Enhancing nutrient cycling and soybean grain yield in sandy soils via early potassium on Urochloa brizantha

ABSTRACT

Sandy soils pose significant challenges for nutrient management owing to their low nutrient retention capacity and the high-intensity rainfall characteristic of tropical regions. This study aimed to evaluate the effects of K fertilization timing on nutrient release from Urochloa brizantha (Hochst. ex A. Rich.) R.D. Webster residues and the subsequent impact on soybean grain yield in sandy soils. The experiment was conducted in Paraíso das Águas, Mato Grosso do Sul State, Brazil, during the 2019/2020 growing season. Treatments included two K rates applied to U. brizantha mulch (0 and 100 kg ha^-1 of K; 0 and 120 kg ha^-1 of K₂O) and four K rates applied to soybean (0, 25, 50, and 100 kg ha^-1 of K, which correspond to 0, 30, 60, and 120 ha^-1 of K₂O) Residue decomposition and nutrient release dynamics were monitored using litter bags collected at intervals of 0, 15, 30, 60, and 90 days. Early K application to U. brizantha significantly enhanced nutrient release, particularly K, which showed rapid release with a half-life of 16 days. Soil K was strongly stratified in the 0.00–0.05 m layer, with no evidence of substantial downward movement beyond 0.40 m within the experimental period. Instead, U. brizantha residues served as a nutrient reservoir, effectively releasing phosphorus (P) and potassium (K) to the subsequent soybean crop, aligning nutrient availability with crop demand. The highest soybean grain yield occurred, reaching 3,573 kg ha^-1 especially when 100 kg ha^-1 of K (120 kg ha^-1 of K₂O) was applied to U. brizantha and no additional K was applied to soybean; this grain yield was higher than that of to 3,176 kg ha^-1, observed with direct K application to soybean without early K. These findings underscore the benefits of system-based fertilization strategies in enhancing nutrient use efficiency and sustaining crop productivity in sandy soils.

Keywords
potassium fertilization; sandy soils; system fertilization; cover crop; Urochloa brizantha

INTRODUCTION

Potassium (K) is an essential macronutrient for crop growth, playing key roles in photosynthesis, osmotic regulation, enzyme activation, and stress tolerance. In tropical agricultural systems, K management is particularly challenging due to high rainfall, elevated temperatures, and intense biological activity, which accelerate nutrient cycling and can increase nutrient losses. The adoption of conservation systems such as no-tillage (NT) has altered nutrient dynamics, concentrating K near the soil surface and influencing its availability to crops (Johnson et al., 2022).

Sandy soils present an even greater challenge for K management. Their low cation exchange capacity (CEC), reduced organic matter content, and weak buffering capacity favor downward nutrient leaching in the soil profile, potentially limiting K availability during critical crop growth stages. These characteristics demand fertilization strategies that synchronize nutrient supply with crop demand while minimizing losses.

Cover crops, especially tropical forages such as Urochloa brizantha, are widely recognized for their high biomass production and nutrient recycling. When used in NT systems, U. brizantha captures residual soil K and stores it in plant tissues, later releasing it through residue decomposition. This gradual nutrient release can benefit subsequent crops and improve nutrient use efficiency in integrated systems.

The concept of “system fertilization,” in which nutrients are applied to the preceding cover or forage crop rather than directly to the cash crop, has gained attention in Brazil for P management. However, studies on early or anticipated K fertilization in such systems are scarce, especially in sandy soils and with soybean as the main summer crop. Recent findings by Echer et al. (2020) demonstrated that early K application to Urochloa ruziziensis before cotton in sandy soil increased fiber yield by ~16 % and maintained fiber quality, indicating efficient K recycling. Echer et al. (2023) further showed that early K application to ruzigrass improved soil K balance—both exchangeable and non-exchangeable forms—up to 40 days after desiccation, benefiting the subsequent cotton crop. While these studies validate the principle of early K fertilization in cover crops, they were conducted with cotton under different production contexts, and no published studies to date have evaluated this strategy in a Urochloa–soybean sequence in sandy soils under NT.

We hypothesized that early K fertilization in Urochloa brizantha increases biomass production and K accumulation in the cover crop, thereby enhancing K release to the subsequent soybean crop and improving soybean grain yield in sandy soils under no-tillage. We aimed to evaluate the effects of applying K to Urochloa brizantha during the off-season on biomass production, K accumulation and release from residues, and subsequent soybean grain yield under NT.

MATERIALS AND METHODS

The experiment was conducted from April 2019 to April 2020 in Paraíso das Águas, Mato Grosso do Sul, Brazil, specifically at the Primavera farm (19°14'5.63" S; 53°14'39.26" W), at an elevation of 677 m with slightly undulating relief. The climate in the region is classified as tropical with a dry season (Aw) according to the Köppen classification system. The meteorological data for the experimental period are presented in figure 1. The experimental site has a sandy soil classified as Neossolo Quartzarênico órtico típico according to the SiBCS, which corresponds to a Typic Quartzipsamment, with a particle-size composition of 870 g kg^-1 sand, 80 g kg^-1 silt, and 50 g kg^-1 clay. Initial chemical properties of the soil at a layer of 0.00-0.20 m were: pH(CaCl₂) 5.5, organic matter (OM) 17.0 g dm^-3, P (Resin) 21.0 mg dm^-3, K⁺ 0.12 cmol_c dm^-3, Ca²⁺ 2.2 cmol_c dm^-3, Mg²⁺ 0.5 cmol_c dm^-3, Ca-Mg-K Saturation 56.3 %, and cation exchange capacity (CEC) = 5.7 cmol_c dm^-3. Soil pH was determined in CaCl₂, organic matter by wet oxidation, P by ion-exchange resin extraction, K by flame photometry, and Ca and Mg by resin extraction, following standard soil analysis procedures.

Figure 1
Meteorological parameters observed during the experimental period from April 2019 to March 2020 in Paraíso das Águas, Mato Grosso do Sul State, Brazil. Monthly precipitation (mm) is displayed on the primary Y-axis, while maximum and minimum temperatures (°C) are shown on the secondary Y-axis. Key management events are indicated, including Urochloa brizantha planting, KCl application, Urochloa brizantha desiccation, soybean planting, and subsequent KCl applications.

The experimental design for litter bag evaluation was a randomized complete block with four replications, in a split-plot arrangement over time and space. The main plots included two K rates applied to Urochloa brizantha: 0 and 100 kg ha^-1 of K (120 kg ha^-1 of K₂O) (K-Brachiaria 0 and K-Brachiaria 100, respectively). Subplots included K rates applied as topdressing on soybean at rates of 0, 25, 50, and 100 kg ha^-1 of K (0, 30, 60, and 120 kg ha^-1 of K₂O). Although K rates applied to soybean (K-soybean) were included in the statistical model for litter bag evaluation, they did not biologically influence the decomposition rate of U. brizantha mulch, since litter bags had been placed in the field prior to soybean fertilization. The third factor evaluated in the sub-subplots was the litter bag retrieval times at 0, 15, 30, 60, and 90 days after deposition. For grain yield evaluation, only the factors allocated to the main plots and subplots (K rates for Urochloa brizantha and for soybean) were considered.

Urochloa brizantha (cv. Marandu) was broadcast-sown on April 19, 2019, using a leveling harrow at a seeding rate of 8 kg ha^-1 of seed. Urochloa brizantha was used solely as a cover crop. Potassium was applied broadcast on the soil surface to Urochloa brizantha during the off-season, without incorporation, 90 days after sowing, at a rate of 100 kg ha^-1 of K (120 kg ha^-1 of K₂O) in the form of KCl (60 % K₂O). No P or nitrogen (N) fertilizers were applied to the cover crop. Potassium topdressing was applied (90 days after sowing), approximately 25 days before the desiccation of Urochloa brizantha, at a rate of 100 kg ha^-1 of K (120 kg ha^-1 K₂O) in the form of KCl (60 % K₂O). On August 14, 2019, Urochloa brizantha was desiccated using a sequential application of herbicides. The first application consisted of Glyphosate (2,400 g ha^-1 of active ingredient [a.i.]), Chlorimuron (20 g a.i. ha^-1), 2,4-D (165 g a.i. ha^-1), and mineral oil (0.25 % v/v) applied with a spray volume of 100 L ha^-1 using JTT 110-02 nozzles. Ten days later, a second application was made with Paraquat (240 g a.i. ha^-1), Diclosulam (11.7 g a.i. ha^-1), and mineral oil (0.25 % v/v) using the same nozzle configuration and spray volume. This desiccation process ensured complete termination of U. brizantha before to soybean planting.

