The Early Effect of Plant Density on Soil Physicochemical Attributes and Bacterial and Understory Plant Diversity in Phoebe zhennan Plantations

Abstract

The effect of stand density on the soil bacterial community and diversity remains unclear. Spectrophotometry and full-length 16S rRNA sequences were used to determine the effects of planting density on soil physicochemical attributes and the diversity of soil bacterial and understory vegetation in a young Phoebe zhennan plantation at five densities. The findings showed that stand density had significant effects on the total nitrogen, ammonium nitrogen (${\mathrm{NH}}_4^{+}\text{-}\mathrm{N}$), nitrate-nitrogen (${\mathrm{NO}}_3^{-}\text{-}\mathrm{N}$), organic carbon, and the dominance and evenness of shrubs. Candidatus Udaeobacter and Candidatus Soilbacter were the two most common genera across the five stand densities. The density D5 (850 stems/hm²) demarcated from the others with a lower diversity of soil bacteria. Overall, the relatively low- and middle-density plantations were more conducive to complex and stable understory vegetation, bacterial communities, and soil nutrient cycles. The functional categories of the bacterial communities revealed a high proportion associated with chemoheterotrophy, aerobic chemoheterotrophy, and nitrogen fixation. Bacterial diversity and function were significantly influenced by soil pH, ${\mathrm{NH}}_4^{+}\text{-}\mathrm{N}$, ${\mathrm{NO}}_3^{-}\text{-}\mathrm{N}$, total phosphorus, and available phosphorus. However, there were no significant correlations between soil physicochemical attributes, understory vegetation, and bacterial diversity. Therefore, we speculated that the key drivers of the soil bacterial community were the soil physicochemical attributes and that stand density affected the soil bacterial community diversity by changing the soil physicochemical attributes. Overall, P. zhennan plantations with densities below 600 stems/hm² were conducive to complex and stable soil bacterial communities and nutrient cycles.

1. Introduction

In plantation ecosystems, stand density is an essential forest management strategy that influences soil microbial diversity, understory vegetation, and soil characteristics [1,2,3]. The diversity and composition of soil microbial communities vary considerably across temporal and geographical scales in terrestrial ecosystems, and they drive most biogeochemical processes and regulate soil nitrogen cycling [4,5,6]. Previous studies have supposed that soil physicochemical attributes, such as pH, soil organic, organic carbon, and soil nitrogen, are closely related to the communities of soil microbes [4,7,8,9,10,11,12,13]. Research on the driving factors and functions of soil microbes has been the focus of forest management strategies [3]. Some studies have shown that the nutrient cycle, understory vegetation diversity, and microbial community are also influenced by stand density [1,14,15,16,17]. In Chinese fir plantations, high-density plantations decreased the composition and diversity of microbes and the contents of soil pH, N, and P [2,3]. Stand density also affects the metabolic function of soil microbes in plantation systems [15]. Similar results were also reported in the forests of Catalonia, which showed that an increase in density decreased soil bacterial diversity by altering the understory vegetation and litter input [18]. In summary, stand density can influence the community and diversity of soil microbes. Nevertheless, the response of soil physicochemical attributes, understory vegetation, and soil microbes of plantations to density regulation is different with different tree species and tree ages, and the difference in optimal planting density is significant. For instance, Chinese fir plantations (37 years old) with a stand density below 3333 stems/ha are essential for maintaining soil fertility and improving forest ecosystem productivity [3]. Reducing the density (1212 stems/ha) in Pinus massoniana plantations (62 years old) can promote the diversity of understory vegetation and soil fertility [1]. However, 4300 stems/ha is the suitable stand density for 8-year-old Dendrocalamus minor var. amoenus plantations, which is more conducive to soil quality [14]. Therefore, Chen et al. [19] noted a fundamental change in the biomass of understory vegetation after 17 years of density-reduction treatment in Picea spp. forests. As such, a deeper understanding is required to clarify how stand density affects soil physicochemical attributes, understory vegetation, and the composition and diversity of soil microbes. The results could provide fundamental information for plantation management.

