1. Introduction

How and why species assemblages differ from one another along spatial gradients are longstanding concerns in ecology and biogeography (Alahuhta et al., 2017). Beta diversity, a measure of assemblage dissimilarity between sites, provides a promising approach for addressing abovementioned concerns. In particular, beta diversity can be partitioned into two distinct and antithetical components: turnover and nestedness (Baselga, 2010; Carvalho et al., 2012). Turnover signifies dissimilarity that originated from the replacement of some species by others from site to site, whereas nestedness denotes dissimilarity that originated from difference in richness caused by species gain or loss between sites (Baselga, 2012). It is generally asserted that turnover might favor conservation efforts that were allocated to multiple different regions, whereas nestedness implied that conservation efforts should focus on species-rich regions at the expense of species-poor regions (Socolar et al., 2016; Nascimbene and Spitale, 2017). Therefore, exploring patterns of beta diversity and its components could provide insight into origins of assemblage dissimilarity as well as guidance for conservation planning.

Generally, determinants responsible for beta diversity patterns predominantly result from two theories: (1) the niche-assembly theory (Whittaker, 1956; Hutchinson, 1957), which states that assemblage dissimilarity was caused by deterministic processes such as environmental filtering or competitive exclusion, and (2) the neutrality theory (Hubbell, 2001), which states that assemblage dissimilarity results from stochastic processes. These two theories are not mutually exclusive, but cooperate in natural systems (Chase and Myers, 2011), and their relative importance often changes among regions or taxa (Weinstein et al., 2014; McFadden et al., 2019). Evaluating the relative role of these theories in determining beta diversity can help us understand the underlying mechanisms of assemblage dissimilarity and have substantial implications for conservation planning. For example, niche-assembly theory suggests that different region may support different species assemblages owing to varied environments. Conservation efforts should, therefore, cover different types of habitats to sustain regional diversity (Legendre et al., 2005; Angeler, 2013). In contrast, neutrality theory emphasizes the roles of random colonization and extinction in determining species compositions, conservation efforts should focus on maintaining or restoring landscape connectivity to promote species dispersal and gene flow (Kadmon and Allouche, 2007).

Mountains play an important role in global biodiversity conservation and ecological experimentation because they support more than 85% of the world's species and form large-scale environmental gradients along elevation (Körner, 2007; Rahbek et al., 2019; Vetaas, 2021). However, while a large number of studies have focused on alpha diversity (i.e., species richness) with elevation, less attention has been paid to beta diversity along the elevational gradient (Wang et al., 2022). Moreover, compared to alpha diversity, elevational patterns and determinants of beta diversity are much more variable and sometimes even difficult to generalize (McCain and Beck, 2016; Wang et al., 2022). An important reason for the variation in patterns and determinants of elevational beta diversity is that there are various analytical types of beta diversity across spatial gradient. For instance, Anderson et al. (2011) summarized two classes and twelve types of beta diversity, each of which may lead to quite different outcomes and interpretations for the same elevational gradient. Additionally, recent advances in alpha diversity have emphasize the need to incorporate evolutionary information to assess phylogenetic dimension of biodiversity (Mouquet et al., 2012). However, the patterns and determinants of biodiversity of different dimensions may not be congruent along the same spatial gradient (Ding et al., 2021). This is true also for beta diversity. Accordingly, analyses that only use single type or dimension of beta diversity might lead to simplistic or incomplete conclusions regarding origin and mechanism of assemblage dissimilarity as well as biodiversity conservation.

As the world's highest mountains and one of the global biodiversity hotspots, the Himalayas encompass a typical biotic and abiotic elevational gradient and have been the focus of various ecological and conservation studies. In this region, previous studies have found hump-shaped patterns at mid-elevations for bird, mammal, and plant richness, and have proposed various potential determinants for these patterns (Pan et al., 2016; Hu et al., 2017; Liang et al., 2020). However, the patterns and determinants of assemblage dissimilarity along the elevational gradient remain understudied, impeding deeper understanding of maintenance and conservation of biodiversity in this high-profile region. Here, we explored the elevational patterns and determinants of beta diversity for seed plants, based on a detailed field survey in the Gyirong Valley, the longest valley in China's central Himalayas. Specifically, we assessed beta diversity and their turnover and nestedness components of pairwise and multiple analytical types and taxonomic and multiple dimensions simultaneously, and aimed to address following questions: (1) How beta diversity of different types (pairwise and multiple), dimensions (taxonomic and phylogenetic), and components (turnover and nestedness) change along the elevational gradient? (2) What is the major determinant of the elevational pattern of beta diversity of different types, dimensions, and components? (3) What can be inferred from the origin and mechanism of beta diversity in terms of biodiversity conservation?

