Land-use impact on plant communities in semi-natural sub-alpine grasslands of Budalen, Central Norway


Austrheim, G.; Gunilla, E.; Olsson, A.; Grontvedt, E.

Biological Conservation 87(3): 369-379

1999


Sub-alpine, semi-natural grasslands induced by mountain summer farming in Budalen, central Norway, can be divided into two main habitats: small enclosures at each summer farm site, and the pastures surrounding them. Enclosures are exposed to different land-use regimes including grazing, mowing, fertilisation and soil disturbance. Pastures have never been fertilised or ploughed and the current grazing pressure is moderate. All investigated pastures are former haymaking areas. Higher conservation values of pastures compared to enclosures is reflected both in patterns of species richness and the occurrence of vulnerable species. Higher species richness in pastures is related to lower nutrient levels, lower loss of ignition in the soil, and higher levels of pH. Vulnerable species are concentrated in species-rich pastures, and have low local abundance and regional distribution. Enclosures reflect a complexity in land-use, were the effects of mowing, fertilising and ploughing on plant community patterns could not be separated. Tree and shrub species are common in both habitats, and indicate a successional invasion of woody species. Maintenance of landuses that have created semi-natural grasslands in the long term perspective (grazing, cutting of trees and shrubs, and mowing) is necessary to prevent forest succession, and a prerequisite for future conservation of sub-alpine, semi-natural grasslands.

BIOLOGICAL
CONSERVATION
ELSEVIER
Biological
Conservation
87
(1999)
369-379
Land
-use
impact
on
plant
communities
in
semi
-natural
sub
-alpine
grasslands
of
Budalen,
central
Norway
Gunnar
Austrheim*,
E.
Gunilla,
A.
Olsson,
Eli
Grontvedt
Department
of
Botany,
Norwegian
University
of
Science
and
Technology,
NTNU
N-7034
Trondheim,
Norway
Received
28
June
1997;
received
in
revised
form
and
accepted
26
May
1998
Abstract
Sub
-alpine,
semi
-natural
grasslands
induced
by
mountain
summer
farming
in
Budalen,
central
Norway,
can
be
divided
into
two
main
habitats:
small
enclosures
at
each
summer
farm
site,
and
the
pastures
surrounding
them.
Enclosures
are
exposed
to
different
land
-use
regimes
including
grazing,
mowing,
fertilisation
and
soil
disturbance.
Pastures
have
never
been
fertilised
or
ploughed
and
the
current
grazing
pressure
is
moderate.
All
investigated
pastures
are
former
haymaking
areas.
Higher
conservation
values
of
pastures
compared
to
enclosures
is
reflected
both
in
patterns
of
species
richness
and
the
occurrence
of
vulnerable
species.
Higher
species
richness
in
pastures
is
related
to
lower
nutrient
levels,
lower
loss
of
ignition
in
the
soil,
and
higher
levels
of
pH.
Vulnerable
species
are
concentrated
in
species
-rich
pastures,
and
have
low
local
abundance
and
regional
distribution.
Enclosures
reflect
a
complexity
in
land
-use,
were
the
effects
of
mowing,
fertilising
and
ploughing
on
plant
community
patterns
could
not
be
separated.
Tree
and
shrub
species
are
common
in
both
habitats,
and
indicate
a
successional
invasion
of
woody
species.
Maintenance
of
land
-
uses
that
have
created
semi
-natural
grasslands
in
the
long
term
perspective
(grazing,
cutting
of
trees
and
shrubs,
and
mowing)
is
necessary
to
prevent
forest
succession,
and
a
prerequisite
for
future
conservation
of
sub
-alpine,
semi
-natural
grasslands.
1998
Elsevier
Science
Ltd.
All
rights
reserved.
Keywords:
Semi
-natural
sub
-alpine
grasslands;
Plants;
Biodiversity;
Land
-use;
Species
richness;
Vulnerable
species
1.
Introduction
Agricultural
systems
utilising
environmental
resour-
ces
in
mountain
regions
with
limited
production
capa-
cities
have
prehistoric
traditions
in
Norway
(Reinton,
1955;
Kvamme,
1988;
Moe
et
al.,
1988).
Inhabitants
of
permanent
farms
in
the
valleys
with
limited
land
areas
for
fodder
production
in
the
vicinity,
moved
their
live-
stock
to
higher
altitudes
in
the
surrounding
mountains
during
the
summer
season.
