Effect of Nitrogen and Irrigation Application on Water Movement and Nitrogen Transport for a Wheat Crop under Drip Irrigation in the North China Plain

For improving water scarcity and groundwater pollution from agriculture, two-year experiments (2011–2013) with three water levels (0.3, 0.5 and 0.8 evaporation (E) in 20-cm-diameter pans) and four nitrogen (N) levels (120, 140 and 190 kg·ha−1 in 2012 and 120, 190 and 290 kg·ha−1 in 2013) were conducted to study effects of water and N availability on water movement and N transport for a wheat crop under drip irrigation in the North China Plain. The results indicated that under drip irrigation, deep percolation at 1-m depth was stable at 0.5–0.8 E with the same N rate for winter wheat. At 0.5–0.8 E, deep percolation was also relatively stable with increasing N rate from 120 to 140 kg·ha−1 or from 190 to 290 kg·ha−1. The irrigation schedule and N rates only affected N leaching below the root zone of winter wheat (60-cm depth), while the N residual in the soil layer presented more risk to the environment than N leaching. In general, the 290-kg-ha−1 N level was not recommended using drip fertigation for winter wheat in the North China Plain. The empirical equation given by the Ministry of Geology and Mineral Resources was also not recommended for estimating the drainage under drip irrigation. OPEN ACCESS


Introduction
Water scarcity is the major constraint on crop production worldwide, especially in the North China Plain (NCP) [1][2][3][4].As one of the main grain-producing areas in China, the NCP accounts for 30% of the national food production [5].Winter wheat (Triticum aestivum L.) is a major crop in the region, primarily limited by seasonal rainfall [6].Because the majority of precipitation comes after the wheat harvest, the annual precipitation cannot satisfy actual crop evapotranspiration (ET) of 800-900 mm during the wheat growing season [4,7].Supplementary irrigation is thus necessary for wheat production in the NCP.Because of the lack of surface water, groundwater in the region has been seriously exploited for irrigation [8,9].Although the excessive exploitation of groundwater resources leads to groundwater depression, flood irrigation is still the major irrigation method [10].At present, the limited water resources have been unable to meet the needs of supplementary irrigation water fully in this region.Meanwhile, the farmers still believe that increasing nitrogen (N) fertilizer is the only way of obtaining high yields of wheat in the NCP.Indeed, the grain yield of wheat increased with increasing rates of fertilizer application, which were reported from a series of field experiments [11,12].However, it is also found that the excessive application of nitrogen (N) cannot increase crop yields further, but considerably increases N leaching [13][14][15].Overuse of fertilizer not only increased the high cost of production, but also led to degradation of water and soil quality, especially the groundwater [16][17][18].A survey of groundwater nitrate-N (NO3 − -N) concentrations found that about 45% of 600 groundwater samples of the NCP exceeded the World Health Organization limit for NO3 − -N in drinking water of 50 mg•L −1 , with the highest NO3 − -N concentration reaching 113 mg•L −1 [4,19].This illustrated that the deterioration of groundwater quality is now becoming serious due to the heavy use of nitrogen fertilizer in the NCP.
Fortunately, the problems of water scarcity and deterioration of groundwater quality could be moderated by the adoption of water saving strategies and optimization of N and water management practices.However, few studies have been conducted to assess the applicability of water saving strategies (i.e., drip irrigation) for winter wheat in the NCP.Wang et al. (2013) reported that drip irrigation was recommended for winter wheat in the NCP due to improved yield and water use efficiency [20], but no further studies were conducted to access the effects of drip irrigation on deep percolation losses and NO3 − leaching for winter wheat in the NCP.
As known from a series of field experiments in other regions, drip irrigation could reduce deep percolation losses of water and minimize NO3 − leaching [21,22].However, zero leaching loss of N is not likely to be possible in agricultural production [23].Drip irrigation also cannot completely eliminate the deep percolation and leaching of N beyond the root zone [24,25].Proper irrigation schedules and reducing the N rates could keep the leaching losses to a minimum [26].For example, high irrigation frequency could reduce NO3 − leaching and N rates had a significant effect on NO3 − leaching under drip irrigation [5,25].However, these experiments were conducted for sugarbeets, tomatoes and maize.There was still no experimental study about the effects of combining different irrigation and N fertilizer management strategies on deep percolation and NO3 − leaching for winter wheat under drip irrigation in the NCP.
With these background considerations, a two-year field experiment was conducted in the NCP.It was assumed that N volatilization was zero, and we only considered plant N uptake, N residual and N leaching during the experimental periods.We focused on assessing the effect of different N and irrigation levels on water movement and N transport for the wheat crop under drip irrigation in the NCP.Two objectives were established in our study: (1) to examine the effect of different levels of N and irrigation on depth and time distribution of soil moisture and N under drip irrigation during the winter wheat seasons of the NCP and (2) to assess the effect of different levels of N and irrigation on deep percolation and NO3 − leaching under drip irrigation in the NCP during the winter wheat growing season.

