Introduction

Agricultural production is the largest water consumer, accounting for more than 60% of total water consumption in China (The Ministry of Water Resources of the People’s Republic of China 2008). Hence, in resolving water shortage, more and more countries are paying greater attention to agricultural practices. The efficient utilization of available water resources is crucial because although China hosts 22% of the global population, it only has 7 and 6% of the world’s farmlands and water resources, respectively (National Economic and Social Development 2008). Therefore, techniques to save irrigation water and increase crop water-use efficiency (WUE) are necessary.

Levels of nitrogen fertilizer application have increased sharply in north China in recent years (Zhu et al. 2005), resulting in nitrate leaching and groundwater contamination (Rossi et al. 1991; Cameron et al. 1997; Zhu et al. 2004). It was reported that overfertilization in north China is leading to high concentrations of nitrate in groundwater and drinking water (average of 68 mg/L) (Li et al. 2001; Zhu et al. 2004), whereas the percentage of applied N intake by crops is below 40% (Zhang et al. 1996).

Therefore, there is an urgent need to maximize the use of water and fertilizer to minimize nitrate leaching and groundwater contamination, and to achieve optimal agronomic, economic, and environmental benefits.

Surface drip irrigation (DI) introduces water and nutrients to soil surface near roots through emitters. Subsurface drip irrigation (SDI) laterals are buried underground, supplying water and nutrients directly to root zones (Phene and Beale 1979; Lamm 1995; Camp et al. 1997). DI and SDI are modern and enhanced water- and fertilizer-saving methods (Phene et al. 1991; Solomon and Jorgensen 1992; Lamm 1995, 2002; Camp 1998; Ayars et al. 1999).

There are limited comparative studies on SDI and DI under different fertigation conditions. Several studies have compared the effects of SDI and DI on crop yield, proving that SDI generally results in higher crop yield (Hanson et al. 1997; Hanson and May 2004; Gencoglan et al. 2006; Patel and Rajput 2009).

Under SDI or DI, irrigation volume and irrigation intervals have significant impacts on crop growth, development, and yield. Shorter irrigation intervals from 3–6 days could result in higher crop yield and water use (Sezen et al. 2005, 2006; Sensoy et al. 2007). However, different SDI frequencies (1 and 7 days) only have a slight effect on corn yields (Howell et al. 1997).

Irrigation methods affect root distribution and nitrogen utilization. Under DI, roots grow preferentially around the wetted emitter area and are concentrated within the top 40 cm of the soil profile (Oliveira et al. 1996; Machado et al. 2003). Larger concentrations of crop roots around the drip tube are found under SDI (Coelho and Dani 1999; Machado et al. 2003). The center of gravity of the root system in SDI and DI fertigation is near the emitter below the soil surface (Martinez Hernandez et al. 1991). Moreover, SDI treatment results in a wider distribution of roots across the bed compared with DI (Zotarelli et al. 2009). Hence, SDI can reduce percolation below the root zone (Hanson and May 2004) and decrease groundwater NO3 −-N pollution (Phene 1999). Meanwhile, too much N fertilizer application does not increase crop production (Camp et al. 1997; Thompson et al. 2002a, b; Sorensen et al. 2004; Mahajan and Singh 2006; Cabello et al. 2009). SDI has higher soil fertility after potato harvesting compared with DI, although this effect is not significant (Selim et al. 2009).

This study aims to investigate the root development, root distribution, nitrogen distribution, and crop yield of bell pepper under SDI and DI, considering different nitrogen fertilization levels, and to determine best management practices for bell pepper production.

Materials and methods

Site description

Field experiment was conducted at Yuhe Irrigation Experiment Station, Datong, Shanxi Province (40°06′N, 113°20′E, and elevation of 1,052 m above sea level). The soil is gravelly loam with 22.5% field capacity. Groundwater level was kept at about 19 m deep. The climate in Datong is semiarid, with an average annual precipitation of approximately 379 mm and an annual evaporation of about 1,152 mm. More than 60% of the annual precipitation occurs during the growing season, which extends from late May to mid-September. The frost-free period is about 110–130 days. Table 1 summarizes the monthly mean climatic data of 2007 and 2008 compared with the mean long-term data for the experimental district. The growing season temperatures in both years were typical of the long-term means at the site. Precipitations in 2007 growing season were all smaller than the long-term means. Total precipitation in 2008 was coincident with the long-term means. However, precipitations of June and September in 2008 were greater than the long-term means. Precipitation in July 2008 was 68 mm lower than the long-term means. The daily rainfall and temperature at the experimental site in 2007 and 2008 are shown in Fig. 1. The monthly average temperatures in 2008 were lower than those in 2007 except July, and a very low temperature (2.4°C) in May 30, 2008, inhibited seedling establishment.

