Response of plant diversity and soil physicochemical properties to different gap sizes in a Pinus massoniana plantation

Study area

The study area is located in Dongfanghong Forest Farm, Huaying City, Sichuan Province, in the low mountain area of eastern Sichuan (106°45′53.27E– 106°45′55.87E, 30°17′48.45N–30°17′55.09N) (Fig. 1). Elevations range from 544.1 to 574.7 m. Climatic characteristics of the region include a mild climate, abundant rainfall, uneven rainfall and large temperature differences. The multiyear average temperature in Huaying is 17.2 °C. The area has abundant rainfall, with a maximum annual average of 1441.7 mm, a minimum annual average of 854.9 mm, and a multiyear average of 1087.84 mm. The soil type is typically a low-fertility yellow soil (Nitisol) in the study area (Li et al., 2020d). The shrub layer in the forest gaps of different areas consists mainly of Cinnamomum camphora (L.) Presl, Quercus serrata Murray and Mallotus barbatus (Wall.) Muell. Arg. The herb layer is mainly composed of ferns.

Experimental design and data collection

The field plot design was the P. massoniana plantation in the 1870S, which had similar site conditions, the same forest age and the same management history in Dongfanghong Forest Farm. In early August 2017, we used laser distance meter (LDM-80H) to determine forest gap sizes, then marked the P. massoniana that needed to be cut with red paint. The experiment was a complete randomized design, and nine artificial gaps were created with gap sizes ranging from 50 m² to 667 m² (Fig. 1). Furthermore, these gaps were divided into three levels, i.e., small gaps (SG): 50–100 m² (n = 3), medium gaps (MG): 100–200 m² (n = 3) and large gaps (LG): 400–667 m² (n = 3), surrounded by closed canopy transition zones and a 5 m buffer. All gaps were approximately circular and located in similar topography, with aspects southwest-facing, slopes ranging from 7 to 17° . The mean height of gap border trees was 15 m, with a mean DBH of 22 cm. A designated sample plot without any management was also established as the control plot with three replicates. We set each of the control plots to be 20 m ×20 m. The trunks and branches of P. massoniana cut down in the forest gap were all removed from the gap, and plant regeneration in the gap mainly depended on seeds and the soil seed bank.

Vegetation surveys were conducted in October 2019. Five 3 m ×3 m squares were mechanically arranged as shrub squares in the four corners and the center of sample squares within different gap areas of each P. massoniana plantation, and then ten 1 m ×1 m squares were set up in each of the shrub squares as herb squares. We counted and recorded the species name, the number of plants and the branch diameter and height of each plant in the shrub quadrats. The average height and branch diameter of each shrub in the plot were measured using a tape line. In the herb layer, we counted and recorded the species name, overall coverage, average coverage and average height of each species occurring in the herb quadrats. The indicator species are that the important values of shrub layer plants are more than 5% and that of herb layer plants are more than 10% (Wang et al., 2019b).

According to the data obtained from the sample plot survey, we calculated the importance value of the plants appearing in the sample plot and then calculated the diversity index. The important value is an important index in calculating and evaluating species diversity, and the comprehensive value is used to express the relative importance of plant species in the community. The IV value was used to calculate species diversity in each plot (Zhang et al., 2011). We used four α-diversity indices: Richness (R), the Shannon-Wiener index (H), Simpson index (D) and Pielou index (J) (Zhou et al., 2016), which were calculated as follows:

Important value (%): IV = (Relative density + Relative frequency + Relative coverage) /3 (1) (Ming et al., 2020) (2)${\displaystyle R=S}$ (3)${\displaystyle \text{Simpson index}:D=1-\sum _{\mathrm{i=1}}^{\mathrm{s}}{\mathrm{P}}_{\mathrm{i}}^2}$ (4)${\displaystyle \text{Shannon-Wiener index}:H=-\sum _{\mathrm{i}=1}^{\mathrm{s}}{\mathrm{P}}_{\mathrm{i}}ln{P}_i}$ (5)${\displaystyle \text{Pielou index}:J=H/ln\mathrm{S}}$

Where S is the total number of species in the plot; i is species i; and P_i is the relative importance value of species i.

