Keywords
climate matching, heterogeneous landscape, invasive species, lag phase,P. muralis , range expansion
Introduction
The global rise in the number of species introduced to regions beyond
their native range via human-mediated translocation shows no sign of
reaching saturation point (Seebens et al., 2017). While many species
fail to establish or have little negative effect following introduction,
a subset of these do spread and can have significant impact on
economies, human health, native biodiversity and ecosystem services
(Kolar & Lodge, 2001; Vila et al., 2010; Keller, Geist, Jeschke, &
Kühn, 2011). The severity of potential negative impacts (e.g.,
extirpation and extinction of native species) are such that invasive
non-native species (INNS) are justifiably regarded as one of the most
significant threats to biodiversity worldwide (Genovesi, 2009;
Simberloff et al., 2013).
For non-native species to become widespread and potentially damaging
following introduction to new regions, introduced populations must
negotiate the three stages of an introduction–establishment–invasion
continuum (Blackburn et al., 2011). Evaluation of the likelihood of a
species to be transported, establish, and to spread, as well as the
potential for having ecological, economical, and health impacts, forms
the basis of ‘invasive’ risk assessment for alien species (Bacher et
al., 2018; Roy et al., 2019). Although it has been argued that the term
‘invasive’ doesn’t always necessarily equate with a species’ negative
impact (Ricciardi & Cohen, 2007), the potential for damaging effects
inherently increases as introduced species increase in population size
and spread across novel landscapes, thus affecting broader areas and
more ecological communities (Crooks, 2005). As such, there is great
interest in understanding patterns and rates of expansion of introduced
species, and the environmental factors which limit their distributions
(Gallien, Munkemuller, Albert, Boulangeat, & Thuiller, 2010; Roy et
al., 2019).
Following introduction and successful establishment beyond native
ranges, species can further expand their range through local dispersal
processes and/or by jump dispersal events that may be human-mediated
(i.e., deliberate or accidental movement of individuals between
habitats) or natural dispersal (i.e., long distance flight in birds)
(Suarez, Holway, & Case, 2001; Ingenloff et al., 2017). Invading
species typically exhibit several phases in the rate of spread. Firstly,
there is an initial establishment phase where rate of spread is slow.
Secondly, an expansion phase typified by increasing rates of spread, and
finally, a saturation phase when available space is occupied and
expansion rates reach a plateau (Arim, Abades, Neill, Lima, & Marquet,
2006).
A suite of factors influence patterns and rates of range expansion
during these phases: propagule size, dispersal mode, matching of
physiological and ecological traits of invading species with
environmental conditions at the receptor site, vital rates (births and
deaths) species interactions, evolutionary processes, spatial
heterogeneity and temporal variability (reviewed in Hastings et al.,
2005). Furthermore, our ability to assess and predict the temporal
dynamics of invasions is often complicated by the phenomenon of lag
phases, wherein an introduced species remains at low population levels
in the early stages of establishment for a protracted period of time
before the sudden onset of rapid range expansion (see Crooks (2005) for
review of the causes of temporal lags at all stages in the invasion
process). Introduced populations of the northern Racoon (Procyon
lotor ), for example, remained small for a number of years following
introduction to Europe before a population explosion in the mid 1990’s
(Salgado, 2018). Similarly, landscape complexity can result in temporal
and spatial patterns of invasion dynamics that deviate from classic
theory of symmetrical, radial expansion from a central point (diffusion
theory) (Skellam, 1951; Shigesada, Kawasaki, & Takeda, 1995; Kinezaki,
Kawasaki, & Shigesada, 2010). The effects of landscape heterogeneity on
patterns and rates of expansion have been demonstrated in the quick
colonization of areas of suitable habitat in the early stages of the
American mink (Neovison vison ) invasion, compared to uptake in
areas of low habitat suitability in Scotland (Fraser et al., 2015), and
the fluctuating rates in range expansion of Cane toad (Rhinella
marina ) in response to changing environmental conditions in newly
invaded areas of Australia (Urban, Phillips, Skelly, & Shine, 2008).
