Homeostatic evidence of management-induced phosphorus decoupling from soil microbial carbon and nitrogen metabolism

Abstract

The theory of microbial stoichiometry can predict the proportional coupling of microbial assimilation of carbon (C), nitrogen (N), and phosphorus (P). The proportional coupling is quantified by the homeostasis value (H). Covariation of H values for C, N, and P indicates that microbial C, N, and P assimilation are coupled. Here, we used a global dataset to investigate the spatiotemporal dynamics of H values of microbial C, N, and P across biomes. We found that land use and management led to the decoupling of P from C and N metabolism over time and across space. Results from structural equation modeling revealed that edaphic factors dominate the microbial homeostasis of P, while soil elemental concentrations dominate the homeostasis of C and N. This result was further confirmed using the contrasting factors on microbial P vs. microbial C and N derived from a machine-learning algorithm. Overall, our study highlights the impacts of management on shifting microbial roles in nutrient cycling.

摘要

人为管理措施诱导土壤微生物磷与土壤微生物碳氮内稳态解耦的代谢证据

微生物化学计量学理论能够预测微生物对碳(C)、氮(N)、磷(P)的同化作用具有比例耦合性。这种耦合性被量化为稳态值，用于计算内稳性系数(H)。C、N、P的H值协变量表明，微生物对C、N、P的同化作用是耦合的。因此，在本文中我们利用全球数据集研究微生物C、N、P的H值在不同生物群落中的时空动态变化。研究结果表明，土地利用和人为管理措施导致P 与C、N的代谢在时间和空间上解耦，并且我们利用结构方程模型(SEM)分析结果表明，土壤因子主导P的微生物动态平衡，而土壤元素含量则主导C和N的微生物动态平衡。利用机器学习算法得出的微生物P与微生物C和N的对比因子也证实了这一结果。总的来说，我们的研究强调了在营养循环中人为管理措施对微生物角色转变具有重要的影响。

INTRODUCTION

Although microbial biomass comprises ~1% of soil organic matter (Xu et al. 2013), it plays an important role in carbon (C), nitrogen (N), and phosphorus (P) cycling in terrestrial ecosystems (van der Heijden et al. 2008; Zechmeister-Boltenstern et al. 2015; Spohn 2016). Microbial nutrient demand is jointly determined by microbial growth and the stoichiometric imbalance between microbial biomass and the nutrient availability in the environment, which is defined as homeostasis (Sterner and Elser 2002; Sistla and Schimel 2012; Hessen et al. 2013; Garcia et al. 2021). Microbial homeostasis is essential for regulating C, N, and P cycles in terrestrial ecosystems (Cleveland and Liptzin 2007; Sinsabaugh et al. 2009). Ecosystem-level C, N, and P stoichiometry have received considerable attention in recent years (Zechmeister-Boltenstern et al. 2015; Delgado-Baquerizo et al. 2017), and there have been studies on homeostasis of animals (Frost et al. 2006), plants (Giling et al. 2012), and soil microbial enzymes (Sinsabaugh et al. 2009). However, the dynamics of microbial homeostasis in C, N, and P cycles remains elusive (Wakelin et al. 2013; Leff et al. 2015), despite growing interest in the role of microbes in soil stoichiometry (Zechmeister-Boltenstern et al. 2015).

Human activities and climate change have fundamentally altered C, N, and P cycles (Greaver et al. 2016), often leading to more accelerated and loosely coupled cycles (Steffen et al. 2004). A better understanding of microbial stoichiometry in the soil is essential for evaluating these cycles. However, because of high spatiotemporal variation, we still know little about the patterns and underlying mechanisms of microbial homeostasis (Xu et al. 2013, 2015; Yu et al. 2021). In particular, the influence of varying environments on C, N, and P homeostasis in soil microbes provides an opportune way to investigate the coupling/decoupling of nutrients in microbes (Cavicchioli et al. 2019). For example, the influence of landscape management and disturbance can have an important, but poorly understood, influence on the strength of microbial C, N, and P stoichiometric homeostasis (H) (Berthrong et al. 2013; Chen et al. 2022).

