Invasive plants have a delayed and longer flowering phenology than native plants in an ecoregional flora

Abstract

Background and Aims

Flowering phenology has been suggested as an important factor to explain invasions of non-native plant species. The success of invasive species might be enhanced by flowering at different times (the vacant niche hypothesis) or flowering for longer periods (the niche breadth hypothesis) than native species. However, comprehensive regional assessments of the flowering phenology of invasive and native floras are lacking in the literature. In this study, we evaluated the flowering phenology of invasive and native plant species pools to test the above-mentioned hypotheses within a biogeographically meaningful region.

Methods

We investigated the start, end and length of flowering between the invasive and native floras that occur at the same elevation range in the Cantabrian Mixed Forests ecoregion (north-west Iberian Peninsula), a biogeographical hotspot for invasive plants in south-west Europe. We also accounted for species habitat preferences and climatic and biogeographical origin of the invasive species.

Key Results

We found a mismatch in flowering time between the ecoregional invasive and native floras. Invasive species had a delayed and longer flowering phenology in comparison to native species. These differences in flowering time were more pronounced in man-made habitats and in invaders from temperate and tropical regions.

Conclusions

Our results are consistent with the vacant niche hypothesis; the asynchrony in flowering time could allow invaders to exploit a temporally empty niche. Our results are also consistent with the niche breadth hypothesis, suggesting that invasive species exhibit a longer flowering period than natives, which might allow them to have prolonged access to resources. Future studies should explore the phenological patterns of invasive and native species across biogeographically relevant regions to enhance our understanding of large-scale invasion events.

INTRODUCTION

Phenology, i.e. the timing of periodic life-history events, is key to successful survival and reproduction in plants and has major implications in structuring communities across space and time (Wolkovich and Cleland, 2011; Xavier et al., 2019). Plant phenology results from adaptations to avoid environmental stress and to maximize resource acquisition that ensure the viability of pollen and seeds (Pau et al., 2011; Reeb et al., 2020). For example, selection on flowering time is driven mainly by environmental cues, such as temperature and photoperiod, but it can also be influenced by mutualists (pollinators and seed dispersers) and antagonists (floral pathogens and predispersal seed predators) (Elzinga et al., 2007; Wolkovich et al., 2014). Thus, the timing of flowering is vital for plant fitness, determining the availability of potential mates, pollination success, seed production and seed dispersal and predation (Wilke and Irwin, 2010).

Flowering phenology has been suggested as an important factor to explain invasions of non-native plant species owing to its effect on plant fecundity (Baker, 1974; Pyšek and Richardson, 2007; Wolkovich and Cleland, 2011). To invade a new region successfully, non-native plants must overcome the novel abiotic and biotic barriers, or ecological filters, to allow survival and reproduction (Richardson et al., 2000). Surpassing these barriers involves flowering at the appropriate time of the year, which will depend on the ecological requirements of the plant (Godoy et al., 2009b). Species often exhibit genetic inertia on flowering development because flowering phenology is an evolutionary adaptation selected to avoid unfavourable environmental conditions in the regions where the plants originally evolved (Rathcke and Lacey, 1985; Herrera, 1992; Godoy et al., 2009b). Thus, if the novel environment is similar to that of the original range, non-native species are more likely to overcome the reproduction barrier. However, if the environmental conditions of the new range differ from those of the native range, the success of non-native species depends on the extent of these environmental differences, the inherent ability of the species to flower in varying conditions and its capacity to adapt through evolution or phenotypic plasticity (Thuiller et al., 2005; Godoy et al., 2009a).

