Global trade and travel activities are leading to an increasing accumulation of alien species across regions of the world (CBD, 2014; Seebens et al., 2017; Beaury et al., 2021). The establishment and subsequent spread of these species in non-native regions have major negative impacts on economic activities and human wellbeing (IPBES, 2019; Diagne et al., 2021), and is among the major threats to biodiversity and species extinction worldwide (Clavero and Garcia-Berthou, 2005; Bellard et al., 2016). Recovering from ecosystem degradation has been one major motivation of global restoration effort and current policies (e.g., UN Decade on Ecosystem Restoration and EU Green Deal). However, managing and controlling invasive alien plant species (i.e., the subset of alien plant species that become established and cause negative impacts; hereafter “IAP”) can be expensive, laborious, and often technically difficult to achieve success (Marais et al., 2004; Haubrock et al., 2021). In fact, previous actions against IAP have not been sufficient to counteract their increasing globalization. To support these actions, substantial research has been done to understand the main dispersal pathways (Pyšek and Richardson, 2010), and spatial distribution patterns of IAP (Vilà et al., 2010). However, research on the distribution patterns of IAP or alien species in general has, thus far, focused mostly on large scales, like global or national studies (Chytrý et al., 2009; Seebens et al., 2017; Essl et al., 2019; Capinha et al., 2020), but detailed knowledge on the patterns and drivers of invasion at local scales—which are key for supporting management and restoration efforts—remains missing for many regions and taxonomic groups.

Here we use a comprehensive local scale data set to examine and describe the geography of invasions by alien plants in riparian ecosystems of continental Portugal. Continental Portugal harbors at least 3,314 vascular plant species (de Sequeira et al., 2012), of which at least 772 are alien (de Almeida, 2018) and 113 are listed as invasive species in the Portuguese law (Law No. 92/2019, 2019). Previous works have addressed plant invasions in Portuguese riverine systems (Bernez et al., 2006; Aguiar and Ferreira, 2013), however, how IAP are distributed in riparian ecosystems across the country remains largely unknown, making it difficult to identify areas under higher pressure from invasions and for which management or restoration efforts would be most needed. This occurs despite riparian ecosystems often being biodiversity hotspots (Duarte et al., 2004; Stella et al., 2013) providing a range of ecosystem functions and services related to water quality, microclimate, structural habitat, a bottom-up energy provider for the food web, and riverbank stability (Naiman et al., 2005), and being one of the most susceptible habitats to invasion (Richardson et al., 2007), often hosting a high diversity of IAP (Vilà et al., 2007; Chytrý et al., 2008).

In light of previous findings for other regions and taxa, a few hypotheses can be made about the drivers of invasion patterns in Portuguese riparian ecosystems. Propagule pressure resulting from proximity to urban areas and human activities may lead to the introduction of cosmopolitan, ornamental, nitrophilous, or cultivated species, originating hotspots of alien plant species (Hruska et al., 2008; Hulme, 2009; Pyšek et al., 2010, 2020). Land-use change is also an important driver of biodiversity change (Sala, 2000), disrupting ecological equilibrium and creating opportunities for alien species to establish (Richardson et al., 2007). Similarly, natural regular disturbances and the diversity of microhabitats, as well as movement of organisms, nutrients, and sediments create establishment opportunities and make riparian areas vulnerable to biological invasions (Pyšek and Prach, 1994; Catford et al., 2014). Seasonal patterns of summer droughts and winter floods, typical of Mediterranean rivers, lead to nutrient pulses and extreme natural disturbances, conditions that are also considered to favor invasions (Gasith and Resh, 1999; Davis et al., 2000; Stella et al., 2013). Finally, climatic variation across wide spatial extents is also a likely relevant factor in determining the distribution of alien plant species (Thuiller et al., 2008).

In this context, the goals of this study are to (i) map the richness of invasive alien plant species in riparian ecosystems of continental Portugal, (ii) assess the human and natural factors related to spatial variation in the values of this variable, and (iii) predict richness values along water courses and at the municipal level for the whole study area.