In this study, we calculated H values of soil microbial C, N, and P, which encapsulate the strategies of living organisms for coping with limiting soil resources (Sterner and Elser 2002). The values of H can be readily quantified by growing an organism across a wide range of elemental ratios (x), measuring the organism’s resulting elemental composition (y) and then plotting the log-transformed values of each to find the slope (1/H) of the resulting relationship, as follows:

where x and y represent the elemental concentrations (the ratio of element density to dry soils) in soils and microbes, respectively (Xu et al. 2015), rather than the elemental ratios (Sterner and Elser 2002). Anthropogenic activity and climate change alter soil microbial activity and nutrient availability, e.g. through land conversion, N input, and water management, as a result, determining if microbial responses to soil nutrients are consistent with their H values under different soil conditions could inform predictions for how soil microbes, and the ecosystems in which they are embedded, respond to global change and anthropogenic activities (Ramirez et al. 2012; Leff et al. 2015). To address the predictive ability of H of plants in response to their soil resources, Yu et al. (2011) estimated H for foliar N (H_N), P (H_P), and N:P (H_N:P) of 11 plant species, which helped distinguish species. Here, however, we rather addressed community-level homeostasis patterns in soil microbial C, N, and P. To do so, we extend the meaning of the H value through its temporal and spatial variation by compiling 4152 data points of soil microbial C, N, and P from the literature (see Supplementary Material).

Our primary objective was to compare the difference in the H values of C, N, and P in managed ecosystems (e.g. pasture and croplands) versus more natural ecosystems (e.g. forests, shrublands, grasslands). That is, we explored how the stoichiometric flexibility of the organisms regulates the plasticity of macro-scale ecosystem processes. Specifically, we aimed to test two hypotheses: (i) microbial homeostasis varies at both spatial and temporal scales and (ii) management promotes the divergence of microbial C, N, and P homeostasis, indicating an elemental decoupling in microbial mechanisms. We expected that human activities would increase biological activity in the soil and, therefore, the availability of nutrients that are more biologically influenced (mainly C and N). At the same time, we expected that human activities would reduce the relative dominance of nutrients linked to the geochemical processes of P cycling, causing a stoichiometric imbalance in the nutrient cycles associated with C and N (Tagliabue et al. 2011; Kong et al. 2018). We calculated the H values of microbial C, N, and P to evaluate differences between environments with high versus low microbial growth.

MATERIALS AND METHODS

Data analyses

To examine variation in elemental H values of microbes in different environments, we built on our previously published literature compilation (Xu et al. 2013, 2015) by extending the search until December 2019. Our criteria for screening the data remained the same as our previous compilation and was (i) the soil microbial biomass (at least one of soil microbial biomass C, N, or P) must be reported; (ii) the reported soil microbial biomass C, N, and P must be less than soil organic C, total N, and total P, respectively. In total, we retrieved 4152 data points across 14 global biomes from 432 studies published between the 1970s through 2019. Most studies provided geographic coordinates, which we used to retrieve long-term climate conditions for the study site. For those data points where geographic coordinates were not reported, we searched for coordinates based on the names of sites, states, and countries. For each site, we retrieved climate data (soil temperature, ST; soil moisture, SM; air temperature, AT; precipitation, Pr) from the 1961–90 average provided by the University of East Anglia Climatic Research Unit (http://www.cru.uea.ac.uk/). We recorded vegetation distribution data from each site using several data sources: global pasture and cropland data were taken from Ramankutty et al. (2008); wetland distribution data were taken from Aselmann and Crutzen (1989); the spatial distribution of biomes other than wetlands were taken from Ramankutty and Foley (1998). We aggregated studies into 12 major biome types, and then categorized them as managed ecosystems (including cropland and pasture) and natural ecosystems (including boreal forest, temperate coniferous forest, temperate broadleaf forest, tropical/subtropical forest, mixed forest, grassland, shrub, tundra, and desert). For each sampling site, we also extracted other relevant variables, including soil pH, sampling depth, sampling date, biome type, latitude, longitude, and climate variables. For each study, data were collected only in surface soils, primarily within 0–15 cm, but some were 0–30 cm; we assumed that all measurements were representative of the top 0–30 cm soil profile as in previous compilations of global soil organic C and total N (Batjes 1996; Estenan et al. 2000).

