Thinning Modulates the Soil Organic Carbon Pool, Soil Enzyme Activity, and Stoichiometric Characteristics in Plantations in a Hilly Zone

Abstract

Thinning, a core forest management measure, is implemented to adjust stand density and affect soil biogeochemical processes by changing biotic and abiotic properties. However, the responses of soil organic carbon (SOC), soil enzyme activity (EEA), and stoichiometry (EES) in plantations in hilly zones to thinning have received little attention. To test the hypothesis that thinning has regulatory effects on the SOC pool, EEA, and EES characteristics, field sampling and indoor analysis were conducted 9 years after thinning. Thinning significantly influenced the soil properties, especially in the topsoil, and significantly greater SOC and mineral-associated organic carbon (MAOC) contents were observed in the high-density treatment. The EEAs in the topsoil tended to increase with increasing density. SOC, MAOC, and C to phosphorus (C:P) had the greatest influence on the soil EEAs and EESs. Microbial metabolic limitations tended to change from nitrogen to phosphorus with increasing density. The soil properties, SOC fractions, available nutrients, and elemental stoichiometry drove microbial metabolic limitations and were significantly positively correlated with β-glucosidase, elemental stoichiometry, and EES. This study deepens our understanding of EEAs, SOC, and nutrient dynamics under thinning practices and elucidates how forest tending measures affect soil biogeochemical processes, thereby providing ideas for developing strategies to mitigate the adverse impacts of human interventions.

1. Introduction

Soil organic carbon (SOC) is the foundation for many vital ecosystem functions, such as nutrient transformation, bioturbation, and microbial dynamics [1]. Minor changes in SOC lead to drastic changes in the atmospheric carbon pool and C balance of terrestrial ecosystems [2]. SOC in forests accounts for 45% of the total forest carbon stock and predominantly originates from tree residues, rhizodeposits, and microbial necromass [3,4]. Microbial communities are closely linked to SOC sequestration and nutrient dynamics, and microbial necromass contributes 35% of the SOC in the topsoil [5]. This linkage is mediated by tree communities and microenvironments, i.e., soil characteristics, microclimates, and intraspecific competition, which are sensitive to forest management practices [6,7]. Thus, understanding how forest management measures affect forest soil C dynamics from a mechanistic perspective is crucial.

Plantations in certain regions are established for greening barren hills (i.e., public welfare forests in China), and thinning supports better stand conditions and ecological restoration [8]. A national greening project was carried out on these rocky slopes in the 1960s. Overcultivation was common, which seemingly resulted in a reasonable density and ensured sufficient survival, but the stands far exceeded the appropriate density as the trees grew. As trees age, tree layers occupy more upper space and have a greater advantage in utilizing resources, potentially leading to a decrease in understory community diversity [9]. Thinning, a core forest management measure, is employed to reduce stand density to create a better growth environment for the remaining trees and understory vegetation [10]. In general, competition for various resources decreases following thinning, resulting in greater plant productivity. The response of aboveground performance (tree growth rates, timber production, and understory biomass and diversity) to thinning has received considerable attention [11,12]. Thinning alters the microenvironment and the intensity of water and nutrient acquisition by trees and subsequently affects soil ecological processes; however, related research lags behind that on aboveground parts [13]. Therefore, understanding how thinning affects the soil C cycle, nutrient supply, microbial functional diversity, and the associated impact factors will broaden our understanding of stand density regulation strategies [14].

The increase in solar radiation received by the soil and the decrease in canopy interception after thinning have a significant impact on soil C turnover; these influences begin with increases in light penetration and soil temperature and then include changes in soil physicochemical properties and microbial activities [15,16]. Soil extracellular enzymes derived from roots, soil fauna, and microorganisms are involved in multiple biological processes and are capable of converting macromolecular organic matter into nutrients to facilitate vegetation nutrient acquisition and the soil C cycle [1,17]. Notably, thinning has a significant effect on microbial community structure and extracellular enzyme activities (EEAs), and related ecoenzymatic stoichiometries (EESs) provide opportunities to identify microbial metabolic limitations [18]. Moreover, the SOC content alone is not sufficient to quantify the effect of thinning on soil C cycling, and attention should also be given to its labile and recalcitrant fractions (particulate organic carbon, POC; mineral-associated soil organic carbon, MAOC) [19].