A total of 960 plant species were found in riparian ecosystems of continental Portugal. In 382 sites 97 alien species were found and of these 34 are also considered as invasive plant species. The highest number of alien species found in a single plot was 15, and the highest number of invasive alien species was 10. Of the 382 sites, IAP were recorded in 297 sites (77.75%). The IAP species recorded more frequently were Bidens frondosa (207 sites), Conyza bonariensis (131 sites), Phytolacca americana (96 sites), and Arundo donax (83 sites) (Supplementary Table 1).

The model results suggest that richness of invasive plant species is negatively associated with altitude, and positively with the size of upstream catchment and percentage of urban area (Figure 2 and Table 2). The model explained a substantial amount of variability in richness values, as indicated by an adjusted-R² of 0.63.

The k-fold validation procedure recorded a mean absolute error (MAE) of 1.03 ± 0.1 species and relative absolute error (RAE) of 0.63 ± 0.06 (Table 3). The error values of the LOOCV show similar values (Table 3). The relatively small deviations between predicted and observed richness values support the robustness of the model and its use for predictive purposes.

The prediction maps show that there is a higher diversity of IAP in the central and northwestern, lowland, regions of continental Portugal (Figures 3, 4). Low richness values were predicted for river sections in inland regions, regions south of Lisbon, and for mountainous areas. Our model predicts that for 43.1% of river sections of continental Portugal there are one to two invasive plant species, for 11.5% there are three to five, and for 2.2% there are five or more invasive plant species (Figure 3). At the municipal level, the highest average richness of invasive alien plants occurs around the country’s two main cities, Porto and Lisbon (Figure 4). High richness is generally found throughout the central to north-western continental Portugal. Although the highest numbers are concentrated around the larger cities, nearby areas belonging to the same river basins are also considerably affected. River basins with the highest average IAP richness are those of the rivers Lis (3.42 ± 0.97), Vouga (3.13 ± 0.97), Ave (2.88 ± 1.40), and Mondego (2.72 ± 1.64). The basins that harbor more IAP overall are Douro (20), Mondego (20), and Vouga (18) (Supplementary Table 1).

Our results identified a marked distribution pattern in the richness of invasive alien species in riparian environments of continental Portugal. This pattern could be explained, to a large extent, by a few variables, namely altitude, size of the catchment area and the proportion of urban area in a buffer of 500 m around the target site. In addition, a model using these variables was able to predict the distribution of richness values in unsampled sites with good accuracy and at a high spatial resolution, enabling us to identify the river sections most strongly affected by invasions and for which management or restoration actions are more pressing.

Model results show that the richness of invasive alien plants was higher at sites up to 300 meters above sea level, and far from river sources. At higher altitudes, richness values of invasive alien plants decrease quickly. The relationship between increasing altitude and decreasing IAP richness has been often observed (Pyšek et al., 2002; Chytrý et al., 2005, 2009; Anđelković et al., 2022), and has been attributed to the generally milder climatic conditions found in low-altitude areas (Chytrý et al., 2005), benefiting IAP that mostly originate from warm areas (Pyšek et al., 2003). On the other hand, the IAP found at higher altitudes are mostly generalists with broad climate tolerances and able to thrive across a wide range of altitudes (Alexander et al., 2011; Essl et al., 2019). In our data set, 8 species were found at elevations above 500 m. The two most common were Bidens frondosa and Phytolacca americana; both native to North America and invasive in many temperate regions of the world (Huang and Ding, 2015; Ronzhina, 2017).

The positive relationship found between the size of the contributing catchment area and IAP richness likely reflects a scaling effect played by the dendritic structure of riparian networks, suggesting a key role of hydrochory in the accumulation of IAP (Rodríguez-González et al., 2019). A larger catchment area may result in a higher number of IAP simply because more species accumulate in downstream areas. Additionally, it draws attention to the interconnection of river segments along the same river basin, which therefore should not be considered or treated separately (Fausch et al., 2002; Rodríguez-González et al., 2019). In our study area, the higher Strahler stream segments of the larger river basins Tejo, Mondego, and Douro hold higher IAP numbers, and across all river basins, an increase toward the coast can be observed (Figure 3).

