1. Introduction
In addition to holding the largest commercial cattle herd in the world, Brazil is also the largest exporter of beef [
1]. The main factor making beef cattle a competitive activity worldwide is its production being carried out almost exclusively on pasture. Forage grasses of the species
Panicum maximum are among the main types cultivated in Brazil, having characteristics of high persistence in intensive management and high productivity, as a consequence of photosynthetic and water efficiency, as well as high phenotypic plasticity [
2]. Allied to this, the pastures of this species have high support capacity and animal performance, which can lead to different responses of the animal according to the management of the grazing adapted [
3].
Nitrogen (N) fertilization, coupled with maintenance fertilization with other nutrients essential to plants, is essential to ensure agricultural productivity [
4,
5,
6,
7]. In this respect, one of the main triggers of the pasture degradation process is low adoption of these technologies associated with inadequate management of pasture and grazing. The use of N increases and allows the intensification of the use of primary productivity through a higher stocking rate [
8]. Unfortunately, since the middle of the 20th century, the intensification of plant production based on mineral fertilizers has led to a loss of biodiversity [
9]. Additionally, there is a risk of changes occurring in the physical properties of the soil due to increased stocking rates promoted by the increase in pasture production [
7,
8,
10]. The pressure exerted by animal trampling can compromise the physical characteristics of the soil as a result of increased soil density (SD) and penetration resistance (PR) and reduced porosity [
11,
12].
Soil is considered a common benefit for all of humanity and is responsible for improving water quality, reducing gas emissions and increasing food production [
13]. Carbon (C) and N contents are the main indicators of soil health [
5], as they are responsible for physical, chemical and biological functions and are involved in nutrient retention [
14]. The main form of storage of C and N in pastoral environments is through the root system [
15]. However, most research done on forage plants neglects root growth.
It is therefore paramount that current livestock systems adopt measures that promote eco-functional intensification, taking into account the ecosystem services provided by animal production systems. On the other hand, information is lacking in the literature about the residual effect of N doses on pasture-based animal production systems and the participation of soils in environmental services. The knowledge of soil responses to agricultural management practices is essential for the sustainability and longevity of pasture-based animal production systems.
Thus, the hypothesis investigated in this study is that N has a residual effect on the chemical and physical characteristics and C and N stocks in the soil and in the root system of guinea grass after three consecutive years of using the same N levels on pastures grazed intensively by beef cattle. Accordingly, the objective was to examine the residual effect of N doses on the C and N stocks and on the quality of soil in an area cultivated by guinea grass grazed by beef cattle.
2. Materials and Methods
2.1. Experiment Site and Soil-Climatic Monitoring
The experiment was carried out at Embrapa Beef Cattle, in Campo Grande, state of Mato Grosso Sul (MS), Brazil (20°27′ S and 54°37203′ W, 530 m above sea level) (
Figure 1). Data were collected from November 2017 to May 2018
According to the Köppen classification, the climate of the region is an Aw type (rainy tropical savannah), with a defined dry season occurring from May to September. Precipitation in the area was monitored during the experimental period (
Table 1). Climatic data were obtained from the database of the National Institute of Meteorology [
16] at the Campo Grande-MS station.
The soil of the experimental area is classified as Red Latosol [
17], with clay contents between 30 and 35%. Before the beginning of the experiment (2014 and 2017), soil was sampled at the depths of 0 to 10 and 0 to 20 cm for chemical analysis (
Table 2).
2.2. Experimental Development
Guinea grass (Panicum maximum cv. Mombaça) was established in January 2008 (blocks I and II) and November 2010 (block III) and has been used since then under intermittent grazing by beef cattle. The 13.5-ha experimental area was divided into three blocks, each of which was subdivided into three 1.5-ha modules (one for each N level) and these into six 0.25-ha paddocks. Therefore, within each block there was a repetition of each N level. The pastures received three annual doses of N in the form of urea (100, 200 and 300 kg ha−1 of N) for three consecutive years (2014/2015, 2015/2016 and 2016/2017), in addition to topdressing with 80 kg ha−1 of P2O5 and 80 kgha−1 of K2O. After the rainy season of 2017, maintenance and N fertilization were no longer applied, which characterized the evaluation of the residual effect of N doses. Nitrogen fertilization was divided into three annual applications, the first dose of each treatment being applied together with phosphorus and potassium.