Phoebe zhennan S. Lee et F. N. Wei is a precious tree species in China, and its wood is very valuable and famous with a golden color, which is called the Golden-thread nanmu [20,21,22]. The price of P. zhennan wood in China is higher than 5000 RMB/m³, and the price can reach 10,000 RMB/m³ if the diameter at breast height (DBH) of the tree exceeds 30 cm [23]. Consequently, the area of P. zhennan plantations has increased rapidly due to the high value of the wood. However, unscientific density management has induced low production and poor value in the process of P. zhennan plantation management. Previous studies on plantation ecosystems regarding stand density have mainly focused on fast-growing and high-yielding coniferous plantations [2,3,15,24,25]. Differences in litter between coniferous and broadleaf forests result in different responses to changes in stand density [26]. Given the importance of stand density in soil microbial composition, a better understanding of the soil nutrient cycle would be extremely valuable in P. zhennan plantations. However, no comprehensive analysis of the effects of stand density on soil physicochemical attributes, understory vegetation, and soil microbes or any detailed observations of the correlations between biological and physicochemical properties have been performed in P. zhennan plantations. Therefore, we aimed to provide a scientific reference for density management in P. zhennan plantations to enhance soil-fertility maintenance techniques. Consequently, our work had two objectives: (1) to determine how soil bacterial compositions, soil physicochemical properties, and understory vegetation diversity vary with stand density and (2) to identify the main factors influenced by plant density that drive the soil bacterial community composition and diversity.

2. Materials and Methods

2.1. Study Site

The experimental site was in Pengshan County, Sichuan Province, China (103.969021° E, 30.304569° N), with a mean elevation of 453 m and a mean slope of 10°. The mean annual temperature, mean annual precipitation, and mean annual relative humidity of the experimental site were 18.3 °C, 1110.3 mm, and 76%, respectively.

2.2. Experimental Plot Design and Soil Sampling

P. zhennan plantations were planted in 2015 using 2-year-old seedlings and a randomized block design. Five planting densities were designed, 400 (D1), 500 (D2), 600 (D3), 700 (D4), and 850 stems/hm² (D5), with similar site conditions of topography, altitude, soil type, and slope position. A plantation of Eucalyptus grandis was built before the construction of the P. zhennan plantation in this study site, and the P. zhennan plantation was constructed after felling. Routine tending was carried out three years before afforestation, and there was no artificial interference in the later stage. Three 20 m × 20 m plots per density (three repetitions per density) were selected for understory vegetation investigation and soil sampling (

Table 1. Characteristics of Phoebe zhennan plantations of different stand densities.

Number	Stand Density (stems/hm2)	H (m)	DBH (cm)	P (m)	V (m3)	Altitude (m)	Aspect and Slope
D1	400	7.4 ± 0.2	12.4 ± 0.6	3.4 ± 0.4	0.050 ± 0.005	490	S 8°
D2	500	7.4 ± 0.2	12.4 ± 0.0	3.4 ± 0.1	0.051 ± 0.002	492	S 9°
D3	600	7.8 ± 0.3	12.8 ± 0.9	3.5 ± 0.3	0.056 ± 0.009	493	S 8°
D4	700	7.4 ± 0.1	12.5 ± 0.6	3.3 ± 0.2	0.051 ± 0.004	505	S 8°
D5	850	7.5 ± 0.1	12.0 ± 0.7	3.1 ± 0.2	0.048 ± 0.006	510	S 9°

Note: H—height, DBH—diameter at breast height, P—crown diameter, V—volume, S—the southern slope.

). The height (H, m), diameter at 1.3 m height (DBH, cm), and tree volume (V, m³) were measured in November 2021. The tree volume was calculated using the equation V = 0.000074954DBH^1.884652825H^0.881513308 [27]. In April 2021, five soil samples (0–20 cm) per plot were collected using the “five points sampling” method with a 5 cm diameter earth screw, and then the five soil samples were mixed together to make one composite sample [2,3,28]. The soil samples were divided into two parts: one part was used for 16S rRNA sequencing after passing through a 2 mm sieve to remove plant tissues and rocks, and the other part was used to analyze soil chemical properties after air-drying, grinding, and passing through a 0.15 mm sieve.

2.3. Understory Plant Community Inventory

Fifteen subplots (1 m × 1 m) per density were designed to investigate the herbs. For shrubs, the whole plot was investigated. The shrubs and herbs were identified, and their height, quantities, and coverage were measured. All individual species and species coverage in the shrub and herb plots were identified by three observers working together. The importance value (IV) and Shannon–Wiener (H′), Pielou (Jsw), Simpson (D), and Margalef (DMG) indices, which reflect species richness and diversity, were estimated through the following formulas [29,30]:

IV = (RH + RC + RF)/3

H′ = −∑(Pi)ln(Pi)

D = 1 − ∑(Pi)²

Jsw = −∑PilnPi/lnS

DMG = (S − 1)/lnN

where RH, RC, RF, N, S, and ln indicate the relative height, relative coverage, relative frequency, number of individuals, number of herbs/shrubs, and log base-e, respectively. Pi is the proportion of individuals or the abundance of the ith herb/shrub expressed as the proportion of the total in the community.

