Soil aggregation according to the dynamics of carbon and nitrogen in soil under different cropping systems

The objective of this work was to evaluate total soil carbon and nitrogen, as well as their contents in particulate and mineral-associated C fractions; to determine C stock and sequestration rates in the soil; and to verify the effect of C and N contents on soil aggregation, using different crop rotations and crop sequences under no-tillage. The study was carried out for nine years in a clayey Oxisol. The treatments consisted of different cropping systems formed by the combination of three summer crops (cropped until March) – corn (Zea mays) monocropping, soybean (Glycine max) monocropping, and soybean/corn rotation – and seven second crops (crop successions). Soil samples were taken at the 0.00–0.10-m layer for physical fractionation of C and N, and to determine soil aggregation by the wet method. Soybean monocropping increased C and N in particulate C fraction, while the crop systems with corn monocropping x pigeon pea (Cajanus cajan), corn monocropping x sun hemp (Crotalaria juncea), and soybean monocropping x corn as a crop succession increased total C in the soil. Greater rates of soil C sequestration were observed with soybean/corn rotation and with soybean monocropping, as well as with sun hemp as a second crop. The increase in total N increases soil C stock. Soil aggregation was most affected at particulate C fraction. Increases in soil N promote C addition to particulate fraction and enhance soil aggregation.


Introduction
Agricultural soils of the tropical regions contain less soil carbon than their capacity to store, because of the high soil organic matter mineralization (Lal, 2005).In these regions, Oxisols predominate, with up to 150 Pg C stored in the topsoil layer of 0.20 m (Lal, 2012).Therefore, the evaluation of the impact of crop sequences on carbon and nitrogen dynamics in Oxisols Pesq.agropec.bras., Brasília, v.51, n.9, p.1652-1659, set.2016 DOI: 10.1590/S0100-204X2016000900065 could be a means to mitigate CO 2 and N 2 O emissions, with repercussions on climate and on soil quality (Sá & Lal, 2009).The no-tillage system is a conservationist practice that can minimize losses and increase carbon in tropical soils, from 0.35 Mg ha -1 per year (Bayer et al., 2006) to 1.30 Mg ha -1 per year, which can be intensified with crop rotation (Sá et al., 2013).Soil aggregation and carbon and nitrogen contents are attributes very sensible to soil management and also affected by crop type.Legumes are often used in conservationist soil management systems due to their capacity to biologically fix high quantities of nitrogen and can contribute to soil carbon addition at rates of 0.88 Mg ha -1 per year (Diekow et al., 2005).These additions are commonly linked to increases in soil aggregation, which, in turn, protect soil carbon against microbial decomposition (Tisdall & Oades, 1982;Conceição et al., 2013).Grasses are mainly used in conservationist systems due to their high aboveground biomass yield, with greater C:N ratio (Marcelo et al., 2012a), and to their dense root system, which is associated with intense microbial activity.These croptype features can increase soil carbon at a rate of 0.71 Mg ha -1 per year (Calegari et al., 2008;Martins et al., 2012a).
However, the effects of crop rotation systems and of crop sequences are far from being completely understood.This is unfortunate, since soil conservation should not be treated exclusively as a function of the effects of legumes and grasses on its C and N dynamics (Silva et al., 2010).
Soil aggregate stability and diameter are closely related with soil carbon and nitrogen contents (Tisdall & Oades, 1982;Conceição et al., 2013).Moreover, the physical fractionation of soil organic matter into particulate carbon and mineral-associated carbon may greatly contribute to the understanding of C and N dynamics and of soil aggregation (Diekow et al., 2005).Carbon in particulate fraction represents the lowest carbon stock in Oxisols, although it is the most active fraction and is highly susceptible to soil management and cropping systems (Diekow et al., 2005;Sá & Lal, 2009;Conceição et al., 2013), whereas mineral-associated carbon fraction is the greatest and most stable carbon stock in the soil (Lal, 2005).
The great chemical affinity of Fe and Al oxyhydroxides with soil carbon optimizes aggregation in Oxisols, with the formation of organomineral aggregates (Tisdall & Oades, 1982).The union of these highly stable aggregates originates microaggregates, with diameter size smaller than 250 µm, commonly associated with carbon sequestration in Oxisols (Tisdall & Oades, 1982;Conceição et al., 2013).Furthermore, microaggregates are held together by roots, fungi hyphae, and polysaccharides to form macroaggregates with diameter size greater than 250 μm, which are less stable and more sensible to changes in land use and soil management (Tisdall & Oades, 1982).These macroaggregates can also be formed around particulate carbon, protecting it against microbial degradation (Lal, 2005).However, inside macroaggregates, particulate carbon may be found fragmented into smaller particles, which bind themselves to soil mineral particles to form microaggregates (Tisdall & Oades, 1982).
The objective of this work was to evaluate total soil carbon and nitrogen, as well as their contents in particulate and mineral-associated C fractions; to determine C stock and sequestration rates in the soil; and to verify the effect of C and N contents on soil aggregation, using different crop rotations and crop sequences under a no-tillage system.

