Nitrate leaching in granitic regosol as affected by N uptake and transpiration by corn

Abstract

Field and soil column experiments were conducted to analyze the effects of N uptake and transpiration by corn on nitrate–nitrogen (NO₃-N) leaching in a granitic regosol and to evaluate the contribution of plant growth to the reduction of NO₃-N leaching. In the field experiment, NO₃-N leaching, N uptake by corn and retained N (inorganic N and microbial biomass N) were monitored in conventional planting density (CPD), high planting density (HPD) and no planting (NP) treatments. Nitrogen (N) was applied as (NH₄)₂SO₄ at the rate of 300 kg N ha⁻¹ and corn (Zea mays L.) was sown. In the soil column experiment, 1,500 mg N per column was applied and corn was sown in four treatments: no plants (NP), 1-plant, 2-plants and 4-plants per column. NO₃-N leaching, N uptake and transpiration by corn and retained NO₃-N in the soil were measured. In the field experiment, cumulative NO₃-N leaching in the NP treatment was 208 kg N ha⁻¹ from 38 to 49 days after treatment. In the CPD and HPD treatments, NO₃-N leaching was reduced to 148 kg N ha⁻¹ and 73 kg N ha⁻¹, respectively. Nitrogen uptake by corn and retained N in the soil increased with increasing plant density. Cumulative NO₃-N leaching was negatively correlated with N uptake by corn (r = −0.940, P < 0.01, n = 10). NO₃-N leaching decreased as N uptake by corn increased above 60 kg N ha⁻¹. In the soil column experiment, cumulative NO₃-N leaching decreased with increased planting density because of increased N uptake by corn and the amount of retained NO₃-N in the soil. The amount of retained NO₃-N in the soil was positively correlated with transpiration by corn (r = 0.943, P < 0.01, n = 12). We concluded that NO₃-N leaching from a granitic regosol during the rainy season could be reduced by increasing the planting density because of the increase in N uptake by plants and the increase of retained N in the soil derived from the increase in transpiration by plants.

Key words:

• corn
• granitic regosol
• N uptake
• nitrate leaching
• planting density
• transpiration

INTRODUCTION

To ensure crop yield or as a result of tradition more N fertilizer than necessary is applied to crops. Excess fertilized N is transformed into nitrate and is easily leached from arable soil to pollute rivers and marine and groundwater resources (Andraski et al. 2000; Thomsen et al. 1993; van der Ploeg et al. 1997). This nitrate–nitrogen (NO₃-N) leaching from arable soil is a worldwide concern. If rainfall is high in the period following harvesting of the main crop, as is the case in Europe, the use of cover crops is effective for reducing NO₃-N leaching. Cover crops reduce both the residual N in the soil and the deep percolation of rainfall, consequently reducing NO₃-N leaching (Brandi-Dohrn et al. 1997; Logsdon et al. 2002; Macdonald et al. 2005). It is believed that a decrease in soil N content by plant N uptake and a decrease in deep percolation by transpiration of plants can reduce NO₃-N leaching from arable soil.

The Chugoku district of Japan is characterized as having an Asian monsoon climate with heavy rains in the main cropping season. The heavy rain often occurs after the application of fertilizer for the main crop, which is in the early growing stage. Therefore, NO₃-N leaching occurs easily in soils with high NO₃-N content during cropping. NO₃-N leaching from arable soil depends on soil texture and is greatest in sandy soils (Bergström and Johansson 1991; Geleta et al. 1994; Sogbedji et al. 2000). Granitic regosols have a sandy texture with poor physical, chemical and biological characteristics and are widely distributed in the Chugoku district of Japan (Herai et al. 2006). In addition, NO₃-N leaching from arable soil depends on the rate of N fertilizer application and the soil N content (Decau et al. 2004; Trindade et al. 1997). Large amounts of NO₃-N leaching occur with chemical fertilizer application. To reduce NO₃-N leaching from granitic regosol during the June–July rainy season, two methods are generally considered. The reduction of the NO₃-N level in soil by plant N uptake is one method. This may be enhanced by controlling the planting density and timing of sawing, or by the selection of crops that have a higher rate of N uptake. To reduce the NO₃-N level in soil, it is also effective to use a controlled N release fertilizer (Quiroga-Garza et al. 2001) or a nitrification inhibitor (Ball-Coelho and Roy 1999). Increasing the amount of retained N (inorganic N and microbial biomass N) temporarily in the soil during the rainy season is the other method. This may be controlled by microbial biomass N formation using organic manure or by a decrease in the deep percolation of NO₃-N in the soil with transpiration by plants. The retained N in the soil is available for plants after the rainy season because NO₃-N leaching hardly occurs after the rainy season. Herai et al. (2006) reported that NO₃-N leaching from granitic regosol was less with compost application than with chemical fertilizer application, and that the increase in microbial biomass N with compost application reduced soil inorganic N and consequently reduced NO₃-N leaching.