Platycladus orientalis, an evergreen tree species with excellent ecological value, was widely cultivated during the project, and many plantations are characterized by high density and a lack of effective management. However, the soil ecological processes in these plantations have received little attention, especially changes in the soil C pools and EEAs following thinning, which are poorly understood. This study aimed to (1) characterize differences in soil SOC and SOC fractions; (2) quantify microbial nutrient limitations in stands with different densities and soil depths; and (3) analyze the potential relationships among SOC fractions, soil physicochemical properties, and microbial nutrient limitations. Thinned stands experience a temporary reduction in SOC and then gradually recover over time, but how long this recovery takes has yet to be confirmed. We hypothesize that (1) relatively high stand densities increase soil C storage, (2) the response of labile and recalcitrant soil carbon to thinning varies with soil depth, and (3) thinning has regulatory effects on EEAs and microbial nutrient limitations. The validation of these hypotheses is expected to elucidate the importance of thinning for regulating SOC pools, soil EEAs, and EESs and deepen our understanding of sustainable management strategies.

2. Materials and Methods

2.1. Site Description and Field Investigation

The study site is located at Zhaotuan Forest Farm, Tongshan District, Xuzhou City, Jiangsu Province (34°11′49″ N, 117°30′54″ E, 75 m above sea level). Tongshan District is located in the transition zone between the Yellow River alluvial plain and the hilly region. The local average precipitation is 868.6 mm, the average annual temperature is 15.3 °C, extremely low temperatures occur in January (−13.5 °C), and extremely high temperatures occur in July (39.9 °C; obtained from the National Meteorological Science Data Center from 1991 to 2020; http://data.cma.cn/ (accessed on 23 March 2024)). The average sunshine duration is 2283 h, and the average frost-free period is 210 days.

The plantations are located on Mt. Dong and Mt. Dazhai at Zhaotuan Forest Farm, and the sites were characterized by west slope, leached cinnamon soil, and other site conditions. Pure Platycladus orientalis (L.) Franco plantations were established in 1959 for ecological restoration; thinning was conducted on Mt. Dong in 2014, and no other management measures were implemented. The low-efficiency forest-cutting strategy was adopted, exactly forest cutting with a specific stocking intensity based on ensuring that the remaining trees are arranged as evenly as possible, cut down trees of specific diameter classes, and priority was given to the removal of dumped and dead wood, thus creating a new stand density. The felled stem and logging residues were removed from the sites. The existing densities were obtained through our field investigation, and the final tree densities (tree·ha⁻¹) were 1300 (abbreviated as D1), 2900 (D2), and 3900 (D3) (

Table 1. Characteristics of Platycladus orientalis plantations at different stand densities.

Location	Density (tree·ha−1)	Height (m)	DBH * (cm)	Clear Bole Height (m)	Crown Diameter (m)
					E × W	N × S
Mt. Dong	1300	12.82 ± 0.37a	22.72 ± 3.35a	2.56 ± 0.27b	4.17 ± 0.07a	4.62 ± 0.52a
Mt. Dong	2900	9.88 ± 1.44b	13.82 ± 1.19b	2.51 ± 1.09b	2.70 ± 0.29b	2.71 ± 0.22b
Mt. Dazhai	3900	10.65 ± 0.50b	12.32 ± 2.36b	6.26 ± 0.40a	2.63 ± 0.41b	2.68 ± 0.32b

* DBH, diameter at breast height; E × W, east × west; N × S, north × south. The data were collected at the end of March 2023.

). D1 and D2 had an understory plant community consisting of Bidens pilosa L., Setaria viridis (L.) P. Beauv., Themeda triandra Forssk., Vitex negundo var. heterophylla, and Broussonetia papyrifera. D3 had fewer understory plants, namely, some weak Vitex negundo var. heterophylla plants and almost no herbaceous plants.

