Soil Bioindicators and Crop Productivity Affected by Legacy Phosphate Fertilization and Azospirillum brasilense Inoculation in No-Till Systems

Abstract

Pressure on agroecosystems is increasing with rising agricultural demand, pushing Brazilian agriculture toward more sustainable systems that prioritize soil health. This study aimed to evaluate whether long-term no-till management and inoculation with Azospirillum brasilense influenced soil bioindicators; chemical, biological, and enzymatic attributes; and how these attributes correlated with crop productivity in a rotational system. The experiment also assessed the residual effects of phosphate fertilization (initially applied in 2013 and reapplied in 2020) and its interaction with inoculation on soil phosphorus fractions and crop performance. This study was conducted on Dystrophic Red Oxisol in the low-altitude Cerrado region under 20 years of no-tillage management, using a randomized block design in a 5 × 2 factorial scheme: five phosphorus doses (0, 30, 60, 120, and 240 kg ha⁻¹ P₂O₅) and inoculated or non-inoculated grasses, with four replicates. The results showed that inoculation influenced dry matter (DM) production and nutrient cycling, improving soil health despite lower fertility and total DM. The correlation between bioindicators and productivity suggests that soil health indicators can be used to monitor system sustainability. No consistent effects of inoculation or phosphate fertilization were observed for some crop components, indicating complex interactions under long-term conservationist systems.

1. Introduction

The world population is expected to reach 9.7 billion people by 2050 and peak at the end of the century, in 2100, with approximately 11 billion inhabitants [1]. Therefore, the development of a solid and stable society directly and indispensably depends on food, which is a basic human necessity that can never be ignored. However, the world is already facing high levels of hunger, which could be further aggravated by population growth, affecting society’s quality of life and directly and indirectly impacting global natural resources [2].

The pressure on agroecosystems intensifies with the increasing demand for agricultural products and the growing public demand for high-quality food, more sustainable production systems, and lower environmental impact practices that preserve biodiversity and societal well-being. This scenario presents a major challenge for the global agricultural sector, as the negative impacts of food production for a growing population still cause environmental and ecological imbalances, putting the sustainability of global natural resources at risk [3].

Thus, environmental degradation caused by anthropogenic activities has resulted in significant declines in productivity and soil capacity to fulfill its ecological functions [4]. Soil degradation involves changes in its physical, chemical, and biological properties, such as reduced nutrient levels, organic matter depletion, and alterations in soil density, porosity, structure, aggregate stability, and infiltration capacity [5].

The lack of soil cover or inadequate coverage for prolonged periods has been one of the main factors contributing to soil exhaustion and degradation. In this context, the No-Tillage System (NTS) is a conservationist management practice that increases the volume and quality of soil cover. This system is based on three fundamental principles: no soil disturbance, crop rotation, and permanent soil cover with straw. Therefore, for the success of this system, crop rotation should be based on crops that provide a high volume and variety of crop residues, compensating for the rapid decomposition of organic material in tropical and subtropical climates [6].

To enhance straw production in areas under NTS, the cultivation of grasses in crop rotation is recommended. In this regard, the adoption of Integrated Crop–Livestock–Forestry Systems (ICLFSs) presents a successful alternative, as they combine different agricultural activities, such as livestock, crops, and forestry, in the same area, either in intercropping or rotation. These systems benefit from the synergies among activities through interactions that promote the more efficient use of available biotic and abiotic factors, increasing crop productivity and profitability while contributing to soil quality improvements, mainly due to the large volume of plant residues left by various system components [7].

In addition to ICLFSs under NTS, the inoculation of crops with Plant Growth-Promoting Rhizobacteria (PGPR) contributes to increased productivity. PGPRs represent a vast group of microorganisms that associate with plants without causing productivity losses and promote crop growth and development through direct and/or indirect mechanisms. In recent decades, due to the growing interest in and concern for more sustainable and ecological agriculture, there has been an increased search for production practices that reduce mineral fertilizer use, such as inoculation with PGPR [8].

However, several challenges remain for increasing and stabilizing crop productivity, as tropical soils typically have low nutrient concentrations, with phosphorus (P) and nitrogen (N) being the most limiting elements for production in Brazil [9]. Phosphorus, in addition to having a high fixation and immobilization potential in the soil, also comes from finite and non-renewable sources. Consequently, the debate over the inability to meet future food demands sustainably has been revisited in the scientific community [10].

Given this scenario, research is needed to address the peculiarities of phosphorus dynamics in the soil and its use in more efficient management systems, involving production, the decomposition of residues, and nutrient cycling. These studies aim to increase crop productivity and phosphorus inputs into the soil, making it possible to reduce the use of mineral fertilizers. If conducted responsibly, the agricultural sector has the potential to mitigate the environmental damage caused by food production while contributing to environmental conservation. Therefore, research investigating management practices and technologies related to this issue is of great social, economic, and ecological importance. The significance of these studies, combined with the fact that in more complex production systems, soil attributes can be used reliably to assess soil quality and conservation, underscores the need for further research in this area. Studies evaluating the physical, chemical, and biological characteristics of soils under NTS and/or ICLFSs have been somewhat neglected.

We hypothesized that long-term no-till systems combined with crop inoculation by Azospirillum brasilense modify soil chemical and biological properties in ways that enhance enzymatic activity, making these bioindicators reliable tools for evaluating soil health in sustainable cropping systems. We also hypothesized that such inoculation promotes root development and enhances phosphorus uptake, particularly in low-mobility forms, thereby increasing crop productivity and enabling the more efficient use of residual phosphorus in the soil.

This study aimed to investigate whether long-term conservationist management practices, such as no-tillage and inoculation with A. brasilense, influenced soil bioindicators, including chemical, biological, and enzymatic attributes, and how these attributes correlated with crop productivity. Additionally, it evaluated the residual effects of phosphate fertilization and its interactions with inoculation on soil phosphorus fractions and the productivity of crops grown in rotation systems.

2. Materials and Methods

The experiment was conducted at the Teaching, Research, and Extension Farm (FEPE)—Plant Production Sector, belonging to the Faculty of Engineering of Ilha Solteira (FEIS)/São Paulo State University (UNESP), located in the municipality of Selvíria/ Mato Grosso do Sul (Scheme 1), with geographic coordinates 20°20′45″ S and 51°24′25″ W and an altitude of 350 m.

The relief of the area is characterized as moderately flat, with soil classified as a typical Dystrophic Red Oxisol [11], with a clayey texture, updated according to the Brazilian Soil Classification System [12]. The climate type is Aw, according to Köppen’s classification [13], described as humid tropical with a rainy season in summer and a dry season in winter. The average annual precipitation is 1370 mm and the average temperature is 23.5 °C [14]. During the experiment, climatic data were collected from the meteorological station installed at FEIS (Figure 1).

The experimental area has a history of annual crop cultivation under no-tillage (NT) for 20 years. Before the current study, the following crops were grown in rotation (Scheme 2).

The experimental design was a randomized complete block design (RCBD), in a 5 × 2 factorial scheme; treatments combined five P₂O₅ doses (0, 30, 60, 120, and 240 kg ha⁻¹) with two inoculation levels (with/without A. brasilense), using plots with 0 kg ha⁻¹ of P₂O₅ and no inoculation as controls. Fertilization was applied using monoammonium phosphate (MAP—11% N, 52% P₂O₅), aiming to evaluate phosphorus complexation in assimilable forms. The recommended dose, based on an initial soil analysis conducted in 2013 (Bulletin 100), was 60 kg ha⁻¹ of P₂O₅. The other rates were established as half (30 kg ha⁻¹), double (120 kg ha⁻¹), quadruple (240 kg ha⁻¹), and zero of the recommended dose. To ensure that nitrogen was not a limiting or confounding factor in the evaluation of phosphorus effects, the amount of N provided by MAP was calculated for each P₂O₅ dose, and complementary N was applied using urea (46% N) so that all treatments received a uniform nitrogen supply of 50.77 kg ha⁻¹. The urea rates applied to each treatment were 110.37, 96.57, 82.78, 55.20, and 0 kg ha⁻¹ for 0, 30, 60, 120, and 240 kg ha⁻¹ of P₂O₅, respectively. Phosphorus application was broadcast without incorporation during black oat sowing in 2013 and reapplied in 2020, together with the inoculation (or not) of the grain-producing grasses in the crop rotation (Scheme 2) with A. brasilense. There were four replications per treatment. Each experimental unit (plot) was 3.2 m wide and 10 m long, covering an area of 32 m². The experiment had four blocks for each area (inoculated and non-inoculated), with five P₂O₅ rates per block, totaling an experimental area of 1280 m², excluding the buffer zone surrounding the entire experimental area, which was established to prevent interference from adjacent plots and ensure the integrity of treatment effects.

Before the implementation of the experiment and surface liming, a chemical analysis of the soil was conducted in the 0.00–0.20 m layer. The results indicated an acidic pH (CaCl₂) of 5.0, organic matter content of 25 g dm⁻³, and available phosphorus (P, resin method) of 22 mg dm⁻³. Exchangeable cations included 2.2 mmolc dm⁻³ of potassium (K), 19.0 mmolc dm⁻³ of calcium (Ca), and 14.0 mmolc dm⁻³ of magnesium (Mg). The concentration of sulfur (S) was 3.0 mmolc dm⁻³. Potential acidity (H⁺+Al³⁺) was measured at 36 mmolc dm⁻³. The base sum (SB) was 35.2 mmolc dm⁻³ and the base saturation (V%) was 49%.

The evaluations began in the fallow area with Urochloa brizantha (BRS Paiaguás) (Figure 2), following maize harvest in an intercropping system. Sowing took place on 8 November 2021 using a no-till seeder with a shank-type furrow opener at a depth of approximately 0.07 m, with a row spacing of 0.34 m and a seed rate of approximately 7.8 kg ha⁻¹, with a cultural value (CV) of 60%. Fertilization consisted of 100 kg ha⁻¹ of potassium chloride (KCl) at sowing and 500 kg ha⁻¹ of the 20-00-20 formulation applied as topdressing on 7 December 2021.

To evaluate soil fertility in the experimental units, soil samples were collected on 6 May 2022. Ten simple samples per plot were combined into a composite sample and analyzed for chemical attributes in the 0.00–0.10 m and 0.10–0.20 m layers.

To determine the dry matter yield of Paiaguás grass, plant samples were collected on 15 October 2022 using a 1.0 × 1.0 m sampling frame per experimental plot. The collected plants were placed in labeled plastic bags and weighed using a precision balance to determine fresh matter (FM). Subsamples were then taken, placed in paper bags, and oven-dried at 65 °C for 72 h to determine dry matter (DM), which was extrapolated to yield per hectare. Paiaguás grass was desiccated on 17 October 2022 using glyphosate (1560 g ha⁻¹ active ingredient (a.i.)) to prepare for the next crop in the rotation.

Next, soybean (Glycine max, cultivar 97R50IPRO) was sown (Figure 3) on 8 November 2022. Seeds were inoculated with Bradyrhizobium japonicum (strains SEMIA 5079 and SEMIA 5080) at 185 mL per 50 kg of seed immediately before sowing in a shaded area. Sowing was performed mechanically using a no-till seeder with a shank-type furrow opener at a depth of approximately 0.05 m, with a row spacing of 0.45 m, a seeding rate of approximately 14.4 seeds m⁻¹, and fertilization with 136 kg ha⁻¹ of KCl.

Harvest occurred on 25 March 2023, with morphological and yield assessments taken beforehand, including plant population (PP), determined by counting plants in the useful plot area (3 central rows, each 3 m long, totaling 9 m²) and converting it to plants per hectare, and soybean grain yield (SGY), measured by manually harvesting plants from the useful plot area, threshing them, and weighing the grains. Yield per plot was extrapolated to kg ha⁻¹ and adjusted to 13% moisture.

