The prediction of invasive plant species based on their traits and climate change

(1)

The prediction of

invasive plant

species based on

their traits and

climate change

A critical review about the value of incorporating climate change in the

risk-assessments of invasive plant species

Figure 1. Kudzu (Pueraria lobata) as an example of an invasive terrestrial plant species in the South-East of the United States (L, no date).

Kelly van Leeuwen BSc

Literature review MSc Earth Sciences Supervisor: dr. Elly Morriën Assessor: dr. Carina Hoorn 28 April 2021 Word count: 8095

Abstract

Invasive terrestrial plant species are causing problems all over the world. To limit the impact of invasive alien species (IAS), policies have focused on preventing, controlling and eradicating IAS. To better control these species, predictions of species that will become invasive and IAS that extent their ranges is necessary. One of the aspects that should be taken into account for these predictions is climate change. This critical review focuses on the value of taking climate change into account in risk-assessments of invasive terrestrial plant species. These risk-assessments are based on general models, called species distribution models. By incorporating climate change in these models, the outputs become uncertain because the input data of future climate change is uncertain. In addition, the responses of plants on climate change differ and are based on their own traits. This causes difficulties in the prediction of the distribution and impact of invasive terrestrial plant species under climate change. The improvement of these models can be found in the emerging field of genetics and epigenetics. In this field, the links between DNA, the folding of DNA and IAS under climate change are investigated. As long as this field is expanding, models can be updated. For now, we can use the models and therefore the risk-assessments for prioritising management of IAS.

Keywords – Invasive alien species (IAS),

terrestrial plants, risk-assessments, climate change, traits, management

(2)

1

1. Introduction: invasive alien species

and climate change

Since mankind started travelling around the world, ecological explosions arise everywhere (Elton, 1958). An ecological explosion is the expansion of an organism in such a way that it is out of control. These explosions have increased even more since 1970 after the planet has seen a massive rise in people and goods travelling around the world (Seebens et al., 2017). Plants, animals and other organisms are taken along on these journeys (IUCN, 2017). Some of these organisms settle and become alien species, but some of these alien species become invasive, threatening biodiversity and ecosystems (Bellard, Cassey and Blackburn, 2016; Early et al., 2016). These invasive alien species (IAS) are defined as “species that are introduced, accidentally or intentionally, outside of their natural geographic range and that become problematic” (IUCN, 2015).

IAS have a major negative impact on ecosystems, health and the economy (Bradley et al., 2012; Bellard, Cassey and Blackburn, 2016; Early et al., 2016; Kaky and Gilbert, 2019). IAS are the second most common threat of species extinctions with biological resource use in the first place (Bellard, Cassey and Blackburn, 2016). They change life-supporting ecosystem services, like fresh water, food security and disease regulation (Pejchar and Mooney, 2009). This forces governments to intervene in affected ecosystems. The United States alone spends $120 billion dollars per year on controlling IAS (Pimentel et al., 2001; Pimentel, Zuniga and Morrison, 2005).

The management strategies used by these governments are focused on preventing, controlling and eradicating IAS (Lodge et al., 2016). Preventing eliminates the impact of IAS, controlling improves the detection and eradication limits future impacts (Rout et al., 2011). Each invasive alien species will need its own management strategy based upon their stage of invasion (Lodge et al., 2016). However,

(3)

2 Lodge et al. (2016) state that prevention as a strategy works best.

For prevention, it is necessary to predict the distribution and the impact of IAS and potential IAS (Pagad et al., 2015). At the moment these predictions are mainly done by defining traits that increase the invasiveness (Van Kleunen, Weber and Fischer, 2010; Lodge et al., 2016). Dispersal and distribution of IAS are taken into account as well. Climate change also influences the distribution and impact of IAS, an aspect that is not always considered in risk-assessments (Bradley, Wilcove and Oppenheimer, 2010; SmithAndrea et al., 2012; Liu et al., 2017).

Assessing and quantifying the impact of climate change on IAS is important, and for this reason the International Union of Conservation of Nature (IUCN) stated that climate change should be incorporated in risk-assessments of IAS (IUCN, 2017). However, climate change is versatile and includes aspects like increasing temperature, extreme weather conditions and changes in precipitation (Hulme, 2014; Djebou and Singh, 2016; Knutson et al., 2017). This raises the question, how valuable is it to include climate change in the predictions on distribution and impact of IAS in the existing risk-assessments? To address this question, I will perform a critical review on existing literature on climate change and potential invasive plant species. My focus will be on terrestrial plant species around the world, because climate change is not limited to borders. Besides this, models used for the predictions of the distribution and impacts of IAS are general and can be applied at every place in the world as long as there is data available. Terrestrial plant species are chosen for their knowledge on invasiveness and responses to climate change.

The review starts with the definition and description of IAS. Thereafter, I will discuss the current risk-assessments based on traits, dispersal and distribution. Next, I will talk about the general effects of climate change on

the distribution and impact of IAS. The next chapter will discuss how we can predict the spread and impact of potential invasive species based on existing models. Then, I will talk about how we can implement the knowledge about the effect of climate change on IAS in the risk-assessments. This will lead to a final conclusion and discussion.

2. Methodology

In this critical review I used Web of Science and Google Scholar. Web of Science is applied to search for the most relevant articles in the field. I looked for these articles in Google Scholar to search for cited and citing literature. The search terms used in Web of Science differed between chapters. In Appendix A, a table with the terms for each chapter are presented. My search focussed on articles between 2000 and 2021 to compare the changes in knowledge in the field. Only articles within biological categories and journals were selected to exclude non-relevant articles. The articles that remained were filtered on their title, abstract and times cited. A selection of the most relevant articles was left and used in Google Scholar to check the cited and citing literature. These articles were also filtered on title, abstract and times cited to obtain the most significant articles. The most recent articles were used, but no time range was set for the search on Google Scholar.

3. Description IAS

Looking closely into risk-assessments of IAS under future climate change requires knowledge about IAS. First, I will discuss the different terms and definitions which are used in invasion biology. Thereafter, hypotheses on how species become invasive are discussed. Finally, I will focus on how plants travel to new environments and become invasive.

3.1 Debate on definition

It is important to have a clear definition on invasive species to prevent miscommunication and misinterpretation of IAS (Chew and Laubichler, 2003; Pyšek et al., 2004).

(4)

3 Ambiguous terms affect the monitoring and the management of IAS (Humair et al., 2014). These ambiguous terms are a consequence of three different discussion points along experts about the definition of IAS (Humair et al., 2014).

There are multiple terms used as synonyms for IAS which causes ambiguity (Colautti and MacIsaac, 2004; Rodriguez, 2006). Exotic, non-native, non-indigenous, alien, imported, introduced and naturalized are all used synonymously for invasive alien species. However, as Colautti and MacIsaac (2004) explain, these terms can refer to different stages of invasiveness. They claim that it is better to refer to ‘stage-based’ terminology than to ambiguous terms.

