Changes in soil microbial metabolic limitations after half-century forest restoration in degraded tropical lands

Due to increasing anthropogenic pressure, over half of the world’s tropical forests are reforested or afforested secondary forests or plantations. The recovery pace and potential of these forests depend largely on soil microbially-mediated biogeochemical cycling. Here we measured soil extracellular enzyme activities and quantified microbial metabolic limitations using a vector analysis in a bare land (BL, representing the original state before restoration), two afforested sites [i.e. a restored secondary forest (MF) and a managed Eucalyptus exserta plantation (EP)] and a nearby undisturbed forest (UF) in south China. Results showed that soil microbial metabolisms were co-limited by carbon (C) and phosphorus (P) across the four forests. Both microbial C and P limitations were higher in BL than UF. Microbial C limitation significantly reduced after restoration only in MF when compared to BL, but it was still higher than that in UF. Interestingly, microbial P limitation significantly enhanced after restoration in both EP and MF when compared to BL, and it did not differ between the two restored forests. Structural equation modeling (SEM) showed that microbial C limitation was primarily attributed to microbial C use efficiency, while microbial P limitation was co-driven by plant biomass, microbial C use efficiency and soil P availability. These findings suggest microbial C limitation could be gradually recovered after forest restoration in southern China, which would facilitate soil organic carbon accumulation. However, the enhanced microbial P limitation after forest restoration underlines the necessity to develop optimal P management in these restored forests.


Introduction
Tropical forests host an estimated 50% of terrestrial global biodiversity and account for roughly 35% of terrestrial net primary productivity (Jobbagy andJackson 2000, Pan et al 2011).Despite such importance, tropical primary forests are undergoing widespread loss due to increasing anthropogenic pressure.To combat these ongoing losses, many tropical degraded lands were reforested or afforested secondary forests or plantations, and the the extent of area has already overtaken that of tropical primary forests (Poorter et al 2016).Understanding regrowth trajectories of these restored forests after human disturbance is therefore essential to predict future land-cover change, regional forest productivity, tree species distributions and C cycle-climate feedbacks (Canadell and Raupach 2008).However, the recovery pace and potential of these restored Forest depend largely on soil microbially-mediated biogeochemical cycling (Crowther et al 2014).So far, information regrading the effects of forest restoration on microbially-mediated biogeochemical cycling remains very limited.
Soil microbial communities are integral parts of terrestrial ecosystems, contributing significantly to nutrient cycling processes and overall soil fertility.Microorganisms play a critical role by secreting extracellular enzymes to activate essential nutrients, thereby alleviating their metabolic limitations (Sinsabaugh et al 2010).The metabolic activities of microorganisms not only provide nutrients to plants, but also compete with plants for nutrients when nutrients are scarce (Inselsbacher et al 2010, Sinsabaugh andShah 2011).In oligotrophic ecosystems, nutrient competition between microbes and plants is more pronounced (Cui et al 2018).Therefore, microbial metabolic limitation can serve as an important indicator of soil nutrient availability status.Soil enzyme chemometrics can be used to assess the demand and utilization efficiency of different nutrients by soil microbes, and it has been widely used to characterize the relative limitation of energy and nutrients for soil microbes (Peng and Wang 2016, Zhou et al 2020).Sinsabaugh et al (2010) conducted a meta-analysis and found that globally, the ratio of logarithmically transformed activities of carbon (C), nitrogen (N), and phosphorus (P) acquisiting enzymes tends to approach 1:1:1.Deviations from this ratio indicate limitations imposed by C, N, or P Moorhead et al (2013,2016) proposed an enzyme stoichiometry vector analysis model.It quantifies microbial demand for C, N, and P by plotting the allocation proportions of enzyme activities as vector lengths and angles.An increase in vector length corresponds to an increase in the demand for C acquisition enzymes, while a steeper vector angle indicates an increase in the demand for P acquisition enzymes (Moorhead et al 2013).The enzyme chemometric analyses has been commonly used to assess the characteristics of microbial metabolic limitations represented by C, N, or P (Tapia-Torres et al 2015, Cui et al 2018, Xu et al 2022, Zhang et al 2023).
Soil microbial metabolic limitation may be influenced by multiple abiotic and biological factors (Li et al 2022, Zhao et al 2024).Studies have shown that climatic factors such as precipitation and mean air temperature can affect the metabolic restriction of soil microorganisms (German et al 2012, Xu et al 2017, Cui et al 2019).Additionally, soil physical and chemical properties, such as pH and moisture, have direct and indirect effects on metabolic activity of microorganisms by altering the concentrations and availability of C, N, and P in the soil (Peng and Wang 2016, Romanowicz et al 2016, Chen et al 2019, Qiu et al 2021).Soil microbial metabolic limitations may also be affected by biological factors, such as aboveground vegetation characteristics (He et al 2020), vegetation biomass (Zhao et al 2014), litter input (Alves et al 2010), and litter C:N ratio (Zhou et al 2019).These factors can affect the nutrient acquisition strategies of microorganisms and thus alter their metabolic limitations.During long-term vegetation recovery processes, changes in environmental factors such as soil nutrients and vegetation types may directly or indirectly impact the metabolic limitations of soil microbes.Deng et al (2019) found that changes in microbial metabolic restriction are the result of interactions of plants, soil, and microorganisms.Although the impacts of biological and abiotic factors on soil microbial metabolic limitations after restoration have been extensively studied, they remain controversial due to differences in vegetation type and geographical location.The key factors that control microbial metabolic limitation following tropical forest restoration remained unknown.
The objective of this study was to evaluate the effects of forest restoration on soil microbial metabolic limitations in a degraded tropical region of south China.We addressed this issue in a bare land (BL) as a control and two restored forest sites over 63 years: a managed Eucalyptus plantation (EP) and a restored secondary forest (MF).A nearby undisturbed primary forest (UF) was also selected to test the recovered potential of microbial metabolic limitations after restoration.We measured soil extracellular enzyme activities (EEAs) and quantified microbial metabolic limitations using a vector analysis in all four forests.We also used regression and SEM to identify the potential controls on microbial metabolic limitations and associated drivers.We seek to address the following two questions: (1) How does soil microbial metabolic limitation vary after forest restoration?
(2) What are the key factors that control microbial metabolic limitation after forest restoration?Findings from this study will deepen our understanding of the ecological implications and consequence of restoration activities on soil health, and provide crucial insights for developing optimal nutritent managements to accelerate forest restoration pace.

