INTRODUCTION:

In the state of Rio Grande do Sul (RS), lowland soils, with about 20% of the total area, cover 71% of rice (Oryza sativa L.) production in Brazil (^{CONAB, 2017}). Rice production with flooding is predominant in these areas (^{SOSBAI, 2016}). However, the rice monoculture for long time favors soil degradation, increasing pathogens population, and weed infestations (^{MARCHESAN et al., 2004}). In these cases, crop rotation with sorghum (Sorghum bicolor) or soybean (Glycine max) is recommended to breakdown the pests and diseases cycles (^{SILVA & PARFITT, 2004}). Compared to rice residues, soybean and sorghum cultivation in these environments will result in the addition of crop residues with distinct chemical composition e.g. C/N ratio and the soluble organic fraction content. In this context, ^{REDIN et al. (2014}) verified that grasses produce about three times more root biomass than legumes. Additionally, reported the presence of larger amounts of cellulose, hemicellulose and lignin for sorghum residues than soybean residues. Therefore, the quality of crop residues in rotation influences the decomposition and availability of nutrients in the soil (^{CHIVENGE et al., 2011}; ^{AITA & GIACOMINI, 2003}). In short term, it is expected that the addition of these residues with different characteristics will cause changes in the intensity of mineralization-immobilization process and, in the long term, alter the compartments of organic N in the soil.

Soil type also has influence on the N dynamics along with the residue’s quality. Clayey soils often contain more amounts of soil organic matter (SOM) (KLEBER et al., 2007) and microbial biomass (^{SIX et al., 2006}). This is because clayey soils tend to have lower C transformation rates than sandy soils due to higher organic-complexation by promoting physical protection against microbial attack (^{CHIVENGE et al., 2011}). Therefore, it is assumed that crop rotation, depending on soil characteristics, changes the SOM dynamic and consequently, the soil organic N content. The SOM and the organic N can be studied by fractionating in order to isolate fractions of different chemical nature and/or susceptibility to decomposition. Among the different methods, the physical fractionation (densimetric, granulometric or adaptation of both) have been used to understand the SOM dynamics when associated with the soil mineral fraction, and could also be used as a strategy to understand the organic N dynamics (^{ROSCOE & MACHADO, 2002}; ^{CHRISTENSEN, 2001}).

Although a number of different procedures may be used, one is known for some operational simplicity, used by ^{BALESDENT et al. (1991}) with adaptations of ^{DIOCHON et al. (2016}) that allows to isolate three very contrasting fractions: >250 µm, called coarse sand fraction (CSF) primarily consists of organic residues in the earlier stages of decomposition (^{CHRISTENSEN, 2001}); 53-250 µm, fine sand fraction (FSF) consists the partially decomposed plant material; and <53 µm, called silt+clay fraction (SCF), which is the most stable SOM fraction being constituted by organo-complexes with silt and clay particles (^{CHRISTENSEN, 1996}).

In crop rotation systems with flooded rice, it is essential to consider the crop residues decomposition in aerobic and anaerobic cycles. According to findings of ^{LUCE et al. (2014}), the relationship between quality of residues and its interaction with soil mineral fraction, it is expected to be higher initial increment of N in the soil with high quality residues, such as soybean, stabilized as SOM fractions associated with smaller particles. However, as this fraction also contains mineral N adsorbed, it is possible that this form of N is more easily lost due to anaerobic conditions in a next flooding cycle for rice cultivation.

This study aimed to evaluate the N present in the physical fractions of SOM after addition of rice, sorghum, and soybeans residues, using labeled ¹⁵N in two soils with different textures.

MATERIALS AND METHODS:

In this study, samples were used from an incubation experiment carried out at greenhouse of the Department of Soil Science at Federal University of Santa Maria (UFSM). We evaluated six treatments formed by the combination of three residues (root + shoot) plants (rice, sorghum, and soybeans) enriched with ¹⁵N isotope and two soil types (Planossolo Háplico Eutrófico gleissólico - SXe_glei and Planossolo Háplico Eutrófico arênico - SXe_aren) (^{EMBRAPA, 2013}). Both soils collected up to 0 to 20 cm for experiment and differ in their particle size distribution: SXe_glei with 173, 287 and 540 g kg^-1 of clay, silt, and sand, respectively, having 1.8 times more silt+clay than SXe_aren with 99, 155 and 746 g kg^-1 of clay, silt, and sand, respectively. A factorial experiment (three types of residues × two types of soils) with complete randomized design (CRD) was used with three replicates per treatment. A control treatment (soil without addition of residues) was also installed to calculate the isotopic abundance.

