Species traits and invasion history as predictors of freshwater fish’s invasion success in Europe

doi:10.21203/rs.3.rs-2064811/v1

Download PDF

Research Article

Species traits and invasion history as predictors of freshwater fish’s invasion success in Europe

https://doi.org/10.21203/rs.3.rs-2064811/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Despite the growing literature on the topic of freshwater fish invasion, few studies employed a comprehensive analysis of the stages characterizing the invasion process (i.e. invasion pathway), thereby demanding a deeper knowledge to avoid incomplete and unbiased conclusions and give support to adequate management strategies. The aim of the present study was to provide a complete analysis of the species traits and invasion history leading an alien freshwater fish species to successfully pass through the invasion pathway stages in European river basins. To predict how likely a freshwater fish species moves from each stage of the invasion pathway - release, establishment, spread and impact - to the next one, Generalized Linear Mixed Models were run, using as predictors 23 functional and ecological species traits, and seven variables describing introduction history for 127 established alien freshwater fish species in Europe. Results showed that the release and spread stages were primarily driven by variables related to the invasion history (i.e. the type and number of causes of introduction). The establishment stage was mainly driven by functional and ecological traits, while the impactstage was driven both by functional and ecological traits and invasion history. By identifying the main drivers of alien species success at a given invasion stage, this study provides scientific knowledge to design target management actions to specific stages of the invasion thus contributing to control their populations more effectively.

aliens species

ecological traits

functional traits

invasion pathway

invasion stage

non-native species

Invasive Alien Species (IAS; sensu Regulation 1127/2014; European Union 2014) threaten native biota and ecosystems functioning worldwide from local to continental scale, negatively affecting ecosystems health and services (Vitule et al. 2009; Simberloff et al. 2013), while accounting for annual economic losses of billions of euros (Cuthbert et al. 2021).

Freshwater ecosystems are among the most threatened and impacted by biological invasions (CBD 2001; Reid et al. 2019). Competitive exclusion (Rowles and O’Dowd 2007) and predation (Griffen and Delaney 2007) by IAS towards native species are responsible for dramatic change in native communities (Crooks 1998). In European freshwater systems, introduced fish species contribute to taxonomic homogenization, with potential negative effects on ecosystem functions (Sommerwerk et al. 2017), as well as representing the second most common threat to native freshwater fish (Costa et al. 2021). Freshwater IAS have also been pointed to have a strong association with ecosystems degradation, with a higher proportion of water bodies invaded by IAS in the moderate to bad ecological status categories (i.e. based on the Water Framework Directive, WFD 2000/60/EC, EU, 2000; Vandekerkhove et al. 2013).

Introduced species become invasive when they successfully overcome barriers (environmental filters) related to different stages of the invasion pathway: transport, release, establishment and spread (Williamson 1996; Kolar and Lodge 2001; Sakai et al. 2001). A fifth stage – the impact stage - is also considered by some authors (e.g. García-Berthou 2007), although others claim that impacts are of more subjective nature and not necessarily an outcome preceded by the spread stage (Blackburn et al. 2011). Identifying the traits acting in each stage of the invasion pathway is critical to predict the invasive potential of alien species (Ribeiro et al. 2008) as well as to identify management targets to control their establishment and spread. Moreover, each stage should be examined as independent from one another, because each stage has been shown to be mediated by a certain combination of traits and a trait promoting the success of one stage may, in contrast, compromise the success of other stages (García-Berthou 2007).

The identification of freshwater fish traits involved in their invasion success, have been previously addressed by researchers, though most studies have been primarily focused on the establishment, spread and impact stages (e.g. Kolar and Lodge 2002; Marchetti et al. 2004a; Ribeiro et al. 2008; Howeth et al. 2016). The establishment stage is the most studied invasion stage for fish, and it has been found to be related with reproductive strategy, size and distance from the native range, diet breadth, environmental tolerances (García-Berthou 2007; Peoples and Midway 2018) and invasion history (e.g. previous invasion success in Marchetti et al. 2004a; Ribeiro et al. 2008). Similarly, the spread stage has been shown to be affected by previous invasion success, low parental care (Ribeiro et al. 2008) and larger adult size/longer lifespan (Moyle and Marchetti 2006, Ribeiro et al. 2008), while the size of the native range has led to opposite conclusions between different study areas: small size of native range in Moyle and Marchetti (2006), Ribeiro et al. (2008); large size of native range in Peoples and Midway (2018). Lastly, the impact stage has been shown to be driven by previous invasion success, physiological tolerance, non-omnivorous diet (Moyle and Marchetti 2006; Ribeiro et al. 2008) and small native range size (Ribeiro et al. 2008).

The transport and release stages are poorly documented and are thus difficult to address (Marchetti et al. 2004a, García-Berthou 2007). Indeed, records about stages that precede successful introductions and, in particular, unsuccessful introductions, are very limited or strongly biased (García-Berthou 2007). One step to overcome this data shortage would be obtaining more detailed information on propagule pressures that, along with abiotic factors and a comprehensive database of species traits, enables understanding a species’ selection through the several stages of the invasion process (Hayes and Barry 2008). Therefore, the assessment of the early stages of the invasion pathway is one of the research questions that need to be addressed in the context of invasive alien fishes (García-Berthou 2007; Buckwalter et al. 2020). Nonetheless, the first stage, transport, was not considered herein since there was no data available at the European scale.

This study investigates the combination of functional and ecological traits (sensu Violle et al. 2007; De Bello et al. 2021) and invasion history that best correlates with the progressive success of alien species along the invasion pathway in European rivers. Here, both IAS and alien freshwater species were considered. For this purpose, predictive models for each of the main invasion stages - release, establishment, spread and impact stages – were developed, using functional and ecological traits along with invasion history information of 127 alien fish species in Europe as predictor variables. The impact stage was included as the last stage, assuming that, despite being more adequately defined as an outcome, it is likely to be driven by the same set of drivers as the previous stages. The applied value of this study is twofold: first, it will support the definition of more effective management guidelines to prevent the success of alien fish species in freshwater habitats, by setting different operational targets (e.g. culling) depending on the stages that a particular alien species still has to endure; second, it will help predicting the future success of new alien fish species throughout the invasion pathways. To the best of authors knowledge, this study provides the most comprehensive approach to find associations between species traits and invasion success of alien freshwater fish species, by addressing the most complete sequence of invasion stages and accounting for a large number of traits at the continental scale.

2.1. Alien fish species

This study considers 127 alien fish species, belonging to 81 genera and 29 families that have been recorded as established in freshwater systems of one or more European countries (Fig. 1). Species were selected based on their establishment status retrieved from several sources [e.g. GISD; Pagad et al. 2018; Frose and Pauly 2020; all species traits, references and otherwise, are available in the OSF online repository https://osf.io/p3qkj/; Marcolin et al. (2022)]. A species was selected when it was reported as established in at least one European country. Whenever inconsistent information between two or more sources (i.e. one source reporting that the species is established in a country and the other source reporting as not established) were found, priority was given to the source reporting the newest reference. Scientific names follow the classification proposed by Fricke et al. (2020).

