patterns in the data and include only those traits that contribute to invasion success,
as opposed to all traits considered important by experts. For example, for alien cacti
it was found that native range size is a predictor of both impact and invasiveness in
South Africa (Novoa et al. 2016a). The statistical approach can take into account
interactions between traits and lead to much simpler tools than trait-scoring, but the
underlying methods and mechanisms are not always easily understood and therefore
less likely to be supported by managers and policy-makers (Keller and Kumschick
2017). Furthermore, due to the models including a very limited number of variables,
no conclusion can be reached if data for any of the few selected variables is lacking.
Decision Trees These can be a subset of the statistical approach, but also be based on
traits considered as important by experts. Similar issues therefore arise in this approach,
which comprises of several questions of which each answer leads either to another
question or a decision regarding a species’ risk. If the answer to any question is not
known, no decision can be reached, or tenuous assumptions are made. For that reason,
this approach has not been used extensively. However, Tucker and Richardson (1995)
suggested a decision tree approach to assess risks of tree invasions to the Fynbos
Biome.
Rapid Screening The rapid screening approach can use elements of the previous
three approaches, with the focus on a quick assessment of often a large number of
species where little information is available. Rapid screening can be used as a standalone risk assessment, to prioritise species for more detailed risk analysis or to create
watch lists. Faulkner et al. (2014) applied a rapid screening tool to almost 400 species
to create a watch list of alien species for South Africa. Their tool included only three
aspects: (1) the invasiveness of the species elsewhere, (2) climate match, and
(3) tourism and trade data as a proxy for propagule pressure.
Mechanistic Approach The mechanistic approach is not based on traits per se, but it
follows a species through the invasion process and assesses the likelihood that it will
cross certain barriers (e.g. those proposed by based on Blackburn et al. 2011), and
whether it has the potential to have an impact. For example, if a species is unlikely to
find suitable climatic conditions and habitats in a new region, it would be unlikely to
establish and subsequently become invasive, which can reduce its risk despite
potential impacts. Such an approach is implemented in Belgium (D’hondt et al.
2015).
Detailed Approach The detailed approach, as the name suggests, requires a substantial amount of data on the alien species’ ecology, biology and behaviour, as well as the
recipient environment and the interaction between all these factors. It often requires
additional research to fill knowledge gaps, which may include interviews with experts.
Since it also includes management considerations and stakeholder perceptions, it
resembles more closely a risk analysis approach. It has been used in Canada to assess
risks of the highly contentious Asian carp species (Mandrak and Cudmore 2004).
Which approach is most useful depends on the circumstances and purpose. For
example, rapid screening is quick and easy. However, results of such assessments
provide limited knowledge on the mechanisms of invasion and expected behaviours
20 Analysing the Risks Posed by Biological Invasions to South Africa
577
as opposed to all traits considered important by experts. For example, for alien cacti
it was found that native range size is a predictor of both impact and invasiveness in
South Africa (Novoa et al. 2016a). The statistical approach can take into account
interactions between traits and lead to much simpler tools than trait-scoring, but the
underlying methods and mechanisms are not always easily understood and therefore
less likely to be supported by managers and policy-makers (Keller and Kumschick
2017). Furthermore, due to the models including a very limited number of variables,
no conclusion can be reached if data for any of the few selected variables is lacking.
Decision Trees These can be a subset of the statistical approach, but also be based on
traits considered as important by experts. Similar issues therefore arise in this approach,
which comprises of several questions of which each answer leads either to another
question or a decision regarding a species’ risk. If the answer to any question is not
known, no decision can be reached, or tenuous assumptions are made. For that reason,
this approach has not been used extensively. However, Tucker and Richardson (1995)
suggested a decision tree approach to assess risks of tree invasions to the Fynbos
Biome.
Rapid Screening The rapid screening approach can use elements of the previous
three approaches, with the focus on a quick assessment of often a large number of
species where little information is available. Rapid screening can be used as a standalone risk assessment, to prioritise species for more detailed risk analysis or to create
watch lists. Faulkner et al. (2014) applied a rapid screening tool to almost 400 species
to create a watch list of alien species for South Africa. Their tool included only three
aspects: (1) the invasiveness of the species elsewhere, (2) climate match, and
(3) tourism and trade data as a proxy for propagule pressure.
Mechanistic Approach The mechanistic approach is not based on traits per se, but it
follows a species through the invasion process and assesses the likelihood that it will
cross certain barriers (e.g. those proposed by based on Blackburn et al. 2011), and
whether it has the potential to have an impact. For example, if a species is unlikely to
find suitable climatic conditions and habitats in a new region, it would be unlikely to
establish and subsequently become invasive, which can reduce its risk despite
potential impacts. Such an approach is implemented in Belgium (D’hondt et al.
2015).
Detailed Approach The detailed approach, as the name suggests, requires a substantial amount of data on the alien species’ ecology, biology and behaviour, as well as the
recipient environment and the interaction between all these factors. It often requires
additional research to fill knowledge gaps, which may include interviews with experts.
Since it also includes management considerations and stakeholder perceptions, it
resembles more closely a risk analysis approach. It has been used in Canada to assess
risks of the highly contentious Asian carp species (Mandrak and Cudmore 2004).
Which approach is most useful depends on the circumstances and purpose. For
example, rapid screening is quick and easy. However, results of such assessments
provide limited knowledge on the mechanisms of invasion and expected behaviours
20 Analysing the Risks Posed by Biological Invasions to South Africa
577