Table 1 Historical monthly and growing season climatic data of the experimental area
Full size table
Fig. 1
figure 1

Meteorological data during the crop growth season. a Cumulative and daily average rainfall during the crop season and b daily average temperature

Full size image

Experimental treatments and field preparation

The field experiment was conducted using a randomized complete block design with eight treatments, including two irrigation techniques (SDI, DI) and four fertilization levels of 0, 75, 150, and 300 kg/ha N (N0, N75, N150, and N300, respectively) (Table 2). Each treatment was replicated thrice. There were 24 plot measuring 7 m × 3 m (6 rows per plot) (Fig. 2).

Table 2 Experimental treatments
Full size table
Fig. 2
figure 2

Experimental plot arrangement. IS stands for isolation strip; N 1, N 2, N 3, and N 4 stand for nitrogen levels of 0, 75, 150, and 300 kg/ha, respectively

Full size image

The test crop was Tongfeng 16,a local variety of bell paper. Two-month-old pepper plants were transplanted in the field 40 cm in rows and 50 cm between plants on May 20, 2007, and May 21, 2008. The plants were irrigated with either SDI or DI systems installed prior to planting. The laterals were installed between every other crop row at a space of 1 m, whereas the SDI laterals were buried at a depth of 20 cm between crop rows. Water was supplied every 4 days using laterals (Netafim super Taphoon 125) with 1.1 L/h of drippers discharge at a spacing of 40 cm.

Soil water content was measured by a portable time domain reflectometry (TDR). Three access tubes were placed at a distance of 0, 25, and 50 cm from lateral pipes for each plot at a depth of 1 m. A total of 24 PVC access tubes were installed (Fig. 3). Volumetric water content for all treatments at 0–20, 20–40, 40–60, 60–80, and 80–100-cm layers was measured before and after each irrigation and after each rainfall.

Fig. 3
figure 3

Layout of trime PVC tubes

Full size image

To measure leaching, a lysimeter (Fig. 4) was installed for each treatment. The lysimeter is a rectangular plastic box measuring 70 × 50 × 90 cm. The top of the lysimeters was buried at 10 cm below the soil surface. The soil was back-filled in layers to maintain its original structure. Drainage water was collected from the outlet at the bottom of lysimeters using a bottle placed in a separate tube next to the lysimeters and measured once every 2 days. Two crops were transplanted in each lysimeter.

Fig. 4
figure 4

Structure of a simple lysimeter

Full size image

Before the plants were transplanted, three sampling points were randomly selected from the entire experimental field to determine basic soil parameters, including soil texture, bulk density, field capacity, and permanent wilting point (Table 3).

Table 3 Average soil physical parameters at the experimental site
Full size table

Irrigation application

Irrigation volume was determined every 4 days based on the difference between estimated ET p and measured effective rainfall. Penman–Monteith’s formula, multiplying crop coefficients from FAO-56 (Allen et al. 1998), was used to estimate ET p. Weather data were collected from an automatic weather station, 20 m away from the field site.

Evapotranspiration estimation

Actual crop evapotranspiration was estimated using the following water balance equation:

$$ ET_{\text{c}} = \Updelta W + I + P - R - D $$
(1)

where ∆W is the change of soil water storage (mm), I is irrigation amount (mm), P is precipitation (mm), R is surface runoff (mm), and D is the deep percolation (mm). The calculation soil layer was set at 0–100 cm.

The ∆W was estimated using the change in soil water content in the soil profile. Surface runoff was ignored, and D was considered as the volume of water drained from the lysimeters.