In ten plants from the useful area per plot, the following parameters were determined: plant height (PlT), measuring the average distance from the soil surface to the upper extremity of the soybean using a graduated tape; height of the first pod insertion (HFPI), measuring the average distance from the soil surface to the first pod insertion using a graduated tape; number of pods per plant (NPP), obtained by counting the total number of pods in the ten plants and subsequently calculating the average per plant; number of grains per plant (NGP), obtained by counting the total number of grains in the ten plants and subsequently calculating the average per plant; and the mass of 100 grains (M100), determined by counting one hundred grains per plot using an electronic counter, measuring their mass with a precision electronic balance (0.01 g), and adjusting moisture to 13% (wet basis).

Subsequently, grain sorghum (cultivar 84G05) was intercropped with Urochloa brizantha (BRS Piatã) (Figure 4). Sowing occurred on 3 May 2023, with only sorghum seeds inoculated with A. brasilense (strains Ab-V5 and Ab-V6 at 2 × 10⁸ colony-forming units (CFUs) mL⁻¹) at 100 mL per 50 kg of seed immediately before sowing in a shaded area. Sorghum was sown at an approximate 0.05 m depth with a 0.45 m row spacing and a seeding rate of about 7.9 seeds m⁻¹. Piatã grass was sown at an approximate 0.07 m depth with a 0.34 m row spacing and a seed rate of 6.0 kg ha⁻¹ (CV 60%). Fertilization included 300 kg ha⁻¹ of 20-00-20 (NPK) topdressing applied on 2 June 2023.

The harvest took place on 28 August 2023, at which time the morphological and productivity evaluations of the crops were carried out. For the sorghum crop, the following parameters were assessed: plant population (PP), obtained by counting the number of plants in the useful area of each plot (three central rows of 3 m, totaling 9 m), later converted into the total number of plants per hectare, and sorghum grain yield (SGY), where the plants in the useful area of the plots were manually harvested and threshed and the grains weighed to determine productivity per plot. This was then extrapolated to SGY in kg ha⁻¹, correcting moisture to 13% (wet basis).

In ten plants from the useful area per plot, the following parameters were determined: plant height (PlH), evaluating the average distance from the soil surface to the upper extremity of the sorghum using a graduated measuring tape; panicle length (PL), obtained by measuring the average panicle length using a graduated measuring tape; basal stem diameter (BSD), measured at the basal third of the stem using a caliper graduated in millimeters, with the average diameter being calculated; 1000-grain mass (M1000), obtained by counting one thousand grains per plot using an electronic counter and determining the 1000-grain mass on a precision electronic balance (0.01 g), correcting moisture to 13% (wet basis); and fresh matter yield (FM) and dry matter (DM) of leaves, stem, and panicle separately, where ten plants from the useful area were manually separated into leaves, stem, and panicle. The plant material was then weighed to obtain the FM, placed in paper bags, and taken to a forced-air oven at 65 °C for 72 h to obtain the DM. The values were then extrapolated to production per hectare.

Simultaneously, the Piatã grass intercropped culture was evaluated for dry matter yield determination. Using a 1.0 × 1.0 m sampling frame, plants within 1 m² of each plot were collected, placed in properly labeled plastic bags, and weighed on a precision balance to determine FM. After this procedure, subsamples were taken, placed in paper bags, and taken to a forced-air oven at 65 °C for 72 h to obtain DM, which was later extrapolated to production per hectare.

To evaluate soil fertility characteristics, biological attributes, enzymatic activity, and soil P fractions in the experimental units, ten simple soil samples were collected on 26 September 2023 to form a composite sample per plot for subsequent analyses at depths of 0.00 to 0.10 m and 0.10 to 0.20 m.

After harvesting the intercropping system, the area remained fallow with Piatã grass (Figure 5). On 6 October 2023, for the dry matter yield determination of Piatã grass, a 1.0 × 1.0 m sampling frame was used to collect plants within 1 m² of each plot. These plants were placed in properly labeled plastic bags and weighed on a precision balance for FM determination. Again, after this procedure, subsamples were taken, placed in paper bags, and taken to a forced-air oven at 65 °C for 72 h to obtain DM, which was later extrapolated to production per hectare. The Piatã grass was desiccated on 13 October 2023 using the herbicide glyphosate (1560 g ha⁻¹ of active ingredient) to proceed with the next crop rotation.

For soil biological attribute evaluations, the following were determined: microbial respiratory activity (MRA), according to the methodology proposed by [15] and microbial biomass carbon (MBC), according to the methodology proposed by [16], with both analyses (MRA and MBC) performed at the Soil Microbiology Laboratory of the “Luiz de Queiroz” College of Agriculture, University of São Paulo (ESALQ/USP—Piracicaba/SP). The metabolic quotient (qCO₂) was calculated as the ratio between MRA and MBC, as proposed by [17], and total organic carbon (TOC) was determined using the method proposed by [18], performed at the Soil Microbiology Laboratory of the Brazilian Agricultural Research Corporation (Embrapa Cerrados—Brasília/DF).

For soil enzymatic activity analyses, β-glucosidase (ßG), aryl sulfatase (AS), and acid phosphatase (AF) were determined using methodologies proposed by [19]. These determinations were performed at the Soil Microbiology Laboratory of the Brazilian Agricultural Research Corporation (Embrapa Cerrados—Brasília/DF).

Soil health was assessed using the four-quadrant model proposed by [20], which evaluates soil carbon modifications (loss, gain, or stability) based on the relationship between TOC and the average activity of the ßG and AS enzymes per unit of TOC.

Soil phosphorus fractionation into labile phosphorus (Lab), moderately labile phosphorus (Mod Lab), non-labile phosphorus (N Lab), inorganic phosphorus (Pi), organic phosphorus (Po), and total phosphorus (P tot) was performed according to the methodology proposed by [21], with adaptations. Analyses were conducted at the Soil Chemistry Analysis Laboratory of the “Luiz de Queiroz” College of Agriculture, University of São Paulo (ESALQ/USP—Piracicaba/SP).

Soil fertility analyses included chemical attributes such as phosphorus (P), sulfur (S), potassium (K), calcium (Ca), magnesium (Mg), potential acidity (H+Al), base sum (BS), cation exchange capacity (CEC), base saturation (V), aluminum saturation (m), organic matter content (OM), and hydrogen potential values (pH), conducted according to the methodology proposed by [22] at the Soil Fertility Laboratory of São Paulo State University “Júlio de Mesquita Filho” (UNESP—Ilha Solteira/SP).

Soil chemical analyses were expressed on a volumetric basis (mmolc dm⁻³), in accordance with the standard methodology adopted in Brazil [9]. This approach is widely used in Brazilian laboratories, where soil samples are processed as air-dried fine earth (TFSA), involving drying, gentle disaggregation, and sieving through a 2 mm mesh. Given this preparation, a bulk density of 1.0 g cm⁻³ was assumed, which was conventionally applied under these conditions to represent the mass–volume relationship of TFSA. This value enabled the approximate conversion of nutrient concentrations to a mass-based expression (e.g., mmolc kg⁻¹), commonly used in the international literature. The bulk density was not measured in the field, as it did not apply to TFSA, but the adopted value aligned with official Brazilian procedures and ensured comparability of the results.

The physical fractionation of soil organic matter included particulate organic carbon (POC) and mineral organic carbon (MOC), following the methodology proposed by [23]. Total organic carbon (TOC) was analyzed according to the methodology described by [24], conducted at the Soil Physics and Geology Laboratories of São Paulo State University “Júlio de Mesquita Filho” (UNESP—Ilha Solteira/SP).

Soil sampling was carried out using a Stihl BT 45 soil auger. For biological attributes, enzymatic activity, and phosphorus fractionation, soil was collected at a 0.00 to 0.10 m depth, whereas, for soil chemical attributes and the physical fractionation of organic matter, samples were collected at 0.00 to 0.10 m and 0.10 to 0.20 m depths. Samples were placed in plastic bags, air-dried, de-clodded, sieved through a 5 mm mesh (fine air-dried soil), and then subdivided for each analysis.

For the comparison of soil chemical attribute analyses between the 2021/2022 and 2023/2024 crop seasons in the 0.00–0.10 m and 0.10–0.20 m layers, a calculation was performed to determine the difference between the analysis levels and, thus, the percentage increase or decrease in soil chemical attributes between seasons was calculated.

The data were subjected to analysis of variance (ANOVA) using the F-test at a 5% significance level (p < 0.05) to evaluate the effects of the treatments. When significant effects were observed, treatment means were compared using the LSD post-hoc test at the same significance level, based on its suitability for pairwise comparisons in factorial experiments with balanced designs. Regression analysis was applied to evaluate the response to phosphate fertilizer doses, and the model (linear or quadratic) was selected based on the significance of the coefficients and the coefficient of determination (R²). Statistical analyses were conducted using SISVAR^® software (Version: 5.6) [25], which assumed data normality and the homogeneity of variances, conditions met by the dataset according to the preliminary diagnostic tests performed.

3. Results

Regarding the phosphorus dose variation, a significant quadratic relationship was observed for soil phosphorus content in the 0.00–0.10 m layer, as illustrated in Figure 6. The regression analysis revealed a maximum P content at the calculated dose of 192.5 kg ha⁻¹ of P₂O₅, suggesting a saturation point beyond which additional phosphorus did not lead to proportional increases in soil P levels. This finding is essential to understanding phosphorus dynamics under no-till systems, as it highlights the limit of soil retention capacity and may reflect factors such as plant uptake, microbial immobilization, or sorption. Thus, the regression model supports more precise recommendations for phosphorus management in tropical soils.

For the inoculation factor, potassium (K), base sum (BS), cation exchange capacity (CEC), aluminum saturation (m%), organic matter (OM), and hydrogen potential (pH) did not show significant differences. For phosphorus (P), calcium (Ca), magnesium (Mg), and base saturation (V%), there was an increase in areas without inoculation, whereas only for sulfur (S) and potential acidity (H+Al) did the inoculated area provide higher means.

In

Table 1. Mean values of soil chemical attributes in the 0.00–0.10 m layer after maize harvest intercropped with Paiaguás grass, in an area under a no-till system, influenced by the residual effect of phosphate fertilization and inoculation of preceding grasses with A. brasilense. Selvíria/MS, 2021/2022.

Treatment	P	S	K	Ca	Mg	H+Al	BS	CEC	V	m	OM	pH
	mg dm−3		mmolc dm−3						%		g dm−3	(CaCl2)
P2O5 doses (kg ha−1) (P)
0	5	6	3.8	28	20	36	51.3	86.9	59	2	26	4.7
30	6	6	3.9	24	19	37	43.4	80.4	54	3	25	4.5
60	7	6	4.7	28	17	33	50.0	83.5	60	2	27	4.9
120	7	6	4.1	24	16	38	43.4	81.1	54	3	25	4.7
240	7	6	3.7	25	17	37	46.1	83.7	55	3	25	4.5
Inoculation (I)
Without	7 a	5 b	3.9	28 a	18 a	34 b	49.9	83.7	59 a	2	25	4.7
With	6 b	7 a	4.1	24 b	16 b	39 a	43.8	82.5	53 b	3	26	4.6
Test—P	0.002 *	0.89 ns	0.10 ns	0.15 ns	0.11 ns	0.26 ns	0.13 ns	0.15 ns	0.08 ns	0.88 ns	0.17 ns	0.60 ns
Test—I	0.001 **	0.001 **	0.53 ns	0.001 **	0.001 **	0.001 **	0.83 ns	0.43 ns	0.001 **	0.09 ns	0.86 ns	0.35 ns
Test—P × I	0.08 ns	0.24 ns	0.24 ns	0.36 ns	0.41 ns	0.29 ns	0.34 ns	0.88 ns	0.30 ns	0.89 ns	0.65 ns	0.96 ns
CV%	13.6	17.9	18.7	16.7	14.2	11.6	12.8	5.4	9.6	43.4	8.1	10.5

Means followed by different letters in the column, for distinct treatments, differed from each other according to the LSD post-hoc test at the 5% probability level. ** and *: significant at 1% and 5% probability by the LSD post-hoc test, respectively. ns: not significant. CV: coefficient of variation; P: phosphorus; S: sulfur; K: potassium; Ca: calcium; Mg: magnesium; H+Al: potential acidity (hydrogen plus aluminum); BS: base saturation; CEC: cation exchange capacity; V: base saturation index; m: aluminum saturation; OM: organic matter; pH: hydrogen potential.

