Effects of Prescribed Burning on Species Diversity of Understory in Pinus yunnanensis Forests of Southwestern China

Abstract

The Pinus yunnanensis forest of southwestern China represents a unique and ecologically critical vegetation type, historically shaped by fire disturbances. To mitigate catastrophic wildfire risks, prescribed burning has been widely implemented as a management tool in these ecosystems. However, its effects on plant community structure and biodiversity remain insufficiently quantified. To investigate the specific changes in plant community characteristics caused by prescribed burning, this study was conducted in the Pinus yunnanensis forest in Zhaobi Hill, Xinping county. Our results revealed that prescribed burning induced differential effects on understory communities while exerting negligible effects on canopy tree composition. In the shrub layer, the number of shrub species decreased from 26 to 20, accompanied by a complete extirpation of arboreal saplings. Dominance hierarchies shifted markedly, transitioning from Lithocarpus mairei and Pinus yunnanensis regeneration cohorts in unburned plots to fire-adapted species Duhaldea cappa and Craibiodendron stellatum. Concomitantly, the average height of shrubs had a significant reduction in burning plots. Contrastingly, the number of herb species increased from 30 to 37 in burning plots, with non-significant alterations in abundance, height, and importance values. Prescribed burning significantly decreases the α species diversity of shrubs, but only has minimal effects on the α species diversity indices of herbs. Overall, prescribed burning appears to be the primary factor affecting the species diversity index of shrubs, while altitude, forest structure, and soil nutrient content exert greater influences on the species diversity index of the herbaceous layer. Prescribed burning was the dominant factor shaping the community structure and species diversity of the shrub layer, and the missing saplings of trees in the shrub layer might influence future forest succession in the long term.

1. Introduction

Pinus yunnanensis is a unique and indigenous coniferous tree species found in southwestern China. The Pinus yunnanensis forest represents the dry subtropical regions of western China and is widely spread across the Yunnan-Guizhou Plateau, with central Yunnan being at its core area [1]. This forest type is abundant in Yunnan Province and has significant carbon storage potential [2,3]. While most of the Pinus yunnanensis forests today are man-made, having been established through air-seeding in the late 1970s, a few original pristine forests still exist in remote areas of northwestern and central Yunnan [2]. The Pinus yunnanensis forest represents a category of forests that have evolved from subtropical evergreen broad-leaved forests as a result of prolonged droughts and frequent wildfires [1,2,3]. This forest type is particularly vulnerable to fire disturbances due to its inherent characteristics and environmental conditions. Specifically, the natural shedding of lower branches by Pinus yunnanensis trees facilitates fire propagation, while the dry and warm climate of the regions where these forests are located further exacerbates their susceptibility to fire [4,5]. Prescribed burning, a widely used technique in various regions, refers to controlled fires that aim to reduce fire hazards in forest ecosystems and modify the characteristics of fire disturbances by reducing the fuel loads on forest surfaces [6,7,8,9,10,11,12,13]. This practice is commonly employed in different ecosystems, including the South African fynbos [7], woodlands and forests of Australia [9,10,11,12,13], fire-prone areas in Mediterranean regions [14,15], fire-suppressed forests in the USA [16,17,18,19,20], and fire-prone ecosystems in China [3,21,22,23].

Prescribed burning has an impact on the structure of plant communities and the diversity of species in forests, and these effects vary depending on the different plant strata and types of vegetation [10,12,13,17,18,22,24,25,26,27,28,29,30,31,32,33]. Although low-intensity prescribed burning has minimal effects on the overstory strata of forests and herbaceous plants, it significantly reduces the cover and stem density of shrubs in the midstory [6,17,18,19,20,23,24,25,27,28,29,33]. Frequent prescribed burning has been found to have detrimental effects on the frequency and relative abundance of important shrubs in fynbos ecosystems [18,24]. However, in dry sclerophyll forests in southeastern Australia, burning may have a positive impact on the species richness of plants less than 1 m in height, while negatively affecting taller plants. It could also reduce the density and canopy cover of the overstory, leading to the loss of woody seedlings and saplings, but increasing the abundance and richness of understory species [18]. In broadleaf forests, burning has been shown to significantly enhance the richness of vascular plants, exotic plants, and herbs [27,34]. Similarly, in oak savanna ecosystems, burning has been observed to increase canopy openness and the percentage of vegetative cover in the understory [29]. On the other hand, the effects of burning on the diversity of vascular plant species and exotic species in Arizona Ponderosa pine forests are more subtle [28]. Understory vegetation plays a crucial role in various aspects of the forest ecosystem, including nutrient cycling, water conservation, forest climate regulation, and animal habitats [17,18], and the changes in the community structure and species diversity in the understory are essential for the overall stability of the forest ecosystem.

Prescribed burning primarily affects understory plants, while its impact on the overstory is generally minimal [22,24,34]. The changes in understory vegetation structure caused by burning may influence the future development of the forest ecosystem [18]. While Pinus yunnanensis forests generally have a lower abundance of understory vegetation, the importance of understory for the stability of forest ecosystems should not be overlooked [3]. However, the effects of prescribed burning on the species composition and diversity of the understory in Pinus yunnanensis forests remain uncertain. We hypothesized that prescribed burning would significantly affect the species composition and plant heights of the understory vegetation, and the impact of burning on the canopy strata of Pinus yunnanensis forests should be minor. It is possible that these effects could be influenced by other factors, such as altitude and soil nutrient levels. To investigate this further, we conducted a comprehensive plant study in both the prescribed burning area and an adjacent unburned area in the Pinus yunnanensis forests of central Yunnan. Our study aimed to address the following questions: (1) How does prescribed burning influence the plant community structure and composition of Pinus yunnanensis forests? (2) What are the differential effects of prescribed burning on the arboreal, shrub, and herbaceous layers within Pinus yunnanensis forests? (3) To what extent do other factors, such as altitude or soil nutrient levels, interact with the effects of prescribed burning on the plant community structure of the understory in this region?

2. Materials and Methods

2.1. Study Area

This study was conducted in Pinus yunnanensis forests, which were aerially seeded in Zhaobi Hill, Xinping county, Yunnan province, southwestern China (Figure 1). The region has a subtropical humid monsoon climate with distinct wet and dry seasons, an average annual temperature of 18.1 °C, an annual mean precipitation of 869 mm, and an annual total sunshine duration of 2838 h. The dominant soil type in this region is red soil, and the Pinus yunnanensis forest was established through afforestation by aerial seeding in the 1970s. The vegetation community of the forest includes three layers: the arboreal layer, the shrub layer, and the herb layer. The arboreal layer includes Pinus yunnanensis, Keteleeria evelyniana, Cyclobalanopsis glaucoides, Castanopsis orthacantha, Castanopsis delavayi, and Quercus acutissima. The shrub layer and herb layer constitute the understory; plants in the shrub layer include Vaccinium duclouxii, V. fragile, Lyonia ovalifolia, Craibiodendron stellatum, and Duhaldea cappa, and the herb layer is mainly composed of Anaphalis sp., Arundinella setosa, Capillipedium assimile, Elsholtzia rugulosa, E. bodinieri, Eulalia pallens, E. quadrinervis, Hedyotis sp., Leontopodium sp., Pteridium aquilinum, Erigeron sp., Pseudognaphalium sp., and Laggera alata. Pinus yunnanensis forests are middle-aged and originated from natural Pinus yunnanensis forests enhanced by aerial seeding and management measures during the 1970s. Prescribed burning has been a consistent practice in this region since 2016 as a preventative measure against forest fires (Figure 2). These prescribed burns typically occur in late January or early February, with low-intensity fires being utilized between 2016 and 2019 and discontinued in 2020 and 2021. The burned area is restored for three years after the last prescribed burning period.