Materials and Methods
The study was carried out in the municipality of Jaboticabal, in the state of São Paulo, Brazil (21°14'S, 48°17'W, at 550 m of altitude).The experimental site is in a tropical/megathermal zone, with Aw climate, according to Köppen's classification, with dry and warm winters, and wet and hot summers.The mean annual temperature is 22°C, with 1,430 mm annual rainfall, considering an average of 30 years.The soil of the experimental area is a clayey Latossolo Vermelho eutrófico (Oxisol) according to the Brazilian soil classification (Santos et al., 2013).At the 0.00-0.10-msoil layer, the contents of clay, silt, and sand were, respectively, 556, 63, and 381 g kg -1 .Some mineralogical properties of this soil have been previously described in Martins et al. (2009).
Before the experiment, in 2002, the soil was prepared with subsoiling down to 0.40-m depth; and plowing and disking were used to incorporate 1.5 Mg ha -1 of lime to the soil.In June 2005, 1.0 Mg ha -1 lime was applied to soil surface with no incorporation, in order to increase base saturation up to 65% (Marcelo et al., 2009).
The experiment was conducted using a randomized complete block design, with strip plots and three replicates.Each block had 21 plots, consisting of three summer cropping systems combined with seven crop sequences (Figure 1).The summer crops were grown between November and March, and were repeated every year in the same plots.They formed the following cropping systems: corn (Zea mays L.) monocropping; soybean [Glycine max (L.) Merr.] monocropping; and soybean/corn rotation.Second crops were sown in March and were also repeated every year in the same plots, consisting of the following crop sequences (second crops): corn, sunflower (Helianthus annuus L.), oilseed radish (Raphanus sativus L.), pearl millet [Pennisetum americanum (L.) K.Schum], pigeon pea [Cajanus cajan (L.) Millsp.],grain sorghum [Sorghum bicolor (L.) Moench], and sun hemp (Crotalaria juncea L.).
Soil fertilization for all second crops consisted of 200 kg ha -1 N-P 2 O 5 -K 2 O 8-20-20, applied at sowing.For summer crops, the base fertilization was performed according to the soil chemical attributes, determined on the soil test of the previous growing season, which can be found in Marcelo et al. (2009).
Soil sampling was done after nine years from the experiment establishment, in October 2011, before sowing the summer crops of the next growing season (2011/2012).In each plot, three undisturbed soil samples (5.0-cm diameter x 5.0-cm height) were taken at the 0.00-0.10-msoil layer.Additionally, 20 disturbed soil samples were taken randomly from each plot, to perform a composite sample, at the 0.00-0.10-mlayer.Each composite sample was divided into two portions: the first was used to select soil aggregates with diameter between 6.30 and 4.00 mm, with the same moisture content in the field; and the second portion was air-dried during 48 hours and ground to pass through a 2-mm sieve.
To determine soil aggregate mean weight diameter, a 25-g portion of soil aggregates, with diameters of 6.30-4.00mm, was transferred to a set of sieves in decreasing order of mesh sizes: 4.00, 2.00, 1.00, 0.50, 0.25, and 0.125 mm.The sieve set was directly immersed -with the aggregates without pre-wettingin water, in an apparatus for vertical oscillation (Yoder, 1936), during 15 min, adjusted to 31 cycles per min.
From the second portion, 30 g were used for a particle-size fractionation.This sample was placed in a 200-mL plastic bottle with a sodium hexametaphosphate solution (5 g L -1 ), and three agate balls (5-mm diameter) were used to improve soil mechanical dispersion.The content was shook overnight, for 16 hours, in a horizontal shaker, at the frequency of 50 rpm.The sample was passed through a 250-µm sieve placed above another one of 53 µm.Fractions were selected as in Koutika et al. (2001): from 2,000 to 250 µm, medium particulate carbon; from 250 to 53 µm, fine particulate carbon; and <53 µm, mineral-associated carbon.The remaining material in each class was washed, ovendried at 50ºC, weighed, and grounded to pass a 105-µm mesh sieve, for carbon and nitrogen determination.Another 4 g of the bulk soil were oven-dried at 50ºC and ground to pass the same 105-µm mesh sieve.Portions between 0.275 to 0.300 g, from each sample, were weighed to determine total carbon and total nitrogen concentrations, using dry combustion at 900°C in a carbon and nitrogen analyzer.Soil density was determined (Grossman & Reinsch, 2002) to calculate total carbon stock by equivalent soil mass (Ellert & Bettany, 1995).In the present study, native forest soil properties, such as total carbon concentration (30.16 g kg -1 ) and soil bulk density (0.92 kg dm -3 ), were used as a baseline to analyze total carbon stocks in the treatments.In addition, the rate of carbon sequestration was determined by the difference between actual C contents minus the C content of 11.02 g kg -1 -obtained before the beginning of the experiment in 2002 -, divided by nine years.
The obtained data were subjected to the analysis of variance using the F-test, and means were compared by Tukey's test, at 5% probability.Also, Pearson's correlation was used to evaluate the relationship of total nitrogen and aggregate mean weight diameter with total carbon and carbon contents at medium particulate, fine particulate, and mineral-associated fractions.

