests. In the 30–60-cm layer, the TN values of CK, T1 and T2 were 0.42, 0.41 and 0.45 g kg−1, respectively and were improved by −1.26% and 6.01% (P > 0.05), respectively. In the 60–90-cm layer, the TN values of CK, T1 and T2 were 0.27, 0.32 and 0.31 g kg−1, respectively, and were improved by 21.39% and 16.46% (P < 0.05), respectively. The results show that soil nitrogen leached from the topsoil into the deep layer and finally into the groundwater and that nitrogen losses increased with nitrogen application. Contrasted with a CK of 12.67 g kg−1 in the 0–30-cm layer, the SOM of T1 and T2 were increased to 0.95 g kg−1 and 1.41 g kg−1, respectively, in 2013. Therefore, swine manure application can increase soil organic matter; most of the tests support the results. Manure application can increase organic carbon in the plow layer and plow pan (P < 0.05); TN, TP, TK and available nutrients were clearly increased. Soil porosity increased by 11%, and the water stability of soil aggregates was stronger than CK. However, the tensile strength of the soil was clearly reduced. MBC and MBN were 2.1 and 1.5 times greater than CK in the plow layer35.

Cropping patterns and soil nitrate leaching loss

Soil nitrogen leaching losses were affected by cropping patterns, which include the crops and the cropping system. Nitrate leaching losses in the paddy field were greater than in the winter wheat field; the primary reason was the presence and quantity of irrigation. Nitrate leaching losses increased with increased irrigation. There was greater nitrate leaching loss in the seedling stage because of the immature crop roots; soil nitrate could not be absorbed more rapidly and completely than required in the seeding stage. A new fertilizing technique was named “the technique of fertilizer nitrogen backward”, which means that the fertilizer rate is reduced in the early stage and increased in the middle stage so that the total dose is not changed36. The top 4 out of 12 cropping patterns, which were winter wheat-rice, spring wheat-Chinese cabbage, winter wheat-silage corn and winter wheat-oil sunflower, clearly reduced the residual soil nitrate, so repeated cropping and intercropping reduced nitrate leaching losses and protected shallow groundwater37. The traditional cropping system has a long fallow from early October to early March or mid-May (150 or 220 days), and the nitrate leaching losses were purely nitrogen (0.69–0.81 kg ha−2)38. Nitrate leaching losses at the vegetable farm were greater than at the crop farm because of the fertilization rate. In agroforestry systems, tree roots can absorb soil N in the deep soil and reduce the amount of nitrogen that leaches into the groundwater39. The nitrogen peak was in the spring formative period for forestry and was in the grain formative period for crops40. For cereal crops, application of 200 kg ha−2 of nitrogen in one season should be considered, and the application of organic and inorganic fertilizer for the wheat-corn rotation should not exceed 400 kg ha−2 (12,000 kg ha−2 yield) on the North China Plain41.

The shallow groundwater and soil nitrate leaching loss

Nitrate leaching losses in the 90-cm layer have a greater chance of migrating into water because the groundwater level is only 120–200 cm high during the planting season. The amount of leakage water was nearly 100–120 cm in the paddy field each year, and the amount of nitrate lost through leaching losses was 8.00–8.69 kg ha−2 y−1 of pure nitrogen. If the leakage water and nitrate lost to leaching enter the groundwater together, the nitrogen concentration in the leachate would be 0.67–0.87 mg l−1; the loss was almost equal to 2.96–3.85 mg l−1 for nitrate leachate. The nitrate concentration exceeded 10 mg l−1 in some places, and it was greater in the planting season. Zhao demonstrated that nitrate and organic nitrogen concentrations in groundwater were 11.85–26.12 and 0.64–5.89 mg l−1, respectively, under greenhouse vegetables and 0.12–4.97 and 0.03–1.00 mg l−1, respectively, under farmland in the Ningxia irrigation zone42. The dominant N leaching form was nitrate (accounting for 57.3–92.0% of TN), and the secondary form was dissolved organic nitrogen (7.8–42.5%); only approximately 1% of nitrite was leached42. Ju noted that greenhouse vegetable farming caused soil nitrate accumulation and leaching; the groundwater nitrate concentration was 9–274 mg l−1 and the soil nitrate concentration was 270–5038 kg ha−2 in the 0–90-cm soil layer in Huimin County, Shandong Province, where nitrate concentration in 99% of the groundwater was more than 10 mg l−1  43. Song analyzed soil nitrate leaching losses of greenhouse vegetables that reached 152–347 kg ha−2 in the 0–100-cm soil layer in Shouguang County, Shandong Province, which was closely related to groundwater pollution44. Nitrate in groundwater was demonstrated to come from farmland soil because of large applications of fertilizer nitrogen45.

Conclusions

Manure application at 6,000 to 9,000 kg ha−2 does not cause an obvious increase or decrease in soil nitrate leaching. Therefore, in the Ningxia irrigation zone, some type of manure application should be practiced as this study found that rice yield is increased from 12.26% to 11.55%, and the winter wheat yield is increased from 9.32% to 12.52%. In addition, the application of manure makes it possible to reduce the use of nitrogen fertilizer, crop yields do not drop, and soil nitrate leaching losses do not increase.