Soybean was sown on November 13, 2019, using the hybrid 98Y30 RR2 Pioneer®. The seeds were treated with fungicides, insecticides, and inoculated with strains of Bradyrhizobium japonicum (strains 5079 and 5080), combined with Azospirillum brasilense (strains AbV5 and AbV6), to enhance biological N fixation and promote plant growth. This inoculation strategy aimed to improve soybean nodulation and overall grain yield under the study conditions.

Planting was performed with a fertilizer drill at a spacing of 0.45 m, a seeding depth of 0.03 m, and a plant population of 260,000 plants ha^-1. No P fertilizer was applied to either soybean or U. brizantha, as soil P levels (21.0 mg dm^-3, resin) were considered high according to Pauletti and Motta (2017). Potassium topdressing was applied 30 days after planting at rates of 0, 25, 50, and 100 kg ha^-1 of K (0, 30, 60, and 120 kg ha^-1 of K₂O) in the form of KCl (60 % K₂O ha^-1).

To evaluate the dry matter decomposition rate and nutrient release, litter bags were used (Aita and Giacomini, 2003; Giacomini et al., 2003; Gama-Rodrigues et al., 2007; Vitti et al., 2008; Rossi et al., 2013, 2016; Santana et al., 2011; Potrich et al., 2014; Santos et al., 2014; Torres and Pereira, 2014). The litter bags were made of nylon fabric with 2 mm mesh openings, measuring 0.20 × 0.20 m. Each litter bag contained 100 g of dry matter from Urochloa brizantha, randomly collected from the field (before desiccation). Residues of Urochloa brizantha were oven-dried at 65 °C until constant weight to determine initial dry mass and nutrient content. Subsequently, 10 g of this dried material were placed in litterbags (0.20 × 0.20 m, 2 mm mesh) and positioned on the soil surface under field conditions. Four litterbags were placed per subplot, each considered an independent replicate for statistical analysis, and collected at the predetermined sampling dates. The interval between U. brizantha desiccation and soybean sowing was approximately 90 days. Litter bags were placed at soybean sowing to synchronize the decomposition period with the soybean cycle, following the methodology of Rossi et al. (2013). After field collection, the litter bag contents were dried in an oven at 65 °C until constant weight and then weighed on a precision scale.

The decomposition rate was assessed by the difference between the initial weight and the remaining amount over the evaluation periods. Samples were ground in a Wiley mill to pass through a 2-mm sieve. The N, P, and K contents (g kg^-1 dry matter) in Urochloa brizantha and soybean plant tissue were determined after drying at 65 °C and grinding. Nitrogen was determined by the Kjeldahl method, P by colorimetry using the molybdenum blue method, and K by flame photometry after acid digestion, following the procedures described by Tedesco et al. (1995). Cumulative nutrient release was estimated using a mass-balance approach that integrates dry matter loss and nutrient content over time. For each period, the nutrient released (kg ha^-1) was calculated as the difference between nutrient stocks at consecutive sampling times, using equation 1.

Eq. 1 $\mathrm{Nutrient}\mathrm{release}\left(\mathrm{t}\to \mathrm{t}+1\right)=({\mathrm{DM}}_{\mathrm{t}}\times {\mathrm{C}}_{\mathrm{t}})-({\mathrm{DM}}_{\mathrm{t}+1}\times {\mathrm{C}}_{\mathrm{t}+1})$

in which: DM is the estimated dry matter (kg ha^-1) obtained from the exponential decay model and C is the nutrient content (g kg^-1) in the residue. The cumulative nutrient release for N, P, and K was obtained by summing the values released across successive intervals.

The dry matter decomposition rates and nutrient release rates from Urochloa brizantha straw were estimated using a nonlinear regression model fitted to the observed values. The fitted model is represented by equation 2.

Eq. 2 $\mathrm{X}={\mathrm{Ae}}^{-\mathrm{kt}}+\left(100-\mathrm{A}\right)$

in which: X represents the percentage of dry matter or nutrient content remaining at time t (days); and kt is the decomposition constant of the more readily decomposable compartment (A). In the exponential decay model fitted to nutrient release data, A represents the asymptotic maximum nutrient release (g kg^-1) and was estimated as a fitted parameter using nonlinear regression.

The same equation used to determine the decomposition rate was applied to evaluate the nutrient release rate by replacing X with N, P, and K contents (g kg^-1 dry matter) in plant tissue. The cumulative release of N, P, and K was estimated by the differences between the initial nutrient amounts in the residue and the amounts at each collection date.

Using the decomposition constant, the time required for 50 % of the biomass to decompose (half-life; T_1/2) of the remaining cultural residues was also estimated according to the equation 3, proposed by Paul and Clark (1989), in which 0.693 is the natural logarithm of 2 and k is the decomposition constant.

Eq. 3 ${\mathrm{T}}_{1/2}=0.693/\mathrm{k}$

During the experimental period, soil available P and K were determined using the Mehlich-1 extractant, and mineral N (nitrate + ammonium) was determined after extraction with 1 mol L^-1 KCl. Three samplings were performed before soybean planting: I - before K fertilization of Urochloa brizantha; II - after K fertilization of Urochloa brizantha (July 4, 2019); III - after K fertilization on the soybean crop (October 12, 2019). Samplings I and II were performed at the layers of 0.00–0.20, 0.20–0.40, 0.40–0.60, and 0.60–0.80 m to evaluate potential nutrient movement after fertilization, considering the deeper root system of Urochloa brizantha. Sampling III used the layers of 0.00–0.10, 0.10–0.20, and 0.20–0.40 m to obtain greater resolution of nutrient distribution in the upper soil profile, where soybean roots are concentrated. Soil samples were collected in the crop row using a Dutch auger. Samples were taken from three points within each experimental unit, homogenized to form a composite sample. After field collection, samples were dried in a forced-air oven at 55 °C until constant weight, then ground in a specialized soil mill using a 2 mm mesh. Available P and K were determined using the Mehlich-1 extractant, with P quantified by colorimetry (molybdenum blue method) and K by flame photometry. Mineral N (NO₃^- + NH₄⁺) was determined after extraction with 1 mol L^-1 KCl. Nitrate was quantified by colorimetry using the phenoldisulfonic acid method, while ammonium was determined by steam distillation with MgO and Devarda’s alloy, according to Tedesco et al. (1995)

Soybean grain yield was determined from a 2 m² area located in the center of each subplot, avoiding border effects. Grain moisture was measured with a moisture meter and adjusted to 130 g kg^-1. Grain weight was recorded using a precision scale (±0.01 g).

All data were subjected to analysis of variance (ANOVA) using the appropriate linear model for split-plot or split-split-plot designs, considering blocks as random effects and fertilization factors and soil depth as fixed effects. The assumptions of normality and homogeneity of variances were verified using the Shapiro–Wilk and Levene’s tests, respectively. When significant effects were detected (p≤0.05), treatment means were separated by Fisher’s LSD test.

For soybean grain yield and U. brizantha biomass, ANOVA was performed considering the interaction between K₂O applied to U. brizantha and K₂O applied to soybean. For soil chemical properties, multifactor ANOVAs tested the effects of K₂O applied to U. brizantha, K₂O applied to soybean, soil depth, and their interactions.

For residue dry mass and nutrient release (N, P, and K), nonlinear regression analyses were performed using the single-exponential decay model described in equation 2. The half-life was calculated according to equation 3. Each regression point corresponded to the mean of four replicates per treatment × sampling date (n = 4). Model fit and parameter significance were evaluated based on regression coefficients (p≤0.05), the coefficient of determination (R²), and residual analysis. All statistical analyses were carried out in Statgraphics Centurion (Statgraphics Technologies, Inc.), and regression curves and graphs were produced in SigmaPlot (Systat Software, Inc.).

RESULTS

Table 1 presents the significance levels (p-values) of the main factors and their interaction on the productivity of Urochloa brizantha residue and soybean grain yield. Table 1 summarises the effects of these factors on decomposition and nutrient-release variables across sampling dates. In the following subsections, we describe the results in detail.