2.2. Soil Sampling

The soil sampling was conducted at the end of March 2023. Three 10 m × 10 m plots were randomly established in each plantation at a distance of more than 20 m, and the distances between the plantations all exceeded 1 km. Five samples were collected from two soil layers (0–10 cm and 10–20 cm) via a cutting ring in each plot, and the sampling points were equally spaced on two diagonal lines. The five soil samples from the same layer were combined into a single composite, equally mixed, and sieved through a 2 mm sieve after removing the fine roots and litter. A portion of each composite was stored at −80 °C, and the other portion was air-dried in a ventilated and dry room for subsequent measurement and analysis.

2.3. Soil Physicochemical Analyses

The soil pH was measured via potentiometry. The SOC content was determined via our previously described method [20]. The POC and MAOC contents were determined via the modified Walkley–Black method after specific pretreatment. Briefly, 10 g of air-dried soil was placed into a conical bottle containing 30 mL of sodium hexametaphosphate [(NaPO₃)₆, 5 g·L⁻¹] and reciprocally oscillated for 18 h (90 r·min⁻¹). The dispersions were passed through a 53 μm sieve; the part under the sieve (<53 μm) was the MAOC, and the part left on the sieve (53–2000 μm) was the POC. Both parts were collected, dried, weighed, and used for further analyses [21]. The total nitrogen (TN), total phosphorus (TP), and total potassium (TK) contents were determined via the Kjeldahl method, Mo-Sb anti-spectrophotometer method, and acid dissolution and flame atomic absorption spectrometric methods. The available nutrients, including ammonium N (AN), nitrate N (NN), available P (AP), and available K (AK), were determined via our previously described methods [22]. The microbial biomass C, N, and P (MBC, MBN, and MBP) contents were determined through the chloroform fumigation method [23].

2.4. Enzymatic Stoichiometry and Soil Microbial Metabolic Limitations

The extracellular enzyme activities were determined by corresponding ELISA kits, i.e., β-glucosidase (BG) activities were determined via a soil β-1,4-glucosidase ELISA kit (N-acetylglucosaminidase (NAG), leucyl aminopeptidase (LAP), and alkaline phosphatase (ALP); Shanghai Yanqi Biotechnology Co., Ltd., Shanghai, China). Furthermore, the EES was calculated via previously described formulas [24]. Vector analysis was applied to measure the extent of soil microbial C and nutrient limitations via analyzing the stoichiometric ratios of the extracellular enzymes [25].

$$\mathrm{L}\mathrm{e}\mathrm{n}\mathrm{g}\mathrm{t}\mathrm{h}=\sqrt{{[BG:\left(BG+ALP\right)]}^2+{[BG:\left(BG+NAG+LAP\right)]}^2}$$

$$\mathrm{A}\mathrm{n}\mathrm{g}\mathrm{l}\mathrm{e}°=\mathrm{D}\mathrm{E}\mathrm{G}\mathrm{R}\mathrm{E}\mathrm{E}\mathrm{S}\left\{\mathrm{A}\mathrm{T}\mathrm{A}\mathrm{N}2\left[\mathrm{B}\mathrm{G}:\left(\mathrm{B}\mathrm{G}+\mathrm{A}\mathrm{L}\mathrm{P}\right),BG:\left(BG+NAG+LAP\right)\right]\right\}$$

Microbial C limitation was represented by the vector length (L), with a longer L indicating greater C limitation. Microbial N/P limitation was represented by the vector angle (A), with A > 45° indicating P limitation and A < 45° indicating N limitation.

2.5. Statistical Analysis

The soil physicochemical properties, SOC fractions, and EEAs at different stand densities were analyzed via one-way analysis of variance via SPSS software (n = 3, version 22.0). Redundancy analysis (RDA) was conducted to determine the influence of soil properties on EEAs, EES, and microbial nutrient limitation. Pairwise correlation analysis (Pearson’s) of the soil properties, EEAs, and stoichiometry was performed, and correlation coefficients were determined via a color gradient. The contributions of soil properties, EEA, and stoichiometry to the variation in vector L and vector A were revealed via partial Mantel tests. Origin 2022 software was used for graphing.

