Planificación

invasion success 3.3 Species traits as predictors of invasion success 1858 1924 1861 1841 1856 1855 1858 1883 1914 1807 1809 1821 1924 1898 1870

Fig. 3.2 The shrub Lantana camara, an example of a very successful invader, was deliberately transported from its native

range (shaded area) to widely dispersed subtropical and tropical locations where it spread and increased to pest propor- tions. Year of arrival is shown where invasions have occurred. (After Cronk & Fuller, 1995.)

66 PART 1 ECOLOGICAL APPLICATIONS AT THE LE VEL OF INDIVIDUAL ORGANISMS

The success of some invasive taxa has a strong element of predictability. Of 100 or so introduced pine species in the USA, for example, those that have successfully spread into native habitats are characterized by small seeds, a short interval between successive large seed crops and a short juvenile period (Rejmanek & Richardson, 1996). It is interesting to note in passing that conifer species classifi ed as rare and endangered, such as Pinus maximartinezii, are the precise converse of their invader cousins, having large seeds, long juvenile periods and long intervals between large seed crops (Richardson & Rejmanek, 2004). Among pine taxa, in other words, the invaders tend to have r-selected traits while the endangered species exhibit K-selected traits (Box 3.1).

Grotkopp et al. (2002) delved more deeply into the suite of traits responsible for invasiveness in pine species. They germinated seeds of 29 species under carefully standardized laboratory conditions and harvested seedlings periodically after each species had reached the stage of expanding their fi rst true leaves. Then for 10 weeks they recorded the weights of leaves and other plant parts, and measured the com- bined area of the seedlings’ leaves, as well as their mean width and density (dry weight per unit volume). These data allowed them to calculate a number of plant performance traits (including, among others, relative growth rate (RGR: mg gplant−1

day−1) and leaf area ratio (LAR: cm2

leaf gplant−1)). Finally, they performed a Principal

Components Analysis (akin to the ordination analyses you encountered fi rst in Section 2.2.3) to explore the relationships between plant traits and invasiveness (Figure 3.3).

The invasive pine species are clearly clumped together to the right on PCA axis 1 and the noninvasive taxa to the left in Figure 3.3. Invasiveness is correlated most strongly, and positively, with relative growth rate (also to the right on axis 1), and negatively with seed mass and juvenile period (to the left on axis 1). In other words, invaders tend to have high growth rates, small seeds and short juvenile periods (i.e. short generation times). The infl uential high relative growth rate is itself a consequence of high net assimilation rates, leaf mass ratios and, in particular, spe- cifi c leaf areas. Our example shows how the details of plant physiology and morphol- ogy can help explain, and predict, invasion potential.

In New Zealand there is a precise record (similar to that of the pines in Section 3.3.1) of successes and failures of attempted bird introductions. These species were shipped out by nineteenth-century European colonists to make their new surround- ings more like ‘home’.

Sol and Lefebvre (2000) found that invasion success of the introduced birds increases with introduction effort (number of attempts and number of individuals introduced since European colonization). This is hardly surprising because the success of even the best fi tted of potential invaders is probabilistic rather than defi - nite. Invasion success is also higher for species whose young do not need to be fed by their parents (such as game birds), species that do not migrate and, in particular, birds with large brains. Animals with relatively large forebrains are thought to cope better with environmental complexity and to respond more rapidly to changes in their environment. In fact, the successful bird invaders have more reports in the international literature of adopting novel food or feeding techniques (mean for 28 species 1.96, SD 3.21) than the unsuccessful species (mean for 48 species 0.58, SD 1.01). For example, among closely related species the successfully invading rook 3.3.1 Species traits predict invasive conifers 3.3.1 Species traits predict invasive conifers 3.3.2 Invasion success – the importance of fl exibility 3.3.2 Invasion success – the importance of fl exibility

LIFE-HISTORY THEORY AND MANAGEMENT CHAPTER 3 67

(Corvus frugilegus) has fi ve reported novel behaviors while the unsuccessful Corvus monedula only has two. Similarly, the successful house sparrow (Passer domesticus) and mallard duck (Anas platyrhynchos) have ten and fi ve reported novel behaviors, respectively, compared with just two each for their unsuccessful counterparts Passer montanus and Anas penelope.

In this disparate group of birds, and in contrast to the pine species, r/K status does not appear to predict invasive potential. Success seems to be linked with high competitive status (analogous to Grime’s C categorization in Box 3.1), an advantage that is conferred by behavioral fl exibility.

