Small gaps support higher understory diversity in subtropical Pinus massoniana Lamb. plantations

1 Introduction

Forest gaps—canopy openings resulting from the death or removal of canopy trees—are among the most common and significant small-scale disturbances in forest ecosystems (Watt, 1947). Gap creation disrupts the closed canopy, altering understory light, temperature, and humidity regimes while reshaping the understory microenvironment via litter decomposition and soil moisture and nutrient dynamics. Extensive research demonstrates that gap size, shape, age, and within-gap position jointly govern the intensity and duration of microenvironmental variation, thereby influencing regeneration strategies and diversity patterns among plant life forms (Hou et al., 2024; Lu et al., 2023).

In natural forests, canopy gaps generally result from tree mortality or windthrow. Their high structural heterogeneity, together with biological legacies such as propagule sources and coarse woody debris, creates diverse microhabitats. By contrast, plantation gaps are often associated with harvesting, pest outbreaks, or disease and typically occur in structurally homogeneous stands. Thus, gaps of similar size may exhibit contrasting ecological responses across natural forests and plantations because of differences in stand structure, regeneration conditions, and biological legacies (Turquin et al., 2025). Close-to-nature interventions, including gap-based management, have therefore become an important focus of plantation restoration and management (Lyu et al., 2021a; Liu et al., 2020). Appropriately designed gap management may improve understory light and thermal conditions, facilitate the regeneration of broadleaved trees, shrubs, and herbs, and promote succession toward greater structural complexity and functional stability (Lyu et al., 2024). However, these outcomes depend not only on gap size but also on stand type, disturbance origin, soil conditions, and propagule availability (Tong et al., 2024; Tinya et al., 2025).

Studies of plantation canopy gaps indicate that understory plant diversity varies with gap size. In Pinus massoniana plantations, intermediate-sized gaps were associated with relatively high shrub-layer diversity, whereas herb-layer diversity showed a more complex response (Lyu et al., 2021b). Similarly, intermediate-sized gaps in Japanese cedar and hinoki cypress plantations on Mount Lushan were associated with greater understory species and phylogenetic diversity (Mao et al., 2025). However, previous research has focused primarily on gap-size effects, leaving the joint influences of gap size and within-gap position (center, transition, and edge zones) poorly understood (Wen et al., 2024; Zhang et al., 2025). Mechanistic interpretations have also emphasized microclimatic factors, particularly light and temperature, while the relationships of understory diversity with soil nutrients and litter characteristics remain understudied (Song et al., 2025; Li X. C. et al., 2025). Thus, further research is needed to clarify within-gap spatial heterogeneity and the environmental drivers of understory plant diversity in plantations.

In China’s subtropical region, Pinus massoniana plantations dominate red-soil hilly areas and are central to regional timber production and ecological security. However, prolonged monoculture, high-density planting, and production-oriented management have often led to soil impoverishment, site degradation, and dominance by a limited set of understory species (Li et al., 2020). Xingguo County, Jiangxi Province – a historically severe soil-erosion area—has relied heavily on P. massoniana for artificial vegetation restoration (Zhan et al., 2022). In recent years, natural disturbances, particularly pine wilt disease (caused by the pine wood nematode), have greatly increased gap formation in P. massoniana plantations (Pan et al., 2021). Therefore, we investigated P. massoniana plantations in Xingguo County, Jiangxi Province, China, to elucidate the relationships among gap size, within-gap position, microenvironmental conditions, and shrub- and herb-layer diversity. Specifically, the objectives were to: (1) Quantify the effects of gap size and within-gap position on microclimate and understory species diversity. Here, we hypothesize that gap size has a stronger influence on both microclimate and understory species diversity than within-gap position. (2) Identify key environmental drivers and their influence pathways for understory diversity. Here, we hypothesize that gap size primarily affects understory species diversity indirectly, mediated mainly through microclimate and soil factors. Our findings offer a scientific basis for the ecological restoration and sustainable management of P. massoniana plantations.