3. Results

3.1. Soil Property Characteristics

The soil properties vary with stand density, and the trends of the changes in the total and effective contents of the same element are basically consistent (Figure 1). The pH gradually increased with soil depth and stand density; a greater TK content was observed in D2, and a greater AK content was observed in D3. The highest TN and AN were observed in D3, whereas the highest NN was observed in D1. The density significantly influenced the soil TN, AN, and NN in the topsoil (0–10 cm; p < 0.05). The density significantly influenced the soil TP and AP in both soil layers and exhibited similar patterns (p < 0.05). The mean TP content in D3 was 32%–36% greater than that in D1 in the two soil layers. Generally, the highest elemental stoichiometry values were found in D1; these values tended to decrease with increasing density in the two soil layers, and only the C:P and N:P ratios significantly differed between these densities in the subsoil (10–20 cm; p < 0.05).

3.2. SOC and Fraction Contents

The highest SOC, POC, and MAOC contents were found in D3 in both soil layers. The SOC, MAOC, and POC contents decreased by 26.35%–59.01% with increasing soil depth at the three stand densities. The proportion of MAOC was always greater than that of POC and tended to increase with increasing soil depth at the three densities, increasing by 4.57%–10.02% (Figure 2). SOC and MAOC were significantly different among the different densities at the 0–10 cm depth (p < 0.05, Figure 2). The POC content did not significantly differ among the different densities in either soil layer (p > 0.05).

3.3. Microbial Biomass C, N, P, and Stoichiometry

Among the microbial biomass indices, only the MBC content differed significantly in the 0–10 cm soil layer (p < 0.05, Figure 3). The MBP content decreased as the stand density increased, resulting in 34.59% and 43.82% decreases in D3 relative to D1 in the two soil layers, respectively. The MBC and MBN contents tended to increase as density increased, resulting in 36.31% and 71.29% increases in D3 relative to D1 and 26.63% and 37.45% increases in the two soil layers, respectively. MBC:MBN gradually decreased with increasing density, whereas MBC:MBP and MBN:MBP showed the opposite trend (p < 0.05).

3.4. Soil EEA and EEA Stoichiometry

Figure 4 illustrates the differences in EEA among the different stand densities, where the mean EEAs tended to increase with the increasing density. The BG, NAG + LAP, and AP activities in D3 were 38.16%, 87.67%, and 38.76% greater, respectively, than those in D1. In the 10–20 cm soil layer, the BG and ALP activities of D3 were 18.51% and 93.06% higher than those of D1, reaching 858.59 and 1985.02 nmol·g⁻¹·h⁻¹, respectively. The EEAs decreased with increasing soil depth, with BG decreasing by 45.16%–59.67%, NAG + LAP decreasing by 28.43%–49.22%, and ALP decreasing by 13.09%–15.75%. The enzyme C:N ratio was consistently <1 among the different densities and decreased slightly with increasing soil depth, decreasing by 1.57%–8.55%. The enzyme C:P and N:P ratios gradually decreased with increasing stand density, whereas significant differences were detected only in the subsoil (p < 0.05).

3.5. Soil Microbial Metabolic Limitations

Vector length (L; microbial C limitation) presented different variation patterns in the two soil layers; for example, D2 presented the highest and lowest L values in the 0–10 (1.40 ± 0.06) and 10–20 cm (1.26 ± 0.04) soil layers, respectively. D1 always had relatively high values (1.40 ± 0.04 and 1.32 ± 0.05, respectively). L was greatest in D1 and D2, indicating severe C limitations. The vector angle in D1 was always less than 45° (N limitation), whereas that in D2 was greater than 45° in the two soil layers (P limitation). The vector angle of D3 showed different patterns in the two soil layers (Figure 5).

3.6. Soil Properties Controlling EEA, Stoichiometry and Microbial Nutrient Limitation

The first two ordination axes of the RDA explained 62.54% of the total variation in the relationships between soil EEA and stoichiometry and soil properties. Soil SOC, MAOC, and C:P had the longest arrows, which indicated their greater ability to explain changes in EEA, stoichiometry, and microbial nutrient limitations (Figure 6). A Mantel test was subsequently performed to verify the relative significance of the soil properties, SOC fractions, EEAs, and EESs in determining microbial nutrient limitations (Figure 7). Vector L was extremely significantly positively correlated with BG, C:N, C:P, N:P, the enzyme C:N, and the enzyme C:P (p < 0.01). The vector angle was positively correlated with NN, ALP, C:P, and N:P and with the enzyme C:N, C:P, and N:P ratios (p < 0.01).

