Skógræktarritið - 15.05.2001, Side 104

ence on herbivore behaviour and
sapling damage (e.g. Hester et al.
1996). Such experiments tend to
be labour-intensive and expen-
sive to set up, with fencing costs
and animal management, but
they can yield informative, field-
based results. Slightly more con-
trolled may be an experiment
limited to a single vegetation
type, with small-area (ha or less)
control of precise timing and
density of herbivore occupancy
(days and weeks). These experi-
ments are also very labour inten-
sive, but can enable isolation of
specific factors which are gener-
ally confounded in more variable
field situations. At the smallest
scale (e.g. greenhouse, growth
chamber, animal enclosure),
completely controlled experi-
ments can be designed where
most factors are tightly con-
trolled (e.g. Millard S- Hester in
press; Fig 2). The advantage of
this type of experiment is that
one can isolate specific factors
and examine them in great
detail, giving important insights
which cannot normally be gained
from less controlled experi-
ments. However, the disadvan-
tage is that such experiments can
be so far removed from reality
that the behaviour of herbivores
and/or saplings is not always rep-
resentative of the field situation.
There is an important place for
these experiments, but they need
to be designed with care and
closely linked to relevant field
experiments. To summarise,
there is an important role for
integrating all these approaches,
as they can complement each
other if well designed.
Complexity to simplicity -
returning to management needs
in this section I will consider how
to translate the above complexity
into simplicity, so as to address
the needs of land managers
Figure 3. Diagrammatic representa-
tion of the Decision Support tool
HILLDEER (MLURI 1998).
and/or policy makers, as outlined
earlier in this paper. it should be
clear how the above research
approaches can contribute to
developing the required under-
standing about herbivore impacts
on forest regeneration, but the
needs of managers/policy makers
are to have simple and easy to
use tools to aid decision-making
on appropriate herbivore num-
bers to achieve a range of desired
end-points. Therefore the
research results will only be of
direct value if they can be trans-
lated into something that people
will actually be able to apply and
benefit from. One approach is the
development and use of Decision
Support Tools (DSTs) which can
be run on personal computers,
with a small information input
that is either: (a) easy for the user
to collect, or: (b) already in the
DST (e.g. countrywide climatic
data, Iand cover data).
The main requirements for a
DST to be successful are: (I) they
need to be easy and inexpensive
to use; (2) they need to use
robust data, and include infor-
mation on the degree of uncer-
tainty associated with the out-
puts - this is a crucial issue, as
computer output can create the
illusion of being unquestionable;
(3) the data which the user needs
to input should also be easy to
collect; and (4) finally, the output
needs to be appropriate, practi-
cal and straightforward.
One example of the underlying
framework of a grazing DST is
shown in Figure 3 (MLURI 1998).
It was developed by a team at the
Macaulay Land Use Research
institute to predict habitat use
and impacts of different densi-
ties of red deer within large
upland areas of Scotland (open
range vegetation only). As indi-
cated in the diagram, it incorpo-
rates a range of vegetation types,
disturbance, other herbivores
(sheep and rabbits) and climate,
thus achieving the aim of inte-
grating a range of information at
different scales to produce sim-
ple outputs. The input require-
ments are straightforward: for
example the approximate areas
of different vegetation types
(available from a whole Scotland
dataset held at the institute) and
the numbers of other herbivores.
The user can manipulate deer
numbers to investigate how
habitat use might change, and
can predict the impacts of differ-
ent deer densities on the range
of vegetation types present. It
also links with a population
dynamics model so that predic-
tions on how different culis will
affect reproductive rates and sex
ratios can be obtained, both in
the short and longer term.
DSTs can be very useful if well
designed, and this is one of the
best ways to integrate the kind of
complex information described
above and to make it actually
work for practical land manage-
102
SKÓGRÆKTARRiTIÐ 2001 l.tbl.