Although NO₃-N leaching is reduced with compost application in the short term, in the long term compost application increases inorganic N derived from the decomposition of organic N accumulated in the soil, and there is a concern that as much or more NO₃-N leaching occurs as with chemical fertilizer application (Maeda et al. 2003). In addition, farmers often apply chemical fertilizer because it is easier to use than compost. In these situations, it is considered that the effect of plant growth on NO₃-N leaching becomes more important. In addition, the increases in N uptake and water-use efficiency with the development of root systems (Benjamin et al. 1996; Kristensen and Thorup-Kristensen 2004), the increase in microbial biomass with plant growth, and the rate of immobilization of N (Joergensen 2000) and denitrification with increasing microbial biomass (Qian et al. 1997) may contribute to reduce NO₃-N leaching. Therefore, it is considered that plant growth, which results in N uptake, transpiration and development of the root system, is an important factor affecting NO₃-N leaching from soil if heavy rain occurs in the main cropping season.

The objectives of this study were to analyze the effects of N uptake and transpiration by plants on NO₃-N leaching from a granitic regosol, and to evaluate the contribution of plant growth to the reduction of NO₃-N leaching. For this purpose, a field experiment was conducted using a N application of 300 kg N ha⁻¹ and three levels of planting density of corn. In addition, to analyze the effect of transpiration by plants on NO₃-N leaching, a soil column experiment was conducted with four levels of planting density.

MATERIALS AND METHODS

Field experiment

Experimental site

The field experiment was carried out at the Hiroshima University experimental field, Hiroshima Prefecture, Japan (34°23′N, 132°42′E). The experimental area falls within the Asian monsoon area and has a mean annual air temperature of 13°C and a mean annual precipitation of approximately 1,500 mm (25 year mean) with heavy rainfall in June–July. The weather data for the experimental site were obtained from a weather station located 2.6 km from the site. The soil was a granitic regosol with a loamy sand texture containing less than 10% clay content. Details of the soil chemical characteristics at initiation of the site were reported by Herai et al. (2006).

Experimental design

The experimental plot of chemical fertilizer (CF) – conventional planting density treatment (CPD of 60,000 plants ha⁻¹) with 80 cm row spacing and 20 cm plant spacing was initiated in 1996, and has been planted with corn as a summer crop. NO₃-N leaching, shoot dry matter and N uptake by corn, soil inorganic N and microbial biomass N were monitored in the CF-CPD treatment from May 2003 to September 2005. In the field experiment of 2005, a CF-high planting density treatment (HPD of 110,000 plants ha⁻¹) with 40 cm row spacing and 20 cm plant spacing, and a CF-no planting (NP) treatment were prepared in the same field block. The CF-NP and CF-HPD treatments have been kept free of vegetation since 1996, except for the CF-HPD treatment in 2002. Corn and oats were planted in the CF-HPD treatment in 2002. A plot measuring 5 m × 8 m was used for each treatment with two replicates.

Corn (Zea mays L. cv. snowdent127S) was sown on 21 May 2003, 18 May 2004 and 26 May 2005 in the CF-CPD and on 26 May 2005 in the CF-HPD treatment. Nitrogen was applied as (NH₄)₂SO₄ at 300 kg N ha⁻¹ in each treatment. An available P level of 100 kg P ha⁻¹ was maintained with superphosphate and fused magnesium

Table 1 Chemical characteristics of the soil before fertilizer application in the field experiment (2005)

	pH		C	N	Inorganic N (mg N kg^−1)		Olsen-P (mg P kg^−1)
	H2O	KCI	Total (g kg^−1)		NH4-N	NO3-N
CF-NP	77	6.0	2.5	0.3	1.9	0.0	24.8
CF-CPD	6.8	6.3	5.0	0.5	1.8	5.5	46.8
CF-HPD	7.9	5.6	2.7	0.3	2.4	0.0	26.1
Values are means (n = 3). CF-NP, chemical fertilizer-no planting; CF-CPD, chemical fertilizer-conventional planting density; CF-HPD, chemical fertilizer-high planting density.