Study sites
The study was conducted at the Xiaoliang Tropical Coastal Ecosystem Research Station, Chinese Academy of Science, in Guangdong Province, China (21 • 27 ′ N, 110 • 54 ′ E).The annual average temperature is 23.9 • C and the mean annual precipitation is 1637 mm from 2012 to 2023 based on the long-term monitoring of Xiaoliang Tropical Coastal Ecosystem Research Station.The soil is classified as a latosol and formed from granite (Yao et al 1984).The highest elevation is about 50 m (Yu and Pi 1985).Until the 1950s, the area was covered by evergreen broad-leaved seasonal rainforests.However, due to excessive human interference, an area of nearly 400 km 2 suffered severe erosion, resulting in significant erosion of the topsoil layer, exposure of the subsoil, and lack of vegetation cover on the surface.The fertility of the soil drastically declined.Most of the natural vegetation in the area was destroyed, with only a small portion of pristine forest (UF) preserved for over 200 years (Yu andPi 1985, Ren et al 2007).
In order to restore degraded land, scientists launched a afforestation campaign in 1959 and established two types of artificial forests and one BL as a control in three geographically similar catchments.The two forests consist of a managed Eucalyptus exserta plantation (EP, 3.9 ha) and a secondary mixed forest (MF, 3.8 ha).A Eucalyptus EP was established on BL in the early 1960s.Forests using mixed native species (MF) were promoted from EP after clearcutting in 1974 and are dominated by Carallia brachiata, Aphanamixis polystachya, Schefflera octophylla, Carallia brachiata, Symplocos chunii, Acacia auriculiformis, Photinia benthamiana, and Cinnamomum burmanni, Lygodium japonicum, Ophiopogon japonicus, and Nephrolepis cordifolia (Mao et al 1992).We also selected a nearby undisturbed forest as a reference site, the forest has been preserved by residents for approximately 200 years.UF is dominated by Sterculia lanceolata, Cinnamomum camphora, Cryptocarya chinensis, Syzygium levinei, Syzygium hancei, Schefflera octophylla, Aquilaria sinensis (Wu et al 2021).More detailed information on the study site and forest restoration is included in Yao et al (1984) and Ren et al (2007).

