Skógræktarritið - 15.05.2001, Qupperneq 135

and Habitat Action Plans (part of
the UK Biodiversity Action Plan)
include costed and phased
expansion targets for different
native woodland types. In addi-
tion, there is an increasing
awareness that more benefits
could be accrued from an expan-
sion in native woodlands if a
long-term strategic approach
were adopted. This forms the
basis of the philosophy of Forest
Habitat Networks (FHN).
Research undertaken by Scottish
Natural Heritage (Peterken et al
1995) has demonstrated that the
spatial distribution of new wood-
land can have a major impact on
its value for biodiversity and con-
servation.
Scottish Natural Heritage
(SNH) has identified the re-estab-
lishment and restoration of native
woodland as one of the most
important steps towards a sus-
tainable future for the natural her-
itage. This paper describes the
development and application of a
modeling tool which is providing
SNH, and to a limited extent
other groups, with assistance
towards achieving this objective.
METHODS
Very little is known about the
potential distribution and extent
of different woodland types to
guide native woodland expan-
sion at regional and local levels.
Most surviving native woodlands
tend to be highly fragmented and
are often radically altered. Com-
parisons with historical recon-
structions from palaeobotanical
studies (Bennett, 1996) are likely
to be of limited relevance since
site conditions have been modi-
fied by factors such as climate
change and environmentai pollu-
tion, the removal of the original
forest cover, and agricultural cul-
tivation of soils.
A more realistic approach is to
predict woodland distribution for
current environmental condi-
tions using site suitability mod-
els. At MLURI a Native Woodland
Model (NWM) is being devel-
oped, with support from Scottish
Natural Heritage, which links
expert knowledge on woodland
and scrub habitat requirements
with digital biophysical data to
predict the occurrence and distri-
bution of a range of woodland
communities. The woodland
mapped broadly corresponds to
'present-natural', i.e. the native
species inherited from primeval
conditions, taking into account
site and climatic changes to the
present day.
Data
Two digital data sources are used
in the model: the 1: 250 000 scale
National soils map (MISR 1984)
and the 1: 25 000 scale Land
Cover of Scotland 1988 (LCS88)
dataset (MLURI 1993). Both data
sets contain a range of informa-
tion relevant to the prediction of
NVC woodland communities.
The 1: 250 000 scale national soil
map comprises 580 soil map units,
differentiated on geological (soil
assocation), pedological (compo-
nent soils) and physiographic cri-
teria (landforms). Each soil map
unit also has a number of vegeta-
tion communities ascribed to it,
but this is not a criterion used to
distinguish one soil map unit
from another. The majority of the
580 soil map units are soil com-
plexes particularly over much of
the central, western and northern
Highlands. i.e. they contain two
or more soil types.
Tfie 1:25 000 Land Cover map pro-
vides information on land cover
existing in 1988. It is the first
ever national census of land
cover in Scotland and was cap-
tured from the visual interpreta-
tion of aerial photographs. The
hierarchical classification allows
for 126 single land cover features
including all the major semi-nat-
ural vegetation communities.
There are over 1000 mosaic cate-
gories used largely to describe
the heterogeneous semi-natural
vegetation resource. in total, veg-
etation mosaics rather than sin-
gle categories cover approxi-
mately 30% of Scotland. It pro-
vides much more robust and
detailed land cover information
than the soil map.
The two datasets were overlaid
within a Geographic information
System forming a new integrated
dataset that contains several
thousand soil/land cover combi-
nations. These combinations,
which are essentially a descrip-
tion of the present site condi-
tions, form the basis of the
Native Woodland Model (NWM)’s
predictions. Each combination is
allocated to an NVC woodland
type or to a mosaic of NVC types
(see below), based on the rela-
tionships between biophysical
properties and woodland require-
ments. It is important to note
that the NWM predicts the poten-
tial for woodland and scrub types
under current soil and vegetation
conditions, i.e. with no or mini-
mal intervention.
Modelling
The National Megelalion Classification
(NVC) has been used as the basis
for most of the woodland cate-
gories described, predicted and
mapped in this project. Nineteen
major UK woodland types and six
scrub communities, each with a
distinctive mix of trees, shrubs,
field and ground flora, are
described in the NVC (Rodwell
1991). The range of woodlands
within the NVC are associated
with different habitats which are
described in terms of climatic
zone, soil types, terrain and
topographic position. These rela-
tionships are described in vary-
ing degree of complexity and
SKÓGRÆKTARRITIÐ 2001 l.tbl.
133