2.2. Field Measurements

In January 2022, field surveys were conducted in unburned and planned burned areas. Fourteen sample plots were established in each area, with each plot consisting of three 20 m × 20 m subplots. The distance between each plot was approximately 200–500 m, while the distance between each subplot was less than 100 m. This resulted in a total of 84 subplots. All subplots had a similar aspect, and other basic information, such as altitude, slope, and canopy density, showed minimal variation between prescribed burning and unburned plots (Figure 1 and Figure 2). The survey plots were strategically placed at least 50 m from the road, with elevations ranging from 1950 to 2200 m. These plots were predominantly located on sunny and moderate slopes (

Table 1. Basic information of sample plots.

Plots	Altitude(m)	Slope (°)	Canopy Density	DBH (cm)	Basal Area (m2/ha)	Height (m)	Tree Density (Stems/ha)	CBI Index
CK	2043 ± 42 a	13.0 ± 4.0 a	0.43 ± 0.05 a	17.50 ± 1.92 a	3515.06 ± 918.52 a	8.27 ± 0.93 a	1235.29 ± 361.75 a	0.75
PB	2019 ± 33 a	10.0 ± 2.0 a	0.38 ± 0.04 a	18.37 ± 2.56 a	3295.66 ± 859.53 a	8.93 ± 1.40 a	1163.39 ± 461.65 a	0

The values in the table are mean ± standard deviation, and different letters indicate a significance level of α = 0.05. The CBI index estimation is presented in Appendix A Table A1.

). Within each 20 m × 20 m plot, the diameters at breast height (DBH) of individual plants, with a DBH greater than 2 cm, were measured and recorded. Additionally, tree height and height under lower branches in the tree layer were recorded. According to the methods for large forest plot surveys [35,36], in each 20 m × 20 m plot, five 5 m × 5 m quadrats were set in a clover shape to investigate the height, crown breadth, and number of plant species in the shrub layer. Furthermore, within each of the five shrub quadrats, five 1 m × 1 m sub-quadrats were set to investigate the number and coverage of plants in the herbaceous layer. Lastly, five soil samples were collected from the 0–10 cm soil layer to measure the chemical stoichiometric characteristics of soil C, N, P, and K.

2.3. Laboratory Analysis

All trees and shrubs in all sample plots were classified based on their respective species, following the guidelines provided in the Yunnan Flora and the unpublished Xinping Flora. For herbs, identification was carried out at the specific species level, and when it was not possible to identify the species of the herbaceous plant, only the genus was identified (

Table A2. Species family genus in the shrub layer of Pinus yunnanensis forests.

Species	Family Name	Generic Name	Group
Vernonia cumingiana	Asterceae	Vernonia	Native perennial shrub
Duhaldea cappa	Asterceae	Duhaldea	Native perennial shrub
Craibiodendron stellatum	Ericaceae	Craibiodendron	Native perennial shrub
Vaccinium bracteatum	Ericaceae	Vaccinium	Native perennial shrub
Rhododendron simsii	Ericaceae	Rhododendron	Native perennial shrub
Lyonia ovalifolia	Ericaceae	Lyonia	Native perennial shrub
Vaccinium fragile	Ericaceae	Vaccinium	Native perennial shrub
Vaccinium duclouxii	Ericaceae	Vaccinium	Native perennial shrub
Campylotropis hirtella	Fabaceae	Campylotropis	Native perennial shrub
Lithocarpus mairei	Fagaceae	Lithocarpus	Native perennial shrub
Cyclobalanopsis glaucoides	Fagaceae	Cyclobalanopsis	Native perennial shrub
Quercus acutissima	Fagaceae	Quercus	Native perennial shrub
Castanopsis delavayi	Fagaceae	Castanopsis	Native perennial shrub
Lithocarpus hancei	Fagaceae	Lithocarpus	Native perennial shrub
Quercus variabilis	Fagaceae	Quercus	Native perennial shrub
Engelhardtia spicata	Juglandaceae	Engelhardtia	Native perennial shrub
Melastoma polyanthum	Melastomataceae	Melastoma	Native perennial shrub
Ficus tikoua	Moraceae	Ficus	Native perennial shrub
Morella rubra	Myricaceae	Morella	Native perennial shrub
Myrica nana	Myricaceae	Myrica	Native perennial shrub
Anneslea fragrans	Pentaphylacaceae	Anneslea	Native perennial shrub
Pinus yunnanensis	Pinaceae	Pinus	Native perennial shrub
Keteleeria evelyniana	Pinaceae	Keteleeria	Native perennial shrub
Raphiolepis indica	Rosaceae	Raphiolepis	Native perennial shrub
Osteomeles schwerinae	Rosaceae	Osteomeles	Native perennial shrub
Rubus calycinus	Rosaceae	Rubus	Native perennial shrub
Symplocos sumuntia	Symplocaceae	Symplocos	Native perennial shrub
Schima wallichii	Theaceae	Schima	Native perennial shrub
Tetrastigma yunnanense	Vitaceae	Tetrastigma	Native perennial shrub

and

Table A3. Species family genus in the herb layer of Pinus yunnanensis forests.

