The Use of Integrated Crop–Livestock Systems as a Strategy to Improve Soil Organic Matter in the Brazilian Cerrado

Abstract

This study aimed to analyze the carbon (C) stock and stabilization of soil organic matter in particulate- and mineral-associated fractions across different land use systems after 32 years of experimentation in the Brazilian Cerrado. The experiment was established in 1991 and was performed in Planaltina-DF. The treatments evaluated included continuous pasture with monoculture grasses; integrated crop–livestock systems under no tillage; continuous cropping under no tillage; minimum tillage; and the preservation of the native Cerrado biome in its original condition. Soil sampling was performed to a depth of 30 cm. Carbon and nitrogen (N) stocks were quantified for the years 2001, 2009, 2013, and 2023, with soil organic matter fractionation performed on samples from 2023. Land use change resulted in significant losses of soil C and N in areas managed with conventional soil preparation practices. Systems that promote plant diversity, such as integrated crop–livestock systems, enhanced soil C and N stocks (72.8 and 5.5 Mg ha⁻¹, respectively) and increased both particulate organic matter and mineral-associated fractions, most of which were in more stabilized forms. Integrated crop–livestock systems are management practices that offer an effective alternative to present methods in terms of combating climate change and supporting ecosystem sustainability.

1. Introduction

Land use change (LUC), particularly the conversion of large areas of the original Brazilian Cerrado into productive land through deforestation and burning, has led to a reduction of approximately 48% in native vegetation and the release of significant amounts of C into the atmosphere [1,2]. In this context, practices such as no tilling, crop rotation, and the use of integrated crop–livestock systems are recommended to increase food production per area, reduce C emissions associated with LUC, and to help achieve the targets set by the Paris Agreement [3,4].

Most soil C is stored in the form of soil organic matter (SOM) [5]. This C is a dynamic component, with exchanges occurring between the soil and the atmosphere, where plants act as intermediaries between these two compartments. Atmospheric C, in the form of CO₂, is captured and stored in plant biomass through photosynthesis [6,7,8]. As this biomass decomposes through biological activity, part of the C returns to the atmosphere via a process known as SOM mineralization [9,10]. In a given land use system, the processes of SOM deposition and decomposition reach a dynamic equilibrium after a few decades [11], which explains the relatively stable SOM stocks in areas of native vegetation.

The LUC can result in accumulation, reduction, or no significant change with regard to SOM levels. The extent of SOM loss depends on the intrinsic properties of the input material, which influence the recalcitrance of decomposing plant matter. Grass species generally have greater biomass production potential and yield residue rich in recalcitrant compounds, leading to slower decomposition [12]. In contrast, legume residues are rich in sugars and proteins, which stimulate soil nutrient cycling by increasing microbial activity, thereby accelerating decomposition and mineralization [13]. Crop rotation with grasses and legumes has the potential to increase C and N stocks and stabilize SOM [14,15,16].

Other factors influencing SOM stability include its interactions with soil, such as occlusion within aggregates and organomineral associations, the latter being the most stable and long-lasting [17]. The most direct way to assess the effects of LUC is by quantifying C stocks before and after the change, using native vegetation with similar climate and soil conditions as a reference for initial stocks. However, uncertainties persist about SOM stability, particularly regarding vulnerability to climate change and management practices [18,19].

One technique that provides important information about soil carbon stability is SOM fractionation, which enables the study of the quality and quantity of organic matter in different compartments and links the effects of its presence to changes in soil and environmental properties [20]. Particulate organic matter (POM) and mineral-associated organic matter (MAOM), quantified through physical fractionation, provide information with which to assess soil carbon stability in relation to different management practices [13,21]. Generally, POM consists largely of plant residue fragments, roots, and hyphae in various stages of decomposition, which still retain recognizable cellular structures. On the other hand, MAOM forms through several pathways, primarily via the mineral adsorption of low-molecular-weight compounds [16,22]. Since POM is more susceptible to decomposition than MAOM, the proportions of these fractions may indicate the vulnerability of production systems to C losses resulting from potential changes in management practices. However, these attributes are influenced by factors such as residue quality, the C/N ratio, the climate, management practices, and intrinsic soil characteristics like mineral composition and texture [23,24,25,26].

To evaluate the effects of the different land uses common in the Brazilian Cerrado biome on the physical fractions of SOM in a Yellow Oxisol with sandy loam texture, Gmach et al. [27] found that biome native vegetation and pasture, established with Urochloa brizantha under continuous stocking, had greater POM values in the soil surface layer (0–10 cm). Analyzing the granulometric fractions of SOM in a Red Oxisol with a clayey texture from areas managed under monoculture pasture and integrated crop–livestock (ICL) systems, Loss et al. [28] reported greater POM values in the native forest (up to 10 cm deep) and ICL (up to 40 cm deep) systems. The authors attribute these results to the greater contribution made by plant residues (straw and root system) in these areas. The management duration of the sites in these studies ranged from 2 to 17 years. Thus, long-term experiments were necessary to assess the variations of C and N in the soil of the Brazilian Cerrado, however, this process is slow and expensive.

Assessing soil C stocks after 22 years of converting Brazilian native Cerrado biome for various land uses, such as monoculture pasture, crops, and ICL systems, Sant’Anna et al. [29] reported the potential of the integrated system to accumulate C over time. Therefore, we aim to investigate what happened to soil C and N in the same area of the study of Sant’Anna et al. [29] in the 10 years since the last sampling was conducted. Our hypothesis is that the increase in soil C provided by the integrated crop–livestock systems is associated with greater production of grass plant residues and the diversity of species present. This increased residue accumulation is expected to be linked to an increase in the POM fraction, which is more vulnerable to losses if management conditions change. To further this understanding, the present study assesses C and N stocks and SOM stabilization through the POM and MAOM fractions under different land use systems after 32 years of experimentation in the Brazilian Cerrado biome.

