Representation of species on the list and differences between regions
The list (Appendix S2), comprising 622 species, represents a tiny proportion of the global woody plant flora that comprises probably around 60,000 (current estimates in the literature range from 50,000 to 100,000) species of ‘trees’ and approximately the same number of ‘shrubs’ (perhaps only 30,000 species). Using these rough numbers, we suggest that only between 0.5% and 0.7% of the global pool of tree and shrub species are currently clearly invasive outside their natural range.
Cursory examination of the list reveals a strong bias in favour of temperate species with obvious usefulness to humans and a strong bias against tropical species. Colonial history has played an important part in the dissemination of woody plants around the world (Crosby, 1986; Spongberg, 1990; Taylor, 2009; Laws, 2010). Consequently, the positions of regions in Fig. 4 and the level of similarity between regions are clearly influenced by historical/cultural factors over the past few centuries. More recently, intentional and co-ordinated transfers for specific purposes such as forestry (in the broad sense) and horticulture have dominated invasion pathways, and these are starting to blur the effects of older introductions. Woody plants from Australia (especially species in the families Fabaceae, Myrtaceae and Proteaceae) have been very widely moved around the world (many of them recently) and are fairly well represented on the main list, (c. 8% of species in Appendix S2), and on the list of widespread invaders (7 of 38 species in Table 2; 18%). There has been far less movement of woody plants from some other parts of the world with rich tree and shrub floras, notably China. This trend is, however, very likely to change in the next few decades (Kunming Institute of Botany, 2003; Normile, 2004).
Numbers of invasive alien trees and shrubs vary considerably between regions of the world, although it is difficult to determine whether the patterns reflect the real extent of invasions and to what extent the patterns are affected by different levels of reporting and the availability of accurate data on the status of species in different regions. Most regions with > 100 known invasive trees and shrubs (Table 1) are places with long histories of introductions and where invasions are generally well studied. Regions with < 100 species are places generally under-represented in terms of the intensity of research on biological invasions (Pyšek et al., 2008). The pattern probably reflects predominantly the magnitude of introductions and plantings (high propagule pressure) and the level of effort devoted to reporting on invasive species, rather than any real difference between regions in overall invasibility. Introductions that have taken place only in the last few decades (see e.g. Grimshaw & Bayton, 2009; Shulkina, 2004; Wharton et al., 2005) have not had time to generate invasions, and there is undoubtedly a substantial ‘invasion debt’ in all regions, especially more affluent regions.
An important feature of the list is the strong under-representation of many well-known families with a large proportion of woody species. Such families that have not (yet) contributed many invasive species include Anacardiaceae [850 species, including c. 200 Rhus sensu lato (including Searsia and Toxicodendron)]; Annonaceae (2100); Betulaceae (140 species, including 35 species in Alnus and 35 in Betula); Burseraceae (640), Chrysobalanaceae (530 species); Combretaceae (525 species); Dipterocarpaceae (535 species); Ericaceae [3850 species, including c. 1000 Rhododendron (650 in China) and 860 Erica species]; Ebenaceae (575); Euphorbiaceae (6500 species, > 60% are trees and shrubs); Fagaceae (970 species, including 34 in Nothofagus and 530 in Quercus); Lauraceae (2550 species); Lecythidaceae (325); Magnoliaceae (221 species); Malvaceae (including Bombacaceae, Sterculiaceae and Tiliaceae; 5000 species, mostly trees and shrubs); Meliaceae (650); Moraceae (1150 species; 850 Ficus spp.); Myristicaceae (520); Proteaceae (1775 species, including 77 Banksia, 149 Hakea and 103 Protea species); Rubiaceae (11,000 species, > 95% of them are trees and shrubs); Sapindaceae (1450 species, including 114 Acer species); and Sapotaceae (975) (numbers of species from Mabberley, 2008).
