Page 26 - NZDPM 33/2017

Nesting habitats of Black Stork (Ciconia nigra L.) in Ukrainian forest zone … 25

smaller buffers because in some parts of the study area the nests are located very close to
each other (sometimes less than in 500 m) and we had no possibility to check inhabitance
of every nest. Of course, many of those nests are used by the same pairs, but we cannot
distinctly separate them. Moreover, in Ukraine 500 m buffers are proposed as a zone of
strict protection of the species, so it seemed reasonable for us to use such a distance for the
analysis.

Fig. 2. Example of a forest structure analysis in 500 m buffers around nests (in this
particular case – the comparison of areas of different forest types; the areas were calculated
in GIS and stored in the attributive table).

In total, in the analysis we have used location information from 100-108 nests and
attributive information about the forest age, type, productivity, main tree species and
vegetation-humidity types from 148000 forest plots (so-called "vydil" – the smallest forest
unit in Ukraine). A large number of forest fragments ("vydil") considered and their
comparatively small size allows to estimate forest structure in an appropriate and correct
way. On the one hand, a small size of the plots allows to describe local habitats in a detailed
way, without major generalization. On the other hand, a big number of such plots provides
an excellent sample size. This is crucial to produce robust and precise conclusions about
Black Stork habitat preferences on a macro level.
In addition, it should be born in mind that it is easy to make false conclusions if a
general forest structure is not considered while analyzing habitat preferences of the species,
because distribution of nests in the forests of different types can be predicted by a
frequency of occurrence of those forest types. Therefore, when judging about Black Stork
habitat preferences we tried to estimate habitat selectivity or avoidance by comparing the
frequency of distribution of nests in the forests of different types and the forest types
proportional composition (compared by area) in 500 m buffers around nests with the
general forest structure (percentage by area) in the region. We have also generated 108
random points in the study area (using ArcGis 10.5 "Create Random Points" tool) and