, the most evident changes related to inoculation can be observed for phosphorus (P) and calcium (Ca) contents, which were 14.3% lower in inoculated plots. Additionally, significant reductions were found for magnesium (Mg) and base saturation (V%), with 11.1% and 10.2% decreases, respectively. In contrast, sulfur (S) and potential acidity (H+Al) showed increases of 28.6% and 12.8% in inoculated treatments. These findings highlight that microbial inoculation influenced soil acidification and nutrient dynamics, especially for elements involved in phosphorus cycling.

In the 0.10 to 0.20 m layer (

Table 2. Mean values of soil chemical attributes in the 0.10 to 0.20 m layer, after maize harvest intercropped with Paiaguás grass, in a no-till system affected by the residual effect of phosphate fertilization and inoculation of preceding grasses with A. brasilense. Selvíria/MS, 2021/2022.

Treatment	P	S	K	Ca	Mg	H+Al	BS	CEC	V	m	OM	pH
	mg dm−3		mmolc dm−3						%		g dm−3	(CaCl2)
P2O5 doses (kg ha−1) (P)
0	4	6	2.7	16	14	36	33.1	69.2	48	5	19	5
30	4	6	2.3	16	15	37	32.9	69.9	47	3	19	5
60	5	7	2.2	17	14	36	32.4	68.0	48	4	19	5
120	5	7	2.3	18	14	40	34.1	74.2	46	6	19	5
240	5	7	2.3	17	13	37	32.2	68.9	47	7	19	5
Inoculation (I)
Without	6 a	5 b	2.3	18 a	15 a	35 b	35.3 a	70.9	50 a	5	19	5
With	3 b	8 a	2.4	15 b	13 b	39 a	30.6 b	68.2	44 b	5	19	5
Test—P	0.19 ns	0.77 ns	0.58 ns	0.79 ns	0.48 ns	0.36 ns	0.93 ns	0.13 ns	0.95 ns	0.10 ns	0.72 ns	0.42 ns
Test—I	0.001 **	0.001 **	0.86 ns	0.001 **	0.001 **	0.040 *	0.001 **	0.28 ns	0.001 **	0.06 ns	0.23 ns	0.14 ns
Test—P × I	0.17 ns	0.26 ns	0.39 ns	0.29 ns	0.34 ns	0.84 ns	0.38 ns	0.36 ns	0.69 ns	0.12 ns	0.35 ns	0.42 ns
CV%	25.1	20.1	27	17.8	13.3	12.5	13	6.9	11.9	23.8	7.6	3.2

Means followed by different letters in the column, for distinct treatments, differed from each other according to the LSD post-hoc test at the 5% probability level. ** and *: significant at 1% and 5% probability by the LSD post-hoc test, respectively. ns: not significant. CV: coefficient of variation; P: phosphorus; S: sulfur; K: potassium; Ca: calcium; Mg: magnesium; H+Al: potential acidity (hydrogen plus aluminum); BS: base saturation; CEC: cation exchange capacity; V: base saturation index; m: aluminum saturation; OM: organic matter; pH: hydrogen potential.

), there was also no interaction between the sources of variation for the soil chemical attributes. For the residual effect of phosphorus doses, no significant regression adjustment was observed. Regarding inoculation, P, Ca, Mg, BS, and V% had higher levels in the non-inoculated areas, while only S and H+Al showed higher values in the inoculated areas. The other attributes were not affected by the inoculation of grain-producing grasses in rotation.

As shown in Table 2, the inoculation of preceding grasses significantly influenced soil chemical attributes in the 0.10–0.20 m layer. Phosphorus (P) content was reduced by 50% in inoculated plots compared to non-inoculated ones, while magnesium (Mg), base sum (BS), calcium (Ca), and base saturation (V%) also decreased by 13.3%, 13.3%, 16.7%, and 12%, respectively. In contrast, sulfur (S) and potential acidity (H+Al) showed marked increases of 37.5% and 10.2% in the inoculated treatments. These variations highlight the effect of inoculation on nutrient mobility and soil acidification processes. Furthermore, in this soil layer, phosphorus content remained low, cation exchange capacity (CEC) was high, organic matter (OM) showed medium levels, and pH was considered low, following the classification in Bulletin 200.

Comparing the results of the two analyzed layers, it can be observed that the levels of P, K, Ca, Mg, BS, CEC, V%, and OM decreased in deeper layers, whereas m% and pH increased. S levels increased with higher P₂O₅ doses (starting from 60 kg ha⁻¹) and also in the inoculated area. Only H+Al remained constant at both depths.

The dry matter yield of Paiaguás grass did not show any interaction or significant difference for the residual effect of phosphorus doses or the inoculation of preceding grain-producing grasses in rotation (

Table 3. Mean values for fresh matter (FM) and dry matter (DM) of Paiaguás grass in a no-till system affected by the residual effect of phosphate fertilization and inoculation of preceding grasses with A. brasilense. Selvíria/MS, 2021/2022.

Treatment	FM	DM
	kg ha−1
P2O5 doses (kg ha−1) (P)
0	11,781	4097
30	13,849	4938
60	12,960	4663
120	13,732	4444
240	12,979	3814
Inoculation (I)
Without	11,944	4065
With	14,176	4717
Test—P	0.83 ns	0.57 ns
Test—I	0.07 ns	0.17 ns
Test—P × I	0.86 ns	0.55 ns
CV%	19.69	13.52

ns: not significant. CV: coefficient of variation; FM: fresh matter; DM: dry matter.

).

For soybean yield components and productivity (

Table 4. Mean values for plant height (PlH), first pod insertion height (HFPI), number of grains per plant (NGP), number of pods per plant (NPP), plant population per hectare (PP), grain yield (SGY), and 100-grain weight (M100) of soybean grown under a no-till system affected by the residual effect of phosphate fertilization and inoculation of preceding grasses with A. brasilense. Selvíria/MS, 2022/2023.

Treatment	PlH	HFPI	NGP	NPP	PP	SGY	M100
	cm		n°			kg	g
P2O5 doses (kg ha−1) (P)
0	112	25	198	108	117,284	3343	16
30	105	23	256	127	114,815	3328	16
60	107	26	279	139	118,518	3714	16
120	107	25	252	124	101,543	3637	16
240	109	25	230	114	117,901	3614	17
Inoculation (I)
Without	112 a	25	233	116 b	122,222	3529	16
With	105 b	25	253	129 a	105,802	3526	17
Test—P	0.654 ns	0.535 ns	0.0001 **	0.001 **	0.728 ns	0.251 ns	0.567 ns
Test—I	0.012 **	0.571 ns	0.066 ns	0.006 **	0.075 ns	0.984 ns	0.139 ns
Test—P × I	0.963 ns	0.516 ns	0.011 *	0.026 *	0.9857 ns	0.053 ns	0.712 ns
CV%	8.00	11.11	13.69	11.98	24.70	16.80	9.95

Means followed by different letters in the column, for distinct treatments, differed from each other according to the LSD post-hoc test at the 5% probability level. ** and *: significant at 1% and 5% probability by the LSD post-hoc test, respectively. ns: not significant. CV: coefficient of variation; PlH: plant height; HFPI: height of first pod insertion; NGP: number of grains per plant; NPP: number of pods per plant; PP: plant population per hectare; SGY: grain yield; M100: 100-grain weight.

), there was a significant interaction between the residual effect of P₂O₅ doses and the inoculation of preceding grasses (Table 4). The breakdown for the number of grains per plant (NGP) (

Table 5. Breakdown of interactions—P₂O₅ doses × inoculation for the number of grains per plant in soybean grown under a no-till system affected by the residual effect of phosphate fertilization and inoculation of preceding grasses with A. brasilense. Selvíria/MS, 2022/2023.

Number of Grains per Plant
Treatment	Inoculation
P2O5 Doses (kg ha−1)	With	Without
n°
0	214	183
30	230 B	282 A
60	251 B	309 A
120	227 B	277 A
240	246	215

Means followed by different uppercase letters in the row differed from each other according to the LSD post-hoc test at the 5% probability level.

) showed higher means in non-inoculated plots with P₂O₅ doses of 30, 60, and 120 kg ha⁻¹. However, no significant regression model was observed in the breakdown of the interaction for the number of pods per plant (NPP) and NGP due to the residual phosphorus effect (Table 5 and

Table 6. Breakdown of interactions—P₂O₅ doses × inoculation, for the number of pods per soybean plant in a no-till system, considering the residual effect of phosphorus fertilization and the inoculation of preceding grasses with A. brasilense. Selvíria/MS, 2022/2023.

Number of Pods per Plant
Treatment	Inoculation
P2O5 Doses (kg ha−1)	With	Without
n°
0	113	102
30	113 B	142 A
60	123 B	155 A
120	116	131
240	113	114

Means followed by distinct uppercase letters in the row differed from each other according to the LSD post-hoc test at the 5% probability level.

).

Significant differences were found for the inoculation factor in plant height (PlH) and the number of pods per plant (NPP), where the inoculated area had the highest PlH values, while the non-inoculated area showed the highest NVP values. No differences were observed for the inoculation factor in the other yield components.

It can also be observed in Table 4 that in the plots where grasses were inoculated in the crop rotation, plant height (PlH) was 6.3% higher than in the non-inoculated plots. Conversely, the number of pods per plant (NPP) decreased by 10% in the inoculated areas.

There was an interaction between the residual effect of P₂O₅ doses and the inoculation of preceding grasses for NPP (Table 6). The breakdown of this interaction showed higher means in non-inoculated plots with P₂O₅ doses of 30 and 60 kg ha⁻¹.

For sorghum yield components and productivity (

Table 7. Mean values for plant height (PlH), panicle length (PL), stem diameter (BSD), plant population (PP), grain yield (SGY), and thousand-grain weight (M1000) of grain sorghum intercropped with Piatã grass in a no-till system affected by the residual effect of phosphate fertilization and inoculation of preceding grasses with A. brasilense. Selvíria/MS, 2023.

Treatment	PlH	PL	BSD	PP	SGY	M1000
	cm		mm	n°	kg ha−1	g
P2O5 doses (kg ha−1) (P)
0	89	32	15.7	63,425	824	186
30	84	29	15.8	73,148	960	195
60	86	31	14.7	66,203	791	187
120	88	31	16.4	66,666	869	190
240	86	32	15.6	70,833	855	189
Inoculation (I)
Without	85	31	15.7	68,703	952 a	185 b
With	88	31	15.6	67,407	767 b	194 a
Test—P	0.156 ns	0.072 ns	0.244 ns	0.261 ns	0.470 ns	0.443 ns
Test—I	0.051 ns	0.291 ns	0.841 ns	0.663 ns	0.004 **	0.008 **
Test—P × I	0.524 ns	0.983 ns	0.765 ns	0.480 ns	0.005 **	0.686 ns
CV%	4.50	6.15	9.02	13.70	21.97	5.37

Means followed by different letters in the column, for distinct treatments, differed from each other according to the LSD post-hoc test at the 5% probability level. **: significant at 1% probability by the LSD post-hoc test. ns: not significant. CV: coefficient of variation; PlH: plant height; PL: panicle length; BSD: stem diameter; PP: plant population; SGY: grain yield; M1000: thousand-grain weight.