A soil drill (five cm in diameter) was used in each crop tree management stand to randomly arrange five sampling points in an “S” shape to collect soil samples. Soil samples at depths of 0–10 cm and 10–20 cm were collected separately at each sampling point and mixed evenly. Approximately 1.5 kg of each sample was selected according to the quartet method, packed into plastic bags and taken back to the laboratory within 24 h. Fine roots, fine gravel and litter were manually separated from the soil. Then, the soil samples were air-dried and passed through two mm and 0.149 mm sieves. A subsample was dried at 105 °C for 48 h to constant weight to measure the soil gravimetric moisture content and soil bulk density (Luo et al., 2018). The organic matter content of the soil was determined by hydration with the potassium dichromate oxidation-colorimetric method (Yang et al., 2020). A 1:2.5 water and soil mixture was whisked together for 10 min with a glass rod, left to stand for 1 h, and then measured for pH with an electronic meter (Li et al., 2020d). Soil hydrolysable nitrogen was measured by the alkali-hydrolysed diffusion absorption method (Chen et al., 2010). Soil hydrolysable nitrogen was measured by the alkali-hydrolyzed diffusion absorption method (Rich, 1958). Soil available phosphorus was measured by extracting subsamples with 0.03 M NH₄F –0.025 M HCl. (Mariotte et al., 2020).

Statistical analyses

Single factor analysis of variance (one-way ANOVA) and multiple comparisons (LSD) were used to test the soil index and plant diversity index. Two-way ANOVA was used to test whether there was a significant interaction among the different forest gaps, different soil layers and soil physicochemical properties. Significant differences were detected at p < 0.05. Pearson correlation analysis was carried out between the plant diversity index and the soil physicochemical properties. Redundancy analysis was carried out by using CANOCO5.0 software to compare the correlation between the plant composition of the shrub and herb layers and environmental factors in different forest gaps. All data were processed by Excel software, Origin Pro 8.0 was used to create the figures, and statistical analysis was performed using SPSS 20.0.

Soil physicochemical properties

According to Table 1, soil pH was not significantly different in the different forest gaps, and it was acidic and fluctuated in a small range, but there was a significant difference between the two soil layers (Table 2). Compared with the control, the content of soil organic matter increased significantly under the middle gap and small gap, and the content of soil organic matter reached the maximum in the 0–10 cm soil layer of the medium gap, which increased by 80.64%. There was no significant difference in soil-available phosphorus content between the different soil layers of each treatment. The content of soil hydrolyzed nitrogen was significantly affected by different forest gap treatments and interactive treatments between different forest gaps and different soil layers (Table 2). Compared with the control, the content of soil hydrolyzed nitrogen significantly decreased by 50% in the small gap, increased by 74.96% in the medium gap, and increased by 2.94% in the large gap. The content of soil-available potassium between different soil layers was significantly lower than that of the control. Soil moisture showed significant changes among different forest gaps, between different soil layers and under the interaction between them (Table 2). The soil bulk density decreased significantly compared with the control, and the surface soil reached the minimum in the medium gap.

Plant diversity and community composition

The composition of shrub layer plant forest gaps was different, and the dominant species also changed gradually (Table 3). A total of 21 shrubs appeared in the control treatment, mainly Cinnamomum camphora (L.) Presl, Aralia chinensis (L.), Urena lobata (L.), Ardisia pusilla A. DC. and Cryptomeria fortunei Hooibrenk ex Otto et Dietr. Under SG treatment, the number of plant species in the shrub layer was 28, and the dominant species gradually evolved into Cinnamomum camphora (L.) Presl, Aralia chinensis (L.), Urena lobata (L.) and Paulownia fortunei (Seem.) Hemsl. There were 35 shrub species in the MG treatment, which were the most abundant in all treatments. Cinnamomum camphora (L.) Presl, Aralia chinensis (L.), Urena lobata (L.), Ardisia pusilla A. DC. and Ardisia japonica (Thunb) Blume were the dominant species in this treatment. The dominant species in the LG treatment were Cinnamomum camphora (L.) Presl, Aralia chinensis (L.), Ardisia pusilla A. DC. and Rubus pirifolius Smith, with a total of 29 shrub species.