2. Methods 2.1. Study area and field sampling

The Gyirong Valley of the central Himalayas is located in the southern part of the Tibetan Plateau in China, bordering the northern part of Nepal (28°16′-29°00′ N, 84°56′-85°24′ E, Fig. 1), extending a length of approximately 90 km and spanning an elevation range of 1840–7341 m. Owing to the influence of the Indian Ocean monsoon, the valley exhibits steep environmental gradients and distinct elevational vegetation zones, which can be divided into evergreen broadleaf forests (1800–2500 m), coniferous and broadleaf mixed forests (2500–3300 m), subalpine coniferous forests (3300–3900 m), alpine bush and meadow (3900–4700 m), alpine tundra with sparse herbs (4700–5400 m), and a scree and nival zone above 5400 m.

Our study was conducted along an elevation gradient of the Gyirong Valley, from Resuo village at 1800 m to Mt. Kongtanglamu and Mt. Mala at 5400 m. Elevations lower than 1800 m and higher than 5400 m were excluded from the study because of geopolitical restrictions and the scree and nival zone. We divided the elevation gradient (1800–5400 m) into 12 elevational bands of 300 m. For each band, we established 8 quadrats of 400 m² on the most common vegetation physiognomic types. Consequently, a total of 96 quadrats were established along the elevational gradient (Table S1). In each quadrat, the seed plants were exhaustively inventoried (for 2–4 h by five surveyors) following Fang et al. (2009). Species that could not be identified in the field were taken to the Museum of Beijing Forestry University for identification. A total of 524 seed plants belonging to 85 families and 320 genera were recorded and used in subsequent analyses.

2.2. Phylogenetic reconstruction

The phylogeny of seed plants recorded in our sampling quadrat was based on the phylogeny of Zanne et al. (2014), which represents the current understanding of the relationships between major clades of seed plants and has been widely used in previous studies (e.g., Diaz et al., 2019; Zu et al., 2019). The phylogenetic tree for seed plants in our sampling quadrat was reconstructed using the online bioinformatics tool "Phylomatic" (Fig. S1; Webb and Donoghue, 2005).

2.3. Beta diversity calculations

Beta diversity was calculated and partitioned following Baselga's (2010) methods. This method requires the calculation of three different dissimilarity indices: (1) the Sørensen dissimilarity index (β_sor), which accounts for the total compositional dissimilarity between communities; (2) the Simpson dissimilarity index (β_sim), which measures only compositional dissimilarity due to species turnover; and (3) the difference between β_sor and β_sim (β_sne), which represents the nestedness-resultant dissimilarity.

For each elevational band, we calculated beta diversity indices of two types (pairwise and multiple) and two dimensions (taxonomic and phylogenetic) simultaneously (Box 1). For pairwise beta diversity, the three indices were calculated and averaged for each pair of the eight quadrats within the same elevational band. While for multiple beta diversity, the three indices were calculated for all the eight quadrats within the same elevational band. The beta diversity calculations were conducted using the "beta.pair", "phylo.beta.pair", "beta.multi", and "phylo.beta.multi" functions of the "betapart" package in the R environment (Baselga and Orme, 2012).

Box 1

Calculations of total dissimilarity (bsor) and its turnover (bsim) and nestedness (bsne) components of pairwise and multiple beta diversity.

Additionally, to assess the contribution of turnover and nestedness components to total dissimilarity, we divided β_sne by β_sor to yield the β_ratio. A β_ratio greater than 0.5 indicated that total dissimilarity is largely driven by nestedness; otherwise, it is largely driven by turnover (Peixoto et al., 2017; Pinto-Ledezma et al., 2018). To assess the relationship between taxonomic and phylogenetic dissimilarity along the elevational gradient, we calculated the deviation of phylogenetic beta diversity from taxonomic beta diversity (hereafter β_dev), using the following formula: β_dev = 1−(phylogenetic β_sor/taxonomic β_sor) (Qian et al., 2020). The greater the value of β_dev, the less variation in phylogenetic composition with respect to variation in taxonomic composition between sites, and vice versa (Peixoto et al., 2017; Pinto-Ledezma et al., 2018).

