The data collection comes partly from literature search and partly from field measurement. We searched relevant peer-reviewed published journal articles from 2005 to 2020 from Web of Science, Google Scholar, and CNKI. The key words used in the search were: “natural forest,” “planted forest,” “soil physical and chemical properties.” 182 relevant papers were retrieved. Then the data was screened in the literature using the following criteria: (1) The study should provide the latitude and longitude of the plot, and whether the forest in the plot is a natural forest or a planted forest. (2) The study should provide complete site and forest information, including the plot location, tree species, age of the stand, tree DBH, and tree density. (3) The study should provide or allow the calculation of the mean value, standard deviation, or standard error of the data on the physical and chemical properties of the soil in the plot. (4) The study presents the results of field studies rather than retrospective or simulation studies. For the articles meeting our standards, soil nitrogen, soil available phosphorus and soil pH indexes in the physical and chemical properties of 0–20 cm soil surface were extracted. If a study has multiple sampling depths from 0 to 20 cm at the same site, we treat these observations as independent samples. In these articles, stand characteristics such as age, stand density and DBH of each plot were collected.

Based on field surveys, 21 sites were selected to conduct actual measurements of soil nutrients data in 298 forests. At each site, at least four 30 × 30 m forest plots were selected, each representing the typical zonal vegetation of the area. We recorded the latitude, longitude, elevation and slope of each site for comprehensive analysis, and recorded the site location, tree species, age, DBH and tree density in real time. In the actual sampling, we used the same method as in the literature to measure the physical and chemical properties of soil. In each forest plot, a five-point sampling method was adopted. Using a soil corer, soil samples were collected from different directions at depths of 0–30 cm. The soil samples were air-dried at room temperature. After air drying, roots, stones, and other debris were removed. The samples were then gently ground only and sieved through a 2 mm mesh, and preserved for the determination of soil N and available P contents as well as pH. The soil N content was determined by the Kjeldahl method. KCl solution was added to the soil samples to extract NH₄⁺ from the soil. At the same time, NO3⁻ in the soil is extracted by adding saturated CaSO₄ solution (Verma and Sagar, 2020). Based on the off-line column extraction method, the soil samples were extracted using HCl and MH₂SO₄ solutions, and then the available phosphorus content in the soil samples was analyzed by using the FI amperometric method (Jakmunee and Junsomboon, 2009). And soil pH was measured in a 1:2.5 soil and deionized water mixture for each sample using a pH meter (PHS-3C, Lei ci). Before measurement, the soil solution was shaken for 30 min and then boiled for 5 min. The Soil Inorganic Carbon (SIC) was determined using the 08.53 calculator (M1.08.53.E, Eijkelkamp). Here, SIC refers to the carbonate value measured by the amount of carbon dioxide (CO₂) emission during digestion with strong acid (Hong et al., 2019).

A t-test at a significance level of 0.05 was used to examine the differences in soil N, available P, and pH between plantation and natural forests. Significance analysis of differences was conducted in the R software package (version 4.3.1) “agricolae.” Principal Component Analysis (PCA) for dimensionality reduction of leaf functional traits data was performed using the R package “pcaMethods,” with the two principal components (PC1 and PC2) obtained representing key leaf functional traits (Homeier et al., 2021). Among them, PC1 is the first principal component extracted from leaf functional traits, and PC2 is the second principal component. Climate factors include mean annual temperature (MAT), mean temperature of the coldest month (MACT), mean temperature of the hottest month (MAHT), mean annual precipitation (MAP), mean annual evaporation (MAE), and mean annual sunshine duration (ASD). Stand factors include stand age, mean diameter at breast height, and stand density.

A linear mixed-effects model was employed to explore the impacts of climate factors, stand factors, key leaf traits, and nitrogen deposition on soil nutrients in plantation and natural forests, while fully considering the effect of random effects on the outcomes. This analysis was conducted using the R package “lme4” (Gurevitch et al., 2018). The R package “linkET” was used to test for multicollinearity among various potential influencing factors, and the correlations between these factors and soil N, available P, pH were visualized as heatmaps (Zhao et al., 2014).