Another confounding factor is the use of biogeographical ranges and the impact of IAS in the definition. Wilson et al. (2009) state that dominant species and biological invasions should be seen separately. This means that invasive species are by definition non-native which implies that the biogeographical range should be taken into account when defining IAS. Colautti and MacIsaac (2004) agree with Wilson et al. (2009) that biogeographical ranges should be included in the definition, but others question this (Warren, 2007; Valéry et al., 2008, 2009). Warren (2007) claims that including these ranges in the definition of IAS is a product of culture and xenophobia, because people are afraid of IAS crossing the borders of their natural ranges. Therefore, according to Warren (2007), impact is the only factor that should be included in the definition of IAS. Valéry et al. (2008, 2009) agree and point out that native species can also be invasive as they cause similar impacts as non-native species. However, it seems that invasiveness is not correlated with the impact of IAS and thus should not be used to connote each other (Ricciardi and Cohen, 2007).

Finally there is a debate around the definitions used for native and non-native. As Pyšek et al. (2004) mention, the difference

between native and non-native is defined by their residence time. According to Colautti and MacIsaac (2004), the deviation is linked to the stage of invasiveness and is independent on origin or residence time. Whether a species is seen as native, also changes through climate change, globalisation and mass disturbance of ecosystems (Hill and Hadly, 2018). The difference between native and non-native is therefore not clear.

The opposing views on the definition of IAS are all intertwined. Native and non-native are not necessarily linked to biogeographical ranges, but are used as synonyms for invasive (e.g. Colautti and MacIsaac, 2004). It is important to keep in mind that every expert has its own definition of IAS and that this influences risk-assessments (Humair et al., 2014). Nevertheless, the search terms for this review will include all the words used as synonyms for IAS to include the most important publications.

3.2 Hypotheses describing how species

become invasive

As with the definition of IAS, there is also discussion about how species become invasive. In 2009, 29 different invasion hypotheses were summarized (see figure 2 for an overview) (Catford, Jansson and Nilsson, 2009). However, it is crucial to know how species become invasive, because management is based upon this knowledge (Jeschke et al., 2012).

For this review, three of the most frequently used hypotheses in the literature are discussed (Lowry et al., 2013). The first hypothesis is on disturbances in the invading ecosystem and is called the intermediate disturbance hypothesis (Wilson, 1994; Townsend, Scarsbrook and Dolédec, 1997; Seebens, Essl and Blasius, 2017). The hypothesis implies that environments with low levels of disturbances are species poor, because areas with little disturbances are easy to invade which lowers the survival changes of native species. Ecosystems with high levels of

(5)

4 disturbances on the other hand are also species poor because they are difficult to invade which makes it hard for species to sustain. In other words, the available niche spaces for invading species are high in areas with low disturbances. The second hypothesis is about community species richness and is better known as the biotic resistance hypothesis (Elton, 1958; Levine and D’Antonio, 1999; Levine, Adler and Yelenik, 2004). The higher the species richness

in an ecosystem, the higher the resilience of that ecosystem and thus the harder it is to invade such an ecosystem. The last discussed hypothesis is the enemy release hypothesis (Elton, 1958; Colautti et al., 2004; Liu and Stiling, 2006; Heger and Jeschke, 2014; Jeschke, 2014). The hypothesis suggest that invasive species can flourish because they do not have any natural enemies in the introduced ecosystem. Catford, Jansson and Nilsson (2009)

Figure 2. Overview of the 29 invasion hypotheses proposed by Catford, Jansson and Nilsson (2009) (Jeschke, 2014). In this review the red lined hypothesis are summarized: intermediate disturbance (DS), biotic resistance (BR) and enemy release (ER).

(6)

5 offer a more detailed overview of the invasion hypotheses.

All the hypotheses have been tested, but no consensus has been reached (Liu and Stiling, 2006; Catford et al., 2012; Jeschke et al., 2012; Fox, 2013; Heger and Jeschke, 2014). The reason for this is that variables, like habitat type and species, affect the support or rejection of a hypothesis (Jeschke et al., 2012; Heger and Jeschke, 2014). However, these hypothesis influence risk-assessments of invasions and it is therefore important to keep in mind that experts have alternative hypotheses (Jeschke et al., 2012). These hypotheses also account for terrestrial plant species. The question still remains how plants travel to new environments.

3.3 Invasiveness in plant species

The old study area of plant dispersal laid ground in the books of of Ridley (1930) and Van der Pijl (1982), which summarize different ways of plant dispersal. This can be roughly divided in two groups: natural dispersal and human-induced dispersal. The natural dispersal of plants is caused by wind, water and animals (Ridley, 1930; Howe and Smallwood, 1982; Van der Pijl, 1982). For example, some seeds can travel long distances by floating or wind travelling. Other seeds are dispersed by animals in their faeces or in their furs.

Human-induced dispersal occurs when humans transport goods and themselves to other places and take diaspores, parts of plants or whole plants with them, intentionally and unintentionally (Ridley, 1930; Van der Pijl, 1982; Seebens et al., 2017). The intentional dispersal is mainly a result of gardening activities. The unintentional dispersal is caused by diaspores or parts of plants that stick to clothes or goods. Another unintentional way of dispersal is travelling along with extreme weather conditions, like typhoons (Nathan, 2006). Human activities that enforced climate change contributed to these conditions. Chapters 4 and 5 will elaborate on this topic.

Some plants will have a higher probability on long distance dispersal than others depending on their own dispersal characteristics (Ridley, 1930; Van der Pijl, 1982; Nathan, 2006). However, this does not mean that every dispersed diaspore or piece of a plant grows, reproduces and even becomes invasive. The growing process and the reproduction depend on environmental factors like light, nutrient availability and water (Van der Pijl, 1982). If the conditions are right, this still does not mean that a species becomes invasive (Jeschke et al., 2012). The different invasion hypotheses from chapter 3.2 examine on the circumstances that should be met for a species to become invasive (see Catford, Jansson and Nilsson (2009) for 29 hypotheses). Although not every alien species becomes invasive, humans increase the probability of invasions by dispersing plants over large distances (Seebens et al., 2017).

4. Current measures on invasiveness

As previously mentioned, terrestrial plant species have multiple ways of becoming invasive. However, the question remains if and how we can predict which species are going to be invasive and where they are going to be invasive. This chapter will investigate this by examining risk-assessments of invasive plant species. First, I will discuss traits that are accompanied with invasiveness. Thereafter, the ways of dispersal with a higher chance of becoming invasive are described. Finally, I will discuss places and habitats which are most vulnerable for invasions by plant species.