2.4. Soil Physicochemical Analyses

Soil pH, soil moisture (SM), soil bulk density (SBD), organic carbon (SOC), total nitrogen (TN), ammonium nitrogen (${\mathrm{NH}}_4^{+}\text{-}\mathrm{N}$), nitrate-nitrogen (${\mathrm{NO}}_3^{-}\text{-}\mathrm{N}$), total phosphorus (TP), and available phosphorus (AP) were tested through standard soil physical and chemical analyses [31]. A pH meter was used to measure the soil pH. The ring-knife soil samples were dried in a constant temperature oven to determine the SM and calculate the SBD. The Walkley–Black K₂Cr₂-H₂SO₄ wet oxidation method was used to measure SOC. TN was measured using the Kjeldahl method. The phenol sulfonic acid colorimetric process and the KCl extraction–indophenol blue colorimetric technique were utilized to quantify ${\mathrm{NO}}_3^{-}\text{-}\mathrm{N}$ and ${\mathrm{NH}}_4^{+}\text{-}\mathrm{N}$. TP was measured using HClO₄-H₂SO₄ colorimetry, and AP was determined using acid solution–molybdenum antimony resistance.

2.5. Amplification of Full-Length 16S rRNA Genes

Despite previous studies that used specific regions of the 16S rRNA gene to assign taxonomic assignments showing that taxonomic assignments are highly sensitive to specific regions, high-throughput analysis revealed unintended missing classifications with low accuracy, especially at the genus and species levels [32]. Therefore, more precise identification methods are required to determine the complexities of soil microbial communities. Recently, full-length 16S rRNA sequencing has greatly contributed to obtaining better resolution in soil microbial taxonomy [33]. Hence, the soil bacterial communities and diversity in P. zhennan plantations were analyzed by full-length 16S rRNA sequencing. The soil genomic DNA was extracted using a Soil Genomic DNA Kit (TIANGEN, Beijing, China), and then the full-length 16S rRNA gene was amplified. The PCR for each sample was run in triplicate and performed in a total volume of 30 μL containing 1.5 μL DNA, 10.5 μL H₂O, 15.0 μL KOD OneTM PCR Master Mix (Beijing Biolink Biotechnology Co., Ltd., Beijing, China), and 3.0 μL primers (27F: 5′-AGRGTTTGATYNTGGCTCAG-3′ and 1492R: 5′-TASGGHTACCTTGTTASGACTT-3′). The PCR protocol was as follows: 95 °C for 2 min, 25 cycles at 98 °C for 10 s, 55 °C for 30 s, 72 °C for 90 s, and 72 °C for 2 min. The contaminants and undesired PCR products were removed through gel-based size selection with 1.8% gel electrophoresis. DNA damage repair, end repair, and ligation of hairpin adapters were performed according to the manufacturer’s instructions. The binding concentration and the on-plate loading concentration of DNA template libraries were optimized using an Agilent Bioanalyzer 2100 (Agilent Technologies, Santa Clara, CA, USA). Then, Sequel Binding Kit 2.0 (Pacific Biosciences, Menlo Park, CA, USA) was used to bind the libraries to sequel polymerase 2.0. The data collection per sample was performed in a single Sequel SMRT Cell 1 M v2 with a 30 h movie time using Sequel Sequencing Kit 2.0 (Pacific Biosciences, Menlo Park, CA, USA).

2.6. Sequencing Analysis

Original subreads were corrected to generate circular consensus sequencing (CCS) reads using SMRT Link (version 8.0). Then, the CCS reads were screened using Mothur v.1.39.5 [34], and sequences outside the range <1200 bp to >1650 bp were removed using cutadapt v2.7 [35]. Chimeras were removed using UCHIME v4.2 [34,36]. Operational taxonomic unit (OTU) classification was performed with a 97% identity cutoff using USEARCH version 10.0, and the conservative threshold for OTU filtration was 0.005% [37]. Taxonomic alignment was performed using the SILVA 132 reference with a 90% cutoff value [38]. Alpha diversity analysis included the calculation of various diversity indices (Simpson and Shannon diversity index), richness abundance estimators (Chao1 and ACE), and Good’s coverage, calculated using QIIME [39]. Principal coordinate analysis (PCoA) was used to evaluate beta diversity with weighted and unweighted UniFrac distance metrics [33]. The functional profiles of the microbial communities were predicted by FAPROTAX [40].

