Soil organic matter of aggregates physicogenic and biogenic in areas under no-tillage system in the Cerrado, Brazil

The aim of this study was to evaluate i) the different cover crops contribution used in no-tillage system (NT) to biogenic aggregation; and ii) the influence of aggregate formation pathways on the compartmentalization and the soil organic carbon origin. Two areas managed under NT with different implementation times (6 and 18 years, NT06 and NT18, respectively) and cover crops were evaluated, totaling six sampling areas: NT06, millet (NT06MI); NT06, brachiaria (NT06BR); NT06, sunn hemp (NT06SH); NT18, millet (NT18MI); NT18, brachiaria (NT18BR); NT18, and sunn hemp (NT18SH). In each sampling area, five pseudo-replicates were collected in the 0.00-0.05 and 0.050.10 m layers. The samples were air-dried and sieved using sieves with 9.7 and 8.0 mm mesh, and the aggregates retained within this interval were selected. The percentage of each type of aggregate (physicogenic and biogenic) was quantified. Total organic carbon (TOC) and the natural abundance of δC (‰) were analyzed and the physical fractionations of SOM were performed: particulate organic carbon (POC) and mineral-associated organic carbon (MAOC) and density fractionation (free light fraction carbon, FLFC). Physicogenic aggregates were quantified in greater proportion, except for the areas of NT06BR and NT18BR in the 0.00-0.05 m layer. The biogenic aggregates Research, Society and Development, v. 10, n. 5, e39910515012, 2021 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v10i5.15012 2 showed the highest contents of TOC, POC, MAOC, FLFC and more negative values of δC. The use of grasses, especially Brachiaria spp., as cover plants in NT after 6 and 18 years of adoption favors the formation of aggregates through the biogenic pathway and they influence the compartmentalization and origin of stored organic carbon.


Introduction
Improvement in soil quality combined with decrease in production costs are characteristics associated with conservation agriculture. The use of conservation systems for soil management aims to increase the sustainability of agriculture in socioeconomic aspects, generate competitiveness for agribusiness, ensure food safety and quality, and preserve the edaphic environment. In Brazil, among these systems the most widely adopted is no-tillage (NT), mainly in areas of extensive agricultural production, such as the Cerrado.
The efficacy of NT is related, among other factors, to the quantity and quality of residues produced by cover plants and their persistence on the soil (Andrade et al., 2018). In the Cerrado, the rapid decomposition of plant residues compromises the maintenance of cover on the surface soil layer (Torres & Pereira, 2013). In this context, it has been sought to use cover plants with high phytomass production and low decomposition rate to make up crop rotation or succession schemes, so that the mineralization of soil organic matter (SOM) and nutrient cycling is slower (Boer et al., 2008;Torres et al., 2008).
Another important factor regarding the analysis of NT, besides the use of different cover crops, is to evaluate the response over the time of implementation of the system. Studies have shown that NT can alter the edaphic attributes, with the succession of crops, due to the continuous supply of organic material on soil surface, derived from plant residues, the beneficial action of the root system of the introduced plants, and the protection of soil surface (Andrade et al., 2018). For this fact, it is interesting to highlight the importance of the time of adoption of NT for the occurrence of changes, such as the accumulation of SOM, improvement of aggregation, nutrient cycling and increased biological activity.
To evaluate these factors involved in the success of NT in Cerrado areas, some edaphic attributes have been used as indicators of soil quality, including the changes in SOM aggregation and dynamics. Aggregate formation mechanisms involve physical and chemical (physicogenic aggregates) and biological (biogenic aggregates) processes (Loss et al., 2014), and the differentiation between morphological types of aggregates is performed according to their genesis or formation pathways (Bullock et al., 1985;Pulleman et al., 2005;Batista et al., 2013).
By evaluating the SOM fractions, it is possible to verify how carbon contents are distributed in the different morphological types of aggregates. From the physical point of view, the SOM compartments that are evaluated are the granulometric fraction (particulate and mineral-associated organic carbon) (Cambardella & Elliott, 1992), and density fraction (light fraction organic carbon) (Sohi et al., 2001).
To complement the study of SOM fractions, the technique of natural abundance of δ 13 C (‰) has been applied in aggregates. From this technique, it is possible to identify the origin of stored carbon that favored its formation, and these differences are due to the natural isotopic signature between plants of photosynthetic cycle C3 (-24 to -34‰) and C4 (-6 to -19‰) (Smith & Epstein, 1971).
From the above, and based on the hypothesis that different cover crops cultivated over time under conservation system of soil management can promote changes in the genesis of aggregates and in the organic matter associated with them, the aim of this study was to evaluate i) the contribution of different cover crops used in no-tillage system to biogenic aggregation; and ii) the influence of aggregate formation pathways on the compartmentalization and origin of soil organic carbon.