phosphate. Potassium was applied at 200 kg K ha⁻¹ as K₂SO₄ and dolomite (containing 40% CaO and 15% MgO) was applied at 1 Mg ha⁻¹ in each treatment. The chemical characteristics of the soil before fertilizer application in the CF-CPD treatments in 2003 were: pH 6.9 (water), total carbon 4.4 g kg⁻¹ and total nitrogen 0.4 g kg⁻¹. The chemical characteristics of the soil in each treatment in 2005 are presented in Table 1.

Two replicate capillary lysimeters (height × width × length of 500 mm × 300 mm × 600 mm: DAIKI, DIK-6900 COMH-9-30, Kounosu city, Saitama, Japan) were used for each treatment and drainage water was sampled from 80 cm below the soil surface. Drainage samples from the lysimeter were collected 24 h after rainfall (at 28, 36, 43, 48, 54, 62, 83 and 103 days after treatment [DAT] in 2003, at 7, 15, 34, 45, 72, 91, 98 and 107 DAT in 2004, and at 20, 38–41, 44–47, 49, 54, 67, 85, 89 and 109 DAT in 2005). After measuring the drainage volume, drainage samples were stored at 4°C until NO₃-N analysis.

Soil was sampled before fertilizer application and subsequently at 0, 9, 15, 30, 60 and 90 DAT in each year. Soil samples were collected at a depth of 0–10 cm and sieved (< 2 mm) before being divided into two subsamples; one sample was stored at 4°C for soil NO₃-N, NH₄-N and microbial biomass N analysis and the other was air-dried for additional chemical analysis. Ten plants were collected from each of the two replicates of each treatment and the fresh weight of leaves, stems and grains was determined at 15, 30, 49, 60 and 90 DAT. Four plants in each treatment were dried at 80°C for at least 48 h before being weighed. The dried samples were finely ground and stored for chemical analysis.

Soil column experiment

The soil column experiment was conducted from 15 July to 11 August 2005 in a greenhouse. The four treatments used differed in planting density: zero (no plant), one, two and four plants per column (NP, 1-plant, 2-plant and 4-plant treatments). All treatments were replicated six times making a total of 24 soil columns and half of the soil columns were sampled before water application on 1 August. The columns were made of cylindrical polyvinyl chloride (PVC) tubes (inner diameter 20 cm and length 80 cm). The bottom of the PVC tube was capped with a PVC plate that had a hole covered with nylon mesh that was connected to a plug with a tube for collecting drainage samples. The soil column was filled with a sieved (< 2 mm) granitic regosol consisting of 60 cm of subsoil and 15 cm of topsoil. The bulk density of the soil used in the soil column experiment was similar to that of the field soil (1.3 g cm⁻³).

On 15 July, phosphorus and potassium were applied as superphosphate at 200 kg P ha⁻¹ and as K₂SO₄ at 200 kg K ha⁻¹, respectively, and dolomite was applied at 500 kg ha⁻¹ for the 15 cm topsoil in each treatment to adjust the soil pH to 6.0. Nitrogen was applied to the soil surface using Ca(NO₃)₂4H₂O at 1,500 mg N per column (corresponding to 476 kg N ha⁻¹) to avoid the influence by nitrification of ammonium sulfate. After N application, corn was sown in the 1-plant, 2-plant and 4-plant treatments. From 31 July to 9 August, deionized water corresponding to 40 mm of precipitation was applied every day to the soil surface in each soil column. The total amount of water applied to the soil column (400 mm) was equal to the average amount of precipitation received during the rainy season in this area. Evapotranspiration from the soil column was calculated by measuring the weight of the soil column every day before water was applied. Plant weight during the water application period was estimated using a growth curve and the plant weight at the beginning and end of the water application period, then subtracted from the weight of the soil column. The difference in evapotranspiration between the NP treatment and the 1-plant, 2-plant and 4-plant treatments was assumed to be transpiration by the plant. Drainage samples were collected 24 h after water was applied and stored at 4°C until NO₃-N analysis. On 1 and 11 August, shoots were cut at the soil surface and roots and soils were sampled from the topsoil and every 10 cm of the subsoil in each soil column. Roots were sieved from the soil and washed with deionized water. The fresh weights of shoots and roots were measured and dried at 80°C for at least 48 h before being weighed. The dried samples were finely ground and stored for chemical analysis. Soil samples were stored at 4°C until NO₃-N analysis.

