植被演替转变了退化喀斯特生态系统中表层土和上层喀斯特裂隙土的土壤质量主控因素

《CATENA》：Vegetation succession shifts soil quality controls across surface and epikarst fissure soils in a degraded karst ecosystem

　　**Abstract:**- Vegetation succession increased the Soil Quality Index (SQI) across both surface and fissure soils.- Vertical position constrained the magnitude of karst soil-quality recovery.- Layer-specific Minimum Data Set (MDS) indicators differed among surface and fissure compartments.- Nutrient and enzyme modules showed the strongest associations with SQI.**Introduction:**Soil quality describes the capacity of soil to sustain biological productivity, regulate water and nutrient cycling, buffer environmental change, and maintain ecosystem functions. Because these functions emerge from coupled physical, chemical, and biological processes, multi-indicator assessment is generally more informative than diagnosis based on a single property (Karlen et al., 1997; Bünemann et al., 2018). Soil quality indices derived from a Total Data Set (TDS) or a Minimum Data Set (MDS) are widely used to synthesize complex soil information into comparable metrics. Their reliability, however, depends on whether indicator selection, scoring, and weighting reflect the dominant processes operating in the target soil-landscape system (Andrews et al., 2002; Gan et al., 2024; Tan et al., 2025).Karst ecosystems provide a distinctive setting for soil quality assessment because the soil mantle is commonly shallow, discontinuous, and strongly coupled with the epikarst zone. Carbonate lithology, rocky outcrops, fissure networks, and rapid water exchange generate strong vertical and lateral heterogeneity in soil water storage, nutrient retention, root distribution, and organic matter turnover (Williams, 2008; Jiang et al., 2014). In the karst rocky desertification regions of Southwest China, vegetation degradation, soil erosion, and slow soil formation interact to reduce ecosystem stability. Ecological restoration can partly reverse these trends by increasing vegetation cover, reducing soil loss, and enhancing carbon and nutrient accumulation, but recovery is constrained by the discontinuous soil-epikarst structure (Jiang et al., 2014; Zheng et al., 2024).Vegetation succession is expected to improve soil quality through linked aboveground and belowground pathways. Increasing plant cover can reduce raindrop impact and surface erosion, while litter inputs, root turnover, and microbial activity promote soil organic matter accumulation, aggregate formation, nutrient cycling, and enzyme-mediated transformations (Yan et al., 2019; Lu et al., 2022). These changes are particularly important in degraded karst landscapes, where small improvements in soil storage and nutrient retention can alter plant establishment and ecosystem functioning. Recent studies have shown that vegetation restoration in karst areas modifies soil nutrients, microbial communities, enzyme activities, organic nitrogen fractions, aggregate stability, and carbon sequestration potential, indicating coordinated changes in several soil functions rather than a uniform increase in nutrient concentrations (Lu et al., 2022; Zheng et al., 2024; Shu et al., 2025; Chu et al., 2026).Despite these advances, many assessments still treat sampled intervals as ordinary depth horizons. This convention is problematic in karst landscapes because soil below the surface mantle often occurs within fissures, pockets, or discontinuous epikarst voids rather than as a continuous subsoil horizon. Fissure-associated soils may differ from surface soils in water retention, rock contact, exchangeable base-cation supply, root accessibility, and microbial habitat conditions (Williams, 2008; Peng et al., 2020; Covington et al., 2023). Recent karst restoration studies further indicate that plant community composition, soil physical structure, and nutrient availability are tightly linked in thin carbonate-derived soils (Tan et al., 2025; Xing et al., 2026). Consequently, an MDS selected for a whole profile may overrepresent near-surface processes and underrepresent constraints specific to fissure soils. A unified MDS assumes that comparable functional gradients and indicator weights apply throughout the sampled profile. In a karst system, however, uniform selection and weighting may obscure differences between the continuous surface mantle and discontinuous fissure infill. This layer-specific MDS approach, rather than a