2.2. Fish species traits and invasion history

For each alien freshwater fish species, 23 species traits (nine functional traits and fourteen ecological traits), and seven variables describing the introduction history of the species were identified (Table 1). Data were retrieved from multiple sources including research studies, data sources and fish index manuals (Marcolin et al. 2022). Whenever inconsistent information between two or more sources were found, priority was given to the source reporting the newest reference. Furthermore, when the inconsistent information was reported by two references of the same year of publication, the “Precautionary Principle” (PP) approach (Cooney 2004) was followed: in face of inconsistent information on alien fish traits, the selected trait class or value was the one that potentially had higher impacts on native ecosystems following the introduction of the alien fish species. For example, since the adult trophic status of Vimba vimba is reported either as omnivore or invertivore by different sources, the PP approach led to classify the species’ diet as omnivore since it can be more impactful on the native fauna than an invertivore one (Oberdorff et al. 2002). Information was successfully retrieved for 90% (3417 out of 3810) of the trait records among all species (Supplementary Information Table S1). For 30% of the species (38 out of 127), all the traits were retrieved.

For many species traits, all the information required for each species was not available. In such cases, firstly it was attempted to consider the phylogenetic similarity (Troia and McManamay 2019), i.e. assigning a trait value, level or category of species belonging to a common genera or family with known traits. Whenever more than one of such species was available, the above PP approach was followed. In addition, whenever possible and needed, fish morphological characteristics were used to infer ecological traits [e.g. from the body shape of Cobitis bilineata was derived the feeding habitat as benthic; Chan 2001; Haas et al. 2010; Franssen 2011; Branco et al. 2013; Ingram 2015; all references for traits and traits values can be found in Marcolin et al. (2022)]. Among the resulting missing trait records, 90% (181 out of 201) were filled following the procedure based on the PP approach and retrieving information according to the phylogenetic similarity and morphological characteristics.

In case were not possible to fill the trait, trait levels were assigned to an intermediate trait score (e.g. the missing value for the Maximum fecundity of the Trinectes maculatus was assigned as ‘3’ out of five possible levels) (Branco et al. 2008). Only the remaining 10% (20 out of 201; 0.5% of the 3810 records) were filled through the assignment of records to the central class of the trait, based on the levels of that specific trait.

Five traits had a high level of missing data (Acid tolerance, Habitat degradation tolerance, Water quality tolerance general, Water quality tolerance oxygen, Water quality tolerance toxic), thus were aggregated into a single combined General tolerance trait. Following the PP approach, the level of the combined trait (tolerant, intermediate, or intolerant) resulted from the highest tolerance score found among the five tolerance traits. After aggregating the tolerance traits into a single variable, the completeness of the records increased to 94% (3101 out of 3302).

Finally, whenever the available information was in the form of interval values (e.g. Female maturity) the PP approach was also followed by selecting the value leading to a higher impact on native ecosystems (Marcolin et al. 2022).

2.2.1. Invasion pathway

Species traits selection/filtering was investigated for the release, establishment, spread and impact stages of the invasion pathway. To model the influence of the different species traits on the invasive potential of alien species, proxies of the four invasion stages were used as response variables (Marchetti et al. 2004a; Ribeiro et al. 2008). Specifically, (1) the propagule pressure as a proxy for the release stage; (2) the fraction of European countries with established populations where the species was released and where established populations have been reported as a proxy for the establishment stage; (3) the number of European countries with established populations as a proxy for the spread stage; and (4) the invasibility status as a proxy of the impact stage (Table 1). As a surrogate measure of the propagule pressure, the number of introduction events corresponding to the number of records of introduction reported of a species in European countries of the study area considered were used, retrieving the information from Froese and Pauly (2020) and AquaNIS database (http://www.corpi.ku.lt/databases/index.php/aquanis). This approach was based on the method used by McGregor et al. (2012), Procheş et al. (2012) and Peoples and Midway (2018) where the authors used, as a surrogate of propagule pressure, the total number of entries/abstracts of human use of a species in discipline-specific literature databases [e.g. “species name AND aquaculture” in Peoples and Midway (2018)]. The Fraction of European countries with established populations was calculated as the proportion between the number of European countries where a species was introduced and established, and the number of European countries in which the species was introduced but not necessarily established. In this way, this variable is a measure of establishment success that is independent from the total number of countries with established populations. The variable was based on the same sources used to assess where a species was established or not (see section 2.1 above). The invasibility status had a binary nature with 0 representing alien species for which there is no evidence of being invasive, and 1 representing IAS. A species was considered to be invasive whenever information was found (e.g. GISD, Pagad et al. 2018; Frose and Pauly 2020) that the species status is invasive in, at least, one country. However, most species were evaluated a priori, following the method of Vilizzi et al. (2019).

2.3. Data analysis

To avoid problems derived from multicollinearity among variables, a correlation tests was ran using the Pearson correlation coefficient. All variable pairs showing r > |0.7| were considered to be correlated. Variance Inflation Factor (VIF; Zuur et al. 2007) was also computed in a stepwise fashion by excluding at each step the variable with the highest VIF and computing the VIF again in the following step until all the VIF of the remaining variables were lower than 2.5 (Johnston et al. 2018). A correlation coefficient of 0.70 was found between maximum adult size and maximum lifespan and, based on the stepwise VIF, the maximum adult size was discarded from all models.

An empirical modelling approach based on Generalized Linear Mixed Models (GLMM) was followed, both to hierarchize the importance of each variable at different invasion stages and to build predictive models for each invasion stage. GLMM was used to account for the non-independence among traits induced by phylogenetic similarity (Ruesink et al. 2005), using taxonomic classification at the family level as a random factor. Different family distributions and link functions were used according to the type of response at each stage (Table 1): Gaussian errors and link = 1 (multiple linear regression) for the release stage (propagule pressure; continuous), Poisson errors and log link (Poisson regression) for the spread stage (Number of European countries with established populations; discrete), and binomial errors and logit link (logistic regression) for the establishment (Fraction of European countries with established populations; proportion) and impact stages (Invasive status; binary). All predictor variables coded at the ordinal scale were treated as numeric in the models. Along the sequence of analyses, the outcome of each stage was assumed to be dependent on the outcome of the previous stage(s). Hence, at each stage the effect of variables was tested while controlling for the effect of the previous stage by adding the response variables of all previous stages to the models of the subsequent stages as candidate independent variables. First, the probability of a species to be introduced (release stage), was modelled as measured by the propagule pressure, using functional and ecological traits as potential predictors, as well as three variables related to the introduction history (cause of introduction, number of causes of introduction and prior invasion success outside Europe; Table 2). Then, in a second model, the probability of a species to establish populations in Europe (establishment stage) was modelled using the Fraction of European countries with established populations, as the response variable. In this step, the same predictors were used but now adding also the response variable of the release stage (propagule pressure; Table 2). In the next modelling step, prior invasion success at European countries (spread stage), was modelled now adding also the response variable of the previous stage (Fraction of European countries with established populations) as a candidate predictor variable. Finally, the invasive status (impact stage) was used as the response variable (binary variable: invasive/non-invasive) adding the response variable of the spread stage (number of European countries with established populations). Nonetheless, the analysis of the impact stage was an exception: the response variable of the release stage (propagule pressure) was not included as a candidate variable in the model since its correlation with the spread stage response variable was higher than the established threshold (r = 0.9).

