Aust J Crop Sci. 19(03):274-283 (2025) | ISSN:1835-2707

https://doi.org/10.21475/ajcs.25.19.03.p281

Temporal changes in basic cations in soils under cocoa cultivation in the Brazilian Atlantic forest

Ádson dos Santos Oliveira¹, José Olímpio de Souza Júnior¹, Júlio César lima Neves², Dário Ahnert¹

¹Universidade Estadual de Santa Cruz, Campus Soane Nazaré de Andrade Ilhéus, Bahia, Brasil, ²Universidade Federal de Viçosa, Campus Universitário, Viçosa, Minas Gerais, Brasil

*Corresponding author: Adson dos Santos Oliveira

Abstract: Low base saturation causes abiotic stress in crops. However, it can be a significant factor for increasing yields when adjusted to appropriate levels. This study aimed to assess the temporal alterations of basic cations in two soil layers under cocoa tree cultivation in the Brazilian Atlantic Forest biome. For this purpose, a seven-year field experiment was established, consisting of 19 experimental areas (EAs), each of which received superficial applications of lime and agricultural gypsum. The changes on basic cations levels were measured at doses of 0.36-1.09 and 0.29-2.69 t ha^-1 limestone and gypsum, respectively, after 0, 40, 52, 72, and 87 months of application. In the first two years, the cocoa trees were fertilized with the N-P-K 16-16-16 formula in doses of 640 and 1,025 kg, respectively, divided into four applications. Nutrition was done with monoammonium phosphate, urea, triple superphosphate and potassium chloride as source. The results suggests that in most EAs, lime did not affect Ca²⁺ and Mg²⁺ levels, consequently, causing no difference on the base sum and base saturation in the 0–20 cm layer. However, gypsum led to increase of Ca²⁺ levels in both soil layers, as well as an increase of Mg²⁺ and K⁺ levels in the 20–40 cm layer. Variations in K⁺ levels were associated with potassium fertilization. Cocoa producers should apply gypsum to enhance the chemical properties of poor soils by increasing exchangeable base levels at greater depths. This improvement promotes deeper root growth, enabling cocoa trees to access a larger soil volume and better withstand water stress during drought periods.

Keywords: Gypsum mobility, soil base, superficial liming.

Abbreviations: EAs_Experimental Areas

Introduction

The cocoa tree (Theobroma cacao L.) has been used since 1100 BC. It is native to South America which then introduced to Central America, Africa, and Asia, (Motamayor, 2002; Henderson et al., 2007). Currently, cocoa is an agricultural commodity that generates an annual revenue over than US$120 billion (FAOSTAT, 2024). The species grows only in tropical regions, engaging six million farmers for its cultivation (Olwig et al., 2024), primarily in highly weathered soils (Fageria and Baligar, 2008).

Soil acidity is the main constraint to plant production in tropical regions (Hue, 2022; Anbessa, 2023). In Africa the soil acidity has been increasing regularly, affecting large agricultural regions of Ivory Coast and Ghana (Agegnehu et al., 2021) and other main cocoa producers worldwide (ICCO, 2024). Acidic soils have low pH, high Al³⁺ levels, and low levels of basic cations such as Ca²⁺, Mg²⁺, and K⁺ (Rheinheimer et al., 2018; Basak et al., 2022). Higher Al³⁺ levels promote disruption of leaf cell nuclear membranes and deformity of root epidermis cells (de Almeida et al., 2015). The low base saturation observed in acidic soils is a major cause of abiotic stress in cultivated plants (Andrade et al., 2021). Low basic cations levels associated with the Al³⁺ toxicity reduces water and nutrient uptake rates (Shetty and Prakash, 2020; Bolan et al., 2023), exposing the crops to abiotic stress mainly during drought periods (Zandoná, 2015.

Soil superficial acidity can be mitigated by lime, which is also a source of Ca²⁺ and Mg²⁺. For sub superficial acidity, the improvement of chemical environment has been achieved by agricultural gypsum, increasing Ca²⁺ level and reducing Al³⁺ saturation, which consequently enhances productivity, especially during drought periods (Bossolani et al., 2022a, b; Caires et al., 2023). Some studies have shown that gypsum reduces Cd²⁺ bioavailability in soils with high CEC, probably due to ions Ca²⁺ from gypsum to have a counteracting effect on cadmium Cd²⁺ mobilization and ion competition at the root surface (Argüello et al., 2022). The Cd²⁺ high level is harmful for human health. It also poses a challenge to the economies of cocoa-producing countries by reducing their competitiveness in the international market

Table 1. Initial values of pH and chemical attributes^1/ used to calculate the lime and gypsum amendment of two soil layers and the respective lime and gypsum doses applied in 19 experimental areas (EAs), in the state of Bahia, Brazil.

1/ pH in H₂0; T (cation-exchange capacity (CEC) at pH 70); V (base saturation); Al³⁺ (exchangeable aluminum); and t (effective CEC) 2/ humid to subhumid.

(Maddela et al., 2020; Santander Ruiz et al., 2021, Reyes-Pérez, et al., 2023).

Cocoa trees experience reduced nutrient uptake, translocation, and accumulation, when grown in acidic soils (Ribeiro et al, 2013; Rosas-Patiño et al., 2017). The species has higher nutritional requirements compared to perennial tropical crops such as coffee and rubber trees (Baligar and Fageria, 2005; Souza Júnior et al., 2018). Cocoa high productivity is related to high levels of foliar nutrients, which are seriously affected by acidity, leading to loss of productivity (Marrocos et al., 2020).