single unified MDS, is essential for capturing the shifting functional controls along the soil-epikarst continuum.This study examined soil-quality recovery along a vegetation succession gradient in a degraded karst watershed of Southwest China. Four vegetation stages, from grassland (GL) to shrub-grassland (SG), shrubland (SH), and secondary forest (FT), were compared across three depth-defined positions measured from the ground surface: surface soil (SS, 0–20 cm), fissure soil at 20–40 cm (SFS), and fissure soil below 40 cm (DFS). The study integrated univariate statistics, PCA-based MDS selection, SQI calculation, RDA/PERMANOVA, random forest analysis, and PLS-SEM to test three hypotheses: (i) soil indicators and SQI would improve from grassland to secondary forest, but the magnitude of improvement would differ among soil-layer positions; (ii) the composition and weights of the MDS would differ among surface soil and the two fissure-soil positions; and (iii) the relative contributions of physical, nutrient, enzymatic, and geochemical attributes would vary across vegetation stages and soil-layer positions.**Section Snippets:****Study Area and Experimental Design:**The study was conducted in the Xinfazhai watershed (0.33 km2; 1281–1393 m), Zhenning County, Guizhou Province, Southwest China (105°49′30″–105°49′51″ E, 26°06′53″–26°07′23″ N), a typical catchment affected by karst rocky desertification (Fig. S1). The region has a humid subtropical monsoon climate, with a mean annual temperature of 14.03 °C and mean annual precipitation of 1277 mm. The landscape is characterized by peak-cluster depressions and well-developed epikarst fissures.**Joint Variations of Soil Indicators Along Successional and Vertical Gradients:**Two-way ANOVA showed significant effects of vegetation stage and soil-layer position on 15 of the 17 indicators (P < 0.05; Fig. 1, Fig. 2, Fig. 3, Fig. 4). Soil-layer position did not significantly affect MC, and vegetation stage did not significantly affect pH. Significant vegetation stage × soil-layer interactions occurred for SOM, URE, and INV, indicating that the magnitude of vegetation-stage differences varied among soil positions (Supplementary Table S2). From GL to FT, MC and STP...**Vegetation Succession Improved Soil Quality but Did Not Remove Vertical Constraints:**Soil quality increased from GL to FT across all three soil positions, but both TDS-SQI and MDS-SQI remained lower in fissure soils than in surface soil. This pattern is consistent with studies showing that vegetation restoration increases organic inputs, improves soil structure, and promotes nutrient retention in degraded karst ecosystems (Yan et al., 2019; Zheng et al., 2024; Tan et al., 2025). The persistent depth gradient indicates that vegetation recovery was superimposed on strong vertical constraints.**Conclusions:**Vegetation succession was associated with improved soil quality across surface and epikarst fissure soils, but the magnitude of recovery declined with depth. Later vegetation stages generally had better physical conditions, larger nutrient pools, and higher enzyme activities, while SQI remained lower in fissure-associated soils than in the continuous surface mantle. Soil-layer position therefore remained an important constraint on recovery.The indicators and weights selected for the MDS differed...**CRediT Authorship Contribution Statement:****Xinru Zhang:** Writing – original draft, Visualization, Software, Methodology, Investigation, Formal analysis. **Weiwei Dong:** Visualization, Formal analysis. **Youjin Yan:** Writing – review & editing, Supervision, Project administration, Funding acquisition, Conceptualization. **Lina Mei:** Writing – review & editing, Visualization, Formal analysis. **Liangjie Wang:** Writing – review & editing, Software, Formal analysis. **Fengling Gan:** Writing – review & editing. **Quanhou Dai:** Supervision, Resources, Funding.**Declaration of Competing Interest:**The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.**Acknowledgement:**Financial support for this research was provided by the National Natural Science Foundation of China (42007054, 42467045) and the China Postdoctoral Science Foundation (2020M673296). The authors are grateful to Nanjing Forestry University for funding this work and to the anonymous reviewers for their insightful comments.**Authors:**Xinru Zhang | Weiwei Dong | Youjin Yan | Lina Mei | Liangjie Wang | Fengling Gan | Quanhou Dai