Table 1.
The number of PAs classified as occupied, at risk, in situ refugia and ex situ refugia for each of the 12 focal species.
Fig 1.
Maps illustrating the number of species finding (A) in situ and (B) ex situ refugia in each NNR or SSSI across England, Scotland and Wales.
Shading intensity corresponds to the number of species within each site finding refugia within the PA.
Table 2.
The results of a Bayesian generalised linear mixed model to examine parameters associated with the number of species with ex situ refugia.
Table 3.
The results of a Bayesian generalised linear mixed model to examine parameters associated with the number of species with in situ refugia.
Fig 2.
Effect plots illustrating the relationship between PA covariates and the estimated number of species (sum of occupancy probabilities) finding refugia within the PA.
The covariates included are mean elevation (A and B), elevational range (C and D) and area (E and F). Figures on the left show the relationships with the number of species with in situ refugia, while those on the right depict species with ex situ refugia in each PA. Lines of best fit derived from a Bayesian generalised linear mixed model are represented with 95% credible intervals (green for in situ refugia and blue for ex situ refugia). All graphs demonstrate a significant positive correlation.
Fig 3.
Forest plot from the Bayesian generalised linear mixed model examining how variables affect the number of species with (A) in situ refugia and (B) ex situ refugia.
For each variable, the plot shows the mean effect, with error bars showing 95% credible intervals. Variables are indicated to have a significant positive effect if their credible intervals are above zero, a significant negative effect if their credible interval is below zero and no effect if the credible interval overlaps zero.