4.1 Trait-based

In 1965, the first paper about the ‘ideal weed’ was written by Baker (1965). In this paper 14 different traits of an ideal weed were summarized. Nowadays this idea is seen as far too simplistic (Pyšek and Richardson, 2007). Studies and reviews show opposing views about the ability to predict invasiveness based on traits. One of these views is proposed by Pyšek and Richardson (2007). Their review of 18 studies found that some traits were explanatory for invasiveness, e.g. timing of

(7)

6 flowering, while others were not, e.g. propagule size. Van Kleunen, Weber and Fisher (2010) found in their meta-analysis of 117 studies that there is a significant difference between invasive and native species in the six traits they investigated. Rejmanek and Richardson (1996) also found evidence for traits that predict the invasiveness of pines. Mean seed mass, minimum juvenile period and mean interval between large seed crops were predictors for invasiveness.

Other studies show no evidence for traits that could explain invasiveness. Leffler et al. (2014) compared differences in traits between native species and between native and exotic species. They found that differences in traits between native species are almost as big as differences between native and exotic species. Furthermore, Daehler (2003) found evidence that invasive species were more likely to have higher leaf area, lower tissue construction costs and greater phenotypic plasticity. However, there was no significant difference in the other three traits they investigated. Just like Leishman, Thomson and Cooke (2010), Daehler (2003) concluded that invasiveness might not be a result of the traits of an invader, but a consequence of environmental conditions like resource availability and altered disturbances.

Specific traits that play a role in invasiveness are not coherent (see figure 3 for an overview of traits that have been investigated in the light of invasiveness). The reason for this is that studies use different methods which result in different outcomes (Pyšek and Richardson, 2007). There is for example variation in scale, in analytical tools and in the subjects that are compared (native-invasive, invasive-invasive, etcetera). In addition, the traits that increase invasiveness vary between contexts, because factors like community ecology also influence invasiveness. (Drenovsky et al., 2012; Leffler et al., 2014). Therefore three hypotheses on general plant traits that contribute to invasiveness are founded. The

two-sides-of-the-same-coin hypothesis explains that invasive species should be on the other side of the spectrum within a trait than a species which is near extinction (Blackburn and Jeschke, 2009). The try-harder hypothesis explains that invasive species should have other traits than the locals to better cope with the environment (Tecco et al., 2010). The join-the-locals hypothesis tells the opposite: invasive species must be able to live in the

Figure 3. Overview of traits that have been investigated in the light of invasiveness of eight different studies and reviews (Rejmanek and Richardson, 1996; Daehler, 2003; Pŷsek and Richardosn, 2007; Pŷsek, 2009; Van Kleunen, Weber and Fisher, 2010; Richardson and Rejmanek, 2011; Leffler et al., 2014; Moravcová et al.,2015). The broadest term of a trait is mentioned, for example maximum height was shifted under plant height.

(8)

7 same environment as the locals and should therefore have the same traits (Tecco et al., 2010). These hypotheses are only supported by contradicting evidence (Blackburn and Jeschke, 2009; Tecco et al., 2010).

4.2 Dispersal

The discussion about predicting invasiveness based on plant specific traits continues when investigating dispersal mode as a predictor (Pyšek and Richardson, 2007). For example, Richardson and Rejmanek (2011) showed that trees and shrubs dispersed by birds had the highest chance of naturalization, while Moravcová et al. (2015) show that animals in general were the best dispersers of IAS. However, they added that this only accounts for species which are too small to compete with other species and which have small seeds. Another example of the discussion about dispersal mode as a predictor for invasiveness is that Pyšek et al. (2009) conclude that the ability of multiple dispersal modes increases the invasiveness, while Gillespie et al. (2012) summarize that long-distance travelling and the chance to establish in a new environment depend on circumstances during the journey and the ability to survive at other places. The reason why these studies draw different conclusion on the effect of natural dispersal modes on invasiveness, might lie in definition and methodology variations between the studies (Richardson and Pyšek, 2012; Moravcová et al., 2015). Naturalised and invasive are not the same and studies look at many different factors in different ways.

Although experts discuss the effects of natural dispersal modes on invasiveness, they agree on the effect of human-induced dispersal on invasiveness. Climate change, driven by humans, causes extreme storms which helps to disperse diaspore all over the world (Nathan, 2006). In addition, diaspores are directly dispersed by humans through forestry, horticulture, agroforestry, food dissemination (Richardson and Rejmanek, 2011) and gardening (Heywood, 2011). These human-induced dispersal modes vanish the natural

biogeographic boundaries (van Kleunen et al., 2015). This increases the probability of the introduction of new alien species. The chances that these species become invasive are also affected by humans (Nathan, 2006; Heywood, 2011; Richardson and Pyšek, 2012). Due to disturbances of the land, terrestrial alien species have a higher chance of becoming invasive (Richardson and Pyšek, 2012). In addition, global and climate change increase the likelihood that garden species can survive outside greenhouses (Heywood, 2011). Humans thus greatly affect the introduction of new alien species and the likelihood that they become invasive.

4.3 Distribution

Humans are not the only factor that influence invasiveness. The native distribution ranges of species also affect the probability of invasion (Rejmánek, 1996; Richardson and Rejmanek, 2011), because species regularly invade regions with the same climate as in their native range (Richardson and Pyšek, 2012). Pyšek et al. (2009) add that the combination of specific traits and a large native range increases this chance. The native range is affected by species’ traits which means that traits and distribution are influencing each other (Pyšek et al., 2009). Jimenez-Valverde et al. (2011) agree on this and add that models that predict places with a high chance of invasion have to include information on distribution, traits that link to invasiveness and locations which are already invaded. Besides the effects of traits on the distribution and introduction of invasive species, inherent vulnerability of places can increase the probability of the introduction of IAS.

Inherent vulnerability can be found in disturbed landscapes which makes it possible to invade regions with other climates (Bradley et al., 2010). Inherent vulnerability can also be at a more global scale. For example, the Southern hemisphere is the biggest acceptor of new IAS (van Kleunen et al., 2015). North-America and Australia are also more vulnerable than Europe, probably because of isolation, dry

(9)

8 climate and less adaptation to disturbed soils in Australia and North-America (van Kleunen et al., 2015). This makes it easier for newly introduced species to settle and invade. Australia and North-America are together with southern Africa, the Pacific Islands and New Zealand the most vulnerable for invasions (Richardson and Rejmanek, 2011; van Kleunen et al., 2015; Pysek et al., 2017). Australia, the Pacific Islands and New Zealand are islands which have the hardest task to withstand invasions because of their isolation (Pysek et al., 2017). Isolation, disturbances and the native climatic boundaries of species are thus influencing invasiveness. Again, humans play a crucial role in plants invasions (see Box 1 for an example of invasions in Chile).