There were a few species of herbaceous plants in the control treatment, most of which were pteridophytes, and there were 6 species in the herb layer (Table 4). There were 17 species of herbs in treatment SG. Pteridium aquilinum (L.) Kuhn and Setaria plicata (Lam.) T. Cooke were the dominant herbs of this treatment. The MG treatment contained the most abundant herbaceous plant species, with a total of 20 species. Lophatherum gracile Brongn., Microlepia hancei Prantl and Setaria plicata (Lam.) T. Cooke were the dominant species in this treatment. There were 11 species of herbaceous plants in the LG treatment, and the dominant species were Dryopteris fuscipes C. Chr., Miscanthus sinensis Anderss., Arachniodes simplicior (Makino) Ohwi and Setaria plicata (Lam.) T. Cooke.

Figure 2 shows that, compared with the control, the richness index of the shrub layer in different gaps increased significantly (P < 0.05). The richness index of the shrub layer was the highest in the MG treatment. As shown in Table 5, the Shannon-Wiener index, Simpson index and Pielou index of the shrub layer in all treatments were significantly higher than those of the control (P < 0.05). The order of them was MG > SG > LG > CK. The Simpson index and Pielou index of the herb layer in different forest gaps were significantly lower than those of CK (P < 0.05). The Simpson index was CK > MG > SG > LG, and the Pielou index was CK > MG > LG > SG. The Shannon-Wiener index was significantly higher for CK (P < 0.05), but the MG treatment had the largest value.

Relationship between soil physicochemical properties and undergrowth plant diversity

As shown in Fig. 3, the soil-available potassium content was closely related to undergrowth plant diversity, and there was a significant negative correlation between the soil-available potassium content and the plant diversity of shrub layers (P < 0.01). The Simpson index of herb layers was influenced by many soil factors and was positively correlated with the available potassium, soil moisture content and bulk density of topsoil (P <0.05). There were significant positive correlations between the Pielou index and the available potassium content, bulk density of topsoil and soil moisture content of bottom soil (P <0.05). The richness index of the shrub layer showed a positive correlation not only with hydrolysis nitrogen content (P < 0.05) but also with soil organic matter content in topsoil (P < 0.01). The richness index of herb layers was significantly positively correlated with bottom soil pH (P < 0.05). Bottom soil bulk density of 10–20 cm was significantly negatively correlated with plant diversity (P < 0.05) and with herb layers (P <0.01).

The redundancy analysis of the relationship between species composition in the shrub and herb layers and soil environmental factors under different forest gaps showed that the zoning of each treatment was obvious. The three forest gap treatments were visually separated from control, indicating that there were great differences in species composition and habitats (Fig. 4). Species of the shrub and herb layers are represented by blue arrow lines, and soil environmental factors are indicated by red arrow lines.

The first principal component axis and the second principal component axis explained 46.86% and 24.71% of the species composition of the 0–10 cm soil layer, respectively, and the total explanation rate reached 71.57% (Fig. 4A). The first principal component axis and the second principal component axis explained 55.06% and 26.54% of the species composition of the 10–20 cm soil layer, respectively, and the total explanation rate reached 81.60% (Fig. 4B). All these indicated that soil physicochemical properties could explain the differentiation of the spatial and characteristics of the vegetation community under the forest. Soil available potassium, bulk density and soil organic matter made a great contribution to the species composition in shrub and herb layers, which were the main factors significantly affected the distribution of species composition (Table 6).

