One of the challenges in ecological research is how to accurately predict future environmental changes on the abundances and spatial distributions of species, and the geographical patterns of biodiversity in the future (Cardinale et al., 2012; Urban et al., 2016). Changes in abiotic/biotic conditions or disturbance regimes can all impact population dynamics and community structure. Finding the likely driving factors of these changes is helpful in predicting the responses of species populations or biotic communities and in designing mitigation strategies to reduce the harmful effects of environmental change and human activities. Determinants on the distribution of species or plant communities explored by most of the previous studies are about abiotic factors (Harrison et al., 2020; Madrigal-González et al., 2020), yet evidence of how biotic factors shape the distributions of species or their assemblages at different spatial scales have received less attention (Araújo and Luoto, 2007). Ecological strategies, representing plants exhibiting similar adaptions to environmental conditions, can reflect tactics employed by species for growth and survival (Walker, 1992). An ecological strategy spectrum is actually the proportions of species with different ecological strategy types in a biotic assemblage. Similar to a life-form spectrum (Raunkiaer, 1934), it summarizes the enormous complexity of species with diverse ecological strategies into a few general recurrent groups. It can reflect combinations of species’ adaptation tactics coping with specific environmental conditions, which can provide an effective way for comparing how species composition of vegetation differs in their spatial distributions across different biogeographic regions and for detecting the key drivers of these differences. Consequently, exploring the impacts of biotic factors on ecological strategy spectra of forests can further enhance our understanding of the population adaptation performance of plant species and the resultant community features and improve our ability in predicting species responses to environmental changes.

A well-recognized species adaptation strategy scheme is Grime’s competitor, stress-tolerator, ruderal (CSR) theory (Grime, 1974, 1977, 2001; Grime and Pierce, 2012). Based on the trade-off among the competition, stress, and disturbance resulting from natural selection, three principal strategies characterized by the occurrence of combinations of different traits. Specifically, C-selected competitors are characterized with traits to aid resource preemption surviving in high-productivity habitat; S-selected stress-tolerators are characterized by traits in response to highly unproductive and abiotically variable habitat; and R-selected ruderals are characterized with traits in the face of frequent disturbances. This theory simultaneously explains both economics (a trade-off in resources investment in “acquisitive” to “conservative” growth) and size (organs and whole plants) as fundamental gradients of plant evolution (Diaz et al., 2016; Pierce and Cerabolini, 2018). Yet, due to the lack of a simple protocol for identifying the ecological strategies of species, the utilities of CSR theory have been debated for many years (Westoby, 1998; Craine, 2005). Since functional traits within and among plant species can provide a surrogate for some fundamental ecological processes (McGill et al., 2006), identifying species ecological strategies based on plant functional traits has received more and more attention (Pierce and Cerabolini, 2013; Diaz et al., 2016; Pierce et al., 2017). Based on functional traits, Grime’s CSR framework has been applied widely in ecological research (Caccianiga et al., 2006; Schmidtlein et al., 2012), e.g., examining species coexistence (Pierce et al., 2014) and correlation of species richness and productivity (Cerabolini et al., 2016). For promoting a wider range of applications and broader generalization of the CSR theory in biogeographically distinct contexts worldwide, Pierce et al. (2017) proposed a CSR ordination method (i.e., StrateFy). This method used only three-leaf functional traits, i.e., leaf area (LA), specific leaf area (SLA), and leaf dry matter content (LDMC), to classify plant species into 19 ecological strategy categories and ordinate them in CSR triangle. The protocol has been experimentally tested by manipulating resource availability and disturbance (Li and Shipley, 2017), and it has been used to investigate community assembly processes or vegetation responses to disturbance (Rosado and de Mattos, 2017; Han et al., 2021).

