1 Introduction

Sustainable agricultural practices and the vegetable oil supply are contemporary concerns. Environmental disturbances result mostly from inappropriate agricultural activities, heavy dependency on fossil fuels, and uncontrolled exploitation of natural resources. The improper development and management of monocrop pastures are among the harmful agricultural practices, because both can speed up land degradation, especially on soils with low fertility (Kichel et al. 2019). On the other hand, the significant increase in global demand for vegetable oils—>50% in the last decade—is mainly driven by food-based consumption and by the increasing interest in biodiesel as a form of carbon-neutral fuel (Lü et al. 2011; Hajjari et al. 2017). However, most vegetable oils used in biofuel production come from edible oil crops, including palm oil, soybean, and rapeseed (Rodionova et al. 2017).

The growing demand for food production may force an increase in land requirements for traditional oil crops, causing the replacement and displacement of non-energy crops (including pasture) into natural forest areas and agricultural intensification in the same arable lands. These direct or indirect land use changes may exacerbate environmental disturbances instead of mitigating them (Nguyen et al. 2017).

Agroforestry, including the silvopastoral system (SPS) and hortipastoral system (HPS), is a land use practice with great potential to sequester more carbon than traditional farming systems. These types of intercropping have the potential to remove more atmospheric CO2 due to the different physiological and growth patterns of the plant species composing the system (Yin et al. 2017). In Brazil, as in most of the tropical areas of the world, silvopastoral systems and hortipastoral systems usually use C4 grasses, such as Brachiaria, Panicum, and Setaria spp., as forage. Brachiaria grasses are considered shade tolerant; however, some Brachiaria species, viz., B. decumbens, may show significant physiological, morphological, and growth changes when cultivated under low light availability (Gomes et al. 2011; Silva et al. 2016). Therefore, the appropriate arboreal component and spacing between species must be chosen, since the microenvironmental conditions below the trees’ canopy will be modified, affecting the performance (i.e., dry mass gain) of the understory species, either positively or negatively.

The macauba palm (Acrocomia aculeata (Jacq.) Lodd. ex Mart.) is a ubiquitous palm tree in South America. In Brazil, the palm is well adapted to the adversities of savanna-like regions and degraded pastures. This species is currently in the spotlight as an emerging oil crop because the pulp and seeds of its drupaceous fruit has high oil content. Its high estimated productivity (20 to 24 t fruit ha−1) (Motoike and Kuki 2009) makes it a potential source of biomass for the biofuel, oil-chemical, and food industries (Montoya et al. 2016). Additionally, the fruit’s solid residues can generate coproducts of considerable market value, such as pellets, charcoal, protein, and fiber cakes. All these advantages, in addition to the capacity to grow in drought-prone and disturbed areas, make this palm a viable alternative crop for agroforestry arrangements.

To test the feasibility of A. aculeata–B. decumbens intercropping, this pioneering study analyzed the effect of different planting densities on microclimatic conditions and the ability of forage biomass availability in the understory to withstand low to moderate grazing pressure. Additionally, we estimated the carbon sequestration efficiency of the consortium.

2 Material and methods

The study was conducted at an experimental farm belonging to the Federal University of Viçosa (Universidade Federal de Viçosa) located at an altitude of approximately 1244 m (20° 40′ 14″ S and 42° 30′ 47″ W) in the Araponga municipality, Minas Gerais, Brazil. The region has a subtropical humid climate (Köppen classification: Cwa), with rainy summers and dry winters. The soil in the area is generally Red-Yellow Latosol (Santos et al. 2006) with a sandy clay texture. The primary treatments in the intercropping system consisted of four different spacings between the macauba palm (Acrocomia aculeata (Jacq.) Lodd. ex Mart) and Brachiaria decumbens cv. Basilisk (T5x4, T6x4, T7x4, and T8x4) plus the pasture areas (Tpasture only). The primary treatments were arranged in four blocks (BL1; BL2; BL3; BL4) as were the areas of pasture only (M1, M2, M3, M4) (Fig. 1). The blocks had different surface areas and numbers of plants: BL1, 3674 m2 and 94 plants; BL2, 4268 m2 and 104 plants; BL3, 3585 m2 and 93 plants; and BL4, 3444 m2 and 134 plants. Within each block, the arrangement of primary treatments was randomized, with no sequences from smallest to largest or vice versa. Intercropping was established in 2009, covering 1.4 ha without artificial irrigation. By the time of forage sowing, the palm trees (seed-propagated from wild palm trees found in the state of Minas Gerais) were 2 years old and approximately 2 m tall.

