Land-Use Changes the Chemical and Physical Properties of an Oxisol in the Brazilian Cerrado

This research aimed to evaluate the effect of different land-uses on physical and chemical properties of Oxisols under cerrado conditions in central region of Goiás, Brazil. The data were analysed in a randomized experimental design in a 4 × 3 factorial arrangement with five replications. The land-uses were: 1-Annual crops, 2-Pasture, 3-Sugarcane and 4-Native forest (control). The second factor consisted of three sampling depths: 0-0.05, 0.1-0.2 and 0.2-0.3 m. The physical attributes analysed were: total clay (TC), natural clay (NC), bulk density (BD), particle density (PD), total porosity (TP), degree of flocculation (DF), soil resistance to penetration (SRP), gravimetric water content (%GWC). Chemicals attributes analized: pH in water (pH), calcium (Ca), magnesium (Mg), aluminium (Al), potential acidity (H+Al), potassium (K), phosphorus (P), soil organic matter (SOM), cation exchange capacity (T) and base saturation (V%). Annual crops showed higher K levels (0-0.3 m) and lower SOM values (0-0.3 m), Ca and Mg (0-0.05 m) affecting T in relation to the forest. The use sugarcane showed higher values of BD, PD and SRP, and smaller values of TP, SOM, P and T. On the other hand, the values of Ca and Mg are kept close to those of the control. The attribute V% is preserved for annual crops, sugarcane and pasture with values higher than those verified in native forest. The use pasture reduced TP and P and Ca, Mg and T. The land-uses studied reduced soil quality compared to the forest, in descending order were sugarcane > annual crops > pasture.


Introduction
The Cerrado is the second largest Brazilian biome in territorial extension, over which the largest conversion of native forest areas to agricultural and/or livestock activities has occurred in recent decades in Brazil. Although this conversion has generated economic progress for this region, the inadequate use of the soil by these economic activities has accelerated the processes of soil degradation.
Variations in terms of intensification of agricultural systems are present in the region of Ceres, state of Goiás (Brazil). In this Cerrado region, pastures (native and cultivated), sugar cane cultivation, fruit growing, annual crops (mainly corn, sorghum and soybean) and olericulture (Marques, 2013) predominate. In addition, the soils of this region, in general, have good natural fertility and good water retention capacity; however, they have undulating relief in large part, which hinders agricultural mechanization and favours the occurrence of water erosion. Annual crops Area used predominantly for corn, bean, soybean and sorghum crops for more than 30 years in the conventional system with heavy harrow followed or not by level harrow. Soil correction every five years (on average) and application of 300 to 500 kg ha -1 of NPK to each plantation. It is located in the UTM coordinates: E = 650,208.78 m and N = 8,302,657.07 m, Datum WGS_1984, Zone 22S, Central Meridian-51. Total area of 1.4 hectares.

Sugarcane
The area has been used for more than 15 years with cultivation with manual harvesting and burning. The soil is corrected and fertilized by means of liming and plastering every 5 years on average and mineral fertilization. The area has terraces for erosion control. The different land-uses were selected in an area under the same soil type, in this case, Typic Hapludox (Oxisol), classified according to the USDA Soil Survey Staff (2014) (Figure 1). In each area corresponding to land-uses, disturbed and undisturbed soil samples were collected to determine, respectively, chemical and physical attributes, at the three depths mentioned. Sampling was performed in a zigzag pattern, with an average distance of 10 meters between collection points (repetitions), with three simple samples per point, from which a composite sample originated (disturbed samples). The chemical and physical analyses of the soil were carried out according to the methodology of Donagema, Campos, Calderano, Teixeira and Viana (2011) with some modifications.
