Modelling climate suitability across the UK
A total of 1331 lizard sightings (presence) data (76 from online portal,
52 from postcard returns, 1203 from visual surveys) were used to develop
relative habitat suitability maps at the UK national extent using MaxEnt
v3.3.3k software (Phillips, Anderson, & Schapire, 2006). Podarcis
muralis has demonstrated rapid adaptive responses following
introduction to cool climates, with ability to evolve broader thermal
tolerance at dispersal fronts (Litmer & Murray, 2019), and prolonged
embryo retention and faster embryonic growth at low temperatures -
compared to ancestral states (While et al., 2015). For this reason, we
only used P. muralis presence records from the introduced UK
range in our MaxEnt model, as they best reflect the current climatic
tolerance of the species, and, secondary introductions from established
populations are the primary concern for the species extending its range
in the UK. We focused on seasonal averages of six climatic variables
(predictors) at 5km resolution (Met Office, 2018) that have most
relevance to wall lizard biology and therefore likely to influence
distribution (Wirga & Majtyka, 2015) (Table 1). These six variables
were refined from an initial input of 13 climate variables through an
iterative process of removal/retention to limit covariate correlation
(Spearman’s rank correlation; correlated if rs ≥ 0.6)
and maximise model performance (Glover-Kapfer, 2015). We kept parameter
settings in MaxEnt the same for modelling at the national and local
levels (see Appendix 1 for detailed method). The following Maxent
parameters were used: Features used = Linear; Output format = logistic;
Random Test % = 20; Regularization Multiplier = 2; Max number of
background points = 10000; Replicates = 5; Replicated run type =
Crossvalidate.