The experimental units were mounted in PVC tubes with 50 mm diameter and 140 mm height. Each cylinder was individually wrapped with tissue type “voile” at the bottom in order to allow water percolation and filled with 235.6 g of soil, resulting in a layer of 10 cm and medium density 1.15 g cm³. In all treatments, 1.38 g (7 Mg ha^-1) of shoot (stems + leaves) and 0.39 g (2 Mg ha^-1) of root residues were added. The N contents measured in residues were 9.95, 6.7, and 16 g kg^-1 while calculated C/N was 41.2, 61.3, and 26.8 for rice, sorghum, and soybean, respectively.

The soils were kept under free drainage system up to180 days with irrigation using distilled water every two weeks simulating a rain of 20 mm. After 180 days, the cylinders were sealed and soils were maintained under flooding conditions until 360 days. Three cylinders from each treatment were sampled at 30, 180 and 360 days of incubation. At each evaluation date the remaining residues and roots were collected from the soil surface and separated manually from the soil. Then, the residues and the soil were dried in an oven at 60 °C for dry matter determination and subsequently finely ground and analyzed for total N content by an elemental analyzer (Flash EA 1112, Thermo Finnigan, Milan, Italy) and ¹⁵N isotope through an isotope ratio mass spectrometer (DELTA V Advantage, Thermo Fisher Scientific, Bremen, Germany) coupled with the elemental analyzer.

Another subsample was carried out for soil particle size by physical fractionation described ^{BALESDENT et al. (1991}) with adaptations according to ^{DIOCHON et al. (2016}). Three fractions were isolated: >250 µm, called coarse sand (CSF) mainly consisted of organic residues in the earlier stages of decomposition; 53-250 µm, called fine sand (FSF) that consisted of primarily materials at an intermediate stage of decomposition; and <53 µm, called silt+clay fraction (SCF) which corresponded to organic matter associated with the minerals; however, due to the aqueous method, in the later fraction (<53 µm) would also be mineral N, but not strongly protected in the soil. Fractions were dried at 60°C, weighed and homogenized, and then milled in a mortar to determine the total N content and ¹⁵N isotopic abundance, as previously mentioned for residues and soil.

From the results, ¹⁵N was calculated from remaining residues (shoot + roots) and soil from each sampling date using following equation:

Nrem %= Nrem15Nad15 ×10015

Where,¹⁵Nrem% represents the percentage of ¹⁵N remaining in the residues and the soil,¹⁵Nrem is the amount of ¹⁵N (mg kg^-1) remaining in residues and in soil, and Nad the amount of ¹⁵N added with crop residues (mg kg^-1). The percentage of ¹⁵N residue arising in each fraction was obtained by the following equation:

%Nfr= Nfr15Nrem soil15 ×10015

Where, %¹⁵Nfr is the percentage of ¹⁵N calculated for each fraction,¹⁵Nfr is the quantity of ¹⁵N in each fraction (mg kg^-1), and ¹⁵Nrem soil is the amount of ¹⁵N in the soil (mg kg^-1).

Normality assumptions were checked for errors and homogeneity of residual variance through the Kolmogorov-Smirnov and Levene, respectively. The analysis of variance (ANOVA) considering 2×3 factors (soils and crop residues) from the data ¹⁵N remaining crop residues, soil, CSF, FSF and SCF fraction at each evaluation time (30, 180 and 360 days after ¹⁵N application) was performed. Means were compared using the Tukey test at 5% significance level using Microsoft Office Excel application and Sisvar software (^{FERREIRA, 2014}).