To meet the cocoa nutritional demands and avoid soil depletion, it is necessary to increase productivity in existing areas instead of deforesting to establish new plantations. The Brazilian Atlantic Forest biome is home to 500 thousand hectares of cocoa (Chiapetti, 2018), cultivated in an agroforestry system called "cabruca", which consisted of planting cocoa trees under native trees presenting rich biodiversity (Sambuichi et al., 2012).

These plantations have great potential to produce sustainable cocoa beans, but soil fertility management remains a challenge to overcome. Seeking to develop strategies to manage the soil fertility of local plantations, this study aimed to evaluate the temporal alterations on chemical attributes related to exchangeable basic cations (Ca²⁺, Mg²⁺, K⁺, base sum, and saturation) in two soil layers (0–20 cm and 20–40 cm). This study was conducted in the Brazilian Atlantic Forest across 19 experimental areas (EAs) cultivated with cocoa trees, after superficial application of lime and agricultural gypsum.

Results

Temporal alterations for Ca²⁺

There was no substantial alteration in Ca²⁺ levels over time across eight EAs at 0–20 cm layer (Table 2). Three of experimental areas (EAs) received lime ranging from 0.54 to 0.98 t ha^-1, two received lime + gypsum from 0.46 to 1.09 t ha^-1 and from 0.60 to 1.06 t ha^-1, respectively, while one received 0.73 t ha^-1 gypsum, and two did not receive any inputs (Table 1).

In 0–20 cm layer, 10 EAs exhibited significant alteration in Ca²⁺ levels over time. All of them had the same behavior, exhibiting a quadratic regression with a positive linear coefficient and a negative quadratic coefficient. (Table 2). Among these EAs, three received limes at rates between 0.36 and 0.54 t ha^-1, five received both lime and gypsum at rates of 0.62 to 1.04 t ha^-1 and 0.29 to 2.39 t ha^-1, respectively, and two did not receive the inputs (Table 1). The Ca²⁺ levels decreased over time only in EA 19. For the 20–40-cm layer, there was no considerable alteration in Ca²⁺ levels over time at 12 EAs. Six EAs exhibited an increase, followed by a decrease in Ca²⁺ levels over time. All received gypsum ranging from 0.29 to 2.39 t ha^-1. The Ca²⁺ levels showed a linear reduction only in EA 9 did over time.

Temporal alterations for Mg²⁺

For the 0–20-cm layer, no significant variations were observed in the Mg²⁺ levels in 10 EAs (Table 3). Among these, five received lime ranging from 0.36 to 0.98 t ha^-1, three received lime plus gypsum from 0.46 to 1.04 t ha^-1 and from 0.60 to 2.39t ha^-1,m respectively, one EA received 0.73 t ha^-1 gypsum, and one EA did not receive any inputs (Table 1).

Table 2. Regression equations for Ca²⁺ levels over time^1/ in two layers from soils cultivated with cocoa trees across 19 experimental areas (EAs) in Bahia, Brazil.

In this layer, five EAs exhibited significant variations in Mg²⁺ levels over time. All of them exhibited a positive linear coefficient and a negative quadratic coefficient, indicating an initial increase followed by a subsequent decrease in Mg²⁺ levels over time (Table 3). Among these five EAs, two received lime (0.62 and 1.09 t ha^-1) plus gypsum (1.06 and 1.96 t ha^-1), one EA received 0.54 t ha^-1 lime, and two EAs received none of the inputs. Finally, in four EAs, the Mg²⁺ levels decreased linearly over time, even in three EAs that received lime doses ranging from 0.70 to 1.04 ha^-1 (Table 1).

For the 20–40-cm layer, there was greater consistency in the Mg²⁺ levels over time compared to the superficial layer, as in 13 EAs, a significant variation in the Mg²⁺ levels were observed over the 87 months of evaluation (Table 3). Six of these EAs exhibited a quadratic model fit. It is essential to note that all six EAs received gypsum, with doses ranging from 0.29 to 2.69 t ha^-1 (Table 1). There was a linear reduction in Mg²⁺ levels over time only in EA 13.

Temporal alterations for K⁺

The K⁺ levels in the 0–20-cm layer did not vary significantly across nine EAs. For five EAs, a quadratic behavior was observed, indicating an initial increase followed by a later decrease in K⁺ levels over time. In four EAs, a positive linear behavior was observed. Finally, in just one EA, the K⁺ levels decreased linearly over time (Table 4).

In the 20–40 cm soil layer, K⁺ levels mostly presented no significant change over time. There was positive linear

behavior only in three EAs and only one EA exhibited a quadratic behavior for the K⁺ level over time.

Temporal alterations for sum of base (SB)

The temporal behavior SB was similar to those observed for Ca²⁺ and Mg²⁺ (Tables 2 and 3). Thus, there was a prevalence of no variation in SB over time, especially in the subsurface layer. In contrast, the predominant significant model was quadratic with a positive linear coefficient and a negative quadratic coefficient, when there was variation.

V% is the percentage of the CEC at pH 7.0 (T) that is occupied by basic cations (SB), with the remaining portion occupied by potential acidity (H+Al³⁺). In other words, V% is highly dependent on SB, but also on potential acidity. Therefore, the behavior of both variables is similar, with the 20–40-cm layer exhibiting identical behavior across all 19 EAs, 14 of which exhibited no variation over time and five of which exhibited significant quadratic models (Tables 5 and 6), with these five EAs having received gypsum (Table 1).