With these data, we then calculated the strength of stoichiometric homeostasis (H) of microbial C, N, and P according to the model (Sterner and Elser 2002),

where y is the microbial C, N, and P concentration (mmol kg⁻¹), x is the soil C, N, and P content (mmol kg⁻¹), and d is a constant. We obtained H and d using linear regression. For temporal data, we took the C, N, and P values as the concentration in every site every year (sampling time: 1997–2017). For each year, there are at least 15 data points; if there were fewer than 15 data points for a year, we deleted that year’s value to keep the value of the year’s H. Across sites, we classified the element concentration of C, N, and P into 16 latitudinal bands at 5-degree intervals and calculated H values of C, N, and P from each band.

We also calculated the H values of each biome type (Supplementary Fig. S1) and the average concentration of total soil and microbial biomass element pools (Figs 2 and 3, S3). We further divided biome types into two broad categories (managed ecosystems and natural types), to evaluate whether H values differed among biome types when we found a significant difference (P < 0.05) between the regression slopes (1/H) for different biomes using analysis of covariance. We used one-way analysis of variance to examine differences in microbial and soil element concentration among managed and natural ecosystems. Prior to analyses, we log₁₀-transformed nutrient concentrations to improve the distribution and homogeneity of variance, but the back-transformed means into the original units for reporting.

Machine-learning method with the Gray CoRrelation Analysis

We used the Gray CoRrelation Analysis (GCRA) machine-learning method (Tsai and Hsu 2010) to integrate the objectives and provide a relative contribution of the environmental factors. Gray system theory puts forward the concept of relational degree analysis, which analyzes the central relationship between the individuals in the system through a particular method, finds out the most critical factors affecting the system, and grasps the main aspects of the contradiction (Tsai and Hsu 2010). Specifically, it describes the relative changes among individuals in the system’s development process (i.e. the relativity of the changing size, direction, and velocity). The basic idea of the Gray relational analysis is to determine whether the geometric shapes of sequence curves are closely related according to their similarity. Using the GCRA method, we set the H values of C, N, and P in managed and natural ecosystems as the parent sequence and the other five environmental factors (including microbial—L_mic, climate—L_cli, edaphic—L_eda, nutrient—L_nutr, and pH—L_pH) as the subsequences. Each subsequence is divided into two time periods: 1978–97 and 1998–2017. From this, we calculate the relative correlation between environmental factors (L_mic, L_cli, L_eda, L_nutr, and L_pH) and each H value, and then rank all environmental factors based on their relative contribution to every H value (Table 1).

First, we devise the sequence of the system as follows:

where X0 is the parent sequence, Xi is the subsequence, and Xi(k) is the observed data of factor Xi at time k. The calculation of the Gray relational degree generally includes the following steps: (i) transformation of the original data; (ii) identify the differential sequence; (iii) calculate the maximum and minimum difference between the poles; (iv) calculate the correlation coefficient; (v) find the correlation coefficient. The details are as follows:

• Step 1: Transforming the raw data

The dimensions (or units) of factors in a system are not necessarily the same. Therefore, dimensions (or teams) must be eliminated from the original data and converted into comparable sequences. Take the initial valued transformation as an example, then

• Step 2: Finding the difference sequence

• Step 3: Finding the difference between the maximum and the minimum difference

• Step 4: Finding the correlation coefficient

The parent sequence of recorded data transformation is {X0(t)}, and the subsequence is {Xi(t)}. At time k, the correlation coefficient L0i(k) between the parent sequence {X0(k)} and the subsequence {Xi(k)} can be calculated using the following formula:

The correlation coefficient reflects the closeness (tightness) of the two sequences being compared at a given moment. For example, at the Δmin moment, γ_0i = 1, and at the Δmax moment, the correlation coefficient is the minimum value. Therefore, the range of the correlation coefficient falls 0 < γ ≤ 1.