Box 1

A good example of the distribution and its causes of invasive plant species can be found in Chile (Arroyo et al., 2000). Chile has been part of the Spanish colonial period. Since this period, changes in invasive and alien species have been noticeable due to travelling. A part of Chile has a Mediterranean climate, just like Spain. Plants and propagules that were intentionally or accidently introduced therefore had a high chance of survival. The survival chances increased when taking into account the location of Chile. Chile is surrounded by the Pacific ocean and the Andes. This means that Chile has to deal with the so-called Island Syndrome. Islands are more susceptible for invasions than the mainland (see chapter 4.3). However, Chile has one advantage against invasions and that is its high biodiversity. This high biodiversity includes a relatively large proportion of woodiness and less of annual plants. This low amount of annual plants implies that invasive species which are annual have a higher chance of survival because there is little competition.

Not only the natural circumstances play an important role in invasions in Chile, humans also have a large impact. The introduction of domesticated grazers and rabbits improved invasions. These animals reduced the woody cover and perennial herbs which gives annual plants the opportunity to flourish.

Besides the grazers, the construction of cities and roads gave invasive plants an advantage. The disturbances of the vegetation are devastating for native species and supportive to invasive species.

All these changes in Chile has increased the number of naturalized species since the colonial period with 690. A small part is intentionally introduced for gardens, agriculture and medicinal use. However, the largest part is accidently introduced, mainly by humans.

5. Climate change and its effects

Invasiveness is an interplay between traits, ways of dispersal, distribution patterns and also a slight bit of coincidence. That is what the previous chapter showed. In addition, climate change is part of this interplay. However, it is important to understand in which ways climate change can influence invasions. This chapter will therefore first focus on the definition of climate change. Hereafter, different consequences of climate change will be discussed. Finally, the role of phenotypic plasticity will be pointed out.

5.1 Definition of climate change

To understand the role that climate change plays in affecting invasiveness, it is necessary to define climate change. Climate change is not the same as global change, so globalization will not be taken into account in this review.

In 1992 a book was published that contained definitions on every term that is used in climatology (Maunder, 1992). The definition for climate change is taken in the broadest sense. Climate change is every inconsistency with the general pattern of meteorological elements. There should be a deviation on the long-term mean of climatic parameters. Extreme weather conditions are a consequence of climate change. The IUCN agrees on this definition of climate change (Adhikari et al., 2011). However, they add that extreme weather events can be a result of climate change, but they can also be natural events. The Intergovernmental Panel on

(10)

9 Climate Change (IPCC) also agrees on the fact that climate change is about changes and variability in the means and properties of climatic variables (IPCC, no date). Climate change can be natural or human-induced. In the scientific world this definition is shared between experts (Bushell, Colley and Workman, 2015). However, the definition changes when policy makers and citizens are asked about it (Ibid.). For this review, I will keep the definition of climate change proposed by the IPCC.

5.2 General consequences of climate

change

In the case of the effects of climate change on invasions, climate change is merely seen as temperature rise, precipitation change and more frequent extreme weather events. Climate change is a driver of factors that influences invasions. These drivers blur the definition of invasiveness (Harrington, Woiwod and Sparks, 1999; Thuiller, Richardson and Midgley, 2007; Walther et al., 2009; Kiers et al., 2010).

As Chown et al. (2015) state, there are multiple effects of climate change on invasiveness. To show the complexity of these effects, I will discuss six of them. The first effect implies that sleepers, casual non-native species, can become invasive under climate change, because propagule pressure is increased (Hulme, 2017). Second, climate change can stimulate invasions by benefitting species with particular functional traits, like good seed defence traits (Drenovsky et al., 2012), good dispersal abilities and tolerance for disturbances (Thuiller, Richardson and Midgley, 2007). Third, climate change facilitates colonization and reproduction by extending growing seasons (Walther et al., 2009). Fourth, climate change results in more extreme weather events which increases the movement of species and decreases the biotic resistance of native species (Thuiller, Richardson and Midgley, 2007; Diez et al., 2012). Fifth, the rise of CO2 affects invasions by influencing the nutrient uptake (Thuiller,

Richardson and Midgley, 2007). When CO2 levels rise, the uptake becomes more efficiently and fast growing plants will have the highest advantage. This characteristic is often associated with invasive species (Pyšek and Richardson, 2007). Finally, climate change can adapt trophic systems and ecosystems (Harrington, Woiwod and Sparks, 1999; Kiers et al., 2010). Some plants will flourish because they are eaten less (Harrington, Woiwod and Sparks, 1999) and other plants will flourish because of changing mutualistic relations (Kiers et al., 2010). Mutualistic relations are affected by climate change through the promotion of migration of one of the species in this relation. The relation is ended which can cause extinction of the staying species. However, an invading species can take over the missed role in this relation and sustain the native species. IAS can thus also be an enrichment of an ecosystem (Walther et al., 2009). However, this mainly accounts for generalists species, because these species are not bonded to one specific other species (Schweiger et al., 2010). Invasions could thus be an improvement, but they could also be a deterioration.

5.3 Range shifts as a consequence of

climate change

Besides the general effects of climate change on invasions, range shifts are seen as one of the most important effects. The temperature increase is the main driver for these range shifts (Thuiller, Richardson and Midgley, 2007; Walther et al., 2009; Thomas, 2010; Chen et al., 2011; Petitpierre et al., 2012; Parmesan and Hanley, 2015).

Range shifts are an effect of a warmer and drier climate which makes it easier for tropical and sub-tropical species to survive at higher latitudes (Bradley et al., 2012). Not only tropical plants move due to climate change, 68% of the plants in the United Kingdom did the same (Thomas, 2010). The reason behind this, is the fact that species best survive under climate conditions which are similar to their native range (Walther et al., 2009; Petitpierre

(11)

10 et al., 2012). Petitpierre et al. (2012) found that only 15% of the terrestrial plant species move away further than 10% out of their natural climate conditions. However, the range shifts do have a limit in the form of rough climates, like tundra’s (Thuiller, Richardson and Midgley, 2007). The search for natural climatic conditions is not only latitudinal but also elevational (Chen et al., 2011; Parmesan and Hanley, 2015). On average species move 11 metres in elevational direction and 16.9 kilometres towards the poles per decade (Chen et al., 2011). Morueta-Holme et al. (2015) even found that plants move on average 27 to 32 metres upwards per decade. However, the movement varies between species and is based on their own traits (Chen et al., 2011; Cooke et al., 2013). The responses to climate change thus vary between species (Bradley, Wilcove and Oppenheimer, 2010; Bellard et al., 2013)

Multiple factors influence the relation between climate change and invasions and this is the reason why observations, experiments and theory are not always consistent (Parmesan and Hanley, 2015). For example, Bellard et al. (2013) do not find a correlation between climate change and the distribution of species. Europe, North America and Australia/New Zealand will be invasion hotspots despite of climate change. This has to do with factors like disturbances and vulnerability. Hulme (2017) agrees that climate change is not the only driver of range expansions, humans are as well.