Soil sampling
The experiment was initiated in November 2023.Six quadrats (10 m × 10 m) were randomly arranged with similar slope and aspect from each of the above four study sites (BL, EP, MF and UF) and were separated by 10 m from each other.Five scattered soil samples were collected using a soil auger (5 cm inner diameter) from 0-20 cm soil layer in each quadrat and then were mixed into one sample.Thus, six replications were collected in each study site, obtaining 24 soil samples in total.Before sampling, we removed humus, litter, and other impurities from the topsoil.All soil samples were screened with a 2-mm sieve to remove roots and gravel.Each soil sample was divided into two parts; one part was stored in the freezer at 4 • C for microbial biomass and enzyme activity determination, and the other part was air-dried for measuring soil physicochemical properties.Plant biomass was measured during the survey in 2015 (Wu et al 2021).

Soil properties measurements
The content of soil organic carbon (SOC) and total nitrogen (TN) were assessed by using a highest sample throughput automatic elemental analyser (Vario max cube, Elementar).The content of total phosphorus (TP) was determined using sulfuric-perchloric acid digestion and molybdenumantimony colorimetry.The pH of soil samples was measured using potentiometric method, and the ratio of water to soil was 1:2.5.Soil water content was determined using 10 g fresh soil sample oven dried at 105 • C for 48 h.Soil dissolved organic carbon (DOC) was determined using Martens' method (Martens 1995).Soil dissolved inorganic nitrogen (DIN), including both ammonium nitrogen (NH 4 + -N), and nitrate nitrogen (NO 3 --N) was extracted by mixing 10 g of each fresh soil sample by mixing it with 50 ml of 2 M KCl solution.The concentrations of DIN components were determined using a flow injection analysis system (LACHAT QC 8500 S2, HACH, USA).Soil available phosphorus (AP) was extracted using a solution containing 0.03 mol l −1 NH 4 F and 0.025 mol l −1 HCl.The concentration of AP was then determined spectrophotometrically at 700 nm using the molybdenum-scandium colorimetric method.

Soil microbial biomass measurements
Microbial biomass carbon (MBC) and nitrogen (MBN) contents of fresh soil samples were determined by chloroform fumigation extraction method (Vance et al 1987).Two fresh samples were weighed in glass containers, one was fumigated for 24 h under dark environment in a vacuum desiccator and another sample was maintained as control.50 ml of 0.5 mol l −1 potassium sulfate solution (soil-water ratio 1:5) was added to fumigated and unfumigated samples, respectively.The obtained filtrate was determined by automatic organic carbon analyzer (TOC-VCSH, Shimadzu, Japan).
Soil microbial biomass phosphorus (MBP) was extracted using the fumigation-extraction method (Vance et al 1987).Three portions of fresh soil were taken, the first portion was immersed in 50 ml of Bray-one solution (0.03 M NH 4 F and 0.025 M HCl) for 30 min, the second portion was immersed in 50 ml of Bray-one solution followed by the addition of 1.2 ml of 250 µg g −1 KH 2 PO 4 and then immersed for 30 min, and the third portion was fumigated under dark environment in a vacuum desiccator for 24 h before undergoing the same immersion steps as the first sample.The extraction solution was analyzed using inductively coupled plasma optical emission spectrometry (ICP-OES; Optima 2000 DV, Perkin Elmer, USA).Phosphorus extraction was corrected based on the results of the first and second extractions to eliminate the effect of phosphorus adsorption during the extraction process.Soil microbial phosphorus was calculated as the difference between the results of the first and third parts, with a conversion coefficient of 0.4 for microbial phosphorus.