• Step 5: Calculating the strength of the association

Structural equation model

We used a structural equation model to disentangle the causal pathways through which distribution and environmental parameters influence the global H values of C, N, and P. To do so, we classified factors into five groups: climate (ST, SM, AT, and Pr), edaphic (sand, clay, and silt), soil nutrient (soil organic carbon, SOC; total nitrogen, TN; total phosphorus, TP), microbes (microbial carbon, MBC; microbial nitrogen, MBN; microbial phosphorus, MBP), and pH. These models consider how environmental factors affect the H value of C, N, and P via direct or indirect effects.

RESULTS

Soil and microbial biomass C, N, and P in managed versus natural ecosystems

The C, N, and P contents in soil microbial biomass (Fig. 1a–c) and soil (Fig. 1d–f) in natural ecosystems were significantly higher than those in managed ecosystems. In managed ecosystems, MBC varied from 0.011 to 349.5 mmol kg⁻¹, with an average of 27.86 mmol kg⁻¹ (CV = 107.8%, Fig. 2a); MBN ranged from 0.013 to 32.79 mmol kg⁻¹, with an average of 3.29 mmol kg⁻¹ (CV = 102.8%, Fig. 2b); and MBP ranged from 0.0025 to 3.86 mmol kg⁻¹, with an average of 0.50 mmol kg⁻¹ (CV = 113.3%, Fig. 2c). In natural ecosystems, MBC varied from 0.0017 to 465 mmol kg⁻¹, with an average of 70.97 mmol kg⁻¹ (covariance coefficient; CV = 120.9%, Fig. 3a); MBN ranged from 0.025 to 42.36 mmol kg⁻¹, with an average of 8.4 mmol kg⁻¹ (CV = 106.6%, Fig. 3b); and MBP ranged from 0.0085 to 9.03 mmol kg⁻¹, with an average of 1.90 mmol kg⁻¹ (CV = 113.9%, Fig. 3c).

In managed ecosystems, SOC varied from 50.83 to 7716.67 mmol kg⁻¹, with an average of 1459.9 mmol kg⁻¹ (CV = 82.9%, Supplementary Fig. S2a); soil TN ranged from 3.57 to 857.14 mmol kg⁻¹, with an average of 115.69 mmol kg⁻¹ (CV = 93%, Supplementary Fig. S2b); TP ranged from 0.11 to 127.42 mmol kg⁻¹, with an average of 20.95 mmol kg⁻¹ (CV = 96.1%, Supplementary Fig. S2c). In natural ecosystems, SOC varied from 1.85 to 20 416.7 mmol kg⁻¹, with an average of 4378.1 mmol kg⁻¹ (CV = 103.6%, Supplementary Fig. S3a); TN ranged from 0.79 to 17 799.7 mmol kg⁻¹, with an average of 340.61 mmol kg⁻¹ (CV = 248.8%, Supplementary >Fig. S3b); TP ranged from 0.11 to 88.58 mmol kg⁻¹, with an average of 22.20 mmol kg⁻¹ (CV = 72%, Supplementary >Fig. S3c). The TN in the soil was much more variable than the SOC and TP in nature ecosystems (Supplementary >Fig. S3a–c). The mean and variance of C, N, and P concentrations in soil and soil microbial biomass were lower in managed ecosystems than in natural ecosystems (Fig. 1).