), a significant interaction was observed between the residual effect of P₂O₅ doses and the inoculation of preceding grasses for grain yield (SGY). The breakdown (

Table 8. Breakdown of interactions—P₂O₅ doses × inoculation for grain yield (SGY) of sorghum grown under a no-till system affected by the residual effect of phosphate fertilization and inoculation of preceding grasses with A. brasilense. Selvíria/MS, 2023.

Sorghum Grain Productivity
Treatment	Inoculation
P2O5 Doses (kg ha−1)	With	Without
kg ha−1
0	1022 A	626 B
30	927	994
60	724	857
120	974	764
240	1115 A	594 B

Means followed by distinct uppercase letters in the row differed from each other according to the LSD post-hoc test at the 5% probability level.

) showed higher means in inoculated plots with P₂O₅ doses of 0 and 240 kg ha⁻¹. However, no regression adjustment was observed for residual phosphorus doses, regardless of whether the areas were inoculated or not.

A significant difference was found for the inoculation factor in grain yield (SGY) and thousand-grain weight (M1000). PG was higher in the inoculated area, whereas M1000 showed higher means in the non-inoculated area. No significant differences were observed among treatments for the other yield components.

Additionally, in Table 7, it is noted that in plots where grasses were inoculated in the crop rotation, PG was 19.4% higher than in non-inoculated plots. Conversely, M1000 decreased by 4.6% in inoculated areas.

For the fresh and dry matter yield of sorghum stem, leaves, and panicle (

Table 9. Mean values for fresh and dry matter of the stem (FM S and DM S), fresh and dry matter of leaves (FM L and DM L), and fresh and dry matter of the panicle (FM P and DM P), respectively, of grain sorghum intercropped with Piatã grass in a no-till system affected by the residual effect of phosphate fertilization and inoculation of preceding grasses with A. brasilense. Selvíria/MS, 2023.

Treatment	FM S	DM S	FM L	FM L	FM P	DM P
	kg ha−1
P2O5 doses (kg ha−1) (P)
0	4138	991	1743	526	2095	1334
30	4608	1094	1996	629	2198	1396
60	3976	840	1474	474	1893	1161
120	4553	1085	1692	477	2100	1380
240	4743	1069	1747	534	2115	1367
Inoculation (I)
Without	4384	1014	1650	500	2257 a	1451 a
With	4423	1017	1811	556	1904 b	1204 b
Test—P	0.544 ns	0.228 ns	0.060 ns	0.165 ns	0.788 ns	0.400 ns
Test—I	0.905 ns	0.964 ns	0.132 ns	0.189 ns	0.029 *	0.006 **
Test—P × I	0.226 ns	0.355 ns	0.564 ns	0.683 ns	0.114 ns	0.313 ns
CV%	23.84	24.17	19.11	25.50	23.48	20.03

Means followed by different letters in the column, for distinct treatments, differed from each other according to the LSD post-hoc test at the 5% probability level. ** and *: significant at 1% and 5% probability by the LSD post-hoc test, respectively. ns: not significant. CV: coefficient of variation; FM S and DM S: fresh and dry matter of the stem; FM L and DM L: fresh and dry matter of leaves; FM P and DM P: fresh and dry matter of the panicle.

), no interaction was observed between the analyzed factors or for the residual effect of phosphorus doses. However, for the inoculation factor, a significant difference was found in the fresh and dry matter yield of the panicle, with both showing higher production in the inoculated area, while no significant differences were observed for the other components.

Additionally, in Table 9, it is noted that in plots where grasses were inoculated in the crop rotation, FM P and DM P were 15.6% and 17% higher, respectively, than in non-inoculated plots.

For the fresh and dry matter yield of Piatã grass (

Table 10. Mean values for fresh matter (FM) and dry matter (DM) of Piatã grass intercropped with grain sorghum in a no-till system affected by the residual effect of phosphate fertilization and inoculation of preceding grasses with A. brasilense. Selvíria/MS, 2023.

Treatment	FM	DM
	kg ha−1
P2O5 doses (kg ha−1) (P)
0	13,305	3831
30	13,366	3806
60	14,741	4092
120	15,407	4225
240	18,590	5298
Inoculation (I)
Without	13,938 b	3960 b
With	16,225 a	4541 a
Test—P	0.009 **	0.012 **
Test—I	0.023 *	0.046 *
Test—P × I	0.045 *	0.012 **
CV%	20.13	20.88

Means followed by different letters in the column, for distinct treatments, differed from each other according to the LSD post-hoc test at the 5% probability level. ** and *: significant at 1% and 5% probability by the LSD post-hoc test, respectively. CV: coefficient of variation; FM: fresh matter; DM: dry matter.

), significant differences were observed for the inoculation factor, with both variables showing higher production in the non-inoculated area.

As shown in Table 10, an interaction was observed between the residual effect of P₂O₅ doses and the inoculation of preceding grasses for Piatã grass fresh and dry matter yield (

Table 11. Breakdown of interactions—P₂O₅ rates × inoculation for fresh matter (FM) of Piatã grass in a no-till system, considering the residual effect of phosphate fertilization and the inoculation of preceding grasses with A. brasilense. Selvíria/MS, 2023.

Fresh Matter of Piatã Grass
Treatment	Inoculation
P2O5 Doses (kg ha−1)	With	Without
kg ha−1
0	9825 B	16,785 A
30	14,472	12,260
60	12,470 B	17,012 A
120	15,190	15,625
240	17,735	19,445

Means followed by different uppercase letters in the row differed from each other according to the LSD post-hoc test at the 5% probability level.

and

Table 12. Breakdown of interactions—doses of P₂O₅ × inoculation for dry matter (DM) of Piatã grass in an area under no-tillage system due to the effect of residual phosphate fertilization and inoculation of predecessor grasses by A. brasilense. Selvíria/MS, 2023.

Dry Matter of Piatã Grass
Treatment	Inoculation
P2O5 Doses (kg ha−1)	With	Without
kg ha−1
0	2798 B	4864 A
30	4248	3363
60	3280 B	4903 A
120	4279	4172
240	5193	5403

Means followed by different uppercase letters in the row differed from each other according to the LSD post-hoc test at the 5% probability level.

). The breakdown for fresh matter (Table 11) revealed higher means in non-inoculated plots with P₂O₅ doses of 0 and 60 kg ha⁻¹. However, no significant regression adjustments were found for areas with or without inoculation. The breakdown for dry matter (Table 12) showed higher means in non-inoculated plots with P₂O₅ doses of 0 and 60 kg ha⁻¹. Similarly to fresh matter, no significant regression adjustments were observed for the residual effect of phosphate fertilization.

Regarding inoculation, plots where grasses were inoculated in the crop rotation showed 14.1% and 12.8% lower MV and MS of Piatã grass, respectively, compared to non-inoculated plots, as indicated in Table 10.

Table 13. Mean values of soil chemical attributes in the 0.00 to 0.10 m layer after the desiccation of Piatã grass in a no-till system, considering the residual effect of phosphate fertilization and the inoculation of preceding grasses with A. brasilense. Selvíria/MS, 2023/2024.

Treatment	P	S	K	Ca	Mg	H+Al	BS	CEC	V	m	OM	pH
	mg dm−3		mmolc dm−3						%		g dm−3	(CaCl2)
P2O5 doses (P)
0	10	5	4.1	25	17	33	46.4	79.3	59	4	28	5.1
30	13	5	4.1	28	19	33	51.2	83.7	60	4	28	5.1
60	11	5	4.2	23	16	33	44.2	77.8	57	4	28	5.1
120	12	5	4.1	23	15	33	42.3	75.2	56	6	27	5.1
240	20	5	4.1	28	17	34	47.1	81.0	58	4	29	5.1
Inoculation (I)
Without	16 a	5	4.6 a	26	18	33	48.9	82.2 a	59	4.1 b	29	5.1
With	11 b	5	3.8 b	24	16	33	43.6	76.8 b	57	4.8 a	28	5.1
Test—P	0.0001 **	0.986 ns	0.962 ns	0.102 ns	0.483 ns	0.976 ns	0.545 ns	0.259 ns	0.886 ns	0.007 **	0.311 ns	0.948 ns
Test—I	0.0001 **	0.283 ns	0.000 **	0.072 ns	0.225 ns	0.811 ns	0.127 ns	0.028 *	0.474 ns	0.038 *	0.352 ns	0.510 ns
Test—P × I	0.0001 **	0.359 ns	0.915 ns	0.018 **	0.279 ns	0.590 ns	0.339 ns	0.357 ns	0.453 ns	0.0001 **	0.651 ns	0.760 ns
CV%	20.44	20.15	14.09	16.02	23.24	13.9	23.33	9.79	13.21	22.77	8.27	4.61

Means followed by different letters in the column, for distinct treatments, differed from each other according to the LSD post-hoc test at the 5% probability level. ** and *: significant at 1% and 5% probability by the LSD post-hoc test, respectively. ns: not significant. CV: coefficient of variation; P: phosphorus; S: sulfur; K: potassium; Ca: calcium; Mg: magnesium; H+Al: potential acidity (hydrogen plus aluminum); BS: base saturation; CEC: cation exchange capacity; V: base saturation index; m: aluminum saturation; OM: organic matter; pH: hydrogen potential.

presents the chemical soil attributes for the 0.00–0.10 m layer, where significant interactions were identified between phosphorus rates and microbial inoculation for phosphorus (P), calcium (Ca), and aluminum saturation (m%). Notably, P and Ca levels were higher in non-inoculated plots, particularly at 120 and 240 kg ha⁻¹ of P₂O₅. In contrast, m% was higher in inoculated plots at these same doses. These results reflect the contrasting roles of inoculation and P fertilization in nutrient dynamics and acid–base balance. Additionally, phosphorus and CEC values were classified as low and high, respectively, while OM content was at medium levels and pH remained low, as per the reference standards of Bulletin 200.

Phosphorus content increased with higher P₂O₅ rates. Inoculation with A. brasilense resulted in slightly higher phosphorus levels compared to the non-inoculated treatment, especially at the lower fertilizer rates. The greatest difference was observed at the 240 kg ha⁻¹ rate, with 29.5 mg dm⁻³ in the non-inoculated treatment and 11.2 mg dm⁻³ with inoculation (

Table 14. Breakdown of interactions—P₂O₅ rates × inoculation for soil phosphorus content in the 0.00 to 0.10 m layer in a no-till system, considering the residual effect of phosphate fertilization and the inoculation of preceding grasses with A. brasilense. Selvíria/MS, 2023/2024.

Phosphorus Content in the Soil in the 0 to 0.10 m Layer
Treatment	Inoculation
P2O5 Doses (kg ha−1)	With	Without
mg dm−3
0	11.0	9.5
30	12.7	13.7
60	9.2	11.7
120	9.7 B	13.7 A
240	11.2 B	29.5 A

Means followed by different uppercase letters in the row differed from each other according to the LSD post-hoc test at the 5% probability level.

). Calcium content was higher in the non-inoculated treatments, particularly at the intermediate P₂O₅ rates (30 and 60 kg ha⁻¹). Inoculation with A. brasilense led to lower calcium levels across all fertilizer doses (

Table 15. Breakdown of interactions—P₂O₅ rates × inoculation for soil calcium content in the 0.00 to 0.10 m layer in a no-till system, considering the residual effect of phosphate fertilization and the inoculation of preceding grasses with A. brasilense. Selvíria/MS, 2023/2024.

Calcium Content in the Soil in the Layer from 0 to 0.10 m
Treatment	Inoculation
P2O5 Doses (kg ha−1)	With	Without
mmolc dm−3
0	26.5	23.5
30	22.5 B	33.2 A
60	23.7	22.7
120	21.7	23.7
240	24.2	27.2

Means followed by different uppercase letters in the row differed from each other according to the LSD post-hoc test at the 5% probability level.

).