For model selection, a multimodel inference approach (Burnham and Anderson 2002) was followed. To select a reduced subset of candidate variables, firstly was ran a backward variable selection based on the Akaike’s Information Criterion (AIC) for each response variable (four stages; as similarly done in Anderle et al. 2022). Based on this reduced subset of variables, a set of candidate models (Supplementary Information Tables S2 to S5) was defined following two main a priori assumptions: (1) the success of alien species at each invasion stage always depend on the introduction history and the outcome of the preceding stages; (2) any other variable (functional and ecological traits) was allowed to enter the model, as far it did not exceed a total of five variables in each candidate model, to avoid model overfitting as well as to ease the setting of a number of candidate models that should not be greater than the number of cases (n=127; Burnham and Anderson 2002). The candidate models were ranked according to the Akaike’s Information Criterion. Models with ΔAICc < 2 (difference between model’s AICc and the lowest AICc) were used for model averaging (Bolker et al. 2009; Richards 2008), in which the resulting variable coefficients resulted from averaging the coefficients of the models containing the variable (conditional averaging). The relative importance of the selected variables was assessed using the probability of each variable to be included in the best approximating models, estimated by summing the Akaike’s weights of all candidate models where the variable was included (Burnham and Anderson 2002). To estimate the goodness-of-fit of models, the marginal and conditional R² were computed: the first representing the variance explained by the fixed effects and the second representing the variance explained by the entire model (Nakagawa and Schielzeth 2013). When the final model resulted from averaging the best approximating models, a weighted average of the models’ R² was computed, in which the weights corresponded to the respective Akaike weights (wi).

All the analyses were performed with R software, version 4.0.0 (R Core Team, 2020). The vifstep function of the usdm R package (Naimi 2015) was used to compute the VIF stepwise procedure. GLMM were performed using lme4 R package (Bates et al. 2007). Multimodel inference was performed with the MuMIn R package (Bartoń 2016).

3.1 Release stage model

The release stage model showed a moderate support (wi=0.35, marginal R²=71%). The model resulted from averaging the two candidate models with the highest Akaike weight. All variables had coefficients with confidence intervals that did not contain zero. Migratory behaviour, parental care, the prior invasion success outside Europe, the number of causes of introduction, and the cause of introduction were the most important predictors for alien species to be released (Table 3-4, Supplementary Information Figure S1). In particular, the prior invasion success outside Europe and the number of causes of introduction showed strong positive effects on propagule pressure. According to the dummy variable coefficients of the cause of introduction variable, alien species with release records that were identified as accidental or both accidental and intentional tended to show a lower propagule pressure than species with exclusively intentional releases. Both the migratory behaviour and parental care showed a significant negative effect on the propagule pressure at this stage.

3.3 Establishment stage model

The establishment stage model showed a good support (wi=0.49, marginal R²=53%). This model resulted from averaging the two candidate models with the highest Akaike weight. Among the selected quantitative variables, three - maximum lifespan, spawning habitat preferences and maximum fecundity – had a negative effect while two - temperature tolerance and European origin (European native) - had a positive effect (Table 3, Supplementary Information Figure S2). All the coefficients had confidence intervals not containing zero. Moreover, alien species with release records that were identified as accidental, intentional and through dispersion tended to show a higher establishment probability than species with exclusively intentional releases.

3.4 Spread stage model

The spread stage model showed a good support (wi=0.49, marginal R²=71%). No model averaging was required since the second best approximating model showed a ΔAIC higher than 2. All quantitative variables included in the model showed a positive effect on the number of European countries with established populations (Table 3, Supplementary Information Figure S3). These included four variables: spawning frequency, number of causes of introduction, propagule pressure (release stage response variable) and fraction of European countries with established populations (establishment stage response variable). All had coefficients with confidence intervals that did not include zero. According to the dummy variable coefficients of the cause of introduction variable (Table 3), alien species with release records that were identified as both intentional and through dispersal tended to be established in more European countries than species with exclusively intentional releases.

3.5 Impact stage model

The resulting impact stage model showed a moderate support (wi=0.24, marginal R²=84%). No model averaging was required since the second best approximating model showed a ΔAIC higher than 2. All quantitative explanatory variables had coefficients with confidence intervals that did not contain zero. These included two variables showing a positive effect – maximum lifespan and number of European countries with established populations (spread stage response variable) – and two variables showing a negative effect – migratory behaviour and female maturity (Table 3, Supplementary Information Figure S4). For climate zone (dummy variable), the categories whose coefficient confidence intervals did not include zero were tropical, temperate/subtropical/tropical and temperate/subtropical/cold (Table 3), which means that species belonging to these trait categories tended to have a higher invasiveness than species restricted to the temperate zone (reference category).

This study supports the idea that successful established alien freshwater fish species, which eventually become invasive, tend to show a different combination of characteristics (traits) though the invasion pathway (Ribeiro et al. 2008). This was true even after accounting for the effect of the variables related with the history of alien species introductions. Moreover, by incorporating the probability of success at the previous stages, which is an innovative aspect, this study isolates the traits that drive the species' success at each stage of the invasion pathway. Traits tend to be differently related to each stage of invasion of alien fish in Europe, as well as the relative importance of alien fish introduction history. Both release and spread stages tend to be more related with the introduction history of the species, such as the causes that lead to their introduction or invasion success outside European borders, while the establishment and invasion stages tend to be primarily related with functional traits.