Responses of soil physical and chemical properties to different gap sizes

Except for the significant increase in soil moisture in the middle gaps, the soil physical properties in the gaps were significantly lower than those in the control, which was basically consistent with our previous hypothesis. Soil moisture influences soluble nutrient uptake and microbial activity, which are determined by and closely related to plant growth (Gao et al., 2020). The increase of soil moisture in the middle gaps may be due to the close relationship between soil moisture and plant growth. However, the plant composition in this kind of forest gap is the most, so the soil moisture is significantly higher than that of other treatments. Studies have shown that soil moisture in different sizes of forest gaps will show a complex pattern, and there is no obvious law (Pang et al., 2016). Our study showed that only the soil moisture of the medium gap was significantly higher than that of the control, but the soil moisture of the large forest gap and small forest gap was significantly lower than that of the control (Table 1). For small canopy gaps, the gap formed by artificial felling will destroy the plants under the forest when the felled P. massoniana is removed so that the interception function of the plants in the forest is lower than that of the control and the soil moisture in the gaps decreases (Meyer, Sisk & Wallace Covington, 2001; Simonin et al., 2007). For larger canopy gaps, in addition to more rainfall in the forest, the increased soil water evaporation and solar radiation in the forest gap will also have a negative impact on soil moisture (He et al., 2013; Ma et al., 2014; Han et al., 2020). These contradictory processes are likely to complicate the relationship between soil moisture and gap size. Soil bulk density is crucial in soil development (Mora & Lázaro, 2014). Soil bulk density can not only affect water infiltration and root growth but also play an important role in the maintenance of water capacity and plant nutrient availability (Salazar et al., 2009). In this study, soil bulk density showed a parabola with the change in gap size, reached the lowest point in the medium gap, and then gradually increased, and bulk density typically increased as soil depth increased. The setting of the forest gap improves the compactness of soil.

Soil organic matter and hydrolyzed nitrogen stocks are controlled by the complex interplay of soil physical, chemical, and biological conditions (Li & Nie, 2020a). In this study, compared with other larger forest gaps, soil organic matter and hydrolyzed nitrogen were significantly higher than those of the control, and the opposite trend of soil organic matter and hydrolyzed nitrogen appeared in the small forest gap (Table 1). With the increase in gap area, the content of soil available potassium decreased significantly, which may be because in the P. massoniana plantation, the fallen needles may stay on the forest ground for a long time, resulting in slow decomposition, which may eventually reduce the nutrient absorption of trees. At the same time, withered pine needles can also prevent other plant litter from decomposing and being absorbed by the soil (Chen & Cao, 2014; Ali et al., 2019). In summary, the medium gap can absorb soil water very well, improve the soil bulk density to a great extent and is conducive to the accumulation of soil organic matter and hydrolyzed nitrogen, which plays a positive role in the environment of the plantation. However, the conclusions of this study can only be applied to similar ecological environment. Therefore, the responses of plant diversity and soil physicochemical properties to other environmental conditions are recommended to be studied in the future.

Responses of undergrowth plant diversity to different gap sizes

The gap size is the direct “driving force” for natural regeneration in forest gaps (Zhang et al., 2018). The species composition of shrubs and herbs in the forest gap changed obviously, and they were in the stage of competition and coexistence among populations. Sufficient light and rainwater led to the formation of plant communities under forest gaps of different sizes (Jokela, Siitonen & Koivula, 2019). The initial period of the formation of the forest gap increased the light intensity in the forest, promoted the seed germination of solar plants in the woodland, provided suitable environmental conditions for the settlement of plant seedlings, and reduced the proportion of shade-tolerant plants. However, with the continuous development of the gap, the light transmittance is gradually weakened and the distribution is patchy, which leads to the emergence of different shade-tolerant tree species, making the species composition in the gap more abundant. There were the most abundant plant species in the medium gaps (Fig. 2), and the number of light-loving plants under the shrub layer increased greatly (Table 3). The proportion of shade-tolerant plants in herb layer also increased significantly, and there were many shade-tolerant species that did not appear in other treatments, such as Lophatherum gracile Brongn, Micronesia hancei Prant, Gonostegia hirta (Bl.) Miq, Liriope spicata (Thunb.) Lour., and Aster tataricus L. f. (Table 4). Neutral plants also account for a part of medium gaps. But in the large gaps, the proportion of light-loving plants increased the most, and the important value of Aralia chinensis (L.) was much higher than that of other treatments, and the proportion of Myrsine africana (L.), Acanthopanax trifoliatus (L.) Merr., Euscaphis japonica (Thunb.) Kanitz, Rosa cymosa Tratt. and Rubus corchorifolius L. f. should not be underestimated. Due to the heterogeneity of the gap environment, different species have different responses to the gap, which makes the diversity of shrubs and herbs different in the gap.