Native and non-native species, having evolved in different regions, might exhibit distinct flowering phenology and climate sensitivity that is adapted to the conditions of their native environment (Godoy et al., 2009b; Reeb et al., 2020). For example, differences in phenology between native and non-native species are expected to be greater with increasing dissimilarity in the environmental conditions in which they evolved. The resulting phenological differences between native and non-native species might, in turn, affect community assembly and invasion success (Godoy and Levine, 2014; Reeb et al., 2020). In this regard, there are at least two non-mutually exclusive hypotheses on the role of flowering phenology in non-native species success (Wolkovich and Cleland, 2011). The vacant niche hypothesis proposes that non-native species may invade when there is a temporally empty niche to exploit, allowing them to use temporarily available resources (Elton, 1958). Some authors suggest that non-native species may benefit from establishing earlier in the season than native species, allowing them to access and exploit resources first, avoid harsh environmental conditions and become more competitive within the invaded community (Sale, 1977; Zettlemoyer et al., 2019). However, other studies have reported a contrasting pattern, in which invasive species flower later than natives, suggesting that successful reproductive strategies might vary with the environmental conditions of the recipient ecosystem (Celesti-Grapow et al., 2003; Godoy et al., 2009b; Durham et al., 2017; Ferreras et al., 2023). The niche breadth hypothesis proposes that non-native species might occupy a broader niche by exhibiting longer phenological phases (e.g. longer flowering) in comparison to native species, allowing them prolonged access to resources (Kolar and Lodge, 2001; Zettlemoyer et al., 2022).

There are several studies that have tested differences in flowering phenology between native and non-native species at small spatial scales (e.g. Lediuk et al., 2014; Durham et al., 2017; Xavier et al., 2019; Ferreras et al., 2023). Others have studied a subset of species that co-occur at large spatial scales (e.g. political divisions) (e.g. Godoy, et al., 2009a, b; Maruyama et al., 2024). However, a study comparing all coexisting native and invasive species at the ecoregional scale is lacking in the literature. Ecoregions are ecologically homogeneous units that occur within a country or across several countries, encompassing areas with a similar biogeographical history and recurrent local ecosystems (Bailey, 2004; Loveland and Merchant, 2004). Thus, ecoregions are likely to be invaded by species with similar ecological characteristics, such as specific macroclimatic requirements (Lázaro-Lobo et al., 2024), and provide a holistic framework for studying phenological patterns of native and invasive species within biogeographically meaningful regions.

In this study, we evaluated the flowering phenology of a regional invasive plant species pool (i.e. the whole list of invasive non-native species known in a regional flora) and how they differ from native plant species. As study area, we used the Cantabrian Mixed Forests ecoregion, one of the main hotspots for plant invasions in Europe (Lázaro-Lobo et al., 2024). Initially, we compared all the invasive species identified in the ecoregion with all the native species that occur at the same elevational range. Then, we compared the flowering phenology of species that live in similar habitats, because differences in timing of flowering might be attributed to differences in habitat conditions rather than to variations in flowering strategies (Thies and Kalko, 2004; Godoy et al., 2009a, b). Lastly, we grouped invasive species by their climatic and biogeographical realm of origin to analyse differences in flowering phenology. The specific hypotheses that we tested in this study were: (1) the vacant niche hypothesis, whereby invaders would flower when natives are not flowering; and (2) the niche breadth hypothesis, whereby invaders would have a longer flowering period than natives. We also hypothesized that these differences in flowering phenology between invasive and native species pools would become evident in man-made habitats (owing to their larger proportion of invasive species and human preferences for plants with extended and prolific flowering) and in invaders originated in biogeographical regions with different climatic conditions.

MATERIALS AND METHODS

Study area

The study was conducted in the WWF Cantabrian Mixed Forests ecoregion, available at the RESOLVE Ecoregions dataset (Ecoregions2017 ©Resolve; https://ecoregions.appspot.com/). The extent of the ecoregion largely fits with the Natura 2000 Atlantic biogeographical region in the Iberian Peninsula (https://environment.ec.europa.eu/). The ecoregion includes territories in NW Portugal, NW Spain and SW France, representing ~20 % of the Iberian Peninsula (Fig. 1).

Fig. 1.

(A) Ecoregions of western Europe, obtained from the RESOLVE Ecoregions dataset (Ecoregions2017 ©Resolve; https://ecoregions.appspot.com/). The black rectangle includes the study ecoregion (in green), and part of other ecoregions. (B) Study ecoregion. Darker colours represent higher elevations (≤2626 m). The red dots indicate recorded presences of invasive species (downloaded from different databases; see main text). The flower pictures within each habitat type represent common invasive (left) and native (right) species that occur in such habitats. The red and blue colours represent the flowering phenology for invasive and native species, respectively.