Summer
farms
for
dairy
production
were
established
for
seasonal
utilisation
of
environmental
resources
in
the
mountains.
Maintenance
of
the
lowland
agroecosystems
was
thus
dependent
on
supplementary
resources
provided
by
summer
-farming
in
the
mountains
(Reinton,
1955).
This
mixed
agri-
cultural
system,
characterised
by
a
dual
dependence
on
stock
-raising
and
cultivation,
and
the
use
of
seasonal
habitats
in
mountains,
has
been
important
all
over
the
*
Corresponding
author.
E-mail:
gunnar.austrheim@chembio.
ntnu.no
world
apart
from
the
Far
East
and
North
America
(Price,
1981).
Most
grasslands
in
Norway
are
of
semi
-natural
origin
(Fremstad
and
Elven,
1991).
Traditional
agricultural
practices
such
as
grazing,
mowing,
fuel
-wood
cutting
etc.,
have
produced
grass-
and
herb
-dominated
com-
munities
in
the
lowlands
and
also
in
sub
-alpine
areas
where
mountain
summer
-farming
has
been
practised.
During
the
20th
century
the
traditional
summer
-
farming
management
has
gradually
lost
its
economic
importance.
Two
major
trends
in
land
-use
changes
are
evident
in
sub
-alpine,
semi
-natural
grasslands
(Olsson
et
al.,
1995):
(1)
marginal
hay
meadows
and
grazing
areas
are
left
to
forest
succession,
(2)
former
semi
-natural
grasslands
are
ploughed
and
fertilised.
Similar
processes
are
taking
place
throughout
western
Europe
where
land
-use
changes
have
caused
a
severe
reduction
of
semi
-natural
grasslands
(Willems,
1990).
Semi
-natural
grasslands
in
Norway
have
high
species
richness,
and
include
several
vulnerable
species
(Nordhagen,
1943;
Losvik,
1988,
1993;
Norderhaug,
1988;
Norderhaug,
1996;
Moen,
1990;
Kielland-Lund,
1992).
However,
0006-3207/98/$
see
front
matter
©
1998
Elsevier
Science
Ltd.
All
rights
reserved
P11:
S0006-3207(98)00071-8
370
G.
Austrheim
et
al./Biologica
little
attention
has
been
paid
to
the
properties
of
diver-
sity
and
composition
in
sub
-alpine
grasslands,
and
how
environmental
variables
affect
the
community
patterns.
Today,
the
summer
-farming
landscape
in
mountains
is
exposed
to
a
range
of
different
regimes
of
grazing,
trampling,
mowing,
nutrient
inputs
and
soil
dis-
turbance,
which
effect
the
maintenance
of
high
species
richness
and
conservation
of
vulnerable
species.
This
paper
presents
a
case
study
from
grassland
sites
main-
tained
by
traditional
or
near
traditional
management
over
several
centuries.
The
Budalen
area
in
central
Norway
represents
one
of
the
few
mountain
valleys
in
Scandinavia
where
summer
farming
in
mountains
is
still
practised,
which
gives
the
area
an
intrinsic
value
for
future
maintenance
as
a
cultural
heritage.
The
questions
in
this
study
are
related
to
the
ecological
properties
of
this
system
and
their
conservation
values:
1.
What
are
the
main
patterns
of
species
diversity,
species
composition
and
environmental
variables
in
these
semi
-natural,
sub
-alpine
grasslands?
Trondheim
Conservation
87
(1999)
369-379
2.
How
do
land
-use
versus
soil
and
topography
vari-
ables
affect
plant
community
patterns?
3.
What
is
the
importance
of
different
land
-use
regimes
for
the
conservation
of
plant
biodiversity
in
these
human
-induced
grasslands?
2.
Study
area
The
Budalen
valleys
(Budal
and
Endal)
are
situated
in
the
region
of
Gauldalen,
central
Norway
(Fig.
1).
Mountain
summer
-farming
has
been
important
in
this
region
since
the
16th
century
and
is
still
commonly
practised
(Olsson
et
al.,
1995).
The
human
-induced
landscape
of
Budalen
can
be
divided
into
outlying
land
(pastures)
and
small
grass-
land
enclosures
at
each
farm
site
(mean
<
0.5
ha).
Tra-
ditionally
the
enclosures
were
manured.
Mowing
was
practised
annually
at
the
beginning
of
August
leaving
a
short
period
in
early
spring
and
late
autumn
for
live-
stock
grazing.