2.4. Environmental variables

Eight environmental variables used to examine the influences of climate, topography, and human disturbance on the elevational patterns of taxonomic and phylogenetic beta diversity (Fig. S2).

Climate is well-acknowledged determinant factor for spatial patterns of both alpha and beta biodiversity (Liao et al., 2022). We used four variables, namely, mean annual temperature, mean annual precipitation, mean annual temperature range, and photosynthetically active radiation to characterize the climatic conditions along the elevational gradient. These variables were obtained from six mini weather stations with three data loggers (HoBo Pro-Temp, HoBo Pro-Precipitation, and HoBo Pro-PAR) established at 2457, 2792, 3368, 3740, 4140, and 5230 m (Fig. 1), from 2016 to 2018. We averaged the three-year data of the variables for each station and extrapolated these data to the entire study area using Kriging interpolation in a GIS environment. For each of the elevational band, we averaged the grid values of the interpolated raster corresponding to the eight sampling quadrats for each of the climate variable.

Topography such as slope and elevation may pose physical barriers to seed dispersal, and thereby was regarded as an important factor for determining biodiversity pattern, especially at regional scale (Guo et al., 2017; Song and Cao, 2017). Here, we used focal mean slope and terrain roughness to assess the influence of topography. For each sampling quadrat, where focal mean slope was calculated as the average of slope within a 300 m neighborhood, and terrain roughness was calculated as the difference between the maximum and minimum elevations within a 300 m neighborhood. These calculations were based on a digital elevation data of the study area with 30-m resolution in a GIS environment. The digital elevation data were derived from the International Scientific & Technical Data Mirror Site, Computer Network Information Center, Chinese Academy of Sciences (www.gscloud.cn). As with the climate variables, we averaged the topography variables of the eight sampling quadrats for each of the elevational band.

Human disturbance may also play an important role in determining community composition (Liang et al., 2019; Sebastián-González et al., 2020). Here, we used human population and area of anthropogenic land use to examine the influence of human disturbance, based on the assumption that the larger the human population and area of anthropogenic land use, the greater the frequency of disturbance events occurring. The human population data were derived from the demographics of villages and towns in the Gyirong Valley, which were provided by the authority of the Mount Qomolangma National Nature Reserve. The area of anthropogenic land use was calculated as the area of artificial surfaces and cultivated land as extracted from GlobeLand30 land cover data (http://www.globallandcover.com). The human population and area of anthropogenic land use were allocated to each elevational band according to their extent.

2.5. Statistical analysis

To explore the elevational patterns of beta diversity, we conducted linear and quadratic models with ordinary least squares regression for each dissimilarity index and the mean elevation. The best model was selected based on the lowest value of the corrected Akaike information criterion for small samples. The ordinary least squares regression was conducted using the "lm" function of the "stats" package in the R environment.

To infer the potential process underlying elevational patterns of beta diversity, we used a null model procedure. This procedure was performed by generating 1000 random species composition matrices using the "randomizeMatrix" function with the independent swap algorithm in the "picante" package for R (Kembel et al., 2012), and then recalculating each phylogenetic diversity index based on the matrix. The standardized effect sizes (SES) were calculated for each dissimilarity index to compare beta diversity between observed and random communities as follow:

where obs was the dissimilarity index of observed communities, mean_random and SD_random was the mean and standard deviation of the dissimilarity index of random communities. SES values significantly deviated from null expectations indicate the dominance of non-random assembly (deterministic processes; Webb et al., 2002). Specifically, SES values less than −1.96 indicates that the dissimilarity index of observed community is significantly smaller than that of random expectations, and implies communities clustering primarily driven by environmental filtering; SES values greater than 1.96 indicates that the dissimilarity index of observed community is significantly larger than that of random expectations, and implies communities overdispersion primarily driven by competitive exclusion (Wang et al., 2022; Ding et al., 2024). As well as the magnitude of departure from random expectation, we were also interested in elevational variations in the SES values, because it can inform us about the elevational pattern of beta diversity after controlling the influences of the alpha diversity and gamma diversity. For this, we performed ordinary least squares regression to fit SES values with the elevation.

To assess the relative importance of different environmental variables in determining elevational patterns of beta diversity, we performed the random forests models. This approach was chosen because it does not require strict assumptions in the data and can better handle multicollinearity and nonlinear relationships (Feng et al., 2017). For each dissimilarity index, we run the random forests models 1000 times. The relative importance of each environmental variable was assessed based on the percentage increase in mean squared error (%IncMSE), which quantifies the increase in mean squared error that occurs when a particular variable is randomly permuted. The random forests models were conducted using the "randomForest" function in the "randomForest" package in the R environment (Breiman, 2001).