Chemical analysis

Drainage samples from the lysimeter and soil column were filtered through a 0.2 µm membrane filter (Dismic-13cp, Toyo Roshi, Tokyo, Japan) and the NO₃-N concentration of the samples was determined using ion chromatography. The system was composed of an electric conductivity detector (CM-8000, TOSOH Corporation, Tokyo, Japan) and a thermostatic box (CO-8020, TOSOH Corporation) for the separation column (TSKgel IC-Anion-PW, TOSOH Corporation). The amount of NO₃-N leached was determined by multiplying the drainage volume by the NO₃-N concentration in the drainage samples.

For soil NH₄-N analysis, moist soil containing 10 g dry soil was extracted with 100 mL of 2 mol L⁻¹ KCl solution for 1 h. The extracts were filtered using ADVANTEC No.5C filter paper (Toyo Roshi) and the ammonium concentration was determined using the indophenol blue method with a spectrophotometer (Shimadzu, UVmini-1240, Kyoto, Japan). For soil NO₃-N analysis, moist soil containing 10 g dry soil was extracted with 100 mL of deionized water for 1 h. The extracts were filtered using ADVANTEC No.5C filter paper and the NO₃-N concentration was determined using ion chromatography. The amounts of microbial biomass C and N in the soil were determined using the chloroform fumigation extraction method (Brookes et al. 1985), and the amount of total organic C and total N dissolved in 0.5 mol L⁻¹ K₂SO₄ extracts were determined using an organic C and total N analyzer (Shimadzu, TOC-V CSN TNM-1, Kyoto, Japan). Microbial biomass C (B _C ) and N (B _N ) were calculated from the relationships B _C  = 2.22 × E _C and B _N  = 2.22 × E _N , respectively (Chowdhury et al. 1999), where E _C and E _N correspond to the amount of organic C and total N extracted from the fumigated samples minus the amount in the non-fumigated samples, respectively. Microbial biomass C and N in 0–10 cm soil were determined by multiplying the bulk density (1.31 g cm⁻³). Total C and N in the plant and soil samples were determined using the combustion method with a CN analyzer (Yanaco, MT-700, Kyoto, Japan).

Statistical analysis

Statistical analyses were carried out using the statistical package SPSS for Windows (version 13.0) with the level of significance (P value) set at P < 0.01. Means and standard deviations for each treatment were calculated.

Figure 1  Cumulative nitrate–nitrogen (NO3-N) leaching and shoot dry matter in the chemical fertilizer-conventional planting density treatment in the field experiment from May 2003 to September 2005.

RESULTS

Field experiment

Drainage volume, NO₃-N concentration and NO₃-N leaching

Cumulative NO₃-N leaching and shoot dry matter in the CF-CPD treatment from May 2003 to September 2005 is presented in Fig. 1. Cumulative NO₃-N leaching in the CF-CPD treatment showed a similar trend each year and increased markedly during the period between 30 and 60 DAT. The total amount of rainfall during the corn cropping season was 842 mm (3-year average).

In 2005, heavy rainfall was received 36–48, 81 and 103 DAT (Fig. 2A). The amount of rainfall 36–48 DAT was 308 mm and particularly heavy rainfall of 213 mm was received 36–39 DAT, including more than 70 mm day⁻¹ 37 and 38 DAT and more than 100 mm day⁻¹ of rainfall was observed 81 and 103 DAT. The drainage volume markedly increased from 38 to 41 DAT and small amounts of drainage were observed 44–49, 85–89 and 109 DAT (Fig. 2B). The drainage volume was a little lower in the CF-HPD treatment than the CF-NP and CF-CPD treatments, but the differences were not statistically significant.

The NO₃-N concentration in the drainage water during the period of 38–41 DAT was lowest in the CF-HPD treatment, followed by the CF-CPD treatment and the CF-NP treatment (Fig. 3A). The NO₃-N concentration in the drainage water was a little higher in the CF-HPD treatment than the CF-CPD treatments 44–49 DAT, but the difference was not statistically significant. The NO₃-N concentration in the drainage water remained above 50 mg N L⁻¹ in the CF-NP treatment after 49 DAT, while remaining very low in the CF-CPD and CF-HPD treatments.