Fig. 1
figure 1

Experimental area location and primary treatment arrangement: 5m × 4m (T5x4), 6m × 4m (T6x4), 7m × 4m (T7x4), 8m × 4m (T8x4); and Tpasture only, in four blocks (Bl1, Bl2, Bl3, and Bl4) with four monoculture areas (M1, M2, M3, and M4). Aerial photographies from Google Earth /2016.

Full size image

In December 2017, soil analysis was performed at the soil analysis laboratory of the Federal University of Viçosa (Table 1).

Table 1 Soil analysis of the area study, in December 2017. pH in water, KCl e CaCl - rate 1:2.5 P - K Mehlich-1 Extractor Ca2+ - Mg2+ - Al3+ - Extractor: KCl - 1 mol/LH + Al - calcium acetate extractor 0.5 mol/L - pH 7.0 SB, sum of exchangeable bases; t, effective cation exchange capacity; T, cation exchange capacity pH 7.0; V, base saturation index; m, aluminum saturation index; P-rem, remaining phosphorus.
Full size table

Agricultural practices included pest control and fertilizer application. From 2010 to 2015, all the primary treatments areas and macauba plants (around the ring region) received fertilizer at least once a year (Table 2).

Table 2 Fertilization carried out in the study area in the tree and forage component between 2014 and 2015.
Full size table

2.1 Grazing cycles and evaluation frequency

All evaluations were performed between December 2014 and July 2016, including two rainy seasons and two dry seasons. During the evaluation period, eight grazing cycles were carried out, five in the rainy season (December, January, February and March) and three during the dry season (April, June, and July). One grazing cycle corresponded to the period between the “entrance day of the cattle” and the “re-entrance day of the cattle” into the area after grass regrowth. The duration of each cycle varied according to the time required for the cattle to lower the forage height to approximately 50% of its initial height (~35–45 cm from the soil surface), recorded prior to the entrance of the animals.

2.2 Climatic components: meso- and microclimates

The regional weather variables were monitored by a meteorological station located in Viçosa, Minas Gerais (INMET 2016), between 2014 and 2016. The meteorological records are shown as monthly average values.

The understory microclimatic and physical variables monitored in the intercropping areas and in the pasture-only area were atmospheric temperature (°C), atmospheric relative humidity (RH%), and available understory light (USL%). In each block, the temperature and relative humidity were recorded 30 cm above the soil surface and between the plant rows with a DT-500 datalogger (Instrutherm) during 1-h intervals between 6 am and 6 pm within 10 consecutive days and expressed as average values. The available understory light values were recorded with an AccuPAR LP-80 ceptometer (Decagon Devices, UK) in five distinct places in each primary treatment during rainy and dry seasons between 8 h and 11 h. The five records were obtained between the lines of the macauba plants and pasture only areas directing the photosensitive stem of the AccuPAR LP-80 in five random directions: north, south, east, west, south-west.