2.7. Statistical Analysis

A general linear model (GLM) test was used to assess the effects of stand density on soil physicochemical attributes, understory vegetation diversity, and the diversity and relative abundance of soil bacteria. Then, an LSD test was used to calculate significant differences (α = 0.05). The abundant taxa used to characterize the microbiomes associated with different stand densities were identified by linear discriminant analysis effect size (LEfSe) [41]. Pearson’s correlation coefficient was calculated to assess the correlation of the relative abundance of the bacteria at the genus level (p = 0.05), and the soil physicochemical attributes and understory vegetation diversity and microbial diversity indices were also determined (p = 0.05). Spearman rank correlation analysis was performed on the difference in all species abundances among samples to identify data with a correlation coefficient larger than 0.1 and p value smaller than 0.05, which were used to construct the network. Redundancy analysis (RDA) was conducted using a Bray–Curtis distance matrix plot calculated by the vegan package in R to understand the degree to which the soil physicochemical attributes elucidated bacterial and understory vegetation diversity [42].

3. Results

3.1. Understory Community Structure

The understory vegetation of the P. zhennan plantation was composed of 63 genera of herbs and shrubs (Tables S1 and S2). There were no significant differences in herb community diversity among the five stand densities (

Table 2. Species diversity indices of the shrub and herb layers under five P. zhennan plantation stand densities.

Stand Densities	Herb Layer				Shrub Layer
	H′	Jsw	D	DMG	H′	Jsw	D	DMG
D1	1.86 ± 0.14	0.69 ± 0.05	0.25 ± 0.04	2.32 ± 0.26	1.00 ± 0.47	0.87 ± 0.08 a	0.44 ± 0.18 b	1.12 ± 0.69
D2	2.02 ± 0.26	0.74 ± 0.09	0.21 ± 0.09	2.58 ± 0.36	0.59 ± 0.61	0.47 ± 0.44 ab	0.67 ± 0.34 ab	0.52 ± 0.48
D3	1.93 ± 0.17	0.75 ± 0.08	0.22 ± 0.06	2.20 ± 0.58	0.16 ± 0.28	0.15 ± 0.25 b	0.92 ± 0.14 a	0.25 ± 0.43
D4	1.76 ± 0.25	0.67 ± 0.12	0.24 ± 0.06	2.24 ± 0.38	1.22 ± 0.26	0.89 ± 0.06 a	0.34 ± 0.10 b	1.30 ± 0.52
D5	1.38 ± 0.34	0.51 ± 0.11	0.40 ± 0.15	2.39 ± 0.18	1.29 ± 0.18	0.78 ± 0.05 a	0.37 ± 0.05 b	1.41 ± 0.39
F value	3.146	3.387	2.535	0.481	3.400	5.517	4.103	2.808
p value	0.064	0.054	0.106	0.749	0.053	0.013	0.032	0.085

Note: Data are reported as the mean ± standard error (n = 3). H′—Shannon index, Jsw—Pielou index, D—Simpson index, DMG—Margalef richness index. Within a column, values followed by the same lowercase letter indicate that they did not differ significantly (p ˂ 0.05).

). Within the herb layer, the Simpson value (D) in the D5 stand was obviously higher than that in the D2, D3, and D4 stands. The H′, Jsw, and DMG values first increased and then decreased with increasing stand density. There were significant differences in the Jsw and D values at the shrub level. The Jsw value in the D1, D4, and D5 stands was significantly higher than that in the D3 stand. In contrast, the D value in the D5 stand was significantly higher than that in the D1, D4, and D5 stands. The DMG and H′ values of shrubs tended to decrease first and then increase with increasing stand density.

3.2. Soil Physicochemical Attributes

There were no significant differences in soil pH, soil bulk density (SBD), soil moisture (SM), total phosphorus (TP), or available phosphorus (AP) among the five stand densities (

Table 3. Soil physicochemical characteristics in Phoebe zhennan plantations of different stand densities.

Stand Densities	pH	SBD (g/cm3)	SM (%)	TN (g/kg)	${\mathbf{NH}}_4^{+}\text{-}\mathbf{N}$	${\mathbf{NO}}_3^{-}\text{-}\mathbf{N}$	TP (mg/kg)	AP (mg/kg)	SOC (g/kg)
D1	5.49 ± 0.48	1.76 ± 0.15	0.17 ± 0.00	11.63 ± 1.65 a	6.38 ± 1.03 b	8.45 ± 0.61 b	0.37 ± 0.02	2.13 ± 0.59	23.73 ± 3.86 a
D2	5.31 ± 0.13	1.77 ± 0.07	0.17 ± 0.03	9.31 ± 0.27 b	6.70 ± 0.80 b	8.36 ± 0.48 b	0.30 ± 0.02	2.01 ± 0.25	21.26 ± 1.00 a
D3	5.26 ± 0.51	1.72 ± 0.18	0.15 ± 0.04	11.44 ± 1.28 a	3.63 ± 0.77 c	11.31 ± 0.84 a	0.30 ± 0.10	2.36 ± 0.89	23.53 ± 3.22 a
D4	5.85 ± 0.12	1.83 ± 0.08	0.17 ± 0.02	9.43 ± 0.71 b	6.50 ± 0.76 b	7.16 ± 0.21 b	0.25 ± 0.05	1.79 ± 0.17	16.25 ± 1.89 b
D5	5.08 ± 0.29	1.76 ± 0.14	0.18 ± 0.02	6.23 ± 0.36 c	9.67 ± 0.10 a	12.84 ± 1.67 a	0.30 ± 0.11	2.76 ± 0.21	20.27 ± 2.43 ab
F value	2.149	0.264	0.535	14.060	23.700	20.120	1.015	1.584	3.969
p value	0.149	0.894	0.713	0.000	0.000	0.000	0.445	0.253	0.035