4. Discussion

Thinning always results in new densities, which can affect the microclimate, litter decomposition rate, soil nutrient utilization, and SOC dynamics. This study demonstrated that thinning causes soil pH to approach neutral by increasing the thinning intensity; this finding is similar to the result of increasing soil pH with increasing thinning intensity in acidic soil [26]. A possible reason for this may be that a reduction in the amount of litter input causes rapid decomposition of organic acids and promotes acidification of alkaline soils [27]. In low-density stands, the canopy intercepts less rainfall, and a large amount of rainfall contacts the ground directly, causing soil erosion and leaching and leading to salt-based ion loss and soil acidification. In addition, coniferous tree litter decomposes faster at low densities, and the accumulation of acidic substances increases in the soil [28]. The relatively thick undecomposed litter layer in the high-density stands in our study is consistent with these findings. A reduction in N availability can have a significant negative effect on AP production [29]. The increased P availability with increasing density may be related to increased root density, as roots release low-molecular-weight organic acids that interact with alkaline phosphatase to promote the conversion of organic phosphorus to inorganic phosphorus [30]. Moreover, the stimulation of net N mineralization and nitrification following thinning resulted in increased NN with thinning intensity to a certain extent [31]. The abundance and activity of microbial ammonia oxidizers and net N mineralization rates are stimulated through plant–microbial interactions, thereby altering the nitrogen cycle and increasing the NN content [32,33]. Moreover, thinning reduced nutrient competition between the remaining trees, which provided the remaining trees with greater growth space and stimulated their demand for N, which may have enhanced N mineralization and nitrification [34].

SOC is classified into POC and MAOC, which are thought to have distinct formation and turnover times and biogeochemical properties [35]. Thinning is widely recognized to reduce total carbon storage in forests, regardless of whether C in harvested wood is included; similar results were obtained in this study [36]. Consistent with our first two hypotheses, the SOC content increased significantly with density, and the MAOC content increased only in the surface layer, indicating that high density not only promoted SOC accumulation but also increased the proportion of MAOC [26]. The dynamic balance of SOC pools mainly depends on carbon inputs and outputs involving microorganisms in long-term stable ecosystems [37]. Thinning not only reduces the amount of plant C entering the soil by removing canopy trees but also accelerates SOC decomposition by improving the soil environment [38]. The trends in fine root biomass were consistent with the trends in stand density, and higher root density resulted in greater SOC storage. Severe thinning significantly decreased SOC nine years after thinning in Pinus tabulaeformis plantations, and this result was closely related to N availability, litterfall, and microbes [39]. A reason for the decreased SOC in thinned stands could be that litter quantity plays a more crucial role in affecting SOC storage than understory vegetation does [40]. MBC is a labile and sensitive organic carbon fraction, and its trends were largely consistent with those of SOC following thinning in our study. D3 had significantly greater MBC contents. Thinning reduces above- and belowground litter inputs, and a smaller amount of litter in low-density stands may decrease soil nutrient availability, which in turn increases the difficulty of meeting the nutrient needs of microorganisms and inhibits their reproduction; on the other hand, decreased canopy density after thinning increases the soil temperature and decreases the soil moisture content, thereby decreasing the MBC content [24,41]. However, this influence may vary seasonally, as warmer and wetter microclimates in thinned stands may allow for rapid microbial growth during summer and increase microbial biomass [15]. Thinned stands experience a temporary reduction in SOC and MBC and then gradually recover over time. This recovery time may last 6–20 years and can be affected by tree species, soil type, climate zone, and other factors [42]. Notably, the SOC and MBC contents of D1 were greater than those of D2 in this study, which may indicate that this recovery time was influenced by thinning intensity; i.e., the greater the thinning intensity was, the faster the recovery was [26]. The POC content did not change significantly at the different densities in the three forest types because the POC pool is mainly composed of plant-derived components, which are easily mineralized [21]. Moreover, an increase in the metabolic quotient after thinning promotes microbial substrate availability and decreases POC stability [43]. Additionally, the higher SOC content and MAOC proportion in D1 than in D2 indirectly suggest its leading role in the recovery process, which has been confirmed by a meta-analysis in which the SOC and MAOC contents increased with recovery time [19]. This recovery benefits from greater understory plant biomass, a warmer and wetter microclimate, and faster litter and root decomposition in low-density stands [38]. Thinning practices change stand three-dimensional structure, litter and root exudate inputs, and physical, chemical, and biological processes in soil, and a new equilibrium of SOC pools may be established decades after thinning.

