Effects of topdressing with nitrogen fertilizer on wheat yield, and nitrogen uptake and utilization efficiency on the Loess Plateau

Abstract

Studies from various countries show that topdressing with nitrogen fertilizer can effectively increase winter wheat grain yields. However, information on its effects on wheat yields on the Loess Plateau in China is scarce. Thus, we investigated yields, N concentration and uptake, nitrogen use efficiency (NUE), and partial factor productivity of nitrogen (PFPN) of winter wheat at a site on the Plateau in three consecutive years following four treatments: no N application; 150 kg N ha⁻¹ and 46 kg P ha⁻¹ applied as base dressing; 198 kg N ha⁻¹ and 46 kg P ha⁻¹ applied as base dressing; and 150 kg N ha⁻¹ and 46 kg P ha⁻¹ applied as base dressing + 48 kg N ha⁻¹ applied as topdressing in the spring. Topdressing with N fertilizer resulted in significant increases in wheat yield, wheat biomass production, and N concentrations and uptake in wheat grain, straw, and chaff. PFPN was 6.9%, 15.2%, and 5.5% higher, and NUE 28.1%, 31.6%, and 12.5% higher, in the three years with top dressing than with base dressing alone. Topdressing N fertilizer for winter wheat is recommended on the Loess Plateau.

Keywords:

• winter wheat
• topdressing N fertilizer
• grain yield
• N concentration
• PFPN
• NUE

Introduction

Wheat (Triticum aestivum L.) is the third most commonly cultured grain crop in the world and a major source of food and feed, grown on over 200 million hectares (ha) of land annually (Rajaram & Braun 2008). On the Loess Plateau in China, winter wheat is intensively cultivated, as a monoculture by most farmers, and covers 56% of the arable land, about 4.3 million ha annually (Zhu 1989; Li et al. 2014).

Topdressing with nitrogen (N) fertilizer has been shown to increase winter wheat yields in various countries. Research from Northwest Europe and the USA shows that additional topdressing of wheat with N fertilizer during growth, as opposed to applying N solely during sowing or the fall, improves grain yields (Wuest & Cassman 1992; Sowers et al. 1994; Recous & Machet 1999; Tran & Tremblay 2000). The practice also increases grain yields of durum wheat cultivated under rainy Mediterranean conditions (López-Bellido et al. 2005, 2006). Under these conditions, topdressing with half or one-third of total N fertilizer during stem elongation reportedly results in higher grain yields than applying fertilizer during sowing only, or during both sowing and tillering (Garrido-Lestache et al. 2004; Marino et al. 2011).

Other studies suggest that, in both northern and southern parts of China, topdressing winter wheat with N fertilizer and irrigation could substantially increase grain yields (Cui et al. 2008; Shi et al. 2012). However, many agronomists believe that in parts of the Loess Plateau with a long dry season, for rain-fed crops that cannot be irrigated N should be applied as base fertilizer, hence early application has been encouraged for winter wheat cultivation (Lu 1983; Li and Zhao 1993). Thus, farmers have applied N fertilizers as a base dressing before planting wheat, following practices recommended since the start of the 1980s.

Given the clear evidence that topdressing with N fertilizer can increase grain yields in various countries and climatic zones, further information on effects of the practice on grain yields in parts of the Loess Plateau in China that are not irrigated is required. Thus, the objectives of this study were to increase the scarce information on effects of topdressing with N fertilizer, without irrigation, on N concentrations and uptake in wheat grain, straw, and chaff, N use efficiency (NUE), and partial factor productivity of applied N (PFPN) in a field trial at a dryland site on the Plateau.