2 Materials and methods

2.1 Study area

Xingguo County (26°03′ ~ 26°44’ N, 115°01′ ~ 115°52′E), situated in south-central Jiangxi Province, China, covers 3,215 km². The landscape is dominated by low mountains and hills, dominated by red soils (Ultisols). The climate is subtropical monsoon, characterized by ample rainfall and distinct seasons, featuring a mean annual temperature of 18.8 °C, mean annual precipitation of 1,515.6 mm, and an average frost-free period of 284 days. Forest cover is 72.2%, comprising evergreen broadleaved forests, P. massoniana plantations, and Cunninghamia lanceolata plantations. The geographical location map of the study area is shown in Figure 1.

2.2 Experimental design

Before establishing plots, we conducted extensive surveys in areas of high gap density within P. massoniana plantations. These surveys confirmed that gaps were predominantly formed by stem snapping and tree mortality induced by pine wilt disease. Gap age was estimated from regenerating P. massoniana age and deadwood decay class, indicating formation within approximately 10 years. Therefore, we selected 40 near-elliptical gaps of approximately 10 years old for study from July to August 2023.

We adopted the expanded definition of a gap as the polygon enclosed by border-tree trunks (Schliemann and Bockheim, 2011). Given the strong influence of border-tree height on gap dynamics and microenvironment, gap size was quantified as the D/H ratio, where D is the mean of the major and minor axes, and H is the mean border-tree height. The major axis was defined as the maximum distance between border trees passing through the gap center, and the minor axis as the perpendicular distance through the gap center. Gaps were stratified into five size classes based on prior classifications (Zhu et al., 2015) and site-specific conditions, using upper-bound exclusion: Class I (0.5 ≤ D/H < 0.75, n = 8), Class II (0.75 ≤ D/H < 1.0, n = 8), Class III (1.0 ≤ D/H < 1.25, n = 11), Class IV (1.25 ≤ D/H < 1.50, n = 6), and Class V (D/H ≥ 1.50, n = 7). Gaps were located at elevations ranging from 156 to 310 m, and each size class encompassed plots distributed across slope positions (upper, middle, lower) and aspects (sunny, shady).

In each gap, 2 m × 2 m small quadrats were systematically placed for microenvironmental and vegetation assessments. Specifically, one quadrat was centered in the gap, four were located in the marginal zone (beneath the inward canopy projection of border trees), and 4 quadrats were placed in the transitional zone midway between the center and edge along both axes. Additionally, at ~10 m beyond the edge of each gap, four 10 m × 10 m closed-canopy control plots (CK) were established, each containing five 2 m × 2 m small quadrats. The sampling design is illustrated in Figure 2.

2.3 Field investigation

Gap and CK plot centers were recorded using GPS, with vegetation and site variables surveyed in both plot types. In each 2 m × 2 m small quadrat, we recorded shrub composition, abundance, cover, and height, as well as litter cover (LC) and thickness (LT). A nested 1 m × 1 m subplot centered within each 2 m × 2 m quadrat was used to record herbaceous species, clump counts and percentage cover. At each quadrat center, 0.5 m above ground, we measured light intensity (LI) using a digital quantum sensor, air temperature (AT) and humidity (AH) with a handheld thermo-hygrometer, and soil temperature (ST) and moisture (SM) at 5 cm depth using a portable probe. For each variable, three consecutive readings were taken within 15 s and averaged to represent the quadrat value. To minimize temporal variation, all microclimatic measurements within a plot were conducted synchronously by multiple observers. All measurements were performed around 10:00 a.m. under clear, rain-free conditions to reduce the effects of time and weather variability.

2.4 Soil sampling and determination

Within each 2 m × 2 m small quadrat, representative microsites were excavated for soil profiles to a depth of 20 cm. Morphological features, including soil type and underlying parent material, were documented. Soil samples were collected using an auger from the surface layer (0–10 cm) and subsurface layer (10–20 cm). Soil samples for physical property analysis were collected using a cutting ring, with approximately 10 g of soil transferred into an aluminum box. For chemical property analysis, soil samples collected from five points within the same soil horizon of each small plot were thoroughly mixed, and a representative subsample of about 1 kg was obtained using the quartering method and placed in a ziplock bag.

