Nature Restoration Shifts the Abundance and Structure of Soil Nematode Communities in Subtropical China

Aims Soil nematode community is an important component of the soil food web, which has been widely recognized as a key bio-indicator for assessing the inuence of nature restoration on ecological functions. However, the dynamics of the abundance, diversity and function of soil nematode community remain unclear under different forest succession phases. Methods The soil nematode community of natural secondary forests was investigated using a chronosequence approach. Nature restoration for ve succession stages were sampled in this study to represent a wide range of stand age groups. Results Soil nematode abundance gradually increased with forest stand age, which reached a peak value (574 individuals 100 g -1 dry soil) in the older age classes. In contrast, soil nematode diversity was not affected by forest stand age. Soil available nitrogen and phosphorus were key factors inuencing soil nematode abundance and diversity during forest secondary succession. The plant parasite index decreased with forest stand age, which indicated that ecosystem function and health would be improved as nature restoration progresses. In addition, the structure of soil nematode community was more sensitive to forest secondary succession compared to plant community and soil microbial community. The bottom-up effects of the plant and microbial communities on soil nematode community were important drivers of nematode community structure in subtropical forests. Conclusions Overall, this study demonstrates the active responses of soil nematode community to nature restoration, and highlights the importance of the above-ground and below-ground interactions to the soil food web. BF, bacterivores; FF, fungivores; PF, herbivores; OP, omnivores-predators. Different letters indicate statistically signicant differences among treatments based on the Duncan’s test (P < 0.05).


Introduction
Global losses of natural habitats are rapidly eroding terrestrial biodiversity and are expected to increase the rate of species extinction (Fahrig, 2003;Tilman et al., 2017). Human activities such as logging, hunting, burning and cultivating, have adversely affected terrestrial ecosystems and damaged the environment, threatening the existence and function of many species (Betts et  , which is characterized by forest stand age and forest community structure (Amossé et al., 2016;Sohlenius, 2002). Interactions between plants and soil organisms can greatly vary with forest age (Kardol et al., 2006;van der Putten et al., 2013). Zhang et al. (2015) showed that the abundance and diversity of nematodes was the highest in the middle forest age, and the nematode community was characterized by the changes in the abundance instead of dominant genera in a temperate forest. A better understanding on the responses of soil nematode communities to forest secondary succession would improve our ability to protect biodiversity and strengthen forest services. Nevertheless, the responses of soil nematode community, plant community, soil microbial community and abiotic factor have not been evaluated simultaneously under forest secondary succession in subtropical forests.
Forest secondary succession affects the soil nematode community through multiple pathways. Firstly, forest secondary succession could directly mediate soil nematode by changing the structures of vegetation and soil microbial community (Scherber et al., 2010). A recent study reported that under forest secondary succession, the perennial woody (shrub) plant cover gradually increased but the cover and The objective of this study is to explore how forest secondary succession could in uence the abundance, diversity and community composition of soil nematode community in a subtropical forest. We examined the dynamics of soil nematode under different forest stand ages (4-5 years, early stand initiation; 8-12 years, canopy closure; 18-22 years, stem exclusion; 25-30 years, canopy transition; and 35-40 years, gap dynamics) under natural secondary succession, with primary forests (over 100 years) as references.
We hypothesized that the abundance and diversity of soil nematode will increase with forest stand age, and that the structure of soil nematode community will be mainly regulated by changes in soil microbial community, soil properties and plant community during natural secondary succession.

Site description
This study was conducted in the Wuyi Mountain range of Longyan municipality (24 • 46′ to 25 • 28′N and 116 • 16′ to 116 • 57′ E), Fujian Province, China. The region represents one of the largest natural subtropical forests in southeast China. The study area had a typical subtropical monsoon climate, with a mean annual temperature of ca 20.1℃ and mean annual precipitation of ca 1646 mm. The soil is formed from granite and is classi ed as a red earth according to the Chinese Soil Classi cation System, equivalent to an Oxisol in the USDA Soil Taxonomy. The nature restoration developed through natural regeneration of local tree species after selective harvesting and rewood collection. Three subplots (5 m × 5 m) were randomly set up within each plot to examine the number of plant species, and then the percent cover of plant species level was investigated at the represent sampled stand (Su et al, 2021). The dominate tree species include Castanopsis carlesii, Castanopsis fargesii and Castanopsis ssa. In addition to dominant species, other tree species include Schima superba, Castanopsis faberi, Cinnamomum micranthum, Pinus massoniana, Sloanea sinensis, Cyclobalanopsis glauca, Cinnamomum camphora and Lithocarpus glaber (Su et al. 2021).
The design of the eld experiment has previously been described by Su et al. (2021). In brief, the forest region suffered from severe typhoons, repetitive harvesting and rewood collection in the last few decades. Stand ages were mainly obtained from forest management records. Natural secondary forests standing for ve succession stages were sampled in this study to represent a wide range of stand age groups, namely 4-5 years, early stand initiation (T4-5); 8-12 years, canopy closure (T8-12); 18-22 years, stem exclusion (T18-22); 25-30 years, canopy transition (T25-30); and 35-40 years, gap dynamics (T35-40); with primary forests over 100 years as references (T > 100). The distance between selected stands was over 1 km in order to account for spatial heterogeneity.