Materials and methods

Site description

The field trial was conducted at a site (latitude 35.1°N, longitude 107.6°E, elevation 1215 m above sea level) in the dryland winter wheat growing region of the Loess Plateau in Changwu County, China. The average annual precipitation at Changwu is 578 mm, and the annual average temperature is 9.3°C. Rainfall during the winter wheat growing season amounted to 186 mm in 2009–2010, 199 mm in 2010–2011, and 234 mm in 2011–2012, wheat harvest times in 2009–2010, 2010–2011, and 2011–2012 amounted to 435, 656, and 622 mm, respectively. Thus, precipitation during the wheat growing season accounted for 43%, 30%, and 38% of annual precipitation, respectively. The water table depth is more than 60 m, thus groundwater is unavailable for plant growth. The soil at the study site has a silt loam texture according to the US Department of Agriculture texture classification system. In China, the soil is referred to as a Heilu soil, which corresponds to a Calcarid Regosol according to the United Nations Food and Agriculture and Educational, Scientific and Cultural Organizations (FAO/UNESCO) classification system (FAO/UNESCO 1988). At the start of the experiment the bulk density and pH of the soil were 1.21 g cm⁻³ and 8.1, respectively, field moisture capacity was 20.2%, and the soil contained 14.7 g organic matter, 16.7 mg total N, 3.3 mg nitrate-N, 0.77 g ammonium-N, 14.5 mg available phosphorous (P), and 129.7 mg available potassium per kg.

Experimental design

The field trial was an in-situ experiment: four fertilizer treatments were applied in 2009–2010, 2010–2011, and 2011–2012: (1) 46 kg P ha⁻¹ only, applied as base dressing – with no N application (N₀); (2) 150 kg N ha⁻¹ and 46 kg P ha⁻¹ applied as base dressing, (N_150b); (3) 198 kg N ha⁻¹ and 46 kg P ha⁻¹ applied as base dressing (N_198b); and (4) 150 kg N ha⁻¹ and 46 kg P ha⁻¹ applied as base dressing + 48 kg N ha⁻¹ applied as topdressing in the spring (N_150b+48t). Urea was used as N fertilizer, and ordinary superphosphate as phosphate fertilizer. The experiment was conducted according to a completely randomized block design. Each plot was 8 m long and 8 m wide. The basal dressing treatments were applied and winter wheat (cultivar Changhan 58, which is drought-tolerant, suitable for planting at about 1200 m above sea level in the rain-fed area of the Loess Plateau, and extensively planted by local farmers in Changwu county) seeds were sown on 25 September, 22 September, and 23 September, then the wheat was harvested on 28 June, 27 June, and 27 June in the 2009–2010, 2010–2011, and 2011–2012 growing seasons, respectively. In each case the seeding rate was 150 kg ha⁻¹, and N topdressing was applied on 24 February.

Sampling and measurements

At each harvest a 3 m² (3 m×1 m) sub-plot in the middle of each plot was harvested manually to determine yields and biomass. Straw (stem and leaves), chaff, and grain were determined after oven-drying to constant weight (60°C for 24 h). These subsamples of the grain, chaff, and straw were taken back to the laboratory to determine their N contents using the Kjeldahl procedure.

Calculation

Winter wheat Harvest Index (HI) was calculated as:

(1)

Winter wheat N uptake was calculated by multiplying the N content of each plant component (straw, chaff, and grain) by its corresponding dry mass. Total N uptake was calculated by summing the N uptake of all the wheat components. NUE was determined as:

(2)

where U_N is the total N uptake by wheat (straw, chaff, and grain) with N application (kg N ha⁻¹), U₀ is the total N uptake by wheat without N application, and N_F is the total amount of N fertilizer applied (kg ha⁻¹).

PFPN was calculated as:

(3)

where Y is the wheat grain yield and N is the N fertilization rate.

Statistical analyses

The effects of the treatments on the measured variables were evaluated using one-way analysis of variance. When F values were significant, the least significant difference (LSD) test was used to compare means. In all cases, differences were deemed to be significant if P ≤ 0.05.

Results

Wheat yield responses to topdressing with N fertilizer

Wheat grain yield and biomass varied between the three experimental seasons, with the highest yields recorded in 2011/2012 and the lowest in 2009/2010 under the three N fertilizer application treatments (Table 1). N fertilizer application significantly increased wheat yield and biomass during all three experimental seasons. The N_150b+48t treatment resulted in significantly greater grain yield and wheat biomass than the N_150b and N_198b treatments during the 2009/2010 and 2010/2011 growing seasons. During the 2011/2012 season, yield and biomass were significantly greater under the N_150b+48t treatment than under the N_150b treatment, but the yield and biomass under these treatments did not differ significantly from values under the N_198b treatment. Furthermore, both yield and biomass values were somewhat greater under the N_198b treatment than under N_150b, but there were no significant differences in either yield or biomass of N_198b and N_150b in any of the three experimental seasons. As shown in Table 1, the yield and biomass values under the N₀ treatment decreased from the first experimental year to the third.