Discussion

Alterations in Ca²⁺ and Mg²⁺ levels

Out of the 15 EAs that received lime and/or gypsum, eight exhibited significant changes in Ca²⁺ levels within the 0–20-cm layer throughout the study, exhibiting quadratic models. This indicates the importance of these two inputs as sources of Ca²⁺ with temporally limited effects, as demonstrated by the decrease in Ca²⁺ levels after the inflection point of the quadratic models, which varied

Table 3. Regression equations for Mg²⁺ levels over time^1/ in two layers from soils cultivated with cocoa trees across 19 experimental areas (EAs) in Bahia, Brazil.

between 32 to 53 months. In the 0–20-cm layer, the increase in Ca²⁺ levels, has been documented particularly in no-till farming (NTF) due to surface liming, where there is no soil disturbance (Caires et al., 2015; Crusciol et al., 2016; Fontoura et al., 2019; Crusciol et al., 2022). In NTF, the influence of plant residues on the dynamics of Ca²⁺ has also been observed through the formation of organometallic complexes, which can enhance the mobility of the nutrient (Miyazawa et al., 2002; Pavinato et al., 2017; Rheinheimer et al., 2018).

Several arguments could be presented that may act independently or in conjunction for the The lack of temporal variations in EAs that exhibited no in Ca²⁺ levels despite receiving lime and/or gypsum such as small doses of these inputs (Auler et al., 2019; Li et al., 2019) could be attributed to the following circumstances: concentration of Ca²⁺ from lime in the more superficial layers, thereby diluting the sample from the 0–20-cm layer (Hume et al., 2023); the loss of Ca²⁺ from gypsum through leaching to layers at depths greater than 0–20 cm (Araújo et al., 2016; Rusyanto et al., 2017) dynamic equilibrium in biogeochemical cycling due to natural nutrient inputs (mineralization from the large influx of organic matter from the removal of old cocoa trees and pruning of branches from trees in the cocoa-cabruca agroforestry system which, in some EAs, led to excessive shading); nutrient extraction by the crop (formation of new cocoa trees over the years, immobilizing nutrients in living plant biomass) and export through harvests (some EAs have exhibited yields

exceeding 3,000 kg ha^-1 year^-1 of dried almonds in recent years) (Snoeck et al., 2016; Van Vliet and Giller, 2017).

In the 20–40-cm layer, the dynamics of Ca²⁺ appear to be linked to the application of gypsum on the surface, as all six EAs that exhibited a quadratic behavior (an increase followed by a decrease in Ca²⁺ levels over time) received gypsum. The application of gypsum has proven effective in increasing subsurface Ca²⁺ levels due to the formation of an ionic pair with SO₄^2- (Ramos et al., 2013). In contrast, Butterly et al. (2021) suggested that the presence of plant-derived organic ligands can influence the mobility of Ca²⁺ in the soil. Therefore, this possibility cannot be dismissed in this study, considering the high input of plant material in the EA resulting from the renewal of old cocoa crops. It should be noted that among the EAs that received gypsum, only in two of them (EA 5 and 8) there were no significant changes in Ca²⁺ levels over time in the 20–40-cm layer. This is likely due to the relatively low doses of gypsum used in these EAs, which were 0.73 and 0.60 t ha^-1, respectively (Auler et al., 2019). In this layer, for all EAs that did not receive gypsum, there was no significant variation in Ca²⁺ levels over time, indicating stability over a time frame of 87 months, except for EA 9, which exhibited a linear reduction in Ca²⁺ over time (Table 3), likely because it had the lowest surface Ca²⁺ level, which would allow plants to absorb Ca²⁺ from deeper subsoil layers.

In the 0–20 cm layer, among the 14 EAs that received lime, only three exhibited a quadratic trend of Mg²⁺ levels, with peak achieved between 29 to 40 months. In eight EAs, the Mg²⁺ level was not significantly affected over the months, indicating a balance between the nutrient

Table 4. Regression equations for K⁺ levels over time¹ in two layers from soils cultivated with cocoa trees across 19 experimental areas (EAs) in Bahia, Brazil.

supplied by the lime and its absorption by plants. In three EAs, the Mg²⁺ level decreased over time, suggesting that plant absorption exceeded the supply from the lime. Particularly, for EA 3, the decrease in Mg²⁺ levels in the surface layer may also have been affected by a higher dose of gypsum, which could have contributed to the leaching of Mg²⁺ to subsurface layers (Alves et al., 2021; Tiecher et al., 2019; Vicensi et al., 2020). The Mg²⁺ content may result from the mineralization of organic matter from the removal of old cocoa trees and the pruning of shade tree branches. It would help to explain the initial increases in Mg²⁺ levels in the three EAs that received lime (Agbotui et al., 2024; Asigbaase et al., 2021), as well as in EA 14 and 16, which did not receive lime. In these latter EAs, the peak Mg²⁺ occurred at around 34 months. The lack of significant variation in Mg²⁺ was also observed in two of the five EAs that did not receive lime.

In the 20–40-cm layer, the quadratic behavior of Mg²⁺ levels were observed only in the EA that received gypsum. This indicates that the gypsum applied to the surface promoted the leaching of Mg²⁺ to the subsurface layer, with peak concentrations that were achieved between 29 and 47 months. Changes in subsurface levels of Mg²⁺ are reported in studies involving the surface application of gypsum, due to the formation of an ionic pair with SO₄^2- which imparts mobility to the cation in the soil profile (Bossolani et al., 2020; Vicensi et al., 2020). This growth in subsurface Mg²⁺ levels allowed the expansion of the soil volume by the root system, leading to improved water and nutrient uptake (Bossolani et al., 2022). For cocoa cultivation, this fact is highly significant, as the species is considered demanding in Mg²⁺ (Souza Júnior et al., 2018).