Experimental Site and Field Management
A replicated field experiment was conducted over two wheat growing seasons (2011-2012 and 2012-2013) at the Irrigation Experiment Station of the China Institute of Water Resources and Hydropower Research (39°37′ N, 116°25′ E, 30 m above sea level), which is located in the Beijing region of the NCP (Figure 1).Summer-dominant rainfall happens in this region, contributing more than 70% of yearly rainfall between July and September.The long term mean annual precipitation was 644 mm from 1951-2013, and the mean temperature was 12.2 °C.Near normal precipitation of 137 mm fell during the winter wheat-growing season (April-June) in 2012, and 98 mm fell in 2013, which was 26% less than the average of 133 mm over the last 60 years (Table 1).The precipitation in both years was significantly less than the average of 35 mm (1951-2013) in May (from stem-elongation to Development of fruit), when the crop was more sensitive to water stress [6,27].The precipitation was also less than the average of 22 mm in April in 2013.The mean temperature over the growing season was 20.2 °C in 2012, which was 0.7 °C higher than the 19.5 °C long-term record , whereas the mean temperature of 18.2 °C in 2013 was 1.3 °C lower than the long-term record.The temperature was also lower in April and June of 2013 than the average maximum/minimum air temperatures of the last 60 years for the corresponding period.Mean daily 20-cm-diameter pan evaporation was 3.07 mm in 2012 (20 March-13 June) and 2.98 mm in 2013 (7 April-17 June).Full details of the experimental site and soil were provided in Wang et al. [20], and only details pertinent to this study are presented here.The soil at the experimental site was a silt loam, with organic matter content of 3.5 g•kg −1 , field capacity of 0.31 m 3 •m −3 , bulk density of 1.58 g•m −3 and permanent wilting point of 0.11 m 3 •m −3 (1 m depth).
Field management: Winter wheat (Triticum aestivum L. cv.Zhongmai 175) was sown on 3 October 2011 and 9 October 2012 using a sowing machine (Table 2).The row spacing was about 12.5-15 cm with seedling density of 450-800 plants•m −2 .Irrigation was applied (75 mm) before sowing.At planting, N (as compound fertilizer, N:P2O5:K2O = 15%:15%:15%) was added at a rate equivalent to 83 kg•ha −1 to ensure germination and establishment.Wheat in the plots was harvested on 13 June 2012 and 19 June 2013.The field experiment began after the overwintering stage of winter wheat and the growing season of winter wheat was divided according to BBCH scale (Table 2)

Experimental Design
Irrigations were scheduled according to the cumulative 20-cm-diameter pan evaporation (E) that had occurred since the previous irrigation.Three levels of water, viz., 0.8, 0.5 and 0.3 E (W1, W2 and W3, respectively), were adopted and the cumulative irrigation amounts were 30, 40 and 30 mm before Stem-elongation, from Stem-elongation to Senescence, and after Senescence, respectively, in 2012 and 2013.When the cumulative pan evaporation reached the specified value for irrigation, supplemental irrigation started.The emitters (Netafim Ltd., Tel Aviv, Israel) were spaced 30 cm apart, with a 1.2 L•h −1 flow rate at 1 bar operating pressure.
Field experiments were designed according to a completely randomized block design.Each treatment was replicated three times.The experimental area was 66 × 32 m with a fixed drip-irrigation system and divided into 27 equal plots of 20 × 3 m.