Methodology
The study was conducted in the city of Uberaba (19º39'10.17" S and 47º58'15.65" W), located in the state of Minas Gerais, Brazil. The sampled areas belong to the Federal Institute of the Triângulo Mineiro (IFTM) and are situated within the Cerrado biome, having as main characteristics: average altitude of 795 m; climate identified as hot tropical (Aw), according to Köppen's classification, with rainy season in summer and dry season in winter, and average annual precipitation of 1600 mm; flat to gently undulating relief; and soil classified as Latossolo Vermelho Distrófico (Oxisol), with medium texture (Santos et al., 2018).
For this study, two areas managed under no-tillage system (NT) with different implementation times were evaluated: NT area implemented 6 years ago (NT06), in transition phase; and NT area implemented 18 years ago (NT18), in the consolidation phase. In the areas under NT, three sub-areas with different cover plants were selected: pearl millet (Pennisetum glaucum L. cv. ADR500) (MI), brachiaria grass (Urochloa brizantha cv. Marandu) (BR) and sunn hemp (Crotalaria spectabilis) (SH), sampled in January 2019.
The management of cover plants is conducted in a similar way between the systems, which differ only for the time of implementation. In total, there were six sampling areas, described as: NT06, millet (NT06MI); NT06, brachiaria (NT06BR); NT06, sunn hemp (NT06SH); NT18, millet (NT18MI); NT18, brachiaria (NT18BR); and NT18, sunn hemp (NT18SH). In all subareas, the same cover crops are used in the corn/soybean succession, cultivated on the residues of these crops in relation to the previous year. Soybean (Glycine max L.) was the annual crop that preceded the time of sample collection.
Fertilization of the annual crops is carried out according to the recommendations of Ribeiro et al. (1999), that is: for corn, 400 kg ha -1 of 08-28-16 formulation at sowing (basal fertilization) with 140 kg ha -1 of N and 80 kg ha -1 of K as topdressing, split and applied at 20 and 40 days after planting; and for soybean, 200 kg ha -1 of 00-20-15 formulation + 2.5% of Zn + 2.5% of Mn at sowing (basal fertilization), corresponding respectively to 40 kg ha -1 of P2O5, 60 kg ha -1 of K2O, 5 kg ha -1 of Zn and 5 kg ha -1 of Mn, with the inoculation of the seed.
For each sampling area, five pseudo-replicates (undisturbed samples) were collected in the 0.00-0.05 and 0.05-0.10 m layers, totaling 60 sampling units. Subsequently, the samples were air-dried and subjected to sifting with a set of sieves with 9.7 and 8.0 mm mesh, using only the aggregates retained within this interval.
These aggregates were taken to the laboratory, examined under binocular magnifying glass, manually separated and identified only two morphological types: biogenic aggregates -those in which it is possible to visualize rounded forms, with the intestinal tract of soil macrofauna individuals, mainly Oligochaeta (earthworms) or those in which it is possible to visualize the presence and activity of roots; and physicogenic aggregates -those that showed angular shapes resulting from the interaction between carbon, clay, cations and soil wetting and drying cycles (Bullock et al., 1985;Pulleman et al., 2005).
After identification, the percentage, relative contribution (mass), of each morphological type of aggregate was determined for each area. Subsequently, they were pounded to break up clods and passed through a 2.0 mm mesh sieve, to obtain the air-dried fine earth (fraction for SOM analysis. Total organic carbon (TOC) contents were quantified by wet oxidation of the organic material in the presence of potassium dichromate in acidic medium (Yeomans & Bremner, 1988).
Granulometric and density physical fractionations of SOM were performed using the methods proposed by Cambardella & Elliot (1992) and Sohi et al. (2001), respectively. Organic carbon of the particulate fraction (POC) and free light fraction (FLFC) of SOM were also determined according to Yeomans & Bremner (1988), whereas the organic carbon of the mineral-associated fraction (MAOC) of SOM was quantified by the difference between TOC and POC.
The natural abundance of δ 13 C was evaluated with a Delta V Advantage mass spectrometer (Thermo Fisher Scientific, Bremen, Germany). Measurements of δ 13 C and the determination of carbon percentage were performed in the Laboratory of Research in C and N Biotransformations (LABCEN), Santa Maria -RS, Brazil.
All data for each layer were evaluated for normality of residuals and homoscedasticity by the Shapiro-Wilk and Bartlett tests, respectively. When the assumptions of the tests mentioned above were not observed, the data were transformed by the Box Cox test. Subsequently, the results were analyzed as completely randomized design in a split-plot scheme, being subjected to analysis of variance with application of F test, and mean values were compared by Tukey test at 5% probability level using the R 3.3.1 program (R Development Core Team, 2015) and the package ExpDes.pt. (Ferreira et al., 2013).