The
outlying
pasture
areas
are
char-
acterised
by
a
small-scale
pattern
of
different
habitats
Oslo
`,
Farm
(permanent
settlement)
A
Summer
farm
Bogen
S
Haugen
0
EMI
Enlisaatra
SANDHOGDA
Singsas
Hermo
TAGAFJELLET
BUDAL
7
-0
-o
1
Hestflotssetra
7
MIDDAGSNIPPEN
0
9.
0
Mossetra
„0,1=•••
SAUHOGNA
1
10
km
Fig.
1.
Study
area.
The
valley
of
Budalen,
central
Norway.
G.
Austrhe
m
et
al./Biological
Conservation
87
(1999)
369-379
371
shaped
by
long
term
use
for
livestock
grazing,
hay-
making
and
harvesting
of
leaves
and
lichens,
and
fuel
wood
(Olsson
et
al.,
1995).
Modern
ley
cultivation
with
ploughing
and
sowing
of
ley
species
in
enclosures
is
not
common.
In
1993,
140
summer
farms
with
enclosures
were
registered,
and
20
units
were
still
in
use
with
dairy
production.
Approximately
70%
of
the
enclosed
grass-
land
area
is
being
used
for
grazing
and
haymaking.
Haymaking
on
outlying
land
ceased
in
the
1950s
and
has
been
replaced
by
grazing.
Present
grazing
pressure
(c.
3000
sheep
and
c.
700
cattle
in
1993)
is
high
and
even
exceeds
former
estimates
(Olsson
et
al.,
1995).
Within
the
Budalen
valleys,
41
sites
(27
enclosures
and
14
pastures),
below
the
present
Betula
pubescens
tree
limit
(altitude
c.
900
m)
and
above
the
highest
situ-
ated
permanent
farm
(c.
600
m)
were
investigated.
The
climate
of
Budalen
is
slightly
oceanic
with
a
mean
annual
precipitation
(1961-1990)
of
760
mm
(Forland,
1993).
Mean
annual
temperature
for
the
same
period
for
two
sites
close
to
the
Budalen
area,
was
0.3°C
(Roros)
and
2.5°C
(Berkak)
(Aune,
1993).
This
relatively
cold
climate
gives
a
limited
growing
season
(mean
temperature
>
6°C)
of
140-160
days
(Pahlson,
1984).
The
bedrock
(granite
and
amphibolic
quartz)
is
rich
in
lime,
but
partially
covered
with
quarternary
deposits
(glacial
and
glacifluvial),
and
peat
(Nilsen,
1991).
3.
Methods
3.1.
Field
survey
The
vegetation
of
41
sites,
were
investigated
during
1993
and
1994.
The
enclosure
sites
reflect
different
degrees
of
impact,
varying
from
semi
-natural
grassland
to
ley
(regularly
ploughed
and
fertilised).
All
pasture
sites
are
semi
-natural
grasslands
(Fig.
2).
At
each
site
a
homogeneous
plot
of
100
m
2
was
chosen,
and
within
each
plot
10
randomly
chosen
quadrats
(0.25
m
2
)
were
used
for
vegetation
analysis.
Each
quadrat
was
divided
into
16
sub
-quadrats
for
recording
vascular
plant
fre-
quency.
In
the
pasture
sites,
bryophytes
and
lichens
were
also
recorded.
Nomenclature
follows
Lid
(1985).
Nine
different
environmental
variables
were
recorded
for
each
100
m
2
plot
(Table
1).
Topography
is
expressed
by
heat
index
(HI)
and
altitude
(ALT).
HI
was
calculated
from
aspect
on
a
linear
scale
where
225°
was
considered
most
favourable
and
given
the
value
0,
and
25°
was
considered
least
favourable
and
given
the
value
200°.
The
scaling
thus
implies
that
e.g.
200°
would
have
the
same
HI
as
250°.
Chemical
and
physical
analyses
of
soil
samples
collected
in
1994
include
loss
of
ignition
(LOI),
dry
weight
(DW),
pH,
soluble
phosphate
(P),
potassium
(K)
and
nitrogen
(NO
3
and
NH
4
).
All
soil
analyses
were
performed
at
Agricultural
Centre
for
Analysis,
As,
Nor-
way
following
international
standards
(Krogstad,
1992).
No.
of
sites
15
-
1
0
5
0
Grazing
Grazing,
lertilising
Mowing
O
Mowing,
fertilising
Not
ploughed
<
1900
1901-20 1921-40 1941-60 1961-80
>1981
Ploughing
date
Fig.