Table 1
Analysis of variance for mulch decomposition, nutrient release, and soil nutrient dynamic

Dependent variable	K on Urochloa brizantha (A)	K on soybean (B)	Time/Depth (C)	Interaction A × B	Interaction A × C	Interaction B × C	Interaction A × B × C
N remaining in mulch (%)	0.0000*	0.2539	0.0000	0.0712	0.0033*	0.5339	0.6498
N release dynamics in Urochloa brizantha mulch	0.0000*	0.2512	0.0000*	0.0686	0.0034*	0.6002	0.5419
P remaining in mulch (%)	0.0000*	0.6834	0.0000*	0.0522	0.0000*	0.9303	0.9203
P release dynamics in Urochloa brizantha mulch	0.0000*	0.6834	0.0000*	0.0522	0.0000*	0.9303	0.9203
K remaining in mulch (%)	0.0000*	0.3829	0.0000*	0.1691	0.0050*	0.7851	0.0356*
K release dynamics in Urochloa brizantha mulch	0.3980	0.2564	0.0000*	0.2391	0.5205	0.7826	0.1147
Soil P (before K on Urochloa brizantha )	—	—	0.0000*	—	—	—	—
Soil K (before K on Urochloa brizantha )	—	—	0.0000*	—	—	—	—
Soil P (after K on Urochloa brizantha )	0.3800	—	0.0009*	—	0.3643	—	—
Soil K (after K on Urochloa brizantha )	0.0003	—	0.0001*	—	0.0012*	—	—
Soil P (after K on soybean)	0.2083*	0.0095	0.0011*	0.5467	0.9747	0.4484	0.5210
Soil K (after K on soybean)	0.0000*	0.0279*	0.0000*	0.1566	0.8125	0.0343*	0.7382

Soybean grain yield

Grain yield was significantly influenced by the interaction between K fertilization on Urochloa brizantha (main plot) and K fertilization on soybean (subplot) (p = 0.0138; Table 2). The highest grain yield (3,573 kg ha^-1) occurred when 100 kg ha^-1 of K was applied to U. brizantha, and no K was applied to soybean. In contrast, when 100 kg ha^-1 of K was applied directly to soybean without K on the cover crop, grain yield reached 3,176 kg ha^-1 — about 11 % lower (Figure 2).

For each U. brizantha K level, increasing K rates in soybean (0, 25, 50, and 100 kg ha^-1 of K) did not improve grain yield; in fact, after fertilizing the cover crop, soybean grain yield declined by 7–15 % when K was also applied to soybean. These results indicate that early K application to the cover crop is more effective than direct K application to soybeans.

Table 2
Analysis of variance for soybean grain yield and initial mulch properties

Dependent variable	K on Urochloa brizantha (A)	K on soybean (B)	Time (C)	Interaction A × B	Interaction A × C	Interaction B × C	Interaction A × B × C
Soybean grain yield (kg ha-1)	0.0073*	0.3666	—	0.0138*	—	—	—
Residual dry matter of Urochloa brizantha mulch (kg ha-1)	0.0018*	0.2699	0.0000*	0.0107*	0.4936	0.9858	0.9203

Figure 2
Soybean grain yield (kg ha^-1) as a function of K rates applied to Urochloa brizantha (0 and 100 kg ha^-1 of K; 0 and 120 kg ha^-1 of K₂O) and K rates applied to soybean (0, 25, 50, and 100 kg ha^-1 of K; 0, 30, 60, and 120 kg ha^-1 of K₂O). Data are presented as means ± 159.428, with statistical significance indicated by p-values 0.0138.

Dry matter decomposition

The analysis of residual dry matter (DM) from Urochloa brizantha mulch, as shown in figure 3, indicates that the initial amount of DM at day zero was significantly higher for the treatment with K application (100 kg ha^-1 of K; 120 kg ha^-1 of K₂O), reaching approximately 7450 kg ha^-1, compared to the treatment without K application (0 kg ha^-1), which had around 6,000 kg ha^-1. The additional ~1000 kg ha^-1 of biomass was associated with approximately 25 days of residual effect from K fertilization on U. brizantha.

Figure 3
Residual dry matter (DM) of Urochloa brizantha mulch over time. Data are presented as means ± 331.1180, with statistical significance indicated by p-values 0.0107.

After deposition, dry mass declined sharply in the first 20 days for both treatments, reaching approximately 3.3–3.5 Mg ha^-1. From 20 to 60 days, the rate of decrease slowed, and by 60 days, the treatments converged to similar values (~2.7–2.9 Mg ha^-1), with a slight further reduction until 90 days. The decomposition pattern fitted a first-order exponential decay model, with different equations for each K-Brachiaria level (Equation 4 and 5).

Eq. 4 $\mathrm{Unfertilized}(0\mathrm{K}-\mathrm{Brachiaria}):\mathrm{Y}=43.82+56.18\mathrm{e}^\left\{-0.0842\mathit{\times }\right\}({\mathrm{R}}^2=0.98,{\mathrm{T}}_{1/2}=8.2\mathrm{days})$

Eq. 5 $\mathrm{K}-\mathrm{fertilized}(100\mathrm{K}-\mathrm{Brachiaria}):Y=31.89+54.90e^\{-0.0924\times \left\}\right({\mathrm{R}}^2=0.96,{\mathrm{T}}_{1/2}=7.5\mathrm{days})$

For the treatment without K, residue decomposition followed the equation Y = 43.82 + 56.18 e^{–0.0842 x}, in which 43.82 % is the resistant fraction, 56.18 % is the labile fraction and the decay constant (0.0842 day^-1). In the K-fertilized treatment, the equation Y = 31.89 + 54.90 e ^{-0.0924 x} indicates that 31.89 % is the resistant fraction, 54.90 % is the labile fraction, and the decay constant is 0.3001 day^-1. The half‑life of the labile fraction decreased from 8.2 days (no K) to 7.5 days (100 kg ha^-1 of K), a difference of 0.7 day (≈8 %). While significant, this small difference may have little practical impact.

Nutrient content in the residual mulch

The chemical composition of Urochloa brizantha shoots and the corresponding biomass production under the two potassium fertilization rates are presented in table 3. Potassium fertilization significantly affected all evaluated variables (p<0.05). The application of 100 kg ha^-1 of K (120 kg ha^-1 of K₂O) increased shoot biomass from 6.47 to 7.51 Mg ha^-1, indicating greater forage growth and nutrient accumulation potential. In contrast, the N, P, and K contents in plant tissue were significantly reduced at the higher K rate, suggesting a dilution effect associated with increased dry matter production. These initial nutrient contents were used to calculate the total nutrient stocks in the plant material and to monitor subsequent nutrient release during residue decomposition.

Table 3
Initial mean concentrations of nitrogen (N), phosphorus (P), and potassium (K), and shoot biomass of Urochloa brizantha under two potassium fertilization rates (0 and 120 kg ha^-1 of K₂O)

K2O rate	Biomass	N	P	K
kg ha-1	Mg ha-1	‒‒‒ g kg-¹ ‒‒‒
0	6.47 ± 0.42 b	15.78 ± 0.22 a	9.10 ± 0.23 a	17.07 ± 0.05 a
100	7.51 ± 0.35 a	11.04 ± 0.22 b	7.17 ± 0.23 b	15.01 ± 0.05 b

• Means followed by different letters within the same column differ by the F test (p<0.05).

The nutrient stock in the Urochloa brizantha mulch indicated that only nitrogen (N) was significantly affected by the potassium (K) fertilization rate (p = 0.003). Plants that received 100 kg ha^-1 of K (120 kg ha^-1 of K₂O) stored an average of 81.6 kg ha^-1 of N, whereas those without K application (0 kg ha^-1 of K) stored 94.9 kg ha^-1 of N. There was no significant effect of K rate on phosphorus (P) (p=0.917) or on potassium (K) (p=0.071), with average nutrient stocks close to 54.6 kg ha^-1 of P and 107.0 kg ha^-1 of K, respectively.

Nitrogen remaining and release dynamics

The percentage of N remaining in the Urochloa brizantha mulch (Figure 4a) was significantly affected by the interaction between K-Brachiaria × time (p = 0.0033). In K-Brachiaria 100, N remaining fell from 100 % at day 0 to 83.9 % at 30 days, then rose to 106.2 % at 60 days and 123.3 % at 90 days, evidencing net immobilization at later stages. In K-Brachiaria 0, N remaining varied less over time, with 69.7, 73.3, 82.5, and 81.7 % at 15, 30, 60, and 90 days, respectively.

The corresponding cumulative N release (Figure 4b; kg ha^-1) was significantly influenced by time (p<0.001), while K-Brachiaria (p = 0.171) and the interaction K × time (p = 0.937) were not significant. The amount of N released increased gradually over the sampling periods, averaging 53.6, 55.0, 56.5, and 58.3 kg ha^-1 at 15, 30, 60, and 90 days after deposition, respectively. The K applied to soybean did not significantly affect N dynamics (p = 0.2539).