Variance decomposition analysis was used to explore the contribution (%) of various influencing factors to the spatial variability of soil nutrients in plantation and natural forests, and this analysis was completed using the “vegan” package in R language (Kong et al., 2022). The boosted regression analysis method was employed to quantify the independent contribution of each influencing factor to the spatial variability of soil nutrients in plantation and natural forests, and this analysis was conducted using the “randomForest” package in R language (Yang X. et al., 2021).

To delve deeper into the mechanisms and pathways influencing the spatial variability of soil nutrients in plantation and natural forests, a structural equation model (SEM) was constructed to examine the direct and indirect effects of key influencing factors on soil N, available P, and pH. Observed variables were initially divided into composite variables and incorporated into structural equation models (SEMs). To validate the reliability of the relationship between key ecological factors and soil N, available P, pH, piecewiseSEM was employed to clarify the random effects of sampling points and provide the contributions of environmental predictive factors. These analyses were conducted using the “piecewiseSEM,” “nlme,” and “lme4” packages (Yang L. et al., 2021). The goodness of fit of the modeling results was assessed using the Fisher C test. Based on a significance level of p < 0.05 and satisfactory model fit (0 ≤ Fisher’s C/df ≤ 2 and 0.05 < p < 1.00), the model was progressively adjusted through a stepwise modification process (Figure 1).

The soil available P content in natural forests is approximately 9–15 mg/kg, while in planted forests, it is about 2–12 mg/kg. The available P content in the soil of natural forests is significantly higher than that in planted forests (p < 0.001), with the soil available P content in natural forests being about 71.4% higher than in planted forests. The soil pH range in natural forests is approximately 3–6, whereas in planted forests, it ranges from about 5–7.5. The soil pH in natural forests is significantly lower than that in planted forests (p < 0.001). However, there is no significant difference in soil N content between natural and planted forests (p > 0.05) (Figure 2).

With the increase in Mean Annual Temperature (MAT) and Mean Annual Precipitation (MAP), soil available phosphorus (AP) content significantly increased in both natural and planted forests (p < 0.001), while soil pH significantly decreased. The soil nitrogen (N) content in planted forests did not significantly change with the increase in MAT and MAP, whereas in natural forests, soil N content significantly decreased (p < 0.001). Additionally, under the combined effects of MAT and MAP, Mean Annual Evaporation (MAE) showed a positive correlation with soil N content and pH, but a negative correlation with soil available P content (Supplementary Figures S1–S3). With the increase in nitrogen deposition, the soil N content in both planted and natural forests increased, but the change was not significant; however, soil pH significantly decreased. The soil available P content in planted forests significantly increased with the increase in nitrogen deposition (p < 0.001), while there was no significant change in the soil P content in natural forests (p > 0.05) (Figure 3).

Overall, with the increase in stand density, average diameter at breast height, and stand age, the soil nutrients (N, available P, and pH) in planted forests did not show a significant trend of change (p > 0.05). In natural forests, soil nitrogen (N) content and pH significantly increased (p < 0.001), while soil available phosphorus (AP) content significantly decreased. The impact of stand factors on soil nutrients in natural forests was greater than in planted forests (Supplementary Figure 4).

Compared to climate factors and stand factors, key leaf traits have a relatively minor impact on both planted and natural forests. Overall, the explanatory power of key leaf traits on the spatial variability of soil nutrients is stronger in planted forests than in natural forests (Supplementary Figure 5).

The results of variance decomposition indicate that soil nutrients in both planted and natural forests exhibit strong environmental plasticity (explained R² ranging from 0.51 to 0.941). Compared to stand factors and key leaf trait factors, climate factors contribute the most to the spatial variability of soil nutrients in both planted and natural forests. This finding is further validated in the gbm (Gradient Boosting Machine) model—that is, the independent contribution of each climate factor is significantly greater than that of other types of influencing factors (Figure 4).

