Soil Physical Quality After 21 Years of Cultivation in a Brazilian Cerrado Latosol

Long-term studies aiming soil quality evaluation under different soil management strategies are no common. Long-term evaluations provided more reliable contributions to decision-making and practices adoption. This study evaluated the soil physical quality of a Brazilian Cerrado Latosol after 21 years of three different soil management strategies: disc plowing (DP), no-tillage (NT), and disc harrowing+subsoiling (DHS). In comparison to the reference, a soil from a native Cerrado area, the removal of the original vegetation and the implementation of the three soil management strategies increased the soil bulk density (Bd) and reduced soil porosity, macroporosity, soil organic carbon (SOC) and the size of water-stable aggregates, but did not change the glomalin-related soil protein (GRSP) contents and clay flocculation. Similar effects were diagnosed on soil physical quality when is considered only the three different management strategies, especially on soil porosity, Bd, size of water-stable aggregates, SOC and GRSP contents. Strategies of DP and NT increased soil resistance to penetration in the superficial layers, while the annual use of DHS reduced this soil mechanical characteristic. The NT system did not provide increasing of soil organic carbon in comparison to other management practices evaluated. In conclusion, removing the native vegetation affected soil physical quality, but the Brazilian Cerrado soil is resilient to physical damage even when different intensive farming practices are implemented for more than two decades. The limitation of the NT system in improving the soil physical quality is related to climate conditions that determine the non-maintenance of straw on the soil surface.


Introduction
The Latosols in the Cerrado biome located in the central region of Brazil are well-developed pedogenetically, homogeneous, and present small variations between horizons.The natural fertility of these soils is low; however, physical characteristics such as drainage and depth are adequate, which make them very useful for agricultural production.The soil management systems adopted in this region have been more focused on increasing productivity by adopting technological innovations and intensifying land use.In contrast, the Brazilian Cerrado is considered one of the world's biodiversity hotspots, constituting the second largest biome of the South American continent and covering an area of 2,036,448 km 2 in states in the north, northeast, midwest, south, and southeast regions of Brazil (MMA, 2018).
Loss of soil physical quality is relatively higher in clay soils, where compaction is higher (Horn, 1988).In Brazil, clayed soils are typical in the Cerrado area where technological innovations in agriculture are responsible for success and a significant percentage of the national gross domestic product.Therefore, adopting management and agricultural practices that depreciate the soil physical quality may decrease productivity in this important agricultural region of Brazil.
Soil erosion and degradation are minimized or reduced by using conservative management practices, with an emphasis on no-tillage systems (Wang et al., 2017).According to the Brazilian Federation of No-Tillage Systems, approximately 32 million hectares were planted with summer crops in Brazil before the 2013/2014 harvest under no-tillage systems, and these area size has increased since the beginning of the practice in 1972 (FEBRAPDP, 2018).
Conservation agriculture represented by no-tillage minimizes soil disturbance by reducing the mineralization of organic matter, consequently increasing the organic carbon content and improving soil aggregation and infiltration (Guo et al., 2016).
Impacts resulting from intensive soil management practices have decreased crop productivity, especially when the soil structure is damaged and soil organic matter concentration is decreased (Zhang et al., 2017).In this context, management practices that have a lower impact on the soil and that maintain or increase organic matter concentration are recommended.
The soil organic matter is a critical component associated with the development and maintenance of soil structure and is the focus of different management practices.Several studies demonstrated the role of glomalin-related soil protein (GRSP) in binding soil particles.Sharifi et al. (2018) suggested the use of the ratio between GRSP and soil organic carbon as an indicator of the level of disturbance of crop systems.
The no-tillage system has been highlighted as a conservative soil management practice in Brazil and worldwide, mainly when it is associated with increased soil mulch.This system reduces the risk and rate of erosion and increases soil organic matter, soil infiltration, soil fertility, and overall soil quality (Ogban et al., 2001;Iqbal et al., 2005).
Most studies on the effects of soil management practices on soil quality were short-term and lacked methodological rigor.Long-term studies are rare because of their complexity and costs, which limits the obtaining of results and recommendations with a higher degree of reliability.
The objective of this study was to evaluate the soil physical quality of a Brazilian Cerrado Latosol after 21 years of intensive cultivation under no-tillage and other soil management strategies.