The soil sample determination method is as follows: soil water content (SWC) was determined using the oven-drying method, while soil bulk density (BD), maximum water holding capacity (MW), capillary water holding capacity (CW), field water capacity (FW), capillary porosity (CP), and non-capillary porosity (NCP) were measured using the ring knife method. pH was measured using a potentiometer; organic matter (OM) was determined via potassium dichromate oxidation; total nitrogen (TN) was measured by the Kjeldahl method; total phosphorus (TP) was determined by the molybdenum-antimony colorimetric method; total potassium (TK) was measured by wet digestion using sodium hydroxide; alkaline hydrolyzable nitrogen (AN) was measured via microdiffusion after alkaline hydrolysis; available phosphorus (AP) was measured by the Olsen method; and available potassium (AK) was determined using flame atomic absorption spectrophotometry (Ali et al., 2019). The descriptive statistics of soil and litter-related factors are provided in Supplementary Table S1.

2.5 Statistical analysis

Shrub- and herb-layer diversity was calculated using the Margalef richness index (D), Shannon-Wiener index (H), and Pielou evenness index (J) with the vegan package in R 4.4.2. Indices were computed as follows:

D = ( S − 1 ) / ln N (1)

H = − ∑ i = 1 s P i ln P i (2)

J = H / ln S (3)

where S is the total number of species; Pᵢ is the relative abundance of species i (P_i = N_i/N); and i = 1, 2, …, S.

Environmental factors examined in this study included microclimatic variables (LI, AT, AH, ST, and SM), soil physicochemical properties (BD, SWC, MW, CW, FW, CP, NCP, pH, OM, TN, TP, TK, AN, AP, AK, and C:N, C:P, N:P ratios), litter characteristics (LT and LC), and topographic factors (elevation, slope, aspect, and slope position). Soil physicochemical properties were expressed as the arithmetic mean of values from two soil layers. Qualitative factors were numerically coded as follows: slope aspect was classified into sunny and shady slopes (coded 1–2, respectively), and slope position was classified into upper, middle, and lower slopes (coded 1–3, respectively). Prior to redundancy analysis, all variables were standardized using z-scores to remove differences in measurement units. Variables with a variance inflation factor (VIF) greater than 5 were excluded to eliminate multicollinearity. In the redundancy analysis, forward selection with Monte Carlo permutation tests (999 permutations) was first applied to identify environmental factors that significantly influenced understory species diversity. The selected significant factors were then subjected to redundancy analysis with understory species diversity.

Mean values of all variables were calculated across small quadrats within each gap and CK plot. Subsequently, differences among indices across gap classes were analyzed using one-way ANOVA, and differences between gaps and CK were assessed using paired-samples t-tests. For gaps of the same class, linear mixed-effects models were used to examine differences in indices at different within-gap positions, with position included as a fixed effect and gap class as a random effect (Luo et al., 2020; Messier et al., 2010).

Data normality was first tested using the Kolmogorov–Smirnov (K-S) test in SPSS 25.0. Variables that violated normality assumptions (light intensity and herb-layer species diversity indices) were transformed using the Box-Cox method, with the optimal λ value determined by maximum likelihood estimation. One-way ANOVA and paired-samples t-tests were also conducted in SPSS 25.0. In R 4.4.2, linear mixed-effects models were performed with the lme4 packages. Understory species diversity indices were calculated using the vegan package, and structural equation models were fitted with the piecewiseSEM package. Redundancy analysis (RDA) was conducted in Canoco 4.5. Figures were generated in Origin 2024, with preliminary processing in Excel 2010.

3 Results

3.1 Microclimate among and within gaps

Light intensity increased progressively with gap class, and all gap classes exhibited significantly higher light levels than CK (p < 0.05), indicating that gap formation markedly altered the understory light environment (Figure 3a). Air temperature and humidity exhibited contrasting responses to gap development, with Class III and larger gaps showing significantly higher air temperatures and lower air humidity than Class I and II gaps (p < 0.05). Compared with CK, all gap classes had higher air temperatures and lower air humidity (p < 0.05; Figures 3b,c).

Soil temperature displayed a unimodal pattern across gap classes, whereas soil moisture showed an overall decreasing trend with increasing gap size. Class I gaps had significantly higher soil moisture than the other gap classes (p < 0.05), while differences in soil temperature among gap classes were relatively minor, with no significant differences detected between gap classes and CK (p > 0.05; Figures 3d,e).

