Predicting the risk of aquatic plant invasions in Europe: How climatic factors and anthropogenic activity influence potential species distributions

Accepted Manuscript Title: Predicting the risk of aquatic plant invasions in Europe: how climatic factors and anthropogenic activity influence potential species distributions Authors: Argantonio Rodr´ ıguez-Merino, Pablo Garc´ ıa-Murillo, Santos Cirujano, Roc´ ıo Fern´ andez-Zamudio PII: S1617-1381(18)30023-2 DOI: https://doi.org/10.1016/j.jnc.2018.08.007 Reference: JNC 25655 To appear in: Received date: 17-1-2018 Revised date: 13-8-2018 Accepted date: 13-8-2018 Please cite this article as: Rodr´ ıguez-Merino A, Garc´ ıa-Murillo P, Cirujano S, Fern´ andez-Zamudio R, Predicting the risk of aquatic plant invasions in Europe: how climatic factors and anthropogenic activity influence potential species distributions, Journal for Nature Conservation (2018), https://doi.org/10.1016/j.jnc.2018.08.007 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. 1 Predicting the risk of aquatic plant invasions in Europe: how climatic factors and anthropogenic activity influence potential species distributions Argantonio Rodríguez-Merino*, Pablo García-Murillo*, Santos Cirujano † & Rocío Fernández-Zamudio‡ *Department of Plant Biology and Ecology, Faculty of Pharmacy, University of Seville, Spain, †Real Jardín Botánico (CSIC), Madrid, Spain and ‡Doñana Biological Station (CSIC), Monitoring Team on Natural Resources and Processes, Seville, Spain Corresponding author: Argantonio Rodríguez-Merino, Department of Plant Biology and Ecology, Faculty of Pharmacy, University of Seville, Profesor García González, 2, 41012 Seville, Spain. E-mail address: [email protected] ABSTRACT Predicting where species invasions will occur is one of the greatest challenges in conservation. Freshwater ecosystems are very vulnerable to the introduction of nonnative species for two reasons: (1) there are many routes of introduction by which nonnatives can arrive in freshwater systems; and, (2) freshwater systems are heavily impacted by a wide variety of human activities. Non-native aquatic plants can have harmful effects if they change habitat conditions, alter ecosystem functioning, and/or become key primary producers in invaded ecosystems. In this study, we focused on the potential distribution of non-native aquatic plants in Europe. The main objectives were to (1) identify environmentally suitable areas into which focal species could potentially spread; (2) generate a combined risk map for all the focal species and for the ten most harmful species in Europe; and (3) identify the main physicochemical characteristics of the areas at greatest risk. The results revealed that the potential distributions of nonnative species were best predicted by climatic factors, notably by temperature-related variables. Anthropogenic activity was also a major contributor to the distribution patterns of all the non-native species examined. Areas experiencing high levels of eutrophication, a phenomenon that is strongly associated with anthropogenic activity, were among those at greatest risk of invasions. The approach presented here was ACCEPTED MANUSCRIPT 2 intended to be broadly applicable. For example, it could be used to look at other taxonomic groups, regions, and/or systems. The overarching aim is to provide an effective basis for developing and implementing management and control strategies that can mitigate the effects of current invasions and prevent future ones. Keywords: biological invasions; exotic species; invasion risk; MaxEnt; modelling techniques; species distribution models 1. Introduction Freshwater ecosystems are considered to be major biodiversity hotspots (Strayer & Dudgeon, 2010). They are also one of the most threatened ecosystems in the world (Collen et al., 2014; Dudgeon et al., 2006). Biological invasions, together with habitat disturbance, are among the main causes of biodiversity decline in inland aquatic habitats (Sala et al., 2000; Simberloff et al., 2013). The arrival of non-native species has been facilitated by an increase in eutrophication worldwide, as well as by globalisation (Keller, Geist, Jeschke, & Kühn, 2011). Although substantial resources have been invested in biodiversity conservation, efforts in freshwater systems remain limited because information is lacking and it is challenging to predict invasion risks (Abell et al., 2010; Brundu, 2015). Over the past 20 years, researchers have been describing the ecological problems associated with the arrival and spread of non-native aquatic plants in Europe (Brundu, 2015; Keller et al., 2011; Sheppard, Shaw, & Sforza, 2006; Willby, 2007). Non-native aquatic plants have reached the continent via different introduction pathways. Aquariums and garden ponds have played an important role (Keller et al., 2011), as have waterbirds and boats (i.e., transport of stowaways) (Hulme, 2009; Reynolds, Miranda, Cumming, & Keller, 2015). Unfortunately, introduction pathways are so numerous, and even ubiquitous, that there is little hope of eliminating them completely (Strayer, 2010). Propagule pressure appears to be key to introduction success (the propagule pressure hypothesis), and, consequently, anthropogenic activity is strongly correlated with the presence of non-native species (Hulme, 2009; Pysek & Richardson, 2010). ACCEPTED MANUSCRIPT 3 Non-native aquatic species are, in general, largely successful because they can tolerate a wide range of environmental conditions, and their establishment is facilitated when there is climatic similarity between the native and invaded ranges (the habitat filtering hypothesis; Gallien, Münkemüller, Albert, Boulangeat, & Thuiller, 2010; Melbourne et al., 2006). Other key contributing factors are their high degree of phenotypic plasticity and the absence of natural enemies in invaded areas (adaptation and enemy release hypotheses; Duncan & Williams, 2002; Ren & Zhang, 2009). Non-native aquatic species have negative impacts on natural biodiversity (Ricciardi & Kipp, 2007). Because they can alter ecosystems by modifying habitat structure, substrate, and water composition (Strayer, 2010), non-native species represent a threat to ecosystems and native aquatic species (Havel, Kovalenko, Thomaz, Amalfitano, & Kats, 2015; Simberloff et al., 2013). One of the most relevant consequences of such processes is that native species are displaced (Stiers, Crohain, Josens, & Triest, 2011). For these reasons, it is critical to predict where invasions are most likely to occur and when they will result in the successful establishment of non-native species. Such an early-warning system is important because ecologists know that eradication is usually effective only when efforts are implemented prior to the establishment of non-native species (Crafton, 2015): once non-natives are established in a new habitat, control efforts are expensive, and most species are essentially impossible to eradicate (Thouvenot, Haury, & Thiebaut, 2013). Species distribution models (SDMs) are increasingly being used to identify landscapes that most resemble those in which a given species is currently found. SDMs are based on environmental layers and species occurrence data (Guisan & Thuiller, 2005); they can work with information on a set of species and predict areas at risk of invasion (O’Donnell et al., 2011; Thuiller et al., 2005). For example, climatic variables are widely known to limit the distribution of species (Woodward & Williams, 1987) and remain one of the best predictors of distribution patterns at large spatial scales (Kelly, Leach, Cameron, Maggs, & Reid, 2014). Although few variables exist to model the distribution of inland aquatic species, terrestrial climatic variables are useful when modelling species distributions in inland aquatic environments (Reshetnikov & Ficetola, ACCEPTED MANUSCRIPT 4 2011). Novel approaches to SDMs have demonstrated the importance of including variables that reflect anthropogenic activity (Gallardo, Zieritz, & Aldridge, 2015). The human footprint is an index that helps quantify the effects of humans on the environment (Sanderson et al., 2002) and is correlated with factors that may influence the introduction and distribution of non-native species (Gallardo et al., 2015). These factors may include propagule pressure—which is related to the number of introduction routes and population density—and changes in land use and habitat transformation, which can both impact native biodiversity by leaving ecosystem gaps that can be exploited by non-native species (empty niche and opportunity windows hypotheses; Champion, Clayton, & Hofstra, 2010; Compton, De Winton, Leathwick, & Wadhwa, 2012; Hierro, Maron, & Callaway, 2004; Shea & Chesson, 2002). When the occupation of empty niches by non-native species is combined with different disturbance events (disturbance hypothesis), the entry of other non-native species may be facilitated, producing a domino effect known as invasion meltdown (Hood & Naiman, 2000; Mack, 2003). SDMs yield biogeographical and ecological data that allow the development and implementation of effective management strategies and that provide guidance on how to prioritise the deployment of limited resources (Gordon, Onderdonk, Fox, & Stocker, 2008; Havel et al., 2015; Vander Zanden & Olden, 2008). Although tools exist for establishing the areas in which the management and control of non-native species should be prioritised, no study has yet sought to analyse the broad spectrum of nonnative aquatic plants present across all of Europe. Only region or species-specific studies have been carried out (Gallardo & Aldridge, 2013b; Gallardo et al., 2015; Kelly et al., 2014; Rodríguez-Merino, Fernández-Zamudio, & García-Murillo, 2017). Here, we focused on the current distribution of non-native aquatic plants in Europe (in total 60 species). There were three main objectives: (1) to identify areas that could potentially be colonised by non-native aquatic plants using species-specific SDMs based on bioclimatic and socioeconomic variables; (2) to overlay these species-specific models to generate combined invasion risk maps, revealing the areas of Europe at greatest risk of multiple invasions—maps were created using all 60 species and using the ten most harmful species in Europe; and, (3) to identify the physicochemical characteristics of water bodies and sediments in these areas of greatest invasion risk. ACCEPTED MANUSCRIPT 5 2. Material and methods 2.1. Study area The models developed in this study are global in scale, and Europe was selected as the focal region. Over the centuries, Europe has been a center for international trade by different civilisations and, as a consequence, a large number of non-native species have become established there (Keller et al., 2011). Europe is of particular interest because it is one of the most threatened regions of the world as a result of the species introductions that mainly took place during the late twentieth century (Early et al., 2016). Furthermore, part of Europe falls within the Mediterranean basin, an important biodiversity hotspot whose conservation is crucial (Myers, Mittermeier, Mittermeier, da Fonseca, & Kent, 2000). Finally, both species invasions and habitat disturbance are major causes of biodiversity loss in this region (Sala et al., 2000). 2.2 Species occurrence data Sixty non-native aquatic plants that are present in Europe were modelled (Table A.1). The list came from Hussner (2012) and was modified in the following way: (1) species considered native in any part of Europe were eliminated; (2) species with low sample sizes were excluded to avoid model performance issues (n ≥ 10) (Papeş & Gaubert, 2007); and, (3) species with lower performance values in the SDMs (i.e., under 0.7) were eliminated (Elith & Leathwick, 2007). For each species, all available occurrence data in both the native and invasive ranges were included to avoid underestimating the fundamental niche (Jiménez- Valverde et al., 2011); this approach is the most common one (but see Bradley, Blumenthal, Wilcove, & Ziska 2010). These data were obtained from the Global Biodiversity Information Facility (GBIF, 2016). Only records from 1950 onwards were included in order to match the framework for current climate data (Gillard, Thiébaut, Deleu, & Leroy, 2017; Kelly et al., 2014). The data set was cleaned up by removing erroneous taxonomic occurrences, duplicates, and geographic outliers using R software (R Core Development Team, 2014). Additionally, spatial autocorrelation in the data was reduced to minimise problems with model overfitting (Boria, Olson, Goodman, & Anderson, 2014). The criterion was that the distance between data pairs should be less than 10 km; the same grid cell was used in the resolution of the predictor layers. By avoiding possible biases in data distributions, this filtering protocol allows occurrence ACCEPTED MANUSCRIPT 6 data from databases (e.g., the GBIF) to be effectively used to analyse macroecological patterns (García‐Roselló et al., 2015). 2.3. Predictor layers Nineteen different bioclimatic layers (Table A.2) were obtained from WorldClim-Global Climate Data (Hijmans, Cameron, Parra, Jones, & Jarvis, 2005; Worldclim, 2015). The variables chosen focused on temperature and precipitation and represent annual trends, seasonality, and extremes that are relevant to species survival and distribution (Hijmans et al., 2005). Climatic variables can be used as filters to delimit the potential distributions of non-native species on a coarse scale (Gallardo et al., 2015; Kelly et al., 2014). The resolution chosen here was 5 arc minutes (~10 km at the equator). The human footprint (HFP) was considered to be a suitable way of quantifying anthropogenic activity; it integrates different types of geographic data (Sanderson et al., 2002) and reflects land-use-related disturbances—such as urbanisation, communication routes, and farming—that could significantly influence the distribution of non-native aquatic plants (Gallardo & Aldridge, 2013b; Gallardo et al., 2015; Kelly et al., 2014). HFP data were obtained from the Socioeconomic Data and Applications Center (SEDAC, 2015). Since the resolution level was 30 arc seconds (~1 km), it was transformed to 5 arc minutes and projected using the World Geodetic System (WGS 1984) to achieve equivalence with the WorldClim layers. Dimensionality among variables was reduced using Pearson´s correlation coefficient analysis, which was performed with the raster package in R (Hijmans & van Etten, 2015). Variance inflation factors (VIFs) were used to assess collinearity among variables and to decide which ones to retain; the HH package in R was used (Heiberger, 2015). The number of variables was reduced to eight because the use of many layers, especially those associated with small sample sizes, could increase model overfitting (Heikkinen et al., 2006). The variables that remained in the model were annual mean temperature (Bio 1), temperature seasonality (Bio 4; standard deviation*100), minimum temperature in the coldest month (Bio 6), mean temperature in the warmest quarter (Bio 10), annual precipitation (Bio 12), precipitation seasonality (Bio 15; coefficient of variation), precipitation in the driest quarter (Bio 17), and HFP. Mean temperature and precipitation as well as their seasonal variation can influence species phenology (Forrest ACCEPTED MANUSCRIPT 7 & Miller-Rushing, 2010). The other variables represent different kinds of stress, such as cold stress (Bio 6), heat stress (Bio 10), and drought stress (Bio 17), which can place limits on species distributions (Godefroid, Cruaud, Rossi, & Rasplus, 2015). 