It is noteworthy that for both P and Ca contents, no regression adjustments were found for phosphorus rates in areas with or without A. brasilense inoculation. Regarding aluminum saturation (

Table 16. Breakdown of interactions—P₂O₅ rates × inoculation for aluminum saturation (m%) in the soil in the 0.00 to 0.10 m layer in a no-till system, considering the residual effect of phosphate fertilization and the inoculation of preceding grasses with A. brasilense. Selvíria/MS, 2023/2024.

Aluminum Saturation in the Soil in the 0 to 0.10 m Layer
Treatment	Inoculation
P2O5 Doses (kg ha−1)	With	Without
%
0	3.5 B	5.0 A
30	5.7 A	3.0 B
60	3.5 B	5.2 A
120	6.5 A	5.0 B
240	5.0 A	2.5 B

Means followed by different uppercase letters in the row differed from each other according to the LSD post-hoc test at the 5% probability level.

), higher means were observed in non-inoculated plots at P₂O₅ rates of 0 and 60 kg ha⁻¹. In inoculated areas, higher means were recorded at P₂O₅ rates of 30, 120, and 240 kg ha⁻¹, again without a significant regression model within the breakdown.

For the inoculation factor, sulfur (S), calcium (Ca), magnesium (Mg), potential acidity (H+Al), sum of bases (SB), base saturation (V%), organic matter (OM), and hydrogen potential (pH) did not show significant differences. In contrast, phosphorus (P), potassium (K), and cation exchange capacity (CEC) increased in non-inoculated areas, whereas only aluminum saturation (m%) showed higher means in the inoculated area.

Thus, plots where grasses were inoculated in the crop rotation had 31.2% lower P content compared to non-inoculated plots, as well as decreases of 17.4% and 6.6% for K and CEC, respectively. Meanwhile, for m%, inoculated plots had higher values, with an increase of 14.6%. Still within this layer, P content was considered low, except at P₂O₅ rates of 30 and 240 kg ha⁻¹, where it reached medium levels. CEC was high, OM was at medium levels, and pH was low.

As shown in

Table 17. Mean values of soil chemical attributes in the 0.10 to 0.20 m layer after the desiccation of Piatã grass in a no-till system, considering the residual effect of phosphate fertilization and the inoculation of preceding grasses with A. brasilense. Selvíria/MS, 2023/2024.

Treatment	P	S	K	Ca	Mg	H+Al	BS	CEC	V	m	OM	pH
	mg dm−3		mmolc dm−3						%		g dm−3	(CaCl2)
P2O5 doses (P)
0	8	5	3.6	19	12	33	35.0	68.0	51	6	20	5.0
30	8	5	3.8	20	14	32	37.0	69.8	53	7	22	5.0
60	8	5	3.9	20	13	33	37.3	70.6	52	7	22	5.0
120	8	5	3.5	18	12	33	34.0	67.7	50	8	21	5.0
240	11	6	3.6	19	13	33	35.4	68.6	51	6	21	5.0
Inoculation (I)
Without	9	5	3.9 a	19	12	33	35.2	68.9	50	7.4	20 b	5.0
With	8	5	3.5 b	20	13	32	36.5	68.9	53	6.4	22 a	5.0
Test—P	0.025 *	0.449 ns	0.529 ns	0.917 ns	0.475 ns	0.948 ns	0.884 ns	0.825 ns	0.936 ns	0.004 **	0.628 ns	0.995 ns
Test—I	0.208 ns	0.208 ns	0.012 **	0.623 ns	0.240 ns	0.298 ns	0.647 ns	0.984 ns	0.371 ns	0.050 *	0.006 **	0.451 ns
Test—P × I	0.003 **	0.891 ns	0.253 ns	0.796 ns	0.265 ns	0.533 ns	0.742 ns	0.836 ns	0.728 ns	0.009 **	0.556 ns	0.869 ns
CV%	23.84	20.78	15.64	24.74	16.45	11.31	22.99	8.33	15.91	22.29	8.6	4.93

Means followed by different letters in the column, for distinct treatments, differed from each other according to the LSD post-hoc test at the 5% probability level. ** and *: significant at 1% and 5% probability by the LSD post-hoc test, respectively. ns: not significant. CV: coefficient of variation; P: phosphorus; S: sulfur; K: potassium; Ca: calcium; Mg: magnesium; H+Al: potential acidity (hydrogen plus aluminum); BS: base saturation; CEC: cation exchange capacity; V: base saturation index; m: aluminum saturation; OM: organic matter; pH: hydrogen potential.

, referring to the 0.10–0.20 m soil layer, aluminum saturation (m%) was significantly influenced by the interaction between phosphorus rates and inoculation. Higher m% values were observed in non-inoculated plots at 0 and 240 kg ha⁻¹ of P₂O₅, while inoculated plots showed increased m% at intermediate rates (particularly 120 kg ha⁻¹). These results underscore the nuanced effects of inoculation and fertilization depth-wise. Across this layer, potassium (K) was the only attribute significantly reduced by inoculation (by 10%), while organic matter (OM) content increased by 9.1%. Other attributes were not significantly affected. The P content was still considered low, CEC remained high, OM showed medium levels, and pH was classified as low.

The interaction between P₂O₅ doses and inoculation with A. brasilense significantly affected soil aluminum saturation in the 0.10 to 0.20 m layer. In the absence of inoculation, aluminum saturation remained constant across all P₂O₅ doses, averaging approximately 7.5%. However, in inoculated areas, there was a quadratic response, with the highest aluminum saturation observed at the 120 kg ha⁻¹ P₂O₅ dose (9.7%), and the lowest at the 240 kg ha⁻¹ dose (3.7%). These results suggest that inoculation with A. brasilense alters the aluminum dynamics in the soil profile under different phosphorus fertilization levels (

Table 18. Breakdown of interactions—P₂O₅ rates × inoculation for the soil chemical attribute aluminum saturation in the 0.10 to 0.20 m layer in a no-till system, considering the residual effect of phosphate fertilization and the inoculation of preceding grasses with A. brasilense. Selvíria/MS, 2023/2024.

Soil Aluminum Saturation in the 0.10 to 0.20 m Layer
Treatment	Inoculation
P2O5 Doses (kg ha−1)	With	Without
%
0	5.0 B	7.5 A
30	7.2	7.5
60	6.5	7.5
120	9.7 A	7.5 B
240	3.7 B	7.2 A

Means followed by different uppercase letters in the row differed from each other according to the LSD post-hoc test at the 5% probability level.

).

Additionally, as shown in Table 17, for phosphorus rates in non-inoculated areas (breakdown), a positive exponential adjustment was found for P content (Figure 7), whereas no adjustment was observed in inoculated areas.

Regarding the inoculation factor, only K showed higher values in non-inoculated areas, while, in inoculated areas, only OM presented higher values. The remaining attributes were not significantly affected by the inoculation of grasses in rotation.

Thus, as shown in Table 17, plots where grasses were inoculated in the crop rotation had 10% lower K content compared to non-inoculated plots. On the other hand, for OM content, inoculated plots showed higher values, with an increase of 9.1%. Still within this layer, P content was considered low, CEC was high, OM was at a medium level, and pH was low.

Contrasting the results of the two analyzed layers, it was observed that P, K, Ca, Mg, BS, CEC, V%, and OM contents decreased in deeper layers, whereas m% and pH showed increased values. S content increased with P₂O₅ rates. Only S (with and without inoculation) and H+Al remained constant at both soil depths.

When comparing the analysis of soil chemical attributes from the 2021/2022 and 2023/2024 growing seasons in the 0.00 to 0.10 m and 0.10 to 0.20 m layers, a similar behavior of soil elements was observed (Figure 8).

There was an increase in P, K, V%, m%, and OM for both sources of variation and soil layers from the 2021/2022 season to the 2023/2024 season. For Ca levels and BS values, an increase was observed only in the 0.10 to 0.20 m layer, while, in the more superficial layer, a decrease occurred due to the treatments. It was observed that S, Mg, H+Al, and CEC decreased between seasons.

The different P fractions were classified according to their lability in each extractor. Labile P included the Pi fractions extracted by RTA and the Pi and Po fractions extracted by sodium bicarbonate. Moderately labile P comprised the Pi and Po fractions extracted by 0.1 mol L⁻¹ NaOH, along with the inorganic P extracted with HCl. Meanwhile, non-labile phosphorus consisted of the sum of Pi and Po extracted by 0.5 mol L⁻¹ NaOH and residual Pi and Po.

Table 19. Mean values for soil phosphorus fractionation in the 0.00–0.10 m layer, labile phosphorus (Lab), moderately labile phosphorus (Mod Lab), non-labile phosphorus (N Lab), inorganic phosphorus (Pi), organic phosphorus (Po), and total phosphorus (P Tot) after Brachiaria brizantha cv. Piatã fallow in a no-tillage system, considering the residual effect of phosphate fertilization and the inoculation of preceding grasses with A. brasilense. Selvíria/MS, 2023/2024.

Treatment	Lab	Mod Lab	N Lab	Pi	Po	P Tot
	mg kg−1
P2O5 doses (kg ha−1) (P)
0	13.09	149.68	224.95	42.88	96.99	387.72
30	14.33	149.31	228.21	46.53	92.14	393.77
60	13.78	148.32	220.37	43.72	94.80	382.45
120	15.12	150.92	224.06	48.14	91.74	390.41
240	26.01	167.01	223.43	57.29	98.30	415.40
Inoculation (I)
Without	13.26 b	150.48	217.34	45.54	94.10	381.17
With	19.66 a	155.61	231.06	49.88	95.49	406.33
Test—P	0.0001 **	0.014 **	0.994 ns	0.003 **	0.394 ns	0.521 ns
Test—I	0.0001 **	0.170 ns	0.227 ns	0.069 ns	0.583 ns	0.053 ns
Test—P × I	0.0001 **	0.017 **	0.785 ns	0.002 **	0.850 ns	0.774 ns
CV%	19.49	07.55	15.73	15.26	8.39	10.06

Means followed by different letters in the column for different treatments differed according to the LSD post-hoc test at the 5% probability level. **: significant at 1% probability by the LSD post-hoc test. ns: not significant. CV: coefficient of variation; Lab: labile phosphorus; Mod Lab: moderately labile phosphorus; N Lab: non-labile phosphorus; Pi: inorganic phosphorus; Po: organic phosphorus; P Tot: total phosphorus.

presents the results of soil phosphorus fractionation in the 0.00–0.10 m layer. A significant interaction between phosphorus doses and inoculation was observed, particularly for labile and moderately labile phosphorus forms. In non-inoculated plots, an exponential increase was observed for both fractions with increasing P₂O₅ doses, indicating a strong residual effect of fertilization. In contrast, inoculated plots showed no significant variation, suggesting that inoculation may reduce P availability or alter its stabilization in the soil. Furthermore, inorganic phosphorus (Pi) also increased exponentially in non-inoculated treatments, reinforcing the contrasting dynamics between treatments. Notably, inoculated plots had 32.5% less labile P compared to non-inoculated ones, which may impact nutrient accessibility for crops. These results highlight the complex interaction between microbial activity and phosphorus forms in conservationist systems.

The regression analysis indicated that increasing P₂O₅ doses resulted in a positive exponential increase in both labile and moderately labile phosphorus fractions in the 0.00–0.10 m soil layer under non-inoculated conditions. The moderately labile fraction showed higher phosphorus concentrations compared to the labile fraction, with R² values of 0.9472 and 0.9109, respectively (Figure 9). Additionally, the inorganic phosphorus fraction also increased exponentially with higher P₂O₅ doses, presenting an R² of 0.943 (Figure 10). These results highlight the direct influence of phosphate fertilization on the accumulation of different phosphorus fractions in the soil.

Regarding the inoculation factor, a statistical distinction was found for the labile fraction, with higher labile P levels in non-inoculated areas. However, for the other forms of P in the soil, no effect of this factor was observed. Table 19 also shows that plots with inoculated grasses in crop rotation had 32.5% lower labile P levels than non-inoculated plots.