Relative contribution
Physicogenic aggregates were present in a greater proportion than biogenic ones in the areas of NT06MI, NT06SH, NT18MI and NT18SH in the two layers evaluated, except in NT06BR and NT18BR in the 0.00-0.05 m layer, where the opposite pattern was observed (Table 1). Between the areas of the NT06 and NT18 systems, it is possible to observe a reduction in the proportion of biogenic aggregates over the time of implementation in the two layers evaluated (Table 1).   Same uppercase letter above the bar indicates no difference between the sampling areas for the same type of aggregate and same lowercase letter above the bar indicates no difference between the types of aggregates for the same sampling area (Tukey, at 5% probability level). NT06MI: 6 years + millet; NT06BR: 6 years + brachiaria grass; NT06SH: 6 years + sunn hemp; NT18MI: 18 years + millet; NT18BR: 18 years + brachiaria grass; NT18SH: 18 years + sunn hemp. CV1: Coefficient of variation for types of aggregates; and CV2: Coefficient of variation for sampling areas. Source: Authors.

Total organic carbon
For TOC, differences were observed between the formation pathways, and biogenic aggregates in the areas of NT06BR, NT18SH (0.00-0.05 m) and NT06SH (0.05-0.10 m) had the highest carbon contents (Figures 1 and 2, respectively). Same uppercase letter above the bar indicates no difference between the sampling areas for the same type of aggregate and same lowercase letter above the bar indicates no difference between the types of aggregates for the same sampling area (Tukey, at 5% probability level). NT06MI: 6 years + millet; NT06BR: 6 years + brachiaria grass; NT06SH: 6 years + sunn hemp; NT18MI: 18 years + millet; NT18BR: 18 years + brachiaria grass; NT18SH: 18 years + sunn hemp. CV1: Coefficient of variation for types of aggregates; and CV2: Coefficient of variation for sampling areas. Source: Authors. Table 2 shows the results for organic carbon of SOM physical fractions. Physicogenic aggregates of NT06SH and biogenic aggregates of NT18MI, in the 0.00-0.05 layer, showed higher POC contents in the comparison between the areas, but no differences were observed in the underlying layer. In the comparison between the types of aggregates, the highest POC contents were quantified in the biogenic aggregates of the areas of NT06BR and NT18MI (0.00-0.05 m), NT18MI and NT18BR (0.05-0.10 m).