Table 1. Effects of N application on the biomass and grain yield of winter wheat.

	Grain yield (t ha^−1)			Biomass (t ha^−1)			HI
Treatment	2009/2010	2010/2011	2011/2012	2009/2010	2010/2011	2011/2012	2009/2010	2010/2011	2011/2012
N0	2.94c	2.47c	2.20c	6.52c	5.92c	5.57c	0.45a	0.42a	0.39b
N150b	3.98b	4.01b	6.27b	9.16b	9.47b	12.77b	0.43a	0.42a	0.49a
N198b	4.04b	4.15b	6.48ab	9.48b	9.51b	13.26ab	0.43a	0.44a	0.49a
N150b + 48t	4.32a	4.80a	6.83a	10.68a	10.46a	14.01a	0.40a	0.46a	0.49a

Note: Different letters following the numbers in each column indicate significant (P ≤ 0.05) differences between treatments according to LSD tests.

N₀, 46 kg P ha⁻¹ only, applied as base dressing with no N application; N_150b, 150 kg N ha⁻¹ and 46 kg P ha⁻¹ applied as base dressing; N_198b, 198 kg N ha⁻¹ and 46 kg P ha⁻¹ applied as base dressing; N_{150b + 48t}, 150 kg N ha⁻¹ and 46 kg P ha⁻¹ applied as base dressing + 48 kg N ha⁻¹ applied as topdressing in the spring; HI, Harvest Index (grain yield/biomass).

As also shown in Table 1, the average HI in the three experimental years was 0.44. Under the N₀ treatment HI decreased from the first experimental year to the third. N fertilizer application increased HI from the first experimental year to the third, except in the 2010/2011 growing season under treatment N_150b. There were no significant between-treatment differences in HI in 2009/2010 and 2010/2011, but HI was significantly high under the N_{150b + 48t}, N_198b, and N_150b treatments than under N₀ during the 2011/2012 growing season (Table 1).

N concentration and uptake in wheat grain responses to topdressing with N fertilizer

The N concentration in wheat grain varied between the three experimental seasons, being highest in 2009/2010 and lowest in 2011/2012 (Table 2). The N concentrations of grain, straw, and chaff were significantly higher under the N fertilizer treatments than under the N-free treatment in all three experimental seasons. In addition, the concentrations increased in the order N_150b < N_198b < N_{150b + 48t}, but there were no significant differences in the N concentrations of grain, straw, or chaff between the three treatments including N applications for the 2009/2010 and 2010/2011 season. In the 2011/2012 seasons, grain N concentrations under the N_150b+48t and N_198b treatments were similar, and in both cases were significantly higher than those resulting from the N_150b treatment. Furthermore, the N concentrations in grain, straw, and chaff under the N₀ treatment all decreased from the first to the third experimental year.

Table 2. Wheat N concentration and uptake.

		Grain			Straw			Chaff
	Treatment	2009/2010	2010/2011	2011/2012	2009/2010	2011/2012	2011/2012	2009/2010	2010/2011	2011/2012
Concentration (g kg^−1)	N0	16.3b	16.1b	12.5c	3.3b	2.9b	2.3b	3.8b	3.5b	3.3b
	N150b	22.1a	20.3a	15.6b	5.1a	5.0a	4.8a	5.8a	5.6a	5.3a
	N198b	22.8a	20.7a	18.0a	5.3a	5.0a	5.1a	6.2a	5.9a	5.5a
	N150b + 48t	23.8a	21.5a	18.8a	5.5a	5.4a	5.2a	6.1a	6.0a	5.6a
Uptake (kg ha^−1)	N0	47.9c	39.8c	27.5d	8.7c	7.3c	5.7d	3.6c	3.2c	3.0c
	N150b	88.0b	81.4b	97.8c	20.6b	18.9b	22.9c	8.5b	7.7b	9.2b
	N198b	92.1b	85.9b	116.6b	21.3b	19.6b	25.4b	9.0b	8.4ab	9.9ab
	N150b + 48t	102.8a	103.2a	128.4a	24.8a	23.3a	27.4a	10.0a	9.4a	10.7a

Note: Different letters following the numbers in each column indicate significant (P ≤ 0.05) differences between treatments according to LSD tests.