Figure 4
Nitrogen (N) dynamics in Urochloa brizantha mulch as affected by potassium fertilization rates: (a) percentage of N remaining over time and (b) cumulative N release (kg ha^-1) during residue decomposition at 15, 30, 60, and 90 days after deposition. Bars represent mean values (n = 4). Error bars indicate the standard errors of the mean.

Phosphorus remaining and release dynamics

The percentage of P remaining in Urochloa brizantha mulch (Figure 5a) was significantly affected by the interaction between K-Brachiaria × time (p = 0.0004). In the K-Brachiaria 120 treatment, P remaining declined rapidly from 100 % at day 0 to 69.2 % at 15 days and stabilized near this value for the rest of the period (69.4 % at 30 days, 68.9 % at 60 days, and 68.8 % at 90 days). In the K-Brachiaria 0 treatment, P remaining decreased from 100 to 45.3 % at 15 days, then remained close to this level (45.0–45.2 %) until 90 days after deposition.

The corresponding cumulative P release (Figure 5b; kg ha^-1) was significantly affected by time (p<0.001), while K-Brachiaria (p = 0.667) and the interaction K × time (p = 0.941) were not significant. The cumulative amount of P released increased gradually with time, averaging 39.7, 40.4, 43.0, and 44.6 kg ha^-1 at 15, 30, 60, and 90 days after deposition, respectively. Potassium applied to soybean (K-soybean) did not significantly influence P dynamics (p = 0.2151).

Figure 5
Phosphorus (P) dynamics in Urochloa brizantha mulch as affected by potassium fertilization rates: (a) percentage of P remaining over time and (b) cumulative P release (kg ha^-1) during residue decomposition at 15, 30, 60, and 90 days after deposition. Bars represent mean values (n = 4). Error bars indicate the standard errors of the mean.

The decrease in P remaining in the Urochloa brizantha mulch over time was described using a first-order exponential decay model (Equations 6 and 7), in which Y represents the percentage of P remaining at each sampling time.

Eq. 6 $\mathrm{K}-\mathrm{Brachiaria}0:Y=45.15+54.90e^\{-0.3011\times \left\}\right({\mathrm{R}}^2=0.73,{\mathrm{T}}_{1/2}=2.3\mathrm{days})$

Eq. 7 $\mathrm{K}-\mathrm{Brachiaria}100:Y=69.21+30.79e^\{-0.8933\times \left\}\right({\mathrm{R}}^2=0.49,{\mathrm{T}}_{1/2}=0.8\mathrm{days})$

Potassium remaining and release dynamics

The percentage of K remaining in Urochloa brizantha mulch (Figures 6a and 6b) was significantly influenced by time (p<0.0001), by the interaction K-Brachiaria × time (p = 0.0148), and by K-soybean (P = 0.0126). In K-Brachiaria 0, K content dropped sharply from 100 % at day 0 to 17.1 % at 15 days, and remained low thereafter (17.7 % at 30 days, 14.5 % at 60 days, and 14.4 % at 90 days). In K-Brachiaria 120, the initial decline was also rapid, from 100 to 25.9 % at 15 days, followed by small fluctuations (25.6 % at 30 days, 19.6 % at 60 days, and 23.0 % at 90 days). Across the evaluation period, K-soybean application slightly increased the proportion of K remaining in the mulch, regardless of K-Brachiaria rate, but did not change the overall pattern of rapid early decline.

The decrease in K remaining in the Urochloa brizantha mulch over time was described using a first-order exponential decay model (Equations 8 and 9), in which Y represents the percentage of K remaining at each sampling time.

Eq. 8 $\mathrm{K}-\mathrm{Brachiaria}0:Y=84.51e^\{-0.26093\times \}+15.49({\mathrm{R}}^2=0.99,{\mathrm{T}}_{1/2}=2.66\mathrm{days})$

Eq. 9 $\mathrm{K}-\mathrm{Brachiaria}100:Y=77.38e^\{-0.20608\times \}+22.62({\mathrm{R}}^2=0.95,{\mathrm{T}}_{1/2}=3.36\mathrm{days})$

The corresponding cumulative K release (Figure 6c; kg ha^-1) was significantly affected by time (p<0.001), while K-Brachiaria (p = 0.199), K-soybean (p = 0.282), and their interactions were not significant. The cumulative amount of K released increased gradually with time, averaging 95.7, 95.4, 99.9, and 100.4 kg ha^-1 at 15, 30, 60, and 90 days after deposition, respectively.

Figure 6
Potassium (K) dynamics in Urochloa brizantha mulch over time (15, 30, 60, and 90 days after deposition) as influenced by K application rates to Urochloa brizantha (K-Brachiaria) and K application rates to soybean (K-soybean): (a) percentage of K remaining for treatments with 0 kg ha^-1 of K applied to Urochloa brizantha; (b) percentage of K remaining for treatments with 100 kg ha^-1 of K applied to Urochloa brizantha; and (c) cumulative K release (kg ha^-1) from Urochloa brizantha mulch. Bars represent mean values (n = 4). Error bars indicate the standard errors of the mean.

Soil phosphorus dynamics

Soil P content (Figure 7) was significantly affected only by soil depth in all sampling times: before K application on Urochloa brizantha (p<0.0001), after K application on U. brizantha (p<0.0001), and after K application on soybean (p<0.0001). In all cases, P contents decreased with increasing depth.

Before K fertilization on U. brizantha (Figure 7a), the highest P content was found in the 0.00–0.05 m layer (28.1 mg dm^-3), declining to values below 10 mg dm^-3 in the 0.20–0.40 m layer. After K fertilization on U. brizantha (Figure 7b), the same trend was observed, with the surface layer (0.00–0.05 m) presenting 14.0 mg dm^-3, while deeper layers showed less than 9 mg dm^-3. After K fertilization on soybean (Figure 7c), the P distribution pattern remained unchanged, with 15.1 mg dm^-3 in the 0.00–0.05 m layer and values below 9 mg dm^-3 below 0.20 m. These results indicate that soil P availability was strongly stratified, restricted mainly to the topsoil, and was not significantly altered by K fertilization in either of P typical of no-tillage tropical soils or soybean.

Figure 7
Vertical distribution of soil phosphorus (P) content (mg dm^-3) at different layers (0.00–0.05, 0.05–0.10, 0.10–0.20, and 0.20–0.40 m): (a) before potassium application on Urochloa brizantha, (b) after potassium application on U. brizantha (0 and 100 kg ha^-1 of K), and (c) after potassium application on soybean showing all K rates (0, 25, 50, and 100 kg ha^-1 of K). In (c), the depth-weighted mean (0.00–0.40 m) followed the regression P = 8.08 + 0.0404 × Kdose (R² = 0.62). Bars represent mean ± standard error (n = 4).

After K fertilization on U. brizantha (Figure 7b), soil P content was significantly affected by depth (p = 0.0009) but not by K dose on the cover crop (p = 0.380). Phosphorus contents decreased sharply from approximately 16 mg dm^-3 in the 0.00–0.05 m layer to below 3 mg dm^-3 at 0.20–0.40 m, regardless of K application. Although the K effect was not significant, the figure was maintained to highlight the strong vertical stratification of P typical of no-tillage tropical soils. After K fertilization on soybean (Figure 7c), P was significantly influenced by both depth (p = 0.0011) and K rate (p = 0.0095), whereas K applied to U. brizantha had no significant residual effect (p = 0.2083). To better represent this significant main effect, all soybean K rates (0, 30, 60, 120 kg ha^-1 of K₂O) were included in the figure, even though the interaction K-Brachiaria × K-Soybean was not significant (p = 0.267). Depth-weighted mean soil P (0.00-0.40 m) increased with K rate on soybean, following the linear model (Equation 10). This relationship indicates a moderate positive response, with P rising from 9.49 mg dm^-3 (Control) to 12.93 mg dm^-3 at 120 kg ha^-1 of K₂O.

Eq. 10 $\mathrm{P}=8.08+0.0404\times K_{\mathrm{dose}}({\mathrm{R}}^2=0.62)$

Potassium in soil

Soil potassium (K) content (Figure 8) showed a clear vertical stratification at all sampling times (p<0.0001), with consistently higher contents in the surface layer and decreasing values with increasing depth. The statistical analyses indicated that soil K was affected by depth in all cases, by K application to Urochloa brizantha in the intermediate and final samplings, and by K application to soybean in the final sampling. The interactions between factors were not significant, indicating additive, independent effects.

