Effects of cropland conversion into Masson pine plantation and leaf-use mulberry orchard on soil carbon pools in southwest China

To explore the differences in the responses of soil carbon (C) pools and key C cycle processes following the conversion of cropland to different plantation types, this study adopted the space-for-time substitution method, with three land-use types as research objects including: Masson pine plantation (MAS), mulberry orchard (MUL), and cropland for maize monoculture (MAI). After more than 25 years of cropland conversion, the contents of soil total organic carbon (TOC), particulate organic carbon (POC), and mineral-associated organic carbon (MAOC) were determined. Additionally, the contents of soil lignin phenols and amino sugars were analyzed. Meanwhile, the activities of cellulase, peroxidase, and β-1,4-glucosidase were measured. The results showed that the TOC content in MAS was 21.1 ± 0.5 g kg −1, significantly higher than that in MUL (11.9 ± 0.4 g kg −1) and MAI (10.2 ± 0.1 g kg −1); the POC and MAOC content in MAS were much higher than that in MUL and MAI. The total lignin phenol content in MAS (803 ± 35 mg kg −1) was nearly 7 times higher than that in MUL (116 ± 12 mg kg −1) and MAI (118 ± 6 mg kg −1). The total amino sugar content in MAS (646 ± 25 mg kg −1) was twice as high as that in MUL (361 ± 44 mg kg −1) and MAI (397 ± 40 mg kg −1); the total microbial-derived C in MAS (5.7 g kg −1) was significantly higher than that in MUL (3.2 g kg −1) and MAI (3.8 g kg −1). Structural equation modeling results confirmed that carbon-degrading extracellular enzymes promote the formation of POC and MAOC by regulating the lignin phenol content, with POC acting as the key fraction sustaining soil total organic carbon stocks. This study reveals that converting cropland to MAS can better promote the sequestration and accumulation of soil organic C compared with converting to MUL, and increasing plant-derived carbon inputs is critical for boosting regional soil carbon storage.

The soil carbon (C) pool is the largest C pool in terrestrial ecosystems, and its storage and dynamic changes directly affect the global C cycle and responses to climate change (1). Land-use change is one of the core anthropogenic factors driving the evolution of soil C pools (2). Against the global goal of “carbon neutrality”, exploring ways to enhance soil C sequestration via optimizing land-use structures have become a research hotspot in soil ecology and global change research (3). Cropland represents the land-use type most intensively disturbed by human activities; the conversion of cropland to orchards or plantations can significantly affect the accumulation, transformation, and stabilization processes of soil organic C by altering vegetation coverage, soil erosion, litter input, and microbial community structure, thereby reshaping the function of the soil C pools (4–6). Reforestation of agricultural land is generally considered to increase soil C storage. In tropical, a 14% increase in soil organic C was observed in topsoil areas within 10–25 years after the conversion of former croplands to forests (4). Under China’s Grain for Green Program, its cumulative carbon sink is estimated to reach 2635 Tg C by 2030 (7).

Soil organic C is not a homogeneous compound and can be further divided into active and inert fractions such as particulate organic carbon (POC) and mineral-associated organic carbon (MAOC) (8). POC is susceptible to environmental changes and serves as a sensitive indicator for the short-term dynamics of the soil C pool, while MAOC is closely associated with soil minerals, has a long turnover period, and acts as a key carrier for the long-term sequestration of soil C (9, 10). Different land-use types result in divergent POC and MAOC concentrations in soils. Moreover, less-disturbed lands (e.g., grasslands and woodlands) exhibit higher POC concentrations than croplands (11). Land-use change tends to modify the concentrations and proportions of POC and MAOC by altering ecosystem net primary production and soil properties (12, 13). Soil microorganisms act as the core driver regulating soil C cycling (14). Through alterations in microbially mediated C pools (fungal-driven C, bacterial-driven C) and characteristic biomarkers (e.g., lignin phenols, amino sugars), land-use change profoundly impacts soil C transformation (15–17). Fungi and bacteria participate in the decomposition and transformation of organic matter such as lignin and cellulose by secreting extracellular enzymes (18). The activities of these C hydrolases and oxidases are sensitive to land-use change, directly determining the decomposition rate of organic matter and C pool balance, and serving as important links between vegetation C input and soil C accumulation (19).