The results from the random forest show the significance of the average predictive value of each influencing factor within the random forest, further illustrating that climate factors are the key determinants in driving the spatial distribution of soil nutrients in planted and natural forests at macro scales (Figure 5). Heatmap analysis further explains the significant correlations between soil nutrients and various key factors (Figure 6).

The structural equation model indicates that for planted forests, climate factors can not only directly affect soil nutrient factors but can also indirectly influence soil nutrient factors by affecting key leaf traits, with their direct effects being greater than their indirect effects. However, for natural forests, climate factors only have a direct impact on soil nutrients, with the indirect effects being not significant (p > 0.05) (Figure 7).

We found that the available phosphorus content in the soil of natural forests is significantly higher than that in planted forests. In natural forests, long-term biomass accumulation and decomposition processes may lead to the accumulation of available phosphorus in the soil (Liu et al., 2012). The decomposition of fallen leaves, dead branches, animal remains, and other organic matter can increase the available phosphorus content in the soil (Liu et al., 2021). Additionally, natural forests typically have higher biodiversity, including a variety of plants, microbes, and animals (Stephens and Wagner, 2007). This diversity may promote more efficient available phosphorus cycling and utilization. We also found that the soil pH values in natural forests are significantly lower than those in planted forests. The decomposition of organic matter in natural forests may produce acidic substances, thereby reducing the soil pH value (Jin et al., 2022). These acidic substances include humic and fulvic acids. There was no significant difference in the nitrogen content of the soil between natural and planted forests, mainly because the sources of nitrogen, such as atmospheric deposition (ammonia and nitrogen oxides) and biological nitrogen fixation, are similar in both types of forests (Yang Y. et al., 2021). The nitrogen cycle in the soil is fast and easily influenced by environmental changes. Therefore, the nitrogen content is affected by various factors and does not reflect the differences in forest types as directly as the available phosphorus content does.

Temperature is a key factor influencing soil microbial activity, and microbes play an important role in the soil nutrient cycle (Gao et al., 2015). Higher temperatures typically promote the growth and metabolic activity of soil microbes, accelerating the decomposition of organic matter and releasing more mineral nutrients (such as nitrogen, phosphorus, potassium, etc.). Precipitation directly affects soil moisture content, which in turn influences the solubility and availability of nutrients in the soil (Wang L. et al., 2022; Wang Z. et al.,2022). Moderate precipitation helps dissolve minerals and organic matter in the soil, making it easier for plants to absorb nutrients. We discovered that with the increase in temperature and precipitation, the soil available phosphorus content in both planted and natural forests significantly increased, while pH significantly decreased. The increase in temperature and precipitation not only enhances biological activity and accelerates the decomposition of organic matter, thereby releasing more phosphorus and acidic substances, but also enhances the growth of plant roots and microbial activity, promoting the mineralization of soil available phosphorus, which aids in the more efficient acquisition of phosphorus from the soil (Lu et al., 2015). The increase in precipitation may enhance the leaching of alkaline ions in the soil, thus leading to a decrease in pH value (Wu et al., 2019). In recent years, an increasing number of studies have found that nitrogen deposition significantly affects soil nutrient content, with an increasing trend in soil nitrogen content in both planted and natural forests as nitrogen deposition increases (Tian et al., 2018; Zhu et al., 2015). However, we also made the surprising discovery that the soil phosphorus content in planted forests significantly increases with the increase in nitrogen deposition, while there is no significant change in the soil available phosphorus content in natural forests. Planted forests are usually composed of single tree species, which may have specific adaptability to the absorption and utilization of nitrogen. With increased nitrogen deposition, these tree species may absorb available phosphorus from the soil more effectively through enhanced root activity (Liao et al., 2024). Furthermore, root exudates (such as organic acids) may alter soil pH, thereby increasing the availability of phosphorus (Zhang et al., 2023). Nitrogen deposition also affects the structure and function of the soil microbial community, which plays a key role in the cycling of phosphorus (Ma et al., 2022). In planted forests, the increased nitrogen may promote the activity of certain microbes that can release more available phosphorus (Xia et al., 2020).