Area Description
This study was carried out in an experimental area of the Brazilian Agricultural Research Company (Embrapa Milho e Sorgo) located in the municipality of Sete Lagoas (19°27.408′S and 44°10.939′W; and 786 m of altitude), Minas Gerais State, Brazil.According to Köppen's classification, the climate of the region is type Cwa, with dry winter and hot summer, and temperatures above 22 °C in the hottest month of the year.The soil of the experimental area was classified as clayey Red Latosol (EMBRAPA, 2013) with a very clayey texture (Table 1).The native soil presents limited fertility, but cultivation practices could improve its chemical attributes.
The study area was previously occupied by pastures.From 1995 to 2016, this area was divided into 320 m 2 plots (20  16 m) and used for cultivation under different soil management strategies.In the present study, the areas cultivated using disc plowing (DP), no-tillage (NT), and disc harrowing + subsoiling (DHS) were selected.A nearby and contiguous area with native Cerrado (NC) vegetation (not included in the original experiment) was used as the reference.All evaluated areas, including the NC area, presented similar soil class, slope, landscape position, and face of exposure to solar radiation.
All experimental plots were cultivated with corn, except for NT soils, which was characterized by corn-soybean rotation.The experiment was analyzed considering a completely randomized block design with three replications.
In the DP treatment, the equipment used was a disc plow with three discs (diameter of 32″).A three-shank subsoiler and an intermediate disc harrow with 16 discs (diameter of 28″) were used in the DHS treatment.Leveling procedure was performed in DP and DHS treatment using a leveling apparatus after soil preparation.Herbicide glyphosate was applied before cultivation in NT soils for desiccation.This same herbicide was applied to all cultivated areas during the offseason to manage weed growth.All operations were performed using two tractors (4,150 kg and 5,500 kg).
The corn crop was cultivated (65,000 plants/ha) in all areas using a seed and fertilizer spreader.Annual maintenance and cover fertilization and the application of gypsum and limestone were performed according to technical recommendations based on soil analyses.Phytosanitary control in the study area was carried out by applying insecticides, with two to three applications after cultivation according to technical recommendations.
The corn-soybean rotation was performed in NT treatment.In the years of soybean cultivation, a total of 500 kg/ha of NPK formulation 4-30-16 was applied to this crop at sowing (280.000plant/ha).Millet was grown with irrigation in the NT areas in the off-season in the years 2008 and 2010, and the formed and incorporated biomass was 60 t/ha of green mass.No other crop was used in the off-season of other experiment years.Note.NC = native Cerrado; DP: disc plowing; NT = no-tillage; and DHS = disc harrow + subsoiler.Analyses: pH in water-ratio of 1:2.5;P and K using Mehlich-1 extractor; Ca 2+ Mg 2+ Al 3+ with KCl 1 mol/L; H+Al with calcium acetate 0.5 mol/L, pH 7.0; SB = sum of bases; CEC = cation exchange capability effective (t, at original pH) and at pH 7.0 (T); V = base saturation; m = aluminum saturation; Pd = particle density.All procedures are according to EMBRAPA (2017).

Methods and Techniques
Disturbed and undisturbed samples were collected in July 2016.The disturbed samples from 0.00-0.20 m were obtained using an auger probe.Three composite samples were used per treatment, corresponding to 20 single samples collected per plot.In the laboratory, soil samples were sieved through a 2.0 mm sieve to obtain soil material for the analyses.The undisturbed samples were obtained using volumetric cylinders of approximately 0.05 m in height and diameter.In each treatment, six rings were collected in the center of the layer at a depth of 0.00-0.10m.
The soil mechanical resistance to penetration (RP) was evaluated in the field in February 2017 at a depth of up to 60 cm using a PenetroLOG digital penetrometer (model PLG1020; Falker), and measurements were performed every 0.01 m at a constant speed.The data were extracted from the storage unit using Penetro-LOG software and graphed and expressed by the mean values in 10 cm intervals, as follows: RP1 (0-10 cm), RP2 (10-20 cm), RP3 (20-30 cm), RP4 (30-40 cm), RP5 (40-50 cm), and RP6 (50-60 cm).
Wet aggregate stability was evaluated in samples pre-moistened and shaken in a set of sieves of different mesh sizes.After agitation, the weight of the samples retained in each sieve was used to calculate the mean weight diameter (MWD) and geometric mean diameter (GMD) according to Embrapa (2017).
Soil bulk density (Bd) was determined using the volumetric ring method, and particle density (Pd) was measured using the volumetric flask method.Microporosity (Mi) was determined in undisturbed samples after equilibrium to the -0.006MPa on a tension table.Total porosity (Pt) was estimated by the equation: Macroporosity (Ma) was calculated by the equation: Water-dispersible clay (WDC) was determined using the pipette method.The degree of flocculation (DF) relative to the total clay (TC) content was calculated according to the equation: All analyses were performed according to Embrapa (2017).The soil total organic carbon (TOC) was determined according to the Walkley & Black method (Yeomans & Bremner, 1988).
Physical fractionation of soil organic matter was carried out to obtain particulate organic matter (POM) and mineral-associated organic matter (MOM) according to the methodology proposed by Cambardella & Elliott (1993).The TOC in the MOM (TOC-MOM) was determined according to Yeomans and Bremner (1988).The TOC in the POM (TOC-MOP) was obtained by subtracting the TOC-MOM from the total soil organic carbon.
The specific surface area (m 2 /g) of soil samples was quantified by measuring the water vapor sorption (Quirk, 1955).
The GRSP concentration, including the easily extractable glomalin (EEG) and total glomalin (TG), was determined according to Wright et al. (1996).For EEG processing, autoclaving was performed once.For TG, autoclaving was performed six times until the solution reached a light-yellow color.These two fractions were quantified by the Bradford (1976) colorimetric method using a spectrophotometer and bovine serum albumin as the standard.