Southwest China is rich in biological, water, and mineral resources, with prominent ecological value (20). Driven by projects such as the Grain for Green program and the Transfer of Sericulture from Eastern to Western China, some croplands in this region have been converted into mulberry orchards or pine plantations (21, 22). Pine forests, the typical native vegetation in the region, are characterized by tolerance to infertile soils and high litter inputs (23). Mulberry grows rapidly, adapts readily to poor soils, and processes strong C sequestration capacity (24). Mulberry orchards and pine plantations constitute major land-use types formed after cropland conversion in Southwest China, and both are acknowledged to improve regional ecological conditions. Nevertheless, these two systems differ drastically in anthropogenic management intensity. Pine plantations receive minimal management and usually adopt near-natural conservation strategies, making them unlikely to deliver direct economic benefits in the short term (25). In contrast, leaf-use mulberry orchards are under intensive anthropogenic management, requiring fertilization, pruning, irrigation, and pest control to maximize leaf yields for sericulture. This practice leads to the removal of a large portion of photosynthetically fixed organic matter from the ecosystem, yet it can yield considerable positive economic returns (26, 27). At present, numerous studies have focused on the effects of converting cropland to a single type of orchard or plantation on the soil C pool (28, 29). However, systematic comparative studies on soil C fractions, microbial-driven C pools, and C-transforming enzyme activities after cropland conversion to Masson pine plantations and mulberry orchards in the same region are still scarce. In particular, the microbial regulatory mechanisms underlying the differences in C pools between the two conversion pathways remain unclear, which hinders the provision of precise theoretical support for optimizing regional land-use and enhancing C sink capacity.

On this basis, this study selected three land-use types (cropland, Masson pine plantation, and mulberry orchard) in southwest China as research objects. By determining the contents of soil total organic C and its fractions, characteristics of lignin phenols and amino sugars, and activities of key C-transforming enzymes, we aimed to clarify the differences in responses of soil C pools and key processes after cropland conversion to Masson pine plantations and mulberry orchards, and reveal the influencing mechanisms of different conversion pathways on soil C sequestration.

Study area and site description

The study area was located in Huashan township, Puge county of Liangshan Yi Autonomous Prefecture in Sichuan province, China (27°25′44.30″N, 102°35′24.39″E). The climate is subtropical monsoon and is characterized by a dry season from November to April and a wet season from May to October. The mean annual temperature is 16.8°C and the mean annual rainfall is 1169 mm. The total annual sunshine duration is 2094 h and the frost-free period is 301 days. The topography is mountainous and the study plots were located between 1260 m and 1330 m above sea level. The soils were classified as Luvisol (WRB).

The major land-use types in Huashan township include Masson pine plantation (MAS), mulberry orchard (MUL), and cropland. Moreover, the local cropland is mainly used for maize monoculture (MAI) and the mulberry orchard is used to collect mulberry leaves for feeding silkworms. According to local farmers, the dominant land-use change in Huashan township in past 30 years were (1) MAI transformed to MAS, and (2) MAI transformed to MUL. The development of the sericulture industry through the expand the area of mulberry orchard has brought considerable economic benefits to the local area and has played a significant role in helping local farmers lift themselves out of poverty.

Management practices of each kind of land-use types

For MAI, maize is planted in May and harvested in October annually. The management activities include using herbicides to prevent weeds growth after sowing and fertilization at rate of 200–240 kg N ha –1 yr –1, 45–55 kg P ha –1 yr –1, and 50–70 kg K ha –1 yr –1. P fertilizers, K fertilizers and 40% of N fertilizers are applied as base fertilizers, and 60% of the N fertilizer is applied in V12 stage of maize. After maize harvesting, the maize stalks are usually burned on site or directly left on the soil surface. For MUL, it is usually using herbicides to control weeds twice a year, and pesticides 3 to 4 times a year. Mulberry leaves were collected by hand from April to October to feed silkworm. In December, the mulberry branches above the base of the tree trunk are cut and removed from the mulberry plantation to prevent the breeding of pests and diseases and to facilitate access to the mulberry orchard. Approximately 7500 to 10000 kg ha –1 yr –1 of farmer-made organic manure (fermented from silkworm excrement and mulberry branches) is surface-applied without incorporation into the soil. The fertilization rate is 400–450 kg N ha –1 yr –1, 75–85 kg P ha –1 yr –1, and 90–100 kg K ha –1 yr –1, and they are divided into two applications in March and July. The annual output of fresh mulberry leaves is about 30000 kg ha –1 yr –1, and the annual output of fresh mulberry branches is about 27000 kg ha –1 yr –1. There is usually no human intervention in MAS.