In the 20–40-cm layer, among 10 EAs that did not receive gypsum and two that did, there was no significant change in Mg²⁺ levels over time. Only in EA 13 the Mg²⁺ level decrease linearly over time, possibly because this EA had a low surface level of Mg²⁺, requiring plants to absorb Mg²⁺ from deeper subsurface layers.

Variation in K⁺ levels

Although lime and gypsum are not sources of K⁺, these inputs can affect the dynamics of this nutrient in soils, either by increasing the CEC of the soil through lime, which enhances K⁺ retention, or by enhancing K⁺ leaching, facilitated by the addition of Ca²⁺ and Mg²⁺ from lime and, particularly, by the addition of Ca²⁺ and SO₄^2- from gypsum (Fageria and Knupp, 2014). It should be noted that the EAs were fertilized with K⁺, with the doses determined based on the planting age, expected productivity, and the most recent soil analyses for each EA, following the criteria suggested by Souza Júnior et al. (2018). The four factors that most profoundly influenced the temporal variations of K⁺ were: specific potassium fertilizations for each EA, natural K⁺ input through the mineralization of organic matter from the thinning of old cocoa trees and pruning of shade tree branches (Asigbaase et al., 2021), K⁺ extraction for the formation of new cocoa trees, and K⁺ export through harvests (Kone et al., 2020). It is noteworthy that K⁺ is the mineral nutrient that accumulates the most in the plant biomass of the cocoa tree, as well as in the cocoa itself (Anokye et al., 2021; Souza Júnior et al., 2018).

Table 5. Regression equations for the sum of bases (SB) over time^1/ in two layers from soils cultivated with cocoa trees across 19 experimental areas (EAs) in Bahia, Brazil.

K⁺ in the 0–20-cm layer exhibited distinct behaviors over time among the 19 studied EAs due to the presented factors. In nine EAs, there was no significant variation in K⁺ levels, indicating a balance between the intake and output of the nutrient (Hafizal et al., 2023). In five EAs, quadratic behavior was observed, which can be comprehended by the initial increase in K⁺ due to the addition of K⁺ through fertilization and the mineralization of biomass from the crop renewal process (Agbotui et al., 2024). This was followed by a subsequent decrease in K⁺ levels caused by the extraction and export of K. It is noteworthy that in many EAs, productivity in recent years has surpassed 3,000 kg ha^-1 year^-1, with some reaching close to 5,000 kg ha^-1 of dried cocoa beans. In four EAs, a positive linear trend was observed, with coefficients ranging from +0.568 to +1.439 (Table 4). These coefficients correspond to annual increases of 6.8 to 17.3 mg dm^-3 of K⁺. These EAs initially had the lowest soil K⁺ levels and consequently received the highest annual K₂O doses, which in recent years ranged from 280 to 320 kg ha^-1 year^-1. In contrast, there was a negative linear trend only in EA 10 with a coefficient of -1.425 (Table 4). This indicates an annual decrease of 17.1 mg dm^-3 of K⁺. This EA received lower K₂O applications of just 40 kg ha^-1 year^-1 in recent years due to initially very high soil K⁺ levels in the early years of the experiment.

In the 20–40 cm layer, K⁺ levels remained highly stable over time, as 15 EAs showed no significant fluctuations throughout the 87-month study. Despite this stability, average K⁺ levels varied widely among these EAs, ranging from 8.9 to 83.9 mg dm⁻³ (Table 4). In three EAs, a positive linear trend was observed, with regression model coefficients indicating estimated annual increases in K⁺ ranging from 7.0 to 9.8 mg dm^-3. It is essential to note that these three EAs received the highest gypsum doses, varying from 1.96 to 2.69 t ha^-1 (Table 1), suggesting that gypsum may have enabled the translocation of K⁺ from the surface to the subsurface layer (Moraes et al., 2016). Only in one EA, there was a quadratic behavior observed for the K⁺ level over time, after the application of 1.06 ha^-1 of gypsum.

Variation in sum of bases and bases saturation

The sum of bases is mainly dependent on the Ca²⁺ and Mg²⁺ levels, which represented more than 95% of variation in both layers (de Almeida et al., 2018). Hence, sum of bases tends to be an intermediate behavior between these two nutrients. This is similar to what was observed for Ca²⁺ and Mg²⁺, in which there was predominantly no change in sum of bases over time, either due to the absence of lime and/or gypsum application or because the doses were small (Fageria and Knupp, 2014).

However, in the 0–20-cm and 20–40-cm layers there were differences in the sum of bases in nine and six EAs due to the previously discussed reasons for Ca²⁺ and Mg²⁺. These variations primarily exhibited quadratic models, indicating that the estimated maximum sum of bases occurred between 28 to 41 months for the 0–20-cm layer and between 34 to 48 months for the 20–40-cm layer, following the application of inputs. It should be noted that for the subsurface layer, all six EAs that exhibited variations in sum of bases values received gypsum which causes cation mobility in soil profile (Rampim and Lana, 2015; Tiecher et al., 2019).

Table 6. Regression equations for base saturation (V) over time^1/ across two layers from soils cultivated with cocoa trees in 19 experimental areas (EAs) in the state of Bahia, Brazil.