Socioeconomic influences are important drivers of invasive species accumulation, often more important than the physical environment (Dawson et al., 2017; Essl et al., 2019). Here, we also identified a significant positive relationship between richness of invasive alien plant species and the percentage of urban area surrounding each sampled site. Urban areas can favor invasions in multiple ways, very often through high colonization and propagule pressures that result from ornamental horticultural activities promoting the plantation of alien species in urban parks and gardens (van Kleunen et al., 2018), from accidental introductions resulting from trading or tourism activities (Hulme, 2009), and in general due to global plant trade (Beaury et al., 2021). At the same time, intense human activity also provides conditions that can favor the establishment of alien species, including disturbances or the creation of novel habitats that provide niche opportunities by reducing natural enemies, such as native plant competitors (McKinney, 2006). Although it is not possible to distinguish the relative contribution of these and other urban-related mechanisms in shaping the observed patterns of IAP richness, it seems likely that the number of IAP in the vicinity of these areas will continue to grow driven by high levels of socio-economic activity.

Previous research has found important limitations in the capacity of statistical models to predict regional-scale richness of alien species (Capinha et al., 2018). We found a good accuracy in predicting IAP richness, with error levels substantially below the threshold indicating uninformative predictions. The reasons for the high accuracy achieved here could be related to the lower diversity of human and natural conditions represented in our model compared to those tested in Capinha et al. (2018), and which comprised mostly continental to global extents. Furthermore, our data sample most of continental Portugal, ensuring that predictions are within the range of conditions used for model calibration, thus improving their reliability (Capinha et al., 2018). This provides good support for the use of our predictions to support decision-making in ongoing and future efforts to reduce IAP pressures on riparian ecosystems. In this regard, it is worth mentioning that the model is based on richness data collected between 2003 and 2006—the most recent available—so although the spatial patterns of variation are unlikely to have changed significantly, the current figures of IAP richness may be even slightly higher than predicted, reflecting the continued accumulation of alien plants worldwide as time progresses (Seebens et al., 2017, 2021).

According to the OECD (2019), in order to reach the goal of the EU Biodiversity Strategy for 2030 to decrease the number of Red List species they threaten by 50% (European Commission, 2020), invasive species must be controlled or eradicated by 2030 from 80% of the most important areas of plant diversity in Europe. Future work could thus aim at crossing the richness patterns we uncovered with observed or potential distribution maps of endangered species listed in the recently published Red List of Vascular Plants of Mainland Portugal (Carapeto et al., 2020) to help pinpoint locations where threatened riparian plant species could be under greater pressure from IAP. Similarly, our estimates could also be combined with estimates of richness of native plant species to map the areas where the proportion of alien invasive species on overall species diversity is higher (Catford et al., 2012). Results from these analyses could then be used to inform national governmental entities in the definition of priority areas for conservation or restoration efforts and by local entities, such as municipalities or environmental NGOs in directing future efforts of eradication or introduction prevention of riparian invasive species. Future work could also consider differing invasion pressures caused by individual species. For example, some areas may still be strongly affected by single invasive species like Acacia dealbata, which might propagate very fast and cover wide areas (Lorenzo et al., 2010; Souza-Alonso et al., 2017). Our estimates of species richness weighted by the magnitude of species-level impacts could provide a deeper understanding of the impact of invasion pressures on riparian ecosystems.

We provided a basis for understanding the distribution patterns of invasive alien plant species in riparian areas of continental Portugal. We found a consistent geographical clustering of highly invaded areas in central and northern coastal areas, which appears to be driven, to a large extent, by the joint effect of a high human population concentration, low altitude values, and the convergence of species spread in downstream areas. Our results also showed many contiguous municipalities sharing similar levels of invasion, thus recommending cross-municipal cooperation and coordination when addressing the prevention and management of invasive plants species. Given also the role identified for species spread through the hydrographic network, these efforts would also benefit from the involvement of entities with responsibilities over upstream river stretches (Fausch et al., 2002; Rodríguez-González et al., 2019).

Publicly available datasets were analyzed in this study. These data can be found here: https://www.apambiente.pt/dqa/index.html.

CC, FSD, LB-de-Á, PMR-G, and RP designed the original research. RP did the analytic work with contributions from FSD. All authors contributed to writing up the article.

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

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