Note: Significant differences between the five density soils were determined using a one-way ANOVA at p ˂ 0.05. The data are shown as the mean ± SD (n = 3). SM: soil moisture, SBD: soil bulk density, SOC: soil oxidizable organic carbon, TN: total nitrogen, ${\mathrm{NH}}_4^{+}\text{-}\mathrm{N}$: ammonium nitrogen, ${\mathrm{NO}}_3^{-}\text{-}\mathrm{N}$: nitrate nitrogen, TP: total phosphorus, AP: available phosphorus. Within a column, values followed by the same lowercase letter indicate that they did not differ significantly (p ˂ 0.05).

). The soil of the P. zhennan plantation is acidic soil, and obvious differences were found among the five stand densities (pH = 5.08–5.85). Significant differences were found in total nitrogen (TN), ${\mathrm{NH}}_4^{+}\text{-}\mathrm{N}$, ${\mathrm{NO}}_3^{-}\text{-}\mathrm{N}$, and organic carbon (SOC) among the five density plantations. The concentrations of SOC in the D1, D2, and D3 stands were significantly higher than those in the D4 stands. The concentrations of TN in the D1 and D3 stands were significantly higher than that in the other stands. However, the concentration of ${\mathrm{NH}}_4^{+}\text{-}\mathrm{N}$ in the D5 stand was significantly higher than those in the other stands, and ${\mathrm{NO}}_3^{-}\text{-}\mathrm{N}$ in the D5 and D3 stands was significantly higher than that in the other stands.

3.3. Overview of Full-Length 16S rRNA Amplification

A total of 174,192 effective CCS reads were separated into 15,849 OTUs across all samples, and the mean number of reads per sample was 9941 sequences (range 9108–10,439 sequences) (Table S3). Good’s coverage values were higher than 97%, revealing that most phylogenetic groups in all samples were captured (

Table 4. Richness and diversity indices of bacterial communities in different stand densities.

Stand Densities	ACE	Chao1	Simpson	Shannon	Coverage
D1	1234.57 ± 85.04	1244.27 ± 86.37	0.99 ± 0.00	8.44 ± 0.12	0.97 ± 0.00
D2	1357.93 ± 86.65	1345.38 ± 92.84	0.99 ± 0.00	8.52 ± 0.10	0.97 ± 0.00
D3	1234.59 ± 177.42	1238.40 ± 172.10	0.99 ± 0.00	8.52 ± 0.27	0.97 ± 0.01
D4	1365.42 ± 54.39	1355.03 ± 56.33	0.99 ± 0.00	8.58 ± 0.14	0.97 ± 0.00
D5	1141.21 ± 55.34	1135.91 ± 47.42	0.99 ± 0.00	8.27 ± 0.09	0.98 ± 0.00
F value	2.576	2.374	0.916	1.795	1.469
p value	0.103	0.122	0.492	0.206	0.283

The data are shown as the mean ± SD (n = 3).

). No significant differences were estimated in the richness estimator among the five stand densities. However, the ACE, Chao1, and Shannon values first increased gradually (from D1 to D4) and then decreased (from D4 to D5). The ACE, Chao1, and Shannon values in the D5 stand were obviously lower than those in the D2 and D4 stands (Table 4). Additionally, the PCoA results also indicated that the distribution of the soil bacterial communities was random (Figure 1).

