Systematic review assessing the effects of amendments on acidic soils pH in tea plantations

Literature collection

We utilized the Web of Science (https://www.webofscience.com/wos), springer link (https://link.springer.com/), Google Scholar (https://scholar.scholar-xm.top/) and the China National Knowledge Infrastructure (https://www.cnki.net/) to investigate publications documenting the impacts of employing soil modifiers on the pH of tea plantations soil both domestically and internationally, up until June 2023. Specific search terms were used for literature screening, including “improvement materials”, “biochar” or “crop straw” or “animal manure” or “lime” or “rapeseed cake fertilizer” and “tea plantations soil pH”. Additionally, the reference lists of the retrieved articles were reviewed manually to identify additional relevant literature. In accordance with the data integration demands of the meta-analysis methodology and the objectives of this study, the retrieved original research articles was scrutinized using the following criteria: (1) The experimental conditions (e.g., time, location, management measures) must be are clearly outlined; (2) the soil utilized in the experiment must specifically be from tea plantations; (3) the experiment should encompass at least one treatment set involving the application of soil amendments and a control group with non-amendments, while ensuring consistency in other experimental conditions; (4) at least three replicates of each treatment in the experiment; (5) data regarding the mean value and standard deviation of soil pH must be accessible throughout the experiment. Based on these screening criteria, a total of 63 original research articles and 366 sets of valid data were collected.

Establishment of the database

The pH values were recorded for each article, including the mean for treatment and control, as well as the standard deviation (SD) or standard error (SE), and sample capacity (n). Standard deviation (SD) is an essential input variable in meta-analysis. Therefore, SD values were calculated in cases where the standard error (SE) for soil pH was presented in the manuscripts, using the equation for SD as provided by Brett et al. (2017).

(1) $${\text{SD=SE}}^{\ast }\sqrt{n}$$

Data extraction from selected valid literature involved capturing information such as the title of the articles, the first author, the experiment location, experiment time, soil pH, soil nutrients, the type of soil amendments, the pH and application amount of the soil amendments. A database of the relationship between the modifiers and soil pH was established through an Excel table. In building the database, data displayed in text and table form is directly extracted, while data displayed in graph form is using by GetData Graph Digitizer software. The unit of modifiers application is unified as t/ha. In pot or incubation experiments, unit conversion is 150,000 kg per 667 m² of tillage soil (Zhang et al., 2016). Soil pH values extracted by CaCl₂ and KCl were converted to deionized water extractable pH values by the following formula (Nguyen et al., 2017):

(2) $$\mathrm{p}\mathrm{H}({\mathrm{H}}_2\mathrm{O})=1.65+0.86\mathrm{p}\mathrm{H}(\mathrm{C}\mathrm{a}\mathrm{C}{\mathrm{l}}_2)$$

(3) $$\mathrm{p}\mathrm{H}({\mathrm{H}}_2\mathrm{O})=1.96+0.74\mathrm{p}\mathrm{H}(\mathrm{K}\mathrm{C}\mathrm{l})$$

Data classification

According to “China Soil”, the soil pH is classified as follows: extremely acidic soil (pH ≤ 4.5); strongly acidic soil (4.5 < pH ≤ 5.5); acidic soil (5.5 < pH ≤ 6.5); neutral soil (6.5 < pH ≤ 7.5); alkaline soil (pH > 7.5). The application rates for different amendments are categorized as follows: for biochar, the divisions are M ≤ 10 t/ha; 10 < M ≤ 40 t/ha; 40 < M ≤ 80 t/ha; M > 80 t/ha. For lime, the divisions are M ≤ 1.5 t/ha; 1.5 < M ≤ 4.5 t/ha; M > 4.5 t/ha. In the case of crop straw, the divisions are M ≤ 20 t/ha; 20 < M ≤ 40 t/ha; M > 80 t/ha. Finally, for animal waste, the divisions are M ≤ 4 t/ha; 4 < M ≤ 10 t/ha; M > 10 t/ha. The pH of the amendments was divided into three groups: pH <= 8, pH <= 10, and pH > 10. Then the biochar in the amendments continues to be classified in detail, and the types of biochar are roughly divided into four categories according to the raw materials of biochar: crop straw (e.g., wheat straw, corn straw, rice straw), wood (e.g., tea branches, bamboo), shell residue (e.g., peanut shell, bamboo coconut shell) and animal waste (e.g., pig manure, sheep manure). Regarding the pyrolysis temperature of biochar, if the literature gives a temperature range, it is classified as its average value. The pyrolysis temperature is divided into four ranges: low temperature (≤400 °C), medium temperature (401– 500 °C), medium high temperature (501– 600 °C), and high temperature (>600 °C).

Statistical analysis

Meta-analysis is a statistical method that quantitatively synthesizes multiple research results. It has the capability to analyze comprehensively and quantitatively the effects of different influencing factors, exploring the relationship between results and clarifying the relative contribution of each influencing factor.