Soil Sampling
To assess the soil water and N status, samples were taken from the 0-20, 20-40, 40-60, 60-80, 80-100 and 100-120 cm layers of the soil profile using a hand-held auger.For each plot, three samples were taken immediately prior to sowing, 1 day before an irrigation, 1 day following an irrigation and at harvest.Additional measurements were taken 1 day after a more than 20-mm-rainfall event in 2013.For each sample, the soil sample was divided into two sub-samples.One sub-sample was used to determine the NO3 − -N concentrations and the other sub-sample was used to determine the water content gravimetrically [29].Twenty grams of air-dried soil were extracted with 50 mL of 1 mol•L −1 KCl, and the NO3 − -N concentrations were determined by an Auto Analyzer (colorimetric method; Bran + Luebbe, Hamburg, Germany) [5,30].

Deep Percolation
Method 1 (M1): During the experimental season, the deep percolation loss (DP) was estimated from the unsaturated hydraulic conductivity as a function of the prevailing moisture content and time period.The equation was thus described as [31]: where q is the mean volumetric flux density (mm•day −1 ), ∆t is the time period (days), and K(Se) is the unsaturated hydraulic conductivity at soil depth of 1 m, which was estimated by [32]: where Ks is the saturated hydraulic conductivity; m is the van Genuchten parameter, which is obtained from the soil moisture characteristic curve, and optimized using software RETC [32]; and Se is the relative saturation, which is calculated by: where θ is the soil water content (cm 3 •cm −3 ), θs is the saturated volumetric water content (cm 3 •cm −3 ) and θr is the residual volumetric water content (cm 3 •cm −3 ).Method 2 (M2): According to the soil water balance of the whole experimental period, DP at 1-m depth was determined by [33]: where P is the rainfall (mm), I is the irrigation (mm), ∆W is the change in soil water storage in the soil profile (mm), and R is surface runoff (mm), assumed to be zero as the plots had small preventive bands.ET is the evapotranspiration (mm), which is calculated as: where ET0 is the reference evapotranspiration, which is determined by the FAO Penman-Monteith method [33].Kc is the crop coefficient, which was obtained for wheat from Allen et al. (1998) [33].Method 3 (M3): The empirical equation was given by the Ministry of Geology and Mineral Resources and DP is estimated using a recharge coefficient (α) multiplied by the amount of irrigation (I, mm) for the region as follows [20]: where the coefficient (α) depends on soil texture and on the amount of total irrigation.The coefficient ranges from 0.1 for clay soil to 0.3 for sandy soil.Values of α were taken to be 0.1 for the amount of total irrigation ≤90 mm, 0.15 for irrigation between 90 and 250 mm, and 0.2 for irrigation ≥250 mm for the soils [6].

Nitrate N Loss Assessment
The ceramic suction cups were located at 60-cm and 1-m depth for each treatment to obtain the samples of soil solution [34].Suction cups were collected weekly with additional extractions taken at 12 h after an irrigation or a more than 20-mm-rainfall event.When the extraction began, a vacuum pressure of 30 kPa was applied to each suction cup.Then the solution was extracted for 36 h and the soil solution samples extracted were sent to the laboratory to be frozen in a 50-mL polypropylene bottle.After the harvest, the soil solution samples were analyzed for NO3 − -N concentration using an Auto Analyzer (Bran + Luebbe, Germany) [5].The NO3 − -N leaching losses (LN) were estimated from NO3 − -N concentration in the suction cups and the amount of deep percolation loss as [35]: where CN is the NO3 − -N concentration in the deep percolating water.The calculated results were converted to kg•ha −1 , and then summed over the experimental period.

Statistical Analyses
Statistical analysis was conducted at a significance level of 0.05.The two factors considered were irrigation level and fertilization.The analyses of variance with Duncan's post-hoc test were conducted to test the effect of different fertilization and irrigation treatments on experimental results.Statistical analysis was conducted using SPSS statistical software.