Species	Family Name	Generic Name	Group
Centella asiatica	Apiaceae	Centella	Native perennial herb
Ageratina adenophora	Asteraceae	Ageratina	Native perennial herb
Pseudognaphalium adnatum	Asteraceae	Pseudognaphalium	Native perennial herb
Leontopodium dedekensii	Asteraceae	Leontopodium	Native perennial herb
Anaphalis margaritacea	Asteraceae	Anaphalis	Native perennial herb
Erigeron canadensis	Asteraceae	Erigeron	Native perennial herb
Laggera alata	Asteraceae	Laggera	Native perennial herb
Gerbera delavayi	Asteraceae	Gerbera	Native perennial herb
Crassocephalum crepidioides	Asteraceae	Crassocephalum	Native perennial herb
Asteraceae japonica	Asteraceae	Asteraceae	Native perennial herb
Taraxacum mongolicum	Asteraceae	Taraxacum	Native perennial herb
Gamochaeta pensylvanica	Asteraceae	Gamochaeta	Native perennial herb
Ainsliaea yunnanensis	Asteraceae	Ainsliaea	Native perennial herb
Carex cruciata	Cyperaceae	Carex	Native perennial herb
Cyperus rotundus	Cyperaceae	Cyperus	Native perennial herb
Parochetus communis	Fabaceae	Parochetus	Native perennial herb
Desmodium microphyllum	Fabaceae	Desmodium	Native perennial herb
Heteropogon contortus	Fabaceae	Heteropogon	Native perennial herb
Lespedeza bicolor	Fabaceae	Lespedeza	Native perennial herb
Desmodium concinnum	Fabaceae	Desmodium	Native perennial herb
Juncus effusus	Juncaceae	Juncus	Native perennial herb
Elsholtzia rugulosa	Lamiaceae	Elsholtzia	Native perennial herb
Elsholtzia bodinieri	Lamiaceae	Elsholtzia	Native perennial herb
Clinopodium repens	Lamiaceae	Clinopodium	Native perennial herb
Scutellaria indica	Lamiaceae	Scutellaria	Native perennial herb
Scutellaria barbata	Lamiaceae	Scutellaria	Native perennial herb
Ophiopogon japonicus	Liliaceae	Ophiopogon	Native perennial herb
Arundinella setosa	Poaceae	Arundinella	Native perennial herb
Eulalia pallens	Poaceae	Eulalia	Native perennial herb
Eulalia quadrinervis	Poaceae	Eulalia	Native perennial herb
Themeda triandra	Poaceae	Themeda	Native perennial herb
Capillipedium assimile	Poaceae	Capillipedium	Native perennial herb
Arthraxon hispidus	Poaceae	Arthraxon	Native perennial herb
Bothriochloa ischaemum	Poaceae	Bothriochloa	Native perennial herb
Miscanthus sinensis	Poaceae	Miscanthus	Native perennial herb
Imperata cylindrica	Poaceae	Imperata	Native perennial herb
Pteridium aquilinum	Pteridiaceae	Pteridium	Native perennial herb
Duchesnea indica	Rosaceae	Duchesnea	Native perennial herb
Hedyotis uncinella	Rubiaceae	Hedyotis	Native perennial herb
Hedyotis auricularia	Rubiaceae	Hedyotis	Native perennial herb
Rubia podantha	Rubiaceae	Rubia	Native perennial herb
Galium bungei	Rubiaceae	Galium	Native perennial herb
Hedyotis diffusa	Rubiaceae	Hedyotis	Native perennial herb

).

The sampled soils underwent air-drying and were subsequently passed through a US Standard 60 mesh to exclude small rocks and fine roots. The resulting mineral soils were then analyzed for total soil carbon (TC), soil organic carbon (SOC), total nitrogen (TN), hydrolysable nitrogen (A–N), available potassium (A–K), and soil available phosphorus (A–P). Additionally, the C/N ratio was calculated based on the total carbon and total nitrogen content of the soils. The TC analysis was conducted using a ZYT-2000 analyzer, while SOC was measured using a Dichromate titration method. TN and A–N were determined using the SKD-200 analyzer. A–K was measured using the FP6410 analyzer, and A-P was determined using a UV-visible spectrophotometer analyzer.

2.4. Statistical Analysis

To evaluate the structure and composition of the Pinus yunnanensis forest community, plant importance value, species abundance, and height were measured. The α diversity indices, such as the Shannon–Winner diversity index (H′), Simpson’s dominance index (D), Pielou’s evenness index (E_H), and Margalef’s richness index (F), were calculated to assess community diversity. Additionally, beta diversity indices, including the Jaccard index (β_j), Cody index (β_c), and Morisita–Horn index (C_MH), were used. For the shrub layer, H′, D, E_H, and F were utilized to measure species diversity. Due to the challenges in accurately counting the number of herbs, H′, D, and E_H were employed to assess the species diversity in the herb layers. The diversity index was calculated using the methodology described in Hill [37] and Ma et al. [38].

$$H\prime =-{\displaystyle \sum _{i=1}^S{P}_{i}ln{p}_i}.$$

(1)

$$D=1-{\displaystyle \sum _{i=1}^S{\left(P_i\right)}^2}$$

(2)

$$E_H=H^{\prime }/lnS$$

(3)

$$F=\left(S-1\right)/lnN$$

(4)

$$P_i=n_i/N$$

(5)

$${\beta }_c=\left(g\left(H\right)+l\left(H\right)\right)/2$$

(6)

$${\beta }_j=j/(a+b-j)$$

(7)

$$C_{MH}=2\times {\displaystyle \sum a{n}_i\times b{n}_i/(({\displaystyle \sum a{n}_i^2}}/a^2+{\displaystyle \sum b{n}_i^2/b^2})\times a\times b)$$

(8)

where $P_i$ indicates the proportion of plant species i in a specific sample site, $n_i$ denotes the count of individual plants belonging to species i, S denotes the number of plant species in a sample site, and N denotes the total count of all plant species in the sample site. In the equations of β_c, β_j, and C_MH, $g\left(H\right)$ represents the number of species increasing along the habitat gradient H, $l\left(H\right)$ represents the number of species decreasing along the same gradient, a represents the number of species in sample A, b represents the number of species in sample B, j represents the number of species shared by both sample A and sample B, and ${\mathit{an}}_i$ and ${\mathit{bn}}_i$ represent the coverage of individual plants belonging to species i in sample A and sample B, respectively. Different values of the Jaccard similarity index (β_j) represent different meanings. Values of 0 to 0.25 indicate extremely dissimilar community compositions, 0.25 to 0.50 represents moderately dissimilar, 0.50 to 0.75 indicates moderately similar, and 0.75 to 1.00 represents extremely similar communities. The Morisita–Horn index (C_MH) represents the coefficient of similarity between communities undergoing planned burning, while the Cody index (β_c) represents the dissimilarity of species composition.

We utilized the Kruskal–Wallis (KW) test to identify disparities in species abundance, coverage, height, and species diversity index. Additionally, we employed NMDS to explore differences in understory community structure between the burned area and the unburned area. We identified the influential factors of α species diversity by using Spearman correlation. Furthermore, we conducted Multivariate Analysis of Variance (MANOVA) to analyze the relative contribution of each factor to the α diversity indexes. Finally, we investigated the associations between species composition and various environmental factors, including fires, altitudes, upper vegetation characteristics, and soil nutrients, using the structural equation model (SEM).

NMDS is used to visually evaluate the structure of plant communities and their environmental resources (such as soil organic layer thickness, moisture content, pH, total carbon, nitrogen, and phosphorus). NMDS is a widely used multivariate analysis method, suitable for complex iterative sorting analysis of nonlinear data. It reflects the relationships between entities in a geometric space using the distances between points in a low-dimensional space through dimensional reduction and sorting, and could simultaneously handle multiple dependent and independent variables and visually display the relationships between variables in a sorting plot.

