Introduction

Soil acidity is one of the main constraints to global agricultural production (Gurmessa, 2021; Rheinheimer et al., 2018). About 40 % of the world’s arable land is affected by high acidity, particularly in Africa, Asia, and South America (Sumner and Noble, 2003). Acidic soils are characterized by low nutrient availability and cation exchange capacity dominated by H⁺ and Al³⁺ (Shetty and Prakash, 2020), which toxicity impairs root development reducing nutrient uptake, crop growth, and yield (Borges et al., 2020).

The cocoa tree is tolerant to Al³⁺, however, this element reduces the root and shoot dry weight of cocoa seedlings (Baligar and Fageria, 2005). Liming enhances biomass production in cocoa seedlings and boosts cocoa bean yield. It also increases soil pH, as well as Ca²⁺ and Mg²⁺ levels, and lowers Al³⁺ saturation (Hirzel et al., 2020; Souza Júnior et al., 2018). However, this input has low mobility in the soil. Therefore, its corrective effect is restricted to topsoil layers (Li et al., 2019; Rheinheimer et al., 2018).

Agricultural gypsum is a good alternative to reduce subsoil acidity, especially aluminum saturation (m%). Its mobility in the soil profile reduces Al³⁺ toxicity and increases Ca2⁺ and SO₄^2- concentrations in deeper soil layers (Enesi et al., 2023). Consequently, water and nutrient absorption in deep layers increases (Bossolani et al., 2020). However, the specific effects of gypsum amendment on cocoa trees are still not fully understood, highlighting the need for further research. Nevertheless, in their pioneering study, Nakayama (1986) showed that gypsum amendment improved the development and distribution of the root system of cocoa seedlings.

This study aimed to assess changes in chemical attributes of soil acidity resulting from lime and gypsum surface application in no-till systems cultivated with cocoa tress in the southeastern region of the state of Bahia, Brazil.

Materials and methods

Data collection

Soil samples were collected from nineteen experimental areas (EAs) located in fifteen municipalities across two climate zones: humid and humid-to-subhumid (SEI-BA, 2022), in the southeastern region of Bahia, Brazil (Figure 1). In total, 20 simple samples were collected for each composite sample. The analysis of the 0-20 cm and 20-40 cm layers were used to calculate the doses of lime and gypsum, respectively (Table 1). The study period was 87 months. Time zero (T0) refers to the month of liming and gypsum amendment, followed by subsequent sampling at T40, T52, T72, and T87 months after fertilization. Samples were collected n 0-10 cm, 10-20 cm, and 20-40 cm layers at these four times, and the values for the 0-20 cm layer were calculated by averaging the values of the 0-10 cm and 10-20 cm layers. Two composite samples were collected per EA at T40, while six composite samples were collected per EA at T52, T72, and T87.

Soil attributes

The following soil attributes were analyzed: pH in H₂O; Al³⁺, Ca²⁺, and Mg²⁺ extracted with 1.0 mol L^-1 KCl; potential acidity (H+Al) extracted with 0.5 mol L^-1 Ca(CH₃COO)₂; organic matter (OM) determined using the Walkley-Black method; and remaining P (P-rem), assessed by stirring 60 mg L^-1 P in a 0.01 mol L^-1 CaCl₂ solution (Teixeira et al., 2017). The sand, silt, and clay contents were determined according to Teixeira et al. (2017). The aluminum saturation was calculated as follows (Eq.1):

The results of the preliminary chemical analysis and the respective lime and gypsum doses are outlined in Table 1.

Calculation of lime and gypsum doses

Lime doses were calculated using the base saturation method (Eq. 2). The objective was to increase base saturation (V%) to 75 % in EAs with initial V% values below 65 % in the 0-20 cm layer, considering a reaction depth (d) of 5 cm and 90 % relative Total Neutralizing Power (RTNP) of the soil corrective. For this purpose, the amount was calculated as follows:

Where:

LD = lime dose, in tons per hectare (t ha^-1);

V2 = target base saturation (75 %);

V1 = base saturation of the 0-20 cm layer, in %; T = CEC at pH 7.0, in mmolc dm^-3;

RTNP = relative total neutralizing power (90 %); d = lime reaction depth (5 cm).

