Field data collection
Sampling sites were selected in typical viper habitats in Albania
(V. graeca : Kulmak and Tomorr mountains), Hungary (V.
ursinii : Peszéradacs meadows) and Ukraine (V. renardi : Davydivka
steppe) (Fig. 1). We searched for vipers by walking slowly in the
designated habitat patches in weather conditions suitable for viper
activity. When observing a viper, we recorded the coordinates of the
location with standard GPS precision, the time of observation, and the
age and sex of the individual. Spatial data on fine-scale location of
recorded vipers (hereinafter: points of known viper presence) were
collected for several days at each site (V. graeca : July 31 to
August 12, 2019; V. renardi : September 25 to October 2, 2019;V. ursinii : September 10 to 15, 2019, and April 19 to May 18,
2020). The number of days spent with field data collection was adjusted
to the density of snakes estimated in previous site visits. At the end
of each search period, we recorded vegetation structure at all the
points of known viper presence. To draw a distinction between the parts
of habitats actually used by vipers and the surrounding parts not used
by vipers, a minimum convex polygon (MCP) covering the points of known
viper presence with a 10 m buffer was drawn. Outside the MCP covering
the parts of habitats actually used by vipers randomly selected points
were placed (i.e., at least 10 m far from the closest point of known
viper presence (hereinafter: random points). The randomly selected
points were not in the analyses in case of being placed in areas
inaccessible to the surveyor or apparently unsuitable for snakes such as
rocky cliffs, water bodies or arable fields. The number of random points
was set at five times the number of presence locations.