Intermittent grazing with a variable stocking rate was adopted as the grazing method, with five days of occupation and 25 days of rest in the first five grazing cycles (summer). In the sixth grazing cycle (fall), the pastures were managed with seven days of occupation and 35 days of rest. Fifty-four male Nellore cattle with an initial average live weight of 205 ± 25.8 kg were used. These animals remained in the paddocks throughout the experimental period, whereas a variable number of put-and-take animals was used to adjust the stocking rate during the occupation period so as to maintain pasture height at 40–50 cm in the post-grazing period [
18]. The paddocks were grazed clockwise.
The different N doses resulted in different rates of herbage accumulation between the treatments, which in turn defined different stocking rates throughout the experimental period (
Table 3).
2.3. Soil Chemistry
All sampling has always been carried out in the same pickets previously chosen at random in each module. Soil samples for chemical analysis were collected in May 2018, at three depths (0 to 10, 0 to 20 and 20 to 40 cm). For the collection, 0.5-m-deep trenches were dug in two paddocks of each experimental block. The samples were stored in a plastic bag, which was then labeled and sent to the Soil Laboratory at the Brazilian Agricultural Research Corporation (EMBRAPA-Beef Cattle). The following chemical attributes were evaluated: pH (1:2.5 ratio), potential acidity (H
++Al
3+), phosphorus content, exchangeable potassium, exchangeable calcium, exchangeable magnesium, base saturation, cation exchange capacity (CEC at pH = 7.0), sum of bases and organic matter content [
19].
2.4. Soil Density (SD)
Soil density was evaluated according to the methodology described by Embrapa [
19]. Trenches 120 cm in depth were dug in two paddocks per module. Eight samples were collected from each trench, four under and four between the tussocks of guinea grass, at the depths of 0 to 10, 10 to 20, 20 to 30 and 30 to 40 cm. For the collection, 100-cm
3 soil sampling rings were used.
2.5. Soil Penetration Resistance (PR)
Penetration resistance was measured in the pre- and post-grazing periods by direct reading, using an automated soil compaction meter (PenetroLOG PLG 1020) at 10 points in two paddocks of each module, up to the depth of 60 cm. Additionally, soil samples were collected at the time of the PR assessments to determine moisture at the depths at which PR was assessed, for a later correction of the values [
20].
2.6. Root System
To measure root dry mass, eight samples were collected in two paddocks of each module: four under and four between the tussocks. Each sample was sub-sampled at the depths of 0 to 10, 10 to 20, 20 to 30 and 30 to 40 cm. For collection, a cylindrical auger 4.8 cm in diameter and 10 cm in height was used. Subsequently, wet soil + root samples were packed in labeled plastic bags.
To separate the roots from the soil, the samples were placed on sieves with 2- and 1-mm screens under running water. Roots were oven-dried at 60–65 °C for 72 h or until reaching constant weight and were weighed to determine the DM content. As in the analysis of PR, soil samples were collected to determine moisture at the time of root collections. The soil moisture values were used to adjust the calculations of root dry mass (kg ha−1) and the percentage of root mass distribution (%) in the soil profile.
2.7. C and N Stocks
Soil samples were collected in May 2018 in six 120-cm-deep trenches in each experimental block. These were taken at the depths of 0 to 10, 10 to 20, 20 to 30, 30 to 40, 40 to 60, 60 to 80 and 80 to 100 cm to determine the levels of C and N. In addition to undisturbed samples, samples were also collected to determine SD for the calculations of the C and N stocks.
Soil samples for determining the C and N contents were packed in plastic bags, dried in the air, ground in a mill with 80-mesh sieves and subjected to a combustion process (temperature around 1000 °C). The C and N contents were determined in a C and N auto-analyzer (CN628, Leco). The gases generated in the combustion chamber (CO
2, N
2, H
2O) were separated in a gas chromatography column, which was followed by thermal conductivity detection as proposed by Embrapa [
19]. The stocks in the different soil layers were calculated according to the SD values, as follows: ttock (mg ha
−1) = C or N content (%) × soil density (g cm
−3) [
19].
2.8. Experimental Design and Statistical Analysis
The experiment was laid out in a split-plot, randomized-block design. The residual effect of N doses was allocated to the plot and the soil depths to the subplot. Data were subjected to analysis of variance using the following model: Yijk = μ + Di + Bj + αij + Pk + (D * P)ik + βijk, where Yijk = value observed using dose i, in block j, at depth k; μ = overall-mean effect; Bj = effect of block J; Di = effect of N dose i (i = 100, 200 and 300); αij = effect of the random error attributed to the plot; Pk = effect of depth k; (D * P)ik = interaction effect between dose and depth; and βijk = random error attributed to the subplot.
When determined to be significant by the F test, the residual effects of N doses at the soil depths and their interactions were analyzed by Tukey’s test at 5% significance.