Nutrient management

Organic fertilizer (chicken manure: 11.1 m3/ha) was uniformly applied in all plots as basal fertilizer before land leveling. During the growth season, urea dissolved in irrigation water in a fertilizer tank was applied according to nitrogen fertilization application levels of the different treatments. In 2007, nitrogen application was applied at July 23, August 16, and September 9 for three different growth stages. But in 2008, nitrogen was applied only two times at July 31 and September 3.

Before and after fertilization, soil samples were collected from 0–20, 20–40, 40–60, 60–80, and 80–100-cm soil layers and at 0, 25, and 50 cm distance from the lateral pipes. Soil samples were then pulverized and sifted out using a 2-mm sieve, soil (10 g) was weighed with an electronic balance, and 50 mL KCl solution (1 mol/L) was added to the soil through a pipette. The samples were then shook for 1 h on a shaker and filtrated through a 0.45-μm membrane after clarification. NO3 −-N content in the soil leach liquor was calculated, assuming that all NO3 −-N were dissolved in water. After leaching, dissolved NH4 +-N and NO3 −-N were determined directly by flow injection analysis (Skalar San++).

Yield

Bell pepper was manually picked eight times in 2 years (August 4, August 23, September 12, and September 14, 2007; and July 29, August 24, September 8, and September 12, 2008). Bell pepper yield was determined by harvesting bell pepper from two adjacent center rows in each plot. ANOVA was performed using the SPSS software package (SPSS V17.0). Significant differences between means for different treatments were compared using Duncan’s test at P < 0.05.

Root sampling and analysis

Soil samples containing crop roots were taken from center rows 1 days after harvest. The soil surface sampling area was 40 cm × 50 cm, with a plant at the middle. Samples were taken at three different depths (0–10, 10–20, and 20–30 cm) in 2007 and four different depths (0–10, 10–20, 20–30, and 30–40 cm) in 2008. After washing away soils using a 0.5-mm sieve, crop roots and organic debris were stored in plastic bags at 4°C until further cleaning and then placed in a glass bowl. Crop roots were handpicked and placed in glass dishes. Root length and other root characteristic parameters were determined with using the Winrhizo (Re′gent Instrument Inc., Quebec City, Canada) software and hardware. ANOVA was performed using the SPSS software package.

Results and discussion

Water application

Rainfall during the crop growing season and the volume of irrigation water are shown in Table 4. The crops were irrigated 14 times, and the total irrigation volume was 257 mm in 2007. However, irrigation frequency was reduced to 10 times in 2008 (164 mm).

Table 4 Irrigation application and rainfall (mm)
Full size table

Seasonal patterns of soil moisture

Average soil moisture of different points (0, 25, and 50 cm from the lateral pipes) in different profiles (0–20, 20–40, 40–60 cm) for 2 years (2007 and 2008) under the 150 kg/ha nitrogen treatment is presented in Fig. 5a as an example for the same nitrogen application. Soil water content in the 0–20-cm layer fluctuating violently, and the moisture there under DI N150 was slightly higher than under SDI N150. Soil water content under SDI N150 was higher than under DI N150 in the 20–40-cm layer, and there were also significant fluctuations of the moisture. However, soil moisture remained almost the same level, and fluctuations were not obvious in 40–60-cm layer under SDI and DI.

Fig. 5
figure 5

Soil moisture at different layers. a Soil water content in 0–20, 20–40, 40–60 cm depth under N150. b Average soil moisture of 0–40 cm depth under different nitrogen treatments

Full size image

Figure 5b showed the average soil moisture of three points (0, 25, and 50 cm from the lateral pipes) and two profiles (0–20, 20–40 cm) for 2 years (2007 and 2008) under different nitrogen application amounts. All average soil moistures from 0 to 40 cm depth in whole growth period were maintained at above 50% available soil moisture except SDI N75 under SDI and DI irrigation method.

Root distribution

Table 5 shows a significant difference in RLD between SDI and DI and among the different nitrogen levels. RLD obviously increased with increasing nitrogen levels until the nitrogen level reached 150 kg/ha in both years, and then it sharply decreased. The effect of irrigation methods on RLD was apparent; the RLD of SDI was obviously higher than that of DI.