2.
Land
-use
within
41
sites.
Ploughing
date
is
classified
into
seven
categories.
Fertilising,
mowing,
and
grazing
are
given
as
categorical
variables.
Table
1
Mean
values
and
SD
for
topography
and
soil
variables
in
(1)
all
sites,
(2)
enclosures,
(3)
pastures.
Test
of
differences
between
enclosures
and
pastures
are
given
by
a
Mann
-Whitney
U
-test
Variables
All
sites
Enclosures
Pastures
p
Mean
SD
Mean
SD
Mean
SD
Heat
index
68
38
70
39
65
37
Altitude
(m)
733
41
733
40
733
47
Loss
of
ignition
(%,
g/100
g)
13
3.4
14
2.7
10
2.9
***
Dry
weight
(%,
g/100
g)
63.1
5.5
62.7
5.1
64
6.2
pH
5.5 0.3
5.4
0.2
5.6
0.3
**
P
-AL
(mg/100
g)
2.5
1.2
3
1.1
1.4
0.6
***
K
-AL
(mg/100
g)
13
3.6
14.8
2.9
9.7
2.2
***
NH
4
(mg/1000
g)
10.9
6.9
10.9
7.3
11
6.1
NO
3
(mg/1000
g)
**p
<
0.01,
***p<0.001.
3.2.
Land
-use
Land
-use
(Fig.
2)
was
recorded
by
interviewing
land
owners,
and
from
official
statistics,
and
expressed
by
fi
ve
variables:
grazing,
mowing,
fertilising
and
habitat
category
(pasture
or
enclosure),
plus
ploughing
fre-
quency
classified
into
six
categories
ranging
from
"no
ploughing"
to
"ploughed
after
1981".
3.3.
Data
analysis
Vegetation
diversity
was
expressed
as
species
richness
in
both
enclosures
and
pastures
(mean
per
0.25
m
2
).
Total
species
pool
was
classified
into
fi
ve
groups
based
mainly
on
habitat
information
given
in
Lid
(1985)
(Table
2):
(1)
ALP:
Alpine
species
not
favoured
by
grassland
management,
(2)
ALPGRA:
Alpine
species
favoured
by
grassland
management,
(3)
GENGRA:
372
G.
Austrheim
et
al./Biologica
General
grassland
species,
(4)
WNL:
Weeds,
nitro-
phytes
and
ley
species,
(5)
OSL:
Other
sub
-alpine
and
lowland
species.
Differences
in
species
richness
between
enclosures
and
pastures
for
all
species
groups
were
given
by
a
Mann
—Whitney
U
-test.
Conservation
87
(1999)
369-379
Relations
between
land
-use
and
other
environmental
variables
describing
each
site
were
analysed
using
corre-
lation
analysis
(Kendall's
non
-parametric
test,
Conover,
1980).
Linear
regression
was
used
for
testing
the
rela-
tionship
between
environmental
variables
and
species
Table
2
Classification
of
habitat
"origins"
(Hab.)
for
vascular
plants
in
41
sub
-alpine
grasslands
based
on
Lid
(1985)
and
personal
observations
Hab.
Species
Hab.
Aconitum
septentrionale
3
Phyllodoce
caeruleaa
Salix
glauca
Salix
hastataa
Salix
herbaceaa
Salix
lanataa
Salix
lapponum
Salix
reticulataa
Saxifraga
aizoidesa
2
Alchemilla
alpina
2
Antennaria
alpinaa
Astragalus
alpinusa
Bartsia
alpinaa
Carex
bigelowia
Carex
capillarisa
2
Equisetum
variegatuma
Erigeron
borealisa
Euphrasia
frigidaa
Gentian
nivalisa
2
Gnaphalium
norvegicum
2
Gnaphalium
supinuma
Luzula
frigida
Luzula
spicataa
Pedicularis
oederf
Phleum
alpinum
2
Poa
alpina
Polygonum
viviparum
Primula
scandinavicaa
Saussurea
alpina
S
bbaldia
procumbensa
Silene
acaulisa
2
Thalictrum
alpinum
Veronica
alpina
Viola
biflora
3
Achillea
millefoliuma
Agrostis
capillaris
Ajuga
pyramidalisa
Alchemilla
glabra
Alchemilla
monticola
Alchemilla
sp.a
Alchemilla
subcrenata
Alchemilla
wichuraea
Antennaria
dioica
Anthoxanthum
odoratum
Botrychium
lunaria
3
Campanula
rotundifolia
4
Carex
canescensa
4
Carex
dioicaa
4
3
Carex
flacca
4
Species
Hab.
