1. Introduction
Nitrogen (N) is a key element to ensure maximum crop yields and food security [
1,
2,
3]. All around the world, the use of N fertilizers has been estimated to be ~155 million tons and has boosted crop and food production [
4]. Nitrogen is the nutrient crops require most [
5] and soil net N mineralization is a critical rate for crop development. This process is mainly controlled by the climate and physical and/or biochemical properties [
6]. Nitrogen is taken up by plants preferentially in mineral forms (ammonium and nitrate). The absorbed form of N depends on the crop, although these N forms represent only 2% of the total N in the soil [
7].
On average, about 50% of N fertilizer is recovered by plants and is susceptible to leaching losses, gaseous emissions and surface runoff [
8]. The N fertilizer applied to maize-forage intercropping can be recovered by the forage and maintained in the system. After harvesting, more than 60% of N remains in the system, partially on forage grass and maize residues but mainly on soil. About 12% of the N residual is recovered by soybeans in succession [
9].
Therefore, to improve the sustainability of agricultural systems, the dynamics of this nutrient in the systems should be investigated, and the contribution of cover crops to soil mineral N and the N fertilizer use efficiency (NFUE) of a subsequent cash crop should be considered to define the effective need for nitrogen fertilization [
10,
11]. In this context, the decomposition process of cover crop residues is fundamental because it regulates N mineralization in the soil and should be taken into account in N fertilizer recommendations.
Nitrogen is a nutrient with highly complex dynamics in the soil-plant-atmosphere system [
12] and is directly associated with carbon (C) accumulation in the soil [
13]. Knowledge of these mineral N dynamics in the soil can facilitate decision-making regarding the most reliable, efficient and sustainable use of crop rotation [
14,
15] and fertilization [
7], and put an end to indiscriminate fertilizer use without concern about avoiding environmental and economic losses.
About half of the N fertilizer applied will become available for plant uptake [
2], while the other part will be in the environment [
16,
17]. In maize, 30–60% of N applied in the form of fertilizer can be recovered by the plants [
5], whereas the rest may be lost by leaching in the form of nitrate, ammonia volatilization or even in much smaller quantities during nitrification, especially in the form of N
2O [
18,
19,
20].
The process of soil degradation and N losses can be minimized by certain techniques, such as no-tillage management with crop rotation and practices such as intercropping. These measures increase the availability of mineral N in relation to monoculture and plowed farming systems [
21,
22], and, consequently, the efficiency of N fertilizer use [
14].
In tropical regions, especially in South America, management practices such as the use of intercropped systems and crop–pasture rotation, including forage or legume species with high biological dinitrogen fixation efficiency as cover crops, are viable alternatives for the recovery of degraded lands and to improve soil carbon accumulation [
23]. In Brazil, the double cropping system under no-tillage is mainly based on soybean as the main crop, followed by maize or sorghum as the second commercial crop [
9]. In regions with lower rainfall availability, gramineous species are commonly used as cover crops, such as pearl millet (
Pennisetum glaucum), Brachiaria (
Urochoa sp.) or
Panicum sp., normally from March to mid-June [
24].
The literature reveals that there is still a lack of information on the use of leguminous species in agricultural systems, especially in the modality of crop succession, intercropping or as a companion crop [
25,
26,
27]. In no-tillage systems integrated with leguminous species, the recovery efficiency of N-fertilizer applied by annual crops may be higher than that of monocropping or using gramineous species as cover crops. In no-tillage systems combining a more diverse set of cover crop species, the N available in deeper soil layers, which has the potential to be leached, can be absorbed by the roots of the plants before the roots of the annual crops reach this depth. This interaction between species in the integrated system increases nutrient recycling efficiency, especially N. In general, the transfer of N is greater from legume species to non-legumes than the other way around [
28].