Figure 8
Vertical distribution of soil potassium (K) content (cmol_c dm^-3) at different layers (0.00–0.05, 0.05–0.10, 0.10–0.20, and 0.20–0.40 m): (a) before K application on Urochloa brizantha; (b) after K application on U. brizantha (0 and 100 kg ha^-1 of K), showing higher surface K under 100 kg ha^-1 of K (0.082 vs. 0.058 cmol_c dm^-3 at 0.00–0.05 m) and convergence at layer (~0.020 cmol_c dm^-3); and (c) after K application on soybean, displaying the residual effect of K applied to U. brizantha and the main effect of K applied to soybean (0, 25, 50, 100 kg ha^-1 of K). Although the interaction was not significant, all rates were plotted to illustrate the overall trend. The depth-weighted mean (0.00–0.40 m) followed the regression K = 0.0702 – 0.000207 × K_dose + 0.0000015 × K_dose² (R² = 0.62). Bars represent least-squares means ± standard error (n = 4).

Before K fertilization on Urochloa brizantha, soil K content differed significantly among soil layers (p<0.0001), reflecting the natural stratification typical of conservation systems. Potassium concentrations reached 0.069 cmol_c dm^-3 in the 0.00–0.05 m layer, 0.053 cmol_c dm^-3 at 0.05–0.10 m, 0.039 cmol_c dm^-3 at 0.10–0.20 m, and 0.029 cmol_c dm^-3 in the 0.20–0.40 m layer. This pattern indicates that, even before fertilization, the surface soil already had approximately twice the K concentration of the subsoil. The higher K availability near the surface can be attributed to nutrient cycling from plant residues and the limited mobility of K⁺ in the soil profile, even under sandy-textured conditions.

After K fertilization on U. brizantha, soil K was significantly influenced by K-Brachiaria (p = 0.0012), depth (p<0.0001), and their interaction (p = 0.0247). Potassium contents in the 0.00–0.05 m layer increased from 0.058 cmol_c dm^-3 in the control treatment to 0.082 cmol_c dm^-3 under 100 kg ha^-1 of K (120 kg ha^-1 of K₂O). In the 0.05–0.10 m layer, values were 0.051 and 0.069 cmol_c dm^-3, respectively, while in the 0.10–0.20 m layer, K decreased to 0.035 and 0.040 cmol_c dm^-3, and at 0.20–0.40 m, both treatments converged to ~0.020 cmol_c dm^-3. This behavior demonstrates that the increase in K availability after fertilization was restricted mainly to the surface layer, with limited translocation of the applied K to deeper horizons. The interaction effect confirmed that the surface application of K intensified nutrient concentration in the topsoil, reinforcing the vertical gradient typical of stratified profiles in conservation systems.

At the final sampling, after K fertilization on soybean, soil K was significantly affected by K applied to U. brizantha (p<0.0001), K applied to soybean (p = 0.0267), and depth (p<0.0001), while all interaction terms were not significant (p>0.05). Potassium concentrations decreased with increasing depth, from 0.082 cmol_c dm^-3 in the 0.00–0.05 m layer to 0.041 cmol_c dm^-3 in the 0.20–0.40 m layer, confirming the persistence of stratification. The residual effect of K applied to U. brizantha was evident, with mean K contents of 0.052 cmol_c dm^-3 and 0.075 cmol_c dm^-3 for 0 and 100 kg ha^-1 of K (0 and 120 kg ha^-1 of K₂O), respectively, showing that surface enrichment remained detectable after the subsequent soybean cycle.

The main effect of K applied to soybean also influenced soil K availability, with depth-weighted means (0.00–0.40 m) of 0.072, 0.056, 0.067, and 0.060 cmol_c dm^-3 for 0, 25, 50, and 100 kg ha^-1 of K (0, 30, 60, and 120 kg ha^-1 of K₂O), respectively. This pattern indicates a subtle response to K fertilization on soybean, best represented by a curvilinear relationship. Although the interaction K-Brachiaria × K-Soybean was not significant (p = 0.267), all K rates applied to soybean were represented in figure 8c, combined with both Brachiaria levels, to illustrate the general trend of K distribution and to ensure consistency with the presentation of soil P results (Figure 7).

A quadratic regression was fitted between the depth-weighted mean (0.00–0.40 m) soil K content and the K doses applied to soybean (Equation 11), which adequately represented the curvilinear behavior of the data.

Eq. 11 ${\mathrm{K}}^{+}({\mathrm{cmol}}_{\mathrm{c}}{\mathrm{dm}}^{-3})=0.0702-0.000207\times {\mathrm{K}}_{\mathrm{dose}}+0.0000015\times {{\mathrm{K}}_{\mathrm{dose}}}^2({\mathrm{R}}^2=0.62)$

This model shows a slight decline in soil K up to approximately 70 kg ha^-1 of K₂O, followed by a modest increase at higher doses, suggesting that moderate applications do not substantially enhance residual K in the soil, whereas higher rates may promote stabilization or minor recovery of available K.

These results demonstrate that soil depth is the main determinant of K distribution in the profile, followed by the residual effect of K applied to U. brizantha, which significantly increased surface K and remained detectable after soybean cultivation. The K applied to soybean also contributed to soil K availability, but to a lesser extent, with a quadratic response pattern. The absence of significant interactions between factors indicates that both fertilizations have independent and additive effects on soil K stratification, reinforcing the persistence of surface enrichment and the low mobility of K in sandy soils under no-tillage management.

DISCUSSION

Soybean grain yield

The soybean grain yield observed in the experiment was high under the study conditions, with the highest grain yield achieved when 100 kg ha^-1 of K was applied exclusively to Urochloa brizantha, slightly surpassing the Brazilian average soybean grain yield in 2021 (3,529 kg ha^-1) (Conab, 2019). In contrast, when K was applied directly to soybean at the highest dose (100 kg ha^-1 of K), without prior application to Urochloa brizantha, the maximum grain yield was still lower than that observed when K was applied only to Urochloa brizantha and not to soybean. This outcome suggests that early K application to Urochloa brizantha enhances nutrient availability and soil conditions more effectively than direct K application to soybean.

Moreover, the reduction in soybean grain yields observed when K was applied to soybean in plots where K had already been applied to Urochloa brizantha may be attributed to the potential toxic effects of chloride from the KCl fertilizer on the rhizosphere microbial community, particularly nitrogen-fixing bacteria. Studies have shown that high chloride contents can inhibit N fixation and plant growth in various legumes. Alfalfa exposed to NaCl exhibited reduced shoot and root biomass, with nitrogenase activity significantly inhibited and correlated to decreased shoot growth and total N content (Serraj and Drevon, 1998). Comparative studies on soybean, common bean, and alfalfa revealed species-specific responses to salt stress, with common bean showing higher sensitivity due to greater Na and Cl accumulation in nodules (Serraj and Drevon, 1998). These findings highlight the detrimental effects of chloride on nitrogen-fixing bacteria and plant growth. However, specific references directly linking KCl application and decreased soybean grain yields in field conditions are limited, suggesting the need for further research to confirm these observations in Brazilian soils.

In Cerrado environments, field trials with soybean in sandy to sandy-loam Oxisols have shown that grain yield generally increases with potassium fertilization up to approximately 66–100 kg ha^-1 of K, with smaller or no yield gains at higher rates (Mascarenhas et al., 1981, 1994; Rosolem et al., 1984; Oliveira et al., 1992; Lana et al., 2002; Leal et al., 2015; Júlio et al., 2017). These studies suggest that exchangeable soil K contents around 0.05–0.08 cmol_c dm^-3 (≈ 20–31 mg dm^-3) are generally sufficient to sustain high soybean yields, while values below 0.04 cmol_c dm^-3 (≈ 16 mg dm^-3) may limit productivity. In our study, soil K after fertilization with U. brizantha ranged from 0.052 to 0.075 cmol_c dm^-3 (≈ 20–29 mg dm^-3), matching this adequate range and supporting the high yields observed without in-season K fertilization.

The differential response of soybean grain yield to K management strategies underscores the importance of considering the interaction between K application timing and the specific needs of the crop and the cover crop. The results indicate that applying K to Urochloa brizantha can improve soybean grain yields primarily by enhancing nutrient cycling (Echer et al., 2020; Bortolli et al., 2024). Additionally, Brachiaria cover crops have been reported to improve soil structure and promote root development of subsequent crops (Rosolem and Steiner, 2017).