2.4. Species distribution modelling The maximum entropy algorithm, MaxEnt version 3.3.3k, was used to generate niche models (Phillips, Anderson, & Schapire, 2006). The MaxEnt algorithm is one of the most effective presence-only data algorithms available and has been shown to perform well even when sample sizes are low and there are moderate georeferencing errors (Elith et al., 2006; Mateo et al., 2013; Phillips et al., 2006; Wisz et al., 2008). The SDMs developed in this study were global in scale, and Europe was selected as the focal region. The modelling parameters described in Phillips et al. (2006), Phillips and Dudík (2008), and Elith et al. (2011) were used. The maximum number of iterations was 1,000, and the number of background points was 10,000. To reduce the probability of model overfitting, the multiple regularisation parameter was changed to 2.5 from the default of 1 (Elith et al., 2010; Gallardo & Aldridge, 2013a; Rodríguez-Merino et al., 2017). The models were calibrated using 70% of the occurrence data, and the remaining 30% of the data was used to test the models obtained. In addition, for each model, a 10-fold cross-validation procedure was used to estimate the errors associated with the fitted functions and the predictive performance of the data that had been set aside (Elith et al., 2011). The output was the probability of presence for each species, which took on a value between zero and one (Phillips & Dudík, 2008). Model performance was assessed using the area under the receiver operating characteristic (ROC) curve (Hosmer & Lemeshow, 2000), or the area under the curve (AUC) for short. Calculating the AUC is one of the most common methods for evaluating presence-only data models (Merow, Smith, & Silander, 2013). This metric reflects a model’s ability to discriminate among suitable and unsuitable habitats (Phillips et al., 2006). The AUC was calculated for each model, and the mean AUC was then determined for a set of 10 replicates to obtain a more robust estimate of predictive performance (Barnes et al., 2014). 2.5. Invasion risk maps ACCEPTED MANUSCRIPT 8 The 10th percentile training presence threshold was used to transform continuous maps into presence/absence maps (Jiménez-Valverde & Lobo, 2007; Liu, Berry, Dawson, & Pearson, 2005). This threshold was chosen because it does well at correctly predicting the presence of non-native species and representing species distributions in suboptimal habitats (Kelly et al., 2014; Pearson, Raxworthy, Nakamura, & Peterson, 2007; Reshetnikov & Ficetola, 2011). The highly conservative estimate of this threshold yields more significant and ecologically relevant results (Jarnevich & Reynolds, 2011). A first invasion risk map was created by overlaying the presence/absence maps for the 60 focal species to produce stacked SDMs (Crafton, 2015; Gallardo et al., 2015; Thuiller et al., 2005); ArcGIS was used (ESRI, 2008). This map revealed the cumulative risk of invasion, and also highlighted those areas where the presence of aquatic invasive species is expected to be higher and, as a consequence, the places where invasion risk is expected to be higher. A second invasion risk map was created that just focused on the ten invasive species considered to be most harmful to aquatic environments in Europe by the European Parliament and the European Council (Commission Implementing Regulations EU 2016/1141 and EU 2017/1263 [Table 1]). Table 1 near here The invasion risk maps were then filtered such that only cells containing water bodies were retained (Gillard et al., 2017). The layer containing lakes, reservoirs, and wetlands was obtained from Lehner and Döll (2004), and the layer containing main rivers and tributaries came from the European Environment Agency (European Environment Agency, 2018). Layers were converted to raster format; cells of 5 arc minutes (i.e., study resolution) that contained at least one water body were selected. 2.6. Generalised additive model analysis The relationships between physicochemical variables and invasion risk in Europe were analysed for the 60 focal species using generalised additive models (GAMs). GAM fitting allows predictor variables to be combined using non-specific functions, which is a helpful feature that is not available in other regression techniques (Wood, 2008). This analysis allowed the detection of the factors that prevail in areas at higher risk of invasion, which helps identify the environmental features that are facilitating non-native aquatic plant invasions. ACCEPTED MANUSCRIPT 15 are usually the result of geological conditions (Frick & Norvell, 1984), industrial and household waste may also contribute (Salminen et al., 2005). DOC levels are highest in regions with humic environments, which are not suitable for most of the focal species (Salminen et al., 2005). These regions are mostly located in northern Europe (northern Great Britain, northern Ireland, and Scandinavia), where invasion risk is lower. They are also high in areas where agricultural waste, organic fertilisers, and manure are abundant—the resulting carbon makes its way into nearby water bodies, increasing DOC concentrations (Molinero & Burke, 2009). High DOC levels can have various ecological consequences; for example, they can increase water turbidity or reduce rates of photosynthesis and respiration (Steinberg et al., 2006), limiting the occurrence of aquatic plants. Of particular importance is the use of carbon for photosynthesis, which is a limiting factor in the growth of aquatic plants, because it is generally found in low concentrations in freshwater ecosystems. For this reason aquatic plants are forced to acquire a series of changes to maximise growth rates. Among these mechanisms of carbon acquisition are the development of aerial or floating leaves (morphological changes), or the use of crassulacean acid metabolism, or the C4 metabolism found in the genus Hydrilla or Egeria (physiological and biochemical changes) (Maberly & Madsen, 2002). For example, the use of HCO3- as a carbon acquisition mechanism is the most widespread strategy in aquatic systems of high alkalinity, this fact could explain the affinity of certain species for alkaline rich freshwaters and the difference in the geographic distribution of certain invaders (Maberly & Madsen, 2002). Taken together, our