Competition and Facilitation Co-Regulate Spatial Patterns and Coexistence in a Coniferous and Broad-Leaved Mixed Forest Community in Zhejiang, China

Abstract

Plant interactions have long been the subject of intense research and debate in ecology. Competition and facilitation are known to be the basic drivers of community structure, species coexistence, and forest succession dynamics; however, there remains a lack of understanding of how spatial patterns affect these in the mixed forest community of coniferous and broad-leaved species. In the present study, we studied the roles of competition and facilitation in a mixed forest community in the Zhejiang Province, China. For this, we divided plants measured at a study site according to the size of the diameter at breast height (DBH) into three levels of saplings as follows: saplings (1 cm ≤ DBH < 5 cm), juveniles (5 cm ≤ DBH < 15 cm), and adults (DBH ≥ 15 cm). We used the spatial point pattern method to explore the raw number, spatial distribution, and both intra- and inter-specific correlations of coniferous and broad-leaved species at these different diameter levels. The corresponding findings were as follows. First, the DBH and structure of coniferous and broad-leaved species were similar, showing a distinct distribution in an inverted “J” shape as the number of individuals with a particular diameter declined as DBH increased. Second, when all individuals of a similar diameter range were pooled, coniferous species were clustered at all size scales; however, in this situation, broad-leaved species (evergreen and deciduous) showed a clumped distribution at small scales, and this decreased as spatial scale increased. Third, there were small, but significant positive correlations between large-diameter trees and small-diameter trees of coniferous species and between evergreen broad-leaved species at different diameter levels; however, as the scale increased, the correlation diminished. On any scale, individual deciduous broad-leaved species at different diameter ranges did not have any significant correlation. Fourth, coniferous species had a strong competitive effect on broad-leaved species, and there was a weak promoting effect between deciduous and evergreen broad-leaved species as well. In summary, different tree species occupy similar living spaces, and through competition and facilitation, regulate the spatial pattern and community stability of coniferous and broad-leaved species in mixed forest communities.

1. Introduction

The spatial distribution of a community refers to both the physical locations of all individuals and, at a certain point, in time [1]. Generally, the spatial distribution of plant communities can be divided into three types: aggregated, uniform, and random [2]. Natural communities often exhibit a clumped distribution, which is determined by effects at scales that are either large, such as habitat heterogeneity, or small, such as plant interactions [3]. Interactions between plants, whether inter- or intra-species, determine the structure of the forest community and related, but complex, abiotic and biotic processes [1], and, as a result, significantly affect the spatial distribution patterns of species [2,3,4]. As the two indispensable factors and primary driving forces in succession in forest communities, competition and facilitation strongly impact the mechanisms of species’ spatial coexistence, community structure, and function [5,6,7,8]. For example, in a study regarding the three predominant species on Vancouver Island, Canada, Getzin et al. [4] found that based on the spatial patterns, intra- and inter-specific competition existed between these species. Liu et al. [8], meanwhile, found the spatial pattern of plants in the cold temperate forest in Kanas (Xinjiang, China) was influenced by competition and facilitation; moreover, the effect of facilitation had a larger effect on the growth and survival of trees. Therefore, the dynamic analysis of spatial distribution patterns based on species enables an understanding of the survival strategies of species, such as resource utilization, parasitism, predation, and reproductive characteristics [9,10,11]. Such analysis can reveal key characteristics of the community, interactions between species, and the unique habitat preferences of certain species [9,12,13], which is essential for understanding population dynamics and the processes that form community structure [5,6].

Mixed coniferous and broad-leaved forests, composed of a blend of both coniferous and broad-leaved (deciduous and evergreen) species, are the main forests of high-altitude mountain areas in the Zhejiang Province of China; this mixed community is rich in species composition and structure. How different coniferous and broad-leaved species coexist in stable equilibrium in mixed forest communities and how plant interactions affect the community structure have not been elucidated in detail. Using dynamic monitoring in a mixed forest of coniferous and broad-leaved species in Jingning, Zhejiang Province, we aimed to address these questions by examining the spatial correlation and distribution patterns of conifer and broad-leaved species using Ripley’s K (d) function analysis. Our specific questions are as follows: (1) What is the spatial distribution of different tree species in the mixed forest communities? (2) What interactions exist between different species in the mixed forest communities? (3) Does competition have a prominent role in maintaining species coexistence and community stability in the mixed forest community?