In turn, base saturation is a variable dependent on the sum of bases, as it also accounts for acidic cations (H + Al³⁺), with its value expressed as a percentage. This helps in understanding why, in some cases at 0–20-cm layer, there was no correlation between the behaviors of sum of bases and base saturation. In turn, in the 20–40-cm layer, the behavior of both variables was the same across all 19 studied EAs, with 14 EAs exhibiting no significant variation over time in sum of bases and base saturation, and five EAs exhibiting significant quadratic models; the latter received gypsum.

Materials and methods

Data collection

Nineteen experimental areas (EAs) located in 15 municipalities across the humid and humid-to-sub-humid climate zones of southeastern Bahia, Brazil (SEI-BA, 2022), were sampled with 20 individual samples collected per composite sample (Figure 1). The initial analyses of the 0–20 cm and 20–40 cm layers were used to calculate the lime and gypsum rates (Table 1). The assessment time was 87 months, starting from month zero (T0), which was when lime and gypsum were applied, followed by subsequent sampling at the 40, 52, 72, and 87 months after inputs application. During these four periods, samples were collected from the 0–10 cm, 10–20cm, and 20–4 -cm layers, with the values for the 0–20-cm layer derived by averaging those from the 0–10 cm and 10–20 cm layers. In T40, two composite samples were collected by EA; in T52, T72, and T87, six composite samples were collected by EA.

Soil attributes analyzed

The following attributes were analyzed: pH in H₂O; Al³⁺, Ca²⁺, and Mg²⁺ extracted using 1.0 mol L^-1 KCl; K⁺ extracted using Mehlich^-1; potential acidity (H+Al) extracted using 0.5 mol L^-1 Ca (CH₃COO)₂; organic matter determined using the Walkley-Black method; and remaining phosphorus (P-rem) obtained by shaking 60 mg L^-1 of P in a 0.01 mol L^-1 CaCl₂ solution (Teixeira et al., 2017). The sum of bases (SB) and base saturation (V%) was calculated according to basic cations levels. The sand, silt, and clay levels were determined following the methods described by Teixeira et al. (2017).

Lime and gypsum rates calculation

The lime rates were determined using the base saturation method – V% (Equation 1). The goal was to increase the V% to 75% in EA as it was below 65% in the 0–20-cm soil layer. For this purpose, we considered a reaction depth (p) of 5 cm and an RNV of the amendment of 90%.

\[LR = \frac{\left( V_{2} - V_{1} \right)\ \times \ T}{10\ \times RNV}\ \times \ \frac{p\ }{20}\ \ \ \ (Equation\ 1)\]

In which:

LR = lime rate, in tons per hectare (t ha^-1);

V₂ = desired base saturation, 75%;

V₁ = base saturation of the 0–20 cm layer, in %;

T = cation exchange capacity (CEC) at pH 7.0, in mmol_c dm^-3;

RNV = Relative Neutralizing Value, 90%;

p = lime reaction depth, 5 cm.

The gypsum doses were calculated based on the stoichiometric equation proposed by Souza Júnior et al. (2018) (Equation 2). The goal was to reduce the Al³⁺ saturation (m%) to 20% in the EA, where it exceeded 20%. The layer thickness (LT) was considered to be 30 cm and the percentage of Ca²⁺ in the gypsum (%Ca_gyp) was 16%.

\[GR = \left\lbrack {Al}^{3 +}\ - \left( \frac{m_{t} \times t}{100} \right) \right\rbrack\ \times \ \frac{0.2\ \times LT}{{\% Ca}_{gyp}}\ \ \ \ (Equation\ 2)\]

In which:

GR = gypsum rate, in tons per hectare (t ^ha-1);

Al³⁺ = AI³⁺ level in the 20–40 cm layer, in mmol_c dm^-3;

m_t = saturation of Al³⁺ tolerable by the cocoa tree, 20%;

t = effective CEC in the 20–40-cm layer, in mmol_c ^dm-3;

LT = layer thickness, 30 cm;

%Ca_gyp = percentage of Ca in the gypsum, 16%.

Lime and gypsum application method

The lime and gypsum doses (Table 1) were converted into grams per plant and manually applied by broadcasting, covering the entire soil surface. Lime and gypsum were not mixed into the soil. However, the organic mulch was tilled using a leaf blower to speed up the contact of the inputs with the soil surface.

Fertilization program for the EAs

The initial fertilization process involved applying 500 mg dm^-3 of P (120 grams of single superphosphate), 130 mg dm^-3 of K (5 g of potassium chloride), 40 g of dolomitic lime as a source of Ca and Mg, and 2 dm³ of poultry litter.

For the formative fertilization in the first and second years, 500 and 800 g of the NPK 16-16-16 fertilizer were applied to the soil surface, respectively, divided into four annual installments for all EAs. The canopy projection of the cacao trees was used as a reference for the application.

Production fertilization commenced in the third year and was tailored to each EA based on the respective soil analyses. For N, P, and K, the doses were based on the propositions by Souza Júnior et al. (2018). The doses were divided into four annual installments, using urea, monoammonium phosphate or triple superphosphate, and potassium chloride as sources.

Micronutrients were applied in a single annual dose based on the respective soil analyses of each EA, when needed. Doses used were: 1.0 kg ha^-1 of B (for areas with a B content in the soil less than 0.7 mg dm^-3); 2.0 kg ha^-1 of Cu (for areas with a Cu content in the soil less than 0.7 mg dm^-3) and 2.5 or 5.0 kg ha^-1 of Zn (for areas with a Zn content in the soil between 3.5 and 5.0 mg dm^-3 or less than 3.5 mg dm^-3 of Zn, respectively). The sources used were boric acid and copper and zinc sulfates.