SEM is a multivariate statistical method used to establish, estimate, and test causal relationships. Compared to traditional statistical methods, SEM can study both observable variables and variables that cannot be directly observed (latent variables), and can indicate direct or indirect relationships between variables. SEM could handle multiple dependent variables simultaneously and visually display the relationships between variables in a path diagram. SEM is used to quantitatively evaluate the direct and indirect impacts of factors such as burning and topography on the surface plant communities of Pinus yunnanensis forests and conduct path analysis.

All analyses were performed in R version 4.2.1 [39]. The WX test was conducted using the “Wilcox.test” package, the NMDS analyses were performed using the “vegan” package, Spearman correlation was employed using the “cor.test” package, MANOVA was executed using the “car” and “MASS” packages, and the SEM analysis was conducted using the “sem” package.

3. Results

There were 63 recorded species from 21 plant families in our sample plots. There were 53 species from 13 families in unburned plots, and 59 species from 19 families in burned plots. The most abundant families were Asteraceae (19.44%), Fagaceae (16.67%), and Poaceae (12.50%) in all the sample plots, and the most abundant families were similar in unburned plots and burned plots. The responses of community characteristics and species composition of the Pinus yunnanensis forest to prescribed burning were varied in the arborous layer, shrub layer, and herbaceous layer.

3.1. Effects of Prescribed Burning on Community Characteristics and Species Composition of the Arboreal Layer

There was no difference in the tree species composition of the arboreal layer between prescribed burning (PB) and unburned areas (CK). The sample plots of both PB and CK included Pinus yunnanensis, Keteleeria evelyniana, Quercus acutissima, and Anneslea fragrans. Pinus yunnanensis was the only dominant tree species, and the number of Keteleeria evelyniana, Quercus acutissima, and Anneslea fragrans was almost negligible.

The KW test showed that the DBH (diameters at breast height) and the height of trees (height) were similar between the burned and unburned areas, though the DBH and height were slightly larger in the burned area than in the unburned area (Table A2, Figure 3a,b). On the other hand, the heights under the branch in the tree layers were higher in the prescribed burning plots than in the unburned plots. The mean shrub heights in the unburned plots and prescribed burning plots were 5.68 m and 6.77 m, respectively, indicating a statistically significant difference between the two groups (Figure 3c).

3.2. Effects of Prescribed Burning on the Composition and Structure of Understory Layers

3.2.1. Effects of Prescribed Burning on the Composition and Structure of Shrubs

Although prescribed burning decreased the average height and coverage of shrubs in understory vegetation (Figure 4), it altered the species community composition of shrubs (Figure 5). Prescribed burning significantly decreased the species richness of shrubs (Figure 6). The number of plant species in the shrub layer decreased from 26 in the unburned plots to 20 after prescribed burning. In addition, the abundance of plants within the shrub layer was also affected (Figure 7a). Tetrastigma yunnanense, Lithocarpus mairei, and saplings of five tree species, including Pinus yunnanensis, Quercus acutissima, Quercus variabilis, Castanopsis delavayi, and Schima wallichii, disappeared from the burned plots. The abundance of Morella rubra, Vaccinium fragile, and Keteleeria evelyniana significantly declined. On the other hand, Melastoma polyanthum, Raphiolepis indica, and Osteomeles schwerinae were newly discovered in the burned plots. Compared to the unburned plots, only Duhaldea cappa showed a significant increase in abundance, while the other co-existing species, in both burned and unburned plots, did not show such a substantial change.

Prescribed burning also resulted in a reduction in the height of most plants in the shrub layer. Vaccinium duclouxii, Vaccinium fragile, Rubus calycinus, Lyonia ovalifolia, Keteleeria evelyniana, Symplocos sumuntia, Vaccinium bracteatum, Engelhardtia spicata, and Anneslea fragrans showed significantly lower heights in the burned plots compared to the unburned plots. However, Cyclobalanopsis glaucoides and Ficus tikoua experienced a significant increase in height, excluding the newly found shrubs. Overall, prescribed burning led to a considerable decrease in the average height of the shrub layer, decreasing from 84.60 ± 28.83 cm in the unburned plots to 50.42 ± 27.71 cm in the burned plots (

Table A4. The importance value, plant abundance, and height of the shrub layer in prescribed burning and unburned plots.

Species	Abundance		Height (cm)		Importance Value (IV)
	CK	PB	CK	PB	CK	PB
Craibiodendron stellatum	12.57 ± 11.82	11.08 ± 9.05	75.42 ± 44.9	54.92 ± 28.1	0.35 ± 0.05	0.46 ± 0.15 *
Lithocarpus mairei	24.00 ± 15.66	- **	127.72 ± 64.20	- **	0.47 ± 0.05	- **
Pinus yunnanensis	6.88 ± 7.69	-	51.04 ± 15.97	-	0.35 ± 0.07	-
Morella rubra	3.50 ± 2.50	0.64 ± 0.98	204.24 ± 95.18	114.89 ± 91.46	0.30 ± 0.08	0.22 ± 0.10
Anneslea fragrans	1.50 ± 0.50	1.46 ± 0.86	43.50 ± 22.50	22.50 ± 2.50 *	0.14 ± 0.02	0.13 ± 0.03
Engelhardtia spicata	3.33 ± 3.30	0.35 ± 0.14	106.33 ± 33.99	43.00 ± 0.82 **	0.16 ± 0.03	0.10 ± 0.01
Schima wallichii	3.00 ± 2.00	-	30.50 ± 24.50	-	0.10 ± 0.03	-
Vaccinium bracteatum	7.00 ± 5.10	1.00 ± 1.56	71.44 ± 20.86	36.25 ± 16.14 *	0.22 ± 0.01	0.19 ± 0.10
Symplocos sumuntia	0.50 ± 0.50	0.17 ± 0.37	200.00 ± 12.50	65.00 ± 5.60 **	0.13 ± 0.01	0.04 ± 0.02
Keteleeria evelyniana	1.33 ± 0.94	1.05 ± 0.43	89.50 ± 57.50	22.33 ± 3.80 **	0.24 ± 0.11	0.12 ± 0.04
Cyclobalanopsis glaucoides	1.50 ± 0.50	0.43 ± 0.25	200.60 ± 119.40	210.00 ± 190.00	0.14 ± 0.08	0.24 ± 0.15
Quercus acutissima	1.67 ± 0.94	-	75.78 ± 19.51	-	0.11 ± 0.02	-
Castanopsis delavayi	2.00 ± 1.00	-	72.17 ± 48.17	-	0.19 ± 0.13	-
Lithocarpus hancei	10.00 ± 5.00	4.50 ± 3.50	65.90 ± 15.30	40.81 ± 4.19	0.34 ± 0.03	0.23 ± 0.04
Quercus variabilis	1.00 ± 0.30	-	315.00 ± 13.50	-	0.15 ± 0.02	-
Melastoma polyanthum	-	0.14 ± 0.35	-	65.00 ± 6.80		0.07 ± 0.01
Raphiolepis indica	-	0.37 ± 0.17	-	30.00 ± 0.50	-	0.16 ± 0.05
Osteomeles schwerinae	-	1.00 ± 0.03	-	105.00 ± 11.40	-	0.13 ± 0.02
Tetrastigma yunnanense	1.00 ± 0.02	-	5.00 ± 0.90	-	0.13 ± 0.01	-
Vernonia cumingiana	0.50 ± 0.34	-	243.00 ± 13.50	-	0.07 ± 0.02	-
Myrica nana	0.67 ± 0.47	-	246.50 ± 9.50	-	0.20 ± 0.01	-
Rhododendron simsii	17.00 ± 0.60	-	53.59 ± 1.60	-	0.19 ± 0.04	-
Duhaldea cappa	4.00 ± 2.00	33.45 ± 25.87 **	53.29 ± 17.48	44.23 ± 17.75	0.26 ± 0.07	0.60 ± 0.14 **
Lyonia ovalifolia	5.92 ± 4.68	8.18 ± 11.22	99.44 ± 77.75	55.02 ± 27.58 *	0.45 ± 0.13	0.37 ± 0.08 *
Vaccinium fragile	15.9 ± 12.21	2.80 ± 4.94 *	20.76 ± 11.33	14.31 ± 3.5 *	0.39 ± 0.12	0.24 ± 0.12 **
Campylotropis hirtella	0.50 ± 0.50	0.99 ± 0.86	21.00 ± 1.50	20.25 ± 0.75	0.06 ± 0.02	0.20 ± 0.10
Vaccinium duclouxii	5.25 ± 4.97	3.64 ± 1.38	51.60 ± 16.37	27.18 ± 3.70 *	0.21 ± 0.05	0.16 ± 0.01
Rubus calycinus	6.00 ± 0.40	1.12 ± 0.50	34.80 ± 3.70	18.00 ± 0.90 **	0.69 ± 0.04	0.10 ± 0.02
Ficus tikoua	1.00 ± 0.20	1.00 ± 0.03	7.00 ± 0.50	9.00 ± 0.50	0.03 ± 0.01	0.05 ± 0.02