A range expansion due to climate change or humans does not mean that species become invasive. Invasions are mainly a consequence of native species that can no longer survive under the rising temperature, while newly introduced species can (Bellard et al., 2013). They can therefore take over the land which means that some species move and invade while others just migrate depending on the resistance of native species to rising temperatures (Hällfors et al., 2014). Climate change therefore blurs the definition between migration and invasion (Walther et al., 2009).

Some newly introduced species can survive in new ecosystems while others have to migrate away from ecosystems where they can no longer survive (Ibid.). During the migration or invasion some species rely on humans (Walther et al., 2009; Hällfors et al., 2014). An example is garden plant species (Walther et al., 2009; Bradley et al., 2012). Due to climate change these species have a higher chance of survival outside greenhouses. These species are often called invasive while range expanders are difficult to define. An example of a range expander can be found in box 2.

Box 2

One of the most famous examples of terrestrial plant species that is likely to shift its range under climatic change is Pueraria montana variety lobata better known as Kudzu (Forseth and Innis, 2004). This plant is mainly a problem in the South-East of the United States, where it was first introduced in 1876. Kudzu was brought to the United States for ornamental uses, food for cattle and for helping combat soil erosion. During the 20th_{century people moved to cities and} the spread of Kudzu became uncontrolled. In 2010 is was estimated that Kudzu covered three million hectares in the United States and will increase this coverage with 50.000 hectares per year. Since this last estimate Kudzu has spread itself within five extra states (see figure 4 for the current distribution).

Figure 4. Distribution of Kudzu in North America

(EDDMapS, 2021).

Kudzu interferes with natural and social practises, like forestry, agriculture, historical

(12)

11 landscapes, railroads and power supply. However, it is extremely difficult to eradicate this species and prevent their spread. It is believed that rapid elongation rates, high leaf area indices, high photosynthetic rates and frequent rooting at stem nodes are the main driver of the invasion of Kudzu.

The spread of Kuzdu is mainly limited by the amount of CO2 and low temperatures at the boundaries of their distribution area (Forseth and Innis, 2004; Bradley, Wilcove and Oppenheimer, 2010). Experts raise their concerns about the impact and distribution of this species when the temperature increases and the CO2 levels rise. Kudzu will probably extend their ranges even further under climate change.

5.4 The role of phenotypic plasticity

The effects of climate change on invasions vary between species (Bradley, Wilcove and Oppenheimer, 2010; Bellard et al., 2013). Some species are better able to deal with climate change than others because of phenotypic plasticity (Richards et al., 2006; Nicotra et al., 2010; Franks, Weber and Aitken, 2014). Phenotypic plasticity is defined as ‘the environmentally sensitive production of alternative phenotypes by given genotypes’ (DeWitt and Scheiner, 2004, p. 2). In other words, the phenotype of an organism changes depending on the environmental conditions and the genotypes. This suggests that differences in traits are a consequence of differences in environments (Van Kleunen, Weber and Fischer, 2010). It suggests that some species have traits that are better adapted to climate change than others (Nicotra et al., 2010).

Franks, Weber and Aitken (2014) reviewed 18 articles to show that there is evidence for plastic and evolutionary responses of terrestrial plants to climate change. They mention however that these responses might be insufficient to keep up with the pace of climate change. Nicotra et al. (2010) agree that species plastically respond to

climate change. They describe 11 key plastic traits which are associated with climate change, like height at maturity, seed size and number and water use efficiency. Some of these traits are also linked with IAS (Pyšek and Richardson, 2007). However, some studies do not agree on the link between phenotypic plasticity and climate change (Jump and Peñuelas, 2005; Visser, 2008). Plasticity only supports a species in the short-term and not in the long-term of climate change (Jump and Peñuelas, 2005). In addition, the optimum plasticity cannot be passed to the next generations (Visser, 2008). New generations should therefore again learn what the best adaptations are. Moreover, studies done in the light of phenotypic plasticity and climate change have some drawbacks (Merilä and Hendry, 2014). There are often more drivers for plasticity implicitly investigated and it is not always clear if the responses are based on plasticity or on evolution.

The discussion on the relation between plasticity and climate can also be seen in the light of plasticity and invasiveness. Many biologists propose that phenotypic plasticity has a great influence on invasions (Richards et al., 2006). Invaders often have a higher plasticity than natives (Daehler, 2003). This gives invaders the opportunity to invade in a wide range of environments, but it does not indicate a performance advantage of invaders over natives. Davidson, Jennions and Nicotra (2011) also conclude that invaders have a higher plasticity, but non-invasive species respond similar or even better to resource limitations than IAS. This is interesting for responses to climate change. However, Van Kleunen, Weber and Fischer (2010) do not show any difference in phenotypic plasticity between invasive and non-invasive species. They mention that the reason is that they conducted an averaged comparison of traits and not an individual comparison. This again shows that phenotypic plasticity is an individual trait. Another aspect which affects research in this subject is that some biologists

(13)

12 look at the evolutionary origin, while others investigate the ecological differences (Richards et al., 2006). Biologists believe maintenance of fitness in unfavourable environments and increase of fitness in favourable environments is crucial for these studies to incorporate so that we are able to speak the same language (Ibid.). For now, the role of phenotypic plasticity on invasiveness in the light of climate change remains unknown. Plasticity is an individual trait which makes it difficult to investigate.

6. Predicting distribution and impact

based on climate change

Climate change has multiple effects on invasive species and their distribution (see chapter 5). The question still remains how these effects change the distribution of IAS in the future. In this chapter, I will discuss which methods are used for these predictions. In addition, the drawbacks of the methods used are highlighted.

6.1 Methods

Multiple studies have used different models to predict the distribution of IAS under climate change (Murray, Stokes and Klinken, 2012; Ibáñez et al., 2014; Roger et al., 2015; Seebens et al., 2015; Bezeng et al., 2017). These models can be divided in roughly three groups: correlative approaches, process-based approaches and ensemble modelling approaches (Bellard et al., 2018). Correlative approaches are using presence data of species. Based on this presence a correlation can be found between environmental conditions and a certain species. Process-based approaches use knowledge about distributions, abundances, phenology and climatic data to predict the habitat that will suit a species. The ensemble modelling approaches define the distribution of a species by using different models and overlap them. This gives the opportunity to see where uncertainties about the distribution of species arise.

One of the most frequently used models is the species distribution model (SDM)

(Jeschke and Strayer, 2008; Wiens et al., 2009; Bradley et al., 2010). This model also goes under the name of bioclimatic model, envelope model and ecological niche model. The model can be divided in two groups: the mechanistic and empirical group (Jeschke and Strayer, 2008; Wiens et al., 2009). The mechanistic group is an example of process-based approaches which means that this group uses the literature for predictions on distributions. The empirical group is an example of correlative approaches and uses presence data for predictions on distributions. Both groups of SDM’s can be used to predict the eventual ranges of invasive alien species under climate change (Jeschke and Strayer, 2008). SDM’s gain knowledge about the climatic conditions where invasive species can flourish. By using future climate conditions, predictions can be made about the distribution of IAS in the future.