Soil extracellular enzyme activity measurements
The enzymes activity involved in soil carbon acquiring (β-1,4-glucosidase, BG; cellobiohydrolase, CBH), nitrogen acquiring (β-1,4-N-acetyl-glucosaminidase, NAG, LAP), and phosphatase acquiring (acid phosphatase, ACP) were accessed (Verchot and Borelli 2005).The functions and substrates of the enzymes are detailed in table S1.In the experiment, 2 g of field-fresh soil and 30 ml of sodium acetate buffer solution with a pH of 5.0 were uniformly mixed using a blender.Subsequently, the soil slurry was dispensed into a 96-well deep-well plate.The deep-well plate was then incubated in the dark at 20 • C, with incubation times of 1 h for β-1,4-glucosidase (BG) and ACP, 3 h for leucine aminopeptidase (LAP), 5 h for β-N-acetylglucosaminidase (NAG), and 24 h for CBH.After the incubation period, the optical density was measured at 405 nm using a spectrophotometer (Thermo Scientific Multiskan, Waltham, MA, USA).Enzyme activity was expressed as micromoles per gram of dry soil per hour (nmol g −1 soil h −1 ).
Relative metabolic limitation of microorganisms was assessed by calculating vector length and angle (Moorhead et al 2016).Higher vector length indicating relatively higher microbial carbon (C) limitation.Vector angle was used to assess microbial nitrogen and phosphorus limitation, with vector angles >45 • indicating that microbes were relatively highly P limited, while angles <45 • indicated that microbes were relatively highly N limited.The equations for vector length and angle were calculated as follows: Vector angle = Degrees {arctan [(BG + CBH) /ACP, (2)

Microbial carbon use efficiency calculates
Microbial CUE is the ratio of C in microbial biomass to C uptake, which is calculated based on biogeochemical equilibrium modeling (Sinsabaugh and Shah 2012): S C:P = MB C:P /Soil C:P × (1/EEA C:P ) where EEA C:N was calculated as (BG + CBH)/ (NAG + LAP), and EEA C:P was calculated as (BG + CBH)/ACP.The CUE max was set at 0.6, and the K C:N and K C:P represent half-saturation constants with a fixed value of 0.5 in all model scenarios.

Statistical analyses
One-way analysis of variances and Duncan's multiple range test were conducted using IBM SPSS 21.0 (IBM Corporation, Armonk, NY, USA) to explore the differences in soil physicochemical properties, plant properties, microbial biomass, EEAs, and microbial metabolic limitation, and the significant differences were set at P < 0.05.Mantel test results were used to explore correlations among environmental factors and vector length and angle based on the Spearman correlation coefficient using linkET package in R v4.2.2.Random Forest models were used to identified the most important environmental variables, and the importance of variables was evaluated by classifying multiple decision tree (Breiman 2001).The analyses were conducted using the randomForest package (Liaw and Wiener 2002) in R v4.2.2.The A3 package were used to evaluate the significance of the model and the cross-validation R2, and the rfPermute package was used to assess the importance of each predictor to vector length and angle.The key factors were then combined in the structural equation modelling (SEM) to further explore the possible pathway of factors which controlled the microbial metabolic limitation.The SEM was conducted by using AMOS 24.0 (IBM Corporation, Armonk, NY, USA).

Soil physicochemical and plant characteristics
The results showed that forest restoration had significant impacts on plant biomass (table 1).There was hardly any plant in BL.With restoration, plant biomass increased to 9.1 t ha −1 in EP and to 179.0 t ha −1 in MF, respectively.The highest plant biomass (243.0 t ha −1 ) occurred in UF.The soil physicochemical properties varied significantly among the four forest types (table 1).SOC, TN, TP, NH 4 + -N, AP, SWC and DOC showed an upward trend with forest restoration.The BL and EP had higher NO 3 -N content than MF and UF, while soil pH was lower in MF and UF when compared with BL and EP.The DIN content initially increased to peak in MF and then decreased in UF.The MBC and MBN content showed a significant upward trend with forest restoration, and MBP content in MF and UF was significantly higher than that in BL and EP (table 1).
The stoichiometric characteristics of soil organic matter and microbial biomass varied significantly (P < 0.05, table 1) after forest restoration.Soil C/N, soil C/P and soil N/P increased significantly in EP when compared to BL. Soil C/N decreased significantly in UF when compared with MF (table 1).Soil N/P and Soil C/P did not change significantly between MF and UF.Microbial C/N and C/P increased significantly in EP when compared to BL, and then decreased significantly in MF.Compared with BL and EP, microbial N/P ratio decreased significantly in MF and UF.