Dry matter decomposition of Urochloa residue

Decomposition of Urochloa brizantha residue was significantly influenced by K application, as reflected in the differential rates of initial biomass decay. Potassium application enhanced initial biomass production, which, in turn, influenced subsequent decomposition dynamics. Plant residue decomposition typically follows a two-phase pattern: an initial rapid phase followed by a slower phase (Gilmour et al., 1998; Pržulj and Tunguz, 2022). The first phase is characterized by the breakdown of water-soluble compounds and is influenced by initial N content, while the second phase involves the decomposition of lignin and phenols (Pržulj and Tunguz, 2022).

The accelerated decomposition rate with K application suggests enhanced microbial activity, likely due to improved nutrient availability, which stimulates the breakdown of organic matter. The application of K sources can significantly influence soil microbial activity and carbon dynamics. Studies also indicate that alternative K sources, such as K sulfate (K₂SO₄), can enhance microbial biomass carbon and CO₂ release when compared to KCl (Pereira et al., 2022). Although our experiment tested only KCl, these findings suggest that the accompanying anion may play a role in shaping microbial responses, an aspect that deserves further attention in tropical sandy soils. This faster nutrient cycling is crucial in sandy soils, where maintaining nutrient availability can be challenging due to low organic matter content. The shorter half-life of the residue with K application indicates that this practice can facilitate earlier nutrient release, providing timely benefits to subsequent crops.

This enhanced decomposition not only supports nutrient cycling but also improves soil structure and fertility by increasing organic matter turnover. Effective residue management, including strategic K application, can thus play a vital role in optimizing the sustainability and productivity of agricultural systems, particularly in tropical and subtropical regions, where nutrient turnover is naturally rapid and influences nutrient availability dynamics.

The findings emphasize the importance of considering K management for both crop yield and its role in residue decomposition and soil health. Through its effect on nutrient cycling, K application to Urochloa helps maintain the balance between nutrient release and availability for the subsequent soybean crop. Potassium absorbed and stored in Urochloa biomass is later released during residue breakdown, thereby increasing the soil residual pool of exchangeable K and improving fertilizer use efficiency. As a result, part of the soybean K demand is supplied by recycled K already present in the system, reducing the need for additional mineral K applications in subsequent seasons and, therefore, lowering dependency on external inputs and supporting sustainable nutrient management.

Nitrogen remaining and release dynamics

Nitrogen dynamics differed between treatments. With K fertilization, mulch N decreased initially but later exceeded its initial value, indicating immobilization of soil N driven by microbial activity. Without K, mulch N remained more stable (~69 % of initial values), reflecting slower decomposition and reduced microbial demand. Thus, K accelerated decomposition and increased microbial N demand, causing temporary immobilization that may restrict soybean N supply early, but was later compensated for by residue release.

The observed N immobilization, characterized by values exceeding the initial N content, underscores the dynamic interaction between mulch decomposition and soil N availability. This suggests that microbial communities may be immobilizing N from external sources, including soil, to support their metabolic needs during the later stages of residue breakdown, especially when more recalcitrant compounds, such as lignin, become dominant. The decomposition of high C: N crop residues often leads to N immobilization, in which microbial demand for N exceeds the supply from the residue itself (Knapp et al., 1983; Cao et al., 2020). This process is influenced by the quality of organic carbon, with labile compounds such as glucose stimulating greater microbial N immobilization than recalcitrant compounds (Cao et al., 2021).

Conversely, in the treatment without K application (K-Brachiaria 0), the N content in the mulch showed less temporal variation, averaging 69.1 % from 15 to 120 days post-deposition. This stability suggests a slower decomposition process with reduced microbial activity, likely due to the lower initial nutrient availability, which aligns with the slower rates of organic matter turnover observed in residues with lower K levels.

The dynamics of N release and immobilization depicted in figure 4 further emphasize the complexity of N cycling in cover crop residues under varying K regimes. The occurrence of negative release values, where N immobilization exceeds the initial N content, indicates an intricate balance among decomposition, nutrient release, and microbial nutrient demands. These results indicate that K application not only influences decomposition rates but also significantly alters N dynamics within the system, potentially affecting the timing and availability of N for subsequent crops.

Phosphorus remaining and release dynamics

The slower release of P from Urochloa brizantha mulch in plots receiving K fertilization can be attributed to the increased dry matter production induced by K fertilization, which also slowed mulch decomposition. This pattern mirrors the N dynamics observed, where enhanced biomass production led to reduced decomposition rates; however, unlike N, P did not exhibit immobilization episodes during decomposition. The mulch effectively retained P and released it gradually, which is critical for maintaining a steady nutrient supply to subsequent crops.

The decomposition of Urochloa brizantha mulch and its nutrient release dynamics are influenced by fertilization and plant density. Nitrogen fertilization of U. brizantha increases N and K cycling and release to subsequent crops (Werner et al., 2020). However, P release from U. brizantha straw is not affected by N fertilization or desiccation timing (Werner et al., 2020). Plant density impacts mulch decomposition and nutrient release, with higher densities generally resulting in greater nutrient content (Rocha et al., 2020).

Residue P release (15–29 kg ha^-1 of P≈ 34–66 kg ha^-1 of P₂O₅) was relevant considering that soils presented medium P levels according to the Manual de Adubação e Calagem para o Estado do Paraná (Mehlich extractor). Under these conditions, soybeans require 80–120 kg ha^-1 of P₂O₅. Thus, residues covered ~40–55 % of demand, complementing soil reserves. This substantial release underscores the potential of Urochloa brizantha mulch to significantly contribute to the P needs of subsequent crops, thereby reducing reliance on external P inputs and enhancing nutrient use efficiency within the system.

Furthermore, the observed reduction in soil P levels following K application on Urochloa brizantha highlights the role of the mulch as a P reservoir. This suggests that, after K application, Urochloa brizantha effectively sequesters P within its biomass, shifting the primary P pool from the soil to the plant material. This sequestration reduces immediate P availability in the soil but conserves it within the system for gradual release.

The dynamics observed, with significant declines in soil P concentrations at deeper layers and higher concentrations in the topsoil, emphasize the function of Urochloa brizantha mulch as a surface nutrient sink. This gradual release from the mulch during decomposition ensures a continuous and accessible P supply to subsequent crops, supporting crop growth phases that require steady nutrient input. Notably, the absence of P immobilization episodes, unlike the more pronounced N dynamics, indicates a direct availability pathway, reducing competitive microbial uptake and favoring plant access to P.

In summary, integrating Urochloa brizantha as a cover crop with targeted K applications offers a dual advantage: enhancing biomass production and creating a strategic P reservoir within the cropping system. This approach supports sustainable nutrient management by aligning P availability with crop demand, improving nutrient use efficiency, and reducing dependence on synthetic fertilizers. The capacity of Urochloa brizantha to act as both a nutrient buffer and a slow-release source underscores its valuable role in sustainable soil-plant systems, particularly in environments prone to nutrient loss and lower organic matter content.

Potassium release from Urochloa brizantha residue

The initial release of K from Urochloa brizantha mulch was rapid, with an average release of approximately 86.8 kg ha^-1 of K within the first 15 days after deposition. This amount is equivalent to approximately 104.4 kg ha^-1 of K₂O, reflecting the substantial contribution of K from the mulch to the soil nutrient pool during the early stages of decomposition. As the soil extractor used was Mehlich, the adopted guideline follows the Manual de Adubação e Calagem para o Estado do Paraná (Pauletti and Motta, 2017), which recommends 42-66 kg ha^-1 of K (50–80 kg ha^-1 of K₂O) for soybean. This rapid release aligns with the general behavior of K in plant residues, where the element is typically leached quickly due to its high solubility and mobility.

When comparing the K release dynamics observed for Urochloa brizantha with the unified kinetic release equation proposed by de Bortolli et al. (2024) for different residues, K_(remaining) = 90.2347 × exp⁻⁰·⁰⁶³³ × DAD + 9.77 (R² = 0.84), a similar pattern is evident. Both the Urochloa brizantha residues and the unified model describe a rapid initial release followed by a plateau phase, indicating that the same factors across different residue types predominantly govern K release.

The comparable release patterns suggest that K dynamics are consistently influenced by the initial availability and solubility of K within the plant tissues, irrespective of the specific cover crop used. This reinforces the idea that the timing of subsequent crop planting is crucial to maximize K uptake and minimize leaching losses, particularly in sandy soils where K retention is naturally limited.