2. Materials and Methods

2.1. Location and Experimental Design

The experiment was carried out at the Cerrados unit of the Brazilian Agricultural Research Corporation (Embrapa) in Planaltina, Distrito Federal, Brazil (Figure 1; 15°35′ S, 47°42′ W; 1200 m a.s.l.). According to the Köppen–Geiger classification, the regional climate is classified as Aw (tropical savanna), characterized by dry winters and rainy summers, with a mean annual temperature between 22 and 27 °C and a mean annual rainfall ranging from 1400 to 1600 mm [30].

The soil in this area is classified as an Oxisol (Typic Acrustox) according to the US Soil Taxonomy system and as “Latossolo Vermelho” in the Brazilian classification [31]. Its mineralogical composition consists of 50% gibbsite, 18% goethite, 14% kaolinite, 7% hematite, and 10% quartz and other minerals. This composition influences its ion exchange capacity, water retention, and phosphate adsorption characteristics. Figure 2 presents data on rainfall and the average annual temperature in the Federal District from 2000 to 2023.

Before the experiment area began in 1991, the soil chemical and physical properties in the 0–20 cm layer were characterized as follows: pH_(H2O): 6.0; OM (organic matter): 21.7 g kg⁻¹; P (Mehlich-1): 0.9 mg kg⁻¹; Al³⁺: 0.1 cmol_c kg⁻¹; Ca²⁺ + Mg²⁺: 2.9 cmol_c kg⁻¹; K⁺: 0.1 cmol_c kg⁻¹; fine sand: 258 g kg⁻¹; coarse sand: 76.7 g kg⁻¹; silt: 101.8 g kg⁻¹; and clay: 563.5 g kg⁻¹.

The experimental area consisted of native Cerrado biome, characterized as a typical savanna with shrubs and small trees interspersed within a continuous carpet of grasses [32]. In 1991, the native Cerrado area was cleared and divided into areas with different land use management practices.

To evaluate the treatments, an area of native Cerrado vegetation (NC) was maintained as a reference for comparison with the land use management treatments over time. Each experimental plot area measured 50 m in length and 40 m in width (2000 m²). The land use management systems evaluated were as follows: continuous pasture in grass monoculture (PM); integrated crop–livestock (ICL) systems under no-tillage conditions; continuous cropping under no-tillage (NT) conditions; continuous cropping under conventional tillage, which was changed to minimum tillage (MT) in 2013; and a native Cerrado biome area, which was maintained in its original condition (NC). Each treatment included four replicate plots of 40 × 50 m (N = 4), arranged in a randomized block design.

At the establishment of the experiment, dolomitic lime was applied at a rate of 5.8 Mg ha⁻¹, along with P₂O₅ (98 kg ha⁻¹), K₂O (98 kg ha⁻¹), micronutrients (63 kg ha⁻¹), and gypsum (2.8 Mg ha⁻¹). Subsequently, additional applications of lime were performed across the total area at rates of 1.7, 1.1, and 2.0 Mg ha⁻¹ in 1999, 2006, and 2013, respectively, along with 1.5 Mg ha⁻¹ of agricultural gypsum. The gypsum was incorporated into the conventional tillage systems and applied on the soil surface in the no-tillage systems.

From 1999 to 2004, soybean and maize were alternately cultivated in the NT and MT areas. During the growing seasons from 2004/2005 to 2015/2016, a repeated sequence of maize and cover crops was cultivated in the NT area. Maintenance fertilization was applied in the planting furrow during maize and soybean sowing in both NT and MT areas, consisting of 150 kg ha⁻¹ P₂O₅, 80 kg ha⁻¹ K₂O, 2 kg ha⁻¹ Zn (from ZnSO₄·7H₂O), and 10 kg ha⁻¹ of fritted trace elements (FTE BR 12) as a micronutrient source (3.2% S, 1.8% B, 0.8% Cu, 2.0% Mn, 0.1% Mo, 9.0% Zn, and 1.8% Ca). Annually, from 2004/2005 to 2015/2016, after the maize harvest, the following cover crop species were sown in the NT area: pigeon pea (Cajanus cajan (L.) Millsp), sunn hemp (Crotalaria juncea L.), black mucuna (Mucuna aterrima Merr.), and forage radish (Raphanus sativus L.).

In the PM area, Andropogon gayanus Kunth cv. Planaltina was established as a monoculture from 1991 to 1999, after which it was replaced by Urochloa decumbens (Stapf) R.D. Webster cv. Basilisk. Pasture maintenance fertilization aimed to change the base saturation to 50%, with annual applications of 20 kg ha⁻¹ P₂O₅, 50 kg ha⁻¹ K₂O, and 60 kg ha⁻¹ N. The ICL area, cropping, and pasture were alternated every four years. This rotation allowed the introduction of fertilized crops to restore soil fertility in the pastures. During the livestock phase, rotational stocking was implemented, with a 14-day grazing period followed by a similar rest period. Initially, Andropogon gayanus Kunth cv. Planaltina was used for pasture; this was later replaced by Urochloa brizantha (Hochst. ex A. Rich.) R.D. Webster cv. Marandu. Using Nellore beef cattle, the herbage allowance was maintained at 8 to 10 kg of green herbage mass per 100 kg of body weight and adjusted every 28 days.

2.2. Soil Sampling and Processing

In April 2023, undisturbed soil samples were collected using rings of known volume to assess soil density (Sd) at depths of 0–5 cm, 5–10 cm, 10–20 cm, and 20–30 cm. A trench was excavated in each plot, for a total of five trenches, with two samples collected from each trench wall. Disturbed soil samples were also collected at the same depths using a Dutch auger, with four samples per plot at each depth, totaling 80 samples (five plots × four depths × four samples). All soil samples were air-dried, crushed, and passed through a 2 mm sieve to obtain air-dried fine soil samples, which were stored for later analysis.