Link between environmental factors and microbial metabolic limitations
The mantel test results showed microbial C limitation was positively related to soil pH (figure 4), and was negatively correlated with SOC, TN, TP, NH 4 + -N, NO 3 --N, AP, SWC, DOC, DIN, MBC, MBN, MBP, MBN/MBP, CUE and plant biomass (figure 4).Additionally, the Random Forest model indicated that microbial CUE, NH 4 + -N, MBN, MBC, TN, and

Shifts of microbial carbon limitation and the influncing factors
We observed significant differences in microbial C limitation (vector length) among the four forest types (figure 2(b)).Specifically, microbial C limitation significantly reduced after restoration in MF but not in EP when compared to BL, but it was still higher than that in UF.The microbial CUE value exhibited a strong negative correlation with microbial C limitation (figure 3(a)), consistent with previous studies (Cui et al 2020(Cui et al , 2021)).A higher CUE indicates a higher efficiency of converting exogenous C into microbial biomass, which holds greater potential for long-term C sequestration (He et al 2023).Conversely, lower CUE values imply a relatively higher proportion of carbon released through respiration (Manzoni et al 2012(Manzoni et al , 2017)).Our SEM results further indicated that the shifts of microbial C limitation was primarily attributed to microbial CUE driven by increases in plant biomass (as a proxy for litter input) and soil organic matter.At the BL site, there was hardly any plant and hence little litter input.Moreover, long-term soil erosion led to large soil C loss.Thus, microbial C limitation was the highest   microbial CUE (Kotroczo et al 2014).Thus, forest restoration in MF significantly reduced microbial C limitation compared to BL.The lacking change of microbial C limitation in EP may be attributed to timber felling that reduced litter input.Moreover, as a fastgrowth tree species, Eucalyptus likely led to root priming effects on soil organic matter in order to uptake more nutrient for growth.
Furthermore, we also observed that key nutrients such as TN and TP in the soil played crucial roles in microbial growth and metabolism (table 1).During the forest restoration, plants released nutrients into the soil through root exudates and litter decomposition, providing abundant N and P sources for microbes (Zhu et al 2012, Liang et al 2017).This satisfies the needs of microbial growth and metabolism, which in turn enhances microbial CUE and alleviates microbial C limitation.In addition to SOC, TN, and TP, other nutrients in the soil such as NH 4 + -N, NO 3 − -N and AP also significantly influenced microbial C limitation (figure 4).The increase in these readily available nutrients provided microbes with directly usable nutrients, further enhancing microbial biomass and CUE (Yang et al 2023).