4.1 Traits filtering at each invasion stage

The traits that were found associated with the success of alien fish species throughout the invasion pathway in European rivers, were also shown to have similar effects at different spatial scales and biogeographic contexts. Among these traits, the results stress the importance of three traits: maximum lifespan, migratory behaviour and European native (native from Europe or not). The maximum lifespan was also found in other studies to have a significant effect on the alien fish invasion stages (Marchetti et al. 2004a; Jeschke and Strayer 2006; Liu et al. 2017; Peoples and Midway 2018). In these studies, the trait was found to be important for the establishment, spread and impact stages, most often with a positive correlation with the probability of success in each stage. Maximum lifespan has likely an indirect effect on species’ success at different invasion stages. For example, fishes with longer lifespan typically attain larger maximum body sizes (not considered here as candidate predictor due to collinearity problems), which confers enhanced dispersal abilities and possible increased competitive advantages. Maximum body size is often referred to as an important trait for the establishment stage (Jeschke and Strayer 2006; Moyle and Marchetti,2006; Liu et al. 2017; Peoples and Midway 2018). However, studies differ in their conclusions about the effect of body size and/or lifespan on alien fish invasion success, which found negative (Marchetti et al. 2004a; Ruesink 2005; Vila-Gispert et al. 2005; Moyle and Marchetti 2006), positive (Jeschke and Strayer 2006; Ribeiro et al. 2008; both larger body size and longevity: Liu et al. 2017) or no significant effects on the colonization and spread of alien fish (Buckwalter et al. 2020). Lifespan (and body size) may influence the success of alien fish species in two distinct directions. On one hand, species with shorter lifespan and small body size often share a set of life-history characteristics related to high reproductive capacity, including a short generation time and higher population growth rates (Ruesink 2005), which may favour establishment success. On the other hand, species with longer lifespan tend to show higher fecundity (Marchetti et al. 2004a) and are usually preferred for recreational angling (major trait of introduction in non-native areas; Davis and Darling 2017), commercial fisheries and aquaculture, all of which may favour their success as invasive species. This is supported by the results of this study, which suggest that short-living species tend to be more successful during the establishment stage, while long-living species tend to be more successful in the impact stage. Migratory behaviour, according to the models, tends to be related with a lower success during the release and impact stages. A similar trend was found in alien fish from Californian basins (Moyle and Marchetti 2006). The reasoning behind the most often unsuccessful introductions of migratory fish is probably related to their typical homing behaviour and the difficulties they face by their varying physiological and habitat requirements at different life history stages, as well as from the high mortality they face during migration (Moyle and Marchetti 2006). The European native variable was the third showing a significant effect in more than one stage. European native species were found to have lower chances of being released, although they had a higher probability of successfully establishing populations in non-native European countries. Therefore, results suggest that the number of release records is lower for European species than for non-European species, although the European species are more successfully established once they are released, probably better adapted to similar European climatic conditions. This variable could be interpreted as a proxy of distance to the nearest native source/region, which was found in other studies to be important in the establishment stage (Marchetti et al. 2004b; Ribeiro et al. 2008).

Other traits significantly affected the success of alien freshwater fish species along the invasion pathway, despite their effect being limited to a single invasion stage. This was the case of female age at maturity, parental care, spawning frequency, spawning habitat preferences, temperature tolerance and climatic zone of the native range. A negative association of female maturity trait at the impact stage was found, despite a positive association with maximum lifespan, suggesting that fish with long lifespans but early female maturity are favoured at this stage. Moreover, parental care, another trait related to reproduction, was negatively associated with propagule pressure at the release stage. This result is consistent with the trend found in Iberian Peninsula, where arguably alien species would benefit from having similar life-history strategies to native Iberian species, which typically show low parental care (Ribeiro et al. 2008). Nonetheless, an opposite trend was found in other studies, where high parental care was found to favour fish invasions (Marchetti et al. 2004a; Jeschke and Strayer 2006; Moyle and Marchetti 2006; Drake 2007; Buckwalter et al. 2020). The spawning frequency was also found to be a relevant factor during the spread stage. Accordingly, in other parts of the world, such as North America, fish species that spawn more frequently per season were found to colonize a wider range of non-native watersheds (Peoples and Midway 2018). At the Iberian Peninsula, a lower spawning frequency along the year was more frequently found among successful fish invaders originated outside the Iberian Peninsula, while higher spawning frequency were associated with translocated Iberian fish species (Vila-Gispert et al. 2005). Spawning habitat preference was found in this study to be one of the most important traits for the establishment of alien fish in Europe. The models show that limnophilic spawners tend to establish populations more successfully in opposition to rheophilic spawners. Traits related to habitat characteristics are however often disregarded in the literature that investigate relationships between fish invasiveness and functional traits. In one of the few exceptions, it was shown that most alien fish from Catalonian watersheds tend to prefer lower and more lentic stream reaches, although this refers to the general habitat preferences (Vila-Gispert et al. 2005). Among the traits related to tolerance, temperature tolerance was the only to show a significant effect at the establishment stage. As expected, eurythermal species tended to establish populations more successfully. This is consistent with most studies, which often found that successful fish invaders showed higher thermal tolerance (Kolar and Lodge 2002; Peoples and Midway 2018) or higher physiological tolerance (Marchetti et al. 2004a; Moyle and Marchetti 2006; García-Berthou 2007). Finally, the climate zone of provenance, which reflects the latitudinal extent of the native species range, was found to be positively related with the invasion success during the impact stage. This variable is related with the size of the native range that has been more commonly addressed in other studies and was found to be an important characteristic across different invasion stages (Marchetti et al. 2004a; Jeschke and Strayer 2006; Moyle and Marchetti 2006; Ribeiro et al. 2008; Peoples and Midway 2018; Buckwalter et al. 2020).

4.2 Invasion history as driver of alien fish invasion

This work supports previous studies, which suggested a strong positive association of human-related and invasion history characteristics on the success of alien fish species throughout the invasion pathway (Kolar and Lodge 2002; Marchetti et al. 2004a; Jeschke and Strayer 2006; Ribeiro et al. 2008; Peoples and Midway 2018). These characteristics were estimated as the most important among the several candidate variables for most stages. The cause of introduction was found to significantly affect two invasion stages, namely the release and spread stages, with species introduced accidentally or both accidentally and intentionally having less probability to be introduced or spread than only intentionally introduced species. This characteristic was also found to be important for the worldwide success of alien fish (Ruesink 2005). As expected, the number of causes of introduction also influenced the introduction success, also during the release and the spread stages. The Prior invasion success outside Europe, was found to significantly promote the success during the release stage. Other studies reported the importance of this characteristic for the success of both establishment and the spread of alien fish species (Marchetti et al. 2004a; Ribeiro et al. 2008).

As expected, the proxies of one stage were good predictors of the stages that followed. For example, the propagule pressure, here used as a proxy of release probability, was found to be positively associated to the spread of alien fish and additionally to be among the most important predictor variables during this stage. This variable is often referred to as one of the most positively related with the invasion success (Lockwood et al. 2005), namely during the establishment stage, as found in other studies addressing alien fish (Marchetti et al. 2004a; Ruesink 2005; Ribeiro et al. 2008). The establishment probability, as measured by the fraction of countries with reported releases where confirmed established populations were recorded, was positively associated with the spread stage and found to be an important predictor variable during this stage. Finally, the spread stage, as measured by the number of European countries with established populations, was the most important variable affecting the invasion probability during the impact stage.

4.3. Management recommendations

Release stage was found to be mostly governed by human-related variables, namely the introduction history, thus actions to manage alien species should mainly focus on controlling intentional releases of alien fish, especially for species with a known introduction success outside European borders. In the following establishment stage, management actions are more limited since the human component is not the most important in determining the success of alien species. Nonetheless, the results suggest that control actions should be primarily focused on European-origin fish species with shorter lifespan, low maximum fecundity, limnophilic spawning habits and high thermal tolerance. Culling, using different removal techniques, is a good alternative to eradicate alien fish species at this stage, as this method has been shown to be particularly effective in the very initial stages of invasion, especially for species with low dispersal abilities (Dominguez-Almela et al. 2021). The spread stage was again mostly governed by introduction history and human interest, although also positively associated with fish that showed higher yearly spawning frequency. At this stage, actions to manage alien species should mainly focus on controlling the number of causes of introductions, namely intentional releases and dispersion among basins. Nonetheless, avoiding their success on the previous two stages, in particular for species with higher spawning frequency, would be more effective since, in the initial period of invasion process, alien fish species are still spatially constrained and in low abundances (Rytwinski et al. 2019). Finally, based on the results found for the impact stage, focusing control actions on species that succeeded in the spread stage, would be recommended, in particular for non-migratory fish showing longer lifespans, lower female age at maturity and occupying wider latitudes in their native ranges. At this stage, to increase eradication probability or to constrain invasion (for invasive species with high dispersal ability), long-term management control actions (i.e. culling) should be implemented (Dominguez-Almela et al. 2021).