N₀, 46 kg P ha⁻¹ only, applied as base dressing with no N application; N_150b, 150 kg N ha⁻¹ and 46 kg P ha⁻¹ applied as base dressing; N_198b, 198 kg N ha⁻¹ and 46 kg P ha⁻¹ applied as base dressing; N_{150b + 48t}, 150 kg N ha⁻¹ and 46 kg P ha⁻¹ applied as base dressing + 48 kg N ha⁻¹ applied as topdressing in the spring.

Application of N fertilizer also significantly increased grain, straw, and chaff N uptake in all three experimental years (Table 2). The N_150b and N_198b treatments resulted in similar grain and straw N uptake values, which were significantly lower than the values under the N_{150b + 48t} treatment in 2009/2010 and 2010/2011. In 2011/2012, the grain and straw N uptake values significantly increased in the order N_150b < N_198b < N_{150b + 48t}. Chaff N uptake was similar under the N_150b and N_198b treatments, and significantly higher under the N_150b+48t treatment in 2009/2010. In 2010/2011 and 2011/2012, this variable was significantly higher under the N_{150b + 48t} treatment than under N_150b, but values for N_198b did not significantly differ from the N_150b and N_{150b + 48t} values. In addition, N uptake values for grain, straw, and chaff under the N₀ treatment decreased from the first to the third experimental year (Table 2).

PFPN and NUE response to N fertilizers topdressing

PFPN and NUE values varied amongst the three experimental years, being highest in 2011/2012 and lowest in 2009/2010 (Table 3). Among the N treatments, both PFPN and NUE values increased in the order N_198b < N_{150b + 48t} < N_150b. PFPN values were 6.9%, 15.2%, and 5.5% higher under the N_150b+48t treatment than under the N_198b treatment during 2009/2010, 2010/2011, and 2011/2012, respectively, while NUE values were 28.1%, 31.6%, and 12.5% higher, respectively.

Table 3. PFPN and NUE values under the N fertilization treatments.

	PFPN (kg (kg of N)^−1)			NUE (%)
Treatment	2009/2010	2010/2011	2011/2012	2009/2010	2010/2011	2011/2012
N0	–	–	–	–	–	–
N150b	26.5	26.7	41.8	37.8	38.4	62.5
N198b	20.4	21.0	32.7	31.4	32.1	58.5
N150b + 48t	21.8	24.2	34.5	39.1	43.2	65.8

N₀, 46 kg P ha⁻¹ only, applied as base dressing with no N application; N_150b, 150 kg N ha⁻¹ and 46 kg P ha⁻¹ applied as base dressing; N_198b, 198 kg N ha⁻¹ and 46 kg P ha⁻¹ applied as base dressing; N_{150b + 48t}, 150 kg N ha⁻¹ and 46 kg P ha⁻¹ applied as base dressing + 48 kg N ha⁻¹ applied as topdressing in the spring; PFPN, partial factor productivity of nitrogen; and NUE, nitrogen use efficiency.

Discussion

Wheat grain yield response to precipitation

The highest yields of wheat grain were recorded in 2011/2012, and the lowest in 2009/2010 (Table 1), possibly due to variations in precipitation during the growing season in the experimental area, which were also highest during 2011/2012 and lowest during 2009/2010 (as shown by the data in Materials and methods section). These results are consistent with the positive correlation between wheat yields and precipitation during the nine-month (1 October–30 June) wheat growing seasons from 1945 to 1997 at Pendleton (OR) in the American Pacific Northwest reported by Camara et al. (2003).