In the analysis of soil enzymatic activity (

Table 20. Mean values for soil enzymatic activity in the 0.00–0.10 m layer, arylsulfatase (AS), β-glucosidase (βG), and acid phosphatase (AF) after Brachiaria brizantha cv. Piatã fallow in a no-tillage system, considering the residual effect of phosphate fertilization and the inoculation of preceding grasses with A. brasilense. Selvíria/MS, 2023/2024.

Treatment	AS	ßG	AF
	µg PNP g−1 Solo h−1
P2O5 doses (kg ha−1) (P)
0	141	145	313
30	136	137	311
60	128	143	316
120	136	138	289
240	148	155	339
Inoculation (I)
Without	127 b	127 b	319
With	148 a	159 a	308
Test—P	0.558 ns	0.675 ns	0.568 ns
Test—I	0.008 **	0.0001 **	0.587 ns
Test—P × I	0.103 ns	0.315 ns	0.504 ns
CV%	17.41	18.90	18.34

Means followed by different letters in the column for different treatments differed according to the LSD post-hoc test at the 5% probability level. **: significant at 1% probability by the LSD post-hoc test. ns: not significant. CV: coefficient of variation; AS: arylsulfatase; βG: β-glucosidase; AF: acid phosphatase.

), no interaction was found between the analyzed factors, nor for the isolated effect of residual phosphate fertilization. However, A. brasilense inoculation affected arylsulfatase (AS) and β-glucosidase (βG) enzyme activities, with higher activity in non-inoculated areas. For acid phosphatase (AF), no significant differences were found for the inoculation factor.

Table 20 also indicates that in plots where grasses were inoculated in crop rotation, AS and βG activities were 14.2% and 20.1% lower than in non-inoculated plots.

In the analysis of soil biological attributes (

Table 21. Mean values for soil biological attributes in the 0.00–0.10 m layer, microbial respiration activity (MRA), microbial biomass carbon (MBC), total organic carbon (TOC), and metabolic quotient (qCO₂) after Brachiaria brizantha cv. Piatã fallow in a no-tillage system, considering the residual effect of phosphate fertilization and the inoculation of preceding grasses with A. brasilense. Selvíria/MS, 2023/2024.

Treatment	MRA	MBC	TOC	qCO2
	mg CO2 g−1 dia−1	mg C g−1	g kg−1	μg C-CO2 μg−1 CBM dia−1
P2O5 doses (kg ha−1) (P)
0	0.71	0.24	15	2.9
30	0.62	0.28	15	2.2
60	0.50	0.30	15	1.7
120	0.60	0.31	15	1.9
240	0.75	0.28	16	2.7
Inoculation (I)
Without	0.56 b	0.28	15	2.0 b
With	0.72 a	0.28	15	2.6 a
Test—P	0.538 ns	0.080 ns	0.273 ns	0.341 ns
Test—I	0.000 **	0.821 ns	0.872 ns	0.008 **
Test—P × I	0.082 ns	0.246 ns	0.468 ns	0.141 ns
CV%	11.39	10.23	7.71	16.03

Means followed by different letters in the column for different treatments differed according to the LSD post-hoc test at the 5% probability level. **: significant at 1% probability by the LSD post-hoc test. ns: not significant. CV: coefficient of variation; MRA: microbial respiration activity; MBC: microbial biomass carbon; TOC: total organic carbon; qCO₂: metabolic quotient.

), no interaction was observed between the analyzed factors or for the isolated effect of residual phosphate fertilization. For inoculation, differences were observed in the means of microbial respiration activity (MRA) and the metabolic quotient (qCO₂), with higher values in non-inoculated areas. For the other biological soil attributes, no significant differences were found for inoculation. The plots with inoculated grasses in crop rotation had MRA and qCO₂ levels 22.2% and 23.1% lower than non-inoculated plots.

Figure 11 shows soil health via Chaer’s model, indicating the trends for soil carbon within the treatments adopted in this study, where it can be observed that both the inoculation factor and the residual effect of P₂O₅ doses resulted in high and stable levels based on TOC and MRA.

Still analyzing soil carbon trends in the experimental plots, the TOC and SMEA means indicated that for both the inoculation factor and the residual effect of P₂O₅ doses, values fell within Quadrant 1 (Q1) and Quadrant 4 (Q4), which corresponded, respectively, to healthy soil and recovering soil (Figure 12).

The dry matter yield of Brachiaria brizantha cv. Piatã during fallow after harvesting the intercropped sorghum showed no interaction between the factors and no significant difference for either the residual phosphorus doses or the inoculation of preceding grain-producing grasses in rotation (

Table 22. Mean values of fresh matter (FM) and dry matter (DM) for B. brizantha cv. Piatã in fallow after grain sorghum harvest in an intercropping system under no-tillage, considering the residual effect of phosphate fertilization and the inoculation of preceding grasses with A. brasilense. Selvíria/MS, 2023.

Treatment	FM	DM
	kg ha−1
P2O5 doses (kg ha−1) (P)
0	15,962	5562
30	17,156	6150
60	15,553	5515
120	13,928	5029
240	15,020	5123
Inoculation (I)
Without	14,806	5176
With	16,242	5776
Test—P	0.441 ns	0.388 ns
Test—I	0.195 ns	0.128 ns
Test—P × I	0.791 ns	0.277 ns
CV%	22.06	22.15

ns: not significant. CV: coefficient of variation; FM: fresh matter; DM: dry matter.

).

The total dry matter yield of the system during the research period showed no interaction between the residual phosphate fertilization and inoculation, nor did it show a significant effect of phosphorus dose alone. However, for inoculation, higher productivity was observed in non-inoculated areas (

Table 23. Mean values of dry matter (DM) production of B. brizantha cv. Paiaguás, B. brizantha cv. Piatã (first and second cuts), and sorghum (stem, panicle, and leaves) in a no-tillage system, considering the residual effect of phosphate fertilization and inoculation of preceding grasses with A. brasilense. Selvíria/MS, 2021/2023.

Treatment	DM
	kg ha−1
P2O5 doses (kg ha−1) (P)
0	16,193
30	18,222
60	17,245
120	17,204
240	18,143
Inoculation (I)
Without	16,575 b
With	18,227 a
Test—P	0.183 ns
Test—I	0.007 **
Test—P × I	0.329 ns
CV%	10.43

Means followed by different letters in the column for different treatments differed by the LSD post-hoc test at the 5% probability level. **: significant at 1% probability. ns: not significant. CV: coefficient of variation; DM: dry matter.

). Plots with inoculated grasses in crop rotation had a 9.1% lower total DM than non-inoculated plots.

In

Table 24. Mean values for carbon stock in particulate (POC), mineral (MOC), and total organic carbon (TOC) fractions in the 0.00–0.10 m soil layer in a no-tillage system, considering the residual effect of phosphate fertilization and inoculation of preceding grasses with A. brasilense. Selvíria/MS, 2023/2024.

Treatment	TOC	MOC	POC
	g kg−1
P2O5 doses (kg ha−1) (P)
0	13.8	8.3	5.4
30	15.4	9.6	5.7
60	14.9	8.2	6.7
120	14.4	9.7	4.8
240	16.2	9.6	6.6
Inoculation (I)
Without	15.2 a	9.4	5.2
With	14.6 b	8.5	6.2
Test—P	0.000 **	0.676 ns	0.531 ns
Test—I	0.026 *	0.168 ns	0.397 ns
Test—P × I	0.133 ns	0.442 ns	0.652 ns
CV (%)	3.62	20.15	40.19

Means followed by different letters in the column for different treatments differed by the LSD post-hoc test at the 5% probability level. ** and *: significant at 1% and 5% probability, respectively. ns: not significant. CV: coefficient of variation; POC: carbon stock particulate; MOC: carbon stock mineral; TOC: total organic carbon. Regression equation: TOC = 0.0069x + 14.318, R² = 0.5077.

, carbon stock values for the 0.00–0.10 m layer are presented, showing that total organic carbon (TOC) increased significantly with higher residual phosphorus doses, following a quadratic trend. Among the carbon fractions, particulate organic carbon (POC) and mineral-associated organic carbon (MOC) did not vary significantly with fertilization. For the inoculation factor, TOC was significantly higher in inoculated plots (by 4.1%), suggesting that microbial activity may contribute to carbon stabilization in the soil. Although differences in POC and MOC were not statistically significant, the trend toward greater POC in inoculated areas suggests an early shift in carbon partitioning. These findings reinforce the potential of inoculation to enhance carbon sequestration, particularly when combined with adequate phosphorus management.

For soil carbon storage in the 0.10–0.20 m layer (

Table 25. Mean values for carbon stock in particulate (POC), mineral (MOC), and total organic carbon (TOC) fractions in the 0.10–0.20 m soil layer in a no-tillage system, considering the residual effect of phosphate fertilization and inoculation of preceding grasses with A. brasilense. Selvíria/MS, 2023/2024.

Treatment	TOC	MOC	POC
	g kg−1
P2O5 doses (kg ha−1) (P)
0	18.1	14.7	3.4
30	20.0	17.3	2.7
60	19.4	16.9	2.5
120	17.8	14.9	2.9
240	21.1	18.6	2.5
Inoculation (I)
Without	19.6	17.7 a	3.7 a
With	19.0	15.3 b	1.9 b
Test—P	0.003 **	0.006 **	0.683 ns
Test—I	0.174 ns	0.001 **	0.003 **
Test—P × I	0.214 ns	0.477 ns	0.765 ns
CV (%)	4.91	7.57	36.91

Means followed by different letters in the column for different treatments differed by the LSD post-hoc test at the 5% probability level. **: significant at 1% probability. ns: not significant. CV: coefficient of variation; POC: carbon stock particulate; MOC: carbon stock mineral; TOC: total organic carbon. Regression equation: TOC = 8 × 10⁻⁵x² − 0.0122x + 19.129, R² = 0.4565. Regression equation: MOC = 0.0103x + 15.558, R² = 0.3436.

), a significant adjustment was observed for TOC and MOC due to phosphorus doses, following quadratic and linear trends, respectively, with higher contents in areas with the highest phosphorus dose. Regarding inoculation, significant differences were observed for MOC and POC, with higher values for both fractions in inoculated areas.

Among the soil layers analyzed, it can be observed that for the MOC fraction levels, there was an increase with depth in the soil profile. Consequently, the TOC also showed the same behavior, while, for POC, a decrease in its levels with depth was observed.

From the soil carbon (C) values, it was possible to estimate the soil organic matter content. Based on the premise that organic matter contained 58% organic carbon, the factor 1.724 was used for this conversion. Thus, from the results obtained in Table 24 and Table 25, in addition to the carbon stock (POC, MOC, and TOC), it was also possible to estimate the quantity of particulate organic matter (POM), which corresponded to organic matter associated with the soil’s sand fraction; mineral organic matter (MOM), associated with the silt and clay fraction; and total organic matter (TOM), which corresponded to the sum of POM and MOM, by multiplying the values by the factor 1.724. Since this was a fixed multiplication factor, the statistical significance results remained unchanged.

In summary, this study demonstrated that inoculation reduced nutrient input in soil layers and did not result in differences in the dry matter production of Paiaguás grass or soybean grain yield but increased the grain and dry matter yield of the panicle in sorghum intercropped with Piatã grass, which, in turn, had its dry matter production reduced in inoculated areas, where lower levels of labile phosphorus were also observed in the soil. Similarly, for the activities of arylsulfatase and ß-glucosidase enzymes, as well as for biological attributes such as microbial respiratory activity and the metabolic quotient, inoculation resulted in lower total dry matter production during the course of this study. Finally, the inoculation of grain-producing grasses led to a greater accumulation of carbon fractions and, consequently, organic matter in the soil.