In this study, meta-analysis was utilized to compare the effect of soil amendments on soil pH, requiring the inclusion of quantitative experimental data for effect value indicators. The response ratio (RR) was calculated based on the natural logarithm of soil pH, with lnRR being determined using formula (Hedges, Gurevitch & Curtis, 1999) (4):

(4) $$\mathrm{l}\mathrm{n}\mathrm{R}\mathrm{R}=\mathrm{l}\mathrm{n}(\mathrm{I}/\mathrm{C}\mathrm{K})=\mathrm{l}\mathrm{n}\mathrm{I}-\mathrm{l}\mathrm{n}\mathrm{C}\mathrm{K}$$

In the formula, M represents the average soil pH of the treatment group after the application of modifiers, while CK represents the average soil pH of the control group without modifiers. Additionally, the coefficient of variation (V), weighted factor (W), weighted response ratio (RR), the standard error (S) of RR, and its 95% confidence interval (CI) can be calculated as follows (Luo, Hui & Zhang, 2006):

(5) $$V_{\mathrm{l}\mathrm{n}RR}={\displaystyle \frac{{\mathrm{S}\mathrm{D}}_{\mathrm{M}}^2}{n_{\mathrm{M}}{M}^2}+{\displaystyle \frac{{\mathrm{S}\mathrm{D}}_{\mathrm{C}\mathrm{K}}^2}{n_{\mathrm{C}\mathrm{K}}C{K}^2}}}$$

(6) $$W_{ij}={\displaystyle \frac{1}{V}}$$

(7) $$R{R}_{++}={\sum }_{i=1}^m{\sum }_{j=1}^{k_i}{W}_{ij}R{R}_{ij}/{\sum }_{i=1}^m{\sum }_{j=1}^{k_i}{W}_{ij}$$

(8) $$s(R{R}_{++})=\sqrt{\frac{1}{{\sum }_{i=1}^m{\sum }_{j=1}^{k_i}{W}_{ij}}}$$

In the formula, SD_M² and SD_CK² represent the standard deviation of the treatment group (with modifiers) and the control group (without modifiers), respectively. Similarly, n_M and n_CK denote the number of samples in the treatment group and the control group, respectively. The smaller the standard deviation of the effect value, the larger the weight assigned, and the weight response ratio (the percentage increase or decrease of the treatment relative to the control) and its 95% CI can be converted by exp (RR₊₊−1) × 100% (Nguyen et al., 2017).

If the 95% confidence interval exceeds 0, it indicates that the application of soil amendments has a significant positive effect on soil pH. Conversely, if the 95% confidence interval is below 0, it suggests that the soil amendments application has a negative effect on soil pH. On the other hand, when the 95% confidence interval spans 0, it means that the soil amendments application has no significant effect on soil pH. The data processing described above was carried out using the metafor package in R software. The significant difference among covariate levels was determined through meta-regression analysis using random effects, maximum likelihood estimator, and mixed-effects models (Ebensperger, Rivera & Hayes, 2012; Ferreira et al., 2016). Total heterogeneity for a covariate was calculated using mixed-effects models as follows (Ferreira et al., 2016):

(9) $${\mathrm{Q}}_{\mathrm{t}}={\mathrm{Q}}_{\mathrm{m}}+{\mathrm{Q}}_{\mathrm{e}}$$where Q_t is the overall degree of heterogeneity estimated by the mixed effects model, with Q_m denoting the total heterogeneity between the levels of covariates explained by the model, and Q_e representing the residuals not explained by the model. If 95% of the CI of the covariates do not overlap, it indicates that the levels of the covariates are significantly different from each other (Ferreira et al., 2016). Qm represents the heterogeneity caused by the explanatory variable, and a larger value, indicates a greater influence of the explanatory variable on the effect value. In this study, I² = 99.52%, indicating very diverse and highly heterogeneous effect values across different study cases. Furthermore, our data set on amendments’ effect on soil pH value showed strong heterogeneity with Q_t = 39,055.5480, p < 0.001, so it is necessary to introduce explanatory variables to explain the heterogeneity.

Publication bias test

Meta-analysis is a quantitative assessment of the average effect value, using data obtained from published literature, which may be affected by publication bias. Rosenthal’s fail-safe number (Nfs) is used as a tool to assess publication bias in the entire database. If Nfs > 5n + 10 (n represents the sample volume), it indicates the absence of publication bias (Macaskill, Walter & Irwig, 2001). In this study, Nfs = 40,896,65, significantly exceeding the threshold of 5n + 10. Therefore, it can be concluded that this study is not affected by publication bias, confirming the reliability of our results.

Overall distribution of response ratio to soil pH change in tea plantations

As illustrated in Fig. 1, SPSS26.0 software was utilized to analyze the normal distribution of soil pH change data of the tea plantations treated with soil amendments. The histogram depicts the frequency distribution of the response ratio of soil pH change in the tea plantations, and the fitting curve aligns with a normal distribution (p < 0.001), indicating that the data analyzed in the study are consistent and suitable for meta-analysis.