** Represents p ≤ 0.05, * represents 0.05 < p ≤ 0.10, and “-” indicates that the species was not found in sample plots.

).

The importance value (IV) of many plants in the shrub layer showed changes in the prescribed burning plots compared to the unburned plots, although most of these changes were not easily noticeable (Figure 7b). However, the IV of Duhaldea cappa and Craibiodendron stellatum exhibited a significant increase in the prescribed burning plots, while Lyonia ovalifolia and Vaccinium fragile decreased. The dominant plants in the shrub layer also shifted from Lithocarpus mairei and Lyonia ovalifolia in the unburned area to Duhaldea cappa in the prescribed burning area. Furthermore, the NMDS analysis results provided further evidence of the notable impact of prescribed burning on the vegetation in the shrub layer. This assertion is supported by the graphical representation of NMDS at a 95% confidence interval. It is worth mentioning that the stress value obtained from ranking the NMDS for the shrub layer was 0.135, indicating that NMDS can yield accurate results (Figure 8).

3.2.2. Effects of Prescribed Burning on the Composition and Structure of the Herb Layers

Prescribed burning resulted in an increase in the number of herb species from 30 to 37 (Figure 9). Three species, including Crassocephalum crepidioides, Cyperus rotundus, and Gamochaeta pensylvanica, disappeared in the burned plots. On the other hand, 15 species, such as Parochetus communis, Themeda triandra, Erigeron canadensis, Gerbera delavayi, Centella asiatica, Clinopodium repens, Lespedeza bicolor, Asteraceae japonica, Taraxacum mongolicum, Scutellaria indica, Desmodium concinnum, Hedyotis diffusa, Bothriochloa ischaemum, Imperata cylindrica, and Ainsliaea yunnanensis, were only found in the burned plot. The majority of the newly–found species belonged to the Poaceae and Asteraceae families (Figure 10).

Prescribed burning had minimal impact on the abundance of herbs. Apart from the newly discovered species, no other species showed a significant increase in abundance. Only Pteridium aquilinum exhibited a significant decrease in abundance, apart from undiscovered species (Figure 11a). The average height of herb layers in unburned plots increased from 25.83 ± 9.61 cm to 34.36 ± 22.28 cm in the burned plots (Figure 12 and

Table A5. The importance value, coverage, and height of herbs in prescribed burning areas and in unburned areas.