Wheat HI response to topdressing N fertilizer

According to Fischer (2007), the highest reported HI values for winter wheat are ca. 0.5, and an effective strategy to increase yields is likely to include measures to increase biomass, while maintaining HI as high as possible, by rapidly determining HI and culling progeny with HI lower than 0.45 (Fischer 2011). The average HI in this field trial over three consecutive seasons was 0.44, just below 0.45, but our field experiment showed that moving some of the N fertilization from base application to topdressing could increase grain yields, and thus HI (Table 1). Therefore, we recommend topdressing N fertilization for winter wheat farmers on the Loess Plateau.

Wheat grain yield response to topdressing N fertilizer

Wuest and Cassman (1992) found that adding 45 kg ha⁻¹ N fertilizer late in the growing season, together with 120 kg ha⁻¹ N as a basal application increased spring wheat yields at a site in the Central Valley of California by 12.3%. Similarly, in a field trial at a site in a typical Mediterranean region López-Bellido et al. (2005) found that wheat yields were lower when 150 kg N ha⁻¹ was all applied at sowing, rather than in split applications, and that both the timing and splits significantly affected yields over a three-year period. Yields were highest when a third or half of the total N fertilizer rate (150 kg N ha⁻¹) was applied (as topdressing) at stem elongation. Similarly, Cui et al. (2008) found that a basal application of 65 kg N ha⁻¹ and a further 135 kg N ha⁻¹ at shooting stages significantly increased wheat yield, relative to a basal application of 200 kg N ha⁻¹ in field experiments at 22 irrigated sites in northern China. Shi et al. (2012) also observed increases in wheat yield when applying half of N fertilizer application as topdressing, rather than all as base fertilizer, at irrigated sites in southern China.

In accordance with the findings from the cited experiments, the results from the field trial reported here show that topdressing with N, or applying some of the N fertilizer used for base dressing in topdressing, can significantly increase wheat grain yields (Table 1). In contrast, the results conflict with the common belief that N fertilizer should be applied solely as base dressing for rainfed winter wheat on the Loess Plateau in China, mainly because nitrate-N is not mobilized by rainfall during the long, dry wheat growing season. However, during the approximately five months between wheat sowing and topdressing at the selected date in this field trial, some N will be lost through ammonia volatilization, denitrification, and immobilization by soil microbes. Losses through volatilization are likely to be particularly severe if urea is used as the N source because most soils of the Plateau are calcareous and have high pH, which strongly enhances volatilization (Christianson et al. 1990; Aggarwal & Praveenkumar 1994; Roelcke et al. 2002). Accordingly, Zhang et al. (1992) found that 30–32% of the N applied in fertilizers may be lost through ammonia volatilization in this region. A further 10–30% may be subject to denitrification processes and potentially lost (Williams et al. 1992), and 24–27% could be immobilized by soil microbes (Shen et al. 1989). Furthermore, on the Loess Plateau the soil organic matter (SOM) content is low, with approximately 10 g kg⁻¹ SOM in most of the soils, and extremely low contents in some areas (Nanjing Institute of Soil Science 1978). Thus, there are severe N deficiencies, wheat crops do not get sufficient N nutrients from the re-greening to the grain-filling stages, and grain yields are low. The N topdressing was applied on 24 February in every year during this field trial. This date was selected because it is shortly before the wheat re-greening stage, when frozen soils start melting, and the N fertilizer can potentially dissolve into the gradually melting soil water. If so, it could supply the N nutrients required for wheat growth from the re-greening to the grain-filling stages, thus resulting in higher yields than would occur if N fertilizer was applied only as a base dressing.