Organic carbon from particulate, mineral-associated and free light fractions of SOM
Regarding the contents of mineral-associated organic carbon (MAOC), in the comparison between the areas, the physicogenic aggregates of NT18MI and biogenic of NT18MI and NT18SH had the highest contents of this fraction in the 0.00-0.05 m layer (Table 3). For the underlying layer, higher contents of MAOC were observed in the biogenic aggregates of NT06SH (Table 3). Among the morphological types of aggregates, differences were only observed in the first layer, in which the highest contents of MAOC were quantified in the biogenic aggregates of NT18SH (Table 3).
For the organic carbon of the free light fraction (FLFC), in the comparison between the areas, no differences were verified in the layer of 0.00-0.05 m (Table 3). In the 0.05-0.10 m layer, the highest contents of FLFC were found in the two types of aggregates of NT18MI (Table 3). Among the formation pathways, in almost all areas, except NT18SH, higher contents of FLFC were quantified in biogenic aggregates, for the two layers (Table 3). Table 3 shows the values of δ1 3 C (‰) in the different types of aggregates, and differences were observed only in the 0.00-0.05 m layer. Among the areas, the least negative values of δ 13 C were found in the physicogenic aggregates of NT06MI (-   Means followed by the same uppercase letter in the column indicate no difference between the sampling areas for the same type of aggregate, and means followed by the same lowercase letter in the row indicate no difference between the types of aggregates for the same sampling area (Tukey, at 5% probability level). NT06MI: 6 years + millet; NT06BR: 6 years + brachiaria grass; NT06SH: 6 years + sunn hemp; NT18MI: 18 years + millet; NT18BR: 18 years + brachiaria grass; NT18SH: 18 years + sunn hemp; CV1: Coefficient of variation for types of aggregates; and CV2: Coefficient of variation for sampling areas. Source: Authors.

Proportion of physicogenic and biogenic aggregates
In the relative proportion of the different pathways of aggregate formation (Table 1) The machines used in NT are heavier compared to those used in conventional tillage, which can cause changes in the edaphic environment, mainly in soil structure, favoring compaction in the surface layers, which was pointed out as one of the main negative effects observed in areas of NT (Andrade et al., 2018).
In the areas NT06BR and NT18BR in the 0.00-0.05 m layer, biogenic aggregates were found in a higher proportion than physicogenic ones (Table 1) Regarding the decrease in the percentage of biogenic aggregates in the areas between the NT06 and NT18 systems (Table 1), Ferreira et al. (2020) point out that the reduction in the proportion of this class of aggregates as a function of the time of adoption of NT is an indication that some practices adopted in the system may be negatively affecting the structural quality of the soil or causing reduction in the activity of the edaphic macrofauna, by compaction, which is common in agricultural areas managed under NT.

TOC contents
The highest contents of TOC in the aggregates in the area of NT18MI in the 0.00-0.05 m layer (Figure 1) may be related to the higher phytomass production and low rate of decomposition of residues, verified for millet crop. In the municipality of Uberaba-MG, Torres et al. (2014) evaluated the phytomass production and decomposition of residues from different cover plants in NT. The authors observed that millet cultivars had the highest values of phytomass production and the lowest rates of residue decomposition. This is due to the adaptation of this crop to the climatic conditions of the study region associated with its C/N ratio, above 25:1, which promote a slower decomposition (Torres et al., 2005).
In a study evaluating TOC contents and stocks in the soil after different periods of cultivation under NT (6, 9 and 10 years), Rossetti & Centurion et al. (2015) verified These results suggest that the root system of brachiaria is more efficient in increasing the carbon content of soil aggregates when compared to that of sunn hemp. These data are not in agreement with those found in the present study for the 0.05-0.10 m layer in the NT06SH area ( Figure 2).
The results of TOC among the morphological types of aggregates (Figures 1 and 2)  action of microorganisms, plants and edaphic animals responsible for biogenic aggregation is assumed. These results suggest that these aggregates, besides indicating greater biological activity, also favor carbon accumulation in the soil.