Species	Abundance		Height (cm)		Importance Value (IV)
	Unburned	PB	Unburned	PB	Unburned	PB
Elsholtzia rugulosa	9.01 ± 6.96	6.83 ± 3.50	31.14 ± 14.18	30.83 ± 11.74	0.46 ± 0.06	0.49 ± 0.02
Pteridium aquilinum	30.59 ± 24.48	11.71 ± 10.43 *	60.66 ± 27.85	57.96 ± 21.96 *	0.31 ± 0.12	0.45 ± 0.05 **
Arundinella setosa	33.75 ± 26.40	17.37 ± 7.72	35.96 ± 18.53	68.94 ± 32.78 *	0.52 ± 0.13	0.35 ± 0.13
Ageratina adenophora	11.56 ± 8.02	4.53 ± 9.07	28.70 ± 14.11	16.16 ± 7.07	0.30 ± 0.03	0.29 ± 0.07
Hedyotis uncinella	1.29 ± 1.04	1.45 ± 0.57	8.38 ± 4.69	7.63 ± 4.07	0.28 ± 0.04	0.25 ± 0.03 *
Eulalia pallens	26.57 ± 24.53	17.10 ± 12.55	54.17 ± 26.05	74.66 ± 43.48	0.39 ± 0.15	0.36 ± 0.10
Eulalia quadrinervis	15.68 ± 11.01	16.66 ± 9.49	23.25 ± 11.62	29.42 ± 21.91	0.43 ± 0.11	0.28 ± 0.11 *
Parochetus communis	-	1.65 ± 1.42 **	-	2.73 ± 0.30 **	-	0.34 ± 0.01 **
Carex cruciata	20.92 ± 19.34	19.22 ± 11.5	25.92 ± 9.46	36.35 ± 9.34	0.27 ± 0.07	0.33 ± 0.05
Themeda triandra	-	22.40 ± 12.61 **	-	65.75 ± 33.8 **	-	0.33 ± 0.12 **
Capillipedium assimile	29.61 ± 12.25	12.06 ± 7.31	45.67 ± 17.17	47.25 ± 16.18	0.25 ± 0.06	0.26 ± 0.06
Hedyotis auricularia	0.57 ± 0.49	0.77 ± 0.65	5.33 ± 4.03	9.00 ± 8.39	0.18 ± 0.03	0.18 ± 0.03
Pseudognaphalium adnatum	1.15 ± 0.81	1.31 ± 0.43	6.33 ± 4.03	3.88 ± 0.22	0.14 ± 0.06	0.16 ± 0.04
Leontopodium dedekensii	6.16 ± 5.14	0.64 ± 0.47	18.57 ± 5.95	17.81 ± 14.93	0.28 ± 0.16	0.18 ± 0.01
Desmodium microphyllum	1.68 ± 1.55	0.82 ± 0.40	5.75 ± 2.75	22.97 ± 19.92	0.08 ± 0.02	0.16 ± 0.03 *
Heteropogon contortus	29.77 ± 27.67	4.60 ± 3.78	27.92 ± 1.56	11.50 ± 4.50	0.26 ± 0.12	0.11 ± 0.02
Juucus effusus	13.38 ± 4.49	1.84 ± 1.19	12.17 ± 1.03	15.75 ± 8.61	0.22 ± 0.01	0.15 ± 0.01
Anaphalis margaritacea	3.44 ± 1.68	1.52 ± 1.29	8.88 ± 6.62	11.25 ± 8.32	0.17 ± 0.03	0.16 ± 0.06
Erigeron canadensis	-	1.48 ± 0.99 **	-	5.78 ± 2.27 **	-	0.19 ± 0.01 **
Duchesnea indica	0.09 ± 0.05	3.07 ± 2.66	2.00 ± 0.40	5.78 ± 2.64	0.05 ± 0.02	0.19 ± 0.02
Elsholtzia bodinieri	14.34 ± 13.90	3.67 ± 2.33	5.38 ± 2.22	6.57 ± 2.56	0.15 ± 0.04	0.17 ± 0.06
Rubia podantha	20.81 ± 0.60	2.92 ± 0.56	39.00 ± 0.80	5.88 ± 2.13	0.15 ± 0.02	0.12 ± 0.03
Galium bungei	2.03 ± 1.20	1.84 ± 1.36	4.00 ± 1.00	3.83 ± 1.55	0.10 ± 0.01	0.12 ± 0.01
Laggera alata	3.95 ± 3.34	1.36 ± 1.02	29.50 ± 14.50	22.00 ± 19.91	0.10 ± 0.01	0.13 ± 0.06
Ophiopogon japonicus	0.09 ± 0.03	1.92 ± 1.52	20.00 ± 0.40	18.25 ± 6.25	0.05 ± 0.01	0.08 ± 0.01
Gerbera delavayi	-	6.85 ± 6.19	-	4.50 ± 2.04	-	0.12 ± 0.07
Arthraxon hispidus	17.24 ± 0.70	4.24 ± 0.52	30.00 ± 0.80	20.00 ± 1.70	0.11 ± 0.02	0.09 ± 0.01
Centella asiatica	-	0.03 ± 0.01	-	3.00 ± 0.40	-	0.04 ± 0.01
Crassocephalum crepidioides	0.06 ± 0.05	-	7.00 ± 0.40	-	0.03 ± 0.01	-
Lespedeza bicolor	-	0.90 ± 0.40	-	10.67 ± 5.3	-	0.07 ± 0.03
Asteraceae japonica	-	0.67 ± 0.01	-	3.00 ± 2.00	-	0.05 ± 0.01
Clinopodium repens	-	1.33 ± 0.20	-	4.00 ± 0.30	-	0.06 ± 0.03
Taraxacum mongolicum	-	0.36 ± 0.10	-	3.00 ± 0.06	-	0.03 ± 0.04
Scutellaria indica	-	0.34 ± 0.20	-	2.50 ± 0.03	-	0.03 ± 0.01
Desmodium concinnum	-	0.07 ± 0.04	-	10.00 ± 0.70	-	0.02 ± 0.02
Hedyotis diffusa	-	0.54 ± 0.30	-	4.00 ± 0.05	-	0.03 ± 0.01
Bothriochloa ischaemum	-	0.02 ± 0.01	-	14.00 ± 2.50	-	0.02 ± 0.01
Miscanthus sinensis	-	4.07 ± 0.30	-	40.00 ± 3.20	-	0.06 ± 0.03
Cyperus rotundus	0.46 ± 0.30	-	3.00 ± 0.20	-	0.05 ± 0.01	-
Gamochaeta pensylvanica	1.13 ± 0.4	-	3.50 ± 0.5	-	0.05 ± 0.02	-
Imperata cylindrica	-	20.53 ± 19.00	-	34.17 ± 20.17	-	0.09 ± 0.04
Scutellaria barbata	3.27 ± 2.95	3.23 ± 2.87	19.67 ± 8.96	31.80 ± 3.66	0.34 ± 0.01	0.40 ± 0.07
Ainsliaea yunnanensis	-	0.50 ± 0.03	-	17.00 ± 4.10	-	0.20 ± 0.01

** Represents p ≤ 0.05, * represents 0.05 < p ≤ 0.10, and “-” indicates that the species was not found in sample plots.

). The burning treatment partially altered the importance values of the herb layer plants. In the unburned plots, the species with the highest importance value (IV) was Arundinella setosa, while in the burned plots, it was Elsholtzia rugulosa. Except for Eulalia quadrinervis, Hedyotis uncinella, Pteridium aquilinum, and Desmodium microphyllum, which were no longer present, and the newly added species, most of the coexisting species in the burned and unburned plots showed no significant difference in their importance values (Figure 12b).

The NMDS results strongly indicate that the impact of planned burning on herbaceous-layer plants can be disregarded. This is confirmed by the graphical representation of NMDS and the 95% confidence interval. The stress values assigned to the NMDS of the herb layer were remarkably low (0.186), which highlights the effectiveness of NMDS in producing accurate results (Figure 13).

3.3. Effects of Prescribed Burning on Diversity Indices of Understory Species

The effects of prescribed burning on species diversity indices differed between shrub layers and herb layers (

Table 2. Effects of prescribed burning on the beta diversity index of the understory.

Layer	Treatments	Number of Increased Species	Number of Decreased Species	Cody (βc)	Jaccard (βj)	Morisita–Horn (CMH)
Shrub layer	Unburned–burned	3.000	10.000	6.500	0.552	0.269
Herb layer	Unburned–burned	16.000	3.000	9.500	0.558	0.736

and Figure 14). Prescribed burning had a significant impact on the Shannon–Winner diversity index (H′), Simpson’s dominance index (D), Pielou’s evenness index (E_H), and Margalef’s richness index (F) of shrub layers (Figure 14a and

Table A6. The multiple regression equation of α diversity indexes for the shrub layer.

Shrub	Equation Model	R2	AIC
Shannon	Shannon = 5.595 × Alt + 0.467 × A-K + 0.103 × Density − 0.180 × Fire − 4.283 × pH − 0.180 × A-P − 0.251 × BA	0.630	−74.463
Simpson	Simpson = 3.160 × Alt + 0.067 × Density + 0.240 × A-K − 0.082 × Fire − 0.073 × A-P − 0.120 × BA − 2.356 × pH	0.637	−106.208
Pielou	Pielou = 2.782 × Alt + 0.151 × A-K − 1.528 × pH − 0.108 × A-P − 0.241 × Trees.H − 0.103 × DBH	0.546	−105.754
Margalef	Margalef = 4.733 × Alt + 0.600 × A-K − 0.251 × Fire − 4.817 × pH − 0.273 × BA	0.489	−57.260

Alt represents altitude, Fire represents prescribed burning, Trees.H represents tree height, Density represents tree density, DBH represents tree diameter at breast height, BA represents tree basal area, A-P represents soil available phosphorus, and A-K represents soil available potassium.