Subsamples were then ground according to the method of Arnold and Schepers [33] and used to determine the concentrations of total C, total N, and the natural abundance of ¹³C. These analyses were conducted using a Finnigan DeltaV continuous-flow automated mass spectrometer, coupled to a Costech total C and N analyzer (model ECS4010 Finnigan MAT, Bremen, Germany), at the John Day Stable Isotope Laboratory of Embrapa Agrobiologia. Soil fertility analyses (0–20 cm) were performed following the procedures of Teixeira et al. [34], and the results are presented in

Table 1. Soil fertility characterization for 2023 across different land use management systems.

LUM	Depth	pH	Al	Ca	Mg	K	P
		H2O	cmolc dm3			mg dm−3
PM	0–5	6.35	0.00	4.10	2.17	54.54	1.40
	5–10	6.21	0.00	3.67	1.32	35.00	0.40
	10–20	6.12	0.00	2.60	0.96	21.88	0.17
	20–30	5.85	0.00	1.44	0.66	17.38	0.11
ICL	0–5	5.80	0.01	6.11	2.13	67.04	30.04
	5–10	5.81	0.01	4.42	1.62	37.51	19.20
	10–20	5.52	0.03	2.69	0.83	22.74	3.17
	20–30	4.96	0.14	1.48	0.57	20.46	0.30
MT	0–5	5.64	0.00	4.08	1.14	41.95	31.83
	5–10	5.63	0.02	3.50	1.16	25.40	25.43
	10–20	5.17	0.11	2.21	0.54	28.58	9.44
	20–30	4.96	0.17	1.31	0.37	29.21	2.30
NT	0–5	6.08	0.00	5.21	1.68	66.23	33.34
	5–10	6.26	0.00	4.62	1.60	52.12	38.20
	10–20	5.76	0.00	2.92	0.92	39.30	9.54
	20–30	5.06	0.18	1.29	0.67	32.96	3.24
NC	0–5	4.51	0.84	0.29	0.23	39.79	0.01
	5–10	4.52	0.69	0.06	0.09	24.79	0.01
	10–20	4.65	0.49	0.02	0.05	20.34	0.01
	20–30	4.48	0.63	0.03	0.06	33.30	0.04

LUM: land use management; PM: pastureland area under monoculture; ICL: integrated crop–livestock system; MT: minimum tillage system; NT: no-tillage system; NC: native Cerrado biome. Al—titration method; Ca—atomic absorption; K—flame photometry; Mg—atomic absorption; P—Mehlich 1.

.

2.3. Calculations

Soil C stocks were calculated for the 0–30 cm soil layer. Bulk density measurements were performed to determine the mass of soil in each sampled layer. Any excess soil mass from a specific profile was discounted from its deepest layer (20–30 cm) using the total soil mass associated with the lowest bulk density value (indicating the least amount of compaction). This procedure is mathematically described by Sisti et al. [35]. The soil density used was the average bulk density value obtained from collections made in 2001, 2009, 2013, and 2023 for each depth (

Table 2. Average soil bulk density measured for the years 2001, 2009, 2013, and 2023 across different land use management systems.

Depth (cm)	PM	ICL	MT	NT	NC
0–5	1.01 ± 0.026	0.97 ± 0.016	1.14 ± 0.025	1.08 ± 0.038	0.76 ± 0.035
5–10	1.11 ± 0.022	1.17 ± 0.041	1.11 ± 0.021	1.03 ± 0.039	1.00 ± 0.046
10–20	1.07 ± 0.016	1.11 ± 0.028	1.12 ± 0.019	1.10 ± 0.014	0.95 ± 0.035
20–30	1.19 ± 0.024	1.13 ± 0.039	1.04 ± 0.016	1.14 ± 0.052	1.10 ± 0.032

PM: pastureland area under monoculture; ICL: integrated crop–livestock system; MT: minimum tillage system; NT: no-tillage system; NC: native Cerrado biome. Means ± standard error.

).

The forage grass Urochloa brizantha follows the C₄ photosynthetic pathway, while the native vegetation consists of C₃ plants. The C derived from U. brizantha roots was estimated to have a natural abundance of ¹³C = −12.50‰ [6]. To calculate the proportion (%) of soil C derived from the original forest vegetation (%CdrF), the equation developed by Cerri et al. [36] was applied:

%CdrF = 100 × (δ¹³CCP − δ¹³C-C₄)/(δ¹³C-C₃ − δ¹³C-C₄)

where δ¹³CCP represents the natural abundance of δ¹³C(‰_PDB) in the soil at each depth interval under pasture, while δ¹³C-C₄ and δ¹³C-C₃ denote the δ¹³C values of carbon from the roots of the pasture grasses (−12.50 to U. brizantha) and from the native forest vegetation.

2.4. SOM Particle Size Fractionation

Approximately 10 g of air-dried fine soil (samples collected in 2023) and 30 mL of sodium hexametaphosphate solution (5 g L⁻¹) were shaken for 15 h on a horizontal shaker [37]. The suspension was passed through a 53 µm sieve using deionized water. The material retained on the sieve, consisting of particulate organic matter (POM) associated with the sand fraction, was oven-dried at 50 °C, weighed, ground in a porcelain mortar, and analyzed for C and N using a Delta V Advantage mass spectrometer (Thermo Scientific, Bremen, Germany) coupled with an EA 4010 elemental analyzer (Costech, Valencia, CA, USA at the John Day Stable Isotope Laboratory at Embrapa Agrobiologia. The material that passed through the 53 µm sieve, consisting of mineral-associated organic matter (MAOM) from the silt and clay fractions, was quantified by the difference between total organic C and POM.