N concentration response to wheat grain yield

In this study, the wheat grain yields and grain N concentrations showed opposite temporal trends over the three experimental years: the grain yields were highest and N concentrations lowest in 2011/2012, while yields were lowest and N concentrations highest in 2009/2010 (Tables 1 and 2). This dilution effect is consistent with previous observations of negative correlations between cereal grain yields and grain N concentrations. For example, Wang and Yu (2008) observed an increase in grain yield from 5534 to 8706 kg ha⁻¹ with a reduction in grain N concentration from 29.4 to 23.8 g kg⁻¹ in dryland winter wheat. Similarly, an increase in maize yield potential from 1930 to 1990 in the north-central USA was reportedly accompanied by a reduction in maize grain N concentration (Duvick & Cassman1999). In addition, Hou et al. (2012) found that while the grain yield of maize cultivars released from the 1950s to the 2000s on the north China plain increased chronologically from 6.7 to 11.6 t ha⁻¹, the grain N concentration strongly decreased, from 19.2 to 14.6 g kg⁻¹. Dilution effects in maize grain N concentrations associated with rapid increases in yields have also been detected in other studies (Rajcan & Tollenaar 1999a, 1999b).

N concentration and N uptake response to topdressing N fertilizer

Topdressing with N, or transferring some of the N fertilizer designated to base dressing for use in topdressing, tended to increase N concentrations of grain, straw, and chaff (Table 2). Similarly, Ottman et al. (2000) found that late application of N fertilizer increases the N concentration of wheat grain more effectively than application early in the season. They found that nitrogen applications near anthesis of 0, 3.4, and 6.7 g N m⁻², respectively, resulted in grain N contents of 20.2, 22.3, and 24.6 g kg^–1 in 1995 and 23.2, 24.7, and 26.5 g kg^–1 in 1996. Other studies have also shown that applying N later in the spring is more effective than earlier application for enhancing the N concentration in wheat grain. For example, Brown and Petrie (2006) investigated topdressing N effects of three seasonal treatments on winter wheat (incorporation of 168 kg N ha^–1 in the fall before planting, topdressing 168 kg N ha^–1 early in the spring, and a combination of both treatments). They found spring topdressing N resulted in 1.4–2.3 g kg^–1 higher grain N concentrations than fall application. However, preplanting and topdressing N provided the highest grain N concentration (21.1 g kg^–1). In conclusion, topdressing with N, or transferring some N fertilizer designated to base dressing for use in topdressing, significantly increases N uptake in wheat grain; increasing both grain yields and N grain concentrations.

PFPN and NUE response to topdressing N fertilizer

NUE can also be expressed as the PFPN. The PFPN is a useful index because it does not require measurements of the grain yield in unfertilized plots or N uptake of either grain or plants. In the UK, the average PFPN for winter wheat reportedly increased from 36 kg (kg of N^–1) in 1986/1990 (186 kg of N ha⁻¹ and 6.6 Mg grain⁻¹) to 42 kg (kg of N^–1) by 2006/2010 (186 kg of N ha⁻¹ and 7.8 Mg of grain ha⁻¹) (Fertilizer Use on Farm Crops for Crop Year 2010. 2011). In China, average N application rates have exceeded crop N uptake, resulting in a large N surplus. Consequently, the average PFPN for wheat is 43 kg (kg of N^–1), while for cereal cropping systems the PFPN is only 25−37 kg (kg of N^–1), which is below the world average: 44−72 kg (kg of N^–1)(Zhang et al. 2008). As shown by data in Table 3, the average PFPN in this field trial over three consecutive seasons was 27.7 kg (kg of N^–1), which is lower than the Chinese national average for wheat, and also below the world average for cereal crops. A recent review of worldwide data on NUE for cereal crops from researcher-managed experimental plots reported that, in a single season, N recovery efficiencies from crop fertilizer averaged 57% for wheat (Ladha et al. 2005). The average NUE in this field trial over three consecutive seasons was 45.4% (Table 3), which is clearly lower than the average researcher-managed experimental value. However, our trial was conducted under rainfed conditions on the northwestern drylands in China, where winter wheat production is generally hindered by drought and unfertile soils. Consequently, wheat yields cannot be as high as in other parts of China or the world. Nevertheless, the results indicate that the yield, PFPN, and NUE of wheat grain can be significantly improved by applying some of the N fertilizer typically applied solely as base dressing in topdressing.

Disclosure statement

No potential conflict of interest was reported by the authors.

Funding

This research was financially supported by the Natural Science Foundation of China [contract number 31272250], Modern Agricultural Technical System (MATS), and Special Fund for Agro-scientific Research in the Public Interest [contract number 201103003].