Exotic plant species with longer seed bank longevity and lower seed dry mass are more likely to be invasive in China

Abstract

Globalization of social and economic activities has led to the large-scale redistribution of plant species. It is still unclear how the traits aid the successful invasion of alien species. Here, we downloaded global plant trait data from the TRY-Plant Trait Database and classified alien species in China into four groups: high, medium, need attention and harmless according to their distribution and degree of harm to local plant communities based on existed studies. The relationship between plant functional traits and invasion level was clarified, and we established a prediction model based on plant functional traits and taxonomy. The results showed that species with smaller seeds, smaller individuals, lower special leaf area and longer seed bank longevity (SL) are more likely to be an invasive species after introduction to foreign ecosystems. In summary, exotic species with longer SL and lower seed dry mass are more likely to be invasive in China. We also trained two predictive models to check if we can predict a species’ invasion. Combining the two models together, statistically, we could predict if a species is invasive from its traits and taxonomy with a 91.84% accuracy. This model could help local governments, managers and stakeholders to evaluate shall we introduce some plant species in China.

摘要

种子库寿命较长且种子干质量较低的外来植物物种更有可能在中国成为入侵物种

社会和经济活动的全球化导致植物物种的大规模重新分布。目前尚不清楚植物特征如何促成外来物种的成功入侵。在本研究中，我们从TRY数据库下载了全球植物特征数据，根据现有研究和外来物种对当地植物群落的危害程度，将中国的外来物种分为4组：高、中、需关注和无害。本研究厘清了植物功能性状与入侵程度之间的关系，并基于植物功能性状和分类建立了预测模型。研究发现，种子较小、个体较小、叶面积较小且种子库寿命较长的物种更有可能在引入到外来生态系统后成为入侵物种，具有较长种子库寿命和较低种子干质量的外来物种更有可能在中国成为入侵物种。本研究还训练了两个预测模型，以检查是否能够预测物种的入侵情况。模型可以通过物种的特征和分类来预测其是否具有入侵性，准确率达到91.84%，同时可以帮助地方政府、管理者和利益相关者评估是否应该将某些植物物种引入到中国。

INTRODUCTION

The introduction of exotic species is controversial in the field of ecology and social development. On the one hand, socioeconomic dependencies maintain the presence of exotic species such as plantation forestry and species with food, ornamental and scientific value (Kull et al. 2017). And the socio-ecological keystone species also enhance local biodiversity and are symbols of human progress (Wang et al. 2009). Furthermore, globalized socioeconomic activities cause the large-scale redistribution of plant species, and the question of whether exotic species are invasive and thus a threat to local ecosystems is a concern for ecologists (Pimentel et al. 2000). However, not all species transported outside their native range establish viable invasive populations, which is why uncertainty motivates ecologists to continually search for what differentiates successful invasive exotic populations from all exotic populations.

Several hypotheses were proposed such as ‘Biodiversity resistance hypothesis’ (Elton 1958), ‘Resource opportunity hypothesis’ (Elton 1958), ‘Ideal weeds characteristics’ or ‘Inherent superiority hypothesis’ (Baker 1965), ‘Enemies release hypothesis’ (Shaw and Hatcher 2017), ‘Evolution of increased competitive ability hypothesis’ (Notzold 1995), ‘New weapon’ (Callaway and Aschehoug 2000; Williamson and Fitter 1996), ‘Propagule pressure’ (Matesanz and Sultan 2013; Williamson and Fitter 1996) and ‘Niche opportunity hypothesis’ (Shea and Chesson 2002) to describe the mechanism of alien species invasive success. Ecologists have proposed a methodological framework for invasive plant studies that integrates population characteristics based on the different stages of the invasion process (Blackburn et al. 2011). Despite the debate about each hypothesis, the kernel theory of them is successful invaders have higher competitive and fitness than native species, and invaders gain a competitive advantage due to the chemosensory effect. Interference-like allelopathy increases the success rate of intrusion if interactions between exotic and native species are structured spatially (Allstadt et al. 2012).