Experiment design and soil sampling

We used the space-for-time substitution approach to quantify changes in soil C pool following conversion from MAI to MAS or MUL. Specifically, three replicated plots were established for each land-use type in this experimental design. Based on interviews with local farmers, we selected three plots for MAI which have maintained maize monoculture for more than 25 years; three plots for MAS which were transformed from cropland in 1999, and three plots for MUL which were transformed form cropland in 1998. A critical assumption of the space-for-time substitution approach is that all plots were initially similar with regard to soil characteristics and land-use history such that measured differences in soil characteristics among land-use type can be attributed to recent land-use change (30). We selected plots with consistent altitude, slope, and soil color. Consultation with local experienced agricultural technicians further verified that these plots shared nearly identical water and fertilizer management regimes prior to land transformation. To validate the rationality of this site selection, we additionally compared land-use-independent soil properties (specifically soil texture within the 20–60 cm soil layer) across all experimental plots. Basic soil and topographic characteristics across all land-use types are presented in Supplementary Table 1. No significant differences in soil texture were detected among MAI, MAS, and MUL.

In each plot, 6 sampling points were randomly selected and soil samples were collected from 0–15 cm using a soil auger in August 2024. The 6 soil samples from each plot were pooled in the field to form one composite sample. All soil samples were screened through 2 mm mesh and split into two portions: one was used to determine the soil enzyme activities and the other was air-dried, ground, and used to determine soil chemical properties.

Soil carbon fraction measurement

Soil total organic carbon (TOC) was measured by the dichromate digestion method (31). Soils were fractionated by size after aggregate dispersion following the method described by Cambardella and Elliott (32). A 10 g portion of each sample, sieved at 2 mm, was shaken for 18 h in 5 g L -1 sodium hexametaphosphate. Following dispersion, samples were sieved again at 53 μm using a vibratory sieve shaker. Soil particulate organic matter and mineral-associated organic matter represented the fraction remaining on the sieve (>53 μm) and the fraction that passed through the sieve (<53 μm), respectively. All fractions were oven-dried at 60 °C, weighed, and ground for analyses. The C content of particulate organic matter (POC) and the C content of mineral-associated organic matter (MAOC) were analyzed as described for TOC.

Soil enzymes activities measurement

The activity of cellulase was determined using carboxymethylcellulose sodium as the substrate (33) and the activity was defined as the amount of enzyme that produces 1μmol of glucose per gram of soil sample per hour. The activity of peroxidase was determined using L-3,4-dihydroxyphenylalanine (L-DOPA) substrate with exogenous H 2 O 2 supplemented in reaction mixtures and was defined as the amount of enzyme that generates 1 μmol of L-DOPA oxidation product per gram of dry soil per hour (34). The activities of β-1,4-glucosidase using p-nitrophenyl-β-D-glucopyranoside as the substrate (35), and the specific proceed follow Fang (36). The activities of β-1,4-glucosidase was defined as the amount of enzyme that release 1μmol of p-nitrophenol per gram of soil sample per hour.

Soil lignin phenols measurement

Lignin phenols were determined using alkaline CuO oxidation combined with gas chromatography (37). In short, lignin macromolecules decompose into monomeric phenols by alkaline CuO high pressure digestion. Quantitative ethyl vanillin was added to the oxidized products, acidified with 12 M HCl to pH 1, and then centrifuged at 2500 r min −1 for 30 min. After centrifugation, extraction was performed three times with ethyl acetate, and transcinnamic acid was added and concentrated under N 2 for further analysis. The phenol oxidation products were derived using N,O-bis-(trimethylsilyl) trifluoroacetamide and pyridine to yield trimethylsilyl derivatives before quantification. Finally, the lignin phenol monomer content was determined via gas chromatography. We condensed eight distinct phenols into representative lignin phenols, including vanillyl (Va) (vanillin, acetovanillone, and vanillic acid), syringyl (Sy) (syringaldehyde, acetosyringone, and syringic acid), and cinnamyl (Ci) (p-coumaric acid and ferulic acid)-type monomers.