Consideration of dispersal processes across heterogeneous landscapes is
therefore central to predicting potential for range expansion during the
invasion process (Travis, Harris, Park, & Bullock, 2011; Bocedi,
Zurell, Reineking, & Travis, 2014; Grayson & Johnson, 2018). The
development of platforms for spatially explicit individual-based
modelling (Bocedi, Zurell, et al., 2014; Samson et al., 2017) have
enabled the nested interactions between dispersal, landscape properties,
and population dynamics to be considered in predicting species
distributions, increasing the ecological realism of range expansion
models (Andrew & Ustin, 2010; Ferrari, Preisser, & Fitzpatrick, 2014;
Mang, Essl, Moser, Kleinbauer, & Dullinger, 2018; Hunter-Ayad &
Hassall, 2020).
In this study, we determine the potential for range expansion of the
non-native common wall lizard (Podarcis muralis ) in the UK.Podarcis muralis has a long history of introductions beyond its
native range which covers most of Western and Southern Europe (Gassert
et al., 2013). Many of these introductions have extended its range
throughout continental Europe (Schulte, Gassert, Geniez, Veith, &
Hochkirch, 2012; Wirga & Majtyka, 2015; Šandera, 2017), but the species
also has several populations established in the New World, both in the
United States (R. M. Brown, Gist, & Taylor, 1995) and Canada (Allan,
Prelypchan, & Gregory, 2006). Introduced to Vancouver Island, British
Columbia, in 1970, the species persisted in isolated populations until
2006, but has since spread with alarming speed due to jump dispersal
(human mediated) and natural radial dispersal of 40-70 meters a year in
urban areas (Engelstoft, Robinson, Fraser, & Hanke, 2020){Engelstoft,
2020 #2101}.
To date there is no empirical evidence of negative ecological impacts ofP. muralis introductions in the UK, and there is mixed social
perception and opinion towards the species’ presence (Williams, Dunn,
Quinn, & Hassall, 2019). However, suspected declines in native lizards
through interference and/or exploitation contest have been reported
following introductions of P. muralis , to both Germany (Münch,
2001; Kühnis & Schmocker, 2008; Schulte, 2009) and the UK (Mole, 2010).
There have been multiple introduction events of P. muralis to the
UK both as deliberate releases of captive animals and as cargo
stowaways, with some extant populations having been established on the
UK mainland as early as the 1970s (Michaelides, While, Bell, & Uller,
2013). More recent introductions (1980s onwards) have mostly arisen from
movement of individuals from already established populations (secondary
introduction) or captive-bred animals, rather than directly sourced from
the native range (Michaelides, While, Zajac, & Uller, 2015). The UK
populations represent the species at the northern extent of its range,
with sites having markedly different climatic conditions compared with
the native range. For example, air temperatures during the main activity
season in populations in England are 5–10°C lower than their source
regions in Tuscany and western France (While et al., 2015).
We investigated the potential for range expansion of P. muralisin the UK with models highlighting different (but complementary)
parameters likely to influence spread at two spatial scales. Firstly,
since long distance jump dispersal via translocation is important in
facilitating spread of this species, we aimed to predict the national
extent of the area potentially available for further colonisation by
running a species distribution model (SDM) based on climatic suitability
at these northern extremes. As has been speculated elsewhere, the
ability to survive cold winters is likely limiting to the spread of
introduced Podarcis populations (Burke, Hussain, Storey, &
Storey, 2002). We therefore hypothesised that latitudinal clines in
climate would restrict the area available for northward expansion ofP. muralis via long distance human-assisted translocation in the
UK. Second, to make predictions of population growth and dispersal
patterns, as well as identify environmental features important to range
expansion at a local level, we took a hybrid model approach combining
SDMs, informed by variables characterising 10 local landscapes (i.e.,
microclimate, proximity to geographic features, and habitat type), with
a high resolution (15 x15 m) spatiotemporal individual based model (IBM)
simulating local population and dispersal dynamics. We expected that
landscape characteristics (i.e., configuration and connectivity of
suitable habitat patches), would result in asymmetrical patterns of
predicted dispersal within populations, which in turn, would result in
spatial and temporal variance in patterns of population growth and range
expansion between populations. These analyses allow us to investigate
the proximate and ultimate barriers to spread, as well as simulating the
potential for invasion lag in each population.