Invasive species also benefit from evolutionary history and habitat differences after invasion, undergoing more genetic variation and resulting in ecotypes that are better adapted to environmental conditions, use more resources and have greater resistance or traits to external environmental stresses than the native plant community (Penk et al. 2017). Some of them have unique biological traits or inherent superiority in themselves (e.g. morphology, physiological, behavior and genotype) (Kudo et al. 2011), thus eventually gaining an advantage in competition and successfully invading. The resource-enemy release hypothesis (R-ERH) proposes that resources and release from natural enemies together determine the success rate of invasion (Blumenthal 2005). They invade native plant communities with ‘new weapons’ and thus become more competitive. In recent years, the deliberate introduction of international trade or other soil activities has become a new mechanism of plant invasion (Gagnon et al. 2007; Kueffer 2017).

A functional plant trait is any trait that affects the survival and reproduction of a plant (He et al. 2020) as well as its adaptation to its environment (He et al. 2018; Rejmánek 1996), and it helps to predict the distribution of a species as well as its adaptation to environmental changes (Laughlin et al. 2020; Pollock et al. 2012). From an evolutionary viewpoint, plant functional traits are measurable properties that have been exhibited overtime during their evolution and development and are usually closely related to the optimization of plant productivity or the adaptation of plants to their environment (Sobral 2021). As research has progressed, it has become clear that plants perform their adaptation or invasion success through a combination of functional traits, or that any single trait is achieved through the synergy of multiple functional traits (Küster et al. 2008; Westoby and Wright 2006). Accurately quantifying the trade-offs and dependencies between these multiple traits can help us to better understand the habitat adaptation strategies of plants (Legras et al. 2020).

At the community level, plant traits affect survival and reproduction within and among species. It might seem critical to identify the expected effects of traits on the invasion level of exotic species (Liao et al. 2007). Previous studies also showed that plant size affects invasive capacity by affecting fecundity, intra- and interspecific competition. Taller plants have more sexual reproductive resources available (Liu and Pennings 2019) and are better competitors for natural resources such as light (Keddy 2001). Spartina alterniflora has undergone rapid evolution since its introduction to China, and it exhibits higher fruiting rates than indigenous species. Objective analysis reveals this species’ rapid adaptation and reproductive success in the Chinese environment compared with native species (Liu et al. 2020). The propagule pressure of invasive plants, including the number of seeds, seed germination efficiency (SGE) and seed bank longevity (SL), is a key determinant of invasion success (Cassey et al. 2018; Ning et al. 2021, 2023). The smaller the seeds and the faster the seed germination, the more invasive they are, e.g. Gramineae, Asteraceae, Amaranthaceae and some Pinus spp. Pyšek et al. (2012) conducted a literature review of 287 publications to assess the impact of 167 plant species on resident species, communities and ecosystems and the results showed certain species traits, such as life form, stature and pollination syndrome, may be used to predict impact.

To enhance our capability to prevent and control biological invasions, it is vital to take into account the appropriate habitat of the local community and comprehend the connection between invasiveness and plant traits (Küster et al. 2008; Pyšek and Richardson 2007). According to previous studies and public data, China has 585 exotic plant species, and based on their spreading scale as well as harmfulness to the local community, they were classified into different levels as high, middle, need attention and harmless (details about the classification method can be found in Supplementary Method). Comparative analysis represents one of the key techniques employed in the study of invasiveness. This approach relies on the examination of phylogenetic or functional similarities, taking into account biological traits such as morphology, reproductive capacity, living habits and environmental conditions such as light, humidity and temperature. By scrutinizing these factors in invasive species and local communities, it is possible to predict the potential for invaders and identify invasion sites (Kleunen et al. 2010). Studies of macroecology and plant functional traits provide ideas for predicting the invasiveness of exotic species in a large-scale pattern and making predictions using models (He et al. 2018); and in-depth study on the relationship between plant functional traits and invasiveness is important for early prediction, prevention and control of invasive species expansion, as well as restoration and management of infested ecosystems.