Soil amino sugars measurement

Amino sugars were determined via acid hydrolysis combined with gas chromatography (38). Briefly, Soil samples containing approximately 0.3 mg of N were hydrolyzed with 6 M HCl at 105 °C for 8 h. Myoinositol was added to the hydrolysate as an internal standard, filtered, drying with a rotary evaporator and dissolution into deionized water. The pH value of the solution was adjusted to 6.7 and centrifuged at 3000 r min −1 for 15 min to remove precipitation. The supernatant was freeze-dried, and the remaining solid substance was dissolved in anhydrous methanol and centrifuged at 3000 r min −1 for 15 min. The supernatant was transferred to a bottle, and the solution was dried with N 2. The solution was then mixed with 1 mL deionized water and 0.1 mL N-methylglucamine, shaken, and freeze-dried again. Finally, amino sugar-related derivatives were determined via gas chromatography.

The total amino sugars were summarized as glucosamine (GlcN), galactosamine (GalN), mannosamine (ManN), and muramic acid (MurN). Additionally, bacterial- and fungal-derived C were calculated as previously described (16):

Fungal-derived C (mg kg −1) = (GlcN/179.17–2 × MurN/253.23) × 179.17 × 9

Bacterial-derived C (mg kg −1) = MurN × 45

Here, GlcN and MurN were measured in mg kg −1. The constants 179.17 and 253.23 are the relative molecular weights of GlcN and MurN, respectively. The conversion coefficients for GlcN to fungal-derived C and MurN to bacteria-derived C were 9 and 45, respectively (16).

Statistical analyses

The normality of distribution and equality of variance were first tested using Shapiro–Wilk’s test and Levene’s test, respectively. One-way ANOVA with Tukey’s HSD test was conducted to examine the effects of land-use conversion on soil chemical properties. Segmented structural equation modeling (SEM) analyses were performed using the R package “piecewiseSEM” to assess the relationships among chemical properties and soil C contents. All data analyses were performed using R version 4.3.1. The statistical significance for all the tests was set at P ≤ 0.05. Data presented as mean ± standard error.

Soil lignin phenols content

The contents of soil Va, Sy, and Ci monomer in MAS were significantly higher than those in the MUL and MAI (P < 0.01), whereas no significant difference was observed between the MUL and MAI soils (Figures 1A–C). The total lignin phenol content in the MAS (803 ± 35 mg kg −1) was nearly 7 times higher than that in the MUL (116 ± 12 mg kg −1) or MAI (118 ± 6 mg kg −1) (Figure 1D). Across the three land-use types, no significant difference was detected in the Sy to Va ratio, which ranged from 0.60 to 0.74 (Figure 1E). Similarly, the Ci to Va ratio showed no significant difference, with a range of 0.12 to 0.21 (Figure 1F).

Soil amino sugars content

The contents of GlcN, MurN, and GalN in the MAS were higher than those in the MUL and MAI (Figures 2A, C, D), whereas no significant difference in ManN content was observed among the three land-use types (Figure 2B). The total amino sugar content in the MAS (646 ± 25 mg kg −1) was higher than that in the MUL (361 ± 44 mg kg −1) or MAI (397 ± 40 mg kg −1) (Figure 2E).

Soil carbon fractions

The contents of POC in the MAS, MUL, and MAI were 8.3 ± 0.5 g kg −1, 2.7 ± 0.1 g kg −1, and 1.8 ± 0.01 g kg −1, respectively, with significant differences observed among all three land-use types (Figure 3A). The contents of MAOC in the MUL (7.2 ± 0.1 g kg −1) and MAI (6.6 ± 0.03 g kg −1) were significantly lower than that in the MAS (11.1 ± 0.3 g kg −1) (Figure 3B). The TOC content in the MAS (21.1 ± 0.5 g kg −1) was significantly higher than those in the MUL (11.9 ± 0.4 g kg −1) and MAI (10.2 ± 0.1 g kg −1) (Figure 3C).