POC contents
In the Cerrado of Minas Gerais, Pereira et al. (2012) quantified higher stocks of POC in an area cultivated with corn + sunn hemp and in the fallow area, and for the soybean crop with millet and brachiaria as cover crops, in the layer of 0.00-0.20 m. These results may have occurred due to the balance in the final C/N ratio of the mixture of the residues from the cover plants with those of the main crop, which caused the increase in POC stocks compared to areas cultivated only with grasses (millet or corn) and legumes (soybean or sunn hemp).
Increments in POC contents are generally verified in areas with legumes and grasses, since the biological nitrogen fixation (BNF) performed by legumes contributes to the growth and development of grasses, which in turn react with higher phytomass production and, consequently, greater amounts of dry matter are supplied to the soil (Pereira et al., 2012). Thus, part of this material, along the SOM decomposition process, will be transformed into POC. The results of POC in biogenic aggregates under NT18MI are in agreement with the TOC data (Figure 1), with a similar pattern between the variables in the 0.00-0.05 m layer, probably due to the characteristics inherent to the millet crop.
Evaluating seasonality in Guaíra-PR, Ferreira et al. (2020) found in the dry season higher contents of POC in physicogenic, intermediate and biogenic aggregates in NT after 23 years of implementation (maintenance phase) compared to other systems. According to the authors, these results were found because soils under NT remain with minimal disturbance and constant supply of organic residues, so aggregates are better preserved, enabling greater protection of carbon in this fraction of SOM.
Among the formation pathways, the results of POC are similar to those observed by Rossi et al. (2016) and Schultz et al. (2019). The higher contents of POC in biogenic aggregates (Table 3) indicate the predominance of material with greater lability compared to physicogenic aggregates (Loss et al., 2014), and the incorporation or maintenance of this material is favored in biogenic aggregates due to the agents responsible for its formation (soil fauna and plant roots) (Loss et al., 2014;. These results indicate that biogenic aggregates can promote greater protection of this physical fraction of SOM.

MAOC contents
For MAOC, Ferreira et al. (2020) verified in the rainy season higher contents in the forest and NT areas after 23 years of adoption in the different types of aggregates. The favorable climatic conditions associated with the high availability of food sources found in systems with longer time of adoption, contribute for the edaphic environment to start housing more complex communities of soil fauna individuals, including decomposer organisms. These organisms convert POC into CO2 and more stable forms of carbon in the soil.
Higher MAOC stocks were observed by Pereira et al. (2012) in areas of corn and soybean planted on millet and brachiaria residues, respectively, in the 0.10-0.20 m layer. According to the authors, these results demonstrate that the use of cover crops (grasses and legumes) before annual crops in the Cerrado under NT favors the increase of MAOC stocks.
Regarding the results of MAOC between the areas, the carbon contents quantified in the two types of aggregates in the areas of NT18MI in the layer of 0.00-0.05 m (Table 3) point to the presence of carbon in more recalcitrant fractions due to the longer time of adoption of NT, associated with a vegetation cover with lower rate of decomposition of its residues. For MAOC contents in biogenic aggregates in the NT06SH area, the data were similar to those of TOC in the 0.05-0.10 m layer ( Figure 2).
As for the results of MAOC among the formation pathways (Table 3), these may be related to the capacity that biogenic aggregation has for protection and physical stabilization of the more labile fractions of SOM, such as POC. The highest contents of MAOC in biogenic aggregates indicates that this class is more favored in the formation of the most stable fractions of SOM.
Other authors have also found that higher contents of MAOC were quantified in biogenic aggregates, when compared to physicogenic aggregates (Loss et al., 2014;Rossi et al., 2016;Schultz et al., 2019). Comparing the formation pathways, Ferreira et al. (2020) found differences in MAOC contents only in the rainy season, with higher values of this fraction in biogenic aggregates. However, MAOC does not always function as a good indicator in the measurement of soil quality, since changes in the contents of this SOM compartment require many years to be observed (Carmo et al., 2012), due to the high degree of stability of this physical fraction of SOM.

