Skógræktarritið - 15.05.2001, Blaðsíða 103

importance of different factors in
determining herbivore range
use.
Forage preferences
Once a predictive understanding
of when the animals are likely to
actually visit a particular area is
gained, it is at this second level
that an understanding of diet
choice is fundamental. This con-
cerns not only the relative prefer-
ences for different tree species by
different herbivores, but also the
range of other factors which
affect these preferences, such as
surrounding vegetation and time
of year. Preferences for different
tree species have been relatively
widely studied in a range of
countries (e.g. Mitchell, Staines
& Welch 1977; Van Hees., Kuiters
& Slim 1996), resulting in a rea-
sonable consensus as to which
tree species are most preferred
by large herbivores - examples of
more preferred species are Salix
spp, Sorbus aucuparia, Populus trem-
ula, and examples of less pre-
ferred species are Pinus sylvestris,
luniperus communis, Picea spp. This
is an important issue for a land
manager who wishes to encour-
age regeneration of particular
tree species. Forexample, Pinus
sylvestris is likely to regenerate at
higher deer densities than Salix
spp, and there are good examples
of this pattern in some of the
Scottish estates which have been
reducing deer numbers to allow
forest regeneration (Beaumont et
al. 1994). It is also well estab-
lished that, within a species,
some individual saplings are
more preferred than others, for a
range of reasons such as mor-
phology, visibility and chemical
content (Mitchell etal. 1977;
Danell etal. 1991; Edenius et al.
1995). However, both the above
factors are strongly influenced by
surrounding vegetation; itaffects
food choice, sapling visibility and
accessibility. Finally, time of year
also interacts with all the above;
many factors change through the
year to alter herbivore forage
choice, such as biomass and
digestibility of other forage, and
visibility of the saplings them-
selves.
To summarise, a range of fac-
tors interact even at the level of
forage preferences to determine
choice of diet. The crucial issue
for designing robust manage-
ment options is to understand
and be able to predict how, when
and why these factors might
interact.
Sapling responses to damage
At the smallest scale, factors
affecting sapling responses to
damage are very similar: different
species respond differently to
damage, and the response also
depends on the age and size of
the sapling; younger, smaller
saplings are generally more
badly affected by browsing dam-
age (Chapin et al. 1995). To give
an example, Pinus sylvestris is gen-
erally one of the less preferred
species of browse, but it is also
one of the least tolerant of dam-
age (Miiler et al. 1982; Millard &
Hester in press). Even if it is less
heavily browsed than a species
such as Sorbus aucuparia, which is
very tolerant of damage (Miller et
ai. 1982; Millard & Hester in
press), it may still be more badly
affected. Therefore browsing
preferences for different tree
species need to be considered
along with sapling tolerance.
Timing and severity of damage
are also fundamentally important
and their effects also vary
between different tree species
(Danelletal. 1994, Hester et al.
1996). Furthermore, as with for-
age preferences, surrounding
vegetation, as well as soils, affect
competition and resource avail-
ability, which may all influence a
sapling’s ability to regrow after
damage (Hester et al. 1996).
Therefore, as with foraging
preferences, a complex of factors
all interact at the sapling level,
and all need to be understood to
some degree to be able to make
reliable predictions about brows-
ing impacts on saplings under
different conditions.
Integration
Most research to date has
focused on particular questions
or small groups of questions,
usually within only one of the
three main approaches described
above. Such types of studies are
continuing in many different
countries, often in isolation
rather than as part of a larger
integrated programme. Thus, 1
suggest that the most pressing
need at present is to move
towards much greater integration
of such work at a whole range of
scales; from the landscape,
where a large number of interact-
ing factors require to be studied,
down to the individual plant
where highly controlled experi-
ments can be carried out.
Research at the landscape
level (e.g. Fig 1) has the advan-
tage of describing 'reality'; but
the disadvantage is the huge
variability and the difficulty of
isolating which factors or combi-
nations of factors have caused
the results measured (e.g.
Beaumont et al. 1994). Using
only this approach can lead to
misleading conclusions. It can be
combined with research at medi-
um geographical scales (ha -
km), using semi-controlled
experiments, for example, where
large replicate experimental
plots are set up within a natural-
ly variable landscape, with a
range of factors varied explicitly
(such as herbivore density, tim-
ing of grazing, different sapling
species) to examine their influ-
SKÓGRÆKTARRITIÐ 2001 l.tbl.
101