Variation of microbial C, N, and P homeostasis in managed versus natural ecosystems

We found that the H values of microbial C increased from 1978 to 2017 in both managed and natural ecosystems (Fig. 4a). Likewise, we found that the H values of microbial N also increased through time in both managed and natural ecosystems (Fig. 4b). We also found increases in the H values of P in the natural ecosystem through time, but a disjoint relationship of the H values of P in managed ecosystems through time (Fig. 4c). Specifically, in managed ecosystems we found a highly significant increase between H values and P through time from 1978 to 1997, and a highly significant increase between H values of P through time from 1998 to 2017. The finding indicates that to some extent, soil MBP might be a homeostatic system in managed compared with natural ecosystems. That is to say, the ability of soil microbial C, N, and P to control the variation of elements also increases with time, which is consistent in managed and natural ecosystems.

We analyzed the geographical variation of homeostasis for each element and found a single-peak curve of homeostasis with latitude; first increasing from the equator to mid-latitude and then declining to the poles (Fig. 4e and f). This relationship was consistent in both managed and natural ecosystems, but the pattern was relatively stronger for P in managed ecosystems (Fig. 4f). In addition, negative quadratic relationships were observed between latitude and C, N, and P in managed and natural ecosystems (Fig. 4e and f). And more significant for the P in managed ecosystems (Fig. 4f).

Environmental controls on the microbial homeostasis of C, N, and P

Using the machine-learning method with the GCRA algorithm, we found that a clear P decoupling from C and N in managed ecosystems (Table 1). The GCRA results were consistent with the temporal P-decoupling patterns in managed ecosystems and coupling variation patterns in natural ecosystems (Table 1). We found the strongest influence of microbial factors followed by climate and soil pH. The influence of edaphic factors and nutrients was small. However, we did find differences in the influence of environmental factors on H values of P compared with that of C and N. There were two distinct stages in H value variability of C, N, and P, indicating C–N coupling and P decoupling (Table 1). For example, during 1978–97, the contribution order of controlling factors was: L_pH > L_mic > L_cli > L_nutr > L_eda, for the H value of both C and N. However, the contribution order of P was: L_cli > L_nutr > L_pH > L_mic > L_eda. This indicates that in the early stages, the most important controlling factor on the homeostasis of C and N was pH, while climate was predominantly influential for the homeostasis of P. We also found that the contribution of environmental variables to the homeostasis of C and N was consistent, but differed for P. During the 1998–2017 period, the order of the contributions was: L_pH > L_mic > L_nutr > L_eda > L_cli, for the H values of C and N, but L_pH > L_nutr > L_eda > L_mic > L_cli for the H values of P (Table 1). The order of the contributions of environmental variables to P also showed P-decoupling changes to C and N. We found a predominance of the importance of pH for the homeostasis of C, N, and P. S soil nutrients were the second key factor for P, while microbes were the second key factor for C and N during the 1998–2017 time period.

DISCUSSION

Rather than using the C:N:P ratio (Elser et al. 2010; Tapia-Torres et al. 2015, 2018; Halvorson et al. 2019; Buckeridge and McLaren 2020; Zhang et al. 2020), we use C, N, and P concentrations to calculate H values (Elser et al. 2010; Yu et al. 2011; Hessen et al. 2013; Li et al. 2016; Spohn 2016; Wang et al. 2018; Su et al. 2019). This is because there was no correlation between the soil C:N:P and microbial C:N:P, but there was a strong relationship between the elemental concentration in microbes and soil (Supplementary >Fig. S4; Griffiths et al. 2012). Nutrient regulation of soil microbial feedbacks to disturbances such as land-use can strongly influence nutrient cycling (Bardgett et al. 2008; Sinsabaugh and Follstad Shah 2012). This happens primarily through changes in the substrate stoichiometry induced by variations in microbial biomass (Ågren 2004). Our results show that the H values of C, N, and P varied spatially and temporally in natural ecosystems with various climate and soil factors. Importantly, however, we found that the homeostasis of P was decoupled from C and N in managed ecosystems across space and time; this decoupling phenomenon can be verified by the H value of P, C, and N in managed ecosystems (Supplementary >Fig. S1). This was most likely because of controlling factors impacted on the cycling of C, N, and P, which we discuss in more detail below.