FLFC contents
The labile fractions of SOM are fundamental for cycling of carbon between compartments and of nutrients in a short period of time, in addition to their contribution to the formation and stabilization of soil aggregates (Santos et al., 2013).
Evaluating the contents of light organic matter (LOM) in aggregates collected in areas of NT after 17 years of implementation, with and without crop-livestock integration (NT-CLI and NT without CLI, respectively) and Cerrado area, Loss et al. (2011) found MOL contents similar to those of the Cerrado area in the 0.05-0.10 m layer in the NT-CLI area. This demonstrates that the CLI contributed to the supply of root residues in subsurface and the renewal of the root system, so the supply was equivalent to that of litter produced in the Cerrado area and was more efficient than that of the area under NT without CLI in increasing the LOCM contents. These results are similar to those observed in this study for FLFC in subsurface (Table 3).
Among the morphological types of aggregates, the results of FLFC in biogenic aggregates corroborate the data of POC (Table 3), probably due to the greater lability in both fractions. The similarity of the organic structures present in the particulate organic matter and in the light fraction of SOM was confirmed by Conceição et al. (2007) through optical microscopy analysis. The authors demonstrated that both fractions are basically composed of root fragments and organisms of soil fauna, hyphae and decomposing plant residues. In this study, the results of FLFC and POC may be associated with the potential of the formation pathways for carbon sequestration, in which the type of carbon stored by biogenic aggregates is more soluble than the more recalcitrant carbon stored in physicogenic aggregates (Pinto et al., 2019).

Natural abundance of δ 13 C
In the aggregates of all areas, δ 13 C contents ranged from -15.55 to -17.88‰, indicating predominance of carbon originated from plants with photosynthetic cycle C4 (species that fix CO2 through the PEPCase enzyme), with values between -6 and -19‰ (Table 3).
In areas of NT with and without crop-livestock integration (NT-CLI and NT without CLI, respectively), Loss et al. corroborating the values observed in the present study (Table 3). The use of C4 plants, such as corn, in rotation with soybean, and millet or brachiaria as cover crops, contributes to the formation and stabilization of aggregates, especially for areas with brachiaria (Loss et al., 2011).
In the comparison between the formation pathways, in the biogenic aggregates of the NT06BR and NT18SH areas, the most negative values of δ 13 C were verified in the 0.00-0.05 m layer (Table 3). Loss et al. (2017) in Santa Catarina, Brazil, found no differences in δ 13 C contents between biogenic and physicogenic aggregates. However, Loss et al. (2014) quantified more negative values in biogenic aggregates in areas of NT and secondary forest. The authors related these results to the high contents of TOC in these aggregates and to the isotopic signature of the predominant vegetation in these areas. The data of δ 13 C in biogenic aggregates in this study are similar to those observed by Loss et al. (2014) and Jouquet et al. (2009).

Conclusion
The use of grasses, especially Brachiaria spp., as cover plants in no-tillage system after 6 and 18 years of adoption favors the formation of aggregates through the biogenic pathway. Both compartmentalization and origin of the stored organic carbon are influenced by the aggregate formation pathways.
Highest contents of TOC, POC, MAOC, FLFC and more negative values of δ 13 C were found in biogenic aggregates.
Mainly in areas covered with millet and sunn hemp under systems in consolidation and transition phases, respectively.
Biogenic aggregates were more influenced and benefited by the stabilization mechanisms of the different physical fractions of SOM.