Results and Discussion
Soybean monocropping increased soil carbon and nitrogen contents in the medium particulate carbon fraction, when compared with corn monocropping (Table 1).However, neither summer crops nor second crops affected the contents of carbon and nitrogen in fine particulate and mineral-associated fractions.The increases in carbon and nitrogen by soybean monocropping may be attributed to the low C:N ratio of it residues in comparison with those of corn, which contributes to quickly adding carbon and nitrogen into the medium particulate fraction.The continuous input of crop residues on the soil, under no-tillage, is essential to carbon addition in particulate carbon fraction, which is composed of fresh residues (Sá & Lal, 2009).
Within monocropping, carbon and nitrogen contents showed the lowest values with the second crops corn, sorghum, and sunflower (Table 2).This may Table 1.Soil contents of carbon and nitrogen, as well as C:N ratio, determined in different carbon fractions and in bulk soil, rate of carbon sequestration, and mean weight diameter (MWD) of an Oxisol under three cropping systems combined with seven crop sequences, under no-tillage (1) .

Fine particulate carbon fraction
Mineral-associated carbon Bulk soil Rate of C sequestration  exporting, respectively, 40.85, 44.90, and 37.10 kg ha -1 N (Marcelo et al., 2012a(Marcelo et al., , 2012b)).Besides, pigeon pea and sun hemp fix, annually, 23.5 and 19.5 g kg -1 N, respectively (Marcelo et al., 2012a), which may contribute to the accumulation of soil carbon and nitrogen (Bayer & Mielniczuk, 2008).According to Seo et al. (2006), crops recover 15% of the total N from labeled legume residue, in the first year; and another 55% are recovered from the soil organic N fraction.
Within soybean monocropping, soil under corn, as a crop sequence, had its carbon content increased, in comparison with soil under oilseed radish (Table 2).Plots cultivated with corn as a second crop had higher plant biomass on soil surface (2.7 Mg ha -1 ), with higher nitrogen content (13.10 Mg ha -1 N per year) than that with oilseed radish (0.3 and 0.9 Mg ha -1 N per year, respectively).Considering that the highest NO 3 - contents from oilseed radish residues become available in the first 54 days after plant harvesting (Heinzmann, 1985), this crop may not contribute significantly to nitrogen availability for the summer crop, contrary to what has been observed with corn, which can release NO 3 -during soybean growth.
In general, soil cultivated with corn as a second crop had higher carbon and nitrogen contents, and, among summer cropping systems, soybean monocropping raised C and N, when compared with corn monocropping, at the 0.00-0.10-msoil layer (Table 2).Amado et al. (2006) reported that the soybean-wheat (Triticum aestivum L.) succession provided total organic carbon 46% higher than corn-black oat (Avena strigosa Schreb.),though both successions produced the same quantity of plant residues, probably due to the rapid decomposition of soybean residues, which may have boosted wheat dry matter productivity (Weber & Mielniczuk, 2009).It should be noted that nitrogen availability by legumes cultivated in summer can exceed the needs of plants cultivated as second crops (Weber & Mielniczuk, 2009).
Mineral-associated carbon fraction, followed by fine particulate and medium particulate fractions, had higher total carbon and nitrogen (Figure 2).The highest carbon and nitrogen content verified in the more stable fraction, i.e., mineral-associated carbon, indicates a Table 2. Carbon and nitrogen contents in bulk soil, under three cropping systems combined with seven crop sequences, under no-tillage (1) .