Soil sampling
A 20 × 30 m plot for each sampled stand was randomly set up to represent the stand. Similar climate among these stands provides an ideal chronosequence to study the effects of natural forest succession on ecosystem multifunctionality. Twelve soil cores (0-10 cm) were collected from each plot using an auger (2.5 cm inner diameter) in May 2019. The litter layer was removed prior to sample collection, and twelve soil cores were collected from each plot and composited. After the removal of visible plant residues, soil samples were passed through a 2-mm-mesh sieve, placed in an ice box and transported to the laboratory. Each sample was subdivided into two portions, one for the determination of basic soil properties and the other for nematode analysis.

Analysis of soil variables
Soil moisture was determined by oven-drying at 105 o C for 24 hours. Soil pH was measured using a glass electrode (1/2.5 soil/water mixture) with a pH meter (Mettler Toledo, Greifensee, Switzerland). Soil exchangeable NH 4 + -N and NO 3 − -N were extracted with 1 M KCl (1:5 soil : KCl solution) and determined by an automated ow injection analyzer (Skalar San++, Holland, Netherlands). Soil available phosphorus (P) was extracted with NaHCO 3 and analyzed by a spectrophotometer (TU-1810, China). Soil microbial biomass carbon (MBC) was determined with fumigation-extraction method described by Wu et al. (1990). Soil total carbon (C) and total nitrogen (N) were determined with a CNS Macro Elemental Analyzer (LECO Corp, MI, USA). Soil microbial community was assessed by phospholipid fatty acids (PLFAs).
The PLFAs were extracted from the soil as described by Wan et al. (2014). The resultant fatty acid methylesters of soil samples were separated, identi ed and quanti ed by capillary gas chromatography.
The abundance of each individual fatty acids was expressed as nmol g − 1 dry soil in a given sample against an internal standard (methylester C19:0; Sigma-Aldrich, Taufkirchen, Germany). The structure of microbial community was analyzed for each experimental year with a principal component analysis (PCA) based on the relative molar abundances of the entire fatty acids after standardizing to unit variance.

Extraction and examination of soil nematode
Soil nematode individuals were collected by direct funnel extraction from 100 g fresh soil, as described in a modi ed Baermann funnel protocol recommended by Whitehead and Hemming (1965). Soil nematode  (Bongers, 1990).

Statistical analyses
Differences between forest stand ages were tested using one-way ANOVA followed by Duncan's test. A non-metric multidimensional scaling (NMDS) was employed to visualize the structures of soil nematode, microbe and plant community using the 'vegan' package in R (version 3.5.2). The NMDS1 scores were conducted as the indicators for the community structure of soil nematode, microbe and plant. Structural equation modeling (SEM) was performed to analyze hypothetical pathways that may explain how forest ages affect soil nematode community structure using AMOS version 24.0 (Amos Development Corporation, Chicago, IL, USA). We classi ed all variables into ve functional groups before SEM analyses, namely forest stand age, plant community (NMDS1 of entire plant species), soil environment (i.e. soil pH, bulk density and moisture), soil microbial community (NMDS1 of entire fatty acids) and soil nematode community (NMDS1 of entire nematode genera). All variables were natural log transformed prior to SEM analysis to reduce departure from normality and linearity. The relative contribution of soil nutrients to soil nematode density and diversity was further examined by aggregated boosted tree (ABT) analysis (De'ath, 2007), using the 'gbmplus' package in R. Statistical analyses were performed in SPSS 22.0 (IBM SPSS Statistics for Windows, Version 22.0. Armonk, NY, USA).