Species
Carex
nigra
Carex
ornithopodaa
Carex
pallescens
Carex
paniceaa
Carex
piluliferaa
Carex
pulicarisa
Carex
vaginata
Coeloglossum
viride
Deschampsia
cespitosa
Equisetum
pratensea
Euphrasia
strictab
Festuca
ovina
Festuca
rubra
Filipendula
ulmaria
Galium
boreale
Galium
uliginosum
Gentianella
campestris
Geum
rivale
Gymnadenia
conopseaa
Hieracium
pilosellaa
Hieracium
sp.
Hierochloe
odorataa
Hypericum
maculatuma
Juncus
filiformisa
Leontodon
autumnalis
Leucanthemum
vulgare
Luzula
mult
ora
Melampyrum
pratensea
Molinia
caeruleaa
Nardus
stricta
Parnassia
palustrisa
Poa
pratensis
Potentilla
crantzi
Potentilla
erecta
Prunella
vulgaris
Ranunculus
auricomusa
Rhimmthus
minor
Runiex
acetosa
Rumex
acetosella
Selaginella
selaginoides
Silene
diocia
b
Stellaria
graminea
Veronica
officinalis
Viola
caninaa
Viola
tricolour
b
Achillea
ptarmicab
Anthriscus
sylvestrisa
Cerastium
fontanum
Equisetum
arvensea
Myosotis
arvensisb
Phleum
pratense
b
Plantago
majors
Poa
annua
Ranunculus
acris
Sagina
procumbens
Stellaria
mediab
Taraxacum
sp.
Trifolium
pratense
Trifolium
repens
Veronica
serphyllifolia
Alnus
incanaa
Anemone
nemorosa
b
Betula
nana
Betula
pubescens
Calluna
vulgarise
Cirsium
helenioidesa
Crepis
paludosaa
Dactylorhiza
maculataa
Deschampsia
flexuosa
Empetrum
sp.
Epilobium
palustrea
Equisetum
hyemalea
Equisetum
palustrea
Equisetum
sylvaticuma
Eriophorum
angustifoliuma
Galium
palustrea
Geranium
sylvaticum
Gnaphalium
sylvaticum
b
Juniperus
communis
Listera
ovata
b
Luzula
pilosa
Luzula
sudeticaa
Maianthemum
bifolium
Oxalis
acetosellaa
Paris
quadrifoliab
Pinguicula
vulgaris
b
Pyrola
minors
Rubus
chamaemorusa
Rubus
saxatilis
Salix
phylicifolia
Scirpus
cespitosusa
Solidago
virgaurea
Sorbus
aucupariaa
Tofieldia
pusillaa
Trientalis
europea
Vaccinium
myrtillus
Vaccinium
uliginosum
Vaccinium
vitis-idaea
Viola
riviniana
b
a
Species
found
only
in
pastures.
b
Species
found
only
in
enclosures.
(1)
ALP;
Alpine
species
not
favoured
by
grassland
management,
(2)
ALPGRA;
Alpine
species
favoured
by
grassland
management,
(3)
GENGRA;
General
grassland
species,
(4)
WNL;
Weeds,
nitrophytes
and
ley
species,
(5)
OSL;
Other
sub
-alpine
and
lowland
species
not
favoured
by
grassland
management.
G.
Austrhe
m
et
al./Biological
Conservation
87
(
1999)
369-379
373
richness
at
each
site.
All
tests
were
available
in
the
sta-
tistical
program
package
of
SPSS,
release
7.0
(SPSS
Inc.,
Chicago).
Gradient
structure
in
vegetation
composition
and
its
environments
was
analysed
using
ordination
techniques
performed
by
CANOCO
(ter
Braak,
1987ter
Braak,
1990),
arranging
species,
samples
and
environmental
variables
along
axes.
The
importance
(percentage
variance
explained)
of
each
axis
is
expressed
by
eigenvalues.
Gradient
length
in
standard
deviation
units
is
given
through
non-linear
rescaling
of
each
axis
and
expresses
species
turnover
along
the
axes.
Detrended
Correspondence
Analysis
(DCA)
(Hill,
1979;
Hill
and
Gauch,
1980)
calculate
theoretical
ecolo-
gical
gradients
in
the
grasslands
as
a
function
of
species
and
sample
variation.