Botanists and ecologists working on invasive ecology in China have accumulated decades of research and the results at the national scale are included in the Plant Science Data Center (PSDC) and Invasive Alien Species of China (IASC). According to existed studies and public data, China has 585 exotic plant species and based on their spreading scale as well as their harmfulness to the local community, they were classified into different levels such as high, middle, need attention and harmless. With the resurgence of globalization and international trade in the post-epidemic era, the introduction of species for ecological restoration, landscaping and other purposes will also become more common in international cooperation. Spartina alterniflora, as an example, which was originally introduced to China from the east coast of America to promote sedimentation, control soil erosion and protect embankments. However, over four decades, it spread to become the most dominant salt marsh plant along China’s 18 000-km coastline, currently accounting for more than 60% of the salt marsh area in China, causing serious ecological problems (Nie et al. 2023). So the question now is, can we predict which species will have a major impact on ecosystems before they are introduced?

The aim of the current study is to try to solve the above question by linking the invasiveness of exotic plant species in China with their taxonomy as well as functional traits, and specifically, we will: (i) clarify the relationship between the invasion level of exotic species and taxonomy (family to which each species belongs), (ii) reveal the relationship between plant functional traits and invasion level and (iii) establish prediction models based on plant functional traits and taxonomy to provide a reference for decision making on the introduction of exotic species and the management of invasive species.

METHODS

Data source

TRY is an enduring and worldwide shared database of botanical attribute information encompassing traits documented in the literature or by research species (https://www.try-db.org/TryWeb/Home.php). IASC is an authoritative online information platform for biological invasion communication in China. It includes the Chinese Invasive Alien Species Database, Geographical Distribution Information System as well as cumulative research (Supplementary Fig. S1).

In the current study, we download the list of alien plant species as well as the harmfulness level from IASC and their functional traits data from TRY. Here, we selected six functional traits including seed dry mass (SDM, mg), SL (year), SGE (%), seed number per individual (SNI), plant height (PH, m) and special leaf area (LMA, mg/mm²). Based on previous studies, these traits are most likely linked with species’ invasiveness and those are also fully documented in the TRY database (Supplementary Table S1). According to IASC, we categorized species of high and middle levels as invasive species and others as non-invasive species.

Mantel’s test

Mantel’s test is based on a regression analysis of dissimilarity matrices summarizing all pairwise sample combinations. In the current study, we built the independent matrix by the six selected traits and the dependent variable matrix by species’ invasiveness.

For each matrix X (m × k), a distance or dissimilarity matrix is calculated. This is achieved through the application of a selected measure, such as the Euclidean or standardized Euclidean distance, resulting in two matrices, D1 (m × m) and D2 (m × m). The dissimilarity matrix is comprised of pairwise dissimilarity measures d_ij between all objects (i.e. pathways) (i,j = 1, …, m). As a result, it is symmetric (d_ij = d_ji) and the diagonal elements equate to 0 (d_jj = 0). The most related traits influencing invasiveness are tested by Mantel’s P-values and Mantel’s R-values.

Principal component analysis

Principal component analysis (PCA) is a technique used in multivariate analysis to detect the primary components or factors responsible for the variations within a system. Here, we take the most related traits selected by the Mantel test (SL, SDM, PH and LMA) as original variables, and transform the original set of variables into main components. Each component retains a percentage of the original variance and is extracted in order from most explanatory to least explanatory.