Depth and Time Changes of Relative Soil Moisture
As one of the most important soil moisture parameters, the relative soil moisture is described as θ/θFC (θFC = moisture at field capacity).The temporal variations in the relative soil moisture as affected by different levels of N and irrigation during winter wheat seasons of 2012 and 2013 are presented in Figures 2 and 3.The relative soil moisture variations throughout the experimental seasons were significantly influenced by the depth and irrigation levels, but not by the N levels.Because of the preceding winter irrigation, the relative soil moisture was high in the period 180-192 DAS in 2012 and 183-196 DAS in 2013.Because the ET increased with plant growth, the relative soil moisture decreased gradually after the stem-elongation stages, especially among the treatments with W2 and W3 irrigation levels.Under all treatments, the relative soil moisture variations were larger in the shallow layers (0-60 cm), and diminished as the depth increased, i.e., soil moisture variation was more prominent in the top 0-20-cm layer and less prominent in the 20-40-cm layer.Compared to the W3 treatments, W1 and W2 treatments distributed more water in a greater depth of soil.In particular, minimal effects of irrigation on the soil moisture of the 40-60-cm layer were observed at the W3 irrigation level in 2012 and 2013.A decline in the relative soil moisture of the 60-100-cm layer was observed during the experimental seasons for all treatments.
No significant differences in soil moisture were observed under W1F1, W1F2 and W1F3 treatments in 2012.Similarly, the N application rates had no significant impact on soil moisture at the W2 and W3 irrigation level treatments in 2012.The similar trend towards no significant effect of N levels on the soil moisture was also observed in 2013.

Deep Percolation
Cumulative deep percolation of water (DP) at 1-m depth was estimated from the soil hydraulic conductivity (M1), soil water balance (M2) and empirical equation (M3).The effect of different levels of N and irrigation on cumulative deep percolation during two wheat growing seasons is shown in Tables 4 and 5.The DP was significantly influenced by irrigation and N levels in loam soil during both experimental seasons.With the same N level, the DP rates estimated by M1 were significantly higher at W1 and W2 irrigation levels than that at W3 irrigation level in both years.However, there was no significant difference in DP between W1 and W2 irrigation levels.This is because of the low volume of water applied during the W3 irrigation schedule, whereby more soil moisture in the 80-100-cm layer was extracted by plant roots in the W3 treatments.Under the same irrigation level, the DP decreased with increasing N levels in 2012 and 2013.In particular, the effect of different N levels on the DP was significant among W3 irrigation schedule treatments in both years.Thus, the minimum DP rates of 119 and 52 mm were found at the W3F3 treatment in 2012 and the W3F4 treatment in 2013, respectively.Under W1 and W2 irrigation levels, the DP for F1 treatments was significantly higher than that for F3 treatments, whereas it had no significant difference with that for F2 treatments in 2012.Similarly, under W1 and W2 irrigation schedules, the DP rate for F1 treatments was significantly higher than that for F3 and F4 treatments, while no significant difference in DP was found between F3 and F4 treatments in 2013.losses estimated by the empirical equation (M3) were the lowest for all treatments in 2012.However, the DP estimated by M3 seemed to be higher in 2013 than the amounts in 2012, even though it was drier in 2013.There was higher deviation between the results calculated by M3 and the real DP.

Depth and Time Changes of NO3 − -N Concentrations
The temporal variations in NO3 − -N concentrations under different irrigation and fertilizer levels during the winter wheat seasons of 2012 and 2013 are presented in Figures 4 and 5.The figures show that the NO3 − -N concentration variations in soil layers during the experimental periods were significantly influenced by the depth, irrigation and N levels.Following the application of N as dissolved urea in water, the NO3 − -N concentration in the upper 0-40-cm increased steeply for all treatments on 192 DAS in 2012 and 185 DAS in 2013.The NO3 − -N concentration variation was more prominent in the top 0-20-cm layer and less prominent in the 20-40-cm layer in 2012 and 2013.In the 40-60-cm layer, the NO3 − -N concentration varied slightly during the growth periods of wheat in 2012 and 2013.The NO3 − -N concentration variation in the upper 0-60-cm layer was significantly influenced by irrigation levels in loam soil during two experimental seasons.The peak NO3 − -N concentration in the 0-20-cm layer, present on 192 DAS in 2012 and 185 DAS in 2013, increased significantly with decreasing irrigation level.With the same N level, the NO3 − -N concentration in the 20-60-cm layer was significantly higher at the W1 irrigation level than the concentration at W2 and W3 during the 192-220 DAS in 2012 and 185-220 DAS in 2013.In the late growth period of 2013, the NO3 − -N concentration in the 0-20-cm layer was significantly higher at W2 and W3 than the concentration at W1.There was no significant impact of irrigation levels on the NO3 − -N concentration in the 60-100-cm layer.The NO3 − -N concentration in the 0-40-cm layer increased significantly with the increase in N level during the two experimental periods.At F4 level, the NO3 − -N concentration in the 40-60-cm layer was obviously higher than the concentration at other levels in the late growth periods.Moreover, with the same irrigation level, the NO3 − -N concentration variation in the 20-40-cm layer at F4 level was more significant than the changes at other N levels in 2013.There was no significant impact of N levels on the NO3 − -N concentration in the 60-100-cm layer during two experimental seasons.