In this context, we believe that cover crops in no-tillage systems alter soil mineral N levels and influence the nitrogen fertilizer use efficiency (NFUE) of subsequent maize. Thus, we hypothesized that no-tillage systems with different cover crops with different chemical compositions affect the availability of mineral N (nitrate and ammonium) and NFUE of maize in the Brazilian Cerrado differently. To test this hypothesis, ammonium (NH4+) and nitrate (NO3−) concentrations and NFUE were evaluated in soil under different cover crops grown after maize fertilized with N (WN) and without N (NN) topdressing at the end of the rainy season and the beginning of the following season. The chemical composition of the cover plants was also assessed to help explain N mineralization in this production system.
2. Materials and Methods
2.1. Location and Experimental Design
The experiment was carried out at the Cerrados unit of the Brazilian Agricultural Research Corporation (Embrapa), in Planaltina, Distrito Federal, Brazil (15°35′30″ S, 47°42′30″ W). According to the Köppen–Geiger classification, the regional climate is Aw (tropical savanna), with dry winters and rainy summers, a mean annual air temperature between 22 and 27 °C, and a mean annual rainfall of 1345.8 mm [
28]. Rainfall and mean air temperature from the end of one (6 April 2016) to the beginning of the following rainy season (16 November 2016), when the two soil samplings were taken, are shown in
Figure 1. The mean temperature from April to November was 22.2 °C (±2.3), and precipitation from April to September was 65.4 mm (dry season), and in October and November it was 255 mm (part of the rainy season).
The soil was classified as Typical Acrustox and before the experiment, the chemical and physical soil properties (0–20 cm layer) were determined as follows: pH (H2O) 6.0; organic matter 21.7 g kg−1; P (Mehlich-1) 0.9 mg kg−1; Al3+ 0.1 cmolc kg−1; Ca2+ + Mg2+ 2.9 cmolc kg−1; K+ 0.1 cmolc kg−1; fine sand 258 g kg−1; coarse sand 76.7 g kg−1; silt 101.8 g kg−1; and clay 563.5 g kg−1.
From 1999 to 2004, alternating soybeans and maize were grown in the area. In the growing seasons from 2004/2005 to 2015/2016, a repeated sequence of maize and cover crops was cultivated under no-tillage management. Every year, from 2004/2005 to 2015/2016, after harvesting the maize hybrid 30F53VYHR (February), the following cover species were sown: pigeon pea (Cajanus cajan (L.) Millsp), sunn hemp (Crotalaria juncea L.), black Mucuna (Mucuna aterrima Merr.) and oilseed radish (Raphanus sativus L.).
The experiment was arranged in a randomized block split plot design with three replications. The plots (12 × 8 m) consisted of four cover crops, and the subplots (12 × 4 m) of maize fertilized with N topdressing (WN) or without N topdressing (NN). Two topdressings of 65 kg N ha
−1 were applied as urea in WN, corresponding to 130 kg total N ha
−1 in topdressings plus 20 kg N ha
−1 at sowing, for a total of 150 kg N ha
−1, as recommended by Sousa and Lobato [
29]. In addition, maintenance fertilization was applied in the planting furrow at maize sowing, consisting of 150 kg ha
−1 P
2O
5, 80 kg ha
−1 K
2O, 2 kg ha
−1 Zn (ZnSO
4·7H
2O) and 10 kg ha
−1 FTE BR 12 as a micronutrient source (3.2% S, 1.8% B, 0.8% Cu, 2.0% Mn, 0.1% Mo, 9.0% Zn and 1.8% Ca).
In April 2016, the cover crops Cajanus cajan and Crotalaria juncea were sown at a density of 20 plants m−1, M. aterrima at 10 plants m−1 and R. sativus at 40 plants m−1. The row spacing for all cover crops was 0.5 m. Cover crops were cut at flowering (between May and August 2016), depending on the species: Raphanus sativus was cut in May 2016, Crotalaria juncea in June 2016 and Cajanus cajan (L.) Millsp and Mucuna aterrima in August 2016. The crop residues were left on the soil surface. The maize hybrid 30F53VYHR was sown in November 2016 at a row spacing of 0.75 m and five seeds m−1, i.e., to achieve a density of 66,667 plants ha−1.