2.3 Determination of Brachiaria decumbens aboveground biomass

Prior to starting this work, we carried out preliminary tests to detect potential problems that could occur during the evaluation period. As the areas of the consortium were not very large, we reduced the sampled area of the forage by using a smaller frame (1m × 0.5m) tossed at random and increased the number of evaluations over time. By using this smaller frame, we reduced the chances of underestimating the samples in the event the frame fell too close or in the same area as the previous assessment preventing the collection of plant material from the same portion of the soil in a short period. To estimate the forage dry mass, the metal frame of 1.0 m × 0.5 m was randomly thrown on the pasture areas, and the grass contained within it was cut to 5 cm from the soil surface. The frame was cast only once in each primary treatment/block, with four samples per primary treatment. A subsample of approximately 200 g was taken from the freshly harvested plant material, totaling 800 g per primary treatment on average. These values were repeated in all the grazing cycles performed over 20 months of evaluation, as described in Section 2.1. The leaves and stems from each subsample were separated, weighed, and dried in a ventilated oven at 55 °C until a constant weight was reached. The leaf dry mass (LDM) and stem dry mass (SDM) were used to calculate the leaf/stem ratio (LDM/SDM) and aboveground forage dry biomass (A-FDB = LDM+SDM), expressed per area as t ha−1. The samples for estimating the average aboveground forage dry biomass (A-FDB) (before grazing cycles) were collected eight times, with 8 samples taken before “the entrance day of the cattle” and 8 taken after the cattle exit, corresponding to the eight grazing cycles described in Section 2.1.

2.4 Determination of macauba palm aboveground biomass

The aboveground macauba dry biomass (A-MDB) was estimated by the equation proposed by Frangi and Lugo (1985) for arboreal palm species and expressed as t ha−1:

$$ \mathrm{A}-\mathrm{MDB}=10.0+6.4\times \mathrm{PH} $$

where PH is the average of fifteen macauba plant heights (m) per spacing measured with a clinometer (Haglöf Co., Sweden).

Finally, the estimated aboveground forage dry biomass of the system (A-DBS) was obtained by the sum of A-FDB and A-MDB from the average of eight samples, including the dry and rainy seasons.

2.5 Aboveground stored carbon and CO2 assimilation estimates

The estimated aboveground stored carbon (A-SC) for each intercrop component (macauba palm and Brachiaria grass) was obtained using the formula proposed by Schlesinger (1991) and expressed in t ha-1:

$$ \mathrm{A}-\mathrm{SC}=0.455\times \mathrm{B} $$

where B is either the aboveground forage dry biomass (A-FDB) or the aboveground macauba dry biomass (A-MDB), determined as described previously. The aboveground stored carbon of the system (A-SCS) was calculated by the sum of A-FSC and A-MSC.

Similarly, the aboveground CO2 assimilated by each component—A-FCO2 Assim and A-MCO2 Assim—was determined using their respective stored carbon estimates (A-SC) based on the formula proposed by Rizvi et al. (2011) and expressed in t ha−1:

$$ \mathrm{A}-{\mathrm{CO}}_2=\left[\mathrm{A}-{\mathrm{SC}}_{\left(\mathrm{component}\right)}\right]\times \left(44/12\right) $$

The total amount of aboveground CO2 assimilated in the system (A-CO2 System) was obtained by the sum of A-FCO2 Assim and A-MCO2 Assim.

2.6 Data analysis

The experimental design was in split plots, randomized blocks, with four replicates. For the forage data analysis, five spatial arrangements and two seasons were used to determine the aboveground forage dry biomass (A-FDB) and leaf/stem ratio. The plots (primary treatments) were constituted by five spatial arrangements, four intercropping areas (i.e., macauba palm–Brachiaria grass consortium (T5x4, T6x4, T7x4, and T8x4)) plus the Tpasture only area (Fig. 1), and the subplots (secondary treatments) were constituted by the dry and rainy seasons.

For the analysis of aboveground macauba dry biomass (A-MDB), four spatial arrangements and two seasons were used. The plots (primary treatments) were constituted by four intercropping areas (i.e., macauba palm–Brachiaria grass; T5x4, T6x4, T7x4, and T8x4), and the subplots (secondary treatments) were constituted by the dry and rainy seasons.