Table 5 Root length density (RLD) of bell pepper as influenced by irrigation method and N application levels
Full size table

At the same fertilization level, the root length and percentage of root length in each layer to total root length decreased with soil depths (Table 6). The percentages of root length at 30–40-cm soil depth to total root length under DI and SDI were 1.25% (DI N150) and 2.81% (SDI N150), respectively, indicating that there were almost no bell pepper roots below 40-cm soil depth. Root lengths under SDI N0, SDI N75, SDI N150, and SDI N300 were 1.06, 1.06, 1.46, and 1.07 times longer than those under DI N0, DI N75, DI N150, and DI N300, respectively. The percentage of root length below 10-cm soil depth under SDI N150 was higher than that under DI N150 by 7%. Hence, SDI does not only promote root growth but also results in deeper root development.

Table 6 Bell pepper root length at different soil depths
Full size table

Under the same irrigation methods, the impact of different nitrogen levels on RLD at harvest is shown in Fig. 6. At 0–10-cm soil depth, RLD gradually increased with increasing nitrogen levels. However, at 10–20-cm soil depth, RLD declined sharply when the nitrogen level exceeded 150 kg/ha. These findings imply that too much nitrogen application inhibits root growth in deeper soil layers.

Fig. 6
figure 6

RLD distribution during the 2007 growing season in all treatments

Full size image

NO3 −-N distribution in soils

Figure 7 shows NO3 −-N concentrations in soil 2 days before fertilization (August 14), 2 days after fertilization (August 18), and 22 days after fertilization (September 7).

Fig. 7
figure 7

Vertical distribution of NO3 −-N concentration in soil profiles

Full size image

Before fertilization, there was no significant difference in NO3 −-N distribution between SDI and DI. However, 2 days after fertilization, NO3 −-N concentrations under SDI treatment were distributed with a parabolic curve; the maximum value (14.2 mg/kg) was found at 20–40-cm soil depth. In contrast, NO3 −-N concentrations under DI treatment declined with increasing soil depth, and a maximum concentration (15.7 mg/kg) was obtained at the top soil (0–20 cm). Furthermore, 22 days after the fertilization, NO3 −-N gradually moved downward due to water movement, crop growth, and root activities. The maximum NO3 −-N concentration 22 days after fertilization under SDI and DI occurred at 40–60 cm and 60–80 cm, respectively.

As mentioned earlier, bell pepper roots were concentrated at 0–40-cm soil depth. Nitrogen leaching below 40 cm, which was hardly useful to the plants, was of residual value. The maximum residual NO3 −-N concentration at 40–60 cm under SDI (8.4 mg/kg) was far lower than that under DI treatment (13.8 mg/kg at 60–80 cm).

Residual NO3 −-N concentrations in soil profiles increased with increasing levels of nitrogen fertilizers (Fig. 8), but the residual of N150 was only slightly higher than N75. The NO3 −-N residual concentration for N300 treatment was sharply higher than that for N150 treatment 22 days after fertilization. This tend was found in all nitrogen treatments.

Fig. 8
figure 8

Vertical distribution of NO3 −-N as influenced by different nitrogen levels

Full size image

The data above show that SDI promoted the development of bell pepper roots and favored the establishment of intensive root layers, which can prevent nitrate leaching. At nitrogen level lower than 150 kg/ha, fertigation produced lesser residues.

ET c

Bell pepper plants were transplanted at May 20 in 2007 and May 21 in 2008. The growth seasons last 118 days and 115 days separately in 2007 and 2008.

Bell pepper ET c was calculated by formula (1). D in formula (1), the amounts of drainage water collected from the lysimeter are shown in Table 7.

Table 7 Drainage water amounts (mm)
Full size table

Table 8 shows the cumulative water consumption of bell pepper and ET 0 calculated by Penman–Monteith’s formula during the two growing seasons. The ET c and ET 0 of all treatments in 2008 were lower than those in 2007 due to the lower monthly average temperature in 2008 as compared with 2007. A very low temperature (2.4°C) in May 30, 2008, inhibited seedling establishment.