Statistical Analysis
The data were subjected to analysis of variance to assess differences between the treatments after confirming the normality of residuals using the Shapiro-Wilk test (p < 0.05).Dunnett's posthoc test (p < 0.10) was used to compare the means of the treatments relative to the control area (NC).The mean values in the cultivated areas were compared using Tukey's test (p < 0.10).All statistical analyses were performed using R software version 3.4 (R Core Team, 2017) and the "ExpDes.pt"package developed by Ferreira et al. (2009).

Results and Discussion
The soil physical and chemical attributes are shown in Table 2.The different management practices for more than 20 years changed around one-third of the soil attributes in comparison to the native Cerrado (p < 0.10).

Effects on Soil Structure
The soil structure was used to access the soil physical quality because this parameter is highly sensitive to soil management practices.The native Cerrado area (NC) presented a larger and more stable aggregates as indicated by MWD and GMD values.Beutler et al. (2001) observed a similar result in an adjacent site of the present study.
The authors verified higher GMD in the NC area, although the soil of no-tillage (NT) treatment has presented larger aggregates (> 2 mm) in a proportion similar to that of the NC area at a depth of 0 to 5 cm.
The removal of the native vegetation in the cultivated areas and the use of maintaining soil management practices for two decades reduced the average size of the aggregates, although the differences between the three cultivated areas (NT, DP, and DHS) were not significant.The presence of larger and more stable aggregates increases soil pre-consolidation pressure (Letey, 1985;Dexter, 1991), facilitates water infiltration by increasing macroporosity, and promotes water retention in micropores (Dexter, 1988).
The bulk density (Bd) increased with the land use and management in the three cultivated areas in comparison to the NC area.Nonetheless, we cannot find differences among the cultivated areas.The Bd value in soil with NT (1.33 kg/dm 3 ) was 56% higher than in the NC area (0.85 kg/dm 3 ).Increases in Bd values under NT systems have been reported in the literature (Sheehy et al., 2013;Domínguez & Bedano, 2016).For this reason, is usual to recommend periodic and minimal tillage in NT soils to improve their physical properties.Considering that, Camara and Klein (2005) observed the reduction of Bd values and the increase of water infiltration in soils submitted to the scarification and managed under NT after six years.
The soil porosity (P) was not decreased with cultivation using disc plowing compared with the NC area.However, these soil characteristics was reduced in NT soils and in the soils managed by harrowing and subsoiling.The increase in Bd in NT soils decreased the soil porosity, as reported in previous studies (Silveira et al., 1999;Silva et al., 2008).In contrast, an increase in soil porosity in the surface layer was observed by Loss et al. (2017) in vegetables grown under NT compared to conventional tillage, and this result was attributed to the higher number of roots and consequently higher occurrence of voids in the soil matrix.In turn, Costa et al. (2003) found no differences in total porosity, macroporosity and microporosity in a Brazilian Latosol managed in the long-term under NT and conventional tillage.Note.NC = native Cerrado; DP = disc plowing; NT = no-tillage system; DHS = disc harrow + subsoiler.MWD = mean weight diameter; GMD = geometric mean diameter; Bd = bulk density; WDC = water-dispersible clay; RP = soil resistance to penetration; TOC = total organic carbon; TOC-MOM = TOC of the mineral-associated organic matter fraction; TOC-POM = TOC of the particulate organic matter fraction; and SS = soil specific surface area; Means followed by standard errors.CV = coefficient of variation in the cultivated areas (DP, NT, and DHS).Samples collected at a depth of 0 to 20 cm, except for the evaluation of Bd (0-0.1 m) and RP.The means followed by an asterisk (*) were significantly different from the reference area (NC) using Dunnett's test (p < 0.10).For the cultivated treatments (DP, NT and DHS) the means followed by the same letter in each line were not significantly different using Tukey's test (p < 0.10).
