Nitrogen , phosphorus , and potassium released by decomposition of palisade grass to soybean in succession

The objective of this work was to determine the nitrogen, phosphorus, and potassium released by N-fertilized Urochloa brizantha straw, desiccated before sowing of no-tillage soybean (Glycine max) in succession. The pasture was grown in three paddocks, each fertilized with one N rate (0, 150, or 300 kg ha-1), and assessed at two desiccation periods (60 or 15 days before soybean sowing), in four replicates. Nitrogen fertilization of U. brizantha pasture increases N and K cycling and the release of these nutrients to the subsequent soybean crop. The quantity of K released by U. brizantha straw fertilized with 150 or 300 kg ha-1 N (about 80 kg ha-1 K) offsets that contained in an export of up to 4 Mg ha-1 soybean grains, assuming zero K losses in the system. An early desiccation of the U. brizantha pasture does not improve K uptake by the soybean crop, since 50% of the nutrient is released in the first 20 days after desiccation. The dynamics of P release from U. brizantha straw (about 6 to 10 kg ha-1 during the soybean crop) is not influenced by N fertilization and the time of pasture desiccation.


Introduction
The management of Crop-Livestock Integration Systems (ICLS) in a no-tillage (NT) system can increase the use efficiency of resources, as nutrients, to ensure high yields (Salton et al., 2014). To this end, ICLS is appropriate to increase the profitability and sustainability of soybean [Glycine max (L.) Merrill] yields in the tropical and subtropical regions of Brazil (Franchini et al., 2014).
The use of pasture of the genus Urochloa, integrated with grain production, is a feasible way of improving and sustaining NT systems in the long term (Costa et al., 2014). Forage species of this genus, apart from the high shoot and root biomass production, also favor a longer period of residue conservation on the soil, due to the high carbon/nitrogen ratio, which slows down decomposition and increases the suitability for hotter regions, where decomposition is fast (Franchini et al., 2014).
The cultivation of grain-producing crops after NT grass pastures can raise the N demand in the production system (Mateus et al., 2011). Thus, pasture fertilization with N is a technique that can increase yields of forages and subsequent crops (Costa et al., 2014). The N assimilated by forage plants is associated with carbon chains, intensifying the root and shoot growth, which can improve nutrient uptake and cycling, as P and K (Galindo et al., 2018). In turn, the nutrients released by the decomposing straw contribute to meet the nutritional needs of subsequent soybean (Cavalli et al., 2018).
Desiccation of plants present in the area by the application of non-selective herbicides is one of the main practices of NT systems. According to Franchini et al. (2014) the desiccation time can significantly change the nutrient availability for subsequent soybean, and Costa et al. (2015) suggested a synchronization of the nutrient release from residual straw with the requirement of subsequent annual crops. We hypothesize that nitrogen fertilization of the palisade grass pasture will increase the N, P, and K cycling, the decomposition rate of residual straw, and the release speed of these macronutrients to subsequent soybean, which may change the appropriate fertilization rates for the soybean crop.
In this context, the objective of this work was to determine nitrogen, phosphorus, and potassium released by nitrogen-fertilized Urochloa brizantha straw, desiccated at different times before no-till soybean sowing in succession.

Materials and Methods
The experiment was carried out between March 2016 and March 2018, in the municipality of Londrina, in the state of Paraná (23°11'S, 51°11'W; at an altitude of 620 m). The climate in the area, according to the Köppen-Geiger climate classification, is humid subtropical (Cfa), with an annual mean temperature of 21°C, a mean maximum temperature of 28.5°C in February, and a mean minimum temperature of 13.3°C in July. The soil, with a very clayey texture, was classified as Latossolo Vermelho distroférrico in the Brazilian classification (Santos et al., 2018) or Rhodic Eutrudox in the USA classification (Soil Survey Staff, 2014) and had been cultivated with NT crops -, soybean in the summer and wheat (Triticum aestivum L.) or black oat (Avena strigosa Schreb.) in the winter -, for 15 years before this assay.