3.4. Soil Bacterial Community Composition

In total, 33 phyla, 154 orders, 359 genera, and 434 species were identified in the P. zhennan plantation. At the phylum level, the dominant bacterial communities were Proteobacteria, Acidobacteria, Verrucomicrobia, and Bacteroidetes, which comprised more than 80% of the sequences under the five stand densities (Figure 2A). The relative abundances of Acidobacteria in D5 and Verrucomicrobia and Bacteroidetes in D2 were higher than those in the other stands (Figure 2B). Candidatus Udaeobacter, Candidatus Soilbacter, and ADurb. Bin063-1 were the predominant genera in all samples (Figure 2C). The relative abundances of Occallatibacter and Bryobacter in D3 and Candidatus Udaeobacter in D5, D3, and D2 were higher than those in the other stands (Figure 2D). The results of the correlation analysis showed that the relative abundance of Bryobacter was significantly positively correlated with Occallatibacter and Bradyrhizobium (p ˂ 0.05); nevertheless, there was no significant correlation between Occallatibacter and Bradyrhizobium (r = 0.26, p ˃ 0.05) (Table S4, Figure 3). The dominant species in the P. zhennan plantation were Bradyrhizobium jicamae, Sphingomonas parvus, and Paraburkholderia terrae (Figure S1). Overall, there were no significant differences in the relative abundance of the bacterial community between the five density stands at every classification level (Figure 3). At the genus level, the relative abundance of the top 20 bacterial communities also showed no significant differences among the five density stands (

Table 5. Relative abundance of the top 20 bacterial genera in different stand densities.

Phylum	Class	Order	Family	Genus	D1	D2	D3	D4	D5	F Value	p Value
Verrucomicrobia	Verrucomicrobiae	Chthoniobacterales	Chthoniobacteraceae	Candidatus_Udaeobacter	23.87	27.48	19.68	25.66	20.42	1.074	0.419
		Pedosphaerales	Pedosphaeraceae	ADurb.Bin063_1	7.00	6.49	6.15	5.83	5.07	0.868	0.516
		Chthoniobacterales	Xiphinematobacteraceae	Candidatus_Xiphinematobacter	1.04	2.03	1.58	2.07	6.28	3.930	0.036
Proteobacteria	Alphaproteobacteria	Rhizobiales	Xanthobacteraceae	Bradyrhizobium	4.99	4.59	6.11	5.62	5.59	0.329	0.852
				Pseudolabrys	2.81	2.28	2.01	2.51	1.68	1.101	0.408
		Sphingomonadales	Sphingomonadaceae	Sphingomonas	3.52	2.87	3.27	3.47	1.78	2.825	0.083
		Reyranellales	Reyranellaceae	Reyranella	1.22	1.08	0.85	0.85	1.27	1.094	0.411
	Gammaproteobacteria	Burkholderiales	Burkholderiaceae	Paraburkholderia	1.41	1.47	1.59	1.68	3.62	2.460	0.113
		Betaproteobacteriales	Nitrosomonadaceae	Ellin6067	1.76	2.25	1.18	1.88	1.25	2.896	0.079
		Xanthomonadales	Rhodanobacteraceae	Rhodanobacter	1.07	1.18	1.45	1.32	1.55	0.492	0.742
	Deltaproteobacteria	Myxococcales	Haliangiaceae	Haliangium	1.46	1.47	1.53	1.56	0.79	2.047	0.163
Acidobacteria	Blastocatellia_Subgroup_4	Pyrinomonadales	Pyrinomonadaceae	RB41	2.96	2.84	3.64	6.96	0.60	2.446	0.115
	Acidobacteriia	Acidobacteriales	Acidobacteriaceae_Subgroup_1	Occallatibacter	2.37	2.36	3.21	1.18	4.86	2.139	0.150
				Edaphobacter	1.81	1.54	1.76	1.07	3.05	1.958	0.177
			Koribacteraceae	Candidatus_Koribacter	2.16	1.54	2.13	1.22	2.67	1.108	0.405
		Solibacterales	Solibacteraceae_Subgroup_3	Bryobacter	2.42	2.35	3.15	2.47	2.89	0.535	0.713
				Candidatus_Solibacter	8.87	7.44	9.55	7.55	9.54	0.283	0.882
Nitrospirota	Nitrospiria	Nitrospirales	Nitrospiraceae	Nitrospira	1.31	1.43	2.12	2.20	0.99	1.446	0.289
Bacteroidetes	Bacteroidia	Flavobacteriales	Flavobacteriaceae	Flavobacterium	1.02	2.22	1.71	1.27	0.94	2.152	0.149
Gemmatimonadetes	Gemmatimonadetes	Gemmatimonadales	Gemmatimonadaceae	Gemmatimonas	1.31	1.01	1.52	0.93	1.24	0.502	0.735

).