Gypsum doses were calculated according to the stochiometric equation proposed by Souza Júnior et al. (2018) (Equation 3). The objective was to reduce Al³⁺ saturation (m%) to 20 % in EAs with m% higher than 20 %, considering a layer thickness (LT) of 30 cm and a gypsum percentage (%Cages) of 16 %. For this purpose, the amount was calculated as follows:

where:

GD = gypsum dose, in tons per hectare (t ha^-1); Al³⁺ = Al³⁺ concentration in the 20-40 cm layer, in mmolc dm^-3; mt = tolerable Al³⁺ saturation (cocoa tree = 20 %); t = effective CEC in the 20-40 cm layer, in mmolc dm^-3; LT = layer thickness (30 cm); %Cagyp = percentage of Ca in gypsum (16 %).

Lime and gypsum amendment method

The lime and gypsum doses (Table 1) were converted to grams per plant and manually broadcast, covering the entire soil surface in a no-till system. To enhance contact between the inputs and the soil, the organic mulch was turned over using a leaf blower.

Statistical analysis

To assess the homogeneity of error variances, a one-way ANOVA was performed for each chemical attribute within each EA, subsequently calculating the ratio of the largest to the smallest mean square residual. The values were well above 7, indicating a lack of homogeneity of variances. Therefore, each EA was considered as an individual experiment. Finally, a regression analysis was performed to evaluate the variation of chemical attributes over time (0, 40, 52, 72, and 87 months), testing linear and quadratic coefficients using the mean square error from the ANOVA of each EA. Significance was assessed at up to 10 % probability, according to the F-test. All statistical tests were performed using the R software (R Core Team, 2023).

Results

The pH values of the 0-20 cm layer remained stable over time in eleven of the EAs (Table 2), seven of which had received surface liming at doses ranging from 0.36 to 1.09 t ha^-1 (Table 1). The pH of the 0-20 cm layer decreased over time in eight EAs, seven of which had received surface liming at doses ranging from 0.54 to 1.04 t ha^-1. In contrast, the pH of the 20-40 cm layer remained stable in twelve EAs, while it decreased in seven EAs over time (Table 2).

Al³⁺ concentration in the 0-20 cm layer did not decrease in any of the fourteen EAs that received surface liming; however, Al³⁺ concentration increased over time in ten EAs (Table 3), nine of which had received liming ranging from 0.54 to 1.09 t ha^-1 (Table 1). The nine EAs where Al³⁺ concentration did not change were the same ones where the pH did not change either.

Al³⁺ concentrations in the 20-40 cm layer varied in three of the eight EAs that received gypsum amendment (Table 3). The quadratic models were significant, indicating a decrease in Al³⁺ concentrations, which reached minimum values and subsequently increased over time. In these EAs, the effect of gypsum on reducing Al³⁺ concentrations peaked at 38, 52, and 25 months in EAs 3, 4, and 7, respectively. Al³⁺ concentration increased linearly over time in the 20-40 cm layer in five EAs (Table 3); however, gypsum was applied in only one of them, i.e., EA 12, which received a dose of 2.39 t ha^-1 gypsum (Table 1). Al³⁺ concentrations did not change in eleven EAs, four of which had received gypsum amendment at doses ranging from 0.60 to 1.06 t ha^-1.

m% in the 0-20 cm layer did not vary over time in eleven EAs, seven of which received doses of lime ranging from 0.36 to 1.04 t ha^-1 (Tables 1 and 4). Al³⁺ concentrations did not change either in nine of the eleven EAs mentioned above (Tables 4 and 5). m% increased over time in eight EAs, all of which had received surface liming at doses ranging from 0.54 to 1.09 t ha^-1, except for EA 6.

Aluminum saturation in the 20-40 cm layer varied over time in ten EAs (Table 5), increasing in five of them. In these EAs, gypsum was not applied. m% variation over time was described using a quadratic equation in the five EAs. All of them received gypsum amendments (Table 5 and Figure 2). The mean minimum m% (m%min) was 15.6 %, reached, on average, 41 months after applying gypsum.