Unlike most of the Iberian Peninsula, the study ecoregion is characterized by warm summers, mild winters and generally abundant rainfall throughout the year (Rivas-Martínez et al., 2002; de Castro et al., 2005; Loidi, 2017). These climatic conditions, at the ecotone between the temperate and Mediterranean climate, provide an optimal setting for warm-temperate and humid ecosystems. Together with its varied topography and soil types, the ecoregion fosters a high diversity of ecosystems, being one of the most important areas for biodiversity conservation and carbon storage of the Iberian Peninsula (Jung et al., 2021). However, invasive plants are widely established within the study ecoregion, especially in areas near the coast, owing to the benign climate, high population density and complex international trade connections (Gassó et al., 2012; Fernández de Castro et al., 2018).

Species selection

We used a recently created checklist of 175 invasive alien plant species present in the ecoregion (the ecoregional invasive plant species pool; Lázaro-Lobo et al., 2024), based on expert knowledge, evidence from scientists and managers, and local literature (e.g. Fagúndez and Barrada, 2007; González Costales, 2007; Campos and Herrera, 2009). In the checklist, alien plant species were considered as invasive when they produce reproductive offspring (either by seeds or other propagules) in areas distant from sites of planting/sowing, without direct intervention by humans, independent of their impacts (sensu  Richardson et al., 2000; Pyšek et al., 2004).

To evaluate the elevational range where the invasive species occur, we extracted occurrence points at 1 km resolution (or higher) from global and regional databases. Global databases included the Global Biodiversity Information Facility (GBIF; https://www.gbif.org/; downloaded on 1 June 2023) and iNaturalist (https://www.inaturalist.org/; downloaded on 31 May 2023). Regional databases included administration reports (González Costales, 2007), projects (e.g. LIFE Fluvial, https://www.lifefluvial.eu/; and LIFE Stop Cortaderia, http://stopcortaderia.org/), doctoral theses and hydrographic administrations (e.g. Cantabrian (www.chcantabrico.es/) and Miño-Sil (www.chminosil.es/). We filtered the presence points from these databases to retain: (1) relatively recent records (i.e. those dated after 1950); (2) records with coordinate uncertainty of ≤1 km; and (3) records from suitable data sources (from iNaturalist, we removed non-research grade and cultivated plants; from GBIF, we excluded unknown data sources, in addition to fossil and living specimens, such as those found in botanical gardens). We randomly selected occurrence points that were separated by ≥1 km, to reduce the effect of local clusters of records in subsequent analyses. The elevation of the invasive species occurrence points was extracted from the Shuttle Radar Topography Mission (SRTM) database (https://srtm.csi.cgiar.org/), which was originally produced by NASA and has a spatial resolution of 90 m. Then, we calculated the highest 95th percentile of the elevation values of the invasive species occurrence points (758 m), thus excluding extreme outlier values. Lastly, we identified native species that occur within the study ecoregion at ≤758 m from the Iberian and Macaronesian Vegetation Information System (SIVIM) database (Font et al., 2010). Such database compiles all vegetation-plot data for the study ecoregion (Font et al., 2017).

Data collection

We compiled a range of species traits and biogeographical variables that, according to the literature (Rathcke and Lacey, 1985; Post and Stenseth, 1999; Godoy et al., 2009a, b; Maruyama et al., 2024), are likely to influence the flowering phenology of the selected species. The variables included: (1) taxonomic family; (2) growth form [graminoid or grass-like (flowering plant belonging to the families Poaceae, Cyperaceae and Juncaceae), forb/herb, vine, shrub or tree]; (3) lifespan (annual, biennial or perennial); (4) climatic origin of the invasive species, which was divided into tropical (between the Tropics of Cancer and Capricorn), temperate (between the Tropics and the Arctic and Antarctic circles) and Mediterranean (regions around latitude 40°N or 40°S, described by Cowling et al., 1996); (5) WWF biogeographical realm of origin (Afrotropical, Australasia, Indo-Malaya, Nearctic, Neotropical or Palaearctic. Realms encompass extensive areas of the land surface of the Earth where species have evolved in isolation for long periods. Our study ecoregion is included in the Palaearctic realm, meaning that invasive species that evolved in this realm might exhibit similar traits to the native species. If the native region of an invasive species covers two or more climatic or biogeographical categories, we considered the category where the species is more abundant); and (6) habitat type; we used a curated and habitat-classified regional subset of SIVIM to assign EUNIS habitat types (Chytrý et al., 2020) to the regional native flora. The EUNIS habitats invaded by each invasive species were determined with regional data and expert knowledge (Lázaro-Lobo et al., 2024). We differentiated broad habitat categories: coastal habitats (marshes, dunes, beaches and cliffs), wetlands (peatlands, bogs and continental water bodies), grasslands (lands dominated by grasses, forbs, mosses or lichens), shrublands (shrublands, heathlands, scrub and tundra), forests (forests and wooded lands) and man-made habitats (roadsides, human settlements, industrial developments, etc.).