Funding

This research was supported by FCT – Fundação para a Ciência e a Tecnologia project “Alien Trade: something in the way they move: How global pattern of wildlife trade influence global invasion success” (PTDC/BIA-ECO/30931/2017-POCI-01-0145-FEDER-030931). FM, JR and JS were funded by Fundo Europeu de Desenvolvimento Regional through the Operational Competitiveness Factors Program “COMPETE” and by National Funds through the Foundation for Science and Technology (FCT) within the framework of Alien Trade project (PTDC/BIA-ECO/30931/2017-POCI-01-0145-FEDER-030931). PB has been financed by National Funds through the Foundation for Science and Technology (FCT) I.P. under “Norma Transitória – DL 57/2016/CP1382/CT0020”. LR was supported by Portuguese National Funds through FCT, I.P. (DL57/2016/CP1440/CT0011 and CEECIND/00445/2017, respectively). The Forest Research Centre (CEF) is a research unit funded by FCT (UIDB/00239/2020). The Associate Laboratory TERRA (LA/P/0092/2020) is also funded by FCT.

Conflict of Interest

The authors declare no conflict of interest

Author Contribution

All authors contributed to the study conception and design. FM, PB, JMS and PS collected the data. FM and PS analysed the data with input from all authors. LR and PS provided logistical and financial support. FM and PS wrote the manuscript with contributions from all authors. All authors read and approved the final manuscript.

Data availability statement

The complete dataset used in this study is available in the OSF online repository at: https://osf.io/p3qkj/