results suggest that the most at-risk areas are those with increased levels of eutrophication, as well as those in which propagule pressure is high due to the greater number of introduction pathways. The high degree of disturbance at such locations provides a colonisation opportunity for non-native species (Nilsson & Berggren, 2000) because they are more likely to invade modified or degraded aquatic habitats than more natural habitats (Bunn & Arthington, 2002; Dudgeon et al., 2006). Although SDMs are a common tool for determining the potential distributions of non-native species, they should still be used with care. It is important to note that areas at low risk of invasion on the map could be considered by some authors to be "coldspots" (e.g., Kelly et al., 2014). This interpretation should be made with caution because species distribution models assume niche conservatism over time and space ACCEPTED MANUSCRIPT 16 (Peterson, 2011). Indeed, species’ complete niches are frequently underestimated (Fletcher, Gillingham, Britton, Blanchet, & Gozlan, 2016; Tingley, Vallinoto, Sequeira, & Kearney, 2014) because assessments of habitat suitability are based exclusively on the variables included in the models. The accessibility of other, less favourable areas when more suitable areas are not available is not taken into account. This caveat is important because, given climate change, more areas may become suitable for nonnative species in the not-too-distant future (Gillard et al., 2017; Havel et al., 2015; Hussner et al., 2010; Sheppard et al., 2006). Due to invasive species are typically generalists and have the capacity to tolerate broad climatic conditions (Walther et al., 2009), the effect of climate change is one of the factors that could affect the geographic distribution of invasive species in the future (Hellmann, Byers, Bierwagen, & Dukes, 2008). Increasing temperatures and alterations in flow regimes—caused by precipitation-based changes to surface water quality (Whitehead, Wilby, Battarbee, Kernan, & Wade, 2009)—are expected to alter the potential distributions of different species (Bellard et al., 2013), creating a northward trend (Gillard et al., 2017). These changes could also be one of the main drivers of future biological invasions (Bellard et al., 2013). In the case of aquatic plant species, distribution patterns can be limited by factors such as dispersal barriers, historical biogeography, or biotic relationships (Pont, Hugueny, & Oberdorff, 2005). However, when humans act as vectors, such constraints may become irrelevant (Gallardo & Aldridge, 2013b). In any case, in the type of research undertaken in this study, it is preferable to overestimate potential species distributions (Jiménez-Valverde et al., 2011). Record number is another important element to consider when modelling the potential distributions of non-native species since it could affect the ability to characterise habitat suitability. Past studies have estimated the minimum number of records needed and have found that a greater number of records will lead to more precise models (Hernandez, Graham, Master, & Albert, 2006; Wisz et al., 2008). The number of records used in this study allowed a good level of precision. Increasingly, models of potential species distributions are employing water-body filtering (Gillard et al., 2017). We decided to show both general maps and filtered maps ACCEPTED MANUSCRIPT 17 because layers containing information on water body surfaces remain incomplete. This fact means that valuable information related to habitat suitability is unavailable. For example, Mediterranean temporary ponds, small streams, and artificial water bodies such as irrigation channels and irrigation pools do not currently appear on continentalscale maps; they do, however, potentially provide suitable habitat for invasive species. Such species include E. crassipes and Gymnocoronis spilanthoides, which have both been detected in irrigation channels in Italy (Ardenghi, Barcheri, Ballerini, Cauzzi, & Guzzon, 2016; Brundu et al., 2013) or Pistia stratiotes in the South of Iberian Peninsula (García-Murillo, Dana, & Rodríguez, 2005). Although the data used in the GAMs are reliable, they must be used with caution. First, these data were obtained via interpolation from measurements made at sampling stations distributed throughout Europe. However, they represent annual means (Salminen et al., 2005) and do not reflect any interannual variation or seasonality in the variables of interest. It is also important to recognise that interpolation precision will depend on variable range, which could lead to unrealistic results (Gallardo & Aldridge, 2013a). That said, we feel that these data suffice for a preliminary assessment of the physicochemical characteristics that prevail in the areas facing the greatest invasion risk. In contrast, we do not think it would be appropriate to use these data in the SDMs because data are not available for the species’ entire distribution ranges (native and invasive); indeed, it is recommended that information on the native range be used when modelling of potential distributions (Jiménez-Valverde et al., 2011). 5. Conclusions Our findings have helped identify areas in Europe that are at higher risk of invasion by non-native aquatic plants; we mapped the potential invasion risk by examining the influence of anthropogenic activity and climatic variables on non-native aquatic species colonisation and spread. In general, we wish to highlight the importance of paying attention to areas with mild winters and/or high levels of anthropogenic activity, such as southwestern Europe or the British Channel. These results can guide decision making, both by those managing non-native aquatic plant species and by those responsible for preserving biodiversity in aquatic systems (Strayer, 2010). Indeed, ACCEPTED MANUSCRIPT 18 studies like ours should make it possible to evaluate invasion risks even in remote and poorly studied areas (Hespanhol, Cezón, Felicísimo, Muñoz, & Mateo, 2015). Using modelling techniques to predict the early stages of colonisation and, more importantly, to anticipate the spread of non-native species could be key to efforts aimed at managing and preserving the natural environment (Liu et al., 2011; Thuiller et al., 2005). Acknowledgements We thank Professor Amanda D. Rodewald of Cornell University for her helpful comments on the manuscript. We would like to thank the two anonymous reviewers and Cathal O’Mahony for their helpful comments, which improved considerably the manuscript. The authors have no conflicts of interest. References Abell, R., Thieme, M., Ricketts, T. H., Olwero, Ng, R., Petry, P., … Hoekstra, J. (2010). Concordance of freshwater and terrestrial biodiversity. Conservation Letters, 4(2), 127–136. https://doi.org/10.1111/j.1755-263X.2010.00153.x Akaike, H. (1974). A new look at the statistical model identification. IEEE Transactions on Automatic Control, 19(6), 716–723. https://doi.org/10.1109/TAC.1974.1100705 Alahuhta, J., Heino, J., & Luoto, M. (2011). Climate change and the future distributions of aquatic macrophytes across boreal catchments. Journal of Biogeography, 38(2), 383–393. https://doi.org/10.1111/j.1365-2699.2010.02412.x Araújo, Ferri‐Yáñez, F., Bozinovic, F., Marquet, P. A., Valladares, F., Chown, S. L., & Sax, D. (2013). Heat