3. Results
The interaction effect between N dose and soil depth was not significant (p > 0.05) for the studied variables. Thus, the effects of N doses and soil depths were evaluated and discussed in isolation.
3.1. Soil Chemistry
The N doses did not influence (
p > 0.05) soil pH (5.2 ± 0.02), exchangeable calcium (2.1 ± 0.09 cmolc dm
−3), exchangeable potassium (0.22 ± 0.01 cmolc dm
−3), potential acidity (3.5 ± 0.14 cmolc dm
−3), cation exchange capacity (7.0 ± 0.06 cmolc dm
−3), base saturation (49.3 ± 1.4%) or phosphorus content (4.1 ± 0.74 mg dm
−3). However, the doses affected the levels of exchangeable magnesium (
p = 0.0229); the highest concentrations in the soil were found using the N doses of 100 and 200 kg ha
−1 and for organic matter content (
p = 0.0046), the highest values in the soil occurred under the N doses of 200 and 300 kg ha
−1 (
Figure 2).
The effect of the depths was not significant (
p = 0.0745) for the soil pH (5.2 ± 0.04). The highest values of potential acidity (H Al) (
p = 0.0001) were found at the depths of 0 to 10 and 0 to 20 cm, and the lowest of 20 to 40 cm. The levels of exchangeable potassium (
p = 0.0001), exchangeable calcium (
p = 0.0001), exchangeable magnesium (
p = 0.0001) and base saturation (
p = 0.0016) were higher at depths of 0 to 10 and 0 to 20. The cation exchange capacity (
p = 0.0001), phosphorus levels (
p = 0.0001) and organic matter content (
p = 0.0001) decreased with increasing depth (
Figure 3).
3.2. Soil Density
No difference was observed between the N doses for SD under the tussocks (
p = 0.1025), which averaged (± standard error) 1.22 ± 0.02 g cm
−3; or between the tussocks (
p = 0.4666), where SD averaged 1.25 ± 0.04 g cm
−3. However, soil density varied according to the depths, under (
p = 0.0056) and between (
p = 0.0040) the tussocks. Under the tussocks, SD was equal at the depths of 0 to 10 cm and 10 to 20 cm. The SD values found in these layers were greater than at the depth of 20 to 30 cm, and the latter was greater than at 30 to 40 cm. Between the tussocks, the lowest SD was observed at the depth of 30 to 40 cm and the highest at 0 to 10 cm, which in turn was greater than that observed at 20 to 30 cm, whereas intermediate values were found from 10 to 20 cm (
Figure 4).
3.3. Soil Penetration Resistance
Penetration resistance did not differ (1.4 ± 0.02 Mpa) between the N doses in the pre-grazing period (p = 0.7026). However, the soil depths affected this variable; the highest value (1.6 ± 0.04 Mpa) was found at the depth of 10 cm.
In the post-grazing period, a difference was observed between the N doses (
p = 0.0396) and between depths (
p = 0.0009) for PR (
Figure 5B). The highest mean PR (1.9 ± 0.07 MPa) was found in the pastures that received 300 kg ha
−1 N for three consecutive years. Between the depths, the highest PR was observed at 10 cm (2.7 ± 0.09 MPa).
3.4. Root System
The N doses did not change the root dry mass of guinea grass. Mean values (±standard error) for this variable were 2589.5 ± 466.6 kg ha−1 (p = 0.5613) under and 941.2 ± 276.8 kg ha−1 (p = 0.2773) between the tussocks, respectively.
There was an effect of the depths for the root system variables. The dry mass of roots under the tussocks was greatest at the depth of 0 to 10 cm, which decreased with increasing depth. When the material was sampled outside the tussock, the largest root dry mass was obtained at the depth of 0 to 10 cm and smallest between 30 and 40 cm, with intermediate values found at the other depths (
Table 4).
3.5. C and N Stocks
There was no residual effect of N doses for the average (
p = 0.1019) and total (
p = 0.3677) C stock in the soil (
Table 5) under guinea grass. However, the C stocks fluctuated as a function of the depths (
p = 0.0001), with the highest values detected at 0 to 10 cm and lowest at the depths of 20 to 30 and 30 to 40 cm.
The total N stock did not change (
p = 0.64510) in response to the N doses. However, the average stock was largest (
p = 0.0484) in the soil cultivated with guinea grass under the residual effect of the N dose of 300 kg ha
−1 (
Table 6). As regards the distribution of N in the soil profile, the highest (
p = 0.0001) values were observed at the depth of 0 to 10 cm and the lowest from 20 to 30 cm.