Recursive split tree

It is a decision tree model based on recursive segmentation greedy algorithm to classify data. The dependent variable is a categorical variable, taking the values 1, 2, …, K. In our current work, we have constructed two sets of independent variables. The first set contains the most related functional traits tested by the Mantel test (SL, SDM, PH and LMA), while the second set adds the taxonomic information (the family to which the species belong). Let the region R_m contain N_m observations, $pmk^$ is the frequency of occurrence of class k observations at the m th node as follows:

Divide the observation on the mth node into:

then the most dominant category on that node is obtained. The main criteria for splitting nodes at present are information gain, information gain rate and Gini index, and this study adopts the Gini index for node splitting. Its mathematical expression is as follows:

where c denotes the number of categories in the dataset and p_i denotes the number of samples i as a proportion of all samples. From this formula, it can be seen that the Gini index increases when there is a higher degree of data mixing in the dataset. When there is only one data type in dataset D, then the value of the Gini index is a minimum of 0. If the selected attribute is A, then the formula for the Gini index of the split dataset D is:

where k denotes that the sample D is divided into k parts and the dataset D is split into k D_j datasets. To address the problem of over- and under-fitting the model, the tree needs to be pruned, and cost complexity pruning is introduced here.

Generate a large tree T₀ and consider a subtree $T⊂T0$⁠. A subtree is obtained by deleting internal nodes from the large tree. Denote the number of leaf nodes (final nodes) of T by |T|. Define the cost complexity criterion as follows:

For each α, find a subtree $Tα⊂T0$ to minimize C_α(T), while α plays a role in balancing the size of the tree and the quality of data fitting, with larger α giving a smaller tree and smaller α giving a larger tree. For each α, it can be shown that there exists a unique minimal subtree T_α such that C_α(T) is minimized.

To find T_α we use weakest link pruning: we successively collapse the internal node that produces the smallest per-node increase in

and continue until we produce the single-node (root) tree. This gives a sequence of subtrees, and this sequence must contain T_α. Estimation of α is achieved by cross-validation: we choose the value $α^$ to minimize the cross-validation sum of squares.

In the current work, we simply parameterize the invasiveness as 1 (for high), 2 (middle), 3 (need attention) and 4 (harmless), which means lower values indicate a higher invasive capability. Similarly, we parameterized all the invasive species as 1 and non-invasive species as 2. As shown in Fig. 3, all the functional traits we selected are mutually independent (all the P-values >0.1), which indicates the selected traits give us non-redundant information. All the analyses above are performed in the R software (r-project.org) by self-programming.

RESULTS

Taxonomy of exotic plant species in China

As shown in Fig. 1a, all the exotic species belong to 85 families, and their distribution in different families meets power-law distribution. 490 (over 71.9%) species come from 17 (accounts for 20%) families, of which Compositae (98 species), Fabaceae (80 species), Poaceae (65 species), Solanaceae (31 species), Euphorbiaceae (23 species) and Amaranthaceae (22 species) are the top 6 families. Fig. 1b focuses on the invasive species and the Matthew’s effect is even more significant. 296 invasive species come from 50 families, while 200 of them (over 74%) come from 10 (account for 20%) families. Compositae (50 species), Fabaceae (38 species), Poaceae (36 species), Amaranthaceae (18 species), Solanaceae (15 species) and Euphorbiaceae (13 species) are also in the top 6 families.

To prove the reliability and universality of the above results, we also did parameter estimation based on power-law distribution. We parameterized family simply as the order of the species number they have, and the dependent variable is the proportion of species in a specific family. The fitting curves and fitting equations are shown in Fig. 1a and b. When we considered all the exotic species together, the residual standard error is 0.03, and the achieved convergence tolerance is 4.87 × 10⁻⁶. When we only take invasive species into consideration, the residual standard error is 0.01, and the achieved convergence tolerance is 5.49 × 10⁻⁶. These results support the point that species from different families showed different invasiveness (Supplementary Table S2).

In summary, species taxonomy is one of the potential factors to predict if one species will become an invasive one after introduction.