2.5. Statistical Analysis

For statistical analysis, data were processed using Sisvar software version 4.5 [38], obtained from the Federal University of Lavras, Minas Gerais. Analysis of variance (ANOVA) was carried out using the F test at α = 0.05. The Shapiro–Wilk test assessed the normality of soil C and N data. Fisher’s least significant difference (LSD) test was subsequently applied to compare means at a 5% significance level. Graphs were generated using SigmaPlot, displaying treatment means and the respective standard errors.

Data from the physical fractionation of SOM were tested for homogeneity of error variances and normality using the Shapiro–Wilk test. ANOVA was then performed, followed by a Tukey test (p < 0.05), and we used RStudio version 4.3.0 [39] for statistical analysis. Multivariate analysis was carried out using the principal component analysis method, generating a biplot graph using the prcomp functions from the stats package. A Monte Carlo test was carried out to verify the significant differences of the eigenvalues between sites using the PCA through ade4 package [40].

3. Results

3.1. Soil C and N Concentration

As expected, soil N and C concentrations were greater in the surface layers and decreased with a depth up to 30 cm in all evaluated years (

Table 3. The soil carbon concentration (g C kg⁻¹) at certain depths in the years 2001, 2009, 2013, and 2023 under the same land use management program.

Depth (cm)	PM	ICL	MT	NT	NC
2001
0–5	24.09 ± 0.66	28.13 ± 1.11	25.22 ± 0.59	25.47 ± 0.80	29.67 ± 0.45
5–10	22.61 ± 0.63 ab	24.89 ± 0.46 ab	24.12 ± 0.29 ab	22.39 ± 0.40 b	26.62 ± 0.51a
10–20	20.30 ± 0.71	22.00 ± 0.14	22.33 ± 0.58	21.84 ± 0.21	20.41 ± 0.39
20–30	14.44 ± 0.31 d	18.67 ± 0.31 ab	15.90 ± 0.52 cd	20.41 ± 0.22 a	17.68 ± 0.34 abc
2009
0–5	30.48 ± 0.93 b	36.88 ± 0.24 a	24.13 ± 0.32 c	36.88 ± 0.24 a	29.01 ± 0.51 b
5–10	24.25 ± 0.88	25.53 ± 0.46	24.45 ± 0.24	25.53 ± 0.46	25.18 ± 0.58
10–20	21.31 ± 0.75	21.90 ± 0.61	22.05 ± 0.14	21.13 ± 0.65	21.43 ± 0.33
20–30	17.10 ± 0.53	20.45 ± 1.24	17.45 ± 0.34	21.20 ± 1.17	16.47 ± 0.19
2013
0–5	28.05 ± 0.78 bc	37.78 ± 0.52 a	23.30 ± 1.18 c	31.15 ± 0.75 b	37.58 ± 0.46 a
5–10	23.86 ± 0.39 ab	26.17 ± 0.45 ab	23.97 ± 0.49 ab	22.65 ± 0.24 b	30.03 ± 1.61 a
10–20	21.65 ± 1.32	22.83 ± 0.20	21.29 ± 0.37	20.95 ± 0.25	21.39 ± 0.36
20–30	14.65 ± 1.33	19.73 0.44	17.80 ± 0.35	18.19 ± 0.82	18.04 ± 0.43
2023
0–5	35.53 ± 0.25 ab	39.53 ± 0.51 a	24.83 ± 0.04 c	36.02 ± 0.96 ab	32.54 ± 0.64 b
5–10	22.82 ± 0.58	24.80 ± 0.75	22.69 ± 0.25	23.32 ± 0.42	27.00 ± 0.17
10–20	17.22 ± 0.60 b	23.57 ± 0.77 a	18.91 ± 0.40 ab	20.60 ± 0.20 ab	22.24 ± 0.28 a
20–30	14.32 ± 0.73 b	18.39 ± 0.30 ab	15.96 ± 0.35 ab	17.63 ± 0.44 ab	20.18 ± 0.21 a

PM: pastureland area under monoculture; ICL: integrated crop–livestock system; MT: minimum tillage system; NT: no-tillage system; NC: native Cerrado biome. Means ± standard error followed by the same lowercase letter, in the same row, do not differ statistically from each other (Tukey p < 0.05).

and

Table 4. Soil N concentration (g N kg⁻¹) at certain depths in the years 2001, 2009, 2013, and 2023 under the same land use management program.

Depth (cm)	PM	ICL	MT	NT	NC
2001
0–5	1.33 ± 0.04	1.68 ± 0.06	1.37 ± 0.04	1.39 ± 0.05	1.61 ± 0.04
5–10	1.25 ± 0.02	1.41 ± 0.03	1.27 ± 0.01	1.20 ± 0.02	1.39 ± 0.05
10–20	1.05 ± 0.03	1.11 ± 0.01	1.18 ± 0.03	1.19 ± 0.02	1.05 ± 0.03
20–30	0.84 ± 0.02 c	1.00 ± 0.01 ab	0.84 ± 0.03 c	1.07 ± 0.01 a	0.92 ± 0.01 bc
2009
0–5	1.36 ± 0.07 c	2.27 ± 0.02 a	1.48 ± 0.02 c	1.97 ± 0.06 ab	1.68 ± 0.03 bc
5–10	1.44 ± 0.03	1.52 ± 0.01	1.30 ± 0.01	1.37 ± 0.03	1.42 ± 0.03
10–20	1.25 ±0.04	1.29 ± 0.02	1.14 ± 0.01	1.20 ± 0.01	1.20 ± 0.02
20–30	0.95 ± 0.03 ab	1.12 ± 0.03 a	0.92 ± 0.02 b	1.09 ± 0.02 ab	0.99 ±0.01 ab
2013
0–5	1.66 ± 0.05 b	2.71 ± 0.04 a	2.45 ± 0.01 b	2.25 ± 0.08 a	2.25 ± 0.07 a
5–10	1.43 ± 0.02	1.58 ± 0.02	1.38 ± 0.01	1.36 ± 0.07	1.75 ± 0.06
10–20	1.24 ± 0.07	1.35 ± 0.02	1.14 ± 0.02	1.20 ± 0.08	1.16 ± 0.06
20–30	0.91 ± 0.03	1.15 ± 0.02	0.93 ± 0.01	1.00 ± 0.04	0.95 ± 0.04
2023
0–5	1.98 ± 0.06	2.97 ± 0.05	2.08 ± 0.06	2.67 ± 0.10	2.01 ± 0.10
5–10	1.52 ± 0.05 b	1.74 ± 0.04 a	1.46 ± 0.05 b	1.75 ± 0.08 b	1.94 ± 0.09 ab
10–20	1.10 ± 0.04	1.63 ± 0.11	1.24 ± 0.05	1.15 ± 0.05	1.35 ± 0.03
20–30	0.81 ± 0.03	1.69 ± 0.07	0.90 ± 0.06	0.88 ± 0.05	1.02 ± 0.04