Statistical analysis

The results of the soil chemical attributes were submitted to analysis of variance (ANOVA), performed for each experimental area, given the non-homogeneity of the residual variance between the areas, evaluated based on the quotient of the largest by the smallest mean square of the residue being greater than 7.0, as recommended by Gomes and Garcia (2002) and Pimentel Gomes (2009), for agronomic experiments. This limit is a relaxation of Box's original proposal in relation to the value 4.0. Subsequently, a regression analysis of the soil chemical attributes was performed as a function of the time elapsed after the application of the inputs. The models tested were simple linear and quadratic, and the model that presented significant coefficients at up to 1%, 5% and 10% probability was accepted by the F test, using the mean square of the residue of each ANOVA, and the highest value of the coefficient of determination (R2).

Conclusions

In most EAs, the surface application of lime without soil mixing did not change the levels of Ca²⁺ and Mg²⁺ and, consequently, the base sum and saturation in the 0–20-cm soil layer. Nevertheless, in most EAs, the surface application of gypsum without soil mixing increased the Ca²⁺ levels in both soil layers. Additionally, it led to an increase in the Mg²⁺ and K⁺ levels in the 20–40-cm layer. Variations in K⁺ levels were primarily associated with potassium fertilization.

Acknowledgments

The authors thank the Brazilian cocoa farmers which farms were available for this study, Mondelez International, Universidade Estadual de Santa Cruz (UESC) and Centro de Inovação do Cacau (CIC), under the agreement RD-15540 (financial support).

Statement of contributions

All the authors contributed to the conception and design of the study. Conceptualization, statistical analysis, methodology, and writing original draft were performed by AO. Conceptualization, data acuration, methodology, supervisor, writing review were performed by JS and JN. JS was responsible for writing the review. DH was responsible for founding acquisition and writing review. All authors commented on the previous versions of the manuscript. All authors have read and approved the final manuscript.

References

Anokye E, Lowor ST, Dogbatse JA, Padi FK (2021) Potassium application positively modulates physiological responses of cocoa seedlings to drought stress. Agronomy. 11(3):563.

Agegnehu G, Amede T, Erkossa T, Yirga C, Henry C, Tyler R, Nosworthy MG, Beyene S, Sileshi GW (2021) Extent and management of acid soils for sustainable crop production system in the tropical agroecosystems: a review. Acta Agr Scand. 71:852-869.

Alves LA, Tiecher TL, Flores JPM, Filippi D, Gatiboni LC, Bayer C, Tiecher T (2021) Soil chemical properties and crop response to gypsum and limestone on a coarse-textured Ultisol under no-till in the Brazilian Pampa biome. Geoderma Reg. 25: e00372.

Agbotui DK, Ingold M, Buerkert A (2024) Carbon and nutrient cycling in cocoa agroforests under organic and conventional management. Nutr Cycl Agroecosys. 129:7–20.

Asigbaase M, Dawoe E, Sjogersten S, Lomax BH (2021) Decomposition and nutrient mineralisation of leaf litter in smallholder cocoa agroforests: a comparison of organic and conventional farms in Ghana. J Soil Sediment. 21:1010–1023

Anbessa, B (2023) Review on Potential of Liming for Sustainable Management Some Selected Soil Chemical

Properties and Crop Yield Improved in Tropical Soil. GRES. 1:46–64.

Andrade CA, Ramsdorf I, de Andrea A, Pereira L, Naudi C, Pereira J, de Mello R, Silva Junior C, Eduardo Teodoro P (2021). Physiological response and earliness of soybean genotypes to soil base saturation conditions. J Agron Crop Sci. 207:163–169.

Auler AC, Caires EF, Pires LF, Galetto SL, Romaniw J, Charnobay AC (2019) Lime effects in a no-tillage system on Inceptisols in Southern Brazil. Geoderma Reg. 16: e00206.

Araújo LG, Figueiredo CD, Sousa DD, Nunes RDS, Rein TA (2016) Influence of gypsum application on sugarcane yield and soil chemical properties in the Brazilian Cerrado. Aust J Crop Sci. 10(11):1557-1563.

Baligar VC, Fageria NK (2005) Soil Aluminum Effects on Growth and Nutrition of Cacao. Soil Sci Plant Nutr. 51:709–713.

Argüello D, Dekeyrel J, Chavez E, Smolders E (2022) Gypsum application lowers cadmium uptake in cacao in soils with high cation exchange capacity only: A soil chemical analysis. Eur J Soil Sci. 73(2): e13230.

Basak BB, Sarkar B, Saha A, Sarkar A, Mandal S, Biswas JK, Wang H, Bolan NS (2022) Revamping highly weathered soils in the tropics with biochar application: What we know and what is needed. Sci Total Environ. 822:153461.

Bolan N, Sarmah AK, Bordoloi S, Bolan S, Padhye LP, Van Zwieten L, Sooriyakumar P, Khan BA, Ahmad M, Solaiman ZM, Rinklebe J (2023) Soil acidification and the liming potential of biochar. Environ Pollut. 317:120632.

Bossolani JW, Crusciol CAC, Moretti LG, Garcia A, Portugal JR, Bernart L, Vilela RG, Caires EF, Amado TJC, Calonego JC, dos Reis AR (2022a) Improving soil fertility with lime and phosphogypsum enhances soybean yield and physiological characteristics. Agron Sustain Dev. 42:26.