Many large, particularly tropical, woody genera are clearly under-represented. Examples (with number of known invasive species/total number of species) are Psychotria (0/1850), Piper (1/1050), Rhododendron (1/1000), Erica (4/860), Ficus (4/850), Eucalyptus (7/750), Schefflera (1/600), Ixora (0/560), Quercus (3/530), Ilex (1/400), Vaccinium (1/450), Baccharis (1/350), Clusia (1/300+), Litsea (0/300+), Inga (1/300), Lithocarpus (0/300), Melalaeuca (4/250), Licania (0/220), Magnolia (0/220), Ocotea (0/200), Palicourea (0/200), Persea (1/200), Pouteria (0/200), Shorea (0/200), Terminalia (1/200), Zanthoxylum (0/200), Casearia (0/180), Homalium (0/180), Rinorea (0/170), Lasianthus (0/170), Commiphora (0/150), Oreaopanax (0/150), Calliandra (1/130), Faramea (0/130), Camellia (0/120), Lonchocrpus (0/120), Coccoloba (0/120), Nectandra (0/120), Hirtella (0/110), Hopea (0/100) and Lindera (0/100). On the other hand, some genera are over-represented in our database. These are mostly relatively small genera, e.g. Casuarina (3/17), Schinus (3/33), Ligustrum (8/40), Fraxinus (7/42), Prosopis (5/44), Tamarix (4/54) and Pinus (22/110).
Species from many genera and families have not been sufficiently widely transported and disseminated around the world for long enough and in large enough numbers to give them a chance to invade. This clearly complicates the quest to evaluate the ongoing natural experiment to provide ecological reasons for taxonomic biases in the list. Very few woody plant groups have been surveyed in enough detail to assess the levels of invasiveness in relation to the degree of transport and dissemination outside their natural ranges.
A few taxonomic groups on the list have, however, been sufficiently well disseminated and the determinants of invasiveness well enough studied to allow for at least preliminary judgements to be made regarding the distribution of invasiveness across the whole group. The most notable group in this regard is the clade Pinophyta, for which enough evidence is available to allow for reasonably robust conclusions to be drawn on the determinants of invasiveness, taking into account life-history traits, propagule pressure and facets of invasibility. For this group, a syndrome of life-history traits [small seed mass (< 50 mg), short juvenile period (< 10 years) and short intervals between large seed crops] separates the most invasive species from others with less potential to invade (Rejmánek & Richardson, 1996; Richardson & Rejmánek, 2004). The discriminant function derived from the life-history traits of invasive and non-invasive pines was later incorporated, together with other biological attributes, into general rules for the detection of invasive woody seed plants (Table 6.1 in Rejmánek et al., 2005; Table 1 in Rejmánek, 2011). To date, this is the only risk-assessment procedure based exclusively on biological attributes of tested woody plant species. Although an ecological syndrome associated with inherent invasiveness clearly exists for this group, good evidence has also emerged that the elucidation of the expression of invasiveness in this taxon must incorporate extrinsic factors such as propagule pressure and residence time (Richardson et al., 1994; Procheşet al., 2011). Another group for which considerable insights are now available is Acacia (sensu lato) and in particular taxa in Acacia subgenus Phyllodinae native to Australia (‘Australian acacias’) (Box 1). Eucalypts (the genera Angophora, Corymbia and Eucalyptus in the Myrtaceae) have been exceptionally well disseminated and widely planted for well over a century in many parts of the world. No clear ecological syndromes favouring invasiveness have been discovered in this group (Rejmánek & Richardson, 2011), and surprisingly, few species are listed as invasive (only eight species; Appendix S2; Table Box 1). The extent of invasiveness of eucalypts in particular regions is well explained only by metrics that describe the magnitude and duration of plantings (Rejmánek et al., 2005). We suggest that the situation for pines and eucalypts probably represents opposite endpoints on a continuum from ecological/phylogenetic/taxonomic mediation of invasiveness on the one end (exemplified by pines), to mediation driven primarily by factors related to propagule pressure (with eucalypts as exemplar). Other factors relating to the composition of the list, with implications for understanding current invasions and predicting future invasions, are discussed in the following sections.