), resulting in a decrease in these indices. However, the positive effects of prescribed burning on the H′ and D of herb layers were limited, and it had slightly negative effects on the E_H of herb layers (Figure 14b and

Table A7. The multiple regression equation of α diversity indexes for the herb layer.

Herb	Equation Model	R2	AIC
Shannon	Shannon = −0.181 × SOC + 0.635 × DBH + 0.131 × Density + 1.224 × pH	0.681	−90.089
Simpson	Simpson = 1.436 × Alt + 0.071 × Fire + 0.441 × pH − 0.071 × SOC	0.547	−136.768
Pielou	Pielou = 0.278 × DBH + 0.186 × A-N + 0.056 × Density − 0.192 × SOC	0.615	−124.213

Alt represents altitude, Fire represents prescribed burning, DBH represents tree diameter at breast height, Density represents tree density, SOC represents soil organic carbon, and A-N represents soil hydrolyzable nitrogen.

).

Different β diversity indices showed varying responses in herbs and shrubs to prescribed burning. The Jaccard index (β_j) indicated that there was a moderate similarity in the number of plant species between the areas subjected to prescribed burning and those that were not burned, with slightly higher similarity in the herbaceous layer compared to the shrub layer. On the other hand, the Cody index (β_c) suggested a higher species turnover in the herbaceous layer compared to the shrub layer. However, the Morisita–Horn index (C_MH) indicated a higher similarity in the herbaceous layer in both the burned and unburned areas.

3.4. Analysis of Major Factors Affecting the Species Biodiversity of the Understory

There were notable differences in influential factors between the shrub and herb layers (Figure 15). The MANOVA showed that the diversity index of the shrub layer calculated by Shannon’s index was mainly affected by soil pH, which accounted for 26.08% of its importance. Similarly, Margalef’s index was primarily influenced by pH, with an importance value of 29.81%. When considering Pielou’s index, the dominant influencing factor was found to be altitude, with an importance of 19.53%. On the other hand, fire regimes emerged as the most influential factor for Simpson’s index, with an importance value of 19.92%. Moving on to the herb layer, both Pielou’s index and Shannon’s index were predominantly affected by tree diameter at breast height, with importance values of 30.57% and 26.25%, respectively. Lastly, Simpson’s index in the herb layer was primarily influenced by altitude, with an importance value of 17.51%.

The structural equation model (SEM) revealed the relationships among α species diversity indexes of the understory vegetation and the possible influencing factors (Figure 16). For the shrub layer (Figure 16a), significant negative correlations existed between Simpson’s index and Shannon’s index and fire (prescribed burning), with path coefficients of −0.25 and −0.34 (p < 0.01), respectively (Figure 15a). Additionally, Margalef’s index, Simpson’s index, Shannon’s index, and Pielou’s index displayed significant correlations with soil pH, with path coefficients of −0.52, −0.50, −0.47, and −0.34 (p < 0.05), respectively. Similarly, Simpson’s index, Shannon’s index, and Pielou’s index exhibited positive correlations with altitude, with path coefficients of 0.44, 0.37, and 0.54 (p < 0.01), respectively. For the herb layer (Figure 16b), Simpson’s index had a significant correlation with fire (prescribed burning) and altitude, with path coefficients of 0.46 and 0.72 (p < 0.001), respectively. Additionally, Shannon’s index was found to be significantly influenced by tree diameter at breast height and soil pH, with path coefficients of 0.78 and 0.31 (p < 0.001), respectively, and Pielou’s index was also significantly correlated with tree diameter at breast height and soil A-N, with path coefficients of 0.72 and 0.44 (p < 0.01), respectively (Figure 15b).

4. Discussion

Prescribed burning had minimal impacts on overstory trees of the Pinus yunnanensis forest [21,29,40]. Our findings further demonstrate that prescribed burning exerts no significant alteration on the floristic composition of the arboreal stratum in Pinus yunnanensis forests, with its primary measurable impact limited to an increase in the height below branches. the increase in the height of the first branch of Pinus yunnanensis can be attributed to two additional factors. The warm and dry climate characteristic of the southwestern region, combined with low-intensity fires, has altered the micro-environment within forest gaps, leading to elevated temperatures and reduced humidity. These conditions enhance transpiration rates and induce the natural shedding of lower branches. Furthermore, Pinus yunnanensis exhibits intense intraspecific competition and a pronounced self-pruning effect, which results in excessive canopy closure during the seedling stage. Insufficient light and nutrients cause the lower branches and leaves to fall off. The species’ strong pioneer traits enable the upper canopy to dominate, further suppressing the growth of lower branches. Additionally, the increase in the height of the first living branch of Pinus yunnanensis, coupled with the reduction in shrub layer height following prescribed burning, has collectively diminished the vertical continuity of the Pinus yunnanensis forest in burned plots, thereby significantly reducing the likelihood of crown fires.

Environmental changes and biological characteristics jointly drive the increase in branches below the height. However, the effects of prescribed burning on shrubs were relatively larger [41,42]. Prescribed burning altered species composition in the shrub layer of the understory, which was similar to the effects of frequency fire on shrubby dry sclerophyll forests of southeastern Australia [41], Mediterranean-climate shrublands [8], understory vegetation of Arizona ponderosa pine forests [43], a Pinus halepensis forest in northeastern Spain [42], and Appalachian oak-dominated forests [44].

The varied impacts of prescribed burning on the understory of forests might be because of the differences in forest types and the post-fire time [25,44,45,46]. We found that three years after prescribed burning, species richness of the shrub layer in the prescribed burning plots was still clearly lower than that in unburned plots, which was similar to the effects of prescribed burning on the understory of Pinus halepensis forests in northeastern Spain [11]. On the contrary, understory species richness of Appalachian oak-dominated forests was not affected by burning [44]. Prescribed burning shifted the dominant species from Lithocarpus mairei and Lyonia ovalifolia in the unburned area to Duhaldea cappa in our study, consistent with the influence of periodic burning on the understory composition of an open New England woodland [46]. Saplings of Pinus yunnanensis, Schima wallichii, Quercus acutissima, and Lithocarpus mairei missing from the understory of Pinus yunnanensis forest after prescribed burning may affect forest regeneration processes and the future of the forest ecosystem [14,44].