Functional traits and invasiveness

Fig. 2 shows the difference of species’ functional traits with different invasiveness. SDM, SL and PH show significant differences between invasive species and non-invasive species statistically. In detail, the average SDM of high- and middle-level species is 11.90 and 4.84 mg, respectively, while the values of non-invasive species are 41.37 and 155.14 mg for need attention and harmless levels, respectively. The difference between invasive (7.65 mg) and non-invasive (67.49 mg) species is statistically significant (P < 0.001). The average SL of high, middle, need attention and harmless levels are 5.60, 8.56, 3.63 and 3.33 years prospectively, and the average value of invasive species (7.63 years) is significantly longer than non-invasive species (3.57 years, P < 0.05). The third significant difference is the PH. Species in high and middle levels need attention, and harmless levels are 1.46, 1.19, 5.60 and 4.25 m, respectively, the average PH of invasive species (1.29 m) is significantly (P < 0.001) lower than non-invasive species (4.59 m). Thus, statistically, invasive species have smaller and lighter weeds, longer SL and shorter PH.

SGE, SNPI and LMA do not show differences among invasiveness levels, and invasive species do not show differences with non-invasive species in these three factors. To link the function traits to invasiveness, we also did an independence test for the traits we selected and a Mantel analysis between the trait matrix and invasiveness matrix (Fig. 3).

For the Mantel’s analysis, the functional traits were defined as independent values and parameterized invasiveness as dependent values, and we linked the invasive level with each function trait. As shown in Fig. 3, functional traits do not show explanatory power neither significant correlation to the species’ invasiveness according to Mantel’s P-values and Mantel’s R-values. However, when we classified all the species into invasive and non-invasive groups, the parameterized dependent values showed a positive correlation to SDM, PH and LMA, and a negative correlation to SL. Since we parameterize all invasive species as 1 and non-invasive species as 2, the result indicates species with smaller seeds, smaller individuals, lower LMA and longer SL are more likely to be invasive species after introduction to foreign ecosystems. These results also double checked the analyze above.

Linking functional traits’ matrix to invasiveness

R-Values from Mantel’s regression analysis of each pair are still extremely low, and thus, a single functional trait lacks the power to predict invasiveness. Thus, we conducted PCA analysis for the functional trait matrix containing each one significantly correlated to invasiveness (Fig. 4a). We also overlayed species on the projected two-dimensional surface composed of orthogonal first and second components (Fig. 4b).

The first two components give over 72.7% information of the four-dimensional matrix and SL, SDM, PH and LMA contribute 30.13%, 27.45%, 24.46% and 17.96%, respectively. SL and SDM influence invasiveness more than the other two traits. Moreover, as shown in Fig. 4b, non-invasive species are distributed mainly in the third and the fourth quadrants, which indicates they have relatively shorter seed longevity, heavy seed mass, shorter PH and lower leaf mass per area. Furthermore, seed longevity and seed mass are more important for plants’ invasion. In summary, exotic species with longer SL and lower SDM are more likely to be invasive in China.

Uncertainty arises from the distribution of invasive species, which is relatively evenly distributed throughout the ordination, suggesting that there is no clustering of invasive species across these principal components. The PCA results do not indicate a separation between invasive and non-invasive species. For example, invasive species are likely to have shorter seed longevity, heavy seed mass, shorter PH and lower leaf mass per area, but not all exotic species with these traits are invasive.

Prediction of exotic plant species invasion

Basic statistics, Mantel’s analysis and PCA together indicate that SL, SDM, PH and LMA are four potential functional traits that can predict if a plant species could be invasive for foreign ecosystems, and the demography analysis also indicates species from some families have higher potential to become invasive. Thus, we trained two predictive models based on the recursive split tree method to check if we can predict species’ invasion. 70% of data were randomly selected to train the model and the remaining 30% of data were used to test the model.

Model 1 only takes functional traits into consideration, the overall accuracy is 67.78%, and it is more sensitive to non-invasive species. The accuracy of non-invasive species prediction is over 70% (Fig. 5a). Then we took taxonomy (family of each species belongs to) into consideration and trained Model 2. The results show the overall accuracy dropped to 56.11%, and while it is more sensitive to invasive species, the accuracy of invasive species prediction is over 72% (Fig. 5b). Combining the two models together, statistically and theoretically, we could predict if a species is invasive from its traits and taxonomy with a 91.84% accuracy. This model could help local governments, managers and stakeholders to evaluate shall we introduce some plant species in China.