Variations in soil properties
Soil pH, bulk density, NH 4 + -N concentrations, NO 3 − -N concentrations and MBC were largely in uenced by forest secondary succession (Table 1). Compared to T > 100, soil pH was signi cantly higher in the all succession stages. Soil pH tended to increase with forest stand age. Soil bulk density, NH 4 + -N, NO 3 − -N and MBC concentrations increased with succession stages. There was no signi cant difference in the available P, total C or N among forest stand ages. Soil nematode abundance and diversity Soil nematode abundance varied greatly during forest secondary succession (Table 2 and TableS1). The total nematode abundance increased with forest stand ages, and reached a peak (574 individuals 100 g − 1 dry soil) in the T35-40 age class (Fig. 1a) (Table 2). Similar to total nematode abundance, the abundances of all the trophic groups tended to increase with forest stand ages. Compared to the primary forests, natural secondary forests had lower abundances of PF across all stand ages, while the abundances of BF and OP were higher in the older stand age ( Table 2). By contrast, soil nematode diversity was not signi cantly affected regardless of forest stand ages (Fig. 1b). Table 2 The abundance of soil nematode (individuals 100 g − 1 dry soil) during forest secondary succession

Nematode trophic composition and ecological index
Compared to the primary forests, forest stand age showed contrasting effects on the relative abundances of nematode trophic groups (Fig. 3). The relative abundance of BF was signi cantly lower in the primary forests than in the T18-22 age class whereas that of PF was signi cantly lower in the T8-12 age class. In contrast, the relative abundance of OP was higher in the T25-30 age class than in the primary forests whereas that of FF was not signi cantly affected by forest stand ages. The PPI index decreased with forest stand ages (Fig. 2). Compared to T4-5, the PPI index was signi cantly lower by 8.2-9.3% in the T18-22 and T25-30 age classes.
The structure of soil microbe, nematode and plant communities For soil nematode, NMDS showed clear separation of younger age classes (T8-12 and T25-30) from the older age classes (T35-40 and T > 100), with the other two succession stages (T4-5 and T18-22) clustering together (Fig. 4a). The NMDS ordinations of soil microbial community and plant community showed clear separation at T4-5 and T > 100 age classes from the other age classes (Fig. 4b, c).
Contributions of plant and soil variables to the soil nematode communities ABT analysis indicated that soil NH 4 + -N and available P were the most signi cant predictors for the abundance of soil nematode during secondary succession, with the relative contribution of 33.3% and 26.5%, respectively. Soil available P and NO 3 − -N were the most signi cant predictors for soil nematode diversity during secondary succession, with the relative contribution of 43.2% and 23.5%, respectively (Fig. 5).
The SEM was constructed to explore the direct and indirect effects of forest secondary succession on soil nematode community (Fig. 6). Forest secondary succession signi cantly in uenced plant community structure, soil environment, soil microbial community and soil nematode community. Positive effects of forest stand age on the structures of plant and soil microbial communities, but negative effects on soil environment and nematode community structure, were detected. Forest secondary succession explained 39.9%, 22.1% and 39.7% of the total variances in plant community, soil environment and soil microbial community, respectively. Forest stand age, soil environment and the structures of plant and microbial community pathways explained 57.6% of total variance in soil nematode community structure. We also found that the microbial community structure was the main indirect pathway affecting soil nematode under forest secondary succession (Fig. 6). Overall, forest secondary succession could directly and indirectly in uence soil nematode community through modifying the structures of plant and microbial communities, and the soil environment.