The estimated aboveground dry biomass of the system (A-DBS), estimated aboveground stored carbon in the system (A-SCS) and aboveground assimilated CO2 in the system (A-CO2 System) of all five areas (T5x4, T6x4, T7x4, T8x4, and Tpasture only) were expressed in t ha−1. The analyses were based on the average value of macauba plants plus the accumulation of dry forage biomass from the eight grazing cycles, including the dry and rainy seasons. The collected data were subjected to ANOVA in a randomized block design (DBC), and the averages were compared when necessary using the statistical software SAEG®. The graphics and figures were made using SigmaPlot® 14.0 software.

3 Results

3.1 The understory microclimate in the intercropping system is determined by the macauba palm density

The mesoclimatic conditions during the duration of the study were typical of the region, i.e., rainy summers and dry winters. The average monthly temperature over the evaluation period remained between 15 and 25 °C, which was considered a mild winter for the region. The precipitation and air temperature averages were 776 mm and 4.5 °C, respectively, higher during the rainy season than during the dry season (Fig. 2).

Fig. 2
figure 2

Monthly averages of precipitation and temperature. The records correspond to two rainy seasons and two dry seasons. Rainy season I (December 2014 to March 2015); dry season I (April 2015 to August 2015); rainy season II (September 2015 to March 2016) and dry season II (April 2016 to August 2016).

Full size image

The microclimatic conditions in all five primary treatments varied according to the season and the macauba palm planting density, but there was no interaction between these factors. The average understory temperature in the intercropping areas (T5x4, T6x4, T7x4, and T8x4) varied between the seasons, being 5.9 °C during the dry season and 7.4 °C during the rainy season. However, the understory temperature differences between the closest and widest intercropping areas were lower, not exceeding 1.9 °C. The understory temperature of the intercropping areas increased gradually with the distance between the palms. The temperature between seasons in the Tpasture only area varied by 2.4 °C. This area attained the highest temperatures during the dry season and rainy season, at 31.0 °C and 33.4 °C, respectively (Table 3).

Table 3 Average value and standard deviation of daily temperature (T °C), relative humidity (RH%), and under story luminosity (USL%) of the dry and rainy seasons. *Primary treatments: Macauba palm–Brachiaria grass intercropping system, spacing of 5m × 4m (T5x4), 6m × 4m (T6x4), 7m × 4m (T7x4), 8m × 4m (T8x4), and Tpasture only. The images in the “Bottom view of the canopy” column were taken with a digital camera (Fujifilm FinePix S4500); the lens directed at the sky and positioned parallel to 5 cm from the ground surface.
Full size table

The understory relative humidity had the opposite response to that of temperature, i.e., the greater the distance between macauba palms was, the lower the recorded relative humidity (RH). The relative humidity values were higher in all the intercropping areas than in the Tpasture only area in both the dry and rainy seasons. The maximum and minimum RH values were recorded in the T5x4 and in the Tpasture only areas, respectively, with averages of 70.5% and 53.6% during the dry season and 62.9% and 57.2% during the rainy season (Table 3).

Light availability in the understory decreased in the intercrop areas as the distance between the palms shortened, with a maximum available understory light of 94.3% in the T8x4 spacing and a minimum of 14.4% in the T5x4 spacing. In the T5x4 and T6x4 areas, the overlap of the macauba palm canopy was evident and caused greater shading of the understory forage. In the T7x4 and T8x4 spacings, in contrast to denser spacings (T5x4 and T6x4), light interception by the palm canopy was smaller, allowing a greater passage of light to the understory, regardless of the season (Table 3). However, light availability in these spacings varied between seasons, and the available understory light in the T7x4 and T8x4 spacings increased by approximately 12% and 20% during the dry season, respectively, compared with that in the rainy season.