The residual effect of P₂O₅ doses resulted in an increase in the concentration of this nutrient in the soil, in accordance with the increase in the residual dose. However, it did not result in differences in the dry matter production of Paiaguás grass or in soybean or sorghum grain yield. On the other hand, for the dry matter production of Piatã grass, higher productivity was observed with higher residual P₂O₅ doses, as well as for labile, moderately labile, and inorganic phosphorus levels. For enzymatic activities and soil biological attributes, the residual phosphorus doses in the soil did not result in significant differences. Regarding carbon stock and different soil organic matter fractions, higher values were obtained in areas with the highest residual phosphorus doses.

However, technologies and management practices, such as those presented in the current study, can contribute to more resilient and sustainable agricultural practices, with increased productivity and reduced pressure on production areas. This is important in the light of future scenarios, where there will be increased global demand for food and products from agriculture, in a highly fertilizer-dependent agricultural system.

4. Discussion

4.1. Soil Chemical Properties

Soil pH, in isolation, is the factor that most affects the availability of P in the soil, as values close to 6.5 promote the greatest availability of this nutrient in the soil solution. The presence of other ions in the soil can also interfere with P absorption. For example, Mg has a synergistic effect on P absorption, as it acts as a carrier for P by activating ATPase in membranes and generating ATP during photosynthesis and respiration [26].

This pH effect helps interpret the results of the low P availability (an average of 8 mg dm⁻³) in the soil fertility analyses of the present study (Table 1, Table 2, Table 13 and Table 17), as pH values were close to 5.2. Conversely, Mg levels were high (an average of 15 mmolc dm⁻³), which may have increased P absorption by crops, temporarily immobilizing the element in plant biomass.

However, the results previously presented in Figure 6 show an increase in soil P content with the rising application rates of the treatments, indicating that the ideal dose was 192.5 kg ha⁻¹ of P₂O₅. This pattern reflects the tendency of phosphorus to become fixed in the soil matrix at higher application rates, resulting in decreased availability beyond an optimal threshold. These findings support the hypothesis that phosphorus dynamics are significantly influenced by long-term fertilization strategies in conservationist systems.

In integrated production systems, organic matter input into the soil is higher than in conventional systems, which is closely related to the soil microbial population. Microorganisms play a crucial role in organic matter transformations, using crop residues as an energy and nutrient source for cell formation and development. This process leads to the temporary immobilization of carbon, nitrogen, calcium, magnesium, phosphorus, sulfur, and micronutrients, which are later released into the soil upon microbial death and become available to plants again [27]. This also supports the lower average levels of P, Ca, Mg, and, consequently, BS and V%, observed in inoculated areas of this study.

During the decomposition and mineralization of soil organic matter, H+ ions, organic acids, and S are released, increasing soil acidity [28]. This explains the higher average potential acidity and sulfur content in inoculated areas. Additionally, dry matter production (crop residues) was higher in the inoculated area throughout the crop rotation period preceding this study [29]. The decomposition of this organic material increased soil acidity in these areas and, due to higher grain yield, nutrient export was also greater, which explains the lower nutrient levels in the soil of the areas inoculated with A. brasilense.

4.2. Crop Productivity

Studies evaluating the relationship between phosphorus doses (52.32 and 4.36 mg dm⁻³ of P) in potted plants, without maintenance fertilization, in Quartzarenic Neosol, and the root system development of Urochloa plants found that higher phosphorus doses resulted in an average increase of 172.5% in root mass [30]. Similarly, in a study conducted on Dystrophic Red Latosol, testing phosphorus doses (0.0, 277.5, 555.0, 832.5, and 1110.0 kg ha⁻¹ of single superphosphate) for dry matter yield in maize, greater production was observed at higher doses [31]. This increased plant growth and production with higher phosphorus supply is attributed to more efficient root development, enhancing water and nutrient uptake due to a larger soil exploration area. This leads to better phosphorus distribution in the plant, allowing its translocation and storage in superficial roots and aerial parts [32].

In this context, increasing soil P levels enhances dry matter production in crops, particularly in pastures, due to their dependence on this nutrient for tillering and root development [33]. However, in the present study, no significant yield difference was observed between the control (0 kg ha⁻¹) and the highest residual dose (240 kg ha⁻¹) (Table 3 and Table 22). This can be attributed to the area being under a stabilized no-till system for 20 years, with a diverse crop rotation, involving plants with different root types, sizes, and nutritional requirements. Consequently, this increases nutrient cycling efficiency, as nutrients are absorbed at different soil depths, accumulated in plant tissues, and redeposited into the soil after mineralization, meeting the crop’s phosphorus needs [34].

The bacterium A. brasilense colonizes the interior and surface of roots as facultative endophytic diazotrophs. In addition to fixing nitrogen, it can also promote plant growth by solubilizing phosphate, producing phytohormones (auxins, gibberellins, and cytokinins), and stimulating root metabolism [35]. This microorganism benefits grasses by primarily enhancing root development, which, in turn, improves soil surface utilization for water and nutrient absorption [36]. These findings support previous research showing that A. brasilense facilitated the development of Cynodon (Croastcross) plants during the summer of the second production year, after one year of application [37]. However, in the present experiment, the presence or absence of A. brasilense did not significantly affect Paiaguás grass yield (Table 3). This effect may have been due to the substantial organic residue accumulation after harvest, resulting from the long-term adoption of no-till farming. The abundance of organic material in the soil provides sufficient nutrients for proper crop development [38].

Phosphorus deficiency leads to reduced plant productivity and stature. In soybeans, it lowers the species’ potential during early reproductive stages by reducing flower production and increasing flower abortion [39]. However, in the present study, no grain yield differences were observed among treatments with different residual phosphorus applications (Table 4).

However, as observed in Figure 7, in the area without inoculation, there was a significant quadratic response of soybean grain yield to the residual phosphorus rates, with an ideal dose of 151.6 kg ha⁻¹ of P₂O₅. In contrast, in the inoculated areas, no regression adjustment was observed, indicating that in the absence of inoculation, phosphorus doses influenced soybean yield.

In a study by [39], phosphate fertilization resulted in a quadratic adjustment for soybean yield, increasing at lower doses and declining at higher P₂O₅ doses. This partially aligns with the present study’s results, where, despite the lack of a regression fit, a similar trend was observed due to the residual effect of P₂O₅ doses. Under excessive phosphorus conditions, P can suppress the absorption of cationic micronutrients, particularly zinc (Zn), but also copper (Cu), iron (Fe), and manganese (Mn). Excessive P use can also reduce CO₂ fixation and starch synthesis [40], and this nutrient imbalance may explain the decrease in morphological attributes and crop productivity at the highest residual dose (240 kg ha⁻¹ P₂O₅).

However, under P deficiency, plants accumulate more sugars as a form of potential chemical energy. At the same time, the lack of ATP limits biosynthesis processes, decreasing RNA, starch, and lipid synthesis, leading to protein deficiency and the increased accumulation of soluble nitrogen compounds, which hinders vegetative development. P also regulates photosynthesis and carbohydrate metabolism, mitigating plant growth limitations [9]. This suggests that the soybean crop in this study did not experience P deficiency, as uniform productivity was observed across residual P₂O₅ treatments.

In another study evaluating the effect of broadcast phosphorus doses (40, 80, 120, and 160 kg ha⁻¹ of P₂O₅) on soybean development and yield, no significant differences were observed among treatments regarding grain yield [41]. However, these results are justified by unfavorable climatic conditions during the experiment.

The bacterium A. brasilense has a modest biological nitrogen fixation (BNF) capacity compared to Bradyrhizobium. However, Azospirillum’s primary microbial function is promoting plant growth through phytohormone production, particularly in root systems, enhancing nodulation and BNF performed by Bradyrhizobium. Additionally, it increases root volume, expanding the soil contact surface for nutrient uptake. Thus, soybean plants co-inoculated with Bradyrhizobium on seeds and Azospirillum in the soil (residual effect from previous grass crops in rotation) exhibited earlier and increased nodulation, leading to an average yield increase of 16% [42,43].

In a study conducted by [44], it was also reported that co-inoculation with Bradyrhizobium sp. and Azospirillum in soybeans promoted an 11% increase in root dry mass, a 5.4% increase in the number of nodules, a 10.6% increase in nodule dry mass, and a 3.2% increase in grain yield compared to conventional inoculation with Bradyrhizobium sp.

However, the results obtained in the present study, which showed no significant differences in productivity between inoculated and non-inoculated areas, can possibly be explained by the favorable chemical conditions of the soil. These conditions are a result of the management practices employed in the area, such as the adoption of no-tillage systems and the rotation of different plant species. These practices also contribute to the formation of biopores, which are subsequently reused by successive crops, reducing the plant’s energy expenditure for root development [45]. Additionally, they allow for deeper root growth, leading to a greater volume of explored soil, which ensures proper nutrition and production of the crop.

Given the above, it can be inferred that the lack of significant differences between treatments may be attributed to the high quantity and quality of organic material in the area. This results from a consolidated no-tillage system under an integrated crop-livestock production management, where the chemical and biological balance of the soil is stabilized, thus minimizing the effects of the treatments.

Since up to 80% of the total phosphorus (P) in soils originates from plant residues (straw) and microbial biomass, which, upon decomposition by soil microbiota, converts organic P into mineral P through mineralization [46], this process is likely the reason why sorghum productivity did not show significant differences in response to residual P₂O₅ doses in the present study (Table 7). This is due to the P absorbed by the rotation crops and the accumulation of a large volume of plant material, which, during decomposition, releases the P necessary for the proper development of the crop.

In addition to nutrient availability, soil organic matter (SOM) can retain more water than its own mass, meaning that small increases in SOM can lead to significant gains in the amount of water available for the crop, thus improving drought tolerance. Both the increased water retention and availability due to organic content and the lower soil density, which allows for deeper root growth, enhance the crop’s ability to withstand dry periods with minimal impact on harvest yields [38]. Given that there was little precipitation during this intercropping study (Figure 1), the minimal impact on productivity can also be attributed to the quantity and quality of SOM.

The inoculation of A. brasilense also has positive effects on drought tolerance, as it improves plant morphological characteristics such as root branching, increased root biomass, and higher root hair density, which result in better soil profile exploration for water uptake, in addition to biological nitrogen fixation (BNF) and phytohormone production [47]. All these factors contributed to an increase in sorghum grain yield and panicle dry matter. However, for the production of Piatã grass in intercropping, a decrease was observed in inoculated areas, likely due to competition between crops, as only the sorghum was inoculated, leading to its dominance over the forage.

Conversely, the residual phosphorus doses led to the increased production of Piatã grass. Phosphorus is a limiting nutrient for achieving higher productivity, and, according to Liebig’s Law of the Minimum, this effect is even more pronounced in pastures. A deficiency of this macronutrient results in reduced root development, making plants less drought-tolerant and less capable of absorbing soil nutrients, leading to a lower tillering rate and, consequently, reduced forage mass production [48].

4.3. Phosphorus Fractions and Soil Health

Phosphorus concentration in the soil can increase either through mineral fertilization, which immediately supplies the nutrient to the crop, or organically, where P only becomes available after microbial activity mineralizes SOM, releasing inorganic phosphate ions (available P) [49]. Therefore, areas with a higher volume of organic material and SOM naturally have higher levels of labile P, which aligns with the results obtained in this study. In the non-inoculated area, where there was greater dry matter production (Table 23), the labile P fraction in the soil was 32.5% higher (Table 19).

Plants absorb only the P present in soil solution, where its ionic forms include H₂PO₄⁻, HPO₄²⁻, and PO₄³⁻, with the first form predominating in Brazilian soils due to their low pH [9]. The greater the nutrient availability, the higher the plant absorption. In areas inoculated with A. brasilense, due to the increased root volume, nutrient uptake and temporary immobilization in plant biomass are also higher, explaining the lower labile P content in the inoculated areas of this study.

The response of phosphorus fractions to residual fertilization was also distinct between inoculated and non-inoculated plots. In Figure 9, exponential increases in both labile and moderately labile P were observed in non-inoculated areas, suggesting a more efficient release of bioavailable phosphorus under conventional management. Similarly, Figure 10 shows an increase in inorganic phosphorus (Pi), indicating that phosphorus applied in earlier cycles may persist in forms accessible to plants under no-tillage conditions. These results reinforce the importance of understanding P dynamics under biological interventions.