3. Results 3.1. Elevational patterns of beta diversity

For both pairwise and multiple beta diversity, taxonomic and phylogenetic β_sor and β_sim were best described by second-order regression models, showing left-skewed hump-shaped pattern with peaks at 3600 m around, whereas taxonomic and phylogenetic β_sne were best described by first-order regression models, showing a general increasing pattern (Fig. 2, Table S2). Both pairwise and multiple phylogenetic beta diversity were remarkably low than their taxonomic counterpart, with β_dev less than 0.3 across all elevational bands (Fig. S3). All β_ratio, both pairwise and multiple, taxonomic and phylogenetic, were less than 0.5, indicating that beta diversity of different types and dimensions were largely driven by turnover component (Fig. S4).

3.2. The SES of beta diversity

The SES of beta diversity showed patterns similar to those of observed, suggesting that alpha diversity and gamma diversity have little influence on beta diversity. For both pairwise and multiple beta diversity, the SES of taxonomic and phylogenetic β_sor and β_sim showed hump-shaped patterns along the elevational gradient, and were all significantly less than random expectation (smaller than −1.96, Fig. 3a, b, d, and e). The SES of taxonomic β_sne showed general increasing tendency with elevation, whereas SES of phylogenetic β_sne displayed a U-shaped relationship with elevation, and they were all significantly greater than random expectation (larger than 1.96, Fig. 3c and f).

3.3. Relative importance of each environmental variables

The relative importances of environmental variables were generally similar for beta diversity of different types and dimensions (Fig. 4). Specifically, all climate-related variables, that is, mean annual temperature, mean annual precipitation, mean annual temperature range, and photosynthetically active radiation, showed higher relative importance than other variables in explaining β_sor and β_sim, but lower relative importance in explaining β_sne. By contrary, focal mean slope and terrain roughness have higher relative importance for β_sne following by human population and area of anthropogenic land use.

4. Discussion 4.1. How assemblage composition varies along elevational gradient

When exploring beta diversity along spatial gradients, clarity is needed regarding the conceptual type of beta diversity of interest, because beta diversity of different types may lead to different patterns and interpretations (Anderson et al., 2011). Current studies on beta diversity along the elevational gradient mostly focus on assessing variation in community composition among site with different elevational distance, and found a classic patterns of distance-decay in composition similarity. Comparatively, how variation in community composition itself is changing along the elevational gradient were remain elusive. In the Gyirong Valley, we used both pairwise and multiple beta diversity to assess how variation in community composition itself is changing along the elevational gradient. These two types of beta diversity show consistently left-skewed hump-shaped elevational pattern for taxonomic and phylogenetic total dissimilarity and its turnover component (Fig. 2). A similar pattern was observed in previous studies on the species richness of vascular plants along the same elevational gradient (Liang et al., 2020, 2023). These results provide evidences for a long-standing assumption in studies of community ecology, that is, regions of high turnover are also regions of high species richness. This assumption is generally attributed biotic exchange between different ecozones (Lomolino, 2001). The current mid-elevations of the Himalayas were once lower elevations or even near sea level before the collision of the Indian plate and the Asian plate. After the collision, Gondwanan plants carried by the Indian plate mixed with Laurasian plants that evolved on the Asian plate in the newly uplifted regions (Raven and Axelrod, 1974). It is probably that such historically biotic exchange has left a potential legacy for the higher species richness and species turnover at the mid-elevation.

Turnover was the dominant component of beta diversity, regardless of its types and dimensions, indicating that assemblage dissimilarity of seed plants originated from species replacement rather than species loss or gain. Interestingly, the contribution of turnover component tended to decrease with elevation (Fig. S4). This result corresponds to those of previous studies which found that turnover tends to be higher in regions at low latitudes than at high latitudes (Baselga, 2010; Dobrovolski et al., 2012). Qian et al. (2020) argued that extinction rate is expected to be higher and species ranges tend to be larger in areas with lower mean temperatures and greater temperature variability (i.e., Rapoport's rule, Stevens, 1989), which would lead to decreasing turnover but increasing nestedness of species among sites. Accordingly, our previous study has showed that the mean range size of vascular plant assemblages increases along the elevational gradient in the Gyirong Valley owing to worsening climate conditions and increasing climate variability (Liang et al., 2021), which could also account for the decreasing trend of turnover.