RESULTS AND DISCUSSION:

All variables met the errors normality assumptions and homogeneity of residual variance by the Kolmogorov-Smirnov test (p-value ≥ 0.370) and Levene (p-values ≥ 0.193), respectively, lending credibility to the results for ANOVA and comparisons means using the Tukey test.

The amount of remaining ¹⁵N in crop residue (shoot + roots) decreased throughout the incubation period and was significantly influenced only by the type of residue in all evaluations, with no soil type effect (Table 1). Though, ^{LUCE et al. (2014}) have reported effects of two soils having different textures on remaining ¹⁵N. Residues had a greater mineralization in the sandy soil, while type of soil had no influence on remaining ¹⁵N. Absence of effect of soil on the N release from residues may be due to the fact that, in this research, the residues were deposited on the soil surface instead of incorporation, showing that little contact between residues and soil reduces or cancels their influence on the decomposition of residues (^{CARVALHO et al., 2008}).

At 30 days the amount of remaining ¹⁵N was higher in the rice residues (77.0%) and almost half of ¹⁵N was no longer in soybean and sorghum residues (Table 1). This result should be directly related to the difference in chemical composition (quality) of crop residues. COPPENS (2006) relates the quick N release after adding straw to the soil by releasing soluble and readily degradable substances containing N. According ^{REDIN et al. (2014}), the high content of soluble fraction in soybean residues, particularly in leaves, should have favored the microbial decomposing activity. The effect of the quality of the residues was also observed after 180 days of incubation, and the remaining ¹⁵N amount in the soybean residues, was approximately half (26, 0%) in the average of the two soils than that observed in the rice and sorghum residues (Table 1). After 360 days of incubation, the highest quantity of ¹⁵N remaining was quantified for sorghum residues. According to REDIN et al. (2014), this fact is possibly related to a predominance of more recalcitrant fractions with high levels of lignin and insoluble complex which remained in the grass residue, particularly in stems, prohibiting its decomposition in the later stages than rice residues.

Regarding the amount of ¹⁵N transferred to the soil from the residues, there was isolated effect of crop residues after 30 days of incubation with the soybean crop residue showed higher amount of ¹⁵N in soil (Table 2). The result approves the findings obtained by ^{LUCE et al. (2014}), which showed the chemical quality of residues is an important factor in the N release to the soil at the initial stage of the crop residues decomposition. At 180 days of incubation there was great effect of soil and crop residues, and the soil with higher amounts of clay (SXe_glei) and soybean was one with the highest amount of ¹⁵N (Table 2).

At 360 days of incubation, there was interaction between soil and residues. Rice residues provided significantly higher amounts of ¹⁵N in the soil (SXe_glei) compared to soybean and sorghum; however, the residue did not differ in the sandy soil (SXe_aren). It is possible that this soil treatment with soybean, whose decomposition was more intense in the first 180 days of incubation, the N from residues have been converted to nitrate and; therefore, lost for denitrification during the flooding period, even if the residues continued to be decomposed during this period. Flooded soil condition, in the absence of oxygen, bacteria utilize inorganic compounds oxidized of N instead of O₂ as a final electron acceptor (^{MOREIRA & SIQUEIRA, 2006}). The nitrate is the N form which serves as the electron receiver replacing oxygen (^{PONNAMPERUMA, 1972}), favoring denitrification to N₂O or N₂, escaping to the atmosphere being volatile in nature.

It is also possible to observe that treatment with rice residues, which had decomposed least in the first 180 days, followed releasing N in flooded conditions. However, in the anaerobic cycle, the N may have be kept in the ammonium form, without significant losses in the flooding period, which may explain the N amount increased in the SXe_glei soil than the same soil with soybean at the end of the 360 days of incubation (Table 2). The O₂ absence during flooded conditions stops the nitrification process, favoring the NH₄ ⁺ accumulation in flooded soils that can be retained in the soil CEC, absorbed by plants, immobilized biomass, and/or suffer nitrification under aerobic regions (^{MOREIRA & SIQUEIRA, 2006}).