3.5. Predictive Functional Profiles of Bacterial Communities

Among the putative functions predicted by FAPROTAX, the highest relative abundance of the putative function was related to chemoheterotrophy (28.40%–31.08%) (Figure 4A). The relative abundances of bacteria with aerobic chemoheterotrophic and nitrogen fixation functions were estimated to be 26.93% and 6.87%, respectively. The stand density could affect the putative functions of bacterial communities with putative functions, although the difference in bacteria with putative functions was not significant among the five densities (Figure S2). Overall, the D4 stand had the highest relative abundance of bacteria with nitrification and aerobic ammonia oxidation functions. The relative abundance of bacteria that functioned as animal parasites or symbionts in the D5 stand was higher than that in the other density plantations. The results of the correlation analysis showed that the relative abundance of bacteria with nitrification functions and predatory or exoparasitic bacteria was significantly negatively correlated with ${\mathrm{NO}}_3^{-}\text{-}\mathrm{N}$ and AP but significantly positively correlated with the pH value (Figure 4B).

3.6. Relationship between Soil Physicochemical Attributes and Bacterial and Understory Diversity

The redundancy analysis (RDA) results revealed that there were close relationships between the soil physicochemical attributes and the compositions of soil bacteria, herbs, and shrubs under density control (Figure 5). The RDA explained 35.34%, 33.83%, and 35.44% of the variation in the bacterial composition, herbs, and shrubs, respectively. From the perspective of the relationship between the understory vegetation and soil physicochemical attributes, the distribution frequency of most herbs and shrubs was positively related to the concentrations of ${\mathrm{NO}}_3^{-}\text{-}\mathrm{N}$ and ${\mathrm{NH}}_4^{+}\text{-}\mathrm{N}$ (Figure 5B,C). The bacteria were concentrated in the third and fourth quadrants at the genus level, and the ${\mathrm{NH}}_4^{+}\text{-}\mathrm{N}$, AP, SOC, and TP contents had positive correlations with the genera Bryobacter, Candidatus Solibacter, and Bradyrhizobium. The alpha diversity of bacteria was significantly influenced by pH, ${\mathrm{NH}}_4^{+}\text{-}\mathrm{N}$, ${\mathrm{NO}}_3^{-}\text{-}\mathrm{N}$, TP, and SOC (Figure 6A). However, there were no significant correlations between the soil physicochemical attributes and understory vegetation (Figure 6B,C). The correlation between the understory vegetation and bacterial diversity was also not significant (Figure 6D,E).

4. Discussion

In multiple studies, stand density was one of the key factors that affected soil physicochemical attributes in plantations [1,16]. In Phoebe zhennan plantations, the influence of stand density on total nitrogen (TN), ammonium nitrogen (${\mathrm{NH}}_4^{+}\text{-}\mathrm{N}$), nitrate-nitrogen (${\mathrm{NO}}_3^{-}\text{-}\mathrm{N}$), and organic carbon (SOC) contents was significant. Plants at higher densities need to use and uptake greater amounts of nutrients [43]. Therefore, the contents of TN and SOC in the relatively high-density plantations were lower than those in the relatively low-density plantations of P. zhennan. We also found that the soil pH decreased with increasing density, indicating that soil acidification increased with increasing density, which was caused by organic acids produced by roots in relatively high-density plantations [44]. In acidic soils, phosphate easily precipitates with iron, aluminum, and manganese in the soil, reducing the available phosphorus [45]. Consistent with the results of a previous study, our current study found that the available phosphorus (AP) content reached a maximum with the lowest pH at the highest density.

Stand density and soil physicochemical attributes explained most of the variation in the diversity and composition of the bacterial communities in plantations [4]. Every plant tends to be associated with a specific soil microbial community. In P. zhennan plantations, the dominance of soil bacteria at every taxonomic level is different from that in other tree species plantations [2,3,4]. This may be caused by species specificity and the planting environment. For instance, Verrucomicrobia and Bacteroidetes were the dominant bacteria at the phylum level, but these two bacteria were not dominant in Chinese fir plantations [3]. At the genus level, Subgroup-2-norank, JG37-AG-4-norank, and Acidobacteriaceae-Subgroup-1-uncultured were the dominant bacteria in Hevea brasiliensis plantations [46], but these were not determined in our study. Plantations with higher stand density changed plant detrital inputs and subsequently enhanced the activity of soil microbial communities [4]. There was no significant difference in the relative abundance of bacterial communities among the five stand densities, indicating that the soil bacterial communities are resilient and not sensitive to density control in young P. zhennan plantations. Multiple studies have suggested that density is one of the most important driving factors of soil microbial composition [2,3,4]. Consistent with those of previous studies, our results showed that the diversity of the bacterial community in the relatively low- and middle-density plantations (400–700 stems/ha) was higher than that in the high-density plantations, indicating that P. zhennan plantations with low densities are conducive to more diverse bacterial communities. This may be partly explained by changes in canopy, litter species, and understory vegetation [47,48].

