1. Introduction
In forest ecosystems, soil microorganisms play a critical role in carbon, nutrient cycling, and energy flow [
1]. Microorganisms are also regarded as sensitive indicators of soil quality for their role in soil carbon processes, including soil organic matter decomposition and turnover [
2]. Over the past two decades, soil microbial community and soil carbon process in forest management had been investigated by some studies [
3,
4], which theorized that soil carbon was affected by forest management through soil microbial regulations. Therefore, the effect of forest management on the soil microbial community structure and function have received increased research attention [
5].
Forest thinning is one management strategy with strong potential. High stand densities of trees in pure forest may lead to poor growth and the reduction of forest productivity [
6]. Parts of trees are selectively removed and form gaps to increase the availability of light, water, and nutrients to improve forest microclimatic conditions [
7]. Soil microbial community structure and function are affected by thinning through the correlation between above- and below-ground processes [
8]. Thinning directly affects carbon input (litter and roots), and then alters soil carbon process [
9].
Forest thinning has shown positive or negative effects on soil microorganisms, while there remains no consensus on the influence on soil microbial community. Some previous studies theorized that thinning reduced soil microbial abundance, since large gaps promoted high soil and air temperatures and improved living conditions for bacteria- and fungi-feeding nematodes [
10]. The positive effect of thinning on soil microbial diversity was observed in a Chinese fir plantation by a previous study, which theorized that thinning increased the availability of light, water, and nutrients to ameliorate forest microclimatic conditions [
11]. However, the insignificant influence of thinning on soil microbial community was also supported by other literatures [
12]. It is theorized that the decrease in soil microbial diversity is probably related to reductions in soil organic carbon and microbial abundance and that diversity can be improved by management practices [
13].
Some previous works theorized that microbial communities are correlated with their function, and microbial community structure and function strongly influence soil carbon dynamics [
14,
15,
16]. Soil microorganisms depolymerized soil organic matter by producing extracellular enzymes, which are involved in soil nutrient cycling [
17]. The activities of soil enzymes are a direct expression of soil microbial community structure. The assay of soil enzymes, which can be divided into oxidases and hydrolases activities, gives an indication of functions that can be assumed by the microbial community [
18]. Poor-quality and complex compounds such as lignin is degraded by oxidases which is produced primarily by fungi [
19]. However, the hydrolase, which is produced primarily by bacteria, degrades cellulose and is related to soil carbon acquisition [
20].
Many studies have investigated the effect of thinning on soil organic carbon [
21,
22], however, the regulation of soil microbial community on soil carbon with thinning treatment is still not well understood. After thinning treatment, soil organic carbon is altered following the variation in microbial community [
23,
24]. Therefore, studying the effect of microbial community structure on functions, as well as the microbial effect on soil carbon properties, is necessary for better understanding the belowground processes affecting carbon dynamics [
25,
26].
Thinning may also affect soil microbial communities by altering soil physicochemical factors in forest ecosystems [
27]. The microclimates in different thinning intensity cause the variation in soil water content and pH [
28]. It is known that soil pH has a significant effect on soil microbial communities and soil enzyme activities relating to carbon decomposition [
29]. Meanwhile, soil microbes have the ability to maintain their intracellular pH [
30]. There are many studies of the effect of soil water content on soil microbes [
31,
32]. Soil water maintains soil microbial activities and indirectly influences microbial substrate and oxygen availability. Therefore, it is necessary to note the effect of soil pH and soil water content on soil microbial community structure.
Larix principis-rupprechti Mayr is a prominent plantation tree species in the warm, temperate Taiyue Mountains of Shanxi province. In this paper, four thinning intensities were established to examine soil microbial community structure, function, and soil carbon with thinning in L. principis-rupprechti plantations. We measured the activities of hydrolases to assess soil microbial function relating to cellulose and chitin degrading capacity, and oxidases to assess microbial function relating to lignin degrading capability in driving soil carbon transformation.