NO3 − Leaching and NO3 − Residual
The seasonal N leaching was determined by deep percolation and N concentration in the soil solution.The NO3 − leaching below the root zone (at 60 cm) and in the 100-cm soil layer were presented in Tables 6 and 7 during the two wheat seasons in 2012 and 2013.The results showed that the NO3 − leaching below the root zone was significantly influenced by irrigation and N levels in 2012 and 2013.The uniform precipitation weakened the effects of irrigation levels on NO3 − leaching in 2012, but the NO3 − leaching at the 60-cm layer was still higher at W1 and W2 irrigation levels with the same N level than the amount at W3.A similar trend was also observed in the dryer year of 2013.As expected, the NO3 − leaching at 60 cm significantly increased with increasing N application level in 2012 and 2013.However, both irrigation and N levels had no significant impact on NO3 − leaching at 100 cm in both years, even though the deep percolation at the 100-cm layer was significantly affected by different irrigation and N levels.The NO3 − leaching at 100 cm seemed to be more affected by the initial N concentration in the soil in 2012 and 2013.Comparing to N leaching, more N was maintained in the 0-100-cm soil layer as N residual during the wheat crop seasons.Table 6.Nitrate leaching (NL), initial nitrate concentration (Nini) and nitrogen residual (NR) for all treatments during the wheat season of 2012.Notes: W 1 , W 2 and W 3 : irrigations scheduled according to 0.8, 0.5 and 0.3 cumulative evaporation of 20-cm-diameter pan (E), respectively; F 1 : N application at 120 kg•ha −1 ; F 2 : N application at 140 kg•ha −1 ; F 3 : N application at 190 kg•ha −1 ; NS: not significant at the 0.05 level; *: significant at 0.05 level of significance.
The N residual at 0-100-cm depth was significantly influenced by irrigation and N levels for the winter wheat crop in 2012 and 2013 (Tables 6 and 7).The N residual increased with the increase in N level, whereas it decreased with the increase in irrigation water level during two experimental seasons.No significant difference in N residual was found at W1F1, W1F2, W2F1 and W2F2 treatments, but significantly higher residuals were found at W3F1 and W3F2 in 2012.With the same N level, higher N residuals were observed at the W3 level in 2013.With the same irrigation level, higher N residuals were also found at the F3 and F4 treatments in 2013.In particular, the highest N residual in F4 treatments, observed in the 0-100 cm depth in 2013, may generate more nitrogen leaching during intensive rainfall events or heavy irrigation events after wheat harvest.The N residual in the 0-100-cm soil layer, averaged over all treatments, was 19 kg•ha −1 in 2012, which was clearly lower than the accepted residual N values of 50 kg•ha −1 in the 0-90-cm layer for many European countries [36].However, due to the longer dry spells in 2013, the average N residual reached 175 kg•ha −1 , while the minimum N residual of 99.2 kg•ha −1 at the W1F1 treatment was also substantially higher than the accepted value of 50 kg•ha −1 .Table 7. Nitrate leaching (NL), initial nitrate concentration (Nini) and nitrogen residual (NR) for all treatments during the wheat season of 2013.

Effects of Irrigation on Water Movement and N Dynamics
The loss of nitrogen from irrigation systems posed a serious threat to receiving water bodies [37], and probably caused a considerable increase in nitrate concentration in the groundwater [18].Therefore, proper irrigation water management practices are needed that tactically allocate water and fertilizers to minimizing N leaching to groundwater [38].As is well-known, there are differences between drip irrigation and flood irrigation practice.For example, wetting patterns from a drip irrigation emitter are usually considered as the radius of the wetted volume [39].Drip irrigation can also maintain the desired concentration and distribution of ions and water in the soil [38], while minimizing leaching of N from the root zone of agricultural fields [24].However, the application of water and N in excess of crop requirements contributes to the leaching of water and N below the root zone under drip irrigation.Therefore, optimal irrigation and fertilizer scheduling of drip irrigation is important to minimizing leaching of N from agricultural fields.A clear understanding of water and N dynamics in the soil is also important for the management of irrigation and fertigation under drip irrigation.