Something similar may be the case for certain plant invaders too. Japanese grass (Microstegium vimineum) is an important annual weed of Asian origin that is now widespread east of the Mississippi River in the USA and in many other countries. Gibson et al. (2002) studied four populations in quite different environments in Illinois, USA. They concluded that the species’ invasive potential is fueled by a remarkably fl exible response to local microhabitat conditions, being able to tolerate acidic, sandy and silty soils in both open and shaded locations. Such plastic- ity, akin to the behavioral fl exibility of birds, may be a property of many invaders.

(a) (b) Axis 2 Axis 2 Axis 1 Axis 1 Leaf thickness Seed mass Leaf density NAR LMR RLPR RGR LAR SLA Generation time pinea palustris lambertiana coulteri pinaster ponderosa jeffreyi canariensis nigra caribaea resinosa thunbergii muricata halepensis radiata sylvestris rigida elliottii taeda glabra virginiana contorta banksiana patula strobus cembra sabiniana flexilis torreyana Invasive Unclassified Noninvasive Fig. 3.3 Results of a Principal Components Analysis (PCA) showing (a) the relationships among species traits of 29 species of pine (Pinus) (b) known to be invasive or noninvasive. The positions of the species and the trends in species traits are shown in separate panels for simplicity of presentation. Key for plant traits: relative growth rate (RGR: mg gplant−1 day−1), net assimilation rate (NAR: mg cm−2leaf day−1), leaf mass ratio (LMR: gleaf gplant−1), relative leaf production rate (RLPR: leaf leaf−1 day−1), leaf area ratio (LAR: cm2_{leaf gplant}−1_{) and} specifi c leaf area (SLA: cm2_{leaf gleaf}−1_{). The} invasive and noninva- sive species cluster in different parts of the diagram. (After Grotkopp et al., 2002.)

68 PART 1 ECOLOGICAL APPLICATIONS AT THE LE VEL OF INDIVIDUAL ORGANISMS

Physiological and behavioral fl exibility can be equated to possession of a broad (or generalist) niche (Sections 2.4.2 and 3.2.1).

Species invasion is a complex process with several identifi able stages: (1) transport from the native range; (2) release into a new location; (3) establishment of a self- sustaining population in the new location; (4) spread beyond the site of fi rst estab- lishment; and (5) impact in the receiving community. Both the invasive pine study (Section 3.3.1) and that of the successfully introduced New Zealand birds (Section 3.3.2) fi t into category (4). Generally speaking, managers are most concerned about species like these, regarding as truly invasive those that spread beyond their original establishment site. However, to increase predictability in invasion ecology, Kolar and Lodge (2001) suggest that species’ characteristics should be examined in rela- tion to all the stages of invasion.

Cassey et al. (2004) assessed the fi rst three of the fi ve phases – (1) transport, (2) release and (3) establishment – in an analysis of all of the world’s 350 parrot species. Transporting and selling parrots is big business and populations that establish in new locations are easily spotted and quickly recorded by ornithologists. For these reasons parrots provide an excellent case study to determine species traits that cor- relate with transport (species commonly transported outside their natural range score 1, others 0), release (species intentionally released outside their range score 1, those transported but not released score 0) and establishment (the proportion of releases leading to successful population establishment).

Cassey’s team found that different sets of variables were related to the probability that a species will enter each stage on the pathway through transport to establish- ment (Table 3.2). Large species are more likely to be transported. More signifi cantly, species that are transported and released outside their native range tend to be wide- spread ones that are traded to be kept in captivity or as pets. Presumably this simply refl ects the need for a predictable supply of desirable trade species. Threatened species do not fi gure in the groups that are transported and released (note the nega- tive correlation) – illegal trade in some rare species certainly occurs, but such indi- viduals are not released into new areas.

Of more interest was the fi nding that species with broader diets and larger alti- tudinal ranges (i.e. with broader niches) were more likely to establish in a new area, as were those with longer fl edging periods, perhaps because of the head start gained from an extended period of parental care. Migratory species, as with the New Zealand birds, were less likely to establish. You might have predicted, on fi rst prin- ciples, that migratory species would, like ruderal plants, possess some good colonizer traits and be effective invaders. However, it turns out that the complex requirements of migratory species – innate migration cues, learnt migration routes, appropriate juxtaposition of seasonal habitats – do not suit them for invasions, at least not those of the human-assisted kind.

Pysek et al. (2003) also have something to say about the fi rst three stages of inva- sion based on their analysis of fi rst records of 668 alien plants in the Czech Republic since the year 1500. Deliberately introduced species with utilitarian value (e.g. culinary or medicinal) arrived before ornamental species, and ‘accidental’ arrivals have been most recent. Moreover, species of European origin arrived earlier than those from North America, with Asian and Australasian species turning up in more modern times. In their analysis of the relationship between timing of fi rst records 3.3.3 Separating

invasions into sequential stages – different traits for each?