DISCUSSION

What kinds of species tend to be more invasive?

According to our analysis, the taxonomy of the alien plant is likely to be one of the factors that determine whether an alien plant is potentially invasive. Independent studies in Europe, the Americas and Africa showed similar results. A site study in South Africa found that the dominant families of alien plants were Solanaceae and Fabaceae (Pyšek et al. 2012). Based on information provided by the European Alien Species Information Network (EASIN), the families possessing the highest number of alien taxa are Asteraceae, Poaceae and Fabaceae (Arianoutsou et al. 2021). A separate study conducted in Chile likewise found that the majority of alien plants belong to the families Poaceae, Asteraceae, Fabaceae and Brassicaceae. Based on the current and existed research findings, we can confidently conclude that certain families of species are more likely to become invasive.

Moreover, a prior meta-analysis has discovered that the predictability of plant invasion effects could be influenced by species traits, invasion context and environmental conditions including habitat and biome (Pyšek et al. 2012). Life form, height and type of pollination are a few of the traits that can be used to anticipate the effects of species invasion, no matter the specific habitat or geographical area. Nonetheless, the importance of a plant invasion’s impact on soil resources cannot be estimated based on species traits or environmental context alone, but increasingly hinges on the identification of individual species and their taxonomic affiliation. Site history and location may also be significant factors for the ecological outcomes of invasion. Evaluating the attributes of the species and invasion context is crucial in predicting the impact of invasive plants.

In our current work, we found that the plants with longevous seeds and shorter PH are more invasive. This outcome aligns with the presence of varied life histories and functional approaches that grant successful invasion in distinct scenarios and ecological circumstances (Callaway and Aschehoug 2000; Catford et al. 2019; Marta Carboni 2016; Shiflett et al. 2017). Under harsh environmental conditions, the species with heavier seeds had higher seedling rates (Catford et al. 2019). When the seed life is long, the relative abundance of the species is high (Grubb 1998). Furthermore, research indicates that plants capable of both sexual and asexual reproduction are likely to exhibit higher levels of relative abundance, spread rates and environmental range sizes (Thomas et al. 2019). Plants exhibiting these traits can serve as formidable competitors through the employment of conservative resource utilization strategies (Cornwell and Ackerly 2010).

The TRY database lacks nutrient and root trait data for many species. According to existing studies, invasive plant species often increase biomass, net primary production and nitrogen availability. They also alter nitrogen fixation rates and produce litter that decomposes more quickly than that of co-occurring native species. Exotic plant invasions can double gross nitrification rates by altering the abundance and composition of ammonia-oxidizing bacteria in soil. This can potentially affect ecosystem nutrient budgets (Hawkes et al. 2005). Invasive plant species gain a fitness advantage in the alien range by developing root traits that increase resource uptake and competitive ability, such as increased specific root length and branching intensity. In addition, the root traits of exotic species such as S. alterniflora are highly sensitive to soil nutrient availability. This sensitivity leads to increased productivity through higher root nitrogen concentration and root length density (Liu et al. 2020). Further experiments and analysis of big data are necessary to fully understand how nutrient availability and root traits affect the invasion of exotic plant species.

Seeds and fitness

In the current study, we found longer SL and lower SDM help exotic plant species become more invasive. Longer SL could help the species survive harsh natural conditions for a relatively long time and wait for more niches because of the death of native species. When a species invades a new habitat, it usually involves only a few groups or individuals from the ancestral species. This can lead to a decline in the genetic diversity of new species formed or invasive groups, known as the ‘bottleneck effect’ (Yuri et al. 2017).