2. Study Site and Methodology

2.1. Study Site

The study site was in Jingning Caoyutang (27°31–27°56′ N, 119°30–121°55′ E), situated at the junction of Daji and Jingnan Township in the She nationality Autonomous County (Jingning, Zhejiang Province). With a total area of 5.5 km², the average elevation of this area is 1253 m and the largest peak reaches 1689 m. A central subtropical monsoon climate, composed of wet, hot summers and dry, cold winters, is typical for the region. The average annual rainfall reaches 1918 mm, the frost-free period is 196 days, the annual sunshine time is 1717.6 h, the annual average temperature achieves 11.8 °C, and the forest cover rate is 96.5%; moreover, the soil type is red soil.

2.2. Sample Surveys and Data Collection

Between May and July 2018, we collected data at a study site situated in a typical subtropical mixed forest composed of broad-leaved and coniferous species; moreover, based on the sample construction standards of the Center for Tropical Forest Science (CTFS), a dynamic forest monitoring study site of 1 ha (100 m × 100 m) was set. The entire plot was split into 20 m × 20 m subplots using the adjacent lattice investigation method, and stainless steel tubes were used as markers at four corners of each of the 25 subplots. Using an aluminum plate, we counted living woody plants in each sample with DBH ≥1 cm; moreover, we marked plants at a height of 1.3 m with red paint and recorded the germination status, branching, coordinates, tree height, DBH, and species names of all individually labeled woody plants.

We used woody plant diameters as a measurement of age. Linear regression showed that DBH had a significant and positive correlation with tree height (R² = 0.88, p < 0.001; Figure A1). Therefore, we combined it with the actual diameter level for our analyses, for which we categorized plants into classes based upon DBH: saplings with a 1 cm ≤ DBH < 5 cm; saplings with a 5 cm ≤ DBH < 15 cm; adults with a DBH ≥ 15 cm.

2.3. Point Pattern Analysis Method and Null Model

Using the pair-correlation function g (r), we investigated spatial distributions at the individual and community levels. The K function derived the pair-correlation function g (r); moreover, to substitute the circle in the K function, the ring was utilized in this study. In addition, in the calculation, the cumulative effect was expounded [12]. In this context, to explore the spatial correlation between two species, the paper adopted a double variable g (r) function, and a single variable g (r) function was also utilized to explore the degree of spatial clumping of species.

To a large extent, a species’ distribution is influenced by environmental heterogeneity. In addition, Wiegand et al. [12] found asymmetric competition among species. Considering that for individual species, the survival rate in different habitats is variable, and both habitat association and diffusion limitation are influential. We used the heterogeneous Poisson process for the null model selection to simulate bivariate statistics for study species. As a spatial heterogeneity function, λ(s) demonstrates the relationship between individual density and habitat heterogeneity [8,14] For simulations, the position of one tree species was initially fixed, and then a heterogeneous Poisson process was adopted to randomize the position of the spatial distribution of another tree species with a maximum distance of 60 m [5]. After that, we examined changes in the spatial correlation between the two tree species. In the next step, the position of the latter tree species remained stationary, and the spatial distribution of the former tree species was randomized based upon a heterogeneous Poisson process. In order to obtain a 99% confidence interval, 100 Monte Carlo simulations were carried out for each species pair. For univariate results, the relationship is thought to have a clumped distribution when the observed value exceeds the confidence interval. On the contrary, a confidence interval higher than the observed value indicates a random spatial distribution; moreover, a confidence interval including the observed value is expected for a hyper dispersed or regular pattern. For the bivariate analysis, if the confidence interval did not exceed the observed value, it indicates a positive correlation. Conversely, when the confidence interval is higher than the observed value, it is a negative correlation. Finally, if the confidence interval included the observed value, the two species are considered to not have any significant correlation [5,15]. The “spatstat” package in R 3.2.4 was utilized for all analyses [16].