Discussion

The pH in the 0-20 cm layer did not increase in the EAs that received surface liming (Table 2). This result suggests that the corrective effect was restricted to the topsoil layers, which were not assessed in this study. Furthermore, this result corroborates the findings of other studies, which have shown that the corrective effects of surface liming in a no-till system are limited to the topsoil layers, particularly at depths of 0-5 cm and 5-10 cm (Alleoni et al., 2005; Zandoná et al., 2015). In a field study with cocoa trees, Nakayama (1986) observed that surface liming in a no-till system corrected soil acidity up to a depth of 10 cm, six years after the application of this amendment. The effect of surface liming is restricted to the topsoil layers because lime has low mobility in the soil due to the loss of the accompanying ion (CO₃^2-) during dissolution reactions, which consume H⁺, yielding H₂CO₃. Subsequently, H₂CO₃ dissociates into H O and CO. In addition, the Al³⁺ precipitation 2 reaction with OH2 - derived from lime dissolution forms Al(OH)₃, contributing to the surface effect of this input (Rheinheimer et al., 2018).

The coefficients of the linear regression models for the 0-20 cm layer ranged from 0.007 to 0.013 per month, corresponding to an annual pH reduction rate ranged from 0.084 to 0.156 units (Table 2). In the 20-40 cm layer, the annual pH reduction rate ranged from 0.048 to 0.096. These results suggest that the acidification process is more intense in the topsoil layer.

The soil buffering indicators T, P-rem, OM, and clay did not explain differences in pH variation between soils because their mean values were 96.6 mmolc dm-3, 37.6 mg L^-1, 30.1 g kg^-1, and 42.1 %, respectively, in the eleven EAs where the pH of the 0-20 cm layer did not change over time. These results indicate a lower acidity buffering in these soils than in the soils of the eight EAs whose pH values decreased-1over time, with average values of 102.0-1 mmolc dm-3, with average values of 102.0-1 mmolc dm^-3, 31.5 mg L^-1 , 34.8 g kg^-1 , and 47.9 % respectively.

Climate may be useful to explain these results because seven of the eight EAs whose pH values decreased over time are located in the rainiest climate zone (Tables 1 and Figure 1). High precipitation rates promote the leaching of basic cations, thereby increasing soil acidity. In addition to climate, the decrease in pH values over time in the seven EAs that received surface liming might have been caused by interactions between factors such as the low dose of lime applied to the soil, averaging 0.71 t ha^-1, and OM mineralization.

Neither the Al³⁺ concentration (Table 3) nor the pH (Table 2) changed over time in the 0-20 cm layer in nine EAs, suggesting a close correlation between these two attributes. Al³⁺ solubility and, hence, toxicity, are primarily controlled by the pH. Al³⁺ availability troughs at pH levels above 5.5 (Ruíz et al., 2016). Therefore, despite having received surface liming, some EAs did not show variations in Al³⁺ because the pH did not change.

The m% in the 0-20 cm layer did not change in eleven of the EAs. In ten of them, the mean m% ranged from 0 % to 8.3 % (Table 4). These m% values were low for cocoa trees, according to Souza Júnior et al. (2018). The mean m% was higher (14.9 %) only in EA 9; however, it was still lower than the threshold of 15 % suggested by Baligar and Fageria (2005) as the maximum m% tolerated by the crop. Nevertheless, m% increased over time in this layer in eight EAs (Table 4), reflecting a rise in Al³⁺ concentrations despite surface liming being applied in seven of them at doses ranging from 0.54 to 1.09 t ha^-1 (Table 1). In this context, the increase in Al³⁺ and, thus, in m%, might have resulted from natural soil acidification processes such as OM mineralization, root uptake of basic cations and H+ extrusion, leaching of bases, and H+ and Al³⁺ accumulation in the soil profile (Li et al., 2019). The low mean dose of 0.71 t ha^-1 lime combined with the low mobility of this input might have also contributed to this result (Alleoni et al., 2005; Parecido et al., 2021; Zandoná et al., 2015). The coefficients of the linear regression models indicated an increase in m%, ranging from 0.5 % to 5.6 % per year.

The m% in the 20-40 cm layer decreased in five EAs after surface gypsum amendment (Table 5). The minimum m% (m%min) reached in each EA was estimated based on quadratic models of the variation of m% over time and gypsum dose. The m%min values ranged from 7.2 % to 25.5 %, averaging 15.6 %. This value is close to the target m% of 20 % which was adopted as the tolerable threshold (mt) for cocoa cultivation and used to calculate the doses of gypsum (Equation 3).