3.2 The forage yield is proportional to the amount of light available in the understory

The biomass values of the tree component of the intercropping system varied according to macauba palm spacing but not season. The A-FDB was 0.38, 0.48, 0.83, and 0.62 t ha−1 greater during the rainy season than during the dry season for the T5x4, T6x4, T7x4, and T8x4 intercropping areas, respectively (Fig. 3). The biomass value of the Tpasture only area was 0.24 t ha−1 less during the rainy season (Fig. 3).

Fig. 3
figure 3

Aboveground forage dry biomass (A-FDB) in two seasons, dry and rainy seasons. Primary treatment arrangement: 5m × 4m (T5x4), 6m × 4m (T6x4), 7m × 4m (T7x4), 8m × 4m (T8x4); and Tpasture only. Averages followed by the same letter within the same season, dry season (lower case) or rainy season (upper case), do not differ by Tukey test (P <0.01).

Full size image

The average per sample of A-FDB, A-FSC, and A-FCO2 Assim, expressed in t ha−1, increased concomitantly with the space between the macauba palm trees (Fig. 4a). The lowest values were obtained in the T5x4 and T6x4 spacings, which were approximately 64% smaller than the value obtained in the Tpasture only area. In the T8x4 spacing, which had the widest space between palms, the average A-FDB values were 4.9% smaller—without a significant difference (P<0.05)—than those found in the Tpasture only area. The A-FDB in the T7x4 spacing presented a noteworthy threshold value among the intercropping areas. In contrast, the average A-MDB, A-MSC, and A-MCO2 Assim (t ha−1) were inversely related to palm tree spacing (Fig. 4b). The highest A-MDB value was found in the T5x4 spacing, with a value approximately 12 t ha−1 greater than T8x4 spacing. The estimated values of A-MDB (5-year-old plants) obtained in this study were very similar to those estimated from the data obtained by Moreira et al. (2020) using destructive analysis of 4.8-year-old plants.

Fig. 4
figure 4

Aboveground forage dry biomass (A-FDB), aboveground forage stored carbon (A-FSC), aboveground forage CO2 assimilation (A-FCO2 Assim) (a); averages followed by the same letter do not differ by Tukey test (P <0.10). Aboveground macauba dry biomass (A-MDB), aboveground macauba stored carbon (A-MSC), aboveground macauba CO2 assimilation (A-MCO2 Assim) (b); averages followed by the same letter do not differ by Tukey test (P <0.05). Primary treatment arrangement: 5m × 4m (T5x4), 6m × 4m (T6x4), 7m × 4m (T7x4), 8m × 4m (T8x4), and Tpasture only.

Full size image

The Brachiaria leaf/stem ratio varied between seasons. This ratio was on average 1.12 in the dry season and 0.94 in the rainy season. During the dry season, the ratio was similar among the intercropping areas but differed somewhat (ratio ≤ 1.40), being slightly higher in the T8x4 spacing and the Tpasture only area. During the rainy season, no differences were detected. During the dry season, the leaf/stem ratio values of intercropping areas remained between 0.97 and 1.17. However, the leaf/stem ratio in the Tpasture only area was 1.2 times larger than the highest leaf/stem ratio value in the intercropped areas (observed in spacing T8x4). During the rainy season, the leaf/stem ratios of all primary treatments did not differ from each other. The intercropping area values (T5x4, T6x4, T7x4, and T8x4) remained between 0.83 and 1.14, and the lowest value of 0.83 was found in the Tpasture only area.

3.3 The aboveground dry biomass and efficiency of carbon sequestration will depend on the macauba planting density

The aboveground dry biomass of the system (A-DBS), aboveground stored carbon in the system (A-SCS), and aboveground assimilated CO2 in the system (A-CO2 System) varied among the primary treatments; however, they were consistently higher in the intercropping areas than in the Tpasture only area (Fig. 5).