The flowering phenology of invasive and native species was derived from the start, end and length (duration) of the flowering season, measured in months. Information on flowering phenology was obtained mainly from Flora Iberica (Castroviejo et al., 1986–2021). Flora Iberica is a compilation of all the vascular plants of the Iberian Peninsula and the Balearic Islands, both native and naturalized alien species, created by multiple experts in different taxonomic groups, using herbarium material from universities and research centres in the Iberian–Balearic territory. Given that Flora Iberica includes historical phenological data and that plant phenology can shift in response to climate change (Cleland et al., 2007), we compared the flowering data obtained from Flora Iberica with that from iNaturalist, which includes more recent data (from 2008), finding a high correlation for flowering start (r = 0.95), flowering end (r = 0.94) and flowering length (r = 0.92). Thus, we used historical phenological data from Flora Iberica for the analysis. However, we also used other sources (e.g. https://floraveg.eu/; https://invasoras.pt/; Campos and Herrera, 2009) when flowering data were not available in Flora Iberica.

Data analysis

Statistical analyses were performed using R v.4.3.0. Generalized linear mixed models (GLMMs) were used to evaluate differences in the start, end and length of flowering between invasive and native species. First, we considered the whole data set, with species status (i.e. invasive vs. native) as a fixed factor and with taxonomic family, growth form and lifespan as random factors to account for species traits that might also affect phenology (Reeb et al., 2020). Second, we evaluated differences between native and invasive species occurring in the same habitat type, controlling for the effects of taxonomic family, growth form and lifespan. For each habitat, we kept the native species that occurred in >5 % of the plots. Third, we assessed the flowering phenology of native and invasive species grouped by their climatic and biogeographical realm of origin, controlling for the effects of taxonomic family, growth form and lifespan, as in the previous analyses.

To build the GLMMs, we used the ‘glmmTMB’ package with the appropriate data distribution for each response variable, adhering to the guidelines for mixed models (Zuur et al., 2009). The start and end of flowering were analysed using a Gaussian distribution when the model residuals met the assumptions of normality and homoscedasticity or using a gamma distribution when these conditions where not met owing to skewed data distributions. The length of flowering was modelled as a negative binomial distribution to account for overdispersion. We assessed the goodness of fit of all the statistical models by plotting observed versus predicted values and by plotting residuals versus predicted values.

RESULTS

We found significant differences in flowering phenology between invasive and native species when all species were included in the analyses (Supplementary Data Tables S1 and S2). On average, invasive species started flowering in early May (mid-spring), whereas native species started flowering in early April (early spring; Figs. 2 and 3a; Supplementary Data Table S1). The length of the flowering period was greater for the invasive species (5.4 months, on average) than for the native species (4.5 months; Supplementary Data Table S1). Invasive species finished flowering in early September (late summer), whereas native species ended flowering in late July (mid-summer; Figs. 2 and 3a; Supplementary Data Table S1).

Fig. 2.

Circular histogram of flowering phenology showing the start and end of flowering time for invasive species (left) and native species (right). Each triangle represents the number of species that start or finish flowering in each month.

Fig. 3.

Boxplots showing the start, end and length of flowering for invasive species (red) and native species (blue). We indicate the results using the whole data set (A) and dividing the data by habitat type (B), climatic origin of invasive species (C) and biogeographical realm of origin of invasive species (D).

The same pattern of flowering phenology in invasive and native species was observed when the analysis was conducted separately for each habitat type (Fig. 3). However, we found significant differences only in man-made habitats, where invasive species started and ended flowering ~1 month later, on average, than native species (Supplementary Data Tables S1 and S3).