Environmental controls on homeostasis of C, N, and P in microbes

In managed ecosystems, climate, pH, nutrients, and microbial biomass directly affected H values of C and N (Fig. 5a and b). In contrast, climate, pH, edaphic factors, and microbial biomass directly affected H values of P (Fig. 5c). That is, edaphic factors replaced nutrients as the direct factor affecting H values of P in managed ecosystems (Fig. 5a–c). In the face of disturbances inherent to managed ecosystems, C and N processes should recover more rapidly than P processes (Mooshammer et al. 2017) if they are subjected to continuous external interference. So, in the managed ecosystems, accompanied by the persistent disturbances of anthropogenic activity, the recoverability of P metabolisms would be further weakened, while the recovery ability of C and N is not affected due to their quick metabolism ability (Mooshammer et al. 2017). This results in H values of P shifting to limitation by soil edaphic factors rather than nutrients. But, in natural ecosystems, climate, pH, edaphic, nutrients, and microbial directly affected the H value of C, N, and P (Fig. 5d and e). Without external anthropogenic interference, the metabolism of C, N, and P was directly controlled by these factors simultaneously. The indirect effect of a nutrient on H values of all elements by microbe indicated that increases in soil microbial biomass depend on the abundance of sufficient soil nutrients to maintain the homeostasis of microbial elements required for managed and natural ecosystems. These mechanisms also showed the decoupling of P in managed ecosystems when compared with natural ecosystems.

Mechanisms underlying microbial P decoupling from C and N in managed ecosystems

On the basis of our global synthesis of the empirically derived insights into the important role of microbes in nutrient cycling, we propose a conceptual framework that summarizes the possible mechanisms underlying the P-decoupling homeostasis with C and N in the face of anthropogenic activities (Fig. 6). The Spearman correlation between soil element concentration and microbial concentration was highly significant and positive (for managed ecosystems, C: R² = 0.29, N: R² = 0.27, P: R² = 0.03; for natural ecosystems, C: R² = 0.45, N: R² = 0.27, P: R² = 0.17, all P < 0.0001) and presented a strongly isometric shape except the P in the managed ecosystems (Fig. 6a). The correlation between soil P content and microbial P content in managed ecosystems was significantly positive (R² = 0.03, P < 0.0001). The strong homeostatic relationship between soil microbial biomass and each element in nature supported the notion that soil and microbial C, N, and P are strongly coupled in natural ecosystems. This emphasizes that natural ecosystems have a strong potential to enrich low-concentration elements such as P in microbial biomass. However, managed ecosystems have a weaker potential to enrich low-concentration elements. A possible reason for the decoupling of P from C and N in managed ecosystems might be attributed to variations in the spatial distribution of microbial biomass. By which differences in the spatial distribution of C, N, and P strength the decoupling of H values of P, particularly over the long-time anthropogenic activity in managed ecosystems.

We suggest that the mechanisms for the management-promoted P decoupling from C and N can be explained by the C, N, and P turnover rates. Specifically, turnover rates of microbial P are typically higher than that of C and N (Bicharanloo et al. 2020). Because of the higher microbial turnover rates in managed ecosystems, P accumulates less in microbial biomass and thus varies less with external variation than does C and N (such as irrigation, fertilization, and other management practices in managed ecosystems) (Bicharanloo et al. 2020). Likewise, we found that microbial P was less susceptible to changing soil P in managed ecosystems, providing further explanation for P decoupling from C and N.

There are several likely mechanisms underlying the decoupling of P from C and N in managed ecosystems. First, low microbial P demand due to slower growth and/or decreased microbial biomass in managed ecosystems could have led to lower production of phosphatases relative to that in natural ecosystems (Mooshammer et al. 2017). When C sources are present in the soil, microbes can utilize the carbon sources as an energy and carbon substrate, allowing them to increase their metabolic activities, including the production of organic acids, enzymes, and other compounds that can solubilize P more actively (Hameeda et al. 2006; Patel et al. 2008). The metabolism of energy-rich compounds can allow the necessary energy to decompose more recalcitrant fractions of the soil organic matter (Craine et al. 2007). Furthermore, the mineralization of organic P is a C- and N-consuming process based on the production and release of N-rich exoenzymes (Spohn 2016). In a wide range of managed ecosystems, Marklein and Houlton (2012) showed that N inputs to soil accelerated the activity of phosphatases, indicating that microbial communities allocate more C and N to P acquisition when there is higher N input. Since the release of bound inorganic P is caused by organic acids and siderophores that contain C and N (Jones and Oburger 2011), the solubilization of P should also underlay stoichiometric constraints.