However, 19 ecological strategy categories may be too numerous for identifying the actual ecological responses of species to varying habitats and for clearly defining and explaining how each of the strategy categories responds to specific changes in environmental conditions across a large geographical scale. Consequently, we regrouped the species into only four ecological strategy groups, namely, C-strategy group (C-group), S-strategy group (S-group), R-strategy group (R-group), and intermediate-strategy group (Int-group). The regrouping allowed us to quantify how species were differentiated with environmental changes. To the best of our knowledge, previous studies have applied CSR theory and CSR strategy ordination method StrateFy to discern and compare the distribution of plant ecological strategies in responses to abiotic conditions, such as anthropogenic disturbance (Wang et al., 2018; Han et al., 2021), natural disturbances (Barba-Escoto et al., 2019; Matos et al., 2021), and climatic changes (Rosenfield et al., 2019). However, contrary to the effects of abiotic factors, studies on how biotic factors affect plant species strategies’ distributions at various spatial scales (e.g., landscapes, regions, and continents) have gained less attention in the background of predicting distribution patterns of vegetation in a changing environment.

Each species has its unique morphological, physiological, and phenological characteristics, i.e., having distinct functional traits, which make a species adapt to various environmental conditions. Communities are composed of species with different functional traits; thus, different species combinations in communities will affect their functions in ecosystems and the ecosystem services. Consequently, species-level analyses can provide insights into understandings at the community or ecosystem level (Symstad et al., 2003; Xu et al., 2019). In addition, species diversity is a key and common biotic driver of ecological processes and ecosystem functions (Tilman et al., 2014). Therefore, we selected three species diversity indices in this study, i.e., species richness, Shannon-Wiener index, and stem density (stem abundance), as possible biotic variables of driving variations in ecological strategy spectra of forest vegetation across different climatic zones. Species richness is a count of species within a biotic community, which is frequently used as the primary measure of biodiversity and is usually involved in other metrics of diversity (Burrascano et al., 2018). Abundance is an important measure of species diversity for evaluating species’ impact on local ecosystems (Ehrlén and Morris, 2015). It has also been incorporated into some indicators of biodiversity (Kenderes and Standovár, 2003). Shannon-Wiener index is a synthetic index, involving both richness and abundance to quantify distributions of individuals among different species in a community (Jabot and Chave, 2009; Bernhardt-Römermann et al., 2015). These biotic variables can reflect the impacts of plant interactions on plant community characteristics (Poorter et al., 2015). Hence, exploring the associations between ecological strategy spectrum and species diversity may provide insights into understandings on how biotic factors impact vegetation and their responses to changing environment.

Currently, one major obstacle in analyzing the associations between ecological strategy spectrum and species diversity is a lack of available data on vegetation investigation across different geographical regions. Here, we established 200 0.04-ha forest dynamics plots (FDPs) in four vegetation types of different climates and carried out a detailed investigation of woody plants (e.g., trees and shrubs) and measurement of their three-leaf functional traits. We calculated three species diversity indices (i.e., species richness, Shannon-Wiener index, and stem abundance) as potential biotic variables for driving variations of ecological strategy spectra for forests across different climatic zones. Based on CSR theory (Grime, 1974, 1977, 2001) and the CSR strategy ordination method StrateFy (Pierce et al., 2017), we identified species’ ecological strategies and ordinated them in the CSR triangle, thus obtaining the ecological strategy spectrum of each plot. Our objectives were to assess how species diversity impacted the variation of the ecological strategy spectra among different forest types by (1) assigning each species included in our data set into the CSR triangle to quantify the proportions of species in C-, S-, Int-, and R-groups in each forest type and then identify the ecological strategy spectra of forests across different climatic zones; (2) exploring the associations between the proportions of C-strategists, S-strategists, Int-strategists, and R-strategists with each species diversity index; and (3) assessing the relative importance of the three selected species diversity indices in regulating ecological strategy spectra of forests across different climatic zones. Rising diversity can lead to increasing competition among species (Grime, 2001; Aguilée et al., 2018). Since C-strategists and Int-strategists have more competitive ability than S-strategists characterized with the more tolerant ability for harsh conditions, we hypothesized that C-strategists and Int-strategists are positively associated while S-strategists are negatively associated with species diversity. Compared with the abundances of species and other biotic factors, the presence of species can more reflect variations of community compositions and environmental conditions across the earth (Kier et al., 2005; Mutke and Barthlott, 2005) and provide more to explain the origin and maintenance of species diversity (Currie et al., 2004; Weiser et al., 2007) across different climatic zones. Therefore, we hypothesize that species richness is the paramount biotic determinant for the variation of ecological strategy spectra of forests across different climatic zones.