The experimental area was divided into three paddocks of approximately 1.2 ha. Each one was fertilized with a nitrogen rate (0, 150, and 300 kg ha -1 N) in the form of urea (45% N), by broadcast seeding, applying half the rate in September and the other half in November 2016. From October 2016 to July 2017, the area was grazed regularly and continuously by male cattle (live weight 350-550 kg), maintaining a pasture height of 30 cm. The mean stocking rate per paddock was 4 and 2 AU ha -1 (AU, animal unit = 450 kg live weight), respectively, in the plots with and without N fertilization. After the grazing period, the area was left ungrazed until August of the same year, when two desiccation times (60 and 15 days before soybean sowing -DBSS) were evaluated in each paddock, with four replications. After the desiccation period, the BRS 1010 IPRO soybean cultivar was sowed in a 0.45 m row spacing and a seeding rate of 300 thousand plants per ha, fertilized with 350 kg ha -1 of the commercial fertilizer formula 00-20-20 (N-P 2 O 5 -K 2 O). The soybean seeds were inoculated with Bradyrhizobium elkaniibr in the day of sowing, according to technical recommendations.
The pasture was desiccated with glyphosate (1,500 g ha -1 a.e. in a spray volume of 200 L ha -1 ), applied with a tractor sprayer with fan-type spray tips. At both drying times, the atmospheric and moisture conditions of the soil were adequate for an appropriate functioning of the herbicide.
After the desiccation performed 60 DBSS, straw was collected (from 1 m 2 per plot) seven times (13,41,68,90,111,139, and 177 days after desiccation-DAD); and straw was collected six times (0, 21, 43, 64, 91, and 129 DAD) after drying 15 days before sowing. The samples were placed in paper bags and dried to constant weight in a forced air circulation oven at 60°C and weighed.
To determine the P and K contents in the palisade straw, the collected samples were digested in a microwave (MARSXpress Model, CEM Corporation, Mathews, USA), using 6 mL of 1: 1 aqueous solution of nitric acid (HNO 3 ), and 2 mL hydrogen peroxide (H 2 O 2 ) at 130V. Leaf tissue of the palisade straw (0.25 g) was used and the volume completed to 30 mL (dilution 120x) for measurements, digested in a microwave oven at 1600 w and cooled to room temperature for 20 min. After, an analytical equipment with an inductively coupled plasma atomic emission spectrometer (ICP-OES Optima 8300 Dual View Model, PerkinElmer Corporation, São Paulo, Brazil) was used, for P and K contents determination. This equipment reads the samples simultaneously for all analytes and allows configurations for both axial and radial signal detection.
The N content in the palisade straw were determined by the Kjeldahl method with sulfuric digestion in a dry tube at 350°C for approximately 4 to 5 h, with concentrated sulfuric acid (H 2 SO 4 ) and hydrogen peroxide (H 2 O 2 ); the samples were analyzed in an automatic distillation-titration system (AutoKjeldahl Unit K-370 Model, Buchi Corporation, New Casttle, USA).
The quantity of macronutrients in the straw was determined by the product of the dry weight by the nutrient content in the pasture residue. The nutrient release from straw to soil was calculated using the first derivative of the equations fitted to the data of accumulated nutrient (N, P, and K) release. Polynomial regression analysis was performed for the time after desiccation selecting the models (linear or quadratic) that best fit the data based on the p-value and r 2 , using System for Analysis of Variance (Sisvar) software (Ferreira, 2011).

Results and Discussion
In the presence of nitrogen fertilization (both 150 and 300 kg ha -1 rates), the pasture had produced a straw yield of about 5 to 6 Mg ha -1 until desiccation (both times) (Figure 1). The pattern of straw decomposition was similar in response to the three N rates, varying from 23 to 29 kg ha -1 decomposed straw per day, with an approximate total of 2 Mg ha -1 at the end of the soybean cycle. This shows that Urochloa brizantha 'BRS Piatã' straw is highly persistent in the field. For an Urochloa ruziziensis pasture, maintained at three heights (15, 35, and 50 cm), grazed continuously for six months, Franchini et al. (2014) reported a dry weight of 4.1, 6.8, and 10.7 Mg ha -1 , respectively.
In response to the three N fertilization rates and both desiccation times, the N content in palisade grass straw did not vary significantly during decomposition ( Figure 2). However, the N straw contents differed significantly from the N rates of the pasture. At desiccation 60 DBS, the N contents were 7.0, 10.0, and 12.3 g kg -1 in the pasture treated with 0, 150, and 300 kg ha -1 N, respectively. When desiccating the pasture at 15 DBSS, the N contents of forage fertilized with 0, 150, and 300 kg ha -1 N, respectively, were 7.5, 10.9, and 12.2 g kg -1 . The N contents found here were similar to those observed by Calonego et al. (2012), in a study in which U. brizantha initially contained 11.4 g kg -1 N. Similarly, Cavalli et al. (2018), found an initial content of 12.3 g kg -1 N in U. ruziziensis.