Furthermore, we investigated the correlation between soil microbial community and function, studied how soil pH, soil water content, and shifts in microbial community drive the variation in soil carbon properties. We hypothesized that: (1) moderate thinning has positive effects on soil microbial structure and function; and (2) there were connections between soil microbial community structure and the function, which relates to soil organic carbon turnover; and (3) the variation in soil physicochemical properties influenced soil microbial function, which is regulated by soil microbial community structure, and finally affected soil carbon properties.
2. Material and Methods
2.1. Study Sites
The study was conducted in L. principis-rupprechtii plantations in the Taiyue Mountains (36°35′–36°53′ N, 111°91′–112°04′ E) of Shanxi province in northern China. This region is characterized as a warm temperate, continental monsoon climate with cold, dry winters and hot, wet summers. The mean annual temperature for the area is 8.6 °C. The annual average precipitation is about 600 mm and the rainy season is from June to August. The study site is located at an altitude of 2300 m (with study plots ranging from 2298–2358 m), with a slope of 23° (plots ranging 22–25°) and a northern aspect. The soil type is Haplic luvisols, L. principis-rupprechtii and Betula platyphylla Suk. are the dominant tree species on site. The understory is dominated by Lonicera japonica Thunb., Corylus mandshurica Maxim., Rubus corchorifolius L.f., Rosa xanthina Lindl., and Lespedeza bicolor Turcz.
2.2. Experimental Design and Treatments
In 1982, foveolate site preparation was carried out along contour lines in mountains, and 3-year-old
L. principis-rupprechtii seedlings were planted with an initial density of 3000 trees hm
−2. In 2000, this plantation was thinned and maintained at 2100 trees hm
−2. In 2012, 12 25 × 25 m study plots were established in this area, and the plantation was again thinned, this time to four specified thinning densities: control (CK, 2100 trees hm
−2, thinning density: 0%), light (LT, 1850 trees hm
−2, thinning density: 15%), medium (MT, 1415 trees hm
−2, thinning density: 35%) and high (HT, 1100 trees hm
−2, thinning density: 50%). Each treatment was replicated three times (one of the control plots was destroyed by deforestation), and the plots were spaced at least 10 m apart in order to avoid edge effects. Detailed information for these four sites is shown in
Table S1.
Five blocks of 2 × 2 m were distributed randomly in each study plot. Soil samples were collected on five different dates in August 2015, and April, June, August and October in 2016, with a depth of 0–10 cm at each block using a metal corer with an inner diameter of 5 cm. Soil samples of the same thinning treatment were mixed to create a composite sample. The soil samples were then sieved at 2 mm to remove roots and gravel. One part of the sample was stored at 4 °C for analyses of soil microbial community and soil enzyme activity. Another part of the sample was air-dried and passed through a 0.25 mm sieve for soil physicochemical analyses.
Soil Property Analyses
Soil water content was calculated from the mass loss by oven-drying samples at 105°C to a constant weight, for at least 48 h [
33]. Soil pH was measured by Sartorius PB-10 with a soil solution ratio of 1:2.5. Soil organic carbon and soil total nitrogen was analyzed with an elemental analyzer (Thermo Fisher Scientific, FLASH 2000 CHNS/O, USA). Soil microbial biomass carbon was measured by the fumigation–extraction method [
33]. Each sample was fumigated for 24 h at 25 °C with alcohol-free CHCl
3, using a 0.5M K
2SO
4 extracting agent and measuring with a TOC analyzer (Analytikjena, Multi N/C 3100 TOC, Germany).