To be able to make such predictions, models need data on different climatic variables. This data is often gained by using remote sensing (Pettorelli et al., 2014; Randin et al., 2020). The data is collected by satellites and converted into maps suitable for the analysis of the distribution of IAS. More and more data is available and it has therefore been suggested to use temporal stacking to gain more precise climatic and presence layers (Randin et al., 2020). Temporal stacking means that data from different moments in time are combined to get a layer that holds more complete information.

6.2 Drawbacks of predictions

The different models described in the previous section, suggest that the possibilities for predicting the distribution of IAS under climate change are present. However, a model is in the end a simple representation of reality (Thuiller et al., 2008). Models rely on assumptions and have therefore their limitations. For models predicting IAS distribution under climate change, there are general limitations, limitations specifically for species distribution models and limitations in data gathering.

(14)

13 One of the general limitations is that the performance of a model relies on the model type (Elith and Graham, 2009; Mainali et al., 2015). Some models, like GARP, show better results than others. The performance also relies on the amount of factors taken into account in the model (Thuiller et al., 2008; Hulme, 2017; Bellard et al., 2018). For example, the effect of the stage of invasion on the distribution is not always considered or even not known (Hulme, 2017). Another example is that climate change might also affect the transport and invasion success of species, which is not included in models (Bellard et al., 2018). A more fundamental general constraint of models that predict distribution under climate change, is the question whether species can move out of their natural range. Studies have found examples of species that can (Broennimann et al., 2007; Walther et al., 2009). This means that the assumption that species live in certain environmental conditions is violated. Modelling species distribution with the knowledge of environmental preferences is therefore worthless (Araújo and Peterson, 2012). However, reviews have also shown that many species stay within their natural preferences (Parmesan and Yohe, 2003; Petitpierre et al., 2012). Another general limitation of models that predict IAS distribution under climate change is that environmental factors can have different effects on the distribution and that these factors have different effects in different habitats (Ibáñez et al., 2014). This makes the prediction of IAS distribution difficult. The last limitation is that there is a time-lag between introduction of a new species and the moment the species becomes invasive (Seebens et al., 2015).

The species distribution models and their limitations are criticized too (Jeschke and Strayer, 2008; Elith and Leathwick, 2009; Wiens et al., 2009; Bradley et al., 2010; Nicotra et al., 2010; Araújo and Peterson, 2012; Bellard et al., 2018; Manzoor, Griffiths and Lukac, 2018). The criticism is mainly directed towards the

assumptions of these models. The models assume that there is no biotic interaction, an infinite dispersal rate and an equilibrium between species and their environment (Jeschke and Strayer, 2008; Wiens et al., 2009; Araújo and Peterson, 2012). This means that species can distribute all over the world without the constraints of travelling, predation and diseases as long as the climatic conditions are the same as in their natural environment. This is why experts think that the risk of invasion is often overestimated (Bellard et al., 2018). Another difficulty of species distribution models is that they depend on the factors that are included, the interactions between factors and the scale of the environmental variables (Elith and Leathwick, 2009). The data of environmental factors defines how well experts are able to predict the IAS distribution. The availability of data can put constraints on the prediction of IAS distribution under climate change (Murray, Stokes and Klinken, 2012; Pettorelli et al., 2014; Mainali et al., 2015; Randin et al., 2020). The main limiting factor of input data is the scale of the data (Murray, Stokes and Klinken, 2012; Randin et al., 2020). The environmental data uses bigger scales for the sake of data storage. However, these scales are often too big to precisely estimate the IAS distribution in the future on a local scale. Another issue regarding data needed for models is that it is not always available as some data is not open source and other data is missing because it is not investigated (Pettorelli et al., 2014). The last important constraint of the data gathered for predicting IAS distribution under climate change, is that there are multiple scenario’s for future climate change (Murray, Stokes and Klinken, 2012). Each scenario results in different outcomes. By using uncertain input data, the outcomes also become uncertain. An example can be found in box 3.

Box 3

On the riparian zones in Australia an invasive species called lippia (Phyla canescens) causes problems (Murray, Stokes and

(15)

14 Klinken, 2012). Lippia can outcompete grasses in stressful situations and take over entire pastures. It also influences soil erosion along waterways. This plant species is used as a case study to investigate its distribution under different climatic scenario’s. In this case study Murray, Stokes and Klinken (2012) showed that the distribution of lippia will reduce under drier conditions, while it will increase with a threefold in wetter conditions. However, the results were dependent on other factors, like land-use changes and grazing. This study therefore shows that the vulnerability of Australia to lippia depends on future climate change, but also on the interaction between climatic variables and for example land-use changes. This is case study therefore shows the limitations and opportunities for species distribution models in the light of climate change.

7. Climate change in risk-assessment

Climate is often incorporated in species distribution models to see where invasive species can flourish. However, traits of species can influence the impact and introduction success of species. In this chapter, I will therefore discuss how species traits can be incorporated in the models and thus how experts might be able to include climate change in their risk-assessments. How valuable this is for management of IAS will be discussed in the last section.

7.1 Risk-assessment models

The studies that are found during this critical review only contained two studies that performed risk-assessment of IAS based on their traits and future climate change. Chai et al. (2016) performed a suitability analysis based on past and future climate conditions to see where potential invasive plant species would end up. The only species that were taken into account were species that were already non-native or invasive in the surrounding areas of Alberta. They looked at the traits of those species to investigate what the risk on the current biodiversity of Alberta would be. They

divided the traits under ecological impact, invasive characteristics, dispersal ability, and feasibility of control. It is important to notice that the traits did not influence the distribution of those species, but that the traits are only used to define the impact of the species. This in contrast to the study of Estrada et al. (2016). Estrada et al. (2016) searched for traits that influenced range shifts in the literature. The traits were divided in 5 groups: movement, ecological generalisation, persistence in unfavourable climatic conditions, reproductive strategy, and competitive ability. For each plant a value (low, moderate or high) was assigned to the different groups. This is used as input data for modelling the species distribution.

As these two examples show, it is possible to include traits and future climate change into models to predict the impact and distribution of IAS and other non-native species. Including both variables is important, because the response of IAS on climate change depends on their own traits (Chen et al., 2011; Cooke et al., 2013). However, including all those variables decreases the predictive power of models (Thuiller et al., 2008; Drenovsky et al., 2012). Besides this, it also makes the predictions more difficult. The impact of invaders becomes variable and depends on climatic settings, the invaded environment, the traits of the invading species and the interaction between those (Pyšek and Richardson, 2010). Some variables will work in conjunction which makes it more difficult to predict what will happen (Cooke et al., 2013).