Second, the decomposition of organic matter is influenced by the degree to which ecosystems are managed. For example, under management, there can be less available C and N relative to P during decomposition, which would lead to unbalanced stoichiometry and reduced microbial activity and diversity (Mooshammer et al. 2014; Ning et al. 2022). Although C, N, and P cycling is tightly coupled through microbial immobilization and mineralization (Finzi et al. 2011), microbial P cycling can decouple from C and N following strong disturbances (Mooshammer et al. 2017). Under persistent anthropogenic activities, the reduced availability of C and N may thus lead to an imbalance in the concentrations of C, N, and P. Given the different cell metabolisms in C, N, and P, C, N and P might weaken their coupling, particularly over a long-time scale (Berthrong et al. 2013). Despite the fact that cycling of C, N, and P is tightly coupled through microbial immobilization and mineralization (Finzi et al. 2011), following strong disturbances, microbial P cycling is decoupled from those of C and N as a consequence of differential stresses (Mooshammer et al. 2017).

Finally, the decoupling of P relative to C and N can also be influenced by the cycles of each element. C and N can exist in gaseous compounds, while P is primarily sedimentary (Moses and Pidwirny 2010). As a result, C and N can escape to the atmosphere (Fig. 6b, the green and gray arrows), while the P cycle is more closed (Fig. 6b, the light brown arrows). Although plausible, this mechanism for the P decoupling from C and N is just a possibility and will need experimental and observational data for verification. Thus, understanding and testing hypotheses related to C, N, P, and P decoupling are not simply an academic exercise; they have implications that extend well beyond the decoupling and interactions themselves.

CONCLUSIONS

Overall, our study advances our understanding of the decoupling of P from C and N cycling through metabolic activities in the soil. Given the homeostasis of P in microbial biomass, this also contributes to the correlation between soil nutrients and microbial biomass. We argue that the unifying mechanisms underlying C, N, and P dynamics play an important role in understanding microbial control on soil nutrient cycling and therefore deserve further investigation. C and N primarily enter terrestrial ecosystems through photosynthesis, N fixation, and rainfall deposition, while P enters the system solely from the weathering of parent materials as well as anthropogenic P inputs. Our findings also suggest that the projected intensifying human activity will likely lead to the homeostasis of biologically controlled C and N in the soil, while P could become more decoupled from other nutrients. These homeostatic characteristics will contribute to our understanding of microorganisms and their macro-ecological homeostasis that shapes microbial biogeography at global scales.

Supplementary Material

Supplementary material is available at Journal of Plant Ecology online.

Figure S1: The difference in the H value of carbon, nitrogen, and phosphorus among vegetation types in natural ecosystems and managed lands.

Figure S2: The distribution of total soil organic carbon, total nitrogen and total phosphorus in managed (a-c) ecosystems.

Figure S3: The distribution of total soil organic carbon, total nitrogen and total phosphorus in natural (a-c) ecosystems.

Figure S4: Correlation of carbon, nitrogen, and phosphorus in soil and microbial biomass.

Funding

This study was partially supported by the National Natural Science Foundation of China (32271681) and the Joint Funds of the National Natural Science Foundation of China (U2006215). L.Z. was partially supported by Key Laboratory of Ecology and Environment in Minority Areas (Minzu University of China) and National Ethnic Affairs Commission (KLEEMA202206). X.X. acknowledged the financial assistance provided by the National Science Foundation (2145130).

Conflict of interest statement. The authors declare that they have no conflict of interest.

REFERENCES

Author notes