The POC to TOC ratio in the MAS was 0.39, which was significantly higher than that in the MUL (0.22); in turn, the ratio in the MUL was significantly higher than that in the MAI (0.17)(Figure 3D). Conversely, the MAOC to TOC ratio in the MAS (0.53) was significantly higher than those in the MUL (0.22) and MAI (0.17) (Figure 3E).

The fungal-derived C content in the MAS was 2.0 times that of the MUL and 1.7 times that of the MAI, respectively (Figure 4A). The bacterial-derived C content in the MAS was 1.5 times that of the MUL and 1.3 times that of the MAI (Figure 4B). The estimated microbial-derived C content in the MAS reached 5.7 g kg −1, which was significantly higher than those in the MUL (3.2 g kg −1) and MAI (3.8 g kg −1) (Figure 4C).

No significant difference was observed in the fungal-derived C to TOC ratio among the three land-use types (Figure 4D). The bacterial-derived C to TOC ratio in the MAI (0.16) was significantly higher than that in the MAS (0.1) (Figure 4E). The ratios of microbial-derived C to TOC in the MAS, MUL, and MAI were 0.27, 0.27, and 0.37, respectively (Figure 4F).

Soil enzyme activity

No significant difference was observed in soil cellulase activity among the three land-use types (Figure 5A). In contrast, soil peroxidase and β-1,4-glucosidase activities in the MAS were higher than those in the MUL and MAI (Figures 5B, C). Additionally, soil β-1,4-glucosidase activity in the MUL was higher than that in the MAI (Figure 5C).

Relationships among soil chemical properties and carbon fractions

The structural equation modeling revealed that β-1,4-glucosidase activity had a significant positive path coefficient for total lignin phenol (Std.Est = 0.95, P = 0.001) and cellulase activity exerted a significant positive effect on total amino sugar (Std.Est = 0.41, P = 0.02) (Figure 6). Lignin phenol significantly facilitated the accumulation of POC (Std.Est = 0.99, P < 0.001) and MAOC (Std. Est = 0.99, P < 0.001), while amino sugar had no significant influences on POC and MAOC. POC exhibited a significant positive effect on TOC (Std. Est = 2.18, P = 0.007) (Figure 6).

Soil lignin phenols content among different land-use types

Lignin phenols are representative biomarkers of plant-derived recalcitrant organic carbon, primarily originating from plant litter decomposition and root exudation, and serve as fundamental stable carbon substrates for long-term soil carbon sequestration (39). The significantly higher concentrations of Va, Sy, Ci monomers, and total lignin phenols in MAS (approximately 7-fold higher than in MUL and MAI) strongly indicated a greater magnitude of plant C input under the MAS. This is consistent with regional management practices. MAS receives persistent inputs of recalcitrant lignin-rich litter (leaf litter and twigs) with slow decomposition rates (40), facilitating substantial accumulation of soil lignin phenols. Maize residues contain inherently lower lignin concentrations and abundant labile polysaccharides, leading to their relatively rapid turnover under tilled agricultural conditions (41). In MAI, post-harvest straw burning further constrains the accumulation of lignin-derived phenols in soils. In addition, MAI feature a shorter growing season, which limits annual inputs of soil organic matter. As intensively managed ecosystem, MUL undergo regular leaf harvesting and pruning, whereby nearly all annually produced aboveground biomass is removed from the mulberry fields, greatly limiting aboveground litter inputs. Frequent tillage in MUL and MAI further accelerates the mineralization and decomposition of organic matter (42). Therefore, despite annual additions of organic manure made from silkworm litter and mulberry prunings to MUL soils, concentrations of lignin phenols are still comparable to those in MAI. These findings demonstrate the critical role of continuous plant organic matter inputs in regulating the accumulation of soil lignin phenols, which aligns with previous studies. Soil lignin phenols content showed a consistent rising as tea plantation age increased from 1 to 42 years (43). Smooth vetch cover cropping in citrus orchards increases soil lignin phenols relative to clean tillage by boosting plant organic matter inputs (44).

The ratios of Va, Sy and Ci monomers are generally used to trace vegetation sources (39), whereas no significant variations in lignin phenol monomer ratios (Sy: Va, Ci: Va) were observed across the three land-use types indicated that these indices failed to distinguish the three ecosystems in this study. This result agrees with the findings in subtropical forests; the soil S:V ratios (0.62–0.77) exhibited negligible differences between deciduous and evergreen species (45). This may be because lignin input from plant litter went through different microbial decomposition rates among land-use types, which results in diminished differences in the biochemical signature of lignin in soils (45).