A previous study considered that density control significantly affected the distribution and relative abundance of soil microbes [3,4]. In this study, the diversity of the bacterial community was significantly related to soil pH, ${\mathrm{NH}}_4^{+}\text{-}\mathrm{N}$, ${\mathrm{NO}}_3^{-}\text{-}\mathrm{N}$, TP, and SOC, suggesting that various soil physicochemical attributes can affect the diversity of bacterial communities in P. zhennan plantations. Previous reports have suggested that soil pH is a predictor of soil microbial communities [4,49]. We also observed a significant positive correlation between soil pH and the diversity of the bacterial community. The relative abundance of the bacterial community with putative functions was also significantly positively correlated with the soil pH, such as nitrification and aerobic ammonia oxidation. In terms of nutrients, higher soil N and P contents indicate a higher efficiency of soil microbial decomposition. In contrast, TP, ${\mathrm{NO}}_3^{-}\text{-}\mathrm{N}$, and ${\mathrm{NH}}_4^{+}\text{-}\mathrm{N}$ showed a negative correlation with the diversity of bacteria, indicating that N and P were limited during the bacterial decomposition process in this study. Soil bacteria and understory vegetation compete for available N and P [16]. However, there was no significant correlation between the diversity of bacteria and understory vegetation, suggesting that the effects of the diversity of understory vegetation on soil bacterial community diversity in the early stage of P. zhennan plantations were not obvious. Compared with understory vegetation, litter decomposition and nutrient return played a key role in topsoil physiochemical properties and bacterial diversity, at least in the early stage of P. zhennan plantations. The results of the correlation analysis also showed that there were no significant correlations between the soil physicochemical attributes and the diversity of the understory vegetation. Consequently, bold speculation suggests that soil physicochemical attributes are the main drivers of soil bacterial community composition and diversity in P. zhennan plantation ecosystems.

Microorganisms are important factors that influence nutrient availability in plant–soil feedback systems [50,51]. We found that density management significantly influenced the composition and diversity of soil microbial communities along with soil physicochemical attributes, which is consistent with the results from previous studies [2,3,4,15]. We also found that a handful of key bacterial species were affected by density-dependent P. zhennan plantation soil feedback, although soil bacterial communities had no significant response to changes in stand density. For instance, the abundances of Acidobacteria and Acidobacteriales in the high-density stands were higher than those in the low-density stands, and they were also the dominant bacteria at the phylum and order levels. The activity of Acidobacteria affected soil acidity and nutrient availability and had a negative response to increases in carbon and pH [52,53,54], suggesting the presence of higher soil acidity and lower carbon availability in relatively high-density stands than in relatively low-density stands in P. zhennan plantations. In this study, changes in stand density with changes in soil pH and SOC may have resulted in differences in the relative abundance of Acidobacteria in plantations with different stand densities. Wang et al. [2] suggested that low-density plantations may be more beneficial to soil nutrient cycles. In our study, the relative abundance of bacteria with nitrification and nitrogen fixation functions at low and moderate densities was higher than that at high densities, which indicates that plantations with an intermediate density had a preference for the soil N cycle. The relative abundance of bacteria with putative functions was influenced differently by soil chemical properties [55]. For example, the relative abundance of bacteria with a nitrogen fixation function was significantly positively correlated with TN, but there was no difference among the stand densities. The relative abundance of bacteria involved in nitrification was significantly positively correlated with soil pH, indicating that the management of soil pH is important for soil fertility and the N cycle. Consistent with the literature [15,16,55], these results implied that stand density can alter soil processes via functional bacteria and that soil physicochemical attributes are important for the functional engineering of soil bacteria, at least in early-stage P. zhennan plantations.

5. Conclusions

The main goal of the current study was to determine the effects of shifts in stand density on soil physicochemical attributes, understory vegetation, and the bacterial community across young P. zhennan plantations under different stand densities. Low- and middle-density plantations showed higher contents of soil TN, TP, and SOC and a higher diversity in herb and bacterial communities than high-density plantations, suggesting that the soil ecosystem functions of low- and middle-density plantations are more stable than those of high-density plantations. Therefore, to maintain good soil ecosystem function, we suggest that the planting density of P. zhennan should be controlled below 600 stems/ha. We also found that the effects of pH, TN, ${\mathrm{NH}}_4^{+}\text{-}\mathrm{N}$, ${\mathrm{NO}}_3^{-}\text{-}\mathrm{N}$, and AP on herb diversity and the effects of ${\mathrm{NH}}_4^{+}\text{-}\mathrm{N}$ on shrub diversity were significant. The results of this study provide essential insights that soil physicochemical attributes are indeed factors that drive the formation of understory vegetation and bacterial communities in P. zhennan plantations. This is also an important factor for how stand density influences the diversity of understory vegetation and the bacterial community.