3. Results

3.1. Diameter Structure of Broad-Leaved and Coniferous Species in the Mixed Forest Community

The 1-ha conifer/broad-leaved mixed study site was an artificial replacement plot with 4378 surviving woody plants (DBH ≥ 1 cm) belonging to 22 families and 31 genera. Among them, there were gymnosperms belonging to 3 families, 3 genera, and 4 species, including Taxodiaceae, Taxaceae (1 genus, 1 species) and Pinaceae (1 genus, 2 species). The most dominant families in the plots were Pinaceae (Pinus massoniana), Magnoliaceae (Liriodendron chinense, Magnolia officinalis), Hamamelaceae (Liquidambar formosana), Theaceae (Schima superba) and Fagaceae (Cyclobalanopsis myrsinifolia, Castanopsis sclerophyll). There were 2359 coniferous trees composed of 4 species, 756 evergreen trees composed of 9 species, and 1263 deciduous trees composed of 28 species, which accounted for 53.9%, 17.3%, and 28.8% of the total tree population, respectively. Coniferous and broad-leaved species had similar inverted “J” distribution of trees with different diameters, and the absolute number of trees decreased with increasing tree DBH (Figure 1).

3.2. Spatial Distribution of Broad-Leaved and Coniferous Species

Overall, at a scale of 0–100 m, coniferous species demonstrated an aggregated distribution (Figure 2). The young and juveniles of coniferous species showed a clumped distribution on a very small scale of r ≤ 4 m and random distributions at both a larger scale (r > 10 m) and overall (0–100 m; Figure 2).

All evergreen and deciduous broad-leaved species displayed similar spatial distribution patterns, namely a clustered distribution on a small scale (r ≤ 10 m) and a random distribution on a large scale (r > 10 m; Figure 1). Saplings of evergreen and deciduous broad-leaved species showed a clumped distribution on a small scale (r ≤ 10 m) and a random distribution on a large scale (r > 10 m). Juveniles of deciduous broad-leaved species had a clumped distribution at r ≤ 5 m and a random distribution at r > 5m. Medium-sized and adult deciduous species, both broad-leaved and evergreen, exhibited random distributions on a scale of 0–100 m (Figure 2).

3.3. Spatial Correlation in Coniferous and Broad-Leaved Mixed Forest Communities

Adult and juveniles of the mixed forest community we examined have significantly positive correlations (p < 0.05) with saplings on a small scale (r ≤ 10 m), but there was no significant correlation as the scale increased. Adults showed no significant correlation with juveniles on any scale (Figure 3A–C). For broad-leaved evergreen species, there were significant positive relationships between many of the size classes on a small scale (r ≤ 10 m), including small and saplings, big and saplings, and big and small trees (p < 0.05 in all cases); however, no significant correlations were observed at r > 15m (Figure 3D–F). Broad-leaved deciduous species, on the other hand, exhibited no significant correlations at any scale (Figure 3G–I).

3.4. Intra-Species Spatial Correlation of Different Diameter Categories in the Mixed Forest

3.4.1. Spatial Correlation between Broad-Leaved Species (Deciduous and Evergreen) and Coniferous Species with Different Diameters

In the mixed forest we examined, we found a significant positive relationship (p < 0.05) between coniferous and broad-leaved species (evergreen and deciduous). The smaller the difference in diameters between broad-leaved and coniferous trees, the stronger the negative correlation of spatial separation (Figure 4). Specifically, the young coniferous trees were significantly and positively correlated with broad-leaved species (p < 0.05) on a small scale (r ≤ 10 m), but not on a larger scale (r > 10 m; Figure 4A,G). Juveniles of coniferous species were significantly and positively correlated with saplings of broad-leaved species on a small scale (r ≤ 10 m) and at a 20–30 m scale (p < 0.05), but no significant correlation was observed at r > 35 m (Figure 4B,H). In addition, juveniles of coniferous species significantly and negatively correlated with broad-leaved trees (p < 0.05) at r ≤ 30 m, but no significant correlation was observed at r > 30 m (Figure 4C,I); moreover, adults of coniferous species significantly and positively correlated with saplings of broad-leaved species (p < 0.05) on a small scale (r ≤ 10 m), but there was no significant correlation on a large scale (r > 10 m; Figure 4D,J). Finally, the correlation between adults of coniferous species and juveniles and adults of broad-leaved species was significantly negative at r ≤ 20 m (p < 0.05), but no significant correlation (r > 30 m) was found on larger scales (Figure 4E,F,K,L).