The dynamics of soil macropores (Ma) and micropores (Mi) was affected by soil use and management.The removal of the NC and the implementation of treatments cultivated increased the Mi and decreased Ma.Under the NT system, there was a reduction in Ma values, reaching values of 0.02 m 3 /m 3 .Similarly, Stone and Silveira (2001) observed that Bd and Mi increase whereas Ma was decreased in NT soils in Brazilian Cerrado.Moreover, Costa et al. (2009) verified that Ma values were increased in a Humic Cambisol managed by plowing and harrowing during 15 years in comparison to soils under NT for nine years.
Soil structure affects the growth and development of crop roots.In this context, the analysis of soil Ma is essential because macropores are crucial for soil water infiltration and are the preferred route for root growth (Calonego et al., 2011).Given that the larger pores are the most affected during soil compaction, good management practices should be prioritized, especially for the traffic of agricultural machines (Bergamin et al., 2010).
Although the average size of aggregates (WMD and GMD) was decreased by adopting cultivation practices, even after two decades, these practices did not affect the clay flocculation.The change in inputs, the decrease in 50 cm no-tillage promote higher RP values than the reference area (NC).Soil resistance to penetration with disc plowing and disc harrow and subsoiler was higher than that in the native Cerrado at a depth of 10-50 cm and 20-50 cm, respectively.In the soil superficial layer (0-10 cm), the lower values of RP with DP and DHS may be due to soil tillage during soil preparation for cultivation.
In general, cultivation practices increase the soil resistance to penetration until 50 cm depth.The no-tillage system presented the highest RP values until 30 cm depth whereas DHS-treated soils presented the lowest RP (Table 2).No differences among cultivated areas (NT, DP and DHS) were diagnosed after 30 cm depth.
Treatments with soil revolving (DP and DHS) presented similar soil resistance to penetration to reference area (NC) in the superficial layer studied (RP1).Disc plowing treatment exhibited high resistance to penetration at 21-30 cm depth, coinciding with the depth of cut of disc plow.
The soil resistance to penetration verified at the no-tillage area is coherent with the Bd values.These results suggest the formation of a "no-till pan" as proposed by Reichert et al. (2009), which usually develops under long-term no-tillage and is commonly found at of 7-20 cm depth.These authors characterize this layer with high mechanical strength, high bulk density and low porosity.Hamza and Anderson (2005) highlight the preference of some farmers for conventional tillage over no-tillage systems because although can increase soil organic matter, it is common to find critical values of RP (> 2 MPa) in no-tillage areas and severe restrictions to root growth can be observed.However and in contrast to this restrictions, most parts of Brazilian farmers consider no-tillage promotes economic environmental and social benefits and its adoption is more than 50% of the cultivated area with annual crops (Freitas & Landers, 2014).
Our results show around 3.5 MPa as the higher RP found.The critical value of RP varies among crops and for corn 1.5 to 4.0 MPa are usually considered critical, although values of 1.3 MPa can reduce 50% of plant growth (ROSOLEM et al., 1999).
Soil moisture should be considered when evaluating RP Lower RP values in the native Cerrado and the use of DHS coincided with higher soil moisture (Table 3).In turn, higher RP values in no-tillage and disc plowing treatments were coincident with lower moisture.Therefore, considering that the treatments were close, without evidence of differences in rainfall, the maintenance of higher soil moisture was only achieved in the areas managed with harrow with the subsoiler.
It is well known that resistance to penetration depends on soil Bd and moisture.We can assume that the differences in RP values were due to soil moisture because there were no differences in Bd among the three cultivated areas (Table 2).Therefore, a linear relationship between RP and soil moisture (Figure 2) indicates that a 1% rise in gravimetric moisture decrease in RP of 0.1 MPa.

Table 1 .
Soil chemical and physical characteristics of the study areas (0-0.20 m depth)

Table 2 .
Physical and chemical characteristics of a Brazilian Cerrado Latosol managed using different practices for more than 20 years

Table 3 .
Gravimetric water content during the mechanical resistance to penetration measurement