Light intensity decreased from center to edge across all gap classes, with significantly higher values at the center than at the edge (p < 0.05; Figure 4). No significant differences were detected among within-gap positions for air temperature, air humidity, soil temperature, and soil moisture. These results support Hypothesis 1 in that gap size exerts a stronger influence on microclimate than within-gap position.

3.2 Species diversity of shrub- and herb-layers among and within gaps

The three species diversity indices of the shrub- and herb-layers were calculated using Equations 1–3. Gap size had an overall weak effect on shrub-layer species diversity, with Pielou evenness being significantly higher only in Class IV than in Class V gaps (p < 0.05; Figures 5a–c). In contrast, herb-layer species diversity responded more strongly to gap size. Margalef richness and Shannon-Wiener diversity were significantly higher in Class I than in Classes II-V, while Pielou evenness was significantly higher in Class I than in Classes II, III, and V (p < 0.05; Figures 5d–f). However, neither shrub- nor herb-layer species diversity indices differed significantly between any gap size class and CK (p > 0.05).

Within-gap spatial variation in shrub- and herb-layer diversity indices (Margalef richness, Shannon–Wiener diversity, and Pielou evenness) was negligible, with no significant positional differences in any size class (p > 0.05; Figure 6). Thus, understory diversity exhibited minimal spatial heterogeneity within gaps. These results support Hypothesis 1 in that gap size has a greater effect on species diversity in the shrub- and herb-layers than within-gap position.

Shrub-layer diversity indices showed no correlation with microclimatic variables in either gaps or CK (p > 0.05; Figures 7a,b). In gaps, herb-layer Shannon-Wiener diversity and Pielou evenness correlated negatively with soil moisture (p < 0.05; Figure 7c). In CK, herb-layer Margalef richness and Shannon-Wiener diversity correlated negatively with air temperature and positively with air humidity (p < 0.05), whereas all three indices were strongly negatively associated with soil moisture (p < 0.01; Figure 7d). Overall, gap formation may have weakened the relationships between herb-layer species diversity and microclimatic factors, with soil moisture remaining the microclimatic variable most closely associated with herb-layer diversity.

3.3 Effects of environmental factors on understory species diversity

Forward selection combined with Monte Carlo permutation tests (Table 1) showed that, shrub-layer species diversity in gaps was highly significantly correlated with soil total phosphorus and litter thickness (p < 0.01), and significantly correlated with soil N:P ratio and soil water content (p < 0.05). In CK, shrub-layer species diversity was highly significantly correlated with soil total phosphorus (p < 0.01) and significantly correlated with aspect, soil field capacity, and soil temperature (p < 0.05).

Table 1. Forward selection and Monte Carlo permutation test of environmental factors.

Only the main environmental factors with significance levels less than 0.05 are listed in the table.

Herb-layer species diversity in gaps was highly significantly correlated with soil pH (p < 0.01) and significantly correlated with soil moisture, litter thickness, and soil total potassium (p < 0.05). In CK, Herb-layer species diversity was highly significantly correlated with soil capillary porosity (p < 0.01) and significantly correlated with soil total potassium, aspect, and soil field capacity (p < 0.05).

Redundancy analysis was conducted to examine the relationships between understory species diversity indices and the main environmental factors (Figure 8). In gaps, shrub-layer species diversity was positively correlated with soil total phosphorus and soil N:P ratio, but significantly negatively correlated with soil water content and litter thickness (Figure 8a). In CK, shrub-layer species diversity was positively correlated with soil total phosphorus, field capacity, and slope aspect, but negatively correlated with soil temperature (Figure 8b). Herb-layer species diversity in gaps was positively correlated with soil pH and soil moisture, but negatively correlated with litter thickness and soil total potassium (Figure 8c). In CK, herb-layer species diversity was positively correlated with soil capillary porosity and field capacity, but negatively correlated with soil total potassium and slope aspect (Figure 8d).

Overall, soil total phosphorus and potassium exerted strong effects on shrub- and herb-layer species diversity in both habitats. Furthermore, litter thickness was a key determinant of shrub- and herb-layer diversity in gaps, whereas aspect and field capacity predominated in CK.