freezes niche evolution. Ecology Letters, 16(9), 1206– 1219. https://doi.org/10.1111/ele.12155 Ardenghi, N. M. G., Barcheri, G., Ballerini, C., Cauzzi, P., & Guzzon, F. (2016). Gymnocoronis spilanthoides (Asteraceae, Eupatorieae), a new naturalized and potentially invasive aquatic alien in S Europe. Willdenowia, 46(2), 265–273. https://doi.org/10.3372/wi.46.46208 Barnes, M. A., Jerde, C. L., Wittmann, M. E., Chadderton, W. L., Ding, J., Zhang, J., … Lodge, D. M. (2014). Geographic selection bias of occurrence data influences transferability of invasive Hydrilla verticillata distribution models. Ecology and Evolution, 4(12), 2584–2593. https://doi.org/10.1002/ece3.1120 Bellard, C., Leroy, B., Thuiller, W., Rysman, J. ‐F., Courchamp, F., & Collins, S. (2016). Major drivers of invasion risks throughout the world. Ecosphere, 7(3), e01241. https://doi.org/10.1002/ecs2.1241 ACCEPTED MANUSCRIPT 19 Bellard, C., Thuiller, W., Leroy, B., Genovesi, P., Bakkenes, M., & Courchamp, F. (2013). Will climate change promote future invasions? Global Change Biology, 19(12), 3740–3748. https://doi.org/10.1111/gcb.12344 Boria, R. A., Olson, L. E., Goodman, S. M., & Anderson, R. P. (2014). Spatial filtering to reduce sampling bias can improve the performance of ecological niche models. Ecological Modelling, 275, 73–77. https://doi.org/10.1016/j.ecolmodel.2013.12.012 Bradley, B. A., Blumenthal, D. M., Wilcove, D. S., & Ziska, L. H. (2010). Predicting plant invasions in an era of global change. Trends in Ecology & Evolution, 25(5), 310–318. https://doi.org/10.1016/j.tree.2009.12.003 Brundu, G. (2015). Plant invaders in European and Mediterranean inland waters: profiles, distribution, and threats. Hydrobiologia, 746(1), 61–79. https://doi.org/10.1007/s10750-014-1910-9 Brundu, G., Azzella, M. M., Blasi, C., Camarda, I., Iberite, M., & Celesti-Grapow, L. (2013). The silent invasion of Eichhornia crassipes (Mart.) Solms. in Italy. Plant Biosystems - An International Journal Dealing with All Aspects of Plant Biology, 147(4), 1120–1127. https://doi.org/10.1080/11263504.2013.861536 Bunn, S. E., & Arthington, A. H. (2002). Basic principles and ecological consequences of altered flow regimes for aquatic biodiversity. Environmental Management, 30(4), 492–507. https://doi.org/10.1007/s00267-002-2737-0 Capers, R. S., Selsky, R., Bugbee, G. J., & White, J. C. (2009). Species richness of both native and invasive aquatic plants influenced by environmental conditions and human activity. Botany, 87(3), 306–314. https://doi.org/10.1139/B08-144 Catford Jane A., & Downes Barbara J. (2010). Using multi‐scale species distribution data to infer drivers of biological invasion in riparian wetlands. Diversity and Distributions, 16(1), 20–32. https://doi.org/10.1111/j.1472-4642.2009.00631.x Champion, P. D., Clayton, J. S., & Hofstra, D. E. (2010). Nipping aquatic plant invasions in the bud: weed risk assessment and the trade. Hydrobiologia, 656(1), 167–172. https://doi.org/10.1007/s10750-010-0446-x Collen, B., Whitton, F., Dyer, E. E., Baillie, J. E. M., Cumberlidge, N., Darwall, W. R. T., … Böhm, M. (2014). Global patterns of freshwater species diversity, threat and endemism. Global Ecology and Biogeography, 23(1), 40–51. https://doi.org/10.1111/geb.12096 Compton, T. J., De Winton, M., Leathwick, J. R., & Wadhwa, S. (2012). Predicting spread of invasive macrophytes in New Zealand lakes using indirect measures of human accessibility. Freshwater Biology, 57(5), 938–948. https://doi.org/10.1111/j.1365-2427.2012.02754.x Crafton, R. E. (2015). Modeling invasion risk for coastal marine species utilizing environmental and transport vector data. Hydrobiologia, 746(1), 349–362. https://doi.org/10.1007/s10750-014-2027-x Davis, M. A., Grime, J. P., & Thompson, K. (2001). Fluctuating resources in plant communities: a general theory of invasibility. Journal of Ecology, 88(3), 528– 534. https://doi.org/10.1046/j.1365-2745.2000.00473.x Dudgeon, D., Arthington, A. H., Gessner, M. O., Kawabata, Z.-I., Knowler, D. J., Lévêque, C., … Sullivan, C. A. (2006). Freshwater biodiversity: importance, threats, status and conservation challenges. Biological Reviews of the Cambridge Philosophical Society, 81(2), 163–182. https://doi.org/10.1017/S1464793105006950 ACCEPTED MANUSCRIPT 20 Duncan, R. P., & Williams, P. A. (2002). Ecology: Darwin’s naturalization hypothesis challenged. Nature, 417(6889), 608–609. https://doi.org/10.1038/417608a Early, R., Bradley, B. A., Dukes, J. S., Lawler, J. J., Olden, J. D., Blumenthal, D. M., … Tatem, A. J. (2016). Global threats from invasive alien species in the twentyfirst century and national response capacities. Nature Communications, 7, 12485. https://doi.org/10.1038/ncomms12485 Elith, J., Kearney, M., & Phillips, S. (2010). The art of modelling range‐shifting species. Methods in Ecology and Evolution, 1(4), 330–342. https://doi.org/doi:10.1111/j.2041-210X.2010.00036.x Elith, J., & Leathwick, J. (2007). Predicting species distributions from museum and herbarium records using multiresponse models fitted with multivariate adaptive regression splines -. Diversity and Distributions, 13(3), 265–275. Elith, J., Phillips, S. J., Hastie, T., Dudík, M., Chee, Y. E., & Yates, C. J. (2011). A statistical explanation of MaxEnt for ecologists. Diversity and Distributions, 17(1), 43–57. https://doi.org/10.1111/j.1472-4642.2010.00725.x Elith, Jane, Graham, C. H., Anderson, R. P., Dudík, M., Ferrier, S., Guisan, A., … Zimmermann, N. E. (2006). Novel methods improve prediction of species’ distributions from occurrence data. Ecography, 29(2), 129–151. https://doi.org/10.1111/j.2006.0906-7590.04596.x ESRI. (2008). ArcGIS Desktop: Release 9.3 Redlands, CA: Environmental Systems Research Institute. European Environment Agency. (2018). WISE Large rivers and large lakes. Available at: http://www.eea.europa.eu (last accessed 17 March 2018). Fernández-Zamudio, R., Cirujano, S., Sánchez-Carrillo, S., Meco, A., & García- Murillo, P. (2013). Clonal reproduction of Azolla filiculoides Lam.: implications for invasiveness. Limnetica, 32(2), 245–252. Ficetola, G. F., Thuiller, W., & Miaud, C. (2007). Prediction and validation of the potential global distribution of a problematic alien invasive species — the American bullfrog. Diversity and Distributions, 13(4), 476–485. https://doi.org/10.1111/j.1472-4642.2007.00377.x Fletcher, D. H., Gillingham, P. K., Britton, J. R., Blanchet, S., & Gozlan, R. E. (2016). Predicting global invasion risks: a management tool to prevent future introductions. Scientific Reports, 6, 26316. https://doi.org/10.1038/srep26316 Forrest, J., & Miller-Rushing, A. J. (2010). Toward a synthetic understanding of the role of phenology in ecology and evolution. Philosophical Transactions of the Royal Society B: Biological Sciences, 365(1555), 3101–3112. https://doi.org/doi:10.1098/rstb.2010.0145 Frick, C. R., & Norvell, W. A. (1984). Chemical and physical properties of Connecticut lakes. Bulletin / Connecticut Agricultural Experiment Station (USA). Retrieved from http://agris.fao.org/agris-search/search.do?recordID=US875012488 Gallardo, B., & Aldridge, D. C. (2013a). Priority setting for invasive species management: risk assessment of Ponto-Caspian invasive species into Great Britain. Ecological Applications: A Publication of the Ecological Society of America, 23(2), 