Prescribed burning decreased the height of shrub-layer species, with frequent burning resulting in a decline in species richness for species taller than 1 m and a relative increase of smaller species less than this height in Pinus yunnanensis forest. Even low-intensity, frequent burning can destroy seedlings, and shorter intervals between burns leave insufficient time for tall shrub species to recover [42,47]. In contrast, the number of individual shrubs and species of shrubs is higher in areas of high-intensity fire. Low-intensity fires reduce fuel loads, thereby enhancing light availability and mineral nutrient effectiveness for understory vegetation [48]. Moreover, prescribed burning increases the heterogeneity of ground-level light intensity and soil conditions in forests, which promotes herbaceous plant diversity [49]. Prescribed burning also positively influences litter decomposition, enabling seeds of many plant species to reach the soil layer and facilitating their establishment and growth [12]. Additionally, low-intensity prescribed fires break seed dormancy mechanisms, stimulating germination processes in various plant species [3,50].

We found that compared with the unburned plots, the abundance of Fagaceae species in the shrub layer decreased after the planned burn, while the abundance of Fabaceae species significantly increased (Table A4). This result is similar to those of Knapp et al. [17] and Amoako et al. [51]. Most Fabaceae plants have strong drought and fire resistance. After fire disturbance or other environmental disturbances, although the total nitrogen content in the soil may decrease, the available nitrogen content increases, providing favorable conditions for nitrogen fixation through root nodules in Fabaceae plants, thereby accelerating their growth and increasing the probability of re-sprouting [52,53,54]. Though seed germination of Pinus yunnanensis is stimulated by high temperature and most species in semi-humid evergreen broad-leaved forests can resprout after fire [3], frequent burning may destroy regenerating saplings in shrub and herb layers, thereby altering the species community composition of shrubs. Prescribed burning is not appropriate for young forests. Low-intensity fires tend to significantly consume low shrubs, the herbaceous layer, and the litter layer [47].

Prescribed burning has enhanced the species richness of herbaceous plants in Pinus yunnanensis forests, with the highest increase observed in pioneer plants belonging to the Compositae and Gramineae families. These findings align with the effects of frequent fires on herbaceous vegetation in upland oak forests in the eastern United States [45]. The dominant species in the herb layer remained unchanged in the prescribed burning plots, while the changes in herbaceous understory abundance and height were also minimal, resembling patterns observed in Ponderosa Pine forests [40,55,56]. Herb abundance and diversity in Mediterranean pine forests increased significantly within 8 to 9 years of the planned burns [14]. Vegetative fuels rearrange over short intervals after fire, and herbaceous and ground apomictic cover in southeastern Australia is higher in low-intensity fire areas [50]. Long-term, low-intensity forest fires can be effective in reducing ground fuel loads and increasing canopy openness by reducing plant continuity vertically and horizontally, removing some shrub and tree seedlings, which contributes to herbaceous plant abundance and cover and reduces the incidence of crown fires [49,57]. Prescribed burning substantially elevated the importance value of ferns while diminishing that of Ageratina adenophora (see Table A5). Wang et al. [58] noted that, despite low-intensity fires suppressing the growth of invasive species, they are insufficient to completely eradicate Ageratina adenophora.

Fires could directly impact the structure of forest plant communities and indirectly influence them by altering soil nutrient content [59,60,61]. Fire can significantly alter the physical, chemical, and biological properties of soil, thereby indirectly influencing the structure and function of forest plant communities. Fire can change the physical, chemical, and biological properties of soils [62], and frequent wildfires limit the rooting activity of vegetation, resulting in a significant loss of soil organic carbon SOC, a decrease in elemental N concentrations, and a significant decrease in phosphorus, potassium, sulfur, and others. An evaluation of cation exchange capacity can be found in Refs. [63,64,65]. However, low-intensity fire has a positive effect on promoting plant community regeneration, and Scharenbroch et al. [59] found that low-intensity forest fires resulted in significant increases in soil pH, conductivity, Ca²⁺, Mg²⁺, K⁺, Na⁺, NO³⁻, total nitrogen, and organic carbon in the understory of oak (Quercus) forests in the United States. Moreover, the combustion process promotes the mineralization of organic phosphorus and releases a large amount of soluble inorganic phosphorus, thereby enhancing the absorption and utilization efficiency of nutrients by vegetation and soil microorganisms [66,67]. Our study also indicated that fire is the most influential factor affecting species diversity in the shrub layer: the α diversity index of the shrub layer exhibited a significant correlation with soil pH, A-K, and A-P content, while the α species diversity of the herbaceous layer was closely related to soil pH, SOC, and A-N.

Fires are a key factor influencing the succession of forest ecosystems, and the high temperatures at which forest fires occur can cause massive mortality of soil microorganisms, which in turn alters the composition and community characteristics of soil microorganisms [68,69]. Low-intensity burning induces the decomposition of nutrients in plants and litter into soil organic matter, which increases the soil organic carbon content and thus promotes the propagation and activity of microbial communities [49,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73]. However, we only considered abiotic factors when exploring the driving factors affecting the species diversity of understory vegetation, and we should comprehensively consider both biotic and abiotic factors in the follow-up study to better understand the differences and patterns of changes in species diversity and carbon stocks in the understory vegetation after the fire, and to further understand the effects of forest fire disturbance on the composition, structure, and succession of Pinus yunnanensis communities to provide a scientific basis for the protection, restoration, and management of Pinus yunnanensis forests after fire.

5. Conclusions

We investigated the impact of prescribed burning on plant community structure characteristics of the understory in the Pinus yunnanensis forests of central Yunnan. We compared the plant community structure in a prescribed burning area with an unburned area. Results showed:

(1)

Prescribed burning had little impact on the arboreal layer, only increasing the height under the living branches.

(2)

Prescribed burning induced a marked reduction in shrub layer species richness. Burning significantly altered the structure of the understory vegetation community in the shrub layers of the Pinus yunnanensis forests, declining from 26 taxa in unburned plots to 20 in prescribed burning plots, and it triggered a species shift in dominance hierarchies, with Lithocarpus mairei being supplanted by the fire-adapted species Duhaldea cappa, and the complete extirpation of arboreal saplings. Concomitantly, all biodiversity indices exhibited pronounced declines: the Shannon–Wiener diversity index (H′), Simpson’s dominance index (D), and Margalef’s richness index (F) were significantly reduced post-burn. Furthermore, vertical stratification was simplified, as evidenced by a substantial decrease in mean shrub-layer height.

(3)

Prescribed burning exhibited a modest augmentation in herbaceous species richness (30 to 37 taxa) yet manifested negligible impacts on herb-layer structural parameters. Comparative analyses revealed non-significant alterations in species abundance distributions, vertical stratification (herb height), and α diversity indices.

(4)

Our findings demonstrated that prescribed burning differentially modified the community structure of the understory vegetation in Pinus yunnanensis forest, with pronounced alterations observed in the shrub layer. The compositional shifts in both shrub and herbaceous strata exhibited dependent variability mediated by interacting biotic and abiotic covariates. Prescribed burning is the predominant determinant of the shrub-layer community. Conversely, herbaceous species diversity displayed stronger covariation with canopy architectural parameters and soil nutrient gradients.