A total of 536 (243 in tropical forest, 171 in subtropical forest, 115 in warm-temperate forest, and 7 in cold-temperate forest) species were found across the four study sites. The ecological strategy of each species for each forest type was identified and plotted in the CSR triangle (see the different colored symbols in Figure 1A). These species occupied different positions in the CSR triangle, resulting in different proportions of species in each of the four ecological strategy groups, and thus forming different ecological strategy spectra for different forest types (Figure 1B). Specifically, in tropical forest, S-strategists accounted for the largest proportion, followed by Int-strategists and C-strategists. In subtropical forest and warm-temperate forest, Int-strategists were the most, followed by S-strategists and C-strategists. Cold-temperate forest involved only S-strategists. However, R-strategists only occurred in the subtropical forest.

The linear regression showed that correlations between the proportion of species in each ecological strategy group and each biotic factor differed (Figure 2). The proportions of species in C- and Int-groups were positively associated with species abundance, species richness, and Shannon-Wiener index. Conversely, the proportion of species in the S-group was negatively associated with all the abovementioned three species diversity indexes. For C-group, species richness (SRC = 0.746) and Shannon-Wiener index (SRC = 0.704) emerged as the most important variables (Figures 2A–C). Stem abundance (SRC = −0.679) was the main determinant for the proportion of species in the S-group, followed by the Shannon-Wiener index (SRC = −0.601, Figures 2D–F). Stem abundance (SRC = 0.654) and Shannon-Wiener index (SRC = 0.454) were the most important biotic variables explaining the proportion of species in Int-group (Figures 2G–I). Unlike the other three ecological strategy groups, the proportion of species in the R-group was not significantly affected by the biotic variables in this study (Supplementary Figure 2).

RDA revealed that our three biotic variables were clearly associated with ecological strategy spectra (adj-R² = 65.71, F = 128.16, P < 0.001). Of the total variation in ecological strategy spectra, 43.51% was explained by the first two axes of RDA. The first RDA axis, explaining 39.65% variation in ecological strategy spectra, was negatively associated with stem abundance, species richness, and Shannon-Wiener index. The second RDA axis, explaining 3.86% of the variation in ecological strategy spectra, was positively associated with stem abundance, species richness, and Shannon-Wiener index (Figure 3A). The first two partial RDA axes as well as each of the variables independently had a significant loading (Supplementary Tables 3, 4), indicating that the biotic factors significantly affected variations in ecological strategy spectra of forests.

Variation partition analysis revealed that the variation in ecological strategy spectra explained by the three biotic variables amounted to 65.7%. Each of the variables independently accounted for a significant amount of variation in the ecological strategy spectra: stem abundance (44.3%), species richness (26.4%), and Shannon-Wiener index (40.1%; Figure 3B and Supplementary Table 5). These results suggested that stem abundance was the most important biotic variable in affecting the variation of ecological strategy spectra across different climatic zones.

We report the results from an extensive vegetation investigation across different climatic zones, analyzing ecological strategy spectra of tropical, subtropical, warm-temperate, and cold-temperate forests. This study revealed that species attributing to S-strategists presented the highest proportions in tropical and cold-temperate forests, while species belonging to Int-strategists occupied the highest proportion in subtropical and warm-temperate forests (Figure 1B). The proportions of species in C-, S-, and Int-groups were significantly associated with the three diversity indices (Figure 2), suggesting that species diversity significantly influenced ecological strategy spectra of forests across different climatic zones (Figure 3A). Furthermore, among the three biotic factors considered in this study, stem abundance was the most important biotic driver for variations of ecological strategy spectra of forests across different climatic zones (Figure 3B).