3.3.3 Separating invasions into sequential stages – different traits for each?

LIFE-HISTORY THEORY AND MANAGEMENT CHAPTER 3 69

and species traits, Pysek’s team considered type of life history (annual, biennial, herbaceous perennial, woody perennial), fl owering time, mode of dispersal (water, wind, animal assisted) and Grime’s CSR strategy. The strongest relationship was with Grime’s strategy: species having the unusual but highly fl exible CSR strategy (most features of competitiveness, stress tolerance and ruderality incorporated in a single organism) arriving earlier than CS species, which in turn arrived earlier than C or CR species. The pattern of accumulation of species with each strategy is shown in Figure 3.4. (Note that the early arrival of CSR strategists is not obvious in this fi gure because of the small number of species of this type.) By far the most com-

Variables Transport Release Establishment

Ecological fl exibility Diet breadth _{+* +* +**} Altitudinal range _+* Latitudinal range _{+*** +***} Life history Body mass +* Annual fecundity Age at maturity Fledging period +** Migration −* −** Dichromatism +* Movement International trade _{+*** +*** +**} To be kept in captivity _{+*** +***} To be kept as pets _{+*** +***} Population size/extent Threat status −** −* Population size +** +* Geographical range +*** +**

+, Positive relationship; −, negative relationship.

Asterisks indicate the level of statistical signifi cance of the results for each trait in each year – more asterisks mean a greater level of confi dence in the result: *p < 0.05, **p < 0.01, ***p < 0.001, blank not signifi cant.

Table 3.2 Tests of the

degree to which species variables correlate with the likelihood of parrot species being trans- ported from their native range, released into a new area and establish- ing in the new location. (Modifi ed from Cassey et al., 2004.)

Table 3.2 Tests of the

degree to which species variables correlate with the likelihood of parrot species being trans- ported from their native range, released into a new area and establish- ing in the new location. (Modifi ed from Cassey et al., 2004.)

Fig. 3.4 Increase with

time in the cumulative number of alien plants in the Czech Republic corresponding to Grime’s CSR strategy combinations. Refer to Box 3.1 for defi nitions of C (competitive), S (stress-tolerant) and R (ruderal). (From Pysek et al., 2003.)

160

Cumulative number of species

140 120 100 80 60 40 20 0 1750 1800 1850 1900 1950 2000 C CR R CSR CS S SR

70 PART 1 ECOLOGICAL APPLICATIONS AT THE LE VEL OF INDIVIDUAL ORGANISMS

Establishment

Response (self-sustaining Spread (number Impact (average

variables population) of catchment areas) abundance)

Propagule pressure 1.00 0.83 0.34 (numbers involved in introduction attempts) Physiological 1.00 0.82 0.08 tolerance

Adult maximum size 1.00 0.38 0.09

Distance from 1.00 0.23 0.02

source

Size of native range 1.00 0.10 0.07

Prior invasion 1.00 0.10 0.76

success

Trophic status 1.00 0.00 0.00

Parental care 1.00 0.00 0.00

Table 3.3 Weight of

evidence to suggest that a particular response variable is important in predicting ‘success’ at each stage of invasions of freshwater fi sh in Californian river catchment areas. The larger the ‘weight’, the more important is the variable in predicting invasion across all possible models created; maximum weight is 1.0. (After Marchetti et al., 2004.)

Table 3.3 Weight of

evidence to suggest that a particular response variable is important in predicting ‘success’ at each stage of invasions of freshwater fi sh in Californian river catchment areas. The larger the ‘weight’, the more important is the variable in predicting invasion across all possible models created; maximum weight is 1.0. (After Marchetti et al., 2004.)

monly represented strategies among species now established are competitive (C) or competitive-ruderal (CR) combinations of traits.

Marchetti et al. (2004) take up where the parrot and plant studies leave off. They compiled information on invasions and failed introductions of fi sh in catchment areas throughout California and separate the invasion process into the three fi nal stages of our scheme: (3) establishment (of a self-sustaining population); (4) spread (assessed as number of catchment areas occupied); and (5) impact (assessed as average abundance attained). Their analysis includes a variety of species traits, from physiological tolerance and adult size to prior invasion success. The process involves a comparison of all conceivable models to predict species ‘success’ at the three stages, on the basis of the species traits. The importance, or weight, of particular traits was estimated across all models, and the larger the weight, the more important the variable in predicting the response of interest (Table 3.3).