Shitfs of microbial phosphorus limitation and the influencing factors
Overall, microbial communities were found to be limited by P (vector angle >45 • , figure 2(c)) across all forest types.Our findings align with previous research Our results showed that microbial P limitation could be significantly enhanced after forest restoration in EP and MF compared to BL (figure 2(c)).This could be attributed to higher plant P demand than soil P supply (table 1), thereby exacerbating the competition for P between plants and microbes (Jones et al 2018, Du et al 2020).Our SEM results further indicated that the shifts of microbial P limitation was co-driven by plant biomass, microbial CUE and soil P availability (figure 6(b)).Forest restoration usually decreased soil P loss compared to BL, hence promoting soil TP accumulation and enhancing soil available P content.However, the increased plant biomass after forest restoration may mean that the demand for P by plants is also constantly increasing, even exceeding the supply of soil P. Compared to BL, EP and MF exhibited an increasing trend in microbial CUE (figure 2(d)).When soil nutrients are scarce, microorganisms may enhance CUE to utilize organic carbon resources more effectively for energy acquisition and growth.This may result in microorganisms allocating more energy to growth and metabolic activities rather than P uptake and utilization (Sinsabaugh et al 2013).Therefore, microbial advantages in C acquisition may lead to a greater reliance on limited P resources in the soil, exacerbating microbial P limitation.
Surprisingly, the microbial P limitation was the lowest in UF, where plant biomass was the highest.This was probably due to that forest net primary productivity usually increases rapidly at young ages, peaks at middle ages, and then decreases at old ages to some extent.Thus, the highest plant biomass in UF may have lower rather than higher NPP and plant P demand.Beside, mycorrhizal symbiosis often become stronger as forest mature in UF, which likely also relieve the microbial limitation.However, a direct measurement of plant P demand and soil P supply as M Hu et al well as a evaluation of the role of mycorrhizal symbiosis were required to better clarify the shifts of microbial P limitation in this study.Additionally, the microbial CUE was highest in UF (figure 2(d)).Microbial CUE is an important indicator of microbial efficiency in utilizing carbon sources.A high CUE implies that microorganisms are able to produce more biomass or carry out more metabolic activities while acquiring the same amount of carbon source (Manzoni et al 2012).Therefore, the high CUE of microorganisms and the richer soil nutrients in UF may directly facilitate microbial P acquisition and utilization, reducing microbial P limitation.Furthermore, high microbial CUE may also reduce P limitation by increasing microbial activity and enhancing their decomposition and mineralization of soil organic matter, further releasing P from the soil (Tarafdar andClaassen 1988, Ru et al 2018).

Conclusion
After half-century of forest restoration on severely degraded tropical lands, we assessed the soil microbial metabolic limitations and the associated key drivers.Our results showed that soil microbial metabolism was co-limited by C and P. Microbial C limitation could be reduced after forest restoration particularly in the secondary forest but was still lower than that in the UF, which was mainly driven by microbial CUE.The increased microbial CUE was also positively correlated with plant biomass and soil organic matter.This suggests that forest restoration could facilitate SOC accumulation.Interestingly, microbial P limitation did not recover to the lower direction of the undisturbed primary forest and even was significantly enhanced after restoration in both the Eucalyptus EP and the secondary forest, which were co-controlled by plant biomass, microbial CUE and soil P availability.Therefore, developing optimal P management was required to maintain and improve the recovery pace and potential of these restored forests.Overall, the shifts of microbial C and P limitations and the strong connections of plant-microbe-soil processes revealed here will provide us with useful information for optimal management to improve soil health and achieve restoration success.