Moreover, for both pairwise and multiple beta diversity, taxonomic total dissimilarity and its turnover component were consistently higher than their phylogenetic counterparts, whereas taxonomic nestedness was lower than their phylogenetic counterpart (Fig. 2). This implies that species replacement occurred more frequently among closely related species, and inversely, species gain or loss occurred more frequently among phylogenetically distant species (Wang et al., 2022). Nevertheless, the former would be the most common, since turnover constitutes most of the total dissimilarity. Moreover, we found that the deviation of phylogenetic beta diversity from taxonomic beta diversity tended to increase with elevation (Fig. S3). The results suggest that, with increasing elevation, species replacements increasingly take place within certain lineages due to the harsh environment, leading to a tendency of assemblage to consist of closely related species (phylogenetic clustering, Jin et al., 2015).

4.2. Why assemblage composition varies along elevational gradient

The null model analysis showed that the observed patterns of beta diversity, regardless of its type and dimension, were significantly different from random expectations (Fig. 3). This result provides evidence for the niche-assembly theory, and indicates deterministic processes underlying the assemblage dissimilarity of seed plants along the elevational gradient (Bishop et al., 2015; Ding et al., 2024). Specifically, we found that β_sor and β_sim of different types and dimensions, particularly at lower and higher elevations, were significantly smaller than random expectations (Fig. 3). Moreover, β_sor and β_sim of different types and dimensions showed stronger associations with climate-related variables than with other variables (Fig. 4). Altogether, these results imply that the elevational patterns of seed plant beta diversity and more particularly their turnover components were determined by climate-driven environmental filtering.

Climate has long been suggested to be a strong environmental filter, especially for plants (Zellweger et al., 2017). In mountains, climatic conditions may selectively filter which species can successfully establish and survive at each elevation, leading to species replacements along the elevational gradient in accordance with their physiological tolerance (Bishop et al., 2015). For example, extreme climate such as lower temperature and precipitation at the higher elevations favor species with stress resistance traits (Qian et al., 2019). Most of these species are predominantly concentrated in certain clades such as but not limited to Rosaceae, Leguminosae, and Asteraceae, which is consistent with the phylogenetic clustering of assemblages observed toward higher elevations (Fig. S3). Comparatively, mechanism of climate-driven environmental filtering is more complex at lower elevations, which may be associated with factors such as water or light availability (Pausas and Austin, 2011). For instance, in a study on species assembly of ferns along a tropical elevational gradient, Kluge and Kessler (2011) found that ferns at lower elevations were selected for specific traits that enable the resistance to drought stress. In the Gyirong Valley, higher temperature at lower elevations will lead to higher evaporation rates as shown by previous study (Liang et al., 2020). This might accelerate plant water loss and reduce soil water supply, and thereby make availability of water one of environmental filters for species lack of appropriate traits. However, we need further empirical analysis to support this assumption.

Moreover, our results did not indicate that other factors such as topography and disturbance had no effect on elevational variation in the assemblage composition of seed plants. Conversely, topography and disturbance play a more important role in determining the elevational patterns of β_sne. This is consistent with conventional cognition that compared to species replacement, species gain and loss might stem from other ecological factors such as physical barriers or human disturbance (Legendre, 2014). In addition to environmental variables, biotic interactions like species competition also have influence on elevational patterns of beta diversity especially for their nestedness component, as indicated by higher observed β_sne than random expectations (Fig. 3). Besides, biotic interactions might also be related to the higher SES of β_sor and β_sim observed at mid-elevations, given that biotic interactions are likely more intense in these areas due to higher specie richness (Liang et al., 2020). However, more studies are required if we aim to evaluate the roles of biotic interactions in community assembly along elevational gradient of Gyirong Valley.

4.3. Conservation implications

Owing to its unparalleled elevational range, the Himalayas have a more conspicuous and more complete vertical climatic gradient than most other mountains. Previous studies have demonstrated that climate is the major driver of the hump-shaped pattern of plant richness in the Himalayas (e.g., Manish et al., 2017; Sun et al., 2020; Liang et al., 2020). However, to the best our knowledge, this study is the first to reveal that climate also shapes the hump-shaped pattern of assemblage dissimilarity for seed plants in the Himalayas. The mid-elevations of the Himalayas have long been recommended as the focus of conservation efforts due to their significant role in harboring greater species richness for mammals, birds, and plants (Pan et al., 2016; Hu et al., 2017; Liang et al., 2020). However, as climate-driven turnover was found to be the main process of compositional variation of seed plant assemblages along the elevational gradient, our study implies that different elevations of the Himalayas may harbor different species adapting to local climatic environment. Therefore, in order to maximalize biodiversity conservation, conservation efforts should better cover elevations with different climate types rather than only focus on elevations with highest species richness (Franklin, 1993; Nascimbene and Spitale, 2017). For example, our study implies that compared to other elevations, lower elevations harbor more phylogenetically heterogenous assemblages and therefore are also worthy of conservation attention.