These findings underscore the importance of synchronizing crop nutrient demands with the release dynamics of K from cover crop residues to optimize nutrient use efficiency in integrated crop-livestock systems, as emphasized by de Bortolli et al. (2024).

Potassium dynamics in the soil were evaluated before and after the K application on Urochloa brizantha and soybean. Initial observations showed that soil K content was influenced by depth, with the highest concentrations in the 0.00–0.05 m layer and significantly lower contents in deeper layers. Despite applying 100 kg ha^-1 of K to Urochloa brizantha and an additional 100 kg ha^-1 to soybean in some treatments, no substantial increase in K levels was observed in the deeper soil layers, indicating that the soil was not the primary reservoir of K in this system. The residual effect of K applied to U. brizantha was evident across samplings, maintaining surface enrichment with no significant increase in subsoil K concentrations, confirming the low mobility of K⁺ in the sandy profile under no-tillage.

Subsurface K concentrations (0.20–0.40 m) remained low and stable, while increases occurred only in surface soil. This suggests that no significant leaching occurred during the experimental period. However, we acknowledge that a definitive assessment would require temporal monitoring with soil solution or lysimeter studies.

Calculations based on the K application rates and soil parameters (1 ha to 0.20 m depth with soil density of 1 Mg m^-3) suggest that the expected increase in soil K should have been approximately 0.154 cmol_c dm^-3. However, the actual observed K levels did not reflect this expected increase, reinforcing that the majority of the K was retained in the mulch and plant biomass rather than in the soil itself.

Of the 100 kg ha^-1 of K applied, ~85.5 kg ha^-1 of K were released from residues, ~59 kg ha^-1 of K were exported in soybean grain, and ~18.3 kg ha^-1 of K were retained in soil (0.00–0.40 m). The small remainder is attributed to retention in structural tissues, roots, or other unmeasured fractions. This balance shows that most K cycled rapidly through biomass and residues rather than stabilizing in soil exchange sites.

This behavior is particularly significant in sandy soils, where low cation exchange capacity and high leaching potential limit K retention. The stability of K levels in the soil, despite substantial K inputs, suggests that the applied K was primarily retained within the mulch and plant biomass rather than in the soil. This finding aligns with the rapid initial release of K from Urochloa brizantha mulch, which meets immediate K needs of subsequent crops but does not significantly enrich soil K content over time.

The observed dynamics suggest that in sandy soils, the mulch and plant biomass play a crucial role in modulating K availability. This underscores the value of managing mulch and nutrient inputs strategically to optimize K use efficiency. The Urochloa brizantha mulch, through its role as a K reservoir, mitigates K losses and ensures a steady nutrient supply, which is critical in systems with low soil retention capacity.

CONCLUSIONS

This study demonstrated that early K fertilization via Urochloa brizantha as a cover crop is an effective strategy to enhance nutrient cycling and soybean productivity in sandy soils. Applying K to U. brizantha increased biomass production and promoted the release of essential nutrients to the subsequent crop. The amount of K released from residues exceeded 83 kg ha^-1 of K (100 kg ha^-1 of K₂O ), which is higher than the fertilization range typically recommended for soybean (42–66 kg ha^-1 of K; 50–80 kg ha^-1 of K₂O based on Mehlich extraction). This release was sufficient to meet and even surpass soybean K demand, thereby reducing the need for additional K fertilization.

Phosphorus (P) release from U. brizantha residues was lower when K was applied to the grass, likely due to a dilution effect from greater biomass accumulation. Nevertheless, the cumulative release reached 15–29 kg ha^-1 of P (≈34–66 kg ha^-1 of P₂O₅), partially meeting soybean demand (~54 kg ha^-1 of P₂O₅ for the observed yield) and highlighting the role of residues as a complementary P source.

Overall, the findings indicate that K fertilization targeted to U. brizantha cover crops can optimize nutrient use efficiency by synchronizing nutrient release with soybean demand. This strategy not only sustained high soybean grain yields but also reduced reliance on direct fertilizer inputs, thereby contributing to a more efficient and resilient soil fertility management system.

ACKNOWLEDGMENTS

The authors would like to acknowledge the LabSolos and LAQUA from the Universidade Tecnológica Federal do Paraná (UTFPR), Campus Pato Branco, for the analytical support. This study was carried out with the support of the Federal University of Mato Grosso do Sul – UFMS/MEC – Brazil. This study was supported by the Coordination for the Improvement of Higher Education Personnel – Brazil (CAPES) – Funding Code 001.

• How to cite:
  Kagimura LT, Assmann TS, Assmann AP, Ratke RF, Godoy ML, Severo IK. Enhancing nutrient cycling and soybean grain yield in sandy soils via early potassium on Urochloa brizantha. Rev Bras Cienc Solo. 2026;50:e0240210. https://doi.org/10.36783/18069657rbcs20240210

DATA AVAILABILITY

All data was generated or analyzed in this study.