Movement of Water for a Wheat Crop under Drip Irrigation
As the results showed, irrigation increased the soil moisture at the surface layer gradually and the soil moisture was influenced by the depth and the irrigation amount (Figures 2 and 3), which was consistent with the results of previous studies [27,40].In our study, the rate of N application did not significantly affect the variation in soil moisture (Figures 2 and 3), which was consistent with the results of Kong et al. [41].On the other hand, the water movement was mainly influenced by the irrigation schedule under drip irrigation, which was similar to results of a previous study under flood irrigation [42].This may be because the irrigation influenced the soil moisture directly, while the fertilizer only had minimal indirect effects on moisture [42].Compared to the W3 treatments, W1 and W2 treatments distributed more water in a greater depth of soil.This was consistent with the laboratory experimental results of Li et al. (2011), who showed that a higher irrigation level yielded a larger wetted depth [43].However, our results further illustrated that the wetted depth was stable when the irrigation level was at 0.5-0.8E for winter wheat in the NCP.

Distribution of NO3 − -N in Relation to the Movement of Water
As mentioned above, the NO3 − -N concentration variations in soil layers were significantly influenced by the depth and irrigation schedule, which was inconsistent with the results obtained by Behera and Panda (2009) in a semi-humid sub-tropical region [42].Due to the high concentration of urea with limited irrigation water applied, the NO3 − -N concentration in the 0-20-cm layer increased significantly with the decrease in irrigation level (Figures 4 and 5).Because of the larger wetted depth under the higher irrigation levels, the NO3 − -N concentration in the 20-60-cm layer was certainly influenced by irrigation schedule (Figures 2-5).No significant difference in NO3 − -N in the 60-100-cm layer was found for all N and irrigation treatments under drip irrigation, which was different from the results of Cassel Sharmasarkar et al. (2001) under traditional flood irrigation [21].In addition, The NO3 − -N concentration variation in the 0-60-cm layer was noticeably higher in 2013 than in 2012 under all treatments because of more dry spells during the period 183-239 DAS in 2013 (Figures 4 and 5, Table 1).The reason that the NO3 − -N concentrations were influenced by irrigation was mainly because transport of solute was the functions of convective and diffusive fluxes of water [44].In this study, due to the higher urea concentration, higher N applications significantly increased the NO3 − -N concentration in the 0-40-cm layer and led to higher N residual in the root zone of wheat, which was similar with previous results of Shi et al. (2012) [45].Although the irrigation water and N residual were primarily maintained in the upper 60 cm under drip irrigation, it could not be conclude that there is no deep percolation and leaching of nutrients beyond the root zone in this study.Since the distribution of soil moisture and NO3 − -N concentration was affected by irrigation schedule and N application rate, the deep percolation and N leaching would be different under different irrigation and fertilizer treatments.

Deep Percolation at Different Levels of N and Water under Drip Irrigation
Deep percolation is considered to be one of the main factors determining the amount of N leached [46].Matching irrigation rates with plant uptake ensures efficient water and N uptake while reducing deep percolation losses of water and nutrient [22].In our study, deep percolation increased with increasing irrigation level, but decreased with increasing N application level under drip irrigation (Tables 4 and 5), which was similar to the results of Behera and Panda (2009) [42].The reason for this may be because the soil water and N level affected the growth and amount of roots [47,48].However, in our study, it was still found that the deep percolation for winter wheat in the NCP was stable at the range of water level from 0.5 E to 0.8 E for drip irrigation.Additionally, for the W1 and W2 irrigation schedules, no significant decrease in deep percolation losses were found between N application rates of N120-N140 or N190-N290.This was mainly because root growth rate was mainly affected by soil water level and water flow could be enhanced only to roots exposed to high NO3-levels [48].The deep percolation was noticeably higher in 2012 than in 2013 under all treatments, possibly because of the higher precipitation in early growing season of 2012 (Table 1), when the roots of winter wheat were still fewer.
Deep percolation is difficult to measure in undisturbed soil conditions [49].Indirect methods based on a detailed knowledge of soil water dynamics improved our ability to estimate deep percolation losses from cropping systems [25,42], while indirect methods based on the soil water balance were also widely used to estimate deep percolation losses [50].The results illustrated that the deep percolation was underestimated using the soil water balance method (M2, Tables 3 and 4).This is probably because of underestimation of the water stress coefficient, as it is assumed that water can be stored in the root zone until field capacity is reached.The empirical equation (M3), which was assumed to only relate to the amount of irrigation, obviously neglected the effect of N levels and the precipitation distribution on deep percolation.In general, M3 was not recommended for estimating the deep percolation losses in the NCP under drip irrigation.