After thinning, enzyme activities showed statistically insignificant decreasing trends, and similar trends were observed in previous studies [26]. A well-established theory is that an increase in root biomass could increase the secretion of extracellular enzymes with stand density. However, the results are not entirely consistent, possibly because thinning in different regions produces resources of different qualities that are utilized by microorganisms [44]. Generally, BG (a typical C-hydrolase) significantly affects the C-cycle pathway and the SOC and MBC contents [45]. Stronger enzyme activities induce C- and N-related microbial chemical processes, thereby providing more nutrients for microorganisms and plants [46]. While thinning did not dramatically alter BG activity, similar results showed that BG activity decreased following thinning and that the decomposition rates of fine roots and litter decreased, which restricted the SOC pool input [47]. To a certain extent, the lack of significant differences in the soil EEAs and EESs indicates that the regulatory effect of stand thinning on soil is not sufficient to cause environmental differences that can change the nutrient acquisition propensities of soil microorganisms [1]. In addition, soil microorganisms have a synergistic effect on C, N, and P acquisition in response to stand thinning; thus, no significant changes in stoichiometric characteristics have been observed [1]. In accordance with our hypothesis, C limitation is slightly alleviated, and P limitation is exacerbated with increasing SOC content; similar conclusions have been drawn in grasslands on the Loess Plateau [48]. Vector analysis revealed that the vector L values were <45°, which indicated microbial N limitation at D1. The leaves, dead roots, branches, and needles left in forests after thinning can improve the availability of substrates with high C:N ratios, thereby increasing the C supply and N limitations of microbial communities [36]. As previously proposed, different thinning intensities may result in different recovery stages, where labile C is utilized mainly in the early stage and recalcitrant C is utilized in subsequent stages (as confirmed by a decrease in C hydrolase activity and an increase in C oxidase activity) [18]. Notably, significant effects on most of the measured soil chemical properties caused by thinning were observed only in the topsoil [49]. This study, which was conducted in only one season, has certain limitations, as significant changes in light, temperature, and humidity in the middle of the growing season lead to significant differences in many indicators [1]. After investigating the long-term effects of thinning, Pretzsch (2020) reported that perhaps no active thinning or removal of dying or dead trees is a better solution if the goal is to achieve the maximum volume yield or carbon stock [49]. However, in high-density stands, there is no increase in understory vegetation diversity, aboveground biomass, soil fertility, or soil moisture effectiveness at any age [50]. Thus, for noncommercial forests in China, moderate thinning is still needed to achieve sustainable management and improve understory plant diversity and natural renewal capabilities.

5. Conclusions

The density and soil depth had significant impacts on soil properties, SOC, and microbial nutrient limitations in public welfare plantations in hilly zones. Thinning led to reduced organic matter input and a changed microenvironment, resulting in significantly greater SOC and MAOC contents in the high-density treatment, but significant differences were mainly observed in the topsoil. Interestingly, the EEAs exhibited similar trends, and no significant differences were observed among stand density as well as between soil layers (except ALP in the 10–20 cm layer), and SOC, MAOC, and C:P had the greatest influence on the soil EEAs and EES. Microbial nutrient limitation varied among the different stand densities, and low tree density caused N limitation in the two soil layers. Soil properties, SOC fractions, available nutrients, and elemental stoichiometry drove microbial metabolic limitations, with metabolic limitations significantly positively correlated with C-acquisition enzymes, elemental stoichiometry, and EES. In general, these microbial metabolic limitations reflect time-dependent response strategies after thinning. The possible stages of the post-thinning recovery process affect EEAs and thus soil C, N, and P cycling and ultimately influence forest SOC sequestration. This study provides a basis for assessing EEAs and SOC and nutrient dynamics under thinning practices and contributes to understanding the impact of forest tending measures on soil biogeochemical processes, thus supporting the development of corresponding strategies to mitigate the adverse effects of human interventions.