Fig. 5
figure 5

Aboveground dry biomass system (A-DBS), aboveground stored carbon system (A-SCS), and aboveground CO2 assimilation of the system (A-CO2 System), of macauba palm–Brachiaria grass intercropping system and pasture only. Primary treatment arrangement: 5m × 4m (T5x4), 6m × 4m (T6x4), 7m × 4m (T7x4), 8m × 4m (T8x4); and Tpasture only. Averages followed by the same letter do not differ by Tukey test (P <0.01).

Full size image

The T5x4 area showed the highest values for A-DBS, A-SCS, and A-CO2 System, which were, on average, 2.6 times greater than those of the Tpasture only area. The values of these three parameters were similar between the T6x4 and T7x4 treatments and approximately two times greater than those for the Tpasture only area. The T8x4 treatment presented the lowest values among the intercropping treatments; however, the value was still 1.6 times greater than those observed in the monoculture system (Fig. 5).

Overall, the presence of macauba in the intercropping system contributed to an increase in A-DBS, A-SCS, and A-CO2 System. The greatest values were found in the treatment with closer plant spacing (T5x4), while the smallest values were found in the T7x4 and T8x4 treatments (Fig. 5). As shown before, dry biomass availability values were smaller in the denser treatments (T5x4 and T6x4) due to low light availability. In contrast, aboveground macauba palm biomass (A-MDB), expressed as t ha−1, decreased as planting space widened, but this was certainly due to the lower tree density (macauba plants ha−1) rather than the areal reduction in the dry matter accumulation capacity of the palm.

4 Discussion

Precipitation and air temperature are fundamental factors in local climatic conditions. In Araponga, Minas Gerais, during our study, the seasons followed historical patterns, with marked seasonality: on average, the rainy seasons, which correspond to summer, presented precipitation and temperatures 4.6 and 1.2 times greater than the dry seasons or winter. However, the natural barrier provided by any tree element in an intercropping system can change the understory climatic characteristics (Nicodemo et al. 2004), affecting plant growth in this stratum.

Temperature, relative humidity, and available understory light variations were detected between the two seasons and when comparing the two cultivation systems (intercropping system and pasture only). The diurnal temperature was similar between spacings T5x4, T6x4, and T7x4 and small when compared to that of the Tpasture only area. The difference in the relative humidity between the intercropping and monoculture systems also increased as palm spacing decreased. According to Nicodemo et al. (2004), trees in silvopastoral systems reduce wind speed, keeping atmospheric humidity and temperature more stable and benefiting pasture production. Therefore, the macauba palms in the intercropping system act as buffer agents, maintaining steady environmental conditions in the understory throughout the seasons, particularly during the dry season. Moderate temperatures and stable relative humidity throughout the seasons contribute to lower levels of abiotic stress, providing better conditions for their biological activities and for soil microorganisms (Schulz and Schirmer 2013).

Among the microclimatic parameters in the intercropping understory, luminosity undoubtedly showed the greatest variation, both in relation to the seasons and to the macauba palm planting distance. The low available understory light values recorded during the dry/winter season can be attributed to the earth’s axial tilt, because sunlight incidence on the soil surface is oblique in the Southern Hemisphere during the dry/winter season, unlike during the rainy season, when it is predominantly perpendicular (Langhi 2011). This seasonal change influences intercrop shading patterns, specifically during the first hours of the day (6–9 am) and late afternoon (4–6 pm).

Although the macauba palm has pinnate-plumose leaves and a noncompact semispherical crown, favorable for light transmission through the canopy and into the understory, available understory light variation (quantity and timing) among the four intercrop treatments was high (from 14.4 to 94.3%). This occurred because in the T5x4 and T6x4 treatments, with shorter spacing, the macauba palm crowns overlapped due to higher tree densities, effectively intercepting sunlight; in the treatments with wider spacing (T7x4 and T8x4), the palms acted more as mitigating agents to excess sunlight rather than fully impeding its transmission.