One of the solution that has been found for investigating the relationship between traits and climate change, is the use of genomic techniques (Chown et al., 2015). The genomic structure of species will be compared with survival chances under different climate conditions. This is important because the structure of DNA (epigenetics) is associated with phenotypic plasticity, a trait shown to be important for invasiveness (Nicotra et al., 2015). Individuals from the same

(16)

15 species can have different epigenetics depending on their location which means that individuals from the same species can have different responses to climatic changes. This individual epigenetics are believed to be heritable (Preite et al., 2015). Phenotypic plasticity as a result of epigenetic variation between individual fulfils a leading role in invasiveness and the adaptation to different climates which should therefore be an important focus in risk-assessment models for IAS.

7.2 Management

The studies of Chai et al. (2016) and Estrada et al. (2016) show that it is possible to incorporate climate change and species traits in risk-assessments. However, there is a debate going on about the usefulness of these models in management. Some experts believe that we could use the models to prioritize management (Roger et al., 2015; Chai et al., 2016). The species with largest range shifts and the biggest impacts can be predicted and managed more carefully. Even though ecosystems are complex, the models can help to predict future distributions (Peterson, 2003). However, experts believe that we should limit the purposes for which we use models (Jeschke and Strayer, 2008; Bradley et al., 2010). The models can for example be used for species that are already known for their invasiveness or for the distribution of a species after human introduction. Araújo and Peterson (2012) emphasize that we can only use models for the present and the near future of the distribution of species. For predicting the future in the long term too many uncertain variables are taken into account to get an helpful model. Hulme (2017) is more doubtful about the use of models in risk-assessments. The models fail to incorporate trade and propagule pressure which makes them less useful. The results of the models also vary. Some models predict a decrease in the distribution of species under climate change (Bellard et al., 2013; Roger et al., 2015; Bezeng et al., 2017) while others show an increase (Chai et al., 2016; Bellard et

al., 2018). This has probably to do with the species, the input variables and the location in the world. This means that we should be careful with the application of these models (Thuiller et al., 2008). We should make a trade-off between precision and generality. The more information we put into a model, the less power a model has (Drenovsky et al., 2012). But the less information we put into a model, the less precise the model is. Figure 5 shows an example of what we need to know before we can use the models and how valuable it is to make models for certain species.

One study has found a solution for making the models more helpful (Bradley et al., 2010). By using multiple models and different scenario’s, experts can overlap the outcomes of the models. Uncertainties can be separated from clear range shifts. Models will then be more helpful for risk-assessments of IAS. A last note that should be made, is that even though models can be helpful, the definitions on invasive species and their impact is still not clear (Humair et al., 2014). This influences the risk-assessments and therefore management of IAS.

8. Discussion

In this review I investigated the value of including climate change in risk-assessments of invasive alien plant species. As some studies have shown, it is possible to do so (Chai et al., 2016; Estrada et al., 2016). However the value of this is under debate. The main reason for this debate is that studies show different results in the effects of climate change on the distribution and impact of IAS (Bradley, Wilcove and Oppenheimer, 2010; Bellard et al., 2013). There are different issues affecting these results. The first issue is that input variables are not always sufficient for the analysis (Murray, Stokes and Klinken, 2012; Pettorelli et al., 2014). In some cases the data is not available and in other cases the scales of the input variables do not match the prediction scales. Besides this, the future of climate change is uncertain (Murray, Stokes and Klinken, 2012). Using these uncertain variables

(17)

16 as input data for a model, results in an uncertain outcome. A second reason is that variables sometimes work in conjunction (Cooke et al., 2013). This means that the combination of different variables results in different outcomes of the same model. A third and last issue is that the responses of species to climate change is based on their own traits

(Chen et al., 2011; Cooke et al., 2013). Experts are not sure which traits are helpful and which are not. They also discuss which traits are helpful in the invasion process (Pyšek and Richardson, 2007). This means that the invasion of new areas is context- and species dependent which affect the outcomes of

risk-Figure 5. Schematic overview of the knowledge that has to be known before models can be used for management. Without the knowledge of the interaction between species traits and climatic variables, models can be used for management although the results will be less accurate. For each species group the value of making a mode for this species is allocated (low, moderate or high). This value is based on the risk that a species becomes invasive in the area of interest.

(18)

17 assessments. The use of models for risk-assessments of IAS is therefore questionable.

The outcomes of the models might be different and they might dependent on multiple variables, but experts can still use the models for prioritizing management (Roger et al., 2015; Chai et al., 2016). The models allow us to make rankings of species that are most likely to expand their ranges and have the most impact (see for example Chai et al., 2016). As Chai et al. (2016) explains, it is better to do this for species that are already invasive in surrounding areas or species that are non-native in the area. This species form the greatest threat to the environment and a lot of data is available for invasive species. This makes it easier to model these species. The modelling should also include different scenarios of climate change, so we can see what happens with a species under varying circumstances (Bradley et al., 2010). This is important because the future of climate change is uncertain which makes the outcome of a model also uncertain. By combining the results of different scenarios we will be able to see the uncertainties in the predictions of the risk-assessments (Bellard et al., 2018). Finally, it is crucial to use a general framework for these assessments, so that studies can be compared. The distribution and impact of IAS under climate change will therefore become more clear. An example for such a general framework can be found in the study of Estrada et al. (2016).

Although, this sounds hopeful for the future of risk-assessments of IAS, we should still be careful and we should keep in mind that management cannot fully rely on risk-assessments made by models. For this reason management should also focus on preventing species to become invasive. This can be done by limiting the intentional human-induced dispersal of species, because this is one of the biggest reasons why species can travel over great distances and become invasive (Nathan, 2006; Heywood, 2011; Richardson and Rejmanek, 2011). For example, native species

can be used for horticultures or gardens instead of alien species. Preventing species from becoming invasive can also be done by reducing human disturbances of land (Richardson and Pyšek, 2012). This has a positive effect on native species which makes an ecosystem less vulnerable for invasive species.

The solutions for IAS risk-assessments and management proposed in this review are just the beginning of exploring the options of preventing terrestrial plant species from becoming invasive. Future research is needed and becomes more important in predicting the distribution and impacts of future IAS. Research is needed to investigate the role of species specific traits on surviving under different climate conditions, because species respond differently on climate change based on their own traits (Chen et al., 2011; Cooke et al., 2013). One of the most promising methods are genomic techniques (Chown et al., 2015). There might be a relation between DNA and the survival chances of plants. The structure of DNA influences the phenotypic plasticity of individual plants, a trait often associated with invasiveness (Nicotra et al., 2015). In addition, more research should be done to better understand the effects of different climate variables working in conjunction on future IAS. The reason for this is that the combination of climate variables results in different outcomes of the distribution of species (Elith and Leathwick, 2009; Cooke et al., 2013). Without knowledge on this topic, distribution ranges of IAS are predicted inaccurately. Besides this, we also need an improvement of data availability. Climate variables are not always freely accessible which makes data collection for models difficult (Pettorelli et al., 2014). An improvement can also be made with the scales of climatic variables. These regional or global scales are often too big to make detailed predictions on species distribution on local scales (Murray, Stokes and Klinken, 2012; Randin et al., 2020). Research on climatic

(19)

18 variables on smaller scales therefore needs to be improvement.