Table 8 Cumulative water consumption under different irrigation and fertilization practices
Full size table

The maximum and minimum water consumption were recorded for DI N150 treatment (451 mm) and SDI N0 treatment (301 mm), respectively, in 2007. In 2008, the maximum water consumption was recorded for DI N75 (387 mm), followed by DI N150 and DI N300 (382 mm). The minimum value was at 334 mm for SDI N0. Except for N300 treatment in 2007, all cumulative water consumptions under SDI were lower than under DI.

Daily average water consumptions at different growth stages under different irrigation techniques (N150 treatment) are shown in Fig. 9. During the seedling establishment period, the DI method resulted in higher daily averaged water consumption compared with SDI because of higher evaporation under DI. After entering the blossom and fruit-set period, the daily average water consumption under SDI became higher than that under DI. This result may be attributed to faster root growth under SDI than under DI. At full bearing period, the plants grew vigorously, leading water consumption to reach its maximum. Water consumption under SDI was lower than that under DI, contributing to low plant height and leaf area (Fig. 10). However, daily average water consumption under DI was slightly lower than that under SDI at the late crop growth stages.

Fig. 9
figure 9

Daily averaged water consumption at different growth stages

Full size image
Fig. 10
figure 10

Plant height and LAI for different treatments (2007)

Full size image

There was a polynomial correlation between crop water consumption and nitrogen levels (Fig. 11). ET c increased with increasing nitrogen levels, reaching a maximum value at 150 kg/ha nitrogen level. Thereafter, ET c again declined. Nitrogen became excessive after 150 kg/ha, and too much nitrogen restricts bell pepper growth leading to lower ET c.

Fig. 11
figure 11

Relationship between ET c and nitrogen levels

Full size image

Crop coefficient k c for bell pepper at different growth stages for the experimental site under SDI N150 is shown in Table 9. At the seedling establishment and full bearing periods, k c values were higher than the values recommended by FAO-56. On the contrary, k c at the blossom and fruit-set period was higher than that recommended by FAO-56. Meanwhile, k c at the late crop growth stages was not stable.

Table 9 Calculated kc at each growth stage
Full size table

Yield and water-use efficiency

Bell pepper yields were measured for each treatment in 2007 and 2008. Results are shown in Table 10.

Table 10 Bell pepper yield and WUE for different treatments
Full size table

Yield under SDI was significantly higher than that under DI in both years. It was higher by 4 and 13% in 2007 and 2008, respectively. The maximum yield was obtained under SDI at 150 kg/ha (SDI N150). SDI had a higher WUE than DI by 13 and 21% in 2007 and 2008, respectively. Maximum WUEs were obtained under SDI without nitrogen supply (SDI N0) in 2007 and under SDI with 150 kg/ha (SDI N150) in 2008.

Standard ANOVA was carried out with Duncan’s test at the 0.05 level of significance. The level of fertilizer application showed a significant effect on bell pepper yield. The relationship between yield and the level of fertilization nitrogen was conical, the yield increased with urea fertilization up to a point (150–200 kg/ha N) when fertilization became excessive.

The variance analysis indicated that experimental years (Y), irrigation methods (I), and nitrogen application amounts (N) significantly influenced pepper yields (Table 11). There was no significant interaction effect among Y, I, and N, except between Y × I.

Table 11 Variance analysis of the yield of bell pepper
Full size table

There was a significantly polynomial correlation between bell pepper yield and cumulative water consumption (Fig. 12). Bell pepper yield was improved when water consumption increased at a certain range.

Fig. 12
figure 12

Relationship between yield and ET c

Full size image

Conclusions

Bell pepper yield under SDI was significantly higher than that under DI in both years. The highest yield (46.54 t/ha) was recorded in the case of SDI at 150 kg/ha (SDI N150). The lowest yield (28.11 t/ha) was recorded in the case of DI with no after-manuring (SDI N0).

Compared with DI, all treatments of SDI had lower water consumption. The total root length densities from 0–40-cm soil depth and the percentage of root length below 10 cm under SDI are higher than under DI. SDI does not only promote crop root growth; it also enhances the downward development of roots. There are lesser soil N residues under SDI than under DI. Consequently, the experiment obtained a preliminary conclusion that SDI was more beneficial to the growth of bell pepper than DI.

From the perspective of increasing bell pepper yield and WUE, as well as reducing NO3 −-N leaching, SDI with N application of 150 kg/ha is the optimal fertigation practice.