PM: pastureland area under monoculture; ICL: integrated crop–livestock system; MT: minimum tillage system; NT: no-tillage system; NC: native Cerrado biome. Means ± standard error followed by the same lowercase letter, in the same row, do not differ statistically from each other (Tukey p < 0.05).

). Areas under PM and ICL management showed increases in soil C concentrations over time, with increased of 14 and 11 g C kg⁻¹, respectively, in 2023 compared to 2001. The lowest surface C concentration was observed in the MT areas, where soil C concentration decreased over time, with reductions of approximately 10 g C kg⁻¹ in 2023 compared to 2013 and 3 g C kg⁻¹ compared to 2001.

Surface N concentrations also increased over time for the PM, ICL, and NT management areas (Table 4), with the ICL system showing the greatest N concentration in 2023 (with an increase of 1.29 g N kg⁻¹ compared to 2001). Additionally, an increase in the N concentration was observed in deeper layers under this management system over the years.

The soil N concentration in the surface layer increased in areas managed under MT and in the native Cerrado biome until 2013, followed by a decline. Overall, soil C and N concentrations varied with the land use management systems adopted over the years, with MT showing the lowest levels of soil C and N concentrations in 2023. Additionally, the area managed under PM, with 100% of the recommended fertilization, exhibited lower soil C and N concentrations in the deepest soil layer when compared to the soils under the other land use management systems.

3.2. Soil C and N Stocks and C/N Ratio

Soil C and N stocks for the 0–30 cm layer showed distinct patterns bases under land use management programs (Figure 3). Soil C stock ranged from 51.1 to 72.8 Mg ha⁻¹, while the N stock ranged from 3.1 to 5.5 Mg ha⁻¹, with the greatest C and N stocks recorded in the ICL system in 2023 (Figure 3).

In the ICL system, soil C stock increased over time, with the greatest stock recorded in 2023, contributing approximately 7 Mg ha⁻¹ more C than in 2001 (Figure 3A). Conversely, C stocks declined for the last 10 years under the PM and MT systems, with the MT showing reductions of approximately 14 Mg ha⁻¹ and 8 Mg ha⁻¹ compared to 2013 and 2001, respectively. Compared to the native Cerrado biome, which served as a reference, the average of C stocks up to 30 cm deep were greater for the ICL and NT systems. The native Cerrado biome exhibited greater C stocks than the PM and MT land use management systems.

Soil N stock was greatest in the ICL system, followed by the NT, MT, NC, and PM systems (Figure 3B). The PM system had the lowest soil N stock, with a difference of −2.1 Mg ha⁻¹ compared to the ICL system. Both the ICL and NC systems showed a linear increase in the N stock over time, with the greatest values recorded in 2023 (p < 0.05). In the ICL system, soil N stock increased by approximately 2 Mg ha⁻¹ compared to 2001, while the MT system experienced a decline of about 0.6 Mg ha⁻¹ over the last 10 year compared to 2013. Notably, there was no significant difference in soil C and N stocks between the PM and NT systems in the 0–30 cm layer, with similar levels observed after 32 years of the same land use management.

The C/N ratio values in the 0–30 cm layer ranged from 14 to 19 (Figure 4), which is considered high for most soils, suggesting the presence of recalcitrant C, such as charcoal. The C/N ratio decreased linearly over time, with 2023 exhibiting lower values and with significant differences observed among the PM, ICL, and MT land use management systems.

The ICL system showed the largest reduction in the C/N ratio, decreasing from 18 in 2001 to 14 in 2023. A decrease in the C/N ratio was also observed in the NC system. The natural abundance of ¹³C was analyzed using samples collected in 2023, highlighting the impact of pasture systems on soil C, as evidenced by the results of these assessments (Figure 5).

After 32 years of land use management, there was an increase (i.e., less negative values) in the natural abundance of ¹³C in soil C across all four depths. This indicates the contribution of C derived from pasture (C₄ plants). In the NT and MT systems, the natural abundance of ¹³C ranged between −18 and −19‰, suggesting that part of this contribution comes from the deposition of C₄ plant residues, particularly those from maize crops.

These crops contribute to the accumulation of soil C in deeper layers, such as 20–30 cm. The C stocks from C₃ and C₄ plants in the soil were calculated, revealing that in the 0–5 cm layer, pasture contributed approximately 60% of the C₄-derived carbon (C-C₄), with this proportion decreasing with depth (Figure 6).

In the ICL system, the contributions of C₃ and C₄ plants to the soil C stock were similar in the surface layer, but notable differences emerged in the 10–20 cm layer, where C-C₄ accounted for approximately 60% of the soil C stock. In systems that did not introduce Urochloa, such as NT and MT systems, lower proportions of C-C₄ were observed. Conversely, there was greater protection of C-C₃ in the deepest layer, with more than 70% of C-C₃ present across all analyzed land use management systems.