Discussion
Effects of nature restoration on soil nematode community We found that the basal resources favored the soil nematode community, as well as the all trophic groups, both of which showed increasing patterns with forest ages during forest secondary succession.
These ndings agree well with previous reports (Keith et al., 2009;Morrien et al., 2017), which demonstrated that soil nematode populations peaked in the oldest age classes, mainly due to the increased root activity and ample food sources for soil nematode. For example, van Eekeren et al. (2009) observed that the abundance of herbivorous nematode was positively associated with ne root mass.
Indeed, vegetation biomass accumulates substantially with forest stand ages, providing greater amounts of resources (leaves and litter) to soil nematode community. Consistent with previous studies (Banning et al., 2008; Orozco-Aceves et al., 2017), we found that soil microbial biomass carbon and the abundance of soil bacteria tended to increase with stand ages. Furthermore, litter accumulation and decomposition may improve the microclimates and soil properties (e.g. bulk density, soil moisture and pH), thereby potentially increasing the population of nematode trophic groups (Qin et al., 2019). These ndings indicated that the abundance of soil nematode in secondary succession may eventually reach a level similar to that of primary forest soils in subtropical forests.
Contrary to our expectations, soil nematode diversity did not increase with forest stand age (Fig. 1b). This might be attributed to external disturbances which greatly alter the resource availability for soil nematode, consequently homogenizing these taxa across forest age classes (Francini et al. 2018). There are several possible reasons to explain this phenomenon. Firstly, in the same experimental site, Su et al. (2021) found that natural secondary forests had similar species diversities of shrub and bryophyte across all age classes. Secondly, soil biota in the higher trophic level require a longer time to adapt to changes in habitats than that in the lower trophic level (Valladares et al., 2012). Another possible explanation may be the distinct linkages among guilds in the soil food webs, because different trophic levels participated in differential interactions, including competition, predation, and mutualistic symbiosis (Yang et al., 2018). In this regard, soil nematodes may recover quickly after soil disturbances (Holtkamp et al., 2008) and are unaffected by forest stand age, despite an increase in plant biomass with time.
Our NMDS analysis of soil and plant communities separated intermediate age classes from primary forests. This suggested that the soil nematode and microbial communities showed an unstable structure during natural forest succession (Francini et al., 2018). We also observed that the community composition of soil nematode in the oldest stand age clustered together with the primary forests, although nematode abundance was lower in the T35-40 than primary forests. Soil nematode abundances and MBC were substantially higher in primary forests than in other forests, which is likely due to frequent disturbances in the young forests (Pickett et al., 2009).
Another important nding was that the value of plant parasite index decreased with forest stand ages.
The plant parasite index, which denotes the proportion of plant parasites and free nematode, could re ect the health of soil and plant in the ecosystem (Bongers, 1990). These results suggest that natural ecosystem disturbance would shift the soil nematode community in a forest ecosystem and may threaten its ecosystem services.

Contributions of environmental factors to the soil nematode under nature restoration
In the present study, soil nutrient availability and vegetation growth were improved markedly during forest secondary succession. Soil nematode community composition was generally correlated with changes in abiotic soil parameters, such as pH, C and N availability, and soil moisture (Liu et  We found that soil NH 4 + -N and available P concentrations were key factors in uencing soil nematode abundance during forest secondary succession, and soil available P and NO 3 − -N concentrations for soil nematode diversity. These results agree with the observation that changes in soil nutrient due to forest secondary succession could shape soil nematode density and diversity (Betts et al., 2017;Frouz et al., 2013). Future studies should devote more efforts to soil food web communities for a better understanding of the potential effects of arti cial or natural disturbance on ecosystem function and health in forest soils.
The At the whole system level, the responses of the soil nematode communities to forest secondary succession were much stronger than those of soil microbial community, plant community and soil environment (Fig. 6). Additionally, the responses of soil nematode and microbial communities were more sensitive to environment changes than that of soil properties and plant communities (Wang et al., 2021; Chen et al., 2015; Wilschut and Geisen, 2020). Expanding the knowledge on soil nematode in forest ecosystems would provide insights into sustainable forest management.

Conclusions
Our study provides strong evidence that the abundance of soil nematode gradually increased with forest stand ages, attributable to increased resource availability. Ecosystem disturbance had no effect on soil nematode diversity, but increased plant parasite index. The plant parasite index recovered to a level similar to that of primary forest. A potential shift in functional processes, as indicated by the increased plant parasite index under ecosystem disturbance, may affect ecosystem health in subtropical forests. Soil microbial and plant communities are critical to the community structure of soil nematodes, indicating that the bottom-up effects of the vegetation drive soil nematode community in subtropical forests. Therefore, future studies should consider potential feedback of above-ground and below-ground interactions to gain better understanding of the impacts on the soil food web during secondary succession in subtropical forests. Figure 1 Total nematode abundance (individuals 100 g-1 dry soil) and diversity during forest secondary succession. Different letters indicate signi cant difference between treatments at P <0.05.  Non-metric multidimensional scaling ordinations (NMDS) based on the considered soil nematode community (a), microbe community (b) and plant community (c).

Figure 5
The relative contribution of soil nutrients to the nematode abundance (a) and diversity (b) during forest secondary succession using aggregated boosted tree (ABT) analysis.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. TableS1.xlsx