3.4.2. Spatial Correlation between Broad-Leaved Species and Coniferous Species at Different Diameter Levels

The spatial correlation of broad-leaved species with coniferous species of different diameters was mainly reflected in the middle- and small-diameter classes, as large-diameter individuals of broad-leaved species had no significant effect on coniferous species. Saplings of broad-leaved species significantly and positively correlated with coniferous trees at r ≤ 5 m, but not on a large scale (r > 10 m). Juveniles of broad-leaved species significantly and positively correlated with saplings of coniferous species at r ≤ 5 m, but this was significant and negative at 10–35 m (p < 0.05) and not significant at r > 35m. A significant negative correlation was found between juveniles of broad-leaved species and coniferous species (p < 0.05) on r ≤ 25 m, but no significant correlation was observed at r > 25 m. Large-diameter broad-leaved species did not significantly correlate with coniferous species at any scale (Figure 5).

3.4.3. Spatial Correlation between Evergreen Broad-Leaved Species and Deciduous Broad-Leaved Species

The spatial correlations between deciduous and evergreen broad-leaved species in the medium- and small-diameter grades were similar. Young and juveniles showed a significant positive correlation (p < 0.05) at r ≤ 3 m, a significant negative correlation (p < 0.05) at 20–37 m, and no significant correlation at r > 37m (Figure 6A,B,G,H). Juveniles showed a significant negative correlation (p < 0.05) at r ≤ 27 m and no significant correlation on a large scale (r > 40 m; Figure 6C,I). Adults of evergreen broad-leaved trees showed significant positive correlations with saplings of deciduous broad-leaved trees at r ≤ 6 m, and no significant correlation at r > 6 m (Figure 6D). Adult deciduous broad-leaved trees significantly and positively correlated (p < 0.05) with evergreen broad-leaved trees on a small scale (r ≤ 10 m), negatively correlated at 26−36 m, and no significant correlation was observed at r > 36 m (Figure 6J). There was a significant positive correlation between adults of evergreen broad-leaved trees and juveniles of deciduous broad-leaved trees at r ≤ 5 m, but no significant correlation was observed at r > 5 m (Figure 6E). Adults of deciduous broad-leaved trees significantly and negatively correlated (p < 0.05) with juveniles of evergreen broad-leaved trees at r ≤ 30 m, whereas at r > 30 m, the correlation was not significant (Figure 6K). There was no significant correlation between the adults of deciduous and evergreen broad-leaved species at any scale (Figure 6F,L).

4. Discussion

In this study, it is observed that competition and facilitation processes occurred simultaneously to shape the mixed forest community of broad-leaved and coniferous trees, notably contributing to their stable coexistence [17,18]; however, results also demonstrated that competition was more common than facilitation as a significant force. In particular, coniferous species had a more significant suppressing effect on broad-leaved species, and competition was also observed between evergreen and deciduous broad-leaved species. In contrast, facilitation was observed only weakly on a small scale between different size categories within and between the species. Therefore, it is speculated that the competitive relationship between trees is essential to the coexistence of coniferous and broad-leaved species in these mixed forest communities.

4.1. Population Dynamics of the Mixed Forest Community

Broad-leaved and coniferous species from the mixed forest community had an inverted “J” diameter/age distribution; moreover, both broad-leaved and coniferous tree species propagate using seed dispersion as the main reproduction mode. Saplings and juveniles accumulate in areas with good environment conditions that favor their growth, development, and reproduction. In this paper, the results demonstrated that when all diameter-level individuals were pooled, coniferous species exhibited clumped distributions at any scale, and the degree of being clumped did not grow significantly as the scales increased. Higher light conditions are more favorable [19] for the regeneration and growth of coniferous species, which can rapidly diffuse and reproduce and hence exhibit a random dispersion pattern across the study site. Furthermore, the tree diameter was inversely proportional to the clumped distribution degree, as was previously observed by Omelko et al. [18]; however, on a small scale, broad-leaved species also showed clumped distribution. Notably, as scale increased, the clumped distribution degree also decreased [8,9,20]; it may be that broad-leaved species were mainly small in diameter, the forest recovery was short and slow, and the individual trees were not affected by a density restriction and hence a clumped distribution resulted. Our results are in high agreement with numerous similar studies that found subtropical mixed forest communities of coniferous and broad-leaved species are primarily shaped by the species’ density, spatial scale, and the tree size resulting in a small-scale, intra-specific clumped distribution [4,17,21]; moreover, our results further support those of both Wang et al. [20] and Condit et al. [9] in that in deciduous and evergreen broad-leaved communities where the species types are relatively few, clumped distributions tend to form.