3.3. Physical Fractionation of Soil

The soil C concentration in the POM only showed statistical differences between land use management systems in the 0–5 cm layer (

Table 5. Carbon and N concentration, C:N ratio and ¹³C abundance of the particulate organic matter (POM) and mineral-associated organic matter (MAOM) after 32 years under the same land use management system.

Depth (cm)	PM	ICL	MT	NT	NC	CV
C-POM (g kg−1 of soil)						%
0–5	2.31 b	3.12 a	1.26 b	1.19 b	1.74 b	31.16
5–10	1.83	1.82	2.14	2.02	1.30	30.40
10–20	1.49	1.38	1.46	1.31	1.15	27.39
20–30	0.90	1.13	0.94	0.82	0.79	27.86
N-POM (g kg−1 of soil)
0–5	0.06 b	0.17 a	0.06 b	0.04 b	0.07 b	41.74
5–10	0.04	0.08	0.12	0.10	0.05	41.98
10–20	0.03 ab	0.06 a	0.06 ab	0.05 ab	0.04 b	21.21
20–30	0.02	0.04	0.03	0.03	0.03	39.17
C:N ratio-POM
0–5	39.64 a	18.12 b	23.85 b	32.59 a	23.65 b	13.88
5–10	40.99 a	21.81 c	18.22 c	20.27 c	27.37 b	9.08
10–20	38.69 a	22.15 c	23.93 c	25.54 bc	31.09 b	10.72
20–30	40.18 a	26.53 b	27.46 b	32.04 ab	30.75 ab	15.01
C-MOAM (g kg−1 of soil)
0–5	33.21 a	36.41 a	23.57 b	34.82 a	30.80 a	7.94
5–10	20.99 b	22.98 ab	20.55 b	21.28 b	25.70 a	8.42
10–20	15.72 c	22.19 a	17.44 bc	19.28 abc	21.09 ab	10.66
20–30	13.41 b	17.26 ab	15.02 b	16.80 ab	19.38 a	11.70
N-MOAM (g kg−1 of soil)
0–5	1.92 b	2.79 a	2.02 b	2.63 ab	1.92 b	14.30
5–10	1.48 ab	1.65 ab	1.34 b	1.64 ab	1.89 a	14.13
10–20	1.06	1.57	1.17	1.10	1.31	21.85
20–30	0.79	1.64	0.86	0.86	0.99	23.39
C:N ratio in MOAM
0–5	17.52 a	13.06 bc	11.79 c	13.31 abc	16.23 ab	12.98
5–10	14.32	13.88	15.47	13.27	13.94	12.13
10–20	14.86	14.52	15.04	17.98	16.09	14.96
20–30	16.92	13.43	18.06	20.81	19.68	22.72
13C abundance of carbon in POM (‰)
0–5	−16.09 a	−17.09 b	−21.98 c	−22.43 c	−25.86 d	2.98
5–10	−17.65 a	−19.39 b	−20.35 b	−20.14 b	−25.05 c	3.38
10–20	−18.72 a	−20.30 ab	−21.95 b	−21.71 b	−24.89 c	3.41
20–30	−19.17 a	−21.21 b	−22.56 b	−21.97 b	−24.35 c	3.09
Proportion of C4 carbon in POM (%)
0–5	72.73 a	60.26 a	30.95 b	25.21 b	0	38.37
5–10	58.47	45.60	37.38	41.09	0	44.07
10–20	48.32 a	35.51 b	23.29 b	25.19 b	0	25.47
20–30	43.33 a	25.66 b	15.96 bc	20.73 bc	0	40.02

PM: pastureland area under monoculture; ICL: integrated crop–livestock system; MT: minimum tillage system; NT: no-tillage system; NC: native Cerrado biome. Means followed by the same lowercase letter, in the same row, do not differ statistically from each other (Tukey p < 0.05).

). The C concentration in the POM was greatest in the surface layer and decreased with depth. Notably, the ICL systems showed significantly greater C and N values in POM compared to the other land use management systems.

The percentage of total C in the POM fraction ranged from 3 to 9%, while MAOM accounted for over 90% of the total soil C. When examining the C/N ratio of POM, the PM system showed greater values at all depths studied (p < 0.05), suggesting the presence of materials with characteristics similar to those of the crops in that system. In all treatments, greater C and N concentrations in the MAOM fraction were observed in the topsoil, with values decreasing with depth. In the 0–5 cm layer, the ICL system showed the greatest C and N concentrations in MAOM, followed by NT, PM, NC, and MT systems (p < 0.05). In the deepest layer (20–30 cm), the NC system had the greatest C concentration in MOAM. The ICL system recorded the greatest N concentrations in MOAM, both in the top layer (0–5 cm) and in the deeper layer (10–30 cm). Notably, the PM system had lower C and N concentrations in the 10–30 cm layer compared to the other land use management systems.

When evaluating the origin of C in POM, it was observed that over 50% of the C in pasture systems (PM and ICL systems) originated from pasture vegetation. In the monoculture pasture system, there was a greater contribution of C₄ vegetation to POM carbon, especially in the topsoil layer, where it exceeded 70%. This observation was further supported by the analysis of ¹³C abundance in POM, which showed less negative values in the PM system, followed by the ICL system, indicating that C₄ vegetation had the strongest influence.

3.4. Principal Components Analysis

Principal component aAnalysis (PCA) was performed to investigate the interactions between land use management systems and soil attributes (Figure 7). The PCA explained 69.3% of the variation in the results, with the first principal component (PC1) explaining 50.1% and the second (PC2) explaining 19.2%.