The physical attributes evaluated were total clay (TC), natural clay (NC), degree of flocculation (DF), bulk density (BD), particle density (PD), total porosity (TP), soil resistance to penetration (SRP), and concomitant to this analysis, determination of gravimetric water content (%GWC), at the same depths. The SRP analysis was carried out using a digital penetrograph (Eijkelkamp model), with 10 samples per plot. TC and NC were determined by the pipette method (Day et al., 1965;modified) after dispersion in water (NC) or 0.1M NaOH solution (TC). DF was determined according to the following equation: PD was determined through the volumetric flask method using ethyl alcohol. BD was determined using a cylindrical metal sampler (0.05 × 0.05 m), which was oven-dried at 105 °C and weighed. BD was obtained by the equation:

BD = Ms/Vs
where, Ms is mass of dry soil (g) and, Vs is volume of the sampler (cm 3 ).
TP was obtained by the equation: The chemical attributes evaluated were pH in water, phosphorus (P) (Mehlich I), potassium (K + ), calcium (Ca 2+ ), magnesium (Mg 2+ ), aluminium (Al 3+ ), potential acidity (H+Al), soil organic matter (SOM), base saturation (V%) and cation exchange capacity (T). Ca, Mg and Al were extracted with 1M KCl and, P and K with the Mehlich I solution (0.05M HCl + 0.0125 M H 2 SO 4 ). P was measured colorimetrically by the molybdenum blue method. Aluminum was measured by titration with 0.025 M NaOH; Ca, Mg and Mg by atomic absorption spectroscopy. Potential acidity was extracted with 0.01 M calcium chloride buffered and measured by potentiometry with combined electrode immersed in suspension. SOM was extracted by oxidation with sodium dichromate in sulfuric acid medium, and determined by spectroscopy UV-VIS (650 nm).
The data were checked for normality and homogeneity of variances of the residuals using the SAEG 4.0 software. Then, analysis of variance (ANOVA) was performed, and when there was an effect of the tested factors (land-use and depth), the averages were compared using the Scott-Knott test (P < 0.05). Pearson's correlation analysis (P < 0.05) between the physical and chemical attributes of the soil was also performed.

Results and Discussion
Tables 2 and 3 present summaries of the analysis of variance (ANOVA) for the physical and chemical attributes of the soil, respectively. The different land-uses and soil depth evaluated significantly influenced most of the physical and chemical attributes of the soil. There was a significant interaction of land-use and depth factors for all physical and chemicals soil attributes except, NC, DF and pH with no significant differences for the treatments tested (Table 3). Note. VF: variation factor. LU: land-use. SD: soil depth (SD1 = 0-0.05 m, SD2 = 0.10-0.20 m, SD3 = 0.20-0.30 m). CV(%). Coefficient of variation. SE: standard error. DF: degree of freedom. TC: total clay. NC: natural clay. DF 1 : degree of flocculation. BD: bulk density. PD: particles density. TP: total porosity. SRP: soil resistance to penetration. GWC: gravimetric water content. ns = not significant; * significant (P < 0.05); **significant (P < 0.01).  Table 4 shows the breakdown of the interaction between the factors of land-use and soil depth for significant physical attributes. For the total clay (TC) attribute, it was found that the pasture use had a lower content in the superficial layers of the soil (0-0.05 and 0.1-0.2 m), as well as in the 0.2-0.30 m layer, however, accompanied by sugarcane. It was found that all land-uses (except pasture 0-0.05 m) had TC values greater than 350 g kg -1 and were therefore classified as clayey (Sousa & Lobato, 2004). In such studies, it is always desirable that the TC jas.ccsenet.org Journal of Agricultural Science Vol. 12, No. 9;2020 content be similar between the study areas (Table 4), which indicates their uniformity. However, this result does not always occur, even when working with the same soil class (Gasparetto, Braida, Carneiro, Scariot, & Tabolka, 2009), which is unfavourable because it is an additional cause of variation between treatments. - Note. TC: total clay. BD: bulk density. PD: particles density. TP: total porosity. SRP: soil resistance to penetration. GWC: gravimetric water content. CV: coefficient of variation. Different lower-case letters in the columns within the same soil depth and upper-case letters in the columns between soil depths differ by the Scott-Knott test (P < 0.05).
