Biological invasions are considered one of the most important global problems experienced by ecosystems (Simberloff et al., 2013). Phylogenetic relatedness of the alien species to the existing community members is hypothesized to facilitate [(pre-adaptation hypothesis; (Ricciardi and Mottiar, 2006)] or hinder [(Darwin’s naturalization hypothesis; (Daehler, 2001)] successful establishment of alien species in the community. The alternative views of these two hypotheses (also called Darwin’s naturalization conundrum) are often resolved by considering the gradients of community assembly (Thuiller et al., 2010). Biotic interactions in a similar environment at a small spatial scale and environmental heterogeneity at a large scale influence the phylogenetic relationship between the community members (Procheş et al., 2008). Numerous studies have investigated the phylogenetic relationship between native and alien species to provide evidence supporting these two hypotheses [e.g., (Procheş et al., 2015; Divisek et al., 2018; Omer et al., 2022)].

However, the introduction of an alien species may also occur in a community with existing alien species belonging to different stages of the invasion process, i.e., introduced, naturalized, and invasive. Biotic interactions (e.g., mutualism) between the alien species at a small spatial scale and environmental factors (e.g., changes in habitat conditions) at a large spatial scale may influence the community assembly process (Simberloff and Von Holle, 1999). The majority of studies on the transition of alien species along the invasion process have been conducted at large spatial scales, often at the country level considering whole flora (e.g., Australia (Gallagher et al., 2015), or for a family [e.g., Proteaceae in South Africa (Moodley et al., 2013)]. These studies either considered phylogenetic uniformity between species of different invasion stages (Gallagher et al., 2015) or tested phylogenetic signals in the factors that could influence the transition (Maurel et al., 2016). Therefore, knowledge of the phylogenetic relationship between the alien species of different invasion stages is rare. Understanding the environmental influence on large-scale phylogenetic relationships across the invasion process may provide important insights into the community assembly process.

This study addressed this knowledge gap by characterizing phylogenetic relationships between alien community members at different spatial and taxonomic scales. Ecological traits are often phylogenetically conserved (Donoghue, 2008), and therefore, species in harsher environments tend to be more closely related (Wiens and Donoghue, 2004). At a large spatial scale, environmental filtering (and not biotic interactions) influences phylogenetic relatedness by sorting phylogenetically closely related alien species in a community (Qian and Sandel, 2017). Therefore, we first hypothesized that the phylogenetic relatedness of alien species should vary across the spatial scale, and the environmental influence will be more pronounced at a large spatial scale, where phylogenetically closely related species will form a more clustered assemblage.

Secondly, among the three stages of the invasion process, the invasive species have been found to possess certain ecophysiological traits (e.g., specific leaf area, height) that aid them in adapting to a range of environmental conditions (Gallagher et al., 2015). Moreover, few studies that have compared functional traits based on phylogenetic history between the native and alien species revealed that the successful invaders need to be functionally different from the native species to occupy novel trait space (Divisek et al., 2018). Adaptations in key traits of the invasive species could arise from standing genetic variation and/or rapid evolution of phenotypically plastic traits during range expansion, particularly in response to ecogeographical variations (Prentis et al., 2008; Hodgins et al., 2018). Although such rapid adaptations are often attributed to their ability to occupy novel niches in the introduced range (Pyron et al., 2015), the fundamental niches of the invasive lineages have been conserved (Liu et al., 2020). Therefore, from an evolutionary perspective, we hypothesized that niche conservatism might increase the phylogenetic clustering of the invasive species compared to that of the aliens that failed to naturalize and naturalized species that failed to invade when environmental filters sort phylogenetically closely related alien species in a community.

Thirdly, the observed phylogenetic pattern can vary depending on the taxonomic scope (whole flora or family-specific) and phylogenetic framework (based on phylogenetic trees or a non-phylogenetic approach) (Qian and Sandel, 2017). As species of the same family are more likely to have similar responses to environmental gradients, we expected that the phylogenetic pattern should be more pronounced at smaller phylogenetic extents. Because individual families have different evolutionary histories, the phylogenetic patterns should also vary among them.