At the beginning of the evaluation period, the N straw content was higher after pasture fertilization with nitrogen (Figure 2), due to the higher concentration of this nutrient. Based on the adjusted models, the N straw content at 0 N fertilization and both drying times was 30-40 kg ha -1 , while after fertilization of the pasture with 150 and 300 kg ha -1 N, it rose to approximately 50 kg ha -1 . The daily rate of N release from pasture Pesq. agropec. bras., Brasília, v.55, e01853, 2020 DOI: 10.1590/S1678-3921.pab2020.v55.01853 straw was higher after nitrogen fertilization. When desiccation was performed 60 DBSS, the daily N release rate varied from 0.07 kg ha -1 without nitrogen fertilization to 0.23 kg ha -1 N at the highest rate. At the desiccation 15 DBSS, the daily N release rate varied from 0.23 kg ha -1 without N to 0.28 kg ha -1 when fertilized with 150 kg ha -1 N. At the end of the evaluation period, the straw of the N fertilized pasture desiccated at 60 or 15 DBSS released about to 25 and 30 kg ha -1 of N to soybean while the non-fertilized   forage released about 8 and 25 kg ha -1 , respectively. These results demonstrate that in integrated systems, an adequate fertilization program for crops must be established since the absence of pasture fertilization limits N cycling (Costa et al., 2014). It is worth mentioning that, despite the high amount of straw (about 5 to 6 Mg ha -1 ) produced by the pasture (Figure 1), the N quantity released for subsequent soybean was relatively low, due to the low straw nutrient contents (Figure 2). The N release rate is directly related to the release of C, since the remaining N had the same performance pattern as the biomass contained in the system (Mendonça et al., 2015). Thus, in plant residues with a C/N ratio higher than 25, a strong temporary N immobilization and a relatively slow decomposition speed are usually observed (Calonego et al., 2012). Once N is fixed in organic compounds, it is available for cycling in the plant-straw-soil complex formed in agroecosystems (Xia et al., 2017). Consequently, the nitrogen returning to the soil contained in plant residues accounts for an important portion of the total N taken up by the subsequent crop (Costa et al., 2012).
The time when 50% of N was released from the straw -at the intersection point of the curves of N contained in the straw and the quantity released to the soil -varied from 65 to 150 days after pasture desiccation, and was not reached during the experimental period by straw of pasture without nitrogen fertilization, desiccated 60 days before sowing (Figure 2 B). Similar results were described by Santos et al. (2014), in that the release of 50% N from U. ruziziensis was 128 days. On the other hand, Cavalli et al. (2018) reported a release time of half of the N contained in U. ruziziensis straw of 47 days. The slow N release from palisade grass straw can be a positive factor for the subsequent crop since, in this way, the nutrient can be better exploited by plants during the development cycle.
During decomposition, the straw P content tended to decrease (Figure 3). The P levels, with a quadratic performance, were similar in the straw of pasture fertilized with the three N rates and at both desiccation times, indicating that N fertilization of palisade grass did not significantly influence the P levels in the residual straw. For marandu grass, Primavesi et al. (2006) observed that the P levels did not increase when fertilized with N in the form of ammonium nitrate but decreased after N applied in the form of urea. No significant differences in the P contents of U. brizantha 'Xaraés' straw were observed by Costa et al. (2014) in response to five N rates (0, 50, 100, 150, and 200 kg ha -1 N), corroborating the data of this study.
The dynamics of P release from palisade grass straw was similar in response to the three N fertilization rates (Figure 3). The total amount released by the pasture ranged from 6 to 10 kg ha -1 P 2 O 5 . These values are low compared to the demand of a subsequent soybean crop, by which around 10 kg P 2 O 5 are exported per ton of grain (Tecnologias…, 2013). However, this study indicates that the amounts of P supplied via residual straw of U. brizantha 'BRS Piatã' are rather low. On the other hand, it should be mentioned that the amount of P cycled through the biomass grazed by cattle was not evaluated. The selection of plant species with high efficiency in accumulating P in the plant shoots and subsequent release to the soil by the mineralization of organic compounds is a useful strategy to increase P availability for the plants since this element can undergo changes by the fixation process and be transformed into non-labile forms in the soil (Costa et al., 2012).
The potential amount of P to be released from organic tissues are usually structurally linked to protein molecules and compounds responsible for energy transport, making it available both for root uptake of the subsequent crop and for immobilization in mineral constituents with complex solubility (Mendonça et al., 2015). As also observed in this study for the remaining amount of P in plant residues, Rossi et al. (2013) reported an initial phase of rapid P release followed by a slower one. Practically all P contained in the straw was released to the subsequent soybean, and the period to release 50% of the P from the straw lasted 30 to 65 days. Santos et al. (2014) observed that it took 67 days to release half of the P contained in U. ruziziensis straw.