Anderle M, Paniccia C, Brambilla M, Hilpold A, Volani S, Tasser E et al (2022) The contribution of landscape features, climate and topography in shaping taxonomical and functional diversity of avian communities in a heterogeneous Alpine region. Oecologia 199:499–512
Coastal Research and Planning Institute (2012) AquaNIS. http://www.corpi.ku.lt/databases/index.php/aquanis. Accessed 20 March 2020
Bartoń K, (2016) MuMIn: Multi-model Inference. R Package Version 1.15.6.
Bates D, Sarkar D, Bates MD, Matrix L (2007) The lme4 package. R package version, 2(1), 74.
Blackburn TM, Pyšek P, Bacher S, Carlton JT, Duncan RP, Jarošík V et al (2011). A proposed unified framework for biological invasions. Trends Ecol Evol 26:333-339.
Bolker BM, Brooks ME, Clark CJ, Geange SW, Poulsen JR, Stevens MHH et al (2009) Generalized linear mixed models: a practical guide for ecology and evolution. Trends Ecol Evol 24:127-135.
Branco P, Costa JL, de Almeida PR (2008) Conservation priority index for estuarine fish (COPIEF). Estuar Coast Shelf Sci 80:581-588.
Branco P, Santos JM, Katopodis C, Pinheiro A, Ferreira MT (2013) Pool-type fishways: two different morpho-ecological cyprinid species facing plunging and streaming flows. PloS one 8:e65089.
Buckwalter JD, Frimpong EA, Angermeier PL, Barney JN (2020) Species traits predict stream‐fish invaders in an Appalachian (USA) river basin. Freshw Biol 65:552-564.
Burnham KP, Anderson DR (2002) Model Selection and Multimodel Inference: A Practical Information–Theoretic Approach, second ed. Berlin, Germany: Springer.
CBD (2001) Global Biodiversity Outlook. Montreal, Canada
Chan MD (2001) Fish ecomorphology: predicting habitat preferences of stream fishes from their body shape. Dissertation, Faculty of the Virginia Polytechnic Institute and State University.
Cooney R (2004) The precautionary principle in biodiversity conservation and natural resource management: an issues paper for policy-makers, researchers and practitioners (No. 2). IUCN.
Costa MJ, Duarte G, Segurado P, Branco P (2021) Major threats to European freshwater fish species. Sci Total Environ 797:149105.
Crooks JA (1998) Habitat alteration and community-level effects of an exotic mussel, Musculista senhousia. Mar Ecol Prog Ser 162:137-152.
Cuthbert RN, Pattison Z, Taylor NG, Verbrugge L, Diagne C, Ahmed DA et al (2021) Global economic costs of aquatic invasive alien species. Sci Total Environ 775:145238.
Davis AJS, Darling JA (2017) Recreational freshwater fishing drives non‐native aquatic species richness patterns at a continental scale. Divers Distrib 23(6):692-702.
De Bello F, Carmona CP, Dias AT, Götzenberger L, Moretti M, Berg MP (2021) Handbook of trait-based ecology: from theory to R tools. Cambridge, UK
Dominguez Almela V, Palmer SC, Andreou D, Gillingham PK, Travis JM, Britton JR (2021) Predicting the outcomes of management strategies for controlling invasive river fishes using individual‐based models. J Appl Ecol 58:2427-2440.
Drake JM (2007) Parental investment and fecundity, but not brain size, are associated with establishment success in introduced fishes. Funct Ecol 21:963-968.
European Commission (2014) Commission Regulation (EU) No 1143/2014 of the European Parliament and of the Council of 22 October 2014 on the prevention and management of the introduction and spread of invasive alien species. Bruxelles, Belgium
Franssen NR (2011) Anthropogenic habitat alteration induces rapid morphological divergence in a native stream fish. Evol Appl 4:791-804.
Freyhof J, Kottelat M (2007) Handbook of European freshwater fishes. Cornol,Switzerland
Fricke R, Eschmeyer WN, Van der Laan R (2020) Eschmeyer’s catalog of fishes: genera, species, references. http://researcharchive.calacademy.org/research/ichthyology/catalog/fishcatmain.asp Accessed 10 February 2020
Froese R. Pauly D eds (2020) FishBase. World Wide Web electronic publication. www.fishbase.org Accessed July 2020
García‐Berthou E (2007) The characteristics of invasive fishes: what has been learned so far?. J Fish Biol 71:33-55.
Global Register of Introduced and Invasive Alien Species n.d. http://www.griis.org Accessed 20 November 2019
Griffen BD, Delaney DG (2007) Species invasion shifts the importance of predator dependence. Ecology 88: 3012-3021.
Haas TC, Blum MJ, Heins DC (2010) Morphological responses of a stream fish to water impoundment. Biol Lett 6:803-806.
Hayes KR, Barry SC (2008) Are there any consistent predictors of invasion success?. Biol Invasions 10: 483-506.
Howeth JG, Gantz CA, Angermeier PL, Frimpong EA, Hoff MH, Keller RP et al (2016) Predicting invasiveness of species in trade: climate match, trophic guild and fecundity influence establishment and impact of non‐native freshwater fishes. Divers Distrib 22:148-160.
Ingram T (2015) Diversification of body shape in Sebastes rockfishes of the north-east Pacific. Biol J Linn Soc 116:805-818.
Invasive Species Specialist Group ISSG (2015) The Global Invasive Species Database. Version 2015.1 http://www.iucngisd.org/gisd/ Accessed on 10 March 2020.
Jeschke JM, Strayer DL (2006) Determinants of vertebrate invasion success in Europe and North America. Glob Chang Biol 12:1608-1619.
Johnston RJ, Jones K, Manley D (2018) Confounding and collinearity in regression analysis: a cautionary tale and an alternative procedure, illustrated by studies of British voting behaviour. Qual Quant 52:1957-1976.
Kolar CS, Lodge DM (2001) Progress in invasion biology: predicting invaders. Trends Ecol Evol 16:199-204.
Kolar CS, Lodge DM (2002) Ecological predictions and risk assessments for non-native species. Science 298:1233-1236.
Liu C, Comte L, Olden JD (2017) Heads you win, tails you lose: Life‐history traits predict invasion and extinction risk of the world's freshwater fishes. Aquat Conserv 27:773-779.
Lockwood JL, Cassey P, Blackburn T (2005) The role of propagule pressure in explaining species invasions. Trends Ecol Evol 20:223-228.
Marchetti MP, Moyle PB (2001) Effects of flow regime on fish assemblages in a regulated California stream. Ecol appl 11:530-539.
Marchetti MP, Moyle PB, Levine R (2004a). Invasive species profiling? Exploring the characteristics of non‐native fishes across invasion stages in California. Freshw Biol 49:646-661.
Marchetti MP, Moyle PB, & Levine, R. (2004b). Alien fishes in California watersheds: characteristics of successful and failed invaders. Ecol appl 14:587-596.
Marcolin F, Branco P, Santos JM, Reino L, Santana J, Ribeiro J, Segurado P (2022) Dataset for “Species traits and invasion history as predictors of freshwater fish’s invasion success in Europe”. Retrieved from https://osf.io/p3qkj/
McGregor KF, Watt MS, Hulme PE, Duncan RP (2012) What determines pine naturalization: species traits, climate suitability or forestry use?. Divers Distrib 18:1013-1023.
Moyle PB, Marchetti MP (2006) Predicting invasion success: freshwater fishes in California as a model. BioScience 56(6):515-524.
Naimi B (2015) Usdm: Uncertainty Analysis for Species Distribution Models. R Package Version 1.1-15. http://CRAN.R-project.org/package=usdm.
Nakagawa S, Schielzeth H (2013) A general and simple method for obtaining R² from Generalized Linear Mixed-effects Models. Methods Ecol Evol 4:133–142.
Oberdorff T, Pont D, Hugueny B, Porcher JP (2002) Development and validation of a fish‐based index for the assessment of ‘river health’in France. Freshw Biol 47:1720-1734.
Pagad S, Genovesi P, Carnevali L, Schigel D, McGeoch MA (2018) Introducing the global register of introduced and invasive species. Sci Data 5:1-12.
Peoples BK, Midway SR (2018) Fishing pressure and species traits affect stream fish invasions both directly and indirectly. Divers Distrib 24:1158-1168.
Procheş Ş, Wilson JR, Richardson DM, Rejmánek M (2012) Native and naturalized range size in Pinus: relative importance of biogeography, introduction effort and species traits. Glob Ecol Biogeogr 21:513-523.
R Core Team (2020). R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. https://www.R-project.org/.
Reid AJ, Carlson AK, Creed IF, Eliason EJ, Gell PA, Johnson PT et al (2019) Emerging threats and persistent conservation challenges for freshwater biodiversity. Biol 94:849-873.
Ribeiro F, Elvira B, Collares-Pereira MJ, Moyle PB (2008) Life-history traits of non-native fishes in Iberian watersheds across several invasion stages: a first approach. Biol Invasions 10:89-102.
Richards SA (2008) Dealing with overdispersed count data in applied ecology. J Appl Ecol 45:218-227.
Rowles AD, O’Dowd DJ (2007) Interference competition by Argentine ants displaces native ants: implications for biotic resistance to invasion. Biol Invasions 9:73-85.
Ruesink JL (2005) Global analysis of factors affecting the outcome of freshwater fish introductions. Conserv Biol 19:1883-1893.
Rytwinski T, Taylor JJ, Donaldson LA, Britton JR, Browne DR, Gresswell RE et al (2019) The effectiveness of non-native fish removal techniques in freshwater ecosystems: a systematic review. Environ Rev 27:71-94.
Sakai AK, Allendorf FW, Holt JS, Lodge DM, Molofsky J, With KA et al (2001) The population biology of invasive species. Annu Rev Ecol Evol Syst 32:305-332.
Simberloff D, Martin JL, Genovesi P, Maris V, Wardle DA, Aronson J et al (2013) Impacts of biological invasions: what's what and the way forward. Trends Ecol Evol 28:58-66.
Sommerwerk N, Wolter C, Freyhof J, Tockner K (2017) Components and drivers of change in European freshwater fish faunas. J Biogeogr 44:1781-1790.
Troia MJ, McManamay RA (2020) Biogeographic classification of streams using fish community–and trait–environment relationships. Divers Distrib 26:108-125.
van Kleunen M, Dawson W, Maurel N (2015) Characteristics of successful alien plants. Mol Ecol 24:1954-1968.
Vandekerkhove J, Cardoso AC, Boon PJ (2013) Is there a need for a more explicit accounting of invasive alien species under the Water Framework Directive?. Manag Biol Invasions 4:25.
Vila-Gispert A, Alcaraz C, García-Berthou E (2005) Life-history traits of invasive fish in small Mediterranean streams. Biol. Invasions 7:107-116.
Vilizzi L, Copp GH, Adamovich B, Almeida D, Chan J, Davison PI et al (2019) A global review and meta-analysis of applications of the freshwater Fish Invasiveness Screening Kit. Rev Fish Biol Fish 29:529-568.
Violle C, Navas ML, Vile D, Kazakou E, Fortunel C, Hummel I et al (2007) Let the concept of trait be functional!. Oikos 116:882-892.
Vitule JRS, Freire CA, Simberloff D (2009) Introduction of non‐native freshwater fish can certainly be bad. Fish Fish (Oxf) 10:98-108.
Williamson MH (1996) Biological invasions. London, UK
Zuur AF, Ieno EN, Elphick CS (2010) A protocol for data exploration to avoid common statistical problems. Methods Ecol Evol 1:3-14.

Table 1

Functional and ecological traits, and invasion history information compiled for the alien freshwater fish in Europe. A more detailed description can be found in Marcolin et al. (2022).