352–364. Gallardo, B., & Aldridge, D. C. (2013b). The ‘dirty dozen’: socio‐economic factors amplify the invasion potential of 12 high‐risk aquatic invasive species in Great Britain and Ireland. Journal of Applied Ecology, 50(3), 757–766. https://doi.org/10.1111/1365-2664.12079 ACCEPTED MANUSCRIPT 21 Gallardo, B., Zieritz, A., & Aldridge, D. C. (2015). The importance of the human footprint in shaping the global distribution of terrestrial, freshwater and marine invaders. PloS One, 10(5), e0125801. https://doi.org/10.1371/journal.pone.0125801 Gallien, L., Münkemüller, T., Albert, C. H., Boulangeat, I., & Thuiller, W. (2010). Predicting potential distributions of invasive species: where to go from here? Diversity and Distributions, 16(3), 331–342. https://doi.org/10.1111/j.1472- 4642.2010.00652.x García‐Roselló, E., Guisande, C., Manjarrés‐Hernández, A., González‐Dacosta, J., Heine, J., Pelayo‐Villamil, P., … Lobo, J. M. (2015). Can we derive macroecological patterns from primary Global Biodiversity Information Facility data? Global Ecology and Biogeography, 3(24), 335–347. https://doi.org/10.1111/geb.12260 GBIF. (2016). Global Biodiversity Information Facility. Available at: http://www.gbif.org (last accessed 20 September 2016). Gillard, M., Thiébaut, G., Deleu, C., & Leroy, B. (2017). Present and future distribution of three aquatic plants taxa across the world: decrease in native and increase in invasive ranges. Biological Invasions, 19(7), 2159–2170. https://doi.org/10.1007/s10530-017-1428-y Godefroid, M., Cruaud, A., Rossi, J.-P., & Rasplus, J.-Y. (2015). Assessing the risk of invasion by tephritid fruit flies: intraspecific divergence matters. PLOS ONE, 10(8), e0135209. https://doi.org/10.1371/journal.pone.0135209 Gordon, D. R., Onderdonk, D. A., Fox, A. M., & Stocker, R. K. (2008). Consistent accuracy of the Australian weed risk assessment system across varied geographies. Diversity and Distributions, 14(2), 234–242. https://doi.org/10.1111/j.1472-4642.2007.00460.x Guisan, A., & Thuiller, W. (2005). Predicting species distribution: offering more than simple habitat models. Ecology Letters, 8(9), 993–1009. https://doi.org/10.1111/j.1461-0248.2005.00792.x Hastie, T., & Tibshirani, R. (1990). Exploring the nature of covariate effects in the proportional hazards model. Biometrics, 46(4), 1005–1016. Havel, J. E., Kovalenko, K. E., Thomaz, S. M., Amalfitano, S., & Kats, L. B. (2015). Aquatic invasive species: challenges for the future. Hydrobiologia, 750(1), 147– 170. https://doi.org/10.1007/s10750-014-2166-0 Heiberger, R. J. (2015). Statistical analysis and data display: Heiberger and Holland. R package version 3.1-23. Heikkinen, R. k., Luoto, M., Araújo, M. B., Virkkala, R., Thuiller, W., & Sykes, M. T. (2006). Methods and uncertainties in bioclimatic envelope modelling under climate change. Progress in Physical Geography, 30(6), 751–777. Hellmann, J. J., Byers, J. E., Bierwagen, B. G., & Dukes, J. S. (2008). Five potential consequences of climate change for invasive species. Conservation Biology, 22(3), 534–543. https://doi.org/10.1111/j.1523-1739.2008.00951.x Hernandez, P. A., Graham, C. H., Master, L. L., & Albert, D. L. (2006). The effect of sample size and species characteristics on performance of different species distribution modeling methods. Ecography, 29(5), 773–785. https://doi.org/10.1111/j.0906-7590.2006.04700.x ACCEPTED MANUSCRIPT 22 Hespanhol, H., Cezón, K., Felicísimo, Á. M., Muñoz, J., & Mateo, R. G. (2015). How to describe species richness patterns for bryophyte conservation? Ecology and Evolution, 5(23), 5443–5455. https://doi.org/10.1002/ece3.1796 Hierro, J. L., Maron, J. L., & Callaway, R. M. (2004). A biogeographical approach to plant invasions: the importance of studying exotics in their introduced and native range. Journal of Ecology, 93(1), 5–15. https://doi.org/10.1111/j.0022- 0477.2004.00953.x Hijmans, R. J., Cameron, S. E., Parra, J. L., Jones, P. G., & Jarvis, A. (2005). Very high resolution interpolated climate surfaces for global land areas. International Journal of Climatology, 25(15), 1965–1978. https://doi.org/10.1002/joc.1276 Hijmans, R. J., & van Etten, J. (2015). Raster: Geographic analysis and modeling with raster data. R package version 3.1.23. Hood, W. G., & Naiman, R. J. (2000). Vulnerability of riparian zones to invasion by exotic vascular plants. Plant Ecology, 148(1), 105–114. https://doi.org/10.1023/A:1009800327334 Horlings, L. G., & Marsden, T. K. (2011). Towards the real green revolution? Exploring the conceptual dimensions of a new ecological modernisation of agriculture that could ‘feed the world.’ Global Environmental Change, 21(2), 441–452. https://doi.org/10.1016/j.gloenvcha.2011.01.004 Hosmer, D. W., & Lemeshow, S. (2000). Applied Logistic Regression (2nd ed.). Hulme, P. E. (2009). Trade, transport and trouble: managing invasive species pathways in an era of globalization. Journal of Applied Ecology, 46(1), 10–18. https://doi.org/10.1111/j.1365-2664.2008.01600.x Hussner, A. (2012). Alien aquatic plant species in European countries. Weed Research, 52(4), 297–306. Hussner, A., Van de Weyer, K., Gross, E. M., & Hilt, S. (2010). Comments on increasing number and abundance of non‐indigenous aquatic macrophyte species in Germany. Weed Research, 50(6), 519–526. https://doi.org/10.1111/j.1365-3180.2010.00812.x Hutchinson, G. E. (1975). A treatise on limnology. Vol. 3. John Wiley, New York, NY. Janes, R. (1998). Growth and survival of Azolla filiculoides in Britain I. Vegetative production. New Phytologist, 138(2), 367–375. https://doi.org/doi:10.1046/j.1469-8137.1998.00114.x Jarnevich, C. S., & Reynolds, L. V. (2011). Challenges of predicting the potential distribution of a slow-spreading invader: a habitat suitability map for an invasive riparian tree. Biological Invasions, 13(1), 153–163. https://doi.org/10.1007/s10530-010-9798-4 Jiménez-Valverde, A., Peterson, A. T., Soberón, J., Overton, J. M., Aragón, P., & Lobo, J. M. (2011). Use of niche models in invasive species risk assessments. Biological Invasions, 13(12), 2785–2797. https://doi.org/10.1007/s10530-011- 9963-4 Jiménez-Valverde, A., & Lobo, J. M. (2007). Threshold criteria for conversion of probability of species presence to either–or presence–absence. Acta Oecologica, 31(3), 361–369. https://doi.org/10.1016/j.actao.2007.02.001 Keller, R. P., Geist, J., Jeschke, J. M., & Kühn, I. (2011). Invasive species in Europe: ecology, status, and policy. Environmental Sciences Europe, 23(1), 23. https://doi.org/10.1186/2190-4715-23-23 ACCEPTED MANUSCRIPT 23 Kelly, R., Leach, K., Cameron, A., Maggs, C. A., & Reid, N. (2014). Combining global climate and regional landscape models to improve prediction of invasion risk. Diversity and Distributions, 20(8), 884–894. https://doi.org/10.1111/ddi.12194 Kriticos, D. J., & Brunel, S. (2016). Assessing and managing the current and future pest risk from water hyacinth, (Eichhornia crassipes), an invasive aquatic plant threatening the environment and water security. PLOS ONE, 11(8), e0120054. https://doi.org/10.1371/journal.pone.0120054 Lehner, B., & Döll, P. (2004). Development and validation of a global database of lakes, reservoirs and wetlands. Journal of Hydrology, 296(1), 1–22. https://doi.org/10.1016/j.jhydrol.2004.03.028 Liu, C., Berry, P. M., Dawson, T. P., & Pearson, R. G. (2005). Selecting thresholds of occurrence in the prediction of species distributions. Ecography, 28(3), 385– 393. https://doi.org/10.1111/j.0906-7590.2005.03957.x Liu, X., Guo, Z., Ke, Z., Wang, S., & Li, Y. (2011). Increasing potential risk of a global aquatic invader in Europe in contrast to other continents under future climate change. PLOS ONE, 6(3), e18429. https://doi.org/10.1371/journal.pone.0018429 Lowe, S. J., Browne, M., Boudjelas, S., & De Pooter, M. (2004). 