Figure 2  (A) Precipitation and (B) cumulative drainage volume from the capillary lysimeter in the field experiment (2005). CF-NP, chemical fertilizer-no planting; CF-CPD, chemical fertilizer-conventional planting density; CF-HPD, chemical fertilizer-high planting density. Error bars represent standard deviation.

Figure 3  Effect of planting density on (A) nitrate–nitrogen (NO3-N) concentration in the drainage water and (B) cumulative NO3-N leaching from the capillary lysimeter in the field experiment (2005). CF-NP, chemical fertilizer-no planting; CF-CPD, chemical fertilizer-conventional planting density; CF-HPD, chemical fertilizer-high planting density. Error bars represent standard deviation.

NO₃-N leaching markedly increased from 38 to 41 DAT and very little leaching was observed 44–49, 85–89 and 109 DAT in all treatments (Fig. 3B). Cumulative NO₃-N leaching until 49 DAT in the CF-NP, CF-CPD and CF-HPD treatments was 208, 148 and 73 kg N ha⁻¹, respectively, and NO₃-N leaching decreased with increasing plant density. Total NO₃-N leaching until 109 DAT in the CF-NP, CF-CPD and CF-HPD treatments was 298, 149 and 80 kg N ha⁻¹, respectively. The ratio of cumulative NO₃-N leaching at 49 DAT to the total amount at 109 DAT in the CF-NP, CF-CPD and CF-HPD treatments was 70, 99 and 91%, respectively.

Shoot dry matter and N uptake

Shoot dry matter was 2–4 Mg ha⁻¹ 49 DAT and reached 10–14 Mg ha⁻¹ at harvest in the CF-CPD treatment from 2003 to 2005 (Fig. 1) Shoot dry matter was lower in 2003 compared to 2004 and 2005. In 2005, shoot dry matter 30 DAT was 0.4 Mg ha⁻¹ in the CF-CPD and 1.2 Mg ha⁻¹ in the CF-HPD treatment (Table 2). Forty-nine DAT, shoot dry matter was 1.4–1.7-fold higher in the CF-HPD treatment than in the CF-CPD treatment. Shoot N uptake was 12 kg N ha⁻¹ in the CF-CPD and 40 kg N ha⁻¹ in the CF-HPD treatment 30 DAT and this increased to 55 kg N ha⁻¹ and 103 kg N ha⁻¹, respectively, 49 DAT (Table 2). Shoot N uptake was 1.2–3.3-fold higher in the CF-HPD treatment than in the CF-CPD treatment during the cropping season.

Inorganic N in the surface soil and soil microbial biomass N

Inorganic N in the surface soil (0–10 cm) was similar in all treatments (Fig. 4) and was high (approximately 300 kg N ha⁻¹) up to 15 DAT, and decreased to 100 kg N ha⁻¹ at 30 DAT and became very low after 60 DAT.

Figure 4  Effect of planting density on inorganic N in the surface soil (0–10 cm) in the field experiment (2005). CF-NP, chemical fertilizer-no planting; CF-CPD, chemical fertilizer-conventional planting density; CF-HPD, chemical fertilizer-high planting density. Error bars represent standard deviation.

Table 2 Shoot dry matter and shoot N uptake by corn in the field experiment (2005)

		15 DAT	30 DAT	49 DAT	60 DAT	90 DAT
Soot dry matter (Mg ha^−1)	CF-CPD	0.02 ± 0.0	0.4 ± 0.1	3.5 ± 0.2	5.7 ± 0.5	14.4 ± 0.8
	CF-HPD	0.04 ± 0.0	1.2 ± 0.3	5.9 ± 0.4	9.1 ± 1.4	20.7 ± 2.9
Shoot N uptake (kg N ha^−1)	CF-CPD	0.9 ± 0.2	12.3 ± 1.4	54.6 ± 0.4	59.9 ± 6.3	111.6 ± 7.6
	CF-HPD	1.6 ± 0.3	40.4 ± 9.4	103.2 ± 16.6	97.8 ± 2.1	130.7 ± 21.8
Values are means ± standard deviation (n = 3). CF-CPD, chemical fertilizer-conventional planting density; CF-HPD, chemical fertilizer-high planting density; DAT, days after treatment.