2.2. Soil Sampling and Analysis
The soil for analyses of ammonium (NH4+) and nitrate (NO3−) concentrations was sampled on two dates: on 6 April 2016 (after maize harvest, at the end of the rainy season), before cover crop planting and on 16 November 2016 (before maize planting, at the beginning of the rainy season). These two moments were chosen since the dynamic process of N mineralization is strongly related to the moisture level. As the Cerrado region is characterized by two seasons, rainy and dry, soil sampling was performed to represent these two conditions.
Composite samples of five sub-samples were collected per subplot (WN and NN) from the layers 0–5, 5–10, 10–20, 20–40 and 40–60 cm. The samples were ice-cooled in the field and in the laboratory, they were maintained in a freezer for a week until extraction and analysis. The soil ammonium (NH4+) and nitrate (NO3−) concentrations were analyzed at both samplings for each layer.
The soil mineral nitrogen concentration (mg kg
−1) in the forms of N-NO
3− and N-NH
4+ was determined through extraction in 50 mL of 2 mol L
−1 KCl, according to the method proposed by Bremner and Mulvaney [
30], and analyzed by colorimetry with a Lachat 228 Quikchem flow injection analyzer (Lachat Instruments, 5600 Lindbergh Drive, Loveland, CO 80539, USA). The soil moisture of each sample was determined using the gravimetric method. The gravimetric soil water content was determined after drying the material at 105 °C for 48 h. The soil mineral N concentrations (NO
3− and NH
4+ in mg kg
−1) were calculated based on soil dry weight.
2.3. Plant Analysis and Maize Yield
Crop residues of the cover crops were sampled at flowering within two rectangular iron sampling frames per subplot (0.38 × 0.58 m).
The plant biomass samples were dried at 65 °C, and a subsample was weighed to determine plant dry matter, and the value converted to kilograms per hectare. The difference between the cover crop sample weight before and after drying was the evaporated moisture. A 3 g sample of dry biomass was ground and oven-heated at 105 °C for 8 h. The total N content (TN) was analyzed by colorimetry with a Lachat 228 Quikchem flow injection analyzer (Lachat Instruments, 5600 Lindbergh Drive, Loveland, CO 80539, USA). The N concentration in plant tissues was used to calculate N uptake by cover crops, considering dry biomass.
The dry matter, acid detergent fiber (ADF), neutral detergent fiber (NDF) and lignin concentrations of the cover crop residues were analyzed at 105 °C [
31]. Analyses of neutral detergent (NDF), fiber and acid detergent (FDA) and fiber were analyzed using the ANKON system [
32]. Lignin was analyzed by digestion of the FDA residue with 72% sulfuric acid, which extracts cellulose and hemicellulose, generating lignin and inorganic matter as residues. Cellulose and hemicellulose were calculated as the differences between FDN and FDA residues and between FDA and lignin residues, respectively. The lignin concentration was given by the difference between the acid digestion residue and the ash after burning at 600 °C for 4 h. In November 2016, 4 m rows of each subplot were harvested to determine maize grain yield, and grain moisture was corrected to 13% (w.b.).
2.4. Nitrogen Fertilizer Use Efficiency (NFUE)
Nitrogen fertilizer use efficiency was calculated using the following equation, according to Dobermann [
33]:
N rate is the N topdressing rate applied (130 kg ha
−1 N).
2.5. Statistical Analysis
Before statistical analysis, the normality of the data was checked. Then, analysis of variance with data repeated over time (beginning and end of the rainy season) and space (0–5; 5–10; 10–20; 20–40; 40–60 cm) was performed to assess the effects of the plant species, N application and sampling times, in addition to the interactions between these factors. The data were analyzed using a two-way ANOVA, followed by Tukey’s test, considering
p < 0.05 for total variables, as a post hoc method to detect statistically significant differences among the treatments. This analysis of variance was performed using R version 3.5.0 software [
34].
Principal component analysis (PCA) [
34] was applied to a dataset with 16 rows comprising cover crops with and without N topdressing of maize (WN and NN), NH
4+ and NO
3− concentrations down to a soil depth of 60 cm, maize yield, total N concentration (TN), hemicellulose, cellulose, lignin and the lignin:N ratio.