We tested these three hypotheses for the alien plant species present in China. The history of alien plant introduction into China goes back more than 2000 years, and many aliens have become naturalized and subsequently invasive (Axmacher and Sang, 2013). With increased globalization and the gradual opening of China’s economy to the global markets in the 21st century, the novel introduction of alien species is expected to increase. Deciphering the phylogenetic relationship pattern across the invasion continuum in China has received recent attention. Previous studies have provided important insights into the phylogenetic relationships between native and alien species (Qian et al., 2022a), native and naturalized alien species (Qian et al., 2022b), and naturalized and invasive alien species (Qian et al., 2022c). However, phylogenetic relationships between alien species of different invasion stages and the influence of these relationships on the community assembly process at multiple spatial and taxonomic scales are rarely investigated in a single analysis [but see (Zhang et al., 2021)].

Specifically, the objectives of this study were to determine: 1) the variation of phylogenetic relatedness of alien species across spatial scales (hypothesis 1), 2) whether the three groups of alien species (introduced, naturalized, and invasive) show similar phylogenetic patterns (hypothesis 2), and 3) how the phylogenetic relationships change depending on the taxonomic scale (whole flora and family-specific) of investigation (hypothesis 3). We addressed these objectives by relating phylogenetic patterns of alien species across spatial scale, invasion process, and taxonomic scope with climate as an environmental filter.

The 706 species (introduced = 165, naturalized = 222, invasive = 319) considered in this study belonged to the 397 genera and 94 families. The phylogenetic tree was highly resolved, with the species distributed among major angiosperm clades (Figure 1). Three families having the maximum number of species were Asteraceae (16.15%, n=114: introduced = 17, naturalized = 32, invasive = 65), Fabaceae (12.75%, n=90: introduced = 28, naturalized = 27, invasive = 35), and Poaceae (10.62%, n=75: introduced = 23, naturalized = 20, invasive = 32).

In this study, we tested three hypotheses related to the variation in phylogenetic relatedness between alien species of three invasion categories at four spatial and two taxonomic scales. The findings of our study supported these hypotheses. We observed more clustered assemblages at finer spatial scales for all three alien species categories at species and family levels. Significantly higher NRI and NTI were observed at provincial (PROs) and ecoregion (ECOs) scales, respectively. These findings supported our first hypothesis, based on the concept that closely related species can respond more similarly to the environmental variation that is more noticeable at a finer spatial scale (Procheş et al., 2008).

The second hypothesis, i.e., the phylogenetic measures will vary between the alien species of three categories, with the invasive aliens showing the most clustered assemblage, was supported by multiple findings of this study. Firstly, the presence of introduced alien species in the community caused a clustered assemblage when the NRI values were considered (positive NRI) and an overdispersed assemblage when considering the NTI values (negative NTI). In contrast, positive NTI values were observed for the invasive and naturalized alien species. These findings indicate that although the existing lineages of the introduced species may lead to small mean pairwise distances, they do not always have a close relative present in the community. The positive NTI values for the invasive and naturalized species, on the other hand, resonates with the pre-adaptation hypothesis, which posits that a close relative in the community may help the successful naturalization and invasion of the introduced alien species (Ricciardi and Mottiar, 2006). Secondly, the communities with invasive species had significantly higher NRI and NTI values across the spatial scales, thus being concordant with the previous studies (Zhang et al., 2021; Qian et al., 2022b). Including invasive species created a more clustered community, as evident from the negative NRI and NTI anomalies. Besides, the NRI values were not significantly different between the naturalized and introduced species. This phylogenetic pattern is indicative that invasive species occupy a different phylogenetic niche than the naturalized and introduced species, as also evident from the two-dimensional phylogenetic space. This finding is consistent with a previous observation that established alien species need to be phylogenetically different to become invasive (Divisek et al., 2018). The communities with naturalized and introduced species were found to have high NTI and negative NTI values, respectively, across all spatial scales. Therefore, for successful establishment and naturalization, the introduced species should have an immediate relative present in the community; however, the presence of invasive aliens could facilitate the establishment of other invasive species due to their similar responses to the environment. This finding provides the phylogenetic basis of the invasional meltdown hypothesis (IMH), which posits that positive interactions among invaders may promote secondary invasions (Green et al., 2011).