Soil amino sugars content among different land-use types

Amino sugars are well-established specific biomarkers for microbial residues, representing a stable and crucial component of SOC pools derived from microbial turnover (46). The significantly elevated contents of GlcN, MurN, GalN, and total amino sugars in MAS soils indicated more vigorous microbial metabolism, higher microbial turnover rates, and greater accumulation of microbial residual carbon under the MAS. Combined with the lignin phenol results, these findings establish a clear causal linkage: sufficient plant-derived lignin substrate input in MAS provides abundant carbon sources for microbial growth and reproduction, thereby promoting the continuous accumulation of microbial amino sugar residues and forming a primary “plant carbon input-microbial proliferation” succession pathway. Consistent with this framework, Xu et al. reported significantly elevated soil amino sugar concentrations alongside longer unmanaged durations in Moso bamboo forests, where amino sugar levels exhibited significant positive correlations with aboveground plant biomass (47). The similar amino sugar levels between MUL and MAI further verify that limited plant substrate input restricts microbial biomass turnover and residual carbon accumulation in both land-use systems.

Admittedly, empirical equations for microbial residue carbon calculations adopt a fixed GlcN/MurN molar ratio of 2 for bacterial cells (48), which inevitably creates uncertainties in amino sugar-based estimates of microbial-derived carbon. Nonetheless, this study quantifies fungal and bacterial carbon sequestration using distinct amino sugar profiles, offering valuable evidence for understanding how land-use variation modulates soil carbon storage and stabilization. Given that the turnover rate of microbial-derived carbon is intermediate between labile plant-derived carbon and recalcitrant organic fractions, it can effectively enhance the buffering capacity of the soil carbon pool and mitigate drastic fluctuations in response to environmental changes (49, 50). In MUL and MAI, the low content of microbial-derived C indicated that the soil C pool is mainly maintained by a small amount of stable C fractions, resulting in weaker flexibility and disturbance resistance. This further highlights the advantage of MAS in maintaining soil C pool stability. The ratios of bacterial-derived C to TOC and total microbial-derived C to TOC showed a consistent order: MAI > MUL > MAS. This contradictory phenomenon between absolute content and relative proportion is mainly caused by differences in the TOC base. The MAS has a large TOC storage; although the absolute content of microbial-derived C is high, its relative proportion is diluted. In contrast, the MAI has a small TOC base, so even though the absolute content of microbial-derived C is low, its relative contribution is prominent. Additionally, the proportion of bacterial-derived C in MAI is significantly higher, which is consistent with the ecological principle that anthropogenic disturbances such as tillage and fertilization are more conducive to the proliferation of bacterial communities and the accumulation of bacterial necromass in soil (47, 51). No significant difference was observed in the ratio of fungal-derived C to TOC among the three land-use types, indicating that the relative contribution of fungal-derived C is stable and the role of fungi in the soil C pool is conservative across the three land-use systems. In contrast, the significant variation in the proportion of bacterial-derived C reflects the selective shaping effect of anthropogenic management intensity on bacterial communities.

Soil enzyme activity among different land-use types

Soil C-transforming enzyme activity directly determines the decomposition and transformation rate of organic C, and is an important indicator reflecting the potential of C cycling (52). The MAS had the largest activities of peroxidase and β-1,4-glucosidase among three land-use types. Peroxidase is mainly involved in lignin decomposition (53); the persistent litter input in MAS plantation increases microbial enzyme secretion through substrate induction, forming positive feedback between substrate supply and enzyme activity. β-1,4-glucosidase is responsible for degrading cellulose into available glucose (33); the sufficient C source supply in MAS can meet the metabolic needs of microorganisms, maintaining high enzyme activity. That result is similar with a meta-analysis based on 803 observations find that increased cellulase, β-1,4-glucosidase, and peroidase activity associated with land restoration (19). Benefiting from the organic fertilizer application, MUL showed slightly higher β-1,4-glucosidase activity than MAI. Limited carbon sources and intensive tillage disturbance led to the lowest enzyme activity in MAI. Soil enzyme activity is tightly correlated with soil microbial biomass (54). The elevated amino sugar concentrations observed in MAS suggest a larger soil microbial biomass, supporting its higher soil enzyme activity.

