The group formed by DC3 samples had its highest distribution probability
in several Mediterranean areas, central and southern Europe (including
England and Belgium to Crimea), north-western Africa and western Asia
(Turkey, Syria, Caucasus, Caspian shores). The most optimal distribution
for DC2 was in the eastern Mediterranean, specifically in the eastern
Aegean islands and the Mediterranean coastal zone of Turkey, Lebanon and
Israel. The distribution model of the latter group overlapped with the
distribution range inferred for D. orientalis , which also
presented its optimum in the eastern Mediterranean area, along the
coasts of Lebanon, Syria, Palestine and Israel. The highest
probabilities of potential distribution ranges for the DC1 group are
restricted to the humid areas of the western Canary Isles (Figure 3),
with Madeira and western Morocco showing a lower probability of
occurrence. Overlap was observed between DC2, DC3 and D.
orientalis in the eastern Mediterranean region, and between DC1 and DC3
in the Canary Islands.
The highest niche breadth obtained corresponded to the clade with the
largest modelled distribution projection (i.e., 0.875 for DC3 clade),
followed by DC2 (0.731), D. orientalis (0.516) and DC1 (0.499).
We calculated Schoener’s D and Hellinger’s I indexes as
metrics of niche overlap between pairs of distribution models. The
highest overlaps between current niches were found between the DC3 clade
with DC1 (D = 0.52, I = 0.64) and DC2 (D = 0.48,I = 0.67), whereas D. orientalis showed lower overlap
values with DC2 (D = 0.32, I = 0.60) and DC3 (D =
0.36, I = 0.48). Although niche overlap between DC2 and D.
orientalis clades showed the lowest values, a PCA using the raw
bioclimatic data obtained from the studied samples did not differentiate
between them (Figure 4); a better separation was, however, found between
DC3 and DC1. These results are supported by observed differences between
the four clades for each of the bioclimatic variables (Supplementary
Data Figure S5).