Table 3 Microbial biomass N in the field experiment (2005)

		0 DAT	15 DAT	30 DAT	60 DAT	90 DAT
Microbial biomass	CF-CPD	15.2 ± 1.8	19.9 ± 1.0	22.1 ± 0.3	30.2 ± 4.4	27.9 ± 1.4
N (kg N ha^−1)	CF-NP	15.2 ± 1.8	14.1 ± 2.7	15.4 ± 1.7	16.3 ± 2.6	22.2 ± 1.3
Values are means ± standard deviation (n = 3). CF-CPD, chemical fertilizer-conventional planting density; CF-HPD, chemical fertilizer-high planting density; DAT, days after treatment.

Soil microbial biomass N in the CF-CPD treatment increased slightly up to 30 kg N ha⁻¹ 60 DAT and then remained constant (Table 3). Microbial biomass N in the CF-NP treatment was 14 kg N ha⁻¹ at 15 DAT and slightly increased during the cropping period. The amount of microbial biomass N in the CF-CPD treatment was approximately 6–14 kg N ha⁻¹ higher than in the CF-NP treatment.

Soil column experiment

NO₃-N leaching

NO₃-N leaching from the soil column increased from 3 to 7 DAT (timing of water application) and was then negligible after 7 DAT (Fig. 5). Total amounts of NO₃-N leaching until the end of the water application period were 416, 427, 403 and 374 kg N ha⁻¹ in the NP, 1-plant, 2-plant and 4-plant treatments with the 4-plant treatment being significantly less than the NP treatment.

Drainage and evapotranspiration

The drainage volume from the soil column decreased with increasing planting density. The total drainage volumes during the 10-day water application period were 331, 307, 281 and 266 mm in the NP, 1-plant, 2-plant and 4-plant treatments, respectively. The drainage volume decreased by 24, 50 and 65 mm in the 1-plant, 2-plant and 4-plant treatments compared to the NP treatment (Fig. 6A). The cumulative evapotranspiration from the soil column during the water application period increased from 61 to 118 mm with increasing planting density (Fig. 6B). Assuming that the differences in evapotranspiration between the NP treatment and

Figure 5  Effect of planting density on nitrate–nitrogen (NO3-N) leaching in the soil column experiment. NP, no plant; 1-plant, one plant per soil column; 2-plant, two plants per soil column; 4-plant, four plants per soil column. Error bars represent standard deviation.

the 1-plant, 2-plant and 4-plant treatments resulted from transpiration by corn, transpiration by corn was estimated to be 19−57 mm for 10 days.

NO₃-N concentration in the drainage water

The NO₃-N concentration in the drainage water was highest 4 DAT in all treatments and increased up to 600 mg N L⁻¹ (Fig. 7). In the 1-plant, 2-plant and 4-plant treatments, NO₃-N concentrations were lower 3 DAT and higher 5–6 DAT than in the NP treatment.

Dry matter and N uptake

Dry matter and N uptake of shoots and roots at the beginning and end of the water application period are

Figure 6  Effect of planting density on (A) cumulative drainage volume and (B) cumulative evapotranspiration in the soil column experiment. NP, no plant; 1-plant, one plant per soil column; 2-plant, two plants per soil column; 4-plant, four plants per soil column. Error bars represent standard deviation.

Figure 7  Effect of planting density on nitrate–nitrogen (NO3-N) concentration in the drainage water in the soil column experiment. NP, no plant; 1-plant, one plant per soil column; 2-plant, two plants per soil column; 4-plant, four plants per soil column. Error bars represent standard deviation.

presented in Table 4. Shoot dry matter was 0.1–0.5 Mg ha⁻¹ at the beginning of the water application period and increased to 0.8–2.8 Mg ha⁻¹ at the end. Root dry matter was 0.04–0.2 Mg ha⁻¹ at the beginning of the water application period and increased to 0.4–1.4 Mg ha⁻¹ at the end. Root dry matter was 40–50% of shoot dry matter. Shoot N uptake at the beginning of the water application period was 2.2–20.8 kg N ha⁻¹ and increased to 9.7–25.1 kg N ha⁻¹ at the end. Shoot N uptake corresponded to 0.5–4% and 2–5% of applied N at the beginning and end of the water application period, respectively. Root N uptake at the end of the water application period was 2.9–9.6 kg N ha⁻¹ corresponding to 30–39% of shoot N uptake.