Despite the low genetic diversity caused by bottleneck effects, invasive plants were able to adapt to new habitats (Arnaud et al. 2016). This huge contrast between very low genetic diversity and very high adaptive capacity has been called the genetic paradox of biological invasion (Chuong et al. 2017). Rapid phenotypic changes in the adaptive traits of invasive plants are essential for organisms to thrive in a constantly changing environment (Van’T Hof et al. 2016). Transposable elements (TEs), which are among the most variable components of the genome, can duplicate and insert themselves into novel locations. Environmental alterations, such as climate change, can concurrently modify the amount of copy numbers of transposable factors and their impacts on gene regulation, resulting in new genetic and phenotypic variations that may be adaptively significant (Stapley et al. 2015). Therefore, TE may facilitate the rapid creation of rich genetic polymorphisms, thus acting as a rapid adaptation agent for species. This adaption relies on a large number of offspring. Lower SDM species generally produce more seeds and provide more possibilities for such rapid adaptation. Another field study also indicated native communities with heavier seeds were more heavily invaded (Ernst et al. 2022). The precise causal mechanism of this pattern remains unknown (Liu et al. 2021). However, it is believed to reflect hierarchical disparities in competitive fitness, triggered by dissimilarities in seed size, heightened resistance to invasion pressure during the seedling stage and enhanced availability of microsites for smaller-seeded invasive species. In brief, indigenous species with bulkier seeds displayed an adverse response to the pressure exerted by the invasive ones.

Plant interspecies trade-offs

Invasive flora significantly affects the composition of species, the structure of plant communities and the function of ecosystems (Adler et al. 2018). Trait–environment relationships are generalizably consistent across plant communities and spatial scales with environmental gradients globally (Conti et al. 2018). It has been suggested that the introduction of plants into a new area could result in the emergence of evolutionary changes due to novel selection pressures emanating from both biotic and abiotic factors (Cadotte 2017). This may result in significant ecological disparities between indigenous and invasive plant species. The hypothesis of increased competitive ability evolution suggests that foreign plant lifeforms may successfully break free from the restraints of natural foes in their new environments (Kuebbing and Nu Ez 2016). Invasive plant species often evolve to lose crucial defensive traits, reallocating resources and energy from defence to growth. This can result in higher interspecific competition with native plants, which may be due to their larger aboveground biomass and stronger allelopathy compared with native species (Golivets et al. 2018). It has been reported that invasive plants may use their dense canopies to protect competitors within their invasive range (Kraft et al. 2015). Previous research has indicated that non-native plants can exert significant allelopathic impacts on surrounding vegetation (Rachel et al. 2016). Consistent with the ‘new weapons’ hypothesis, indigenous species within the range of invasive plants exhibited greater sensitivity to allelochemicals from non-native species than indigenous species beyond the range of invasive plants (Funk and Wolf 2016). The study demonstrates, through extensive analysis of big data, that while niche distinctions may require multiple trait combinations for description, fitness disparities that determine competitive edge correspond to differences in the quality of certain individual traits (Sheppard and Schurr 2019). The grade difference of individual traits can affect species growth and seed yield (Li et al. 2015). The higher the quality of seeds produced by the species, the more superior their performance is relative to coexisting alien species (Narwani et al. 2015). Overall, it seems that the traits linked to competitive ability are more influential than niche-specific conditions in affecting the success of alien plants in symbiotic relationships (Sheppard 2019).

Influence of phylogeny

Invader performance increased with distance to resident species, and species-rich communities with more distantly related species contributed positively to invader performance (El Barougy et al. 2020). Functional diversity, as a measure of trait distribution, is thought to capture some aspects of the niche directly and can reflect the breadth of strategies used by the organisms that make up a community. However, the utility of functional diversity as a proxy for the niche hinges on selecting ecologically relevant traits, which is a notoriously difficult task (Ernst et al. 2022).

By integrating over the evolutionary history of a species and the lineage from which it arises, phylogenetic diversity as well as distance is hypothesized to capture further aspects of niche differentiation. However, the relationship between phylogeny and invasion success is still unclear and has been intensively debated and criticized due to weakly supported underlying assumptions (Qin et al. 2020). Phylogenetic diversity will only reflect ecological differences between species at phylogenetic scales at which ecologically important traits are conserved or exhibit phylogenetic signals.