The following time estimates were also calculated: T1, the time required to reach the target m% tolerated by the crop (mt), set at 20 % (Equation 3); T2, the time to reach the minimum m%_min, representing the maximum duration of gypsum effectiveness; and T3, the time until m% rises back to 20 %, indicating the need for gypsum reapplication. The mean T1, T2, and T3 values were 20.5, 41, and 60.5 months, respectively (Table 5). The maximum gypsum effectiveness on m% reduction (T2) ranged from 28 to 51 months after application. This range is higher than the maximum time of 22 months assessed by Nora and Amado (2013) and similar to the maximum time (43 months) assessed by Caires et al. (2000). Moreover, both studies reported that gypsum chemically improves subsurface soil layers by reducing Al³⁺ concentrations and m%. However, the periods assessed in both studies were shorter than those of this study, where the time for gypsum reapplication (T3) ranged from 52 to 80 months (Table 5). This time interval exceeded the residual effect of the gypsum amendment of 60 months reported by da Costa et al. (2016), and the maximum duration of 72 months observed by Pauletti et al. (2014).

The m% in the 20-40 cm layer increased over time in five of the EAs that were not treated with gypsum (Tables 1 and 5). Based on the linear regression models, m% increased at an annual rate ranging from 2.8 % to 8.0 %.

Conclusions

Soil acidification in experimental areas (EAs) under cocoa cultivation varied markedly over a period of 87 months, showing no correlation with soil buffering indicators.

Surface liming in a no-till system did not change the chemical attributes in the 0-20 cm layer; therefore, its effect might have been restricted to the topsoil layers.

Surface gypsum amendment in a no-till system led to a reduction in Al³⁺ concentration and, especially, Al³⁺ saturation in the 20-40 cm layer in some EAs; however, this effect diminished over time.

References

Alleoni, L. R. F.; Cambri, M. A. and Caires, E. F. (2005). Chemical attributes of a cerrado Oxisol under no-tillage as affected by lime application methods and doses. Revista Brasileira de Ciência do Solo, 29(6), 923-934. https://doi.org/10.1590/S0100-06832005000600010

Baligar, V. C. and Fageria, N. K. (2005). Soil aluminum effects on growth and nutrition of cacao. Soil Science and Plant Nutrition, 51(5), 709-713. https://doi.org/10.1111/j.1747-0765.2005.tb00097.x

Borges, C. E.; Cazetta, J. O.; Sousa, F. B. F. and Oliveira, K. S. (2020). Aluminum toxicity reduces the nutritional efficiency of macronutrients and micronutrients in sugarcane seedlings. Ciência e Agrotecnologia, 44, e015120. https://doi.org/10.1590/1413-7054202044015120

Bossolani, J. W.; Crusciol, C. A. C.; Merloti, L. F.; Moretti, L. G.; Costa, N. R.; Tsai, S. M. and Kuramae, E. E. (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. https://doi.org/10.1016/j.geoderma.2020.114476

Caires, E. F.; Banzatto, D. A. and Fonseca, A. F. (2000). Surface application of lime under a no-tillage system. Revista Brasileira de Ciência do Solo , 24(1), 161-169. https://doi.org/10.1590/S0100-06832000000100018

Da Costa, C. H. M.; Crusciol, C. A. C.; Neto, J. F. and Castro, G. S. A. (2016). Residual effects of superficial liming on tropical soil under no-tillage system. Pesquisa Agropecuária Brasileira, 51(9), 1633-1642. https://doi.org/10.1590/S0100-204X2016000900063

Enesi, R. O.; Dyck, M.; Chang, S.; Thilakarathna, M. S.; Fan, X.; Strelkov, S. and Gorim, L. Y. (2023). Liming remediates soil acidity and improves crop yield and profitability-a meta-analysis. Frontiers in Agronomy, 5, 1194896. https://doi.org/10.3389/fagro.2023.1194896

Gurmessa, B. (2021). Soil acidity challenges and the significance of liming and organic amendments in tropical agricultural lands with reference to Ethiopia. Environment, Development and Sustainability, 23, 77-99. https://doi.org/10.1007/s10668-020-00615-2

Hirzel, J.; Undurraga, P.; León, L.; Panichini, M.; Carrasco, J.; González, J. and Matus, I. (2020). Maize grain production, plant nutrient concentration and soil chemical properties in response to different residue levels from two previous crops. Acta Agriculturae Scandinavica, Section B - Soil and Plant Science, 70(4), 285-293. https://doi.org/10.1080/09064710.2020.1725619