Gaps in forest canopies play a major role in local species richness because they increase habitat heterogeneity and accelerate the development of more complex forests (Swartz et al., 2020; Tena et al., 2020). In this study, the significant increase in species richness of the shrub layer in the forest gap may be because sufficient sunlight in the gap can cause more plants to grow under the forest (Wang et al., 2019b). The maximum value of species richness of the shrub layer appeared in the medium gap (Fig. 2). Generally, the herb layer has stronger sensitivity than the shrub layer. However, there was no practical relationship between the species richness index and forest gap sizes in this study, contrary to the previous hypothesis, which may be due to the weak sensitivity of herbaceous plants to the light response in forest gaps (Nowińska, 2010). The Shannon-Wiener index, Simpson index and Pielou index of the shrub layer and herb layer in the forest gap showed the opposite trend. The Shannon-Wiener index, Simpson index and Pielou index of the herb layer were significantly lower than those of the control (Table 5). This result shows that the herb layer is vulnerable to external interference, and the sharp changes in plant species and quantity make the distribution of different species in the biological community more uneven, and the ecological function of herbaceous plants is not prominent (Wang & Zhang, 2019a). The light transmittance in the large area gap is higher than that in the medium or small area gap, so the light environment in the large area gap is too bare and the dry and wet distribution in the gap is uneven such that the plant growth is relatively slow, which is more suitable for the growth of solar plants (Ritter, Dalsgaard & Einhorn, 2005). However, in the medium gap, the environment provided for plant growth is conducive not only to the growth of light-loving plants but also to the growth of shade-tolerant or neutral plants in the process of increasing light. Therefore, the medium gap can maximize the plant diversity under the forest. In a biological system of enormous heterogeneity and complexity, the same forest management measures may also draw different conclusions, so the popularization of our conclusions will also be limited.

Interaction between undergrowth plant diversity and soil physical and chemical properties

It is well known that species composition cause changes in soil properties that lead to complex local interactions between vegetation and soil (Pérez-Bejarano et al., 2010). In the soil-vegetation ecosystem, soil and vegetation interact with each other (Li, Wang & Zhang, 2020c). Different forest ecosystems possess different plant-soil interactions. In our study, the species diversity in the forest gap was closely related to the soil available potassium content, surface soil bulk density and soil moisture (Fig. 3). There was considerable correlation between the content of available potassium in soil and the parent material of the soil itself. Current reports show that potassium limits plant productivity in terrestrial ecosystems due to limited nitrogen and phosphorus at a global level (Sardans & Peñuelas, 2015; Li et al., 2021). In this study, the content of soil-available potassium was negatively correlated with the plant diversity index of the shrub layer but positively correlated with the Simpson index and Pielou index of the herb layer, indicating that available potassium was the limiting factor of the shrub layer. The content of soil-available potassium was easy to eluviate due to rain water erosion in the early stage of the forest gap, and the plant growth of the shrub layer produced negative feedback (Cox et al., 1999). Herbaceous plants have low requirements for soil nutrients, so they can continue to grow even under the loss of potassium ions (Wang et al., 2018). Soil moisture content and soil bulk density are important for soil hydrological processes (Grote et al., 2010). Improvement of these two physical properties plays an important role in guiding forestry management and improving forest productivity (Boluwade & Madramootoo, 2016). Soil moisture is the main factor affecting direct or indirect changes such as rainwater, soil erosion, runoff and plant growth (Wang et al., 2012; Sun et al., 2020). Forest gaps formed by felling can affect the infiltration of the soil moisture layer and deep layer and affect the participation of underground plants and microorganisms in the geochemical cycle (Vereecken et al., 2014). Therefore, the soil moisture content and soil bulk density were closely related to the plant diversity of the shrub layer and herb layer. Soil organic matter is vital to biodiversity to maintain many functions of plant growth (Kooch & Bayranvand, 2019; Kooch, Ehsani & Akbarinia, 2020). Increasing the litter amount in the forest gap accelerated soil microbial decomposition and promoted an increase in the organic matter content in topsoil (Bongiorno et al., 2019). The soil organic matter content increased significantly with shrub layer species richness and affected the degree of increase of the shrub layer species to some extent. Moreover, the soil bulk density, soil available potassium and soil organic matter content also significantly influenced the vegetation distribution in the shrub and herb layers (Fig. 4) (Table 6). Compared with the control treatment, soil physicochemical properties and plant distribution were significantly improved under different forest gap sizes.