DISCUSSION

Relationships between NO₃-N leaching, retained N in soil and N uptake by corn

In the field experiment, a large amount of NO₃-N leaching was observed after more than 300 mm of heavy rainfall 36–48 DAT, including more than 70 mm day⁻¹ for two consecutive days (Figs 2A,3B) in 2005. The drainage volume was not significantly different among the three treatments (Fig. 2B), but NO₃-N concentration in the drainage water significantly decreased in the CF-CPD and CF-HPD treatments compared to the CF-NP treatment 36–48 DAT (Fig. 3A). In addition, the transpiration by corn would be low on a rainy day. The effect of the transpiration by corn on drainage volume was probably negligible under heavy rainfall conditions of more than 100 mm over 1–2 days. Therefore, the main reason for the decrease in NO₃-N leaching in the CF-CPD and CF-HPD treatments was the decrease in NO₃-N concentration in the drainage water. NO₃-N concentration in the drainage water during the rainy season was closely linked to soil NO₃-N contents in the 0–30 cm soil profile (Roth and Fox 1990) and in the 0–120 cm soil profile (Jemison and Fox 1994) under field conditions. In our experiment, inorganic N in the surface

Table 4 Dry matter and N uptake at the beginning and the end of the water application period in the soil column experiment

	Dry matter (Mg ha^−1)				N uptake (kg N ha^−1)
	Shoot		Root		Shoot		Root
	Beginning	End	Beginning	End	Beginning	End	Beginning	End
1-plant	0.1 ± 0.0	0.8 ± 0.1	0.0 ± 0.0	0.4 ± 0.1	2.2 ± 0.3^†	9.7 ± 0.5	n.d.	2.9 ± 0.5
2-plant	0.2 ± 0.0	1.8 ± 0.3	0.1 ± 0.0	0.9 ± 0.1	6.0 ± 0.3^†	17.0 ± 2.2	n.d.	6.6 ± 0.2
4-plant	0.5 ± 0.0	2.8 ± 0.1	0.2 ± 0.0	1.4 ± 0.0	20.8 ± 2.6	25.1 ± 2.4	4.1 ± 0.6	9.6 ± 0.3
Values are means ± standard deviation (n = 3). ^†Values including root N uptake. n.d., not determined.

soil was measured. However, most of the (NH₄)₂SO₄ applied to the soil was converted into NO₃-N by nitrification in the surface soil and NO₃-N easily moved to the subsoil with rainfall percolation, and the inorganic N in the surface soil was very low 30 DAT (Fig. 4). Consequently, the relationship between NO₃-N leaching and inorganic N in the surface soil was not clear. Therefore, the retained N in the soil from 0 to 80 cm was estimated from the rate of N fertilizer application, plant N uptake and NO₃-N leaching. In this experiment, total N and inorganic N in the soil were very low before fertilizer application (Table 1) and the amount of mineralized N from accumulated organic matter in the CF-CPD treatment was less than 1.5 kg N ha⁻¹ for 50 days (data not shown). Hence, the N applied in the preceding cropping season and the N mineralized from accumulated organic matter had a negligible influence on NO₃-N leaching. In addition, there was minimal rainfall before the NO₃-N leaching increase 38–49 DAT, possibly resulting in very small N loss from the denitrification, probably less than 5 kg N ha⁻¹ (Nishio et al. 2002). Therefore, it could be assumed that the retained N in the soil was almost the same as the difference between N applied in fertilizer and the N lost through plant N uptake and NO₃-N leaching. Root N uptake (data not shown) was 10–20% of shoot N uptake, which was calculated using the shoot to root ratio and root N concentration obtained from the field experiment. The whole plant N uptake (the sum of shoot and root N uptake), retained N in the soil (calculated as described above) and cumulative NO₃-N leaching from the soil in each treatment are presented in Fig. 8A–C. Forty-nine DAT, the amount of retained N in the soil tended to increase with increasing plant density, and was 92, 93 and 115 kg N ha⁻¹ in the soil in the CF-NP, CF-CPD and CF-HPD treatments, respectively. When 300 kg N ha⁻¹ of fertilizer was applied, approximately 90 kg N ha⁻¹ of NO₃-N was retained in the soil temporarily; therefore, approximately 210 kg N ha⁻¹ of NO₃-N was available to be leached by rainfall or N uptake by corn during the June–July rainy season. In the CF-NP treatment without

Figure 8  Cumulative nitrate–nitrogen (NO3-N) leaching, whole plant N uptake and retained N in the soil in the (A) chemical fertilizer-no planting, (B) chemical fertilizer-conventional planting density and (C) chemical fertilizer-high planting density treatments in the field experiment (2005).