Second crops (W) (2)  Summer crops (S) strong organomineral relationship (Bayer et al., 2009), which is important for soil carbon sequestration.Soybean/corn rotation and soybean monocropping were the cropping systems with the highest rate of carbon sequestration, especially with sun hemp as a second crop (Table 1).These results emphasize the importance of the input of large amounts of nitrogen by legume crops for a rapid increase in particulate carbon.Martins et al. (2012b) reported that this kind of input contributes to carbon sequestration in Oxisols in tropical regions.The benefits of sun hemp in the accumulation of soil carbon may be attributed to its low C:N ratio, which favors increasing nitrogen availability, necessary for rapid residue processing into the particulate soil carbon fractions.The C:N ratio of sun hemp residues is equal to 23, with the following amounts of nitrogen accumulated, released, and remaining: 110.1, 89, and 21.1 g kg -1 , respectively (Marcelo et al., 2012b).The average rate of carbon sequestration for the crop sequences in the present study was of 0.63 Mg ha -1 per year, higher than the average of 0.35 Mg ha -1 per year found by Bayer et al. (2006), and lower than the one of 0.89 Mg ha -1 C per year reported by Sá et al. (2013).
Corn monocropping with sorghum as a second crop increased the mean weight diameter of soil aggregates (Table 1).Physically, the roots of grasses contribute to soil aggregation by approaching soil mineral particles.
Chemically, they contribute by continuously releasing exudates into the soil.Moreover, the root system has high density and is constantly renovated, which contributes to the formation of macroaggregates (Six et al., 2002).These results agree with previous studies in the same experimental area, where grasses increased soil aggregation by increasing the contents of carbon and easily hydrolysable polysaccharides (Martins et al., 2009).Indirectly, the addition of larger amounts of polysaccharides and pentoses by grasses, particularly of arabinoses and xyloses, serve as a source of energy for microorganisms that act on soil aggregation (Martins et al., 2012a(Martins et al., , 2013)).
The mean weight diameter showed positive correlation (r=0.32) with carbon content in the fine particulate carbon fraction (Figure 3 A).Although this fraction is commonly associated with microaggregation in Oxisols, the rather low correlation observed in the present study indicates that the high contents of oxides and hydroxides of Al and Fe in these soils have a greater role in the formation of microaggregates, which is crucial for fixing soil carbon (Six et al., 2002).The decomposition of particulate carbon produces more recalcitrant compounds, which act in the formation of microaggregates and stabilize carbon in the soil (Tisdall & Oades, 1982;Conceição et al., 2013).
Soil total carbon was positively correlated (r=0.94) with total nitrogen (Figure 3 B).The addition of nitrogen favored the accumulation of carbon, probably due to increasing microbial activity and biomass production by the subsequent crops.Nitrogen added by legumes is more effective to increase soil carbon than that added by mineral fertilizers, which favors the productivity of the following crops (Bayer et al., 2009;Conceição et al., 2013).
Total carbon was also positively correlated with the carbon content in the medium particulate (r=0.41) and fine particulate fractions (r=0.36), as shown in Figure 2 C and D. The correlation between particulate carbon and total carbon can be explained by the continuous addition of crop residue on soil surface.Carbon increases in soils under no-tillage occur, preferentially, in the particulate fraction, which is protected inside the soil aggregates.Previous studies in the same site reported a strong correlation (r=0.90) between particulate carbon and total carbon (Martins et al., 2012b).

Figure 2 .
Figure 2. Distribution of carbon and nitrogen in mineral-associated carbon fraction (MAC), as well as in fine (FPC) and medium particulate carbon fractions (MPC).Means followed by equal letters do not differ by Tukey's test, at 5% probability.

Figure 3 .
Figure 3. Regression analysis between: A, carbon content in fine particulate carbon fraction (POC-F, 250-53 µm) and mean weight diameter; B, total carbon and total nitrogen; C, total carbon and carbon content in medium particulate carbon fraction (POC-M, 2,000-250 µm); and D, total carbon and carbon content in POC-F (n = 63).