The data indicate distinctions between the native Cerrado biome and the other management systems. The native Cerrado biome area shows a strong positive correlation with exchange Al, while negative correlations are observed with pH, Ca, Mg, and P. Conversely, these last attributes exhibit positive correlations with the ICL, MT, PM, and NT areas. The ILC system was positively correlated with total N and C stocks, exchange K, and organic matter fractions, while the PM system was more associated with high C/N ratios.

4. Discussion

4.1. Soil C and N Dynamics

The results illustrate the pattern of C and N in soils managed under different land use management systems in the Brazilian Cerrado biome. These were evaluated 32 years after the experiment began, specifically in 2001, 2009, 2013 and 2023. Generally, greater accumulation of plant biomass and the influence of roots system on the soil surface (0–10 cm layer) resulted in greater C and N concentrations in this layer, regardless of the land use management system used or the year evaluated. In all treatments, total C and N concentrations decreased with depth (Table 3 and Table 4), a pattern commonly observed in soils under natural vegetation, pastures, and no-tillage systems due to the rapid decline in root mass with increasing soil depth [29,41,42].

The adoption of systems with reduced mechanical soil disturbance, such as ICL, NT, and PM systems, led to increased soil C levels over time compared to the MT system. Among these systems, the ICL area consistently exhibited greater C and N concentrations in the surface layer for all the years evaluated. This increase is attributed to the greater contribution of organic residues to both the soil surface and profile, which promotes the accumulation of these nutrients [4,43]. Several studies also report increases in C and N levels with the adoption of the ICL system [28,44,45,46]. One possible explanation is that land use management practices incorporating grass species in no-tillage farming can maintain or enhance SOM content in the surface layers [47,48].

The greater C and N stocks were recorded in the ICL system, with increases of 11% for C and 52% for N compared to the sampling carried out in 2001. This was followed by the NT system, which showed increases of 4% for C and 26% for N over the same period in the 0–30 cm layer. Both systems surpassed the C and N stocks found in the reference area (NC). Conservation systems such as ICL and NT are characterized by the substantial accumulation of plant biomass residues in the surface layer and within the soil profile. This is combined with minimal soil disturbance and reduced biomass decomposition by the soil microbiota, which collectively favor greater C and N stocks [49,50]. Other studies carried out in the Cerrado biome also observed greater C and N stocks in ICL systems [4,28,29]. These findings further emphasize the critical role of plants in influencing C and N stocks within these systems.

In the integrated crop–livestock system, maize crop was growing at the time of soil collection, following previous harvests of soybeans that preceded a four-year pasture period. The pasture contributes significant amounts of residues with a high C/N ratio, enhancing soil coverage and C content. On the other hand, leguminous crops, such as soybeans, provide N for subsequent crops [51]. Thus, the greater diversity of species in the system, along with the increased deposition of plant residues and a slower degradation rate of the residues over 27 years, contributed to the enhancement of both C and N stocks in the soil.

The MT area exhibited lower C and N stocks. This decline can be attributed to its previous management under a conventional tillage system before the transition to minimum tillage in 2013, which may have favored C mineralization. Lal [52] noted that minimum tillage promotes continuous soil coverage and the incorporation of crop residues, thereby enhancing soil C stock. However, the input of plant residues into the soil (labile carbon) can either accelerate or delay the SOM decomposition, a phenomenon known as the priming effect, which intensifies with greater rates of labile C input [53]. Consequently, it is believed that the change in management practices, including the addition of residues with greater N content, has accelerated the mineralization of previously stored C due to the priming effect, resulting in a decrease in C content. In the coming years, this area is expected to reach a new state of equilibrium.

The high C and N stocks in the native vegetation area of the Cerrado biome are attributed to the significant contribution of plant residues to the soil surface, characterized by a high C/N ratio, along with the absence of human intervention [54]. In the current experiment, there was an increase in C and N stocks in these areas over the years, which can be explained by the recent occurrence of fire, suggesting a recovery process. The Cerrado biome has a history of regular fires, and studies indicate that burning can enhance biomass growth and production by facilitating nutrient release from ash [55]. Additionally, it can improve coverage and biodiversity [56]. This combination of factors may contribute to the observed increases in soil C and N stocks, as reported by Abdalla et al. [57] and Manson et al. [58], in pasture areas.

The C/N ratio in the land use management systems ranged from 14 to 19, with the greatest values recorded in the NC area. These values are considered elevated, and similar high C/N ratios have been reported in other Cerrado studies [59,60,61]. Researchers attributed these elevated C/N ratios to the significant presence of pyrogenic C in the soils, a result of the region’s history of regular fires or slash-and-burn practices. The decrease in the C/N ratio over the years is an indication of a stabilization process of SOM, which is related to the large increase in MAOM. Additionally, there were no changes in the plant species used in each treatment over time that could explain the frequency of fresh residues seen at a lower C/N ratio. A potential benefit for agricultural production is an increase in nutrient availability for plants [62].

It was previously believed that charcoal in the soil was inert and resistant to degradation [63,64]. However, studies such as those by Cusack et al. [65] have shown that this residue can degrade over time. For instance, Knicker et al. [66] investigated the impact of burning on soil C stocks in pastureland in the Planalto region of Rio Grande do Sul, Brazil. Their research demonstrated that the contribution of pyrogenic C can diminish over time, influencing soil C dynamics more than initially understood. The authors found that some charcoal was more susceptible to microbial attack than previously assumed, being translocated through the soil solution and accumulating in subsurface layers. Similarly, Hilscher and Knicker [67] assessed the microbial recalcitrance of charcoal added to a Haplic Cambisol following the burning of ryegrass. After 28 months of incubation, the authors observed structural changes in the remaining charcoal, indicating its susceptibility to microbial degradation. Consequently, the increased N levels in these areas over time have enhanced microbial activity, facilitating the breakdown of more recalcitrant materials like charcoal and leading to lower C/N ratios (Figure 4).