Figure 4 .
Figure 4. Drivers of microbial metabolic limitation across the four forest sites.The color gradient denoting pearson's correlation coefficients.
H, Qu Q, Li G, Liu G, Geissen V, Ritsema C J and Xue S 2022 Impact of nitrogen addition on plant-soil-enzyme C-N-P stoichiometry and microbial nutrient limitation Soil Biol.Biochem.170 108714 Xu Z, Yu G, Zhang X, He N, Wang Q, Wang S, Wang R, Zhao N, Jia Y and Wang C 2017 Soil enzyme activity and stoichiometry in forest ecosystems along the North-South Transect in eastern China (NSTEC) Soil Biol.Biochem.104 152-63 Yang X, Duan P, Hicks L, Wang K and Li D 2023 Mechanisms underlying the responses of microbial carbon and nitrogen use efficiencies to nitrogen addition are mediated by
c DIN (mg kg −1 ) 2.43 ± 2.1 a 14.53 ± 6.24 b 54.98 ± 15.11 d 43.91 ± 5.54 c MBC (mg kg −1 ) 11.29 ± 2.21 a 89.55 ± 22.17 b 348.57± 45.68 c 411.4 ± 68.35 d MBN (mg kg −1 ) 4.61 ± 0.66 a 18.28 ± 4.96 b 78.94 ± 9.74 c 90.35 ± 14.01 d MBP (mg kg −1 ) 0.56 ± 0.07 a 2.34 ± 0.5 a 20.04 ± 7.34 b 22.08 ± 8.13 b MBC/MBN 2.46 ± 0.38 a 4.92 ± 0.31 c 4.42 ± 0.17 b 4.55 ± 0.24 b MBC/MBP 20.04 ± 2.28 a 39.67 ± 13.18 b 19.76 ± 8.53 a 20.56 ± 8.04 a MBN/MBP 8.26 ± 1.11 b 8.14 ± 2.96 b 4.44 ± 1.81 a 4.47 ± 1.55 a Plant biomass (t ha −1 ) Ns 9.1 ± 2.0 a 179 ± 3.7 b 243.0 ± 33 b Lu X, Aoyagi R and Mo J 2018 Reconsidering the phosphorus limitation of soil microbial activity in tropical forests Funct.Ecol.32 1145-54 Pan Y et al 2011 A large and persistent carbon sink in the world's forests Science 333 988-93 Peng X and Wang W 2016 Stoichiometry of soil extracellular enzyme activity along a climatic transect in temperate grasslands of northern China Soil Biol.Biochem.98 74-84 Poorter L et al 2016 Biomass resilience of neotropical secondary forests Nature 530 211 Qiu X, Peng D, Tian H, Wang H, Liu X, Cao L, Li Z and Cheng S 2021 Soil ecoenzymatic stoichiometry and microbial resource limitation driven by thinning practices and season types in.Larix principis-rupprechtii plantations in north China For.Ecol.Manage.482 118880 Ren H, Li Z, Shen W, Yu Z, Peng S, Liao C, Ding M and Wu J 2007 Changes in biodiversity and ecosystem function during the restoration of a tropical forest in south China Sci.China Life Sci.50 277-84 Romanowicz K J, Freedman Z B, Upchurch R A, Argiroff W A and Zak D R 2016 Active microorganisms in forest soils differ from the total community yet are shaped by the same environmental factors: the influence of ph and soil moisture FEMS Microb.Ecol.92 fiw149 Ru J, Zhou Y, Hui D, Zheng M and Wan S 2018 Shifts of growing-season precipitation peaks decrease soil respiration in a semiarid grassland Glob.Change Biol.24 1001-11 Sinsabaugh R L et al 2008 Stoichiometry of soil enzyme activity at global scale Ecol.Lett.11 1252-64 Sinsabaugh R L, Hill B H and Shah J J F 2010 Ecoenzymatic stoichiometry of microbial organic nutrient acquisition in soil and sediment Nature 468 122 Sinsabaugh R L, Manzoni S, Moorhead D L and Richter A 2013 Carbon use efficiency of microbial communities: stoichiometry, methodology and modeling Ecol.Lett.16 930-9 Sinsabaugh R L and Shah J J F 2011 Ecoenzymatic stoichiometry of recalcitrant organic matter decomposition: the growth rate hypothesis in reverse Biogeochem 102 31-43 Sinsabaugh R L and Shah J J F 2012 Ecoenzymatic stoichiometry and ecological theory Annu.Rev. Ecol.Evol.Syst.43 313-43 Tapia-Torres Y, Elser J J, Souza V and Garcia-Oliva F 2015 Ecoenzymatic stoichiometry at the extremes: how microbes cope in an ultra-oligotrophic desert soil Soil Biol.Biochem.87 34-42 Tarafdar J C and Claassen N 1988 Organic phosphorus-compounds as a phosphorus source for higher-plants through the activity of phosphatases produced by plant-roots and microorganisms Biol.Fertil.Soils 5 308-12 Vance E D, Brookes P C and Jenkinson D S 1987 An extraction method for measuring soil microbial biomass-C Soil.Biol.Biochem.19 703-7 Verchot L V and Borelli T 2005 Application of para-nitrophenol (pNP) enzyme assays in degraded tropical soils Soil Biol.Biochem.37 625-33 Waring B G, Weintraub S R and Sinsabaugh R L 2014 Ecoenzymatic stoichiometry of microbial nutrient acquisition in tropical soils Biogeochem 117 101-13 Wu W et al 2021 Faster recovery of soil biodiversity in native species mixture than in Eucalyptus monoculture after 60 years afforestation in tropical degraded coastal terraces Glob.Change Biol.27 5329-40 Xu