Flowering differences between invasive and native species were significant when the invasive species originated from tropical or temperate regions, but not when the invasive species was of Mediterranean origin (Fig. 3; Supplementary Data Tables S1 and S4). Tropical and temperate invaders started flowering ~1 month later, on average, than native species. Moreover, tropical and temperate invaders ended flowering 2.5 and 1.2 months later, respectively, than native species. The flowering length of tropical and temperate invaders was 1.8 and 1.5 months longer, respectively, than native species (Supplementary Data Table S1).

Regarding the biogeographical realm of origin of invasive species, we found that those with Afrotropical, Nearctic and Neotropical origin started flowering later than native species (Supplementary Data Tables S1 and S5) and that invaders originated in the Nearctic, Neotropical and Palaearctic realms finished flowering later than native species. Likewise, flowering length was longer in invasive species that came from Afrotropical, Australasian, Indo-Malayan and Neotropical realms, compared with native species (Supplementary Data Table S1).

DISCUSSION

This is, to our knowledge, the first study that compares the flowering phenology of native and invasive species at the ecoregional scale. Large-scale assessments of species phenology within biogeographically meaningful regions (i.e. areas with similar ecological and biogeographical characteristics) are key to providing insights into invasion events across the landscape. Our results suggest strong differences in flowering phenology between invasive and native species that coexist in the Cantabrian Mixed Forests ecoregion, which is consistent with the vacant niche hypothesis (Elton, 1958). Native plants bloom mainly in spring and early summer, whereas invasive plants extend their flowering period into autumn. This asynchrony in flowering time could allow invaders to use a temporally empty niche and avoid competition with native plants during critical periods of their life cycle (Wolkovich and Cleland, 2011).

We also found that the ecoregional invasive species pool starts flowering later than coexisting natives. Previous research suggests that early flowering would allow invaders to have greater access to resources and more opportunities for seed development, while avoiding harsher environmental conditions (the priority effects hypothesis) (Sale, 1977; Wolkovich and Cleland, 2011; Zettlemoyer et al., 2019). Consistent with this hypothesis, multiple studies have shown that invasive species tend to flower earlier than native species (e.g. Cadotte and Lovett-Doust, 2001; Lake and Leishman, 2004). However, other studies found the opposite pattern (Celesti-Grapow et al., 2003; Knapp and Kühn, 2012; Lediuk et al., 2014; Durham et al., 2017; Ferreras et al., 2023). Thus, early flowering is not the only reproductive strategy for successful establishment of invasive species, which might vary with the environmental conditions of the recipient ecosystem (Godoy et al., 2009b). It is therefore expected that changing patterns on flowering phenology differ across regions with different biogeographical settings. Our study ecoregion is characterized by warm and humid summers (Rivas-Martínez et al., 2002; de Castro et al., 2005; Loidi, 2017), which enables plant growth and flowering during summer and early autumn, unlike in nearby Mediterranean climates. Under temperate, humid and relatively cool summers, an early flowering strategy might hinder flower production owing to a higher climatic uncertainty (e.g. occurrence of late frosts), which might damage the plants and flowers (Wilsey et al., 2011). Thus, flowering when the growing season is more advanced might enhance reproduction owing to the improvement of the environmental conditions and floral resources (Ferreras et al., 2023).

Consistent with the niche breadth hypothesis, our results suggest that invasive species exhibit a longer flowering period than natives, which might allow them to have prolonged access to resources, increasing their reproductive output and colonization potential (Kolar and Lodge, 2001; Zettlemoyer et al., 2022). Given that the study area is characterized by a long growing season without a drought season, invasive species might take advantage of stable conditions in water availability. In addition, producing flowers throughout a longer time period might increase the chances of having seeds available when the right conditions for dispersal and establishment occur (Lake and Leishman, 2004). This larger floral display of invasive species, in comparison with natives, has also been registered by previous research and might be linked to the human-mediated introduction of plants with extended and prolific flowering (Cadotte and Lovett-Doust, 2001; Gerlach Jr and Rice, 2003; Lake and Leishman, 2004; Wilke and Irwin, 2010).