The activities of C oxidative enzymes (i.e., peroxidase) and hydrolytic enzymes (i.e., β-1,4-glucosidase) in MAS are notably higher than those in MUL and MAI. Concurrently, significantly greater lignin phenol concentrations were detected in MAS. This suggests that lignin input outpaces decomposition, leading to net accumulation of lignin phenols notwithstanding enhanced potential enzyme activity. Furthermore, our structural equation modelling (SEM) results further demonstrate that extracellular enzyme activity exerts a positive control over the accumulation of soil lignin phenols.

Soil carbon accumulation among different land-use types

There was a significant difference in the increase of soil C content after the conversion of MAI to MAS and MUL more than 25 years. The increase in soil TOC content after MAI conversion to MAS was significantly greater than that after conversion to MUL. As a labile organic C fraction, the high soil POC content in MAS is a direct reflection of high litter input, where fresh organic matter can be continuously supplemented and accumulated. In contrast, MUL and MAI have low C input, resulting in limited storage of labile C pools. As a stable soil carbon fraction, MAOC has a longer turnover time in soil compared to POC (55, 56). The MAOC content in MAS was significantly higher than that in MUL and MAI, indicating that the conversion of MAI to MAS not only increases the total amount of soil C pool but also improves the quality of the C pool, demonstrating the enormous C sequestration potential of local soils. In Sichuan basin, the average net primary productivity for forest ecosystem were over three times than cropland and transform from cropland to forest significantly increased C sequestration (57).

Structural equation modeling elucidated the core regulatory pathways governing soil carbon sequestration. β-1,4-glucosidase activity exhibited a significant positive effect on lignin phenol accumulation, whereas cellulase activity positively regulated amino sugar concentrations, confirming that elevated enzyme activities serve as critical drivers for the accumulation of soil lignin phenols and amino sugars. More importantly, lignin phenol content substantially promoted the accumulation of both particulate organic carbon (POC) and mineral-associated organic carbon (MAOC), and POC further exerted a positive control over total organic carbon (TOC) accumulation. This highlights the vital role of continuous inputs of plant-derived organic matter in enhancing the stock and stability of soil organic matter.

The core reason about the soil C pool of MUL and MAI is lower than that of MAS is insufficient plant litter input. Therefore, increasing the input of organic matter in MUL and MAI is the key measure to improve its soil C pool function and enhance C sink capacity. To address the deficiencies of the existing management of MUL, measures such as mulberry branch returning can be implemented. The pruned mulberry branches are crushed and returned to the soil to increase the input of plant-derived organic C, provide sufficient C sources for microbial growth, and promote the accumulation of microbial-derived C. For MAI, post-harvest maize straw should be prohibited from open burning. Incorporating straw into topsoil represents a preferable strategy to boost organic matter inputs to soils. At the same time, green manure crops such as Chinese milk vetch and ryegrass can be planted in the winter fallow period (58) for MAI and MUL; after green manure incorporation, soil organic matter can be further supplemented, soil microhabitat can be improved, enzyme activity can be enhanced, thereby promoting the quantity and quality of the soil C pool, narrowing the carbon sink gap with MAS.

Following cropland conversion to MAS, ecosystems achieve efficient soil carbon sequestration via upregulated activities of key carbon-degrading enzymes, greater inputs of plant-derived lignin carbon, and enhanced microbial carbon accumulation. In contrast, MUL and MAI suffer from inadequate exogenous carbon inputs and low functional enzyme activities, which leads to weak carbon sequestration potential. This study systematically differentiates the divergent contributions of plant-derived and microbial-derived carbon to soil carbon pools and clarifies the pivotal roles of lignin phenols and functional enzymes in mediating soil carbon sequestration. These findings provide a theoretical foundation for optimizing regional land-use strategies and promoting soil carbon sink capacity. Nevertheless, the present study only addresses static variations in soil organic matter and carbon fractions. Future investigations integrating in-situ carbon flux monitoring and high-throughput microbial sequencing will help further unravel the microbial molecular mechanisms underlying land-use-mediated soil carbon turnover and stabilization.