SEM was used to disentangle interactions among environmental drivers and their direct and indirect influences on understory diversity. The three diversity indices were strongly intercorrelated (Figure 8); consequently, we reduced them via PCA and used PC1 (> 90% variance explained) as a composite metric of understory in the SEMs (Table 2).

Table 2. Principal component analysis of species diversity in the shrub and herb-layers.

The structural equation model of understory species diversity is shown in Figure 9 and Table 3. In gaps, shrub-layer diversity was strongly positively influenced directly by soil total phosphorus (β = 0.47) and was directly negatively influenced by soil moisture (β = −0.45). Additionally, soil moisture exerted a significant indirect positive effect via the soil N:P ratio (β = 0.17), and gap size indirectly affected diversity through the soil N ratio (β = −0.07; Figure 9a; Table 3). This result supports Hypothesis 2 in that gap size indirectly influences shrub-layer species diversity through soil factors. In CK, shrub-layer diversity was driven chiefly by soil total phosphorus (β = 0.34). The dominant indirect pathway was aspect influencing soil total phosphorus (Figure 9b; Table 3).

Table 3. Standardized path coefficients of the structural equation model.

SMA, shrub-layer Margalef richness; SSW, shrub-layer Shannon-Wiener diversity; SPL, shrub-layer Pielou evenness; HMA, herb-layer Margalef richness; HSW, herb-layer Shannon-Wiener diversity; HPL, herb-layer Pielou evenness; LI, light intensity; AT, air temperature; AH, air humidity; ST, soil temperature; SM, soil moisture.

Herb-layer diversity in gaps was shaped mainly by direct negative effects of litter thickness (β = −0.35) and direct positive effects of soil pH (β = 0.30). Litter thickness further acted indirectly via soil pH (β = −0.10), and gap size indirectly influenced diversity via soil moisture (β = −0.08; Figure 9c; Table 3). This result supports Hypothesis 2 in that gap size indirectly influences herb-layer species diversity through microclimate factors. In CK, capillary porosity exerted the strongest direct positive effect (β = 0.52). Aspect acted solely indirectly through capillary porosity and soil total potassium. Capillary porosity further influenced diversity indirectly through field capacity. Gap formation amplified the role of soil total potassium in herb-layer diversity (Figure 9d; Table 3).

4 Discussion

4.1 Effects of gap size and within-gap position on microclimate

Gaps caused the strongest microclimatic shift in light intensity, with smaller changes in air and soil temperature and moisture. Within-gap microclimatic heterogeneity was minimal, except for light. This aligns with the globally recognized buffering capacity of closed canopies against temperature extremes (De Frenne et al., 2019; De Lombaerde et al., 2022). CK maintained stable subcanopy temperature and humidity, producing smaller gap–CK contrasts in temperature than in light. Additionally, the high humidity and modest diurnal temperature range typical of the subtropical monsoon climate further dampened extreme fluctuations in gap temperature and humidity. Prior work indicates that gap impacts on soil temperature and moisture generally strengthen with gap age or size (Horváth et al., 2023; Lenk et al., 2024). Soil temperature showed a unimodal pattern, peaking in intermediate gap classes, whereas soil moisture declined overall but remained significantly higher than CK in intermediate-sized gaps. This accords with meta-analyses demonstrating that gap creation accelerates surface evaporation and litter decomposition, generating pulsed changes in shallow-soil moisture and organic matter across forest types (Hou et al., 2024). In the infertile red soils of our site, coupled moisture–temperature–evaporation processes may have contributed to transient wetting in small gaps but comparatively greater desiccation in large gaps.

Light intensity decreased consistently from gap center to edge, whereas air temperature, air humidity, and soil moisture exhibited no significant positional differences. Unlike the marked microclimatic gradients typical of temperate and alpine forests (Von Arx et al., 2012; De Frenne et al., 2021), our strong light gradient coupled with weak temperature and humidity gradients likely resulted from rainy-season sampling under high humidity and low wind speeds, which curtailed air exchange between gaps and the forest matrix (Dai et al., 2024). Moreover, shallow, instantaneous measurements of soil temperature and moisture prevented detection of diurnal and seasonal dynamics. Overall, gaps in subtropical P. massoniana plantations strongly altered light regimes but had minor effects on air and soil temperature and moisture (Chen and Huang, 2024).