Trait	Type	Categories	Trait value	Definition
Female maturity	Functional	< 2 years/ 2 - 3 years/ 3 - 4 years/ 4 - 5 years/ > 5 years	1 - 5	Age of sexual maturity of female
Length of breeding season	Functional	Count	1 to 12	Number of months
Maximum adult size	Functional	<10 cm/ 11-20 cm/21-40 cm/ 41-80 cm/ 81-160 cm/ > 160 cm	1 to 6	The maximum length individuals are known to achieve in the wild
Maximum fecundity	Functional	<100 eggs/ 100-1000 eggs/ 1000-10000 eggs/ 10000-100000 eggs/ >100000 eggs	1 to 5	The maximum number of eggs per female under natural conditions
Maximum lifespan	Functional	<= 2 years/ 3-4 years/ 5-8 years/ 9-16 years/>16 years	1 to 5	Maximum known age of large individuals in wild populations living under natural conditions
Migratory behaviour	Functional	Resident/Potamodromous/ Amphidromous/Anadromous/Catadromous	1 to 5	Movement of individuals to reach the spawning habitat
Parental care	Functional	Open-substrate spawners/Brood-hiders/Guarders/Bearers	1 to 4	Behaviour of the adults towards the eggs and brood
Reproductive guild	Functional	Phytophils/Lithophils/Phyto-lithophils/ Ariadnophils/Ostracophilic/Pelagophilic/ Lithopelagophilic/Psamnmophilic/Speleophilic/Viviparous	PHYT/ LITH/ PHLI/ ARIAD/ OTH	Spawning substrate where the eggs are released. We chose as specific level the categories more represented in our database and assigned the less represented as other
Spawning frequency	Functional	Semelparity/Spawning not in consecutive years/Spawning once per year/Spawning more than once per year	1 to 4	Number of breeding season over one year
Adult trophic status	Ecology	Piscivorous/Omnivore/Herbivore/Invertivore/Planktivorous	C/O/H/I/P	Based on the main food items for (>50% of adults)
Acid tolerance*	Ecology	Tolerant/Intermediate/Intolerant	ATOL/AIM/AINTOL	Trait assigned following the tolerance to water acidification
Feeding habitat	Ecology	Benthic/Generalist/Pelagic	1 to 3	Zone preference of the water column by the species for feeding
Flow preferences	Ecology	Limnophilic/Eurytopic/Rheophilic	1 to 3	Water flow preference
Habitat degradation tolerance*	Ecology	Tolerant/Intermediate/Intolerant	HTOL/HIM/HINTOL	Based on the tolerance of a species to the physical degradation of the habitat that inhabits
Spawning habitat preferences	Ecology	Limnopar/Euryopar/Rheopar	1 to 3	Water flow preference for the spawning period
Salinity tolerance	Ecology	Freshwater/Freshwater-brackish/ Freshwater-brackish-marine	FRE/FBR/FBM	Salinity degree of tolerance for a species, based on the living habitat living habitat
Temperature tolerance*	Ecology	Eurythermal/Stenothermal	EUTHER/STTHER	Range of temperature which a species can withstand
Trophic level/index	Ecology	Float	2 to 4.7	Express where fish and other organisms tend to operate in their respective food webs. See FishBase (Frose and Pauly 2020)
Water quality tolerance general*	Ecology	Tolerant/Intermediate/Intolerant	TOL/IM/INTOL	Trait representing the tolerance of a species to the usual water quality parameters
Water quality tolerance O2*	Ecology	Tolerant/Intermediate/Intolerant	O2TOL/O2IM/O2INTOL	Trait representing the tolerance of a species to the low dissolved oxygen concentration
Water quality tolerance TOX*	Ecology	Tolerant/Intermediate/Intolerant	TOXTOL/TOXIM/TOXINTOL	Trait representing the tolerance of a species to the toxic contamination of the water
Climate zone	Ecology	Temperate/Tropical/Subtropical/Cold	TEMP/TROP/SUBT/COLD	Distribution range of the species based on either the climate zone or latitudinal range reported by Froese and Pauly (2020). A species can belong to one category or a combination of two/three/four categories
European native	Ecology	Yes/No	0/1	Is the species native from Europe (1) or outside Europe (0)? Whether the species is native or not in at least one country of the study area considered. Used as a proxy for the distance from nearest native area
Cause of introduction	Invasion history	Intentional/Dispersal/Accidental	ACC/INT/DIS	Variable assigned according to the cause of the introduction of a non-native species in Europe. When the information was not present, we based our decision on the ‘Human use’ section present for each species on FishBase (Froese and Pauly 2020). A species can belong to one category or a combination of two/three categories
Fraction of European countries with established populations	Invasion history	Float	0 to 1	Proportion between the number of European countries where a species was introduced and established, and the overall number of European countries in which the species was introduced
Invasive status	Invasion history	Invasive/Non-invasive	INV/NOT	We considered a species invasive whenever we found information that the species status is invasive in, at least, one country. However, most species were evaluated a priori, following the method of Vilizzi et al. 2019
Number of causes of introduction	Invasion history	Count	1 to 5	This trait is based on the seven categories of cause of introduction described in FishBase (Frose and Pauly 2020). The trait is a mere count of the number of categories of introduction per species.
Number of European countries with established populations	Invasion history	Count	0 to 30	Number of European countries in which each species has been introduced and successfully established
Prior invasion success outside Europe	Invasion history	Count	0 to 94	Number of countries worldwide in which each species has been introduced and successfully established
Propagule pressure	Invasion history	Count	1 to 50	This trait was used as a proxy of propagule pressure based on a similar approach done by McGregor et al. (2012), Proches et al. (2012) and Peoples and Midway (2018). We counted the number of events of introduction of a species in countries of the study area considered, based on the information reported in FishBase (Frose and Pauly 2020) and AquaNIS.

* The ‘Acid tolerance’, ‘Habitat degradation tolerance’, ‘Water quality tolerance general’, ‘‘Water quality tolerance O2’ and ‘Water quality tolerance TOX’ were aggregated afterwards in the trait ‘General tolerance’ (see the Material and Methods). Coastal Research and Planning Institute (2012) AquaNIS. http://www.corpi.ku.lt/databases/index.php/aquanis. Accessed 20 March 2020; Froese R. Pauly D eds (2020) FishBase. World Wide Web electronic publication. www.fishbase.org Accessed July 2020; McGregor KF, Watt MS, Hulme PE, Duncan RP (2012) What determines pine naturalization: species traits, climate suitability or forestry use?. Divers Distrib 18:1013-1023; Peoples BK, Midway SR (2018) Fishing pressure and species traits affect stream fish invasions both directly and indirectly. Divers Distrib 24:1158-1168; Procheş Ş, Wilson JR, Richardson DM, Rejmánek M (2012) Native and naturalized range size in Pinus: relative importance of biogeography, introduction effort and species traits. Glob Ecol Biogeogr 21:513-523; Vilizzi L, Copp GH, Adamovich B, Almeida D, Chan J, Davison PI et al (2019) A global review and meta-analysis of applications of the freshwater Fish Invasiveness Screening Kit. Rev Fish Biol Fish 29:529-568.