Reasons for introduction and dissemination
The reasons for introduction and use of non-native plants are important for evaluating the levels and patterns of invasiveness, as cultivation practices fundamentally shape invasion pathways (Wilson et al., 2009b). The use of plants in horticulture provides a very effective means of dissemination, as plants are cultivated and nurtured (protected from effects of disturbance, allowing plants to attain maturity and accumulate large propagule banks), often in large numbers, at scattered foci, often close to a wide range of potentially invasive habitats. Horticultural plants are frequently selected for attributes that are closely associated with invasiveness, such as long-lasting displays of brightly coloured flesh fruits attractive to a wide range of generalist seed dispersers (Reichard, 2011). The large number of species introduced for horticulture in our global list mirrors the dominance of horticultural species in many regional lists of woody invasive plants (Essl et al., 2011). Species used for forestry are selected for fast growth (one of a package of traits typically associated with species with adaptations for rapid colonization and thus inherent ‘weediness’; Grotkopp et al., 2010) and are typically grown in large plantations, allowing for the accumulation of massive propagule banks. Woody plants most widely used in agroforestry are selected for their tolerance of a wide range of conditions, rapid growth and frequently precocious and prolific fruiting and/or seed production. They are often grown in highly disturbed areas. These criteria define the introduction and dissemination pathways for these species. These, and the role of cultivation methods in mediating invasiveness, are fundamental filters that have resulted in the patterns of occurrence shown in Appendix S2 and Fig. 4. There is a significant rank correlation between number of uses and number of areas occupied by invasive tree species (Kendall’s tau corrected for ties = 0.215; P < 0.001), but not for shrub species (P = 0.87). The mean number of uses is slightly, but significantly higher for trees (1.26) than for shrubs (1.08), Mann–Whitney U-test, P < 0.001. This may also be why the mean number of areas occupied by tree species is somewhat larger (2.35) than for shrub species (1.93), Mann–Whitney U-test; P < 0.001.
Careful consideration must be given to these factors when formulating management strategies, because selection criteria and cultivation practices can be modified potentially to reduce future problems with invasive woody plants (Hughes & Styles, 1987; Richardson et al., 2004a,b; Richardson & Blanchard, 2011).
Efficient propagule dispersal is essential for species to progress from naturalization to invasion (Murray & Phillips, 2010; Rejmánek, 2011). The finding that birds are the prevalent seed dispersal agent for both trees and shrubs (Table 3) is in agreement with earlier analyses (Binggeli, 1996) and is not surprising because birds are among the most efficient long-distance vectors of dispersal (Vittoz & Engler, 2007). Moreover, in the tropics, many bird-dispersed species are also bat-dispersed. The second most important mode of seed dispersal among invasive woody plants seems to be wind. However, percentages of invasive trees and shrubs falling into this category are relatively low (Table 3) compared with Binggeli’s (1996) summary. The likely reason for this discrepancy is that Binggeli included among wind-dispersed species, the so-called censer species (species that slowly release seeds from their fruits by shaking in the wind). This category is very often represented by many naturalized species (e.g. Table 7.3 in Specht & Specht, 1999). For example, dry-fruiting Ericaceae, Melastomataceae, Myrtaceae and Rosaceae belong to this category. If the censer mechanism is included under wind dispersal mode, the percentage of wind-dispersed trees and shrubs would be at least 24%. Besides censer, the ‘other’ modes of dispersal in Table 3 include dispersal by mammals (c. 10–20%), water (5–10%), ants (c. 5%) and ballistic (< 5%). At least 3% of the 622 species on our list, particularly species in the families Polygonaceae and Salicaceae, exhibit long-distance dispersal by water because of vegetative establishment of their parts or whole plants. The major conclusion is that bird-dispersed woody invaders always deserve, for many reasons (see Richardson et al., 2000a; Aslan & Rejmánek, 2010), special attention. Dispersal of alien woody species by vertebrates, mainly by birds and bats, is particularly important in the wet tropical forests (Table 8.1 in Rejmánek, 1996; Lobova et al., 2009).
There are no statistically significant differences in the mean numbers of areas occupied by species with different dispersal modes recognized in Table 3. The only significant difference is between bird-dispersed shrubs (1.77 areas on average) and bird-dispersed trees (2.43 areas) (Scheffe test; P < 0.05).