Li, G. D.; Conyers, M. K.; Helyar, K. R.; Lisle, C. J.; Poile, G. J. and Cullis, B. R. (2019). Long-term surface application of lime ameliorates subsurface soil acidity in the mixed farming zone of south-eastern Australia. Geoderma , 338, 236-246. https://doi.org/10.1016/j.geoderma.2018.12.003

Nakayama, L. H. I. (1988). Influência das formas de aplicação de calcário e gesso agrícola sobre o desenvolvimento do cacaueiro. Revista Theobroma, 18(4), 241-246. https://repositorio-dspace.agricultura.gov.br/handle/1/1207

Nora, D. D. and Amado, T. J. C. (2013). Improvement in chemical attributes of oxisol subsoil and crop yields under no-till. Agronomy Journal, 105(5), 1393-1403. https://doi.org/10.2134/agronj2013.0031

Parecido, R. J.; Soratto, R. P.; Perdoná, M. J.; Guidorizzi, F. V.C.; Gomes, G. G.; Paula, R. A. and Gitari, H. I. (2021). Limestone increased coffee yield and profitability more than phosphogypsum or their combination. Agronomy Journal , 113(4), 3586-3599. https://doi.org/10.1002/agj2.20712

Pauletti, V.; de Pierri, L.; Ranzan, T.; Barth, G. and Motta, A. C. V. (2014). Long-term effects of the application of gypsum and lime in a no-till system. Revista Brasileira de Ciência do Solo , 38(2), 495-505. https://doi.org/10.1590/S0100-06832014000200014

R Core Team(2023). R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing. https://www.R-project.org

Rheinheimer, D. S.; Tiecher, T.; Gonzatto, R.; Zafar, M. and 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. https://doi.org/10.1016/j.geoderma.2017.10.024

Ruíz, J. C.; Cerna-Rebaza, L.; Hernández-Villalobos, K.; Silva- Pereda, L. and Vásquez-Cunya, F. (2016). Effect of aluminum and pH on the growth of roots of Phaseolus vulgaris var. caballero under laboratory conditions. REBIOL, 36(2), 4-15. https://revistas.unitru.edu.pe/index.php/facccbiol/article/view/1696

SEI-BA. (2022). Tipologia climática Thornthwaite & Matter. Superintendência de Estudos Econômicos e Sociais da Bahia (SEI-BA). https://ftp.sei.ba.gov.br/Geoinformacao/mapas/estado/mapa_tipclim-tm_ba_2m_1998_cor.pdf.

Shetty, R. and Prakash, N. B. (2020). Effect of different biochars on acid soil and growth parameters of rice plants under aluminum toxicity. Scientific Reports, 10, 12249. https://doi.org/10.1038/s41598-020-69262-x

Souza Júnior, J. O.; Sodré, G. A. and Neves, J. C. L. (2018). Fertilidade do solo, correção da acidez e recomendação de adubação para o cacaueiro. In Souza Júnior, J. O. (ed.). Cacau: cultivo, pesquisa e inovação (pp. 333-377). Editus. https://doi.org/10.7476/9786586213188

Sumner, M. E. and Noble, A. D. (2003). Soil acidification: The world story. In Rengel, Z. (ed.). Handbook of Soil Acidity (p. 27). CRC Press. https://www.taylorfrancis.com/chapters/edit/10.1201/9780203912317-3/soil-acidification-world-story-malcolm-sumner-andrew-noble

Teixeira, P. C.; Donagemma, G. K.; Fontana; A. and Teixeira, W.G. (2017). Manual de métodos de análise de solo (p. 574). Brasília: EMBRAPA. https://ainfo.cnptia.embrapa.br/digital/bitstream/item/181717/1/Manual-de-Metodos-de-Analise-de-Solo-2017.pdf

Zandoná, R. R.; Beutler, A. N.; Burg, G. M.; Barreto, C. F. and Schmidt, M. R. (2015). Gypsum and lime increase soybean and maize yield and decrease drought stress. Pesquisa Agropecuária Tropical, 45(2), 128-137. https://doi.org/10.1590/1983-40632015v4530301