The area under continuous monoculture of tropical grasses pastures showed a decrease in C and N stocks over the years, with lower stocks compared to the MT system in 2023. This finding contrasts with other reports on C stocks in well-managed pasture areas in the Cerrado [68,69]. It was previously believed that land use management systems incorporating substantial plant residue inputs, such as pastures with recommended fertilization and grazing management, would enhance soil C stocks. However, increased aboveground biomass production does not necessarily lead to proportional gains in soil C inputs or accumulation [70]. Despite the large amount of residue in this area, N limitation may restrict C sequestration, as it depends on nutrient availability and the C:N:P stoichiometric ratios in SOM [41,71,72]. Fornara and Tilman [73] proposed that C sequestration in pasture soils could be driven by increases in root mass, as demonstrated in long-term NH₄NO₃ addition experiments.

The ¹³C abundance under the NC biome area was more negative at the soil surface and decreased with depth (Figure 5), indicating a greater proportion of trees and shrubs (C₃ plants). This is typical of this natural vegetation [59,66]. However, the isotopic abundance value in the native area measured at −22‰, suggesting the possible introduction of C₄ species, which was likely stimulated by fire. The land use management systems under PM and ICL exhibited greater ¹³C abundance values (less negative) across all soil layers, which was attributed to the presence of C₄ grasses. The grasses significantly influence soil C dynamics by contributing surface residues and renewing the root system [27]. The C₄ plants contributed approximately 70% to the total C and 60% to the SOM fraction in the PM and ICL systems (Figure 6 and Table 5).

Soil C and N stocks are influenced by the balance between the input and decomposition of plant biomass in the system, along with factors such as climate, soil texture, nutrient availability, and land use management practices [74,75,76,77]. Over time, soil organic C is rarely stable; it may either accumulate or decline, with temperature and rainfall significantly influencing this process [78]. Conducting long-term experiments presents challenges due to the need for consistent management practices.

4.2. Soil Physical Fractionation

The separation of SOM into physical fractions is proposed as a method for better understanding soil C and N responses to global changes [13]. The hypothesis is that, in the Cerrado biome area, agricultural systems characterized by low soil disturbance and greater net primary productivity, such as well-managed pastures and integrated crop–livestock systems, support greater C and N stocks in the soil. In particular, the POM fraction is expected to exhibit the greatest variations, with the native Cerrado biome area serving as a reference. Generally, the greatest C-POM values in the 0–5 cm layer were observed in the following order: the ICL, PM, NC, MT, and NT systems (Table 5). For the deepest layer (20–30 cm), the lowest C-POM values were found in the NC area. These results can be attributed to the greater contribution of plant material (both above and belowground residues) to these systems, coupled with the absence of soil disturbance, as this fraction consists of plant residues at various stages of decomposition [54,74,79,80].

When quantifying the different SOM fractions across various land use management systems in the Brazilian Cerrado, lower POM values were recorded in the ICL area (0–10 cm) compared to the NC biome area, which exhibited greater values [81,82]. This finding contrasts with results from previous studies that reported lower POM values in the ICL system due to its shorter implementation time, which ranged from 3 to 13 years, whereas the current study analyzed an area with 27 years of land use management. Additionally, POM is influenced by factors such as the quality of plant residue, the C/N ratio, climate, land use, and management practices [23]. These factors may contribute to variations in results depending on the specific study environment.

The C-MAOM content in the topsoil layer was greater in the ICL system, remaining comparable to levels found in the NC, PM, and NT systems, which significantly differed from those in the MT system. This finding highlights the contribution of plant residues from pastures and maize crops to the soil surface. These species exhibit a higher C/N ratio, resulting in the slower decomposition of residues, which enhances the C content bound to silt and clay, thereby forming organo-mineral complexes [13]. The greatest fractions of soil C and N across all areas were recorded in the most stable fraction, indicating that these clay-rich soil exhibit greater resilience to changes in land use management.

It is important to note that the MAOM fraction can form through the adsorption of low-molecular-weight compounds onto minerals or through the decomposition and transformation of organic material by soil biota, leading to necromasses or exudates that are subsequently incorporated into the MAOM fraction [22,83,84]. The greater C and N content of MAOM in the ICL system results from the synergistic interaction between grasses and legumes, combined with nutrient availability, facilitating the decomposition of plant residues. Additionally, the clayed texture of the soil in the current experiment promotes the accumulation of these fractions. Evidence supporting this explanation is the lower C/N ratio observed in the soil layers with larger fractions of SOM, suggesting the presence of higher-quality residues. Some authors have noted that plant residues with a low C/N ratio enhance the efficiency of MAOM formation [85,86]. This phenomenon was evident in the ICL system, where a higher C/N ratio in the pasture area did not result in an increase in MAOM.

5. Conclusions

The results of this 30-year study confirm that land use management practices that cause greater soil disturbance lead to more significant losses of C and N stocks compared to more conservationist systems, such as no-tillage and integrated crop–livestock systems. In continuous monoculture pasture, despite the recommended fertilization with P and K, the C stocks have shown a strong downward trend over the last 10 years, indicating limitations in N availability within the system.

The average C/N ratio across all production systems is high (~18). This is primarily due to elevated initial values in the soil under native Cerrado biome area. This phenomenon can be attributed to the high concentration of C compounds resulting from frequent fires, which produce charcoal and other substances. However, the gradual addition of N to these management systems over the years has led to a decrease in the C/N ratio.

The adoption of systems that promote plant diversity, such as ICL system, has resulted in increases of 72.8 Mg ha⁻¹ for C and 5.5 Mg ha⁻¹ for N stocks in the soil. These increases were recorded for the particulate organic matter and mineral-associated organic matter, with most of these compounds existing in a more stabilized form. Considering the land use changes in the Brazilian Cerrado, the integrated crop–livestock system under no tillage is considered an effective strategy for combating climate change and promoting ecosystem sustainability.