Figure 9  Relationship between cumulative nitrate–nitrogen (NO3-N) leaching and shoot N uptake in the field experiment during the rainy season from 2003 to 2005. The data were from measurements taken 49 days after treatment (DAT) in each year. **P < 0.01.

N uptake by corn, 208 kg N ha⁻¹ of NO₃-N was leached 49 DAT, while in the CF-CPD and CF-HPD treatments, N uptake by corn increased to 59 and 112 kg N ha⁻¹, respectively. In addition, the retained N in the soil increased and NO₃-N leaching decreased. The relationship between cumulative NO₃-N leaching and shoot N uptake during the June–July rainy season from 2003 to 2005 is presented in Fig. 9. The data in this relationship were from measurements at 49 DAT in each year. Cumulative NO₃-N leaching and shoot N uptake showed a high negative correlation (r = −0.940, P < 0.01, n = 10). This correlation indicated that increased N uptake by corn during the rainy season was directly reflected in decreased NO₃-N leaching. In the CF-NP treatment, the amount of 62 kg N ha⁻¹ was leached from the retained N in the soil from 49 to 90 DAT. However, in the CF-CPD and CF-HPD treatments, NO₃-N leaching hardly occurred because N uptake by corn increased to 133 and 162 kg N ha⁻¹, respectively. This indicated that N uptake by corn after the rainy season was necessary to reduce NO₃-N leaching after 49 DAT.

Relationships between NO₃-N leaching, retained NO₃-N in soil and transpiration by corn

The effect of transpiration by corn on drainage volume was observed in the soil column experiment with the rate of water application set at 40 mm day⁻¹ for 10 days. Cumulative NO₃-N leaching, the amount of retained NO₃-N

Figure 10  Effect of planting density on cumulative nitrate–nitrogen (NO3-N) leaching, plant N uptake and the amount of retained NO3-N in the soil at the end of water application in the soil column experiment. Error bars represent standard deviation.

Figure 11  Relationship between the amount of retained nitrate–nitrogen (NO3-N) in the soil at the end of water application and cumulative transpiration by corn for 10 days of water application in the soil column experiment. **P < 0.01.

in the soil and N uptake by corn at the end of the water application period in each treatment are presented in Fig. 10. Nitrogen uptake by corn and the amount of retained NO₃-N in the soil increased and NO₃-N leaching decreased with increasing plant density. The relationship between the amount of retained NO₃-N in the soil at the end of water application and cumulative transpiration by corn for 10 days of water application period is presented in Fig. 11. The amount of retained NO₃-N in the soil and cumulative transpiration by corn showed a high positive correlation (r = 0.943, P < 0.01, n = 12) and the amount of retained NO₃-N in the soil increased by 5–20 kg N ha⁻¹ (corresponding to 1–5% of NO₃-N leaching from the NP treatment; 416 kg N ha⁻¹) with 40–60 mm (corresponding to 10–15% of the total amount of precipitation for 10 days; 400 mm) of transpiration by corn. The increase in NO₃-N concentration in the drainage water of the plant treatments tended to be delayed compared to the NP treatment (Fig. 7). This showed that the increase in transpiration by corn delayed the downward movement of NO₃-N in the soil and consequently increased the amount of retained NO₃-N in the soil.

In the field experiment, transpiration by corn during the rainy season was probably negligible because of heavy rain. The retained N in the field soil could be increased if the transpiration rate of corn was high, as in the soil column experiment. NO₃-N leaching hardly occurred after the rainy season under field conditions because of the little amount of rainfall and plant uptake of retained N in the soil. Therefore, it was considered that the increase of retained N in the soil derived from the increasing of transpiration by corn in the field was effective to reduce NO₃-N leaching during the rainy season.

In this study, it can be concluded that NO₃-N leaching from a granitic regosol during the rainy season can be reduced by increasing plant density because of an increase in N uptake by the plants and an increase in retained N in the soil derived from increasing transpiration by plants.

ACKNOWLEDGMENTS

The authors would like to acknowledge the Ministry of Agriculture, Forestry and Fisheries of Japan for financial support. We thank Professor Minoru Yamauchi (National Agricultural Research Center for Western Region) for his valuable advice and discussions.