Interactions between individual trees greatly influences the distribution pattern of tree species [3]; this is illustrated in our finding that coniferous species had a significant and negative correlation with broad-leaved species on a small scale, whereas an opposite, positive relationship was noted for the evergreen broad-leaved species; this could be seen as evidence of the competition and facilitation processes simultaneously regulating spatial patterning of the mixed forest community.

Through intra- or inter-specific density-dependent effects, the distribution and number of trees in a population can be regulated through competition [21,22]. When it comes to the trees with variable life history characteristics or different sizes, facilitation is the main driver as trees can cover newly established seedlings and shrubs, or adults which cover juveniles [22,23]. Under the influence of a diffusion limitation, most young and juveniles gather around the mother tree, forming an intra-specific clumped distribution and inter-specific dispersion [8,24]; however, as the tree diameter grows, the demand for key environmental factors, such as light, water and heat, mineral nutrients, and space, increases [19]. Due to intense intra-specific competition, the phenomenon of plant self-thinning may also cause the population to decrease gradually with the increase in diameter [9,25]. According to Wang et al. [20], competition-induced self-sparing reduces the intensity of the clumped distribution among large-diameter trees; this phenomenon demonstrates that for trees from small to large, the observed increase in excessive dispersion of trees was caused by density-dependent mortality [4,26]. As shown in the results above, the population dynamics and the spatial distribution of coniferous forests in mixed forest communities are controlled by intra-specific competition [21]; however, intra-specific facilitation also plays a significant role in evergreen broad-leaved species. Additionally, we found that a negative correlation was more frequently observed than a positive correlation, and thus, it is boldly speculated that competition effects are more common than facilitating effects.

4.2. Interspecies Interactions Maintain Stable Coexistence of Communities

There is much evidence that plant interactions in forest ecosystems are often dynamic processes that may be either antagonistic or cooperative [27]. Our study found that the competition between middle-sized trees is most intense, probably because at the peak period of growth, plants with highly developed stems and roots have the greatest demand for resources; moreover, the inter-specific competition effect of broad-leaved species on coniferous species was pronounced only among the trees of small and medium sizes, and a significant correlation was not found between larger individuals. Small plants need the same nutrients for growth and may have strong competitive interactions, higher productivity, and higher resource requirements in relatively superior habitats [8,19]; moreover, interspecific competition can prevent the competitive exclusion of dominant individuals, promote the utilization of resources, and reduce abiotic stress [28]. To some extent, evergreen and deciduous broad-leaved species facilitate one another, further proving that the species richness of the entire community becomes increasingly complex. Differences in ecological strategies resulting from the different functional traits (e.g., specific leaf area, specific stem density, etc.) of evergreen and deciduous broad-leaved species [7] may reduce the competitive intensity, whereas seasonal deciduous leaves in the canopy provide more niche space for species regeneration and growth. Different growing season conditions may also reduce competitive exclusion, thus increasing species richness and community diversity [29,30]. With the increase in stress intensity, the transition from competition to facilitation may give less competitive species a chance to survive in the community during the growing season [29]. Inter-specific competition might appear mainly between trees with the same shade needs, whereas inter-specific facilitation could be observed among trees with different shade needs [18]. Changes in plant interactions may lead to the dynamic coexistence of species and promote the succession process of the forest community.

5. Conclusions

In the mixed forest community we examined, the density of trees was large, the composition of tree species was rich, and a complex combination of spatial forest structures was formed, which resulted in different levels of competition in the forest. Coniferous species inhibited the growth of other tree species, whereas deciduous and evergreen broad-leaved species promoted each other and grew together. Theoretical and empirical evidence suggests that under relatively less stressful abiotic conditions, inter-species competition may take prominence in community construction and species coexistence. The results demonstrated in this study provide further insight on the growth strategy and adaptability of tree species in coniferous and broad-leaved mixed forests.