Effects of different types of amendments on soil pH in tea plantations

Figure 2 illustrates the significant impact of various types of soil amendments on the pH of tea plantations, showing varied levels of increase (p < 0.0001). Lime exhibited the highest increase at 18%, followed by animal waste and biochar at 10% and 8%, respectively. In comparison, rapeseed cake and crop straw had a less pronounced effect on the soil pH of tea plantations, with increases of 2% and 5%, respectively.

Effect of lime on soil pH in tea plantations

Figure 3A illustrates a significant increase in soil pH levels resulting from the application of lime to soils with varying initial pH levels (p < 0.0001). When lime was applied to extremely acidic soil (pH ≤ 4.5), the increase in soil pH was 19%, which was higher than that observed in strongly acidic soil (4.5 < pH ≤ 5.5). However, the difference between the two rates was not statistically significant (p > 0.05).

Figure 3B shows that lime application has a significant effect on increasing soil pH in tea plantations (p < 0.0001); however, the magnitude of this increase varies based on the amount of lime applied. Overall, the rise in soil pH was directly proportional to the amount of lime applied. When the lime application exceeded 4.5 t/ha, the soil pH increase was most pronounced, reaching 27%, a significantly higher value than those observed at M ≤ 1.5 (10%) and 1.5 < M ≤ 4.5 (14%) levels (p < 0.0001). Nonetheless, no significant difference was found between M ≤ 1.5 and 1.5 < M ≤ 4.5 (p > 0.05). Figure 3C shows the impact of the experiment type on lime treatment to increase soil pH. In the incubation experiment, the increase of soil pH by applying lime was 25%, a significant difference compared to the field experiment (p < 0.01).

Effect of application crop straw and rapeseed cake on soil pH

Figure 4A illustrates that the application of crop straw can lead to an increase in soil pH in tea plantations. There appears to be a correlation between the quantity of crop straw applied and the rise in soil pH; however, no significant variances were observed among the different application amounts (p > 0.05). Applying rapeseed cake at a rate below 10t/ha in tea plantations soil can also elevate soil pH, although the increase is less compared to other amendments. Conversely, when the application amount exceeds 10t/ha, it will have an adverse effect on soil pH (Fig. 4B). Figure 4C shows that the application of rapeseed cake in pot experiments resulted in a slightly better increase than was achieved in the field experiments, but the difference was not significant (p > 0.05).

Effect of animal waste on soil pH

Figure 5A demonstrates that the application of varying amounts of animal waste to tea plantations can substantially raise soil pH, with no significant difference in the improvement effect observed between different application rates (p > 0.05). Additionally, applying animal waste with different pH levels in the soil could increase soil pH, and the improvement effect was enhanced with increasing animal waste pH, although the difference was not statistically significant (p > 0.05). Furthermore, under different initial soil pH conditions ranged from 5.5 to 6.5, applying animal waste had the most pronounced improvement effect on soil pH, reaching 21%, which was significantly higher compared to initial soil pH levels of 4.5–5.5 and less than 4.5 (p < 0.05) (Fig. 5C). In incubation and field experiment, application of animal waste significantly increased soil pH (9% and 10%), but the difference was not significant (p > 0.05) (Fig. 5D).

Effect of biochar on soil pH in tea plantations

Experimental type

Figure 6A illustrates the impact of different experiment types on the increase of soil pH in tea plantations through biochar treatment (p < 0.01). Regardless of experimental type, application of biochar consistently and significantly increased soil pH. In pot conditions, the maximum increase in soil pH due to biochar application was 11%, which was significantly higher than the increases observed in field (p < 0.0001) and incubation experiment (p < 0.05). Furthermore, the increase in soil pH in field and incubation experiments were 6% and 8%, respectively, with no significant difference (p > 0.05).

Correlation analysis of soil pH and its influencing factors in tea plantations

Within the biochar subgroup analysis, the regression analysis results (Fig. 7) indicated a positive correlation between the soil pH changes and both initial soil pH and biochar pH (p < 0.01) in tea plantations. The acidic soil (5.5 < pH ≤ 6.5) and the application of biochar with higher pH exhibited the most significant effects on soil pH improvement in tea plantations. There were no significant variations in soil pH changes with increasing pyrolysis temperatures of biochar (p > 0.05). At 500 °C, the effect on soil pH was unstable resulting in a wide range of changes. Additionally, a positive correlation was found between soil pH changes and biochar application rates (p < 0.01). For other types of amendments (Fig. 8), a significant positive correlation was observed between changes in soil pH and the application of lime (p < 0.01). However, no substantial correlation was found the application of animal waste, crop straw, and rapeseed cake (p > 0.05). Notably, as the amount of rapeseed cake applied increases, the change in soil pH gradually decreases. Overall, it was found that both the pH and type of biochar, as well as the amount of biochar and lime applied, have a greater influence soil pH change in tea plantations. Utilizing animal waste for pyrolysis, alkaline biochar at temperatures ranging from 501–600 °C, and applying lime are more effective on improving soil pH in acidic soils.