Correlation
analysis
(Kendall
non
-parametric)
was
used
to
examine
the
relationship
between
DCA
axes
sample
scores
and
the
corresponding
value
of
each
environmental
variable.
Canonical
correspondence
analysis
(CCA)
(ter
Braak,
1987ter
Braak,
1990)
was
used
to
produce
constrained
ordination
axes
that
have
a
direct
relationship
with
the
environmental
variables
measured.
Each
variable
was
in
turn
used
as
constrained,
and
the
fraction
of
variation
in
species
data
explained
was
obtained
by
a
rescaled
hybrid
CCA
as
described
by
Okland
and
Eilertsen
(1994).
All
variables
added
were
tested
to
be
statistically
significant
by
means
of
a
Monte
Carlo
permutation
test
given
in
CANOCO.
4.
Results
4.1.
Main
patterns
of
species
diversity
and
species
composition
A
survey
of
vascular
plants
in
both
enclosures
and
pastures
gave
a
total
of
148
species
(Table
2)
of
which
13
species
were
exclusive
for
enclosures
and
62
for
pas-
tures.
Sixty-one
bryophytes
and
lichens
were
found
in
the
pastures.
Average
species
richness
per
0.25
m
2
for
all
sites
was
21.3
for
vascular
plants
and
6.6
for
cryptogams
in
pastures.
Species
composition
reflected
a
mixture
of
alpine
and
lowland
species
with
a
dominance
of
general
grassland
species
(GENGRA)
for
both
enclosures
and
pastures
(Fig.
3).
Alpine
species
favoured
by
grassland
management
(ALPGRA)
together
with
other
sub
-alpine
and
lowland
species
(OSL)
accounted
for
>
20%
in
pastures.
ALPGRA,
OSL
and
WNL
(weeds,
nitro-
phytes
and
ley
species)
were
all
relatively
important
in
enclosures
(10-15%)
while
WNL
species
only
con-
stituted
5%
of
pasture
species.
Other
alpine
species
(ALP)
were
not
important
in
any
habitat.
There
was
a
significant
difference
between
vascular
plant
species
richness
in
enclosures
and
pastures
for
all
species
groups
(Mann
—Whitney
U
-test,
p<
0.001).
30
25
20
TO
5
0
Pastures
Enclosures
no
ALP
ALPGRA
, ,
GENGRA
WNL
OSL
TOT
Fig.
3.
Species
richness
(mean
per
0.25
m
2
)
given
as
a
total
of
all
vascular
plant
species
and
per
species
group,
in
enclosures
(n=27),
and
pastures
(n=14).
(1)
ALP
=
Alpine
species
not
favoured
by
grassland
manage-
ment.
(2)
ALPGRA
=
Alpine
species
favoured
by
grassland
manage-
ment.
(3)
GENGRA
=
General
grassland
species.
(4)
WNL
=
Weeds,
nitrophytes
and
ley
species.
(5)
OSL
=
Other
sub
-alpine
and
lowland
species
not
favoured
by
grassland
management.
Differences
in
species
richness
between
enclosures
and
pastures
were
significant
for
all
species
groups
due
to
a
Mann
—Whitney
U
-test
(p
<
0.001).
Six
vulnerable
grassland
species
with
a
decreasing
abundance
in
Norway
(Kielland-Lund,
1992)
were
all
concentrated
in
the
species
-rich
sites
namely
Primula
scandinavica,
Gymnadenia
conopsea,
Gentianella
cam-
pestris,
Carex
ornithopoda
and
Carex
pulicaris
(Fig.
4).
Botrychium
lunaria
was
an
exception
with
a
more
even
distribution
along
the
species
richness
gradient.
Most
species
have
a
low
frequency
of
regional
dis-
tribution
and
local
abundance;
66%
occurred
in
<
10%
of
the
quadrats
while
mean
abundance
was
<
10%
for
84%
of
the
total
species
pool
(Table
3).
The
relative
proportion
of
ALP,
WNL
and
OLS
species
was
highest
in
the
low
frequency
group
(
<
2%)
while
ALPGRA
and
GENGRA
species
had
a
higher
relative
proportion
in
the
high
frequency
group
(
>
10%).
Abundances
of
ALP,
WNL
and
OSL
species
had
an
even
higher
rela-
tive
proportion
in
the
low
frequency
group
(
<
2%)
compared
to
distribution
values.
Regional
distribution
and
local
abundance
of
vulnerable
species
listed
above
were
low.