2.3. Soil Microbial Community Structure and Function
Microbial community structure was determined using phospholipid fatty acid (PLFA). PLFA method is sensitive in detecting shifts in microbial community structure, with an inexpensive way of assessing the biomass and composition. PLFA provides the advantage of being an indicator of living organisms since it is rapidly hydrolyzed upon cell death [
34]. The lipids in each freeze-dried soil sample were extracted in a single-phase mixture of chloroform, methanol, and phosphate buffer. The abundance of single PLFAs were calculated based on 19:0 internal standard content. After addition of an internal standard, the phospholipid fraction was subjected to a mild alkaline methanolysis, and the resulting fatty acid methyl esters were separated on a gas chromatograph [
35,
36]. The following soil microbial groups were classified using diagnostic fatty acids as the indicator: gram-positive bacteria, gram-negative bacteria, saprotrophic fungi (Sap), arbuscular mycorrhizal fungi (AMF), and actinomycetes (
Table S2). The abundance of soil microbial single PLFAs were used to analyze soil microbial diversity, which was calculated by the Shannon–Wiener index, species richness, and species evenness.
Shannon index of soil microbial community is
where
n is the number of species and
Pi is the measure of
ith species proportional to the total measure of all species.
Species richness of soil microbial community is
where
n is the number of species.
Species evenness of soil microbial community is
where
Pi is the measure of
ith species proportional to the total measure of all species, and S is the number of species.
The activities of phenol oxidase and peroxidase were determined by using DOPA (3,4-Dihydroxy-L-phenylalamine) as the substrate. Soil suspension (1 g fresh soil with 1.5 mL 50 mmol L
−1 sodium acetate buffer) and 2 mL 5 mmol L
−1 L-DOPA were mixed for phenol oxidase assay. The same suspension was used with an addition of 0.2mL 0.3% H
2O
2 for peroxidase analyses. The activities of β-glucosidase (BG),
N-acetyl-β-glucosaminidase (NAG) and cellobiohydrolase (CBH) were measured with p-nitrophenol assays [
33].
2.4. Statistics
Analysis of variance (ANOVA) was used to determine the effect of thinning treatments on soil microbial community structure and function. Meanwhile, Fisher’s least significant difference (LSD) multiple comparison test (
p < 0.05) was used to compare soil microbial community structure and soil enzyme activities between different thinning intensities. Soil microbial composition and enzyme activities were also tested by repeated measures analysis of variance (RMANOVA) for the effects of thinning treatments across five sampling dates. These analyses were performed using SPSS 19.0 [
3].
Redundancy analysis (RDA) was used to assess the correlation between soil microbial composition and soil enzyme activities, as well as the correlation between soil microbial community, soil pH, and soil water content. Based on Monte Carlo permutation with 499 iteration, the RDA was used with forward selection to filter the relative importance of explanatory variables of the microbial composition and soil enzyme activities. Meanwhile, the significant variables were used in the final analyses. Based on the RDA, partial redundancy analysis (pRDA) was used to partition variance of variables. Marginal effects indicate when the variable is used as the only factor. Conditional effects showed the additional variance, and each variable is indicated when it is included in the model. These analyses were completed in CANOCO 4.5 software (Wageningen University and Research Centre, Wageningen, The Netherlands) for Windows.
Structural equation modeling (SEM) was used to test the hypothetical connections between soil physicochemical variables and soil microbial communities, as well as the correlation between microbial community structure and the function, which is relating to soil organic carbon turnover in different thinning treatments. In the path model depicting the hypothesis on the regulatory pathway of soil microbial community structure and function, soil pH and soil water content were considered as the important indicators of changes in soil microbial composition; the abundance of gram-negative bacteria, saprotrophic fungi, and actinomycetes were the key factors to represent the structural attributes of microbial community. We estimated the model parameters by maximum likelihood estimation using Amos 22.0. The adequacy of model fitting was assessed by a χ
2 test (
p > 0.05, CMIN/df < 2), the comparative fit index (CFI > 0.9) and the root square mean error of approximation (RMSEA < 0.05) [
32]. Numbers on arrows are standardized direct path coefficients. R
2 value represents the proportion of total variance explained for the specific dependent variable.