Phylogenetically distant alien species might be more likely to establish themselves in a local community, and phylogenetically similar species could suffer from phylogenetically related parasites that make the area inhospitable for seedlings of closely related species (Dawson 2015). Mixed evidence has been reported on the relative importance of species richness and phylogenetic distance for resistance to plant invasions, with significant negative relationships found between species richness and invasibility as well as for phylogenetic distance and invasions. Therefore, the relationship between phylogenetic distance and community invasibility may depend on the mode of interspecific interactions and the resulting evolutionary trajectory of the recipient community (Qin et al. 2020).

In the current study, based on the APG IV system, the top 10 families with the highest number of species in China come from 10 different orders (Supplementary Fig. S1), which probably indicates that their phylogenetic relatedness is less likely to influence their functional traits and further this analysis results. Recent studies also examine the relationship between invasive and naturalized flowering plant species in China, analyzing their phylogenetic relatedness (Qian et al. 2022). The study tests four hypotheses and collects data from 28 province-level regions in China, analyzing it using two phylogenetic metrics. The study finds that invasive species form a phylogenetically clustered subset of naturalized species, and more harmful invasive species are more strongly clustered. The clustering of invasive species is consistent across different climatic conditions, and both invasive and naturalized species show greater phylogenetic clustering in regions with stressful climates (Joan 2004). The findings have implications for predicting and controlling invasive species based on their phylogenetic relatedness to naturalized species. The study shows that by measuring both tip- and basal-weighted phylogenetic relatedness, invasive species are overwhelmingly phylogenetically clustered with respect to their naturalized species pool. More harmful invasive species are more strongly clustered with respect to their naturalized species pools compared with less harmful invasive species, and these findings have significant implications for invasion ecology and conservation biology. Overall, the study aims to improve our understanding of geographic and ecological patterns of phylogenetic relatedness of invasive and naturalized plant species in China. Another potted plant study also showed the negative responses of Erigeron canadensis to neighboring invasive grasses and forbs increased with increasing phylogenetic distance between the neighbors and E. canadensis. In contrast, the positive responses to invasive legumes did not depend on phylogenetic distance from E. canadensis. These findings suggest that phylogenetic distance plays a significant role in the interactions between native and neighboring invasive species, with closer phylogenetic relationships promoting positive interactions and increasing distances resulting in more negative responses (Ren et al. 2023).

Given the potential and limits of each of these components of diversity, it is important to consider multiple facets and evaluate the degree to which they provide similar versus distinct insights into community assembly generally, and invasion dynamics more specifically.

CONCLUSIONS

Our study found that species taxonomy is one of the potential factors to predict if one species will become an invasive one after introduction. Evaluating the attributes of the species and invasion context is crucial in predicting the impact of invasive plants. In our current work, species with smaller seeds, smaller individuals, lower LMA and longer SL are more likely to be an invasive species after introduction to foreign ecosystems. Meanwhile, the higher the quality of seeds produced by the species, the more superior their performance is relative to coexisting alien species. Overall, it seems that the traits linked to competitive ability are more influential than niche-specific conditions in affecting the success of alien plants in symbiotic relationships.

Supplementary Material

Supplementary material is available at Journal of Plant Ecology online.

Figure S1: Angiosperm phylogeny flowering plant system of exotic plant species in the current study (APG IV).

Table S1: Indicator system for risk assessment of invasive alien plants (based on The Analytic Hierarchy Process).

Table S2: Classification of invasive plant risk levels.

Funding

This research was supported by National Forestry and Grassland Administration Emergency Leading the Charge with Open Competition Project (202302) and the Fundamental Research Funds for the Central University (BLX202250).

Conflict of interest statement. The authors declare that they have no conflict of interest.

REFERENCES

Author notes