Bossolani JW, Crusciol CAC, Momesso L, Portugal JR, Moretti LG, Garcia A, Fonseca M, Rodrigues VA, Calonego JC, dos Reis AR (2022b) Surface liming triggers improvements in subsoil fertility and root distribution to boost maize crop physiology, yield and revenue. Plant Soil. 477:319–341.

Bossolani JW, Crusciol CAC, Merloti LF, Moretti LG, Costa NR, Tsai SM, Kuramae EE (2020) Long-term lime and gypsum amendment increase nitrogen fixation and decrease nitrification and denitrification gene abundances in the rhizosphere and soil in a tropical no-till intercropping system. Geoderma. 375:114476.

Butterly CR, Costello B, Lauricella D, Sale P, Li G, Tang C (2021) Alkalinity movement down acid soil columns was faster when lime and plant residues were combined than when either was applied separately. Eur J Soil Sci. 72:313–325.

Caires EF, Haliski A, Bini AR, Scharr DA (2015) Surface liming and nitrogen fertilization for crop grain production under no-till management in Brazil. Eur J Agron. 66:41-53.

Caires EF, Duart VM, Rodrigues GLA, Bini AR (2023) Crop nutrition and grain yield as affected by phosphorus fertilization and continued use of phosphogypsum in an Oxisol under no-till management. Arch Agron Soil Sci. 69:2370–2385.

Chiapetti J, (2018) Produção de cacau na Bahia: análise da trajetória política e econômica. In: JO Souza Júnior (ed) Cacau: cultivo, pesquisa e inovação. Editus, Ilhéus.

Crusciol CA, Marques RR, Carmeis Filho AC, Soratto RP, Costa CH, Neto JF, de Castilhos AM (2016) Annual crop rotation of tropical pastures with no-till soil as affected by lime surface application. Eur J Agron. 80:88-104.

Crusciol CAC, Bossolani JW, Portugal JR, Moretti LG, Momesso L, de Campos M, Rosolem, CA (2022) Exploring the synergism between surface liming and nitrogen fertilization in no‐till system. Agron J. 114:1415–1430.

de Almeida JA, Ribeiro CF, de Oliveira MVR, Sequinatto L (2018) Mineralogia da argila e propriedades químicas de solos do Planalto Norte Catarinense. Rev Ciênc Agrovet. 17(2): 267-277.

de Almeida NM, de Almeida AAF, Mangabeira PAO, Ahnert D, Reis GSM, de Castro AV (2015) Molecular, biochemical, morphological and ultrastructural responses of cacao seedlings to aluminum (Al ³⁺) toxicity. Acta Physiol Plant. 37:1–17.

Fageria NK, Baligar VC (2008) Ameliorating soil acidity of tropical Oxisols by liming for sustainable crop production. Adv Agron. 99:345–399.

Fageria NK, Knupp AM (2014) Influence of lime and gypsum on growth and yield of upland rice and changes in soil chemical properties. J Plant Nutr. 37(8):1157-1170.

Henderson JS, Joyce RA, Hall GR, Hurst WJ, McGovern PE (2007) Chemical and archaeological evidence for the earliest cacao beverages. P Natl Acad Sci. 104:8937–18940.

ICCO International Cocoa Organization (2024) Quarterly Bulletin of Cocoa Statistics Ivory Coast.

Food and Agriculture Organization of the United Nations (2024) FAOSTAT statistical database [Rome]:FAO.

Fontoura SMV, de Castro Pias OH, Tiecher T, Cherubin MR, de Moraes RP, Bayer C (2019) Effect of gypsum rates and lime with different reactivity on soil acidity and crop grain yields in a subtropical Oxisol under no-tillage. Soil Tillage Res. 193:27-41.

Gomes FP, Garcia CH (2002) Estatística Aplicada a Experimentos Agronômicos e Florestais. Piracicaba: FEALQ.

Hafizal MT, Putra DP, Wirianata H, Nugraha NS, Suparyanto T, Hidayat AA, Pardamean B (2023) Implementation of expert systems in potassium deficiency in cocoa plants using forward chaining method. Procedia Comput Sci. 216: 136-143.

Hue N (2022) Soil acidity: Development, impacts, and management. In: B Giri, R Kapoor, QS Wu, A Varma (ed) Structure and Functions of Pedosphere. 103-Singapore: Springer Nature.

Hume R, Marschner P, Mason S, Schilling RK, Hughes B, Mosley LM (2023) Measurement of lime movement and dissolution in acidic soils using mid-infrared spectroscopy. Soil Tillage Res. 233:105807

Kone K, Akueson K, Norval G (2020) On the production of potassium carbonate from cocoa pod husks. Recycling. 5:23.

Li GD, Conyers MK, Helyar KR, Lisle CJ, Poile GJ, Cullis BR (2019) Long-term surface application of lime ameliorates subsurface soil acidity in the mixed farming zone of south-eastern Australia. Geoderma. 338:236–246.

Maddela NR, Kakarla D, García LC, Chakraborty S, Venkateswarlu K, Megharaj M (2020) Cocoa-laden cadmium threatens human health and cacao economy: A critical view. Sci Total Environ. 720:137645.

Marrocos PC, Loureiro GADA, Araujo QRD, Sodré GA, Ahnert D, Escalona-Valdez RA, Baligar VC (2020) Mineral nutrition of cacao (Theobroma cacao L.): relationships between foliar concentrations of mineral nutrients and crop productivity. J Plant Nutr. 43(10):1498-1509.

Miyazawa M, Pavan MA, Franchini JC (2002) Evaluation of plant residues on the mobility of surface applied lime. Braz Arch Biol Techn. 45:251–256.