In macauba palm–Brachiaria grass intercropping system, variations in forage performance probably occurred in response to the shading level in the understory, as observed in coconut-pasture plantations (Wilson and Ludlow 1991), and not in response to temperature or atmospheric humidity differences. Shade intensity in the understory is described as the main factor influencing the morphology, physiology, growth, and persistence of forage species in silvopastoral systems, especially grasses with C4 metabolism (Nworji 2020).

The aboveground forage dry biomass (A-FDB) was the smallest in the intercrop treatments with smaller spacing (T5x4 and T6x4), whereas in the T7x4 and T8x4 treatments, the values were equally greater or equivalent to those in the Tpasture only area during the rainy season (Fig. 3). The reduction in the A-FDB values can be associated with a larger light interception by the palm canopy in the treatments with smaller spacing, influencing forage growth. Chong et al. (1991) reported a reduction in forage yield of approximately 5.5 t ha−1 year−1 when the light availability under rubber trees dropped more than 50%. The performance and aboveground dry matter production of B. decumbens in the macauba palm–Brachiaria grass intercropping system followed the same tendency described by Smith and Whiteman (1983) for the same species growing under increasing coconut shading (from 100 to 20% light transmission), i.e., light availability and grass growth were closely related. The authors concluded that B. decumbens was one of the most productive and suitable grass species for coconut–pasture systems when the percentage of light transmission was 60–50%, a statement backed up by Wong (1991). Indeed, in our study, the intercrop forage produced a similar yield under available understory light (USL) >75% (T7x4 and T8x4) as when it was grown under direct sunlight (Tpasture only) (Table 3).

A closed tree canopy in an intercropping system, such as in the T5x4 and T6x4 treatments, severely reduces the photosynthetically active radiation (PAR) available to understory species (Table 3), inhibiting their carbohydrate production (Sophanodora and Tudsri 1991). According to Smith and Whiteman (1983), B. decumbens should not be used as fodder at sites with moderate to intense shading (< 50% of light), since this grass would perish under these circumstances if grazing was added into the system. Because the forage species accumulates less dry matter and presents an unfavorable leaf/stem ratio under heavy shading, as occurred with the forage in the T5x4 and T6X4 treatments, livestock would be forced to graze more frequently, having negative effects on both the pasture and animal weight gain.

Regarding the seasons, the aboveground forage available in the intercrop treatments during the rainy season was at least 0.38 t ha−1 (in T5x4) greater than that in the dry season; spacings T7x4, T8x4, and the Tpasture only area displayed (P>0.05) the highest A-FDB values (t ha−1). In general, the conversion of solar energy into biomass in a plant community depends on photosynthetic rates, which correlate with light and water availability (Wu et al. 2019). These resources were most readily available during the summer rainy season, alongside mineral nutrients artificially supplied to all the areas.

Unlike the forage, the dry mass, carbon assimilation, and CO2 fixation of the tree component varied according to the number of plants per hectare. In this sense, the greater the number of macauba plants was, the higher the values of A-MDB, A-MSC, and A-MCO2 Assim. However, the proximity of the macauba plants negatively influences forage productivity, as previously described. Our results suggest that the spacing in the T7x4 and T8x4 treatments allows for adequate growth of the forage species and the maximum number of macauba plants.

According to Soares et al. (2009), the lower light incidence during the dry season can cause morphological changes in the B. decumbens canopy and cause the leaf/stem ratio to increase as a compensation mechanism. On the other hand, intense shading exacerbates etiolation in grasses, and the leaves become larger and thinner in their effort to better capture sunlight (Araújo et al. 2020). The leaf/stem ratio is considered an indirect indicator of forage quality (Moore et al. 2020). As a result, pasture density and digestibility decrease, prompting livestock to seek and forage in open, sunbathed areas. Therefore, the treatments with wider spacing (T7x4 and T8x4) were more suitable for grazing in the macauba palm–Brachiaria grass intercropping system.