The proportion of species in the C-group increased as the species diversity increased (Figures 2A–C). We considered that favorable microclimatic conditions make more species coexist together, which could result in a more intensive competition among the coexisting species, leading to the higher proportions of C-strategists. Previous studies (Pretzsch, 2014; Jucker et al., 2015; Morin, 2015) have demonstrated that community with higher species diversity may characterize a more complex stand structure (Ehbrecht et al., 2017), which usually could buffer the heating and cooling effects of shortwave and longwave radiations (Häckel, 2008), creating a more favorable microclimate (Gebauer et al., 2012; Kunert et al., 2012; Forrester, 2015; Ehbrecht et al., 2017). Component with C-strategy was negatively associated with annual temperature range (Rosenfield et al., 2019) and temperature seasonality (Pierce et al., 2017). Benign or less stressful environments usually result in strong competition (the biotic filtering process) within and among species (Bertness and Callaway, 1994). These imply that favorable microclimate favors species with high competition ability survival and growth. In addition, species distributed in forest communities characterized with high species diversity usually possess higher competitive abilities (Aguilée et al., 2018). Ultimately, the proportion of C-strategists decreased from tropical forest through subtropical forest and warm-temperate forest to cold-temperate forest with decreasing species diversity. This finding indicated that certain species in forest communities with high species diversity would like to employ competitive ecological strategies to adapt to their living environment. In turn, the forest community with more C-strategists means higher species diversity, which may accompany a much higher conservation value.

The proportion of species in the Int-group was positively correlated with species diversity (Figures 2G–I). This may be due to the variable microhabitats resulting from higher species richness and stem abundance. Habitat-heterogeneity hypothesis has demonstrated that higher species diversity often leads to higher environmental heterogeneities within the forest communities (MacArthur and MacArthur, 1961). Higher environmental heterogeneities, in general, generate more variable microhabitats. For instance, light resource is a major limiting factor for the growth of trees. High diversity forests usually have more complex canopy structures, strongly impacting the distribution of light within forest communities (Kira et al., 1969). Each tree species usually has a characteristic crown architecture (Millet et al., 1998). In natural stands with higher species richness/stem abundance, trees present various growth forms, varying in size and crown architecture from understory through subcanopy to canopy or even overtop canopy species. This variation results in a multilayered foliage structure and enhances the structural complexity of the forest canopies, likely leading to the changes in light resources (Ishii et al., 2004) and thus the occurrence of various microhabitats occurring. Other than the effects of certain available resources due to species diversity, small-scale natural disturbance may be another important driver of environmental heterogeneities within forest community. Despite we have focused our sampling on old growth forests (i.e., in the late stage of succession) for minimizing the effects of disturbances on ecological strategy groups. However, natural disturbances are considered a crucial integral part of forest ecosystem dynamics and a major driver for species’ ecological strategies (Ehbrecht et al., 2021). For example, forests with higher species richness/stem abundance can provide more survival opportunities and more diverse food resources for more animal species (Southwood et al., 1979), especially for herbivores. In turn, different herbivores preferring different plant species or different aged stems (Lawton, 1983) cause a different degree of small-scale disturbance in a forest ecosystem. Meanwhile, high-diversity forest communities usually have more small-scale gap disturbances caused by death or breakage of large trees. Finally, uneven distribution of available light and small-scale disturbances may together create environmental heterogeneities and thus variable microhabitats in forest communities. Int-strategists characterized with intermediate between the competitor, the stress-tolerator, and the ruderal have higher adaptive capacity to the changing environmental conditions (Grime, 2001). Therefore, relatively higher stem abundance/species richness promote larger proportion of species in Int-group.