100 of the World’s worst invasive species alien species: a selection from the global invasive species database. Lukács, B. A., Vojtkó, A. E., Mesterházy, A., Molnár, V. A., Süveges, K., Végvári, Z., … Cerabolini, B. E. L. (2017). Growth‐form and spatiality driving the functional difference of native and alien aquatic plants in Europe. Ecology and Evolution, 7(3), 950–963. https://doi.org/10.1002/ece3.2703 Maberly, S. C., & Madsen, T. V. (2002). Freshwater angiosperm carbon concentrating mechanisms: processes and patterns. Functional Plant Biology, 29(3), 393–405. https://doi.org/10.1071/pp01187 Mack, R. N. (2003). Phylogenetic constraint, absent life forms, and preadapted alien plants: A Prescription for Biological Invasions. International Journal of Plant Sciences, 164(S3), S185–S196. https://doi.org/10.1086/368399 Mateo, R. G., de la Estrella, M., Felicísimo, Á. M., Muñoz, J., & Guisan, A. (2013). A new spin on a compositionalist predictive modelling framework for conservation planning: A tropical case study in Ecuador. Biological Conservation, 160, 150– 161. https://doi.org/10.1016/j.biocon.2013.01.014 Melbourne, B. A., Cornell, H. V., Davies, K. F., Dugaw, C. J., Elmendorf, Freestone, A. L., … Yokomizo, H. (2006). Invasion in a heterogeneous world: resistance, coexistence or hostile takeover? Ecology Letters, 10(1), 77–94. https://doi.org/10.1111/j.1461-0248.2006.00987.x Merow, C., Smith, M. J., & Silander, J. A. (2013). A practical guide to MaxEnt for modeling species’ distributions: what it does, and why inputs and settings matter. Ecography, 36(10), 1058–1069. https://doi.org/10.1111/j.1600- 0587.2013.07872.x Molinero, J., & Burke, R. A. (2009). Effects of land use on dissolved organic matter biogeochemistry in piedmont headwater streams of the Southeastern United States. Hydrobiologia, 635(1), 289–308. https://doi.org/10.1007/s10750-009- 9921-7 Myers, N., Mittermeier, R. A., Mittermeier, C. G., da Fonseca, G. A. B., & Kent, J. (2000). Biodiversity hotspots for conservation priorities. Nature, 403(6772), 853–858. https://doi.org/10.1038/35002501 ACCEPTED MANUSCRIPT 24 Nilsson, C., & Berggren, K. (2000). Alterations of Riparian Ecosystems Caused by River RegulationDam operations have caused global-scale ecological changes in riparian ecosystems. How to protect river environments and human needs of rivers remains one of the most important questions of our time. BioScience, 50(9), 783–792. https://doi.org/10.1641/0006- 3568(2000)050[0783:AORECB]2.0.CO;2 O’Donnell, J., Gallagher, R. V., Wilson, P. D., Downey, P. O., Hughes, L., & Leishman, M. R. (2011). Invasion hotspots for non‐native plants in Australia under current and future climates. Global Change Biology, 18(2), 617–629. https://doi.org/10.1111/j.1365-2486.2011.02537.x Papeş, M., & Gaubert, P. (2007). Modelling ecological niches from low numbers of occurrences: assessment of the conservation status of poorly known viverrids (Mammalia, Carnivora) across two continents. Diversity and Distributions, 13(6), 890–902. https://doi.org/10.1111/j.1472-4642.2007.00392.x Pearson, R. G., Raxworthy, C. J., Nakamura, M., & Peterson, A. T. (2007). Predicting species distributions from small numbers of occurrence records: a test case using cryptic geckos in Madagascar. Journal of Biogeography, 34(1), 102–117. Peterson, A. T. (2011). Ecological niche conservatism: a time‐structured review of evidence. Journal of Biogeography, 38(5), 817–827. https://doi.org/10.1111/j.1365-2699.2010.02456.x Phillips, S. J., Anderson, R. P., & Schapire, R. E. (2006). Maximum entropy modeling of species geographic distributions. Ecological Modelling, 190(3), 231–259. https://doi.org/10.1016/j.ecolmodel.2005.03.026 Phillips, S. J., & Dudík, M. (2008). Modeling of species distributions with Maxent: new extensions and a comprehensive evaluation. Ecography, 31(2), 161–175. https://doi.org/10.1111/j.0906-7590.2008.5203.x Pont, D., Hugueny, B., & Oberdorff, T. (2005). Modelling habitat requirement of European fishes: do species have similar responses to local and regional environmental constraints? Canadian Journal of Fisheries and Aquatic Sciences, 62(1), 163–173. https://doi.org/10.1139/f04-183 Pysek, P., & Richardson, D. M. (2010). Invasive Species, Environmental Change and Management, and Health. Annual Review of Environment and Resources, 35(1), 25–55. Quinn, L. D., Schooler, S. S., & Van Klinken, R. D. (2010). Effects of land use and environment on alien and native macrophytes: lessons from a large‐scale survey of Australian rivers. Diversity and Distributions, 17(1), 132–143. https://doi.org/10.1111/j.1472-4642.2010.00726.x R Core Development Team. (2014). R: A language and environment for statistical computing. R Foundation for statistical computing, Viena, Austria. Available at: http://www.R-project.org. Ren, M., & Zhang, Q. (2009). The relative generality of plant invasion mechanisms and predicting future invasive plants. Weed Research, 49(5), 449–460. https://doi.org/10.1111/j.1365-3180.2009.00723.x Reshetnikov, A. N., & Ficetola, G. F. (2011). Potential range of the invasive fish rotan (Perccottus glenii) in the Holarctic. Biological Invasions, 13(12), 2967–2980. https://doi.org/10.1007/s10530-011-9982-1 ACCEPTED MANUSCRIPT 31 ACCEPTED MANUSCRIPT 32 Table 1 The ten most harmful non-native aquatic plants in Europe according to the European Parliament and the European Council and conveyed by the following Commission Implementing Regulations (CIRs): (1) EU CIR 2016/1141 of 13 July 2016 adopting a list of invasive alien species of Union concern pursuant to EU Regulation No. 1143/2014 of the European Parliament and of the Council and (2) EU CIR 2017/1263 of 12 July 2017 updating the list of invasive alien species of Union concern established by EU CIR 2016/1141 pursuant to EU Regulation No. 1143/2014 of the European Parliament and of the Council. Species CIR Alternanthera philoxeroides (Mart.) Griseb. 2 Cabomba caroliniana A. Gray 1 Eichhornia crassipes (Mart.) Solms 1 Elodea nuttallii (Planch.) H. St. John 2 Hydrocotyle ranunculoides L. f. 1 Lagarosiphon major (Ridl.) Moss 1 Ludwigia grandiflora (Michx.) Greuter & Burdet 1 Ludwigia peploides (Kunth) P. H. Raven 1 Myriophyllum aquaticum (Vell.) Verdc. 1 Myriophyllum heterophyllum Michx. 