Moreover, the dominant role of climate in determining biodiversity patterns implies that species in this region might be more susceptible to climate change, which deserves close attention in future conservation efforts. Available evidence suggests that even slight changes in climate may cause the range retractions of mountain species, particularly plants, which are comparatively difficult to track shifting climate envelopes due to their lack of mobility (Lenoir and Svenning, 2015). For instance, Engler et al. (2011) assessed the impacts of climate change on 2632 plant species across all major European mountain ranges and found that 36–55% of alpine species, 31–51% of subalpine species, and 19–46% of montane species will lose more than 80% of their suitable habitats by 2070–2100. If such range retraction occurs in narrow-range specialists-species, it may elevate their extinction rate and consequently leading to increased compositionally homogenization and decreased beta diversity of species assemblages. Conversely, if range retraction happens on ubiquitous generalists-species, it may reduce their occurrence and consequently result in increased compositionally heterogenization and beta diversity of species assemblages (Tatsumi et al., 2020, 2021). Therefore, future studies should disentangle all possible processes behind the dynamic of beta diversity, in order to yield conservation tailored to each circumstance with ongoing climate change (Tatsumi et al., 2022).

4.4. Caveats and limitations

In this study, we built null models and used SES to infer about the type of ecological process that responsible for the assemblage dissimilarity of seed plants along the elevational gradient. However, SES values could be influenced by the statistical performance of null model. This is because when null models are restricted, there is a possibility that the boundary conditions themselves unwittingly introduce constraints on the null distribution of the pattern of interest, leading to the Narcissus effect (Colwell and Winkler, 1984). As a consequence, this could cause an increase in the similarity between the observed species distribution and the null-modelled distribution, which in turn result in artificially high rejection rates for the focal pattern (Type Ⅱ errors) (Ulrich et al., 2017). Given that, our finding that SES of the observed beta diversity deviation varies along elevational gradient might be biased by the Narcissus effect, and it is necessary to remove the effects by acquiring external data about the sizes of the relevant species pools and about the variations in the relevant environmental or ecological factors among the sites (Ulrich et al., 2017).

5. Conclusion

In this study, we explore beta diversity of different types (pairwise and multiple) and dimensions (taxonomic and phylogenetic) simultaneously for seed plant assemblages along the elevational gradient in the Gyirong Valley of the central Himalayas. We found that total dissimilarity, regardless of its type and dimension, were consistently driven by turnover, and both total dissimilarity and turnover shown a hump-shaped elevational patterns. These patterns were significantly less than random expectation and were mostly associated with climate variables, providing evidence for the niche-assembly theory. In summary, our results suggested that assemblage dissimilarity of seed plants was mostly originate from species replacement determined by climate-driven environmental filtering. Moreover, these results imply that conservation efforts should better cover elevations with different climate types rather than only focus on elevations with highest species richness, and call for close attention to the impact of climate change on the dynamic of beta diversity. Our study highlighted that comparisons of beta diversity of different types, dimensions, and components simultaneously could be conductive to consensus on the origin and mechanism of assemblage dissimilarity.

Acknowledgments

This work was supported by the National Natural Science Foundation of China (grant number 31901109) and Guangdong Basic and Applied Basic Research Foundation (grant number 2021A1515110744).

Data accessibility statement

The raw data for this work are available at the Dryad Digital Repository (https://datadryad.org/stash/share/rxetR6pY-IGGY4XT8cj24fnLcS70XozCXVjLsSQ5Qxc).

CRediT authorship contribution statement

Jianchao Liang: Writing – original draft, Software, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Zhifeng Ding: Writing – review & editing, Methodology, Formal analysis, Data curation. Ganwen Lie: Writing – review & editing. Zhixin Zhou: Writing – review & editing. Zhixiang Zhang: Writing – review & editing, Supervision, Resources, Funding acquisition, Conceptualization. Huijian Hu: Writing – review & editing, Supervision, Resources, Funding acquisition, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.pld.2024.07.011.