The mineralization of organic P into inorganic P is related to microbial and plant enzymes involved in this cycling process, which act on SOM. However, there is a dynamic interplay between microbial mineralization and the immobilization of P (P from microbial biomass), making microorganisms both a source and, at times, a sink for rapidly cycling P [49]. This effect can be observed in the relationship between higher ß-glucosidase and arylsulfatase enzyme activity (Table 20) in non-inoculated areas due to the greater volume of dry matter (Table 23), resulting in higher labile P content in non-inoculated areas and the breakdown of the labile, moderately labile (Figure 9), and inorganic P (Figure 10) fractions. This indicates that a higher dry matter volume leads to greater enzymatic activity, enhancing the availability of the aforementioned P fractions. Meanwhile, in inoculated plots, which exhibited greater stability, enzymatic activity was lower and had no effect on the existing P fractionation from prior organic material degradation.

Microbial biomass P (P-mic) constitutes a highly variable fraction of soil organic P, representing, on average, 1 to 10% of total P and 2 to 24% of organic P. Many forms of organic P found in the soil, such as phospholipids and nucleic acids, originate from microbial sources [50]. However, even though P-mic is not readily available to plants, it serves as a dynamic reservoir of P in tropical soils due to its role in the accelerated decomposition of soil organic matter (SOM). Thus, the poorer the system is in available P, the greater the dependence on organic forms, particularly P stored in microbial biomass, which absorbs and immobilizes P from the soil solution and gradually releases it [51].

Highlighting the importance of SOM for the uniformity of crop production in the rotation studied, in areas with non-inoculated crops (A. brasilense), better soil exploration and, consequently, higher P uptake were driven by the greater input of organic material, which provided adequate nutrient supply for plant development (details in Figure 10 and Figure 11). In contrast, in inoculated areas, due to the hormonal effect of A. brasilense [52], even with a lower total dry matter input, greater root growth and improved soil exploration ensured adequate nutrition for the crops in rotation.

The rate of P immobilization depends on the quality of decomposing residues and the availability of carbon and nitrogen in the soil. If the P concentration in residues exceeds cellular demand, an increase in available inorganic P through mineralization is observed [53]. This supports the exponential increase in inorganic P in areas with the highest residual doses (240 kg ha⁻¹ of P₂O₅) (Table 19) and non-inoculated areas (Figure 10), which also exhibited the highest total dry matter productivity (Table 23). Conversely, if these residues have P concentrations below microbial biomass requirements, P must be supplied from the available inorganic P pool in soil solution, thereby depleting soil solution P [53]. This indicates that there was no P deficiency in the experimental area, regardless of treatment, as the crop obtained adequate P for development and production either through enhanced absorption efficiency due to inoculation (inoculated areas) or the greater volume of dry matter in non-inoculated areas. In this context, it is crucial to consider the specific conditions of the experimental area, which may have influenced the soil nutrient dynamics and plant responses.

Considering the absence of significant differences in several soil chemical attributes and crop productivity parameters despite the application of phosphate fertilization, it is important to highlight the peculiar conditions of the experimental area. The long-term no-till management system, established over more than two decades, contributed to the development of a stable and resilient soil structure, with adequate organic matter content and nutrient cycling capacity. According to the law of the minimum, once phosphorus availability reaches sufficient levels for crop development, additional fertilizer inputs may not lead to further increases in productivity. Therefore, the minor effects observed with the application of residual phosphate fertilization can be attributed to the well-balanced fertility of the system, even in control plots, reflecting the efficiency and sustainability promoted by conservationist practices.

4.4. Soil Biological Indicators

The production of phosphatases results from biochemical modifications at the cellular level, primarily triggered by decreased P uptake by microorganisms. This explains the increased enzyme activity in soils with low available P levels [54], indicating that there was no deficiency of this element for the soil microbiota in the present study. Approximately 30 to 50% of soil microorganism isolates exhibit phosphatase activity. For example, nitrogen-fixing bacteria, such as bean rhizobia, have a high capacity to produce phosphatases under P deficiency conditions to maintain high P concentrations in nodules, given the high demand for this nutrient in biological nitrogen fixation [55].

Regarding the activity of arylsulfatase (AS) and β-glucosidase (βG) enzymes, they were more active in non-inoculated areas. This behavior may be attributed to the higher SOM content in these plots, as previously mentioned, increasing the volume and diversity of the soil microbiota. However, there is also the hypothesis of competition between native soil microorganisms and those introduced via inoculation. Essentially, two factors are recognized as the main causes of species reduction or extinction: the loss of original habitat and the introduction of invasive species. In the first case, habitat loss, the high complexity of soil structure and composition provides considerable protection to the soil microbiota. In the second case, species introduction is more closely related to agricultural practices, as the use of organic and inorganic inputs, soil management, and the introduction of plant species can incorporate new microbial species [56].

Microorganisms in recovering soils exhibit more complex interactions than those in degraded areas, suggesting greater stability and environmental resilience in restored soils [57]. This finding indicates that lower enzymatic activity in inoculated areas may also point to the stabilization and balance of the physical, chemical, and biological components of the plots.

The determination of enzymatic activities provides valuable information on soil health, as enzyme activity is influenced by soil carbon stock and changes in land use [58]. Thus, considering a scatter plot showing the relationship between enzymatic activity and total organic carbon (TOC), it is possible to define four soil health states, represented in four distinct quadrants [20]. Evaluating Figure 12, most of the experimental area fell within Quadrant 1 (Q1), representing high-quality soils. This indicates that adopting good management practices over a long period yields positive results in soil recovery, balance, and resilience.

Recent studies have highlighted the crucial role of soil microbial communities in maintaining ecosystem functions and promoting plant productivity. [59] demonstrated that the interactions between microorganisms, soil properties, and plant diversity are essential drivers of soil microbial diversity in the Atlantic Forest biome. Applying this understanding to agricultural systems, the inoculation with Azospirillum brasilense evaluated in our study can be seen as a management strategy that seeks to harness similar synergistic relationships. By modulating microbial community structures and functional diversity, inoculation has the potential to influence nutrient cycling, enzymatic activity, and soil health, contributing to sustainable crop production under long-term no-till systems. This perspective reinforces the importance of adopting practices that promote beneficial microbial interactions as a path to enhancing soil resilience and agricultural sustainability.

Microbial respiratory activity (MRA) is defined as the total sum of all metabolic functions in which CO₂ is produced. Bacteria and fungi are the main organisms responsible for CO₂ release via soil organic matter (SOM) degradation. MRA is closely related to abiotic soil conditions, including moisture, temperature, aeration [60], and substrate availability in the soil (SOM). These factors also influence microbial biomass carbon (MBC) [61]. Due to the large amount of dry matter in the overall experimental area (Table 23), SOM content in the 0.00–0.10 m soil layer remained at intermediate levels, averaging 28 g dm⁻³ (Table 13). Consequently, TOC and MBC stabilized in the experimental plots (Table 21).

The association between MRA and MBC allowed for the calculation of the metabolic quotient (qCO₂), which is the ratio of MRA per unit of MBC over time. This metric estimates microbial substrate use efficiency and serves as a sensitive stress indicator when MBC is affected. Both tools are crucial for understanding transformations and losses in the soil organic compartments [17].

By analyzing MRA along with other soil quality indicators (Table 21), it was observed that MRA was higher in plots that did not receive A. brasilense inoculation. This suggests a system with greater CO₂ loss due to increased microbial activity. However, as MBC and TOC remained stable between inoculated and non-inoculated areas, and the metabolic quotient decreased in the inoculated area, this indicated lower CO₂ loss. This is an interesting outcome for nutrient cycling and carbon sequestration processes in the soil.

These microbial dynamics also help explain the observed changes in soil nutrient availability. In the inoculated areas, both enzymatic activity (β-glucosidase and arylsulfatase) and microbial respiration (MRA) were significantly lower, suggesting reduced microbial metabolic activity. This may have limited the mineralization of organic residues, consequently reducing the release of key nutrients such as phosphorus, calcium, and magnesium into the soil. At the same time, the inoculated plots showed a greater accumulation of organic carbon in stable fractions, indicating that microbial communities may direct resources toward carbon stabilization rather than nutrient cycling. Therefore, microbial activity data provide a valuable framework for understanding the nutrient dynamics observed in this long-term conservationist system.

The physical fractionation of soil organic matter, analyzed in this study as particulate organic carbon (POC), mineral-associated organic carbon (MOC), and total organic carbon (TOC), contributes to understanding its dynamics under conservation management systems [62].

In the present study, TOC levels increased with soil depth, differing from the findings of studies on organic production systems with different crop rotations within a year, where TOC levels decreased at greater soil depths [62]. However, in this study, high TOC levels in deeper layers could be explained by the long-term no-tillage system, which enabled greater organic material accumulation, resulting in higher carbon levels throughout the soil profile.

Inoculating various crops with the bacterium Azospirillum brasilense enhances root biomass production due to its hormonal stimulation [63]. This explains the higher values observed for TOC, MOC, and POC in this study. With greater root volume, more root-derived organic matter was left in the soil after each cultivation cycle. This also accounts for the higher TOC and MOC levels in deeper layers, as organic matter at the surface decomposed rapidly due to edaphoclimatic conditions.

When anthropogenic alterations occur in a native ecosystem, carbon inputs become lower than outputs or soil organic matter decomposition rates accelerate, leading to a reduction in soil carbon stocks [64]. Consequently, the increase in POC and MOC fractions observed in inoculated areas, along with higher TOC levels under greater residual phosphorus doses, indicates a better balance between SOM input and decomposition.

5. Conclusions

After 20 years under a conservationist no-tillage system, the only consistent effect observed was from the inoculation of grasses with Azospirillum brasilense. Although this practice led to greater dry matter production and nutrient export, which slightly reduced soil fertility and total dry matter accumulation, it also contributed to improved soil health, evidenced by enhanced biological indicators. In contrast, phosphorus fertilization alone did not significantly affect the productivity of rotational crops, indicating a limited residual effect in long-term no-till systems.

The high level of soil stability achieved through long-term conservationist management likely contributed to sufficient nutrient availability, particularly phosphorus, reducing the need for additional fertilization inputs. This finding highlights an important practical implication: the maintenance of conservationist systems can minimize fertilizer costs while preserving soil health.

Notably, in non-inoculated areas, higher microbial activity associated with organic matter degradation led to an increased metabolic quotient (qCO₂), suggesting greater carbon loss and reduced sequestration capacity. Inoculated systems showed improved soil carbon fractions and enzymatic activities, especially at higher residual phosphorus doses, contributing to better soil structure and organic matter stabilization across soil depths.

These findings support the hypothesis that bioindicators are sensitive tools for assessing soil health and that inoculation with A. brasilense can enhance phosphorus use efficiency and promote long-term sustainability in tropical cropping systems. Therefore, adopting microbial inoculation strategies alongside the maintenance of conservationist management practices presents a promising approach to optimize nutrient use, enhance soil resilience, and reduce production costs in sustainable agriculture.

Despite the robust experimental design, including replicated plots and detailed soil and plant evaluations, certain limitations should be acknowledged. The study was conducted over a two-year period, which, although sufficient to capture short-term responses, may not fully reflect longer-term dynamics of phosphorus cycling and microbial community shifts under conservationist management. Additionally, the sampling strategy, while adequate for plot-level comparisons, may limit broader extrapolations to highly heterogeneous field conditions. Future research should consider longer monitoring periods, the use of high-throughput sequencing to track microbial community composition changes, and evaluations across different climatic zones and soil types to validate and expand the applicability of these findings. Such approaches would contribute to a deeper understanding of how microbial inoculation and legacy fertilization interact with soil health and crop productivity over time.