The battle against invasive terrestrial plant species is not over yet. We can use models to support management, but more important is that management should focus on reducing the human spread of species all over the world.

Acknowledgement

First of all, I would like to thank my examiner, dr. Elly Morriën, for the great help and suggestions she has for this review. The classes of dr. Carina Hoorn, my assessor, also really helped me with writing this review in a narrative way. Last of all, I would like to thank everyone who gave feedback on my work to improve it. So thank you dr. Elly Morriën, dr. Carina Hoorn, Naomi Buijs BSc, Tijl de Bruijn and Danielle Gomes BSc.

References

Adhikari, A. et al. (2011) ‘Terminologies Used in Climate Change’. IUCN. Available at: https://www.iucn.org/sites/dev/files/import/ downloads/terminologies_used_in_climate_c hange_2011.pdf (Accessed: 5 March 2021). Araújo, M. B. and Peterson, A. T. (2012) ‘Uses and misuses of bioclimatic envelope modeling’,

Ecology, 93(7), pp. 1527–1539. doi:

https://doi.org/10.1890/11-1930.1.

Arroyo, M. T. K. et al. (2000) ‘Plant invasions in Chile: present patterns and future predictions’, Invasive species in a changing world, pp. 385– 421.

Bellard, C. et al. (2013) ‘Will climate change promote future invasions?’, Global Change

Biology, 19(12), pp. 3740–3748. doi:

10.1111/gcb.12344.

Bellard, C. et al. (2018) ‘Insights from modeling studies on how climate change affects invasive alien species geography’, Ecology and

Evolution, 8(11), pp. 5688–5700. doi:

https://doi.org/10.1002/ece3.4098.

Bellard, C., Cassey, P. and Blackburn, T. M. (2016) ‘Alien species as a driver of recent extinctions’, Biology letters, 12(2), p. 20150623.

Bezeng, B. S. et al. (2017) ‘Climate change may reduce the spread of non-native species’,

Ecosphere, 8(3), p. e01694. doi:

10.1002/ecs2.1694.

Blackburn, T. M. and Jeschke, J. M. (2009) ‘Invasion success and threat status: two sides of a different coin?’, Ecography, 32(1), pp. 83– 88. doi: https://doi.org/10.1111/j.1600-0587.2008.05661.x.

Bradley, B. A. et al. (2010) ‘Predicting plant invasions in an era of global change’, Trends in Ecology & Evolution, 25(5), pp. 310–318. doi: 10.1016/j.tree.2009.12.003.

Bradley, B. A. et al. (2012) ‘Global change, global trade, and the next wave of plant invasions’, Frontiers in Ecology and the

(20)

19

Environment, 10(1), pp. 20–28. doi:

https://doi.org/10.1890/110145.

Bradley, B. A., Wilcove, D. S. and Oppenheimer, M. (2010) ‘Climate change increases risk of plant invasion in the Eastern United States’, Biological Invasions, 12(6), pp. 1855–1872. Broennimann, O. et al. (2007) ‘Evidence of climatic niche shift during biological invasion’, Ecology Letters, 10(8), pp. 701–709. doi:

https://doi.org/10.1111/j.1461-0248.2007.01060.x.

Bushell, S., Colley, T. and Workman, M. (2015) ‘A unified narrative for climate change’, Nature Climate Change, 5(11), pp. 971–973. doi: 10.1038/nclimate2726.

Catford, J. A. et al. (2012) ‘The intermediate disturbance hypothesis and plant invasions: Implications for species richness and management’, Perspectives in Plant Ecology, Evolution and Systematics, 14(3), pp. 231–241. doi: 10.1016/j.ppees.2011.12.002.

Catford, J. A., Jansson, R. and Nilsson, C. (2009) ‘Reducing redundancy in invasion ecology by integrating hypotheses into a single theoretical framework’, Diversity and Distributions, 15(1),

pp. 22–40. doi:

https://doi.org/10.1111/j.1472-4642.2008.00521.x.

Chai, S.-L. et al. (2016) ‘Using Risk Assessment and Habitat Suitability Models to Prioritise Invasive Species for Management in a Changing Climate’, Plos One, 11(10), p. e0165292. doi: 10.1371/journal.pone.0165292.

Chen, I.-C. et al. (2011) ‘Rapid Range Shifts of Species Associated with High Levels of Climate Warming’, Science, 333(6045), pp. 1024–1026. doi: 10.1126/science.1206432.

Chew, M. K. and Laubichler, M. D. (2003) ‘Natural Enemies--Metaphor or Misconception?’, Science, 301(5629), pp. 52– 53. doi: 10.1126/science.1085274.

Chown, S. L. et al. (2015) ‘Biological invasions, climate change and genomics’, Evolutionary

Applications, 8(1), pp. 23–46. doi:

https://doi.org/10.1111/eva.12234.

Colautti, R. I. et al. (2004) ‘Is invasion success explained by the enemy release hypothesis?’, Ecology Letters, 7(8), pp. 721–733. doi:

https://doi.org/10.1111/j.1461-0248.2004.00616.x.

Colautti, R. I. and MacIsaac, H. J. (2004) A neutral terminology to define ‘invasive’ species.

Available at:

https://onlinelibrary.wiley.com/doi/full/10.11 11/j.1366-9516.2004.00061.x (Accessed: 22 February 2021).

Cooke, S. J. et al. (2013) ‘What is conservation physiology? Perspectives on an increasingly integrated and essential science’, Conservation

Physiology, 1(1), p. cot001. doi:

10.1093/conphys/cot001.

Daehler, C. C. (2003) ‘Performance Comparisons of Co-Occurring Native and Alien Invasive Plants: Implications for Conservation and Restoration’, Annual Review of Ecology, Evolution, and Systematics, 34(1), pp. 183–211. doi:

10.1146/annurev.ecolsys.34.011802.132403. Davidson, A. M., Jennions, M. and Nicotra, A. B. (2011) ‘Do invasive species show higher phenotypic plasticity than native species and, if so, is it adaptive? A meta-analysis’, Ecology

Letters, 14(4), pp. 419–431. doi:

10.1111/j.1461-0248.2011.01596.x.

DeWitt, T. J. and Scheiner, S. M. (2004)

Phenotypic plasticity: functional and

conceptual approaches. Oxford University Press.

Diez, J. M. et al. (2012) ‘Will extreme climatic events facilitate biological invasions?’, Frontiers in Ecology and the Environment, 10(5), pp. 249–257. doi: https://doi.org/10.1890/110137.

Djebou, D. C. S. and Singh, V. P. (2016) ‘Impact of climate change on precipitation patterns: a comparative approach’, International Journal

The prediction of invasive plant species based on their traits and climate change