2 ACCEPTED MANUSCRIPT 33 Table 2 Contribution of different variables when building the SDMs. The table shows the percent contribution of variable groups (temperature, precipitation, and HFP = human footprint) and the percent contribution of individual variables (Bio 1 = annual mean temperature; Bio 4 = temperature seasonality [standard deviation * 100]; Bio 6 = minimum temperature in the coldest month; Bio 10 = mean temperature in the warmest quarter; Bio 12 = annual precipitation; Bio 15 = precipitation seasonality [coefficient of variation], and Bio 17 = precipitation in the driest quarter). Variable groups Variable Percent contribution (%) Standard deviation Temperature 46.44 16.51 Bio 1 16.04 13.33 Bio 4 16.70 14.57 Bio 6 8.89 8.83 Bio 10 4.81 5.28 Precipitation 31.64 16.34 Bio 12 11.94 12.32 Bio 15 5.51 6.51 Bio 17 14.19 14.07 HFP 21.92 13.52 ACCEPTED MANUSCRIPT 34 Table 3 Estimates from the GAMs analysing the relationships between six physicochemical variables and invasion risk for the 60 focal species. DOC = dissolved organic carbon. R2 (adj) = measure of model fit. Deviance (%) = percentage of variance in the data explained by the model. AIC = Akaike Information Criterion. Significance = ***; p≤0.001. Model Variables p-value R2 (adj) Deviance (%) AIC Nitrate *** 0.337 39.5 126906.1 Phosphorus *** 0.015 1.37 17629.4 Sulphate *** 0.134 15.7 157721.2 DOC *** 0.145 15.3 158236.0 Conductivity *** 0.281 32.6 135873.4 Alkalinity *** 0.294 31.9 136778.2 Full model *** 0.429 48.6 115150.1 Nitrate *** Phosphorus *** Sulphate *** DOC *** Conductivity *** Alkalinity *** ACCEPTED MANUSCRIPT 35 Appendix A Fig. A.1 Presence/absence maps for the ten most harmful species in Europe. In black are the areas in which the species is present, based on the suitability threshold. A. Alternanthea philoxeroides. B. Cabomba caroliniana. C. Eichhornia crassipes. D. Elodea nuttallii. E. Hydrocotyle ranunculoides. F. Lagarosiphon major. G. Ludwigia grandiflora. H. Ludwigia peploides. I. Myriophyllum aquaticum. J. Myriophyllum heterophyllum. Fig. A.2 Presence/absence maps for ten most harmful species in Europe in which cells have been filtered by water-body presence. In black are the areas in which the species is present, based on the suitability threshold. A. Alternanthea philoxeroides. B. Cabomba caroliniana. C. Eichhornia crassipes. D. Elodea nuttallii. E. Hydrocotyle ranunculoides. F. Lagarosiphon major. G. Ludwigia grandiflora. H. Ludwigia peploides. I. Myriophyllum aquaticum. J. Myriophyllum heterophyllum. ACCEPTED MANUSCRIPT 36 ACCEPTED MANUSCRIPT 37 ACCEPTED MANUSCRIPT 38 ACCEPTED MANUSCRIPT 39 Table A.1 List of focal species. The table shows the number of records (N) used to build the individual models, the AUC values (mean ± standard deviation) for MaxEnt model performance, and the percent contribution of variables to species-specific models. Bio 1 = annual mean temperature; Bio 4 = temperature seasonality [standard deviation * 100]; Bio 6 = minimum temperature in the coldest month; Bio 10 = mean temperature in the warmest quarter; Bio 12 = annual precipitation; Bio 15 = precipitation seasonality [coefficient of variation], Bio 17 = precipitation in the driest quarter; Human Footprint = HFP. The most important variable for each species is in bold. Species N AUC±SD Bio01 Bio04 Bio06 Bio10 Bio12 Bio15 Bio17 HFP Alternanthera philoxeroides (Mart.) Griseb. 220 0.963 ± 0.008 18.996 9.434 10.905 1.778 4.130 1.011 29.500 24.246 Aponogeton distachyos L.f. 122 0.987 ± 0.004 13.224 34.765 13.894 9.394 0.788 12.136 5.963 9.837 Azolla caroliniana Willd. 152 0.932 ± 0.010 8.649 5.968 17.431 5.567 6.215 2.546 25.209 28.415 Azolla filiculoides Lam. 1,497 0.929 ± 0.003 23.238 31.333 14.534 0.878 1.180 7.577 1.020 20.241 Bacopa monnieri (L.) Wettst. 480 0.933 ± 0.004 8.181 14.887 38.362 3.829 0.940 1.802 3.430 28.570 Cabomba caroliniana A. Gray 160 0.959 ± 0.010 11.909 12.564 0.152 0.486 9.642 3.821 35.663 25.765 Callitriche deflexa A. Braun ex. Hegelm. 25 0.965 ± 0.018 10.498 52.658 1.926 0.000 0.345 0.029 5.167 29.377 Ceratopteris thalictroides (L.) Brongn. 322 0.946 ± 0.008 18.103 1.852 4.202 30.270 27.611 4.238 5.893 7.831 Crassula helmsii (Kirk) Cockayne 966 0.961 ± 0.002 23.330 31.293 10.736 0.443 0.156 23.568 8.908 1.566 Egeria densa Planch. 278 0.954 ± 0.005 8.394 8.483 13.851 3.396 1.602 1.001 13.163 50.111 Eichhornia crassipes (Mart.) Solms 704 0.914 ± 0.005 18.483 2.569 21.979 3.192 14.841 1.447 4.008 33.481 Elodea callitrichoides (Rich.) Casp. 28 0.996 ± 0.002 1.215 1.931 12.543 7.741 0.198 12.718 14.735 48.920 Elodea canadensis Michx. 928 0.932 ± 0.002 37.405 7.395 1.728 11.101 2.251 3.405 18.391 18.323 Elodea nuttallii (Planch.) H. St. John 1,741 0.932 ± 0.004 18.278 7.971 5.277 0.775 0.387 21.675 18.988 26.647 Gymnocoronis spilanthoides (D. Don ex. Hook. & Arn) DC. 70 0.985 ± 0.004 19.762 11.707 3.964 0.595 15.337 2.629 36.070 9.936 Heteranthera limosa (Sw.) Willd. 260 0.950 ± 0.009 31.217 5.596 1.310 6.763 13.160 8.605 11.037 22.312 Heteranthera reniformis Ruiz & Pav. 345 0.953 ± 0.004 4.499 16.681 14.272 5.671 20.204 1.935 3.701 33.038 ACCEPTED MANUSCRIPT 40 Heteranthera rotundifolia (Kunth.) Griseb. 164 0.960 ± 0.007 31.363 3.856 1.622 7.130 21.461 11.862 9.445 13.262 Heteranthera zosterifolia Mart. 21 0.921 ± 0.033 0.545 30.401 0.833 0.000 50.316 1.861 1.732 14.312 Hydrilla verticillata (L. f.) Royle 421 0.930 ± 0.008 33.646 13.333 1.393 4.907 22.433 2.230 9.675 12.384 Hydrocotyle bonariensis Lam. 346 0.959 ± 0.009 2.645 36.965 15.401 1.487 2.826 0.620 14.059 25.999 Hydrocotyle moschata G. Forst 91 0.994 ± 0.001 15.439 39.021 1.769 3.023 0.017 0.537 37.538 2.656 Hydrocotyle ranunculoides L.f. 380 0.939 ± 0.012 7.438 3.908 30.075 14.641 5.713 1.037 0.296 36.893 Hydrocotyle sibthorpioides Lam. 863 0.954 ± 0.004 21.962 4.587 33.256 0.068 4.221 0.352 34.811 0.742 Hydrocotyle verticillata Thunb. 297 0.937 ± 0.011 36.030 14.506 22.693 1.816 1.211 3.536 4.978 15.231 Hygrophila polysperma (Roxb.) T. Anderson 15 0.927 ± 0.033 5.927 31.155 11.659 0.305 0.218 0.110 0.123 50.503 Lagarosiphon major (Ridl.) Moss 572 0.975 ± 0.002 3.147 16.634 17.706 14.656 1.418 17.270 8.579 20.590 Landoltia punctata (G. Mey) Les & D. J. Crawford 80 0.944 ± 0.017 12.126 14.551 28.894 0.311 4.501 4.904 5.065 29.649 Lemna aequinoctialis Welw. 362 0.915 ± 0.011 55.542 11.274 0.909 4.752 18.286 1.501 2.858 4.879 Lemna minuta Kunth. 1,125 0.950 ± 0.004 17.372 17.085 4.350 4.128 0.654 23.790 6.701 25.921 Lemna perpusilla Torr. 55 0.927 ± 0.047 13.904 18.065 4.618 8.610 30.488 1.585 1.272 21.459 Lemna turionifera Landolt 286 0.959 ± 0.005 61.085 5.077 0.770 2.288 6.329 5.161 6.469 12.821 Ludwigia grandiflora (Michx.) Greuter & Burdet 170 0.980 ± 0.004 5.632 12.277 13.212 0.851 0.495 16.679 15.084 35.771 Ludwigia peploides (Kunth) P. H. Raven 374 0.932 ± 0.009 31.855 8.350 5.107 7.166 2.535 1.806 3.619 39.563 Murdannia keisak (Hassk.) Hand. Mazz. 96 0.979 ± 0.017 1.383 10.827 1.346 1.467 29.181 1.041 20.584 34.170 Myriophyllum aquaticum (Vell.) Verdc. 711 0.953 ± 0.005 16.862 7.206 20.472 1.778 2.384 0.260 17.916 33.123 Myriophyllum heterophyllum Michx. 138 0.974 ± 0.002 11.712 9.465 11.159 1.472 20.186 9.561 22.431 14.015 Myriophyllum verrucosum Lindl. 625 0.952 ± 0.003 8.458 57.887 7.303 0.095 6.859 3.799 14.274 1.325 Najas gracillima (A. Braun ex. Engelmann) Magnus 77 0.955 ± 0.013 2.109 18.850 0.701 2.440 23.842 26.752 1.579 23.727 Najas graminea Delile 98 0.932 ± 0.014 20.151 9.149 2.173 8.722 34.314 12.630 4.097 8.764 Najas guadalupensis (Spreng.) Magnus 484 0.925 ± 0.006 25.681 7.450 3.113 8.356 9.828 4.635 9.686 31.251 Nelumbo nucifera Gaertn. 126 0.957 ± 0.006 19.968 6.633 2.877 6.550 29.853 3.467 6.501 24.152 Nuphar advena (Aiton) W.T. Aiton 43 0.962 ± 0.017 1.757 7.713 3.129 0.026 0.295 3.336 50.175 33.567 Nymphaea lotus L. 14 0.932 ± 0.020 9.035 68.751 2.145 2.743 2.332 0.836 0.761 13.397 Nymphaea mexicana Zucc. 46 0.967 ± 0.025 9.651 23.999 2.236 0.060 0.648 0.272 0.525 62.609 ACCEPTED MANUSCRIPT