Botrychium
lunaria
occurred
in
10.2%
of
the
quadrats,
while
the
other
species
had
a
frequency
<
2.7%.
All
vulnerable
species
had
a
mean
abundance
<
1.2%.
Successional
tree
and
shrub
species
were
com-
mon
in
both
habitats.
The
relative
distribution
fre-
quency
was
higher
in
pastures
as
compared
to
enclosures
for
Betula
pubescens
(39
and
7%,
respectively)
and
Juni-
perus
communis
(13
and
3%,
respectively)
while
Salix
spp.
had
a
higher
distribution
in
enclosures
(52%)
as
compared
to
pastures
(22%).
4.2.
Environmental
variables
and
vegetation
pattern
There
were
significant
differences
between
enclosures
and
pastures
for
four
soil
variables
(Table
1).
Loss
of
374
G.
Austrheim
et
al./
Biological
Conservation
87
(1999)
369-379
ignition,
P
and
K
had
higher
values
in
enclosures
(p
<
0.001),
while
pH
values
were
higher
in
pastures
(p
<
0.01).
DCA-ordination
of
the
total
sample
set
(410
samples)
shows
a
strong
gradient
separating
enclosures
and
pasture
sites
along
axis
1
(eigenvalue
0.43)
with
a
high
species
turnover
(gradient
length
3.45
S.D.
units)
(Fig.
5).
Pas-
tures
(Pas),
grazing
(Gra)
and
pH
are
positively
correlated
with
axis
1,
while
enclosures
(Enc),
P,
K,
ploughing
(Plo),
fertilising
(Fer)
and
mowing
(Mow),
show
a
negative
correlation
(Table
4).
Eigenvalues
and
gradient
length
are
reduced
along
axis
2
(0.21
and
2.42
S.D.
units).
Dry
weight
(DW)
shows
a
high
positive
correlation
with
the
axis
while
K
has
the
highest
negative
correlation
value.
Correlation
analyses
between
fertilising
and
soil
nutrient
content
show
a
significant
relationship
only
for
NH
4
(p
<
0.01,
Table
4).
Distribution
of
samples
in
the
ordination
diagram
of
Fig.
5,
emphasise
the
main
differences
between
pastures
and
enclosures.
Land
-use
variation
within
enclosures
is
classified
due
to
ploughing
and
fertilisation
regimes,
using
"time
since
last
ploughing"
as
the
primary
classification
35
30
25
C
20
'17
(,)
-
a
-
15
a)
co
10
criterion.
Classes
1-9
could
thus
be
considered
to
repre-
sent
a
gradient
in
land
-use
intensity
ranging
from
recently
ploughed
and
fertilised
samples
(class
1)
to
non
-ploughed
and
unfertilised,
grazed
pastures
(class
9).
Except
for
class
1
and
class
9
at
each
end
of
axis
one,
this
intensity
gradient
is
not
clearly
evident
in
the
diagram,
and
enclosure
sam-
ples
from
classes
2-8
are
more
or
less
scattered
(Fig.
5).
4.3.
Variation
explained
by
each
environmental
variable
A
direct
gradient
analysis,
using
a
rescaled
hybrid-
CCA
ordination,
shows
the
importance
of
each
envir-
onmental
variable
for
three
different
sample
sets:
all
sites,
enclosures
and
pastures
(Table
5).
The
largest
fraction
of
variation
of
the
total
sample
set
is
explained
by
differences
between
enclosures
and
pastures
(6.9%),
pH
(6.3%)
and
P
(5.1%).
Each
separate
land
-use
variable
is
important:
ploughing
(4.8%),
grazing
(4.1%),
mowing
(3.6%)
and
fertilising
(3.1%).
Dry
weight
(3.6%)
and
loss
of
ignition
(3.2%)
are
also
important.
Beta
-diver-
sity
(SD
values)
show
a
similar
ranking
for
the
most
important
variables:
pH
(2.83),
P
(2.17)
and
ploughing
—Species
richness
—Sot
lun
-
Pri
sca
Gen
cam
--
Gyn
corn
-
Car
orn
-
pul
10
-
7
0)
_C
ccs
C
4
(D
C
3
7
0
C
C
2
& I
Sites
6
5
Fig.
4.
Abundances
of
six
vulnerable
grassland
species
within
each
site.
Sites
are
ranked
according
to
total
species
richness.
Bot
lun=Botrychium
lunaria,
Pri
sca
=
Primula
scandinavica,
Gen
cam=
Gentianella
campestris,
Gym
con
=
Gymnadenia
conopsea,
Car
orn=
Carex
ornitopoda,
Car
pul
=
Carex
pulicaris.