REFERENCES

• Aita C, Giacomini SJ. Decomposição e liberação de nitrogênio de resíduos culturais de plantas de cobertura de solo solteiras e consorciadas. Rev Bras Cienc Solo. 2003;27:601-12. https://doi.org/10.1590/S0100-06832003000400004
  » https://doi.org/10.1590/S0100-06832003000400004
• Bortolli MA, Assmann TS, Bortolli BB, Maccari M, Bernardon A, Jamhour J, Franzluebbers AJ, Soares AB, Severo IK. Nutrient dynamics in integrated crop–livestock systems: effects of stocking rates and nitrogen fertilization on litter decomposition and release. Agronomy. 2024;14:2009. https://doi.org/10.3390/agronomy14092009
  » https://doi.org/10.3390/agronomy14092009
• Cao Y, He Z, Zhu T, Zhao F. Organic-C quality as a key driver of microbial nitrogen immobilization in soil: A meta-analysis. Geoderma. 2021;383:114784. https://doi.org/10.1016/j.geoderma.2020.114784
  » https://doi.org/10.1016/j.geoderma.2020.114784
• Cao Y, Zhao F, Zhang Z, Zhu T, Xiao HY. Biotic and abiotic nitrogen immobilization in soil incorporated with crop residue. Soil Till Res. 2020;202:104664. https://doi.org/10.1016/j.still.2020.104664
  » https://doi.org/10.1016/j.still.2020.104664
• Companhia Nacional de Abastecimento - Conab. Acompanhamento da safra brasileira de grãos - Safra 2020/21. Brasília, DF: Conab; 2019. Available from: www.conab.gov.br
  » www.conab.gov.br
• Echer FR, Souza Peres VJ, Rosolem CA. Potassium application to the cover crop prior to cotton planting as a fertilization strategy in sandy soils. Sci Rep. 2020;10:20404. https://doi.org/10.1038/s41598-020-77354-x
  » https://doi.org/10.1038/s41598-020-77354-x
• Echer FR, Volf MR, Souza Peres VJ, Hafemann JA, Silva GRA, Rosolem CA. Ruzigrass as cover crop improves the potassium partial balance and use by subsequent cotton. Nutr Cycl Agroecosyst. 2023;127:347-58. https://doi.org/10.1007/s10705-023-10312-x
  » https://doi.org/10.1007/s10705-023-10312-x
• Gama-Rodrigues AC, Gama-Rodrigues EF, Brito EC. Decomposição e liberação de nutrientes de resíduos culturais de plantas de cobertura em Argissolo Vermelho-Amarelo na região Noroeste Fluminense (RJ). Rev Bras Cienc Solo. 2007;31:1421-8. https://doi.org/10.1590/S0100-06832007000600019
  » https://doi.org/10.1590/S0100-06832007000600019
• Giacomini SJ, Aita C, Hübner AP, Lunkes A, Guidini E. Liberação de fósforo e potássio durante a decomposição de resíduos culturais em plantio direto. Pesq Agropec Bras. 2003;38:1097-104. https://doi.org/10.1590/S0100-204X2003000900011
  » https://doi.org/10.1590/S0100-204X2003000900011
• Gilmour JT, Norman RJ, Mauromoustakos A, Gale PM. Kinetics of crop residue decomposition: variability among crops and years. Soil Sci Soc Am J. 1998;62:750-5. https://doi.org/10.2136/sssaj1998.03615995006200030030x
  » https://doi.org/10.2136/sssaj1998.03615995006200030030x
• Johnson R, Vishwakarma K, Hossen MS, Kumar V, Shackira AM, Puthur JT, Abdi G, Sarraf M, Hasanuzzamanal M. Potassium in plants: growth regulation, signaling, and environmental stress tolerance. Plant Physiol Biochem. 2022;172:56-69 https://doi.org/10.1016/j.plaphy.2022.01.001
  » https://doi.org/10.1016/j.plaphy.2022.01.001
• Júlio OLL, Ascari JP, Mendes IRN, Santos ES, Duarte WM, Nied AH. Formas de adubação potássica e produtividade da cultura da soja. Rev Agrarian. 2017;9:149-55.
• Knapp EB, Elliott LF, Campbell GS. Carbon, nitrogen and microbial biomass interrelationships during the decomposition of wheat straw: a mechanistic simulation model. Soil Biol Biochem. 1983;15:455-61. https://doi.org/10.1016/0038-0717(83)90011-1
  » https://doi.org/10.1016/0038-0717(83)90011-1
• Lana RMQ, Hamawaki OT, Lima LMSD, Zanão Júnior LA. Resposta da soja a doses e modos de aplicação de potássio em solo de cerrado. Biosci J. 2002;18:17-23.
• Leal AJF, Valderrama M, Kaneko FH, Alves U, Leal S, Perin A, Luchese KUO. Produtividade da soja de acordo com diferentes doses de cloreto de potássio revestido ou não com polímero. Glob Sci Technol. 2015;8:19-30. https://doi.org/10.14688/1984-3801/gst.v8n1p19-30
  » https://doi.org/10.14688/1984-3801/gst.v8n1p19-30
• Mascarenhas H, Valadares JMAS, Rotta CL, Bulisani EA. Adubação potássica na produção de soja, nos teores de potássio nas folhas e na disponibilidade de potássio em Latossolo Roxo distrófico de cerrado. Bragantia. 1981;40:125-34. https://doi.org/10.1590/S0006-87051981000100012
  » https://doi.org/10.1590/S0006-87051981000100012
• Mascarenhas HAA, Tanaka RT, Pereira JCVNA, Gallo PB, Bataglia OC. Efeito de adubos potássicos na produção de soja. Sci Agric. 1994;51:82-9. https://doi.org/10.1590/S0103-90161994000100013
  » https://doi.org/10.1590/S0103-90161994000100013
• Oliveira FA, Silva JJSE, Vilela L, Sousa DMG. Doses e métodos de aplicação de potássio na soja em solo dos cerrados da Bahia. Pesq Agropec Bras. 1992;27:1485-95.
• Paul EA, Clark FE. Soil microbiology and biochemistry. 2nd ed. London: Academic Press; 1989.
• Pauletti V, Motta ACV. Manual de adubação e calagem para o Estado do Paraná. Curitiba: SBCS/NEPAR; 2017.
• Pereira DGC, Silva MJG, Meira MWP, Santos HSN, Santos Neto JA, Megda MM, Megda MXV. Microbial activity and carbon rates in the soil in response to the application of potassium sources. Acta Sci Biol Sci. 2022;44:e58248. https://doi.org/10.4025/actascibiolsci.v44i1.58248
  » https://doi.org/10.4025/actascibiolsci.v44i1.58248
• Pereira NS, Soares I, Miranda FR. Decomposition and nutrient release of leguminous green manure species in the Jaguaribe-Apodi region, Ceará, Brazil. Cienc Rural. 2016;46:970-5. https://doi.org/10.1590/0103-8478cr20140468
  » https://doi.org/10.1590/0103-8478cr20140468
• Potrich DC, Marchetti ME, Ensinas SC, Serra AP. Decomposição de resíduos culturais de cana-de-açúcar submetidos a diferentes doses de nitrogênio. Semina – Cienc Agrar. 2014;35:1751-60. https://doi.org/10.5433/1679-0359.2014v35n4p1751
  » https://doi.org/10.5433/1679-0359.2014v35n4p1751
• Pržulj N, Tunguz V. Significance of harvest residues in sustainable management of arable land I. Decomposition of harvest residues. Arch Tech Sci. 2022;26:61-70. https://doi.org/10.7251/afts.2022.1426.061P
  » https://doi.org/10.7251/afts.2022.1426.061P
• Rocha JG, Coelho FC, Lelis RT, Santos GCS, Jaeggi MEPC. Mulch decomposition and nitrogen, phosphorus and potassium release in vetiver grass at different plant densities. Pesq Agropec Trop. 2020;50:e64603. https://doi.org/10.1590/1983-40632020v5064603.
  » https://doi.org/10.1590/1983-40632020v5064603
• Rossi CQ, Alves RE A, Fernandes PRT, Pereira MG, Ribeiro RLDA, Polidoro JC. Liberação de macronutrientes de resíduos do consórcio entre mucuna preta e milho sob sistema orgânico de produção. Rev Cienc Vida. 2008;28:01-10.
• Rossi CQ, Pereira MG, Giácomo SGM, Bettam S, Polidoro JC. Decomposição e liberação de nutrientes da palhada de braquiária, sorgo e soja em áreas de plantio direto no cerrado goiano. Semina: Cienc Agrar. 2013;34:1523-34. https://doi.org/10.5433/1679-0359.2013v34n4p1523
  » https://doi.org/10.5433/1679-0359.2013v34n4p1523
• Rosolem CA, Nakagawa J, Machado JR. Adubação potássica da soja em Latossolo Vermelho-Escuro fase arenosa. Pesq Agropec Bras. 1984;19:1319-26.
• Rosolem CA, Steiner F. Effects of soil texture and rates of K input on potassium balance in tropical soil. Eur J Soil Sci. 2017;68:658-66. https://doi.org/10.1111/ejss.12460
  » https://doi.org/10.1111/ejss.12460
• Santana J, Vieira FA, Souto JS, Gondim SC, Fonseca FCPD. Decomposição da biomassa foliar de cana-de-açúcar em um Neossolo na região de Areia-PB. Rev Caatinga. 2011;24:28-32.
• Santos FC, Albuquerque Filho MR, Vilela L, Ferreira GB, Carvalho MCS, Viana JHM. Decomposição e liberação de macronutrientes da palhada de milho e braquiária, sob integração lavoura-pecuária no cerrado baiano. Rev Bras Cienc Solo. 2014;38:1855-61. https://doi.org/10.1590/S0100-06832014000600020
  » https://doi.org/10.1590/S0100-06832014000600020
• Serraj R, Drevon J. Effects of salinity and nitrogen source on growth and nitrogen fixation in alfalfa. J Plant Nutr. 1998;21:1805-18. https://doi.org/10.1080/01904169809365525
  » https://doi.org/10.1080/01904169809365525
• Tedesco MJ, Gianello C, Bissani CA, Bohnen H, Volkweiss SJ. Análises de solo, plantas e outrosmateriais. 2. ed. Porto Alegre: Universidade Federal do Rio Grande do Sul; 1995. (Boletim técnico, 5).
• Torres JLR, Pereira MG. Produção e decomposição de resíduos culturais antecedendo milho e soja num Latossolo no cerrado mineiro. Com Sci. 2014;5:419-29. https://doi.org/10.14295/cs.v5i4.508
  » https://doi.org/10.14295/cs.v5i4.508
• Vitti AC, Trivelin PCO, Cantarella H, Franco HCJ, Faroni CE, Otto R, Trivelin MO, Tovajar JG. Mineralização da palhada e crescimento de raízes de cana-de-açúcar relacionados com a adubação nitrogenada de plantio. Rev Bras Cienc Solo. 2008;32:2757-62. https://doi.org/10.1590/S0100-06832008000700020
  » https://doi.org/10.1590/S0100-06832008000700020
• Werner F, Ferreira AS, Balbinot Junior AA, Oliveira Junior A, Franchini JC, Debiasi H, Silva MAA. Nitrogen, phosphorus, and potassium released by decomposition of palisade grass to soybean in succession. Pesq Agropec Bras. 2020;55:e01853. https://doi.org/10.1590/S1678-3921.pab2020.v55.01853
  » https://doi.org/10.1590/S1678-3921.pab2020.v55.01853

Edited by

• Editor:
  José Miguel Reichert https://orcid.org/0000-0001-9943-2898 and Tales Tiecher https://orcid.org/0000-0001-5612-2849

Publication Dates

• Publication in this collection
  17 Aug 2026
• Date of issue
  2026

History

• Received
  03 Apr 2025
• Accepted
  12 Jan 2026