Table 2

Functional and ecological traits, and invasion history information used as predictor (P) or response variable ( R ) or not used (/) for each model of invasion stage

Variables	Release	Establishment	Spread	Impact
Biological trait
Female maturity	P	P	P	P
Length of breeding season	P	P	P	P
Maximum fecundity	P	P	P	P
Maximum lifespan	P	P	P	P
Migratory behaviour	P	P	P	P
Parental care	P	P	P	P
Reproductive guild	P	P	P	P
Spawning frequency	P	P	P	P
Ecological trait
Feeding habitat	P	P	P	P
Flow preferences	P	P	P	P
General tolerance	P	P	P	P
Salinity tolerance	P	P	P	P
Spawning habitat preferences	P	P	P	P

Variables	Release	Establishment	Spread	Impact
Temperature tolerance	P	P	P	P
Trophic level/index	P	P	P	P
Climate zone - TEMP	P	P	P	P
European native	P	P	P	P
Invasion history
Cause of introduction	P	P	P	P
Fraction of European countries with established species (establishment stage)	/	R	P
Number of causes of introduction	P	P	P	P
Number of European countries with established species (spread stage)	/	/	R	P
Prior invasion success outside Europe	P	P	P	P
Propagule pressure (release stage)	R	P	P	/
Invasive status (impact stage)	/	/	/	R

Table 3

Relations between alien freshwater fishes and variables reflecting functional and ecological traits, and invasion history for the four invasion stages (release, establishment, spread and impact) inferred from generalized linear mixed models (GLMM). For each relation we provide for estimated coefficient and the 95% confidence interval of the predictor variable (The Akaike weight, and the goodness of fit (marginal and conditional R²) are also provided for each model. NS - candidate predictor not included in the final model. REF – reference category of the dummy variables). Blank cells correspond to variables not selected in the final model of a given stage. Variables explanations can be found in Table 1 and Marcolin et al. (2022).

Variables	Release	Establishment	Spread	Impact
Functional trait
Intercept	10.27 (7.93, 12.62)	1.41 (0.82, 2.01)	1.49 (1.3, 1.67)	-0.71 (-2.3, 0.98)
Female maturity	NS			-1.05* (-1.79, -0.16)
Length of breeding season	NS
Maximum fecundity		-0.43* (-0.75, -0.11)	NS
Maximum lifespan	NS	-0.34* (-0.61, -0.07)	NS	0.94* (0.2, 1.61)
Migratory behaviour	-1.35* (-2.35, -0.34)			-0.81* (-1.51, -0.09)
Parental care	-1.82* (-3.01, -0.62)	NS	NS	NS
Reproductive guild		NS	NS
Spawning frequency		NS	0.18* (0.06, 0.29)

Variables	Release	Establishment	Spread	Impact
Ecological trait
Feeding habitat		NS
Flow preferences		NS
General tolerance
Salinity tolerance				NS
Spawning habitat preferences	NS	1.41* (0.82, 2.01)
Temperature tolerance		0.21* (0.01, 0.42)		NS
Trophic level/index				NS
Climate zone - TEMP		REF		REF
Climate zone - TROP		NS		2.39 (-0.14, 4.67)
Climate zone - TEMP/SUBT		NS		1.57 (-0.12, 3.35)
Climate zone - TEMP/COLD		NS		-0.42 (-2.91, 2.1)
Climate zone - SUBT/TROP		NS		1.16 (-1.09, 3.47)
Climate zone - TEMP/SUBT/TROP		NS		3.01 (-0.07, 5.45)
Climate zone - TEMP/SUBT/COLD		NS		2.74* (0.98, 4.68)

Variables	Release	Establishment	Spread	Impact
Climate zone - TEMP/SUBT/TROP/COLD		NS		-0.28 (-4.27, 3.63)
European native	-1.53* (-2.69, -0.36)	0.51* (0.28, 0.74)
Invasion history
Cause of introduction - INT	REF	REF	REF
Cause of introduction - ACC	-3.55* (-6.44, -0.66)	-0.09 (-0.8, 0.61)	-0.35* (-0.62, -0.07)
Cause of introduction - DIS	-0.84 (-4.78, 3.1)	0.47 (-0.48, 1.41)	0.2 (-0.15, 0.53)
Cause of introduction - INT/DIS	-0.06 (-6.52, 6.39)	-0.36 (-1.3, 0.57)	0.49 (-0.01, 0.93)
Cause of introduction - ACC/DIS	2.07 (-2.84, 6.98)	-0.62 (-1.68, 0.44)	0.22 (-0.16, 0.58)
Cause of introduction - ACC/INT	-5.41* (-8.55, -2.26)	-0.13 (-0.77, 0.52)	-0.26 (-0.54, 0.04)
Cause of introduction - ACC/INT/DIS	4.01 (-0.19, 8.2)	0.17 (-0.53, 0.86)	0.11 (-0.15, 0.4)
Fraction of European countries with established species (establishment stage)			0.26* (0.16, 0.36)
Number of causes of introduction	4.66* (3.47, 5.86)		0.15* (0.06, 0.25)

Variables	Release	Establishment	Spread	Impact
Number of European countries with established species (spread stage)				3.31* (1.44, 4.88)
Prior invasion success outside Europe	4.20* (3.13, 5.26)	NS
Propagule pressure (release stage)		NS	0.54* (0.46, 0.62)
Akaike weight	0.35	0.49	0.49	0.24
R²(marginal/conditional)	0.71/0.72	0.53/0.70	0.71/0.71	0.84/0.85

* coefficients whose confidence intervals do not contain zero

Table 4

Importance of each candidate predictor variable considered at each stage, based on the sum of the Akaike’s weights (see text for further details).

Trait	Release	Establishment	Spread	Impact
Functional trait
Female maturity	0.07			0.63
Length of breeding season	0.18
Maximum fecundity		0.83	0.21
Maximum lifespan		0.66	0.07	0.62
Migratory behaviour	0.48			0.62
Parental care	0.72	0.07	0.35	0.18
Reproductive guild		0.01	0.07
Spawning frequency		0.09	0.57
Ecological trait
Feeding habitat		0.06
Flow preferences		0.04
General tolerance	0.05
Salinity tolerance			0.07	0.00
Spawning habitat preferences	0.10	1.00
Temperature tolerance		0.25		0.24
Trophic level/index				0.30
Climate zone		0.00		0.86
European native	0.28	1.00
Invasion history
Cause of introduction	1.00	1.00	0.79
Fraction of European countries with established populations (establishment stage)			1.00
Number of causes of introduction	1.00		0.85

Trait	Release	Establishment	Spread	Impact
Number of European countries with established populations (spread stage)				1.00
Prior invasion success outside Europe	1.00	0.00
Propagule pressure (release stage)		0.00	1.00

Download PDF

Version 1

posted

You are reading this latest preprint version

Species traits and invasion history as predictors of freshwater fish’s invasion success in Europe

Status:

Version 1

Abstract

Figures

1. Introduction

2. Materials And Methods

2.1. Alien fish species

2.2. Fish species traits and invasion history

2.3. Data analysis

3. Results

3.1 Release stage model

3.3 Establishment stage model

3.4 Spread stage model

3.5 Impact stage model

4. Discussion

4.1 Traits filtering at each invasion stage

4.2 Invasion history as driver of alien fish invasion

4.3. Management recommendations

Declarations

References

Tables

Supplementary Files

Status:

Version 1