Overall, with the increase in stand density and stand age, no significant trend in the change of soil nutrients (N, available P, and pH) was observed in planted forests, whereas in natural forests, soil nitrogen (N) content and pH significantly increased, and soil available phosphorus (AP) content significantly decreased. A higher stand density implies increased competition among trees, which could lead to a reduction in nutrient absorption and cycling efficiency (Farooq et al., 2021). In planted forests, this competition might be mitigated by human management practices, such as thinning, thus resulting in no significant change in nutrients (Forrester, 2013). In natural forests, due to the more complex ecosystem, nutrient cycling may be more efficient, leading to a significant increase in soil N content and pH. Additionally, as stand age increases, soil microbial communities and activity may change (Kang et al., 2018), which are crucial for soil nutrient cycling. In planted forests, due to the monoculture of tree species and possible human intervention, this effect might be suppressed. In natural forests, over time, an increase in microbial diversity might promote the cycling of N and increase in pH (Wang et al., 2023).

Compared to climate and stand factors, the impact of key leaf traits on soil nutrients in planted and natural forests is relatively minor. Overall, the key leaf traits in planted forests have a stronger explanatory power for the spatial variability of soil nutrients than in natural forests. Planted forests often undergo more human management, such as regular thinning and fertilization. These management practices may enhance the association between tree leaf traits (such as leaf area, chlorophyll content, photosynthetic efficiency, etc.) and soil nutrients (dos Santos et al., 2006). The growth of natural forests relies more on natural conditions, where interactions are more complex and uncertain. Additionally, in natural forests, different tree species may occupy different ecological niches, each adapting to specific soil nutrient conditions (Gu et al., 2017). This niche differentiation leads to a more diversified relationship between leaf traits and soil nutrients spatially (Sterck et al., 2011). In contrast, due to the lack of such differentiation in planted forests, the leaf traits of tree species are more likely to directly reflect uniform soil nutrient conditions.

The research further reveals that on a macro scale, the soil nutrients in both planted and natural forests exhibit strong environmental plasticity, with climate factors contributing more to the spatial variability of soil nutrients in planted and natural forests compared to stand factors and key leaf traits. Climate is a major driver influencing soil nutrient cycling. Temperature and precipitation directly affect soil microbial activity and the rate of organic matter decomposition, which in turn influences nutrient release and cycling (Yuan et al., 2021). Due to significant variations in climate factors on a macro scale, their impact on soil nutrients often surpasses that of local stand or vegetation characteristics. Climate conditions directly affect plant growth patterns, photosynthetic efficiency, and water use efficiency, which in turn affects the demand for and efficiency of soil nutrient utilization by plants (Xu et al., 2023). On a macro scale, variations in these factors may be more significant than differences within a stand. The major contribution of climate factors to the spatial variability of soil nutrients in planted and natural forests on a macro scale is primarily due to the direct impact of climate conditions on soil biochemical processes and plant growth, and these impacts are more significant and widespread on a large scale than the effects of stand structure and plant leaf traits (Gong et al., 2023).

The direct effects of climate factors on shaping the variability of soil nutrients in planted and natural forests are greater than their indirect effects. The direct impact of climate on soil nutrients is usually more rapid and intense. For example, changes in temperature and humidity can significantly alter microbial activity and nutrient cycling rates in a short period (Zhou et al., 2012). In contrast, the impact of climate on leaf traits is a more complex and time-delayed process (Wu et al., 2015). Changes in leaf traits first affect plant physiological processes, and then these changes indirectly affect soil nutrient cycling. While the indirect effects of climate factors on soil nutrients are also important, their impact is generally more complex, time-delayed, and more concentrated on local plant–soil systems (Faucon et al., 2017).