4.2 Effects of gap size and within-gap position on understory species diversity

In our study, shrub and herb-layer diversity indices did not differ significantly between forest gaps and CK. However, herb-layer diversity peaked in Class I gaps and declined with increasing gap size. This pattern contrasts with findings from other P. massoniana plantations. In Sichuan, China, Lyu et al. (2021b) reported significantly lower herb-layer diversity in gaps than in CK plots but found that intermediate-sized gaps were associated with greater soil fertility and species diversity. Similarly, Liu et al. (2025) observed higher shrub- and herb-layer diversity in artificially created gaps than in CK. These differences likely reflect the disease-induced origin of most gaps in our study (pine wilt disease), which constituted abrupt disturbances. Pre-disturbance stands were likely degraded and homogenized, with impoverished soil nutrients and seed banks, resulting in a lower ecological baseline than artificially created or windthrow gaps (Chen et al., 2024). Additionally, most gaps were approximately 10 years old, still in early successional stages. Gap-facilitated increases in regeneration and diversity typically manifest during mid-to-late succession (Tong et al., 2024). In the short term, dominant species such as Dicranopteris dichotoma may quickly monopolize light and suppress recruitment of subordinate herbs. Herb-layer species diversity in the gaps likely reflects the strong influence of dominant species on community structure and resource allocation patterns, rather than changes in species richness alone.

Shrub and herb-layer diversity indices showed no significant positional differences across gap size classes, reflecting low within-gap spatial heterogeneity. This differs from subtropical broadleaved forests, which commonly exhibit pronounced center-to-edge-to-understory diversity gradients (Ren et al., 2021; He et al., 2019), but supports our observation of weak within-gap resource gradients (temperature and humidity), thus limiting species segregation along microenvironmental gradients.

4.3 Soil nutrients and litter as regulators of understory diversity

In gaps, shrub-layer diversity was positively influenced by soil total phosphorus but negatively by soil moisture, whereas herb-layer diversity was negatively associated with litter thickness and soil total potassium but favored by higher soil pH and moisture. These findings demonstrate that nutrient, moisture, and litter effects collectively overshadowed direct influences of gap size.

Gap formation typically decreases soil organic carbon yet increases available nitrogen and phosphorus, with these effects intensifying as gaps enlarge and age (Tong et al., 2024). This pattern is consistent with the positive association observed between soil total phosphorus and shrub-layer diversity in our study. Higher total phosphorus may indicate a larger potential phosphorus pool that can be gradually mobilized through microbial mineralization, root exudation, and mycorrhizal activity, potentially improving nutrient conditions for shrub establishment. Similarly, Chen et al. (2024) identified total phosphorus as an important correlate of understory plant diversity in Pinus massoniana plantations of different ages. By contrast, Yu et al. (2022) reported divergent associations of soil nutrients with shrub- and herb-layer diversity in a temperate mixed broadleaved-coniferous forest. These contrasting findings suggest that nutrient-diversity relationships may depend on forest type and vegetation layer. Moreover, the negative direct association between soil water content and shrub-layer diversity, together with its positive indirect association through an increased soil N ratio, suggests that moisture may have context-dependent relationships with nutrient availability and plant competition. In precipitation-rich subtropical red-soil regions, excessive moisture induces hypoxia and leaching, favoring flood-tolerant species while reducing overall richness (Du et al., 2025). Conversely, moderate moisture improves phosphorus availability and root uptake (Zheng et al., 2025). As shown by Megahed et al. (2023), geomatics-based recharge mapping highlights the critical role of soil moisture in vegetation dynamics, consistent with our findings in P. massoniana gaps.

In gap herb layers, litter thickness and soil total potassium were negatively associated with species diversity, potentially reflecting the combined effects of physical barriers and nutrient imbalance. A thick litter layer may impede seedling emergence and herb germination while delaying surface warming and limiting light penetration (Harpole et al., 2016). Wang et al. (2022) found that litter accumulation generally reduced seedling emergence and survival, with stronger negative responses at greater litter depths. Elevated soil potassium can alter competitive hierarchies, allowing dominant species with higher resource-use efficiency to suppress subordinates and reduce diversity (Ou et al., 2020). Additionally, Herb-layer diversity increased with soil pH. This accords with evidence that moderate pH increases alleviate nutrient limitations in acidic red soils, promoting species coexistence (Chen et al., 2024; Tang et al., 2019).