Regarding the bulk density (BD), the forest use presented the lowest averages in the superficial layers of the soil (0-0.05 and 0.1-0.2 m). This effect possibly occurred due to the absence of traffic and trampling and the continuous accumulation of organic material on the soil (e.g., leaves and branches of plants), which results in increased soil porosity and, consequently, reduces the BD, confirmed by a significant negative correlation between BD and TP (   At a depth of 0.2-0.3 m, the soil cultivated with sugar cane showed the highest BD, while the other treatments did not differ from the forest. In the cultivation system in which sugarcane is harvested manually and with burning, the traffic of machines is intense due to the various operations carried out. Agricultural machinery traffic increases the BD and decreases the weighted average diameter and the macroporosity in the wheel set line in relation to the planting line, which causes cumulative degradation of the physical quality of the soil over the years of cultivation (Roque, Souza, Barbosa, & Souza, 2010).
The BD values presented for the sugarcane area are still within the limits considered tolerable by Beulter and Centurion (2004) and Beutler et al. (2005), who indicated values on the order of 1.55 kg dm -3 as critical for soils with a clayey texture. This increase in BD can also be explained by the reduction in the content of organic matter in the soil (SOM) in comparison with the forest use. At all depths evaluated, the forest use had a higher SOM content (Table 6).   Mg: magnesium. H+Al: potential acidity. T: cation exchange capacity. SOM: soil organic matter. V%: base saturation. P: phosphorus. K: potassium. CV: coefficient of variation. Different lower-case letters in the columns within the same soil depth and upper-case letters in the columns between soil depths differ by the Scott-Knott test (P < 0.05).
The total porosity (TP) under forest use was greater than that under other land-uses at depths of 0-0.05 m and, together with the use of annual crops, also at a depth of 0.1-0.2 m. At a depth of 0.2-0.3 m, the use of sugarcane had the lowest average, with no differences between the other land-uses. Only the pasture area showed significant differences between depths, where lower TP values were observed within the first 20 cm, which is possibly related to the higher BD found in this layer. The results found by the present study corroborate those of Tormena, Vidigal Filho, Gonçalves, Araújo and Pintro (2004) for total porosity (TP), being lower in the uses of cane, pasture and crops compared with the forest area.
Regarding the SRP, there were only differences between land-uses at a depth of 0.2-0.3 m, where native forest and annual crops presented the highest averages. This is probably due to the "grid foot" resulting from the conventional tillage system in the area of annual crops, and due to the effect of the greater amount of thick roots in the forest area, as they present similar humidity (GWC) at that depth (Table 4). In the 0-0.05 m layer, the SRP values were lower than the critical level (> 2 MPa) except for cane use. In the deepest layers (0.1-0.2 and 0.2-0.3 m), in all types of uses, it can be noted that the values are higher than the limit considered critical characterizing soil compaction. These results corroborate those found by Oliveira, Rocha, Rocha and Fiorillo (2010). Table 6 shows the breakdown of the chemical attributes of the soil that showed a significant difference for the use x depth interaction. In the 0-0.05 m layer of exchangeable aluminium content (Al), there were no difference between the evaluated land-uses; however, in the subsequent layers (0.1-0.2 and 0.2-0.3 m), the content of Al was lower than in the forest. The lower values of Al in annual crops and sugarcane uses may be related to the residual effect of soil correction carried out with the application of limestone. Several studies show that limestone has a residual effect that persists for several years (Cravo, Smyth, & Brasil, 2012;Natale, Rozane, Parent, & Parent, 2012;Wang, Tang, Baldock, Butterly, & Gazey, 2016).
As for the calcium content in the soil (Ca), there were only differences between land-uses at a depth of 0-0.05 m, where the forest showed the highest average. It was also found that the forest use showed lower levels of Ca in the layers of 0.1-0.2 and 0.2-0.3 m. The higher content of this element in the forest use is probably related to the greater release due to the deposition and decomposition of litter. In relation to magnesium (Mg), there were only differences between land-uses in the 0-0.05 m layer, where annual crops showed the highest average. There was only a depth effect of the land on pasture use, with a higher content of this nutrient in the 0-0.05 m layer, which may be related to the deposition and decomposition of grass residues. The Mg content is positively correlated with the soil organic matter (SOM) content (Table 5), according to previous studies (Triantafyllidis, Kosma, & Patakas, 2018).