The similarity of the leaf/stem ratio in all the treatments during the rainy season was probably due to the beneficial effects of higher temperatures as well as improved light and water availability, typical summer traits, and prerequisites for optimal C4 grass performance, which may have favored the higher growth of B. decumbens. During the dry season, the leaf/stem ratio was approximately 16% higher than that in the rainy season; hence, there was greater emission and/or elongation of the leaf blade and less tillering.

The monocotyledonous macauba palm, a nonwoody species, accumulated approximately 25.4 t ha−1 year−1 of aboveground biomass, a reasonable value compared to the standard aboveground biomass accumulation for dicotyledonous tree species (Santana et al. 2008). This yield trait was also found in the same experimental area when the palm trees were only 4 years old. Our results confirmed that arborous palm species are efficient dry matter accumulators and, consequently, are as suitable as true woody species for fixing carbon (Montagnini and Nair 2004). Therefore, in the present study, the higher planting density in the intercropping system did not interfere with the growth or ability of macauba palm to accumulate dry matter per tree, contrary to what was found for timber tree species (Wilson et al. 1990).

The carbon stock and the CO2 assimilation of the system quantified in this study are directly related to aboveground dry matter accumulation. The aboveground macauba dry biomass (A-MDB) of the T7x4 and T8x4 treatments was 26% and 35%, respectively, smaller than that of the spacing T5x4 (Fig. 4b); however, the estimated aboveground forage dry biomass (A-FDB) in the areas with wider spacing was similar to that of the Tpasture only area (Fig. 4a). These results are consistent with the fact that under moderate shade conditions, B. decumbens obtain similar productivity to pastures growing under full sun or in monoculture systems (Gobbi et al. 2009). In areas with wider spacing, the poorer intercropping performance in regard to total stored carbon, total ground biomass, and assimilated CO2 is probably related to lower macauba plant density, as mentioned previously. Therefore, macauba palm is the element of this intercropping system that effectively captures and stores atmospheric carbon. In contrast, the aboveground biomass of a eucalyptus–pasture intercropping system stored only 11.3 t ha−1 of carbon (Sharrow and Ismail 2004), an amount far below that observed in the T8x4 treatment in the present study.

An agroforestry system is indeed a great carbon sequestration complex, where both agricultural and livestock activities can be carried out simultaneously. Based on the results, for the first time, the success of the macauba palm–Brachiaria grass intercropping system in supporting agronomical and environmental services was confirmed. The adopted plant spacing must allow the best growth and carbon assimilation performance for both the palm and the forage species. In the present study, it was shown that planting distances equal to or greater than 7 m × 4 m provide adequate climatic conditions to benefit the growth of C4 forage plants, especially during the rainy season.

We are aware that our experimental area can be considered small compared to other studies on agroforestry systems. Future investigations in larger areas would be crucial to further solidify the information described here and to cover other important aspects of these agroecosystems such as livestock, plant physiology, and nutrition as well as other alternatives for spatial arrangements. Regardless of our limitations, it is worth noting that the literature available on macauba palm is still incipient, which makes every piece of information on this species very valuable for the scientific community and society. Collectively, the opportunity to evaluate intercropped areas planted with A. aculeata and B. decumbens with more than 10 years of age and arranged in different spatial arrangements is rare, which makes this work remarkably unique.

5 Conclusion

The intercropped Acrocomia aculeata–Brachiaria decumbens system can be considered a viable alternative for feeding cattle under low or moderate grazing pressure, but its success will depend on the density of trees. The presence of macauba in the intercropped system influenced the growth and development of the forage; this effect was significant at different spacings but not between the evaluated seasons (dry and rainy seasons). The distance between the macauba plants affected the microclimate conditions of the understory, reflecting the growth and development of the forage component. In the denser areas, there was a significant reduction in the passage of light through the canopy of the macauba plants, leading to different leaf/stem ratios between the treatments with different spacing and the dry and rainy seasons. At lower densities (i.e., 357 and 312 macauba plants ha−1), the values of forage dry biomass were similar to those ​​obtained in the pasture only areas, validating the system as a viable alternative in terms of land use change.