Differences in flowering phenology between invasive and native species at the ecoregional scale became evident in man-made habitats, as we hypothesized initially. These habitats include developed areas, such as roadsides, human settlements and industrial developments, which are subject to periodic disturbances that generate opportunities for the colonization and establishment of invasive plants (Christen and Matlack, 2006; Lázaro-Lobo et al., 2020). In fact, human-made habitats are the main areas for invasive species in the study ecoregion (Lázaro-Lobo et al., 2024), which might increase the probability of finding species with different phenology from the native ones. In addition, invasive species that bloom in periods when native flowers are not abundant are more attractive to humans (Knapp and Kühn, 2012). Thus, late-flowering plants that prolong the flowering season, such as Cortaderia selloana and Helianthus tuberosus, are often selected by humans for ornamental reasons (Kowarik, 2005; Knapp and Kühn, 2012). In contrast, there are many late-flowering invasive plants that were introduced accidentally and have not been dispersed actively by humans. Their successful spread indicates that invasive species with delayed flowering have an advantage, because they occupy a temporally empty niche, regardless of human preferences (Knapp and Kühn, 2012).

The evolutionary history of invasive species in their native regions influenced the observed phenological patterns. Invasive species that originated in tropical and temperate regions had a delayed and longer flowering phenology than native species that evolved at the transition zone between the temperate and Mediterranean climate. In contrast, invaders native to Mediterranean regions had a similar flowering phenology to native species. These results might be explained by the climatic conditions in which plants evolved originally, because flowering phenology is an adaptive trait selected to avoid unfavourable environmental conditions (Rathcke and Lacey, 1985; Herrera, 1992; Godoy et al., 2009b). Plant species native to Mediterranean regions, characterized by cold winters and arid summers, have a limited timeframe in which to complete their life cycle and primarily bloom in spring (Castro-Díez and Montserrat-Martí, 1998; Godoy et al., 2009a). However, temperate and tropical invaders have not faced the evolutionary pressures associated with arid summer conditions and might exhibit a different flowering pattern (Godoy et al., 2009a). In temperate regions, the flowering period can span from early spring to late summer, avoiding the winter frosts, which are the main environmental constraint (Rathcke and Lacey, 1985). In tropical regions, where temperature is not a limiting factor, the flowering period is affected by internal plant mechanisms, pollinator activity and the beginning of the tropical wet season, primarily aligning with the summer season in temperate climates (Wright and Calderon, 1995; Singh and Kushwaha, 2006; Godoy et al., 2009a).

Regarding species traits (taxonomic family, growth form and lifespan), we found that taxonomic family was especially important to explain the variability in flowering phenology. Thus, future research could further explore the effects of evolutionary relationships on phenological patterns of invasive and native species at the ecoregional scale.

Conclusion

The results from this study show a mismatch in flowering time between invasive and native species at the ecoregional scale. Invasive species had a delayed and longer flowering phenology in comparison to native species, especially temperate and tropical invaders occurring in man-made habitats. Phenological differences between invasive and native species may be explained by: (1) the successful spread of invasive species with different phenology from natives; (2) the evolutionary history of species in their regions of origin (e.g. climatic conditions in which plants have evolved originally); and (3) human factors (e.g. introduction history and human preferences for plants with extended and prolific flowering). This temporal difference in flowering could allow invasive species to avoid competition with native plants during critical periods of their life cycle. Late and prolonged flowering might confer a competitive advantage to invasive species by accessing available resources when most native plants have already completed their reproductive cycle. These findings highlight the importance of considering phenological characteristics of invasive and native species in management programmes, because they can significantly influence species dynamics and the ability of native communities to resist invasion. Future research should provide further insights into phenological patterns of invasive and native species within biogeographically meaningful regions to increase our understanding of invasion events at large spatial scales.

SUPPLEMENTARY DATA

Supplementary data are available at Annals of Botany online and consist of the following.

Table S1: Mean values of flowering phenology parameters between invasive and native plant species. Table S2: summary of the generalized linear mixed models for analyses conducted with the whole data set. Table S3: summary of the generalized linear mixed models for analyses conducted for each habitat type. Table S4: summary of the generalized linear mixed models for analyses grouping invasive species by their climatic origin. Table S5: summary of the generalized linear mixed models for analyses grouping invasive species by their biogeographical realm of origin.