The SOM content was influenced by both land-use and depth (Table 6). SOM was lower in sugarcane uses and annual crops at all depths evaluated. The lower SOM content under these two land-uses is directly related to soil management and the type of harvest (burning the cane field), which contribute to reducing SOM over time, as noted by Mendonza et al. (2000).
As expected, there was a higher content in the 0-0.05 m layer and a reduction in the SOM content with increasing soil depth for all soil uses (except annual crops). This is related to the higher surface deposition of plant and animal waste, which was also observed by Centurion, Cardoso, and Natale (2001) and Wendling, Jucksch, Mendonça and Neves (2005) and with the highest concentration of plant roots on the surface (Oliveira et al., 2009).
In relation to base saturation (V%), there were differences between land-uses at depths of 0.1-0.2 and 0.2-0.3 m, where forest use had the lowest average. Among soil depths, there were differences only for forest use, where the 0-0.05 m layer presented the highest average. This absence of differences between cultivated areas may be associated with the use of fertilizers and acidity correctives, which allow neutralization of H and Al and increase the availability of bases (Hunke et al., 2014). The V% values ranged from 35% to 58%, which are considered adequate for these land-uses (Sousa & Lobato, 2004).
Regarding the total cation exchange capacity (T), the values found for the forest area were higher than those of other land-uses at all depths studied ( Table 6). The areas under annual crops and sugarcane showed a reduction in T values with increasing soil depth. These results confirm that the absence of soil disturbance allows conservation of the physical and chemical properties of the soil (e.g., T), which are directly related to the conservation of the SOM, which reaches up to 90% of the soil T (Novais, Smyth, & Nunes, 2007), according to a positive correlation (P < 0.05) between SOM and T ( Table 5).
The phosphorus (P) content was higher in all depths evaluated when using annual crops, and at a depth of 0.2-0.3 m, the P content in sugarcane use did not differ from that of annual crops. This was due to the adopted fertility management strategy, with frequent fertilization and little mobility of this element. In the case of soil under sugar cane cultivation, the higher P content is probably associated with the deeper application with a furrow/fertilizer in planting cane with reforestation of the cane field. It was found that the forest area showed higher values of potential acidity (H+Al) than other land-uses at all depths studied (Table 6). SOM has several functional groups, especially the carboxylic and phenolic groups, which release the H that will constitute the ions of the soil cation exchange capacity. According to Cardoso, Silva, Curi, Ferreira and de Freitas (2011), this indicates that SOM can be an important source of potential acidity in tropical soils, especially in environments where litter accumulation occurs. Similar results were also found by Portugal, Costa and Costa (2010), with higher values observed in the use of forest (9.38 and 8.45 cmolc dm -3 ), indicating that these values are related to the highest levels of SOM. Santos, Marchão, Silva, Silveira and Becquer (2011) also observed higher levels of H+Al in native forest in relation to pasture.
Regarding potassium (K), it is observed that the use of annual crops presented values significantly higher than the other uses studied (Table 6), which can be explained by the residual effect of applications according to the history of use of the area. The K contents in the 0-0.05 m layer under the annual crops and sugarcane uses are considered high, while under pasture, average levels prevail (Sousa & Lobato, 2004). At the other depths, all land-uses (except annual crops) showed average K levels, indicating that fertility management must be adjusted through cultural practices.

Conclusion
Under the conditions of the study, the use of sugarcane reduces the physical and chemical quality of the studied soil (> BD, < TP, > SRP, < SOM, < T and < P), followed by annual crops (> SRP, < Mg, < SOM and < T) and pasture (< TP and < P). There is a deep gradient for the attributes SOM, Mg, T, P and K associated with the use of sugarcane, annual crops and pasture related to soil and fertility management and the dynamics of nutrients. The non-disturbance of the soil as occurs in the use of land for pasture appears to be one of the appropriate practices in the search for maintenance of the physical and chemical quality of Oxisols in the studied watershed.