We found few significant correlations between the phylogenetic relatedness metrics of the three species categories. The regression analysis also revealed the influence of community types and bioclimatic variables on the NRI and NTI values, more prominently at the ecoregion (ECOs) scale. Given that environmental filtering often drives phylogenetic clustering (Webb et al., 2002), these findings indicate that different environmental factors may influence the alien species at different stages of the invasion continuum. In general, the phylogenetic relatedness of the alien species in China was more influenced by annual mean temperature (BIO1) than the annual precipitation (BIO12). With the decrease in temperature, the NTI values increased, as evident from the negative relationship between NTI and BIO1. This finding implies that with increased environmental stress (e.g., low temperature), phylogenetic clustering increased between species sharing similar environmental space (high NTI). A similar influence of cold stress on phylogenetic clustering has been found in other studies on plants (Li et al., 2014), animals (Pellissier et al., 2013), and microbes (Wang et al., 2012). When we analyzed the influence of the environment on the phylogenetic measures for three alien species categories separately, the NRI and NTI values of the communities with the invasive aliens were found to be more influenced by temperature (BIO1). At the other end of the continuum, precipitation (BIO12) influenced the introduced aliens’ phylogenetic measures more. Moreover, compared to the introduced species, more clustered NTIs were observed for the invasive species in response to decreasing temperature when they occurred alone and combined with the naturalized species. This phylogenetic pattern suggests a more specialized response of the invasive aliens to the environmental filters than the naturalized and introduced species, possibly due to a higher level of trait plasticity, which provides the invasives a greater tolerance to different environmental conditions (Gallagher et al., 2015). It is also possible that the closely related invasive species can retain their niche-related traits through evolutionary history and, therefore, can respond similarly in the environmental space. This finding is thus consistent with the phylogenetic niche conservatism hypothesis (Crisp and Cook, 2012), which has been reported previously for both plant (Liu et al., 2015) and animal species (Cooper et al., 2011).

The phylogenetic pattern varied at the family level from that observed at the species level, thereby supporting our third hypothesis. Both the phylogenetic measures were positive for three alien species categories. With the decrease in the phylogenetic resolution, the species of the same family respond to the environmental gradients more similarly, which can explain the clustered assemblage within families. Indeed, compared to the species level, we observed more significant correlations between the NRI values of the three alien species categories at the family level. In each of the three families, the correlation analysis revealed that the NRI values of the three species categories were positively correlated mainly at finer environmental scales (ECOs). This pattern can be explained based on the similar response of all possible species pairs of the same family to the environmental filtering process (Webb et al., 2002). The shared response to environmental filters within families is also supported by the observation that the bioclimatic variables more influenced NRI values; the pattern in contrast to that observed at the species level where the environmental variables more influenced variation in NTI values. These findings provide evidence of the varied environmental influence at different levels of the phylogenetic tree (Qian et al., 2022a). Besides, the NRI values were higher than the NTIs for all three families, creating a more clustered NRI and less clustered NTI. These findings support the observation that NRI tends to be driven by structure deep in the tree, while patterns near the tip more strongly influence NTI (Mazel et al., 2016). Overall, the phylogenetic patterns observed at species and family levels in this study are consistent with a previous meta-analysis of phylogenetic relatedness patterns (Vamosi et al., 2009).

However, the phylogenetic pattern was not always consistent across families. The invasive species of Asteraceae, either occurring alone or in combination with the other two alien species categories, showed a significantly stronger clustered community assemblage. This pattern became weaker and dependent on the spatial scale in Fabaceae, where significantly higher NRI and NTI values were only observed when the invasives cooccurred with the naturalized and introduced species. We observed no significant differences in the phylogenetic measures in Poaceae between the three species categories. The variations in phylogenetic patterns between families can be explained based on the families’ different evolutionary histories (Harris and Davies, 2016) and the species’ different life history traits, which are correlated with the rate of molecular evolution (Smith and Donoghue, 2008). For example, previous studies have found that plant growth form has an inherent phylogenetic signal (Kerkhoff et al., 2006). On the one hand, the majority of the Asteraceae species included in this study were herbs (113 out of 114 species; see Supplementary Data 1), which might have undergone a higher rate of molecular evolution than shrubs and trees (Smith and Donoghue, 2008), thereby creating the observed phylogenetic signal. On the other hand, the three alien species groups of the Poaceae family might have evolved enough times independently to obscure any clear signal of phylogenetic clustering. Besides, the phylogenetic structure is heavily influenced by polyploidy (Rothfels, 2021), which occurs more often in Poaceae than in any other angiosperm family. Nearly 80% of grass species are supposed to have undergone a ploidy change at some point in their evolution (Huang et al., 2022), thereby obscuring the phylogenetic signal. The variations in phylogenetic patterns between Asteraceae and Poaceae can also be generated due to the uneven number of alien species in three invasion categories. Compared to the higher number of invasive aliens in Asteraceae (65 out of 114 amounting to 57.02%), the relatively balanced proportions of alien species of three invasion categories in Poaceae (32 invasive aliens, 20 naturalized aliens, and 23 introduced aliens) may also cause the absence of clear phylogenetic signal.