Key management issues
Management efforts are underway in many parts of the world to reduce problems associated with invasive alien trees and shrubs. These range from ad hoc local-scale efforts to control invasions and mitigate their effects, to national-scale, systematic strategies that integrate all potential options for reducing current problems and reducing the risk of future problems (van Wilgen et al., 2011; Wilson et al., 2011). Details of such operations are available in many publications. Rather than dissecting case studies, we focus on some overarching issues that complicate management. Background on invaded ecosystems, invasion processes, impacts and determinants of invasibility is provided in Appendix S3.
In devising sustainable strategies for managing problems arising from invasions of introduced trees and shrubs, managers and planners must confront several complex challenges. Most widespread tree and shrub invaders were intentionally introduced to the regions where they now cause problems, and most are still useful in parts of the regions where they occur (Kull et al., 2011). Conflicts of interests abound. Especially for forestry and agroforestry, replacing invasive alien species with native or less invasive non-native alternatives has limited potential. For commercial forestry, eucalypts and pines will remain the foundation of exotic forestry enterprises, and options must be sought to reduce invasiveness and to mitigate negative impacts of the key taxa. There is more scope for finding acceptable alternatives for invasive non-native ornamental species, but the nursery trade has substantial financial investments in many countries. The demand for popular ornamentals also has strong cultural ties, and the demand is thus difficult to change quickly. There are also other challenges for managing invasiveness in ornamental plants. In many taxa, different cultivars, hybrids or subspecific entities show very different levels of invasiveness, e.g. Buddleja davidii, Lantana spp. (an ‘aggregate species’) and Pyrus calleryana. Well-known invasive plants that have descended from domesticated plants include Psidium cattleianum, Pyrus calleryana and Coffea arabica. Genera where species identification is problematical resulting in barriers to effective management include Cecropia, Prosopis, Rubus and Ulmus. These factors all complicate the implementation of clear policies. In forestry, invasiveness may change substantially in hybrids and transgenics, with scope for both enhanced and reduced invasiveness. Biotechnology has the potential to reduce the invasiveness of useful trees by producing sterile trees. Although technologically feasible, important barriers exist. For example, the Forestry Stewardship Council prohibits the use of transgenic species (Richardson & Petit, 2005).
Interventions must consider that invasive alien trees and shrubs (other than the economically important taxa discussed above) may serve useful purposes in some situations. For instance, many alien trees and shrubs have strong value as nurse plants for the restoration of degraded natural forests (Lugo, 2004). Increasing land degradation in many parts of the world will increase the need to stabilization and rehabilitation efforts, including the controlled use of non-native species, even those with known or predicted invasive potential. ‘Weediness’ is often welcomed in such cases, and this is difficult to reconcile with biodiversity conservation concerns. Management strategies for invasive trees and shrubs must accommodate such issues. New multidimensional evaluation protocols (Richardson et al., 2009) are needed.
Another factor that must be taken into account is the rapidly changing global market for products from trees and shrubs, including new uses. For example, many alien trees and shrubs are being proposed for wide-scale planting for the production of biofuels, among them known invasive taxa like Azadirachta indica, Eucalyptus camaldulensis, Calotropis procera, Olea europaea, Leucaena leucocephala, Populus spp., Ricinus communis, Salix spp., Triadica sebifera and Zizyphus mauritiana and many species that are very likely to be invasive (Low & Booth, 2007; Gordon et al., 2011). Altered planting configurations, including massive increases in propagule pressure and the number of planting sites and thus foci for launching invasions, to accommodate biofuel production will surely launch many new invasions of more species over a greater area. Consideration must be given to potential invasions when deciding on strategies for biofuels and other emerging markets for wood-based products in different parts of the world (Richardson & Blanchard, 2011).
Climate change provides a huge challenge for managing woody plant invasions. Changing environmental conditions leads to rapid changes in the invasiveness of alien species (Richardson & Bond, 1991; Willis et al., 2010). Consequently, many alien species already present in an area and currently deemed ‘safe’ (non-invasive) may well become invasive. Recent modelling studies have revealed the extreme complexity involved in unravelling the many mechanisms whereby climate change could potentially influence invasion patterns and in using such information to design long-term management plans at the regional or national scale (Kleinbauer et al., 2010; Richardson et al., 2010).