Moraes, VR, Brito OR, Fioretto RA, Moreira A (2016) Changes in chemical properties of an oxisol due to gypsum application. Commun Soil Sci Plan. 47(5): 571-580.

Motamayor JC, Risterucci AM, Lopez PA, Ortiz CF, Moreno A, Lanaud C (2002) Cacao domestication I: the origin of the cacao cultivated by the Mayas. Heredity. 89:380–386.

Olwig MF, Asare R, Meilby H, Vaast P, Owusu K (2024) Introduction: Climate, Cocoa and Trees. In: MF Olwig, A Skovmand Bosselmann, K Owusu (ed) Agroforestry as Climate Change Adaptation. Palgrave Macmillan:Cham.

Pimentel-Gomes F (2009) Curso de Estatística Experimental. 15ª edição, Nobel.

Pavinato PS, Rodrigues M, Soltangheisi A, Sartor LR, Withers PJA (2017) Effects of cover crops and phosphorus sources on maize yield, phosphorus uptake, and phosphorus use efficiency. Agron J. 109(3): 1039-1047.

R Development Core Team (2023) R: A language and environment for statistical computing Vienna, Austria: R Foundation for Statistical Computing.

Rampim L, do Carmo Lana M (2015) Ion mobility and base saturation after gypsum application in continuous soybean-wheat cropping system under no-till. Aust J Crop Sci. 9(6): 468-476.

Ramos BZ, Toledo JPVF, Lima JMD, Serafim ME, Bastos ARR, Guimarães PTG, Coscione AR (2013) Doses de gesso em cafeeiro: influência nos teores de cálcio, magnésio, potássio e pH na solução de um Latossolo Vermelho distrófico. Rev Bras Cienc Solo. 37:1018-1026.

Reyes-Pérez JJ, Pincay-Ganchozo RA, Carrillo-Zenteno MD, Reynel V, Peña-Salazar K, Tezara W (2023) Macronutrient Fertilization and Cadmium Absorption in Two Cocoa Clones. Hortic. 9(11): 1223.

Ribeiro MAQ, Almeida AAF, Mielke MS, Gomes FP, Pires MV, Baligar VC (2013) Aluminum effects on growth, photosynthesis, and mineral nutrition of cacao genotypes. J Plant Nutr. 36(8):1161-1179.

Rheinheimer DS, Tiecher T, Gonzatto R, Zafar M, Brunetto G (2018) Residual effect of surface-applied lime on soil acidity properties in a long-term experiment under no-till in a Southern Brazilian sandy Ultisol. Geoderma. 313:7–16.

Rosas-Patiño G, Puentes-Páramo YJ, Menjivar-Flores JC (2017) Relación entre el pH y la disponibilidad de nutrientes para cacao en un entisol de la Amazonia colombiana. Corpoica Cienc y Tecnol Agropecu. 18:529–541.

Rusyanto A, Utami SNH, Purwanto BH (2017) The leaching of Ca and Mg cations on the application of calcite and gypsum for sugarcane cultivation in Central Lampung ultisols. Int J Agric Environ Res. 3(3): 3142-3154.

Sambuichi RH, Vidal DB, Piasentin FB, Jardim JG, Viana TG, Menezes AA, Baligar VC (2012) Cabruca agroforests in southern Bahia, Brazil: tree component, management practices and tree species conservation. Biodivers Conserv. 21: 1055-1077.

Santander Ruiz W, Garay Montes R, Verde Girbau C, Mendieta Taboada O (2021) Determinación del contenido de cadmio en suelos, frutos, granos fermentados y secos, licor de cacao y chocolate en zonas productoras de la región San Martín. Rev Soc Quím Perú. 87(1): 39-49.

SEI Superintendência de Estudos Econômicos e Sociais da Bahia Estatísticas Salvador

Shetty R, Prakash NB (2020) Effect of different biochars on acid soil and growth parameters of rice plants under aluminium toxicity. Sci Rep-Uk. 10(1):12249.

Snoeck D, Koko L, Joffre J, Bastide P, Jagoret P (2016) Cacao nutrition and fertilization. Sarv. 19:155-202.

Souza Júnior, JO, Sodré GA, Neves JCL (2018) Fertilidade do solo, correção da acidez e recomendação de adubação para o cacaueiro. In: JO Souza Júnior (ed) Cacau: cultivo, pesquisa e inovação. Editus, Ilhéus.

Teixeira PC, Donagemma GK, Fontana A, Teixeira WG (2017) Manual de métodos de análise de solo Rio de Janeiro: Embrapa Solos.

Tiecher T, Denardin LGDO, Martins AP, Pias OH, Nunes LR, Tiecher TL, Dos Santos DR (2019) Dynamics of sulfate and basic cations in soil solution as affected by gypsum fertilization in an Ultisol of Southern Brazil. Arch Agron Soil Sci. 65:1998-2012.

Van Vliet JA, Giller KE (2017) Mineral nutrition of cocoa: a review. Adv Agron. 141: 185-270.

Vicensi M, Lopes C, Koszalka V, Umburanas RC, Kawakami J, Pott CA, Müller MML (2020) Gypsum rates and splitting under no-till: Soil fertility, corn performance, accumulated yield and profits. J Soil Sci plant Nut. 20:690-702.

Zandoná RR, Beutler AN, Burg GM, Barreto, CF, Schmidt MR (2015) Gesso e calcário aumentam a produtividade e amenizam o efeito do déficit hídrico em milho e soja. Pesq Agropecu Tropi. 45:128–137.