When comparing the soils, the soil with higher amount of clay+silt provided better N preservation from released by the residues at 180 and 360 days of incubation (Table 2), which should be attributed to its clay reactivity characteristics such as the adsorption and retention of elements capacity (KLEBER et al., 2007). The SOM can be, physically and chemically stabilized by adsorption on the surface of clay minerals (^{TISDALL & OADES, 1982}). Several different types of functional groups of organic substances can react with soil minerals, forming organo-mineral complexes of various degrees of stability against biological and chemical occurrence (^{ALLISON, 1973}; ^{CANELLAS et al., 2008}). In this regard, the physical protection may also result in the encapsulation of the SOM in soil aggregates or its coating particles of silt and clay (^{SKJEMSTAD et al., 1993}). Thus, the physical protection consists in forming a barrier that prevents or hinders access of the SOM decomposing microorganisms,

In general, there was added ¹⁵N to the soil in three size fractions during the incubation period (Table 3). At 30 days, the soybean residues had the highest contributions to ¹⁵N in the CSF and SCF, with no difference between the crop residues on the FSF. The highest amount of ¹⁵N reported in soybean CSF in the first 30 days of incubation may be associated with increased degradation of the more soluble carbon (and hence organic N). For ^{REINERTSEN et al. (1984}), ^{CHRISTENSEN (1985}) and ^{AITA & GIACOMINI (2003}), the high initial rate of decomposition of crop residues is due to the easy way as organic compounds, especially carbohydrates water-soluble fraction, are used as an energy source of microbial population. According to ^{CHIVENGE et al., (2011}), the decomposition rate of crop residues and the formation of organic material fractions is mediated by microorganisms, succeeding the mineralization, being controlled by the quality of the crop residues (C/N ratio of the residue) and soil texture.

High quantities of N in CSF were obtained in the sandy soil and are in accordance with findings of ^{LUCE et al., (2014}), larger amounts of N in CSF were quantified in sandy soil than in the clay, where occurs most frequent microbial attack on organic material and hence, a larger amount of this fraction is displayed on sandy soil. This is because the CSF includes particulate matter partially decomposed vegetable residues along with microbial by-products that are great sources of N to soil microorganisms (^{GREGORICH et al., 2006}).

The CSF showed increased values of added ¹⁵N for both soils under incubation, with an interaction between soil and residue at 180 days of incubation (Table 3). According ^{ROSCOE & MACHADO (2002}), CSF is only a small part of the total mass of the soil, but can store a significant portion of the total C and N of soil. According to these authors, it is possible to enlighten the constant ¹⁵N increase over the incubation period in which the ¹⁵N contained in the plant material is quantified in this fraction of coarser materials of SOM.

The SCF is the portion that best quantified ¹⁵N in both soils and higher relative proportion among the fractions (Table 3). According to ^{LUCE et al. (2014}), the addition of crop residues stimulates microbial activity which may result in an initial increase in the amounts of ¹⁵N microbial biomass and mineral forms, which will be quantified in the SCF (^{BALESDENT et al., 1991}). This behavior is evidenced in ^{BLANCO-MOURE et al. (2016}), who reported that stable organic fraction constituted the majority of soil organic carbon (80%) and were not significantly affected by soil management. In this paper, the authors emphasized the role of clay in chemical stabilization through complex between minerals and organic substances and the physical protection of the main C preservation mechanism in soils.

Between soils, the SXe_glei presented more ¹⁵N addition for rice and sorghum residues after 180 days of incubation and rice residues at 360 days of incubation (Table 3). These data confirm ^{LUCE et al., (2014}), which postulates that the organic matter fractionation elucidated the fate of N derived from plant residues decomposition and element association with mineral soil matrix. Considering the data obtained in this study and in order that the nitrogen fertilization is based on the amount of soil organic matter, care must be taken when adding crop residues with high chemical quality in lowland soil in rice rotation with species of upland, because the N mineralized can be lost to the environment after the new flooding cycle for irrigated rice.

CONCLUSION:

In the first 180 days of incubation, there was higher amount of¹⁵N in soil derived from soybean crop residue decomposition. However, at the end of the 360 days of incubation, the highest amount of¹⁵N added was verified at treatments with rice crop residues application. The soil with highest silt plus clay content presented greater N retention capacity and the SOM fraction associated with smaller particle size was the main pool of the N derived from crop residues after 360 days of incubation.