Risk of N Leaching at Different Levels of N and Water under Drip Irrigation
It is possible to control N leaching out of the root zone during the growing season with improved irrigation and fertilizer management.In our study, the N leaching at 60-cm depth was influenced by both irrigation schedule and N rates for the wheat crop under drip irrigation (Tables 6 and 7), which was similar to the findings of Behera and Panda [42], but inconsistent with the results of Sexton et al. (1998), who concluded that irrigation management is even more important than fertilizer management in reducing the N leaching losses [51].In addition, the results about N leaching at 100-cm depth illustrated that N leaching at 100-cm depth was not correlated to irrigation rate, N rate, or deep percolation for winter wheat in the NCP (Tables 6 and 7), which was similar with the results obtained by Allaire-Leung et al. in California [52].From the above results, it could be concluded that more N was retained in the 0-100-cm soil layer under drip irrigation for winter wheat in the NCP.It was also found that the N residual increased with the increase in N application level and decreased with the increase in irrigation water level (Tables 6 and 7).As has been demonstrated, the potential risk of N leaching increased with the increase in residual N after the wheat season in the NCP.Under the treatments with W3 irrigation rate, the N residual significantly increased.Likewise, when the N application was more than 190 kg•ha −1 , the N residual significantly increased as well.Crop production in the NCP is the typical double-cropping systems (wheat/maize rotations) and N residual is likely to be reused in the next season of maize.However, the possibility for leaching with N rates of 290 kg•ha −1 was too high to be considered sustainable in the NCP.Because of the higher possibility of leaching, the N rate of 290 kg•ha −1 is not recommended in the NCP.

Conclusions
Field experiments (2011-2013) were conducted in the North China Plain to evaluate the effects of N and irrigation levels on water movement and N transport for winter wheat crop under drip irrigation.Results of this study suggested that the water and N levels only correlated N leaching below the root zone (60-cm depth) positively during the winter wheat crop seasons, owing to the irrigation water and N primarily maintained in the root zone under drip irrigation.The distribution of soil moisture was mainly affected by irrigation schedule, but deep percolation at 1-m depth was significantly influenced by irrigation rate and N application rate, while stable at water levels of 0.5-0.8E with the same N application rate for winter wheat under drip irrigation.Additionally, at 0.5-0.8E, deep percolation was also kept stable at N rates of N120-N140 or at N190-N290.The empirical equation given by the Ministry of Geology and Mineral Resources seemed not to be appropriate for estimating the deep percolation of drip irrigation.Compared to the N leaching, the N residual in the 0-100-cm soil layer presented more risk to the environment under drip irrigation during the winter wheat seasons, but there is a need to further study the potential reuse of N residual during the next maize seasons.However, since the N residual was too high at 290 kg•ha −1 , it is concluded that the N rate of 290 kg•ha −1 should not be recommended for use with drip fertigation for winter wheat in the North China Plain.Improving irrigation and fertigation programs for agriculture irrigated with drip irrigation systems may still need further field experiments to study and model estimates in the North China Plain.

Figure 1 .
Figure 1.The location of the experimental station of the China Institute of Water Resources and Hydropower Research in the North China Plain is marked with a black dot.

Table 1 .
Monthly precipitation, average, minimum and maximum air temperatures (P, T, Tmin and Tmax, respectively) during the growing period (April-June) of experimental years 2012 and 2013 and long term averages (1951-2013) for the growth stages of winter wheat.

Table 2 .
Data and days after sowing of different growth stages for winter wheat during the experiment in 2012 and 2013.

Table 3 .
Application of N, irrigation amounts and evaporative demands of different treatments for winter wheat cultivation in 2012 and 2013.

Table 4 .
Deep percolation for all treatments during the wheat growing season of 2012.

Table 5 .
Deep percolation for all treatments during the wheat growing season of 2013.