Structural equation modeling confirmed that gap size had no significant direct effect on understory diversity, acting mainly indirectly through soil moisture and nutrient status. This contrasts with studies attributing direct control to gap area (Tinya et al., 2025; Zhou et al., 2024), but aligns with long-term studies identifying soil carbon, nitrogen, phosphorus, and moisture as primary succession drivers, where area effects are often overshadowed by multifactor interactions (Wu et al., 2025; Yu et al., 2022). Specifically, gap size indirectly affected shrub-layer species diversity by altering soil N:P ratio. Larger gaps may modify nutrient mineralization and leaching, thereby shifting the relative availability of N and P and influencing shrub species establishment and coexistence (Bagnato et al., 2021). Consistent with this interpretation, Goossens et al. (2026) reported that plant–soil feedbacks depend on nutrient availability and stoichiometry. Gap size also indirectly affected herb-layer species diversity through soil moisture. Greater solar radiation and evaporative demand in larger gaps may reduce soil moisture, thereby limiting the germination and survival of herbaceous species (Koelemeijer et al., 2024). Paulssen et al. (2026) found that herbaceous species richness increased with soil moisture, consistent with the positive relationship between soil moisture and herb-layer diversity observed in our study. In contrast, Lenk et al. (2024) reported that soil moisture increased with gap size in a European mixed floodplain forest. This discrepancy may reflect differences in forest hydrological regimes and the range of gap sizes examined. In floodplain forests, reduced canopy interception may increase soil water inputs, whereas enhanced radiation and evaporative losses may be the dominant controls in our study area.

Overall, restoration of pine wilt disease-affected P. massoniana plantations could prioritize creating Class I gaps (0.50 ≤ D/H < 0.75) while retaining healthy gap-border trees to limit further expansion. Larger gaps should be regularly monitored for Dicranopteris dichotoma cover and the recruitment of other understory species. Where necessary, targeted removal of D. dichotoma could be combined with enrichment planting of native broadleaved species. Soil management should be guided by measurements of pH, moisture, available phosphorus, and available potassium. In phosphorus-deficient areas, native broadleaved species with deep roots and rapidly decomposing litter could be introduced to improve nutrient cycling. Blanket fertilizer application should be avoided, particularly potassium fertilization where potassium limitation has not been demonstrated (Liu et al., 2023). Where thick litter is associated with poor seedling recruitment, localized litter removal could be tested while monitoring understory recovery, soil moisture, and surface erosion (Ma et al., 2023).

Although regenerating P. massoniana age and deadwood decay class are commonly used as indicators for estimating gap age, a standardized, accurate, and universally applicable method for gap age determination has not yet been established. Future studies could incorporate carbon isotope analysis to further improve the precision of gap age estimation (Yan et al., 2022). Besides, future studies should employ automated monitoring systems to enable longer-term continuous observations and controlled experiments, to clarify the scope, mechanisms, and optimal strategies of gap-based management for plantation restoration and multifunctionality, to provide robust guidance for rehabilitating pine wilt disease-affected P. massoniana plantations.

5 Conclusion

Our results showed that compared with CK, forest gaps had significantly greater understory light availability. Gap size was associated with differences in microclimate and shrub- and herb-layer species diversity, whereas within-gap position showed comparatively weak associations with understory diversity. Understory diversity was jointly related to soil nutrients, moisture conditions, and litter characteristics, with gap size acting primarily through indirect pathways involving these factors. In pine wilt disease-affected P. massoniana plantations, creating small gaps may support greater understory plant diversity. These findings can inform understory habitat restoration and sustainable gap management in Xingguo County and ecologically similar regions.

Generative AI statement

The author(s) declared that Generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/ffgc.2026.1933494/full#supplementary-material

Supplementary material Supplementary material at the source

Footnotes

Edited by: Ram P. Sharma, Tribhuvan University, Nepal

Reviewed by: Weijie Yu, Yulin University, China

Daocheng Ma, Guangxi University, China