All eight traits in the analysis are important for predicting species establishment. This is not really surprising as the traits were chosen, according to fi rst principles, because they might be expected to be infl uential. Establishment is more likely for larger species (more competitive), those showing broad physiological tolerances and diets, parental care, a large native range (broader niches), a nearby native source, a record of successful invasions elsewhere and where large numbers were involved in introduction attempts.

Prediction of the propensity to spread, on the other hand, depends mainly on physiological tolerance and adult size, as well as the effort put in to establishing the species in the fi rst place. Once again, and unlike the pine species, successful fi sh invaders do not fall consistently at one end of the r/K continuum. Competitive status (large size) seems to play a role, but the overriding ecological variable is broad physiological tolerance, probably increasing the likelihood, in the probabilistic

LIFE-HISTORY THEORY AND MANAGEMENT CHAPTER 3 71 world of invasion success, that these species would encounter a match with appro- priate environmental conditions.

Finally, the likely size of invader impact, assessed in terms of the abundance achieved by populations, is not well predicted by species traits, but rather in terms of prior invasion success and establishment effort.

We know rather a lot about the relationship between species traits and invasion success but, regrettably, only for a small proportion of the world’s biota. The most informative databases are those concerning taxonomic groups for which informa- tion is available about species that have become successful invaders as well as those that have not. This tends to apply to taxa that are important to particular segments of society, and for which particularly good records have been kept – horticulturalists and gardeners (pine trees and Czech plants), those in the pet trade (parrots), coloni- als who wished to make themselves at home (New Zealand Acclimatization Society – introduced birds) and angling organizations (freshwater fi sh). These databases are enormously valuable, providing some of the best information we have to assess the causes of invader success.

However, managers need to beware of unwarranted generalization. We do see indications of predictability of invasion success for some taxa, related to high repro- ductive output (e.g. pine seed production), fl exible lifestyles (bird behavior), broad niches (e.g. parrots) and competitive strategies (e.g. Czech plants). But exceptions to the ‘rules’ are common and there are many cases where no relationships have been found, prompting Williamson (1999) to wonder whether invasions are any more predictable than earthquakes. Commonly, the best predictor of invasion success is previous success as an invader elsewhere. Looking on the bright side, even this provides invasion managers with useful pointers when prioritizing poten- tial invaders to their regions.

Having discovered the life-history traits that are linked to the probability of invasion or of success in restoration projects, I now turn attention to the suites of traits that can be used to predict the vulnerability of native species to human pressures. This question is vital to conservation biologists who must understand what makes certain species prone to extinction.

Successful invaders naturally have much in common with successful restoration. The only distinction is that we have different names for the species we do (restora- tion) or don’t (invasion) want to become established. In Sections 3.2 and 3.3 you saw that traits associated with niche breadth and ecological fl exibility could some- times predict the probability of successful establishment, as could suites of traits associated with the r/K concept (small, rapidly reproducing vs large, slow-growing, competitive organisms) or the CSR concept (particularly R species with good powers of colonization and C species with strong competitive traits).

Vulnerability to extinction is a different kettle of fi sh because now we wish to ensure that a species is not lost from its native environment in the fi rst place. The list of human pressures that pose risks (and that are responsible for past extinctions, both global and local) is long and diverse. It includes habitat loss and habitat frag- mentation as a result of land-use change (e.g. pastoral grazing, forestry, freshwater habitat disruption), habitat degradation as a result of pollution and overharvesting of wild populations. Moreover, many species are confronted by more than just one 3.3.4 What we know

and don’t know about invader traits

3.3.4 What we know and don’t know about invader traits 3.4 Species traits as predictors of extinction risk 3.4 Species traits as predictors of extinction risk

72 PART 1 ECOLOGICAL APPLICATIONS AT THE LE VEL OF INDIVIDUAL ORGANISMS

of these pressures. In this section I ask whether the varied situations have anything in common. Are there, in fact, suites of species traits that seem to ‘protect’ native species or, alternatively, that make them more likely to succumb to the pressures? And are these traits the same as those that predict successful invasion or restoration?

First, I consider cases where niche breadth and fl exibility play an important role (Section 3.4.1) before turning to the many situations where the r/K concept again seems to come into its own (Section 3.4.2). In a concluding section, the importance of CSR traits, among others, will be considered (Section 3.4.3).

Managers would be better able to prioritize species for conservation intervention if it were possible to predict, on the basis of species traits, those most at risk of extinc- tion. With this in mind Angermeier (1995) analyzed the traits of 197 native fresh- water fi sh in Virginia, USA, paying special attention to the characteristics of the 17