Similarly, the influence of the environmental filtering process on phylogenetic relatedness was not consistent across the three families considered in this study. The effect of cold stress leading to increased phylogenetic clustering (higher NTI values with a decrease in temperature), as observed at the species level and for Fabaceae and Poaceae families, was not observed for Asteraceae invasive and naturalized species. In addition, the correlation analysis revealed a significant positive correlation of both phylogenetic measures between invasive and naturalized aliens of Asteraceae across all spatial scales. These findings indicate that the invasives and naturalized species of the family Asteraceae might have been subjected to the same environmental filtering pressure. This phylogeographic pattern was not observed in Fabaceae, where, for example, the NRI and NTI values of the invasive and naturalized species showed positive and negative relationships, respectively, with annual precipitation. The different growth forms of the Fabaceae aliens (47 herbs, 29 shrubs, and 14 trees; see Supplementary Data 1), compared to the herbaceous growth form of the majority of the Asteraceae species, could explain the varied response of different categories of alien species to the environmental filtering process. Such details of the phylogenetic relationship are obscured with increased phylogenetic extent (e.g., at the tip of the tree) and spatial scale of the investigation by combining species with different response abilities (e.g., Asteraceae and Fabaceae) to environmental filtering processes (Weigelt et al., 2015).

In conclusion, our study showed that the phylogenetic relatedness of the alien species varies across the invasion continuum. The pattern depends greatly on the environment and spatial scale of investigation. The invasive species community showed the most clustered assemblage, irrespective of the spatial scales. The influence of the environmental filtering process also varied between families, more so at finer spatial scales, although there was a consistent tendency for more clustered NRI for all three species groups. Overall, our study highlights the importance of assessing hierarchical phylogenetic patterns of alien plant species of different invasion stages across spatial and taxonomic scales. We have also identified specific scopes of improvements that can further advance our understanding of the phylogenetic changes across the invasion process. Firstly, the uneven number of species across different communities and taxonomic scales might influence the comparative assessment of the phylogenetic measures. However, using abundance-weighted phylogenetic measures addressed this problem to some extent. Secondly, the occurrence data used in this study to estimate the phylogenetic metrics were collected from GBIF and NSII, which often cannot adequately represent the actual distribution of the species in China (Qian et al., 2018). In addition, the availability of new information on the alien flora of China [e.g., (Lin et al., 2022)] can change the current findings. Thirdly, we selected the families with the maximum number of alien species. The phylogenetic pattern at the taxonomic scale may also vary depending on the choice of families due to their different evolutionary histories. Future studies can gain novel insights into the phylogenetic pattern by including families having the maximum number of invasive and/or naturalized and/or introduced species (e.g., families of genera, like Amaranthus, Ipomoea, and Solanum, with many invasive species from this study). Finally, our study also indicated that plant functional traits could influence phylogenetic patterns in response to the environmental filtering process; therefore, relating phylogenetic patterns with functional traits of alien plant species in future studies can provide novel insights into the invasion process. As the phylogenetic structure is often determined by interacting species present in each community, we suggest future phylogenetic analysis should consider the combination of native and alien species at multiple spatial and taxonomic scales to understand better the dynamics of recipient community changes across the invasion continuum.

The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding author.

AB and YH conceived and planned the project. HF, XL, YL, JW, MY, and HP collected the data. AB, FT, HF, and NZ analyzed the data. AB, FT, NZ, and YH contributed to the interpretation of the results. AB wrote the first draft of the manuscript. All authors contributed to the article and approved the submitted version.