In contrast to the proportions of species in C- and Int-groups, the proportion of species in the S-group showed a negative correlation with species diversity (Figures 2D–F). This association between S-group and species diversity was believed to partly result from that S-strategists usually adapt to environments of high stress and low disturbance (Guo et al., 2019). In general, S-strategists have strong tolerance and resistance to stressful environments (Grime, 2001). Environmental stress (e.g., temperature) often restricts species richness (Wang et al., 2011), resulting in a higher proportion of S-strategists. In addition, S-strategists are characterized with low competitive ability for resources compared with C-strategists and their low tolerant ability to disturbances compared with Int-strategists (Pierce and Cerabolini, 2018). However, as we have discussed above, forest communities with higher species diversity theoretically favor the species with higher competitive ability or adaptive ability to variable microhabitats. Moreover, there are trade-offs among the C-, S-, and R-strategies according to the CSR theory, which suggests a compromise must be reached among the three in all species. For example, species with strong tolerance may show weak competitive ability (Bruelheide et al., 2018). Thus, we argued that trade-offs also existed among the C-, S-, Int-, and R-groups within forest communities for species coexistence. That is, the proportion of species in C- and Int-groups increased with species diversity may result in that in S-group decreased. Hence, the proportion of species in the S-group increased with the decrease of species diversity. For R-group and species diversity, as we expected, no significant correlations were found between them (Supplementary Figure 2). We assumed this may be due to the fact that the R-strategy is mostly adopted by herbaceous species, with only few woody species presenting the R-strategy in conditions with high disturbances (Rosenfield et al., 2019). However, we only focused on trees and shrubs and did not involve herbaceous species in this study.

Although all the three species diversity indices significantly affected the ecological strategy spectra of forest vegetation, their contributions were not equal in importance. Contrary to our hypothesis, variance partitioning revealed that stem abundance instead of species richness was the most important factor driving for variations of ecological strategy spectra among the three diversity indices (Figure 3B). This was in line with a previous study on the trait-demography relationships (Chalmandrier et al., 2021). Abundance was a much stronger indicator than the presence alone in species’ interactions in a community (Ehrlén and Morris, 2015). Furthermore, available space and resources for each individual as well as intra- and interspecific interactions to some extent can be reflected by stem abundance in communities (Fortunel et al., 2018). For example, high stem abundance was observed to have increased competition for light and space among individuals (Muscarella et al., 2018). In addition, the more individuals hypothesis (Gaston, 2000) also demonstrated that stem abundance can promote species richness (Madrigal-González et al., 2020). Shannon-Wiener index, which was second in importance of driving the variations in ecological strategy spectrum, is a complex combination of species abundance and stem richness (Jabot and Chave, 2009). These to some extent further confirmed the importance of stem abundance. Besides, the S-strategists and Int-strategists generally accounted for high proportions in the four forest types (Figure 1B) and were strongly affected by stem abundance (Figure 2). The above descriptions may together explain that stem abundance is paramount in promoting differentiations of ecological strategy spectra of forests across different climatic zones.

Our study, for the first time, explored how species diversity impacted shifts in plant strategy spectra of vegetation based on CSR theory across four climatic zones. The findings clearly revealed that species diversity significantly drove the differentiation in ecological strategy spectra of forests across different climatic zones. For instance, species diversity increased C- and Int-strategists and decreased S-strategists. Evidence suggest that only considering the abiotic drivers is unlikely to accurately predict species distributions responses to future environmental changes (Morris et al., 2020). Changes in biotic drivers cannot be ignored as well (Wisz et al., 2013). Our findings further emphasize that biotic factors should be considered for better predicting how species distributions and vegetation variations respond to environmental change. However, contrary to our expectations, this study showed that stem abundance, not species richness, was the most important biotic driver for variations of forest ecological strategy spectrum. Although the study demonstrated that species diversity regulates the ecological strategy spectrum, some limitations in this study should be acknowledged. The major limitation lies in the fact that in this study, we only considered taxonomic diversity (e.g., species richness and abundance) as biotic factors, regardless of functional diversity. Therefore, more studies will be needed to better probe how biotic factors impact plant adaptive strategies in the future to improve our prediction capability for the dynamic of community response to global change. Despite this limitation, the research on CSR theory provided new insights into understanding the roles of species diversity in regulating the compositions of plant communities at regional scales under changing environmental context. This can also provide some guidelines for forest management and vegetation restoration under environmental change settings.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

XH contributed to data analysis, manuscript drafting, and specimen identification. YX and JH performed the filed investigation and specimen identification. RZ designed the study. All authors participated in the writing.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.