The K contents decreased rapidly in the first days after desiccation, in the straw of pasture treated with three N fertilization rates and desiccation after two periods (Figure 4). The cation K is mostly concentrated in the cytoplasm of plant cells and has no structural function, but forms bonds with easily reversible organic complexes, and is therefore easily disassociated from the remaining plant material (Mendonça et al., 2015). Consequently, shortly after pasture desiccation, the tissue concentration of this Pesq. agropec. bras., Brasília, v.55, e01853, 2020 DOI: 10.1590/S1678-3921.pab2020.v55.01853  Pesq. agropec. bras., Brasília, v.55, e01853, 2020 DOI: 10.1590/S1678-3921.pab2020.v55.01853 nutrient decreased markedly, because after the rupture of the plasma membranes it was rapidly carried away by rainwater (Crusciol et al., 2008). It is important to highlight that N fertilization of the pasture induced a significant increase in the K straw content, indicating that N application can favor K cycling in palisade grass. In a study with U. brizantha 'Xaraés', Costa et al. (2008) confirmed these data since the K straw content tends to a quadratic increase in response to increasing N rates. Similarly, Primavesi et al. (2006) also observed an increase in K levels in U. brizantha 'Marandu' pasture under increasing N rates.
The amount of K released from the straw was high, especially when the pasture had received nitrogen fertilization (Figure 4). In the desiccation carried out 60 DBSS, the amount of K released from the straw of unfertilized pasture was 60 kg ha -1 (Figure 4 B), but 80 kg ha -1 from the straw of pasture fertilized with 150 kg ha -1 N (Figure 4 D). The export per ton of grain of this nutrient by soybean is approximately 20 kg K 2 O (Tecnologias…, 2013). Therefore, for a production of four tons of grain, the K released from the straw of N-fertilized pasture would be sufficient to offset the export in soybean, in case of no erosion or leaching losses in the system.
At desiccation 15 DBSS, the accumulated amount of K released from the unfertilized dry pasture was 60 kg ha -1 (Figure 4 A), and 100 kg ha -1 from the straw of the pasture fertilized with 300 kg ha -1 N (Figure 4 E). These results indicate the effect of pasture N fertilization to intensify K cycling in palisade grass. Also, a probable explanation for the increase of K accumulation in palisade grass straw after nitrogen fertilization is the intensified root growth of the forage, increasing K uptake. It is worth remembering that a more developed root system has a greater exploration capacity of the soil profile, which improves the facility of taking up higher nutrient amounts, even at a depth below 0.2 m. The growth of the root system of forage grasses is fast and deep, which may have resulted in the uptake of K leached into subsurface layers (Mendonça et al., 2015;Galindo et al., 2018).
The time to release 50% of the K contained in the straw was 10-20 days. A very early pasture desiccation can reduce the exploitation of K by subsequent soybean, due to the rapid release of this nutrient. For this reason, the best way to supply soybean with K released from the straw is to dry the grass shortly before sowing. The time required to release half of the K content reported by Cavalli et al. (2018) was 16 days for palisade grass. According to Santos et al. (2014), it took only 13 days to release 50% K.
In general, the nutrient contents released from palisade grass straw followed the following order: K> N> P, regardless of N application to the pasture or desiccation time. The quantities of K, N, and P released by U. brizantha straw fertilized with 300 kg ha -1 N can offset 80, 15.7 and 18%, respectively, of the N exported via soybean in a grain yield of 5 Mg ha -1 , considering zero losses in the system. In this context, the K released from U. brizantha straw is essential for oilseed nutrition, and studies are needed to test the possibility of reducing the K rate applied to soybean grown after pasture species of the genus Urochloa.

Conclusions
1. Nitrogen fertilization of Urochloa brizantha 'BRS Piatã' significantly increases N and K cycling and the release of these nutrients to subsequent soybean (Glycine max).
2. The amount of K released from U. brizantha straw fertilized with 150 or 300 kg ha -1 N is sufficient to offset the export in up to 4,000 kg ha -1 of soybean grains.
3. Early desiccation of U. brizantha pasture (60 days before soybean sowing) can disadvantage the K use by soybean since 50% of the nutrient is released in the first 20 days after desiccation.
4. The dynamics of P release from U. brizantha straw is not influenced by nitrogen fertilization and desiccation time.