Table
3
Distribution
(DIS)
and
abundance
(ABU)
of
all
species
(TOT),
and
species
classified
into
different
habitat
groups
(cf.
Table
2).
Species
are
sorted
into
frequency
groups
Frequency
groups
ALP
ALPGRA
GENGRA
WNL
OSL
TOT
DIS
ABU
DIS
ABU
DIS ABU
DIS
ABU
DIS
ABU
DISABU
>
10%
11
11
40
12
47
25
27
13
21
8
3416
<10>2%
22
0
32
40
22
28
33
19
33
21
2826
<2%
66
89
28
48
32
48
40
69
46
71
3858
G.
Austrheim
et
al./Biologica
(2.01).
Nominal
land
-use
variables
have
short
gradients:
enclosures
and
pastures
(1.37),
mowing
(1.02),
fertilising
(0.98)
and
grazing
(0.97).
A
separate
analysis
of
enclosures
and
pastures
sample
sets
(Table
5),
gives
a
different
ranking
of
environmental
variables.
Altitude
(4.8%),
plowing
(4.7%)
and
K
(4.3%)
are
most
important
in
explaining
the
variation
for
enclosures.
P
(4.5%)
is
the
most
important
variable
for
pastures.
NH
4
explains
only
2.6%
while
other
soil
and
topography
variables
explain
between
3.9%
and
3.0%.
vi
Conservation
87
(1999)
369-379
375
4.4.
Environmental
variables
and
species
richness
Environmental
variables
show
a
linear
relationship
with
species
richness.
Linear
regression
of
normally
distributed
soil
and
topography
variables
with
mean
species
richness
(0.25
m
2
)
at
each
site
show
a
significant
relationship
for
four
variables
(Table
6).
Species
rich-
ness
decreases
with
increasing
values
of
P
(r
2
=
0.65)
and
K
(r
2
=
0.42),
but
increases
with
the
values
of
pH
(r
2
=
0.35).
A
decrease
in
loss
of
ignition
is
also
sig-
7
-'---------
0
o
o
0 0
00
0 0
.•
AA
.1
f
A
1,40
0
0
0
00
A
V
A
A
A
V
.
At
y
iv
v
A
Avy
I
t
,
1.,.
A
t
,r1A4*.
0
,
••
01.
1
„.►
0
o
0
0 0
0 0
0
Site
classes
CLASS
1
Ploughed
after
1980
and
fertilised
CLASS
2
Ploughed
1961-80
and
not
fertilised
CLASS
3
Ploughed
1941-60
and
fertilised
CLASS
4
A
Ploughed
1941-60
and
not
fertilised
CLASS
5
v
Ploughed
1921-40
and
not
fertilised
CLASS
6
Ploughed
1901-20
and
not
fertilised
CLASS
7
*
Ploughed
<1900
and
fertilised
CLASS
8
*
Ploughed
<1900
and
not
fertilised
CLASS
9
0
Neither
ploughed
nor
fertilised
0
0
0
00
O
°O
0
0
0 0
0 0
0 0
0
0
00
00
oo
o
a:
bp
,40
O
00
0e
+
3.5
Bag.
0.43
S.D.
3.45
Axis
1
Fig.
5.
DCA-ordination
of
41
sites
classified
due
to
land
-use.
Classes
1-8
=
enclosures,
class
9
=pastures.
Classes
1
and
9
are
encircled.
Table
4
Kendall
correlation
coefficients
for
(a)
DCA
axis
1
and
axis
2
related
to
environmental
variables
and
(b)
land
-use
variables
related
to
soil
and
topography
(a)
Hi
Alt
LOI
DW
pH
NH
4
Mow
Gra
Fer
Plo
Enc
Pas
DCA1
-0.156
0.047
-0.282*
-0.001
0.321*
-0.762*** -0.459***
-0.077
-0.487***
0.448**
-0.456*** -0.644***
-0.672***
0.672***
DCA2
0.163
-0.244*
-0.156
0.308*
0.106 0.135
-0.333**
0.061
-0.064
0.311*
0.131
0.152
-0.016
0.016
(b)
Mow
Gra
Fer
Plo
Enc
Pas
Hi
Alt
LOI
DW
pH
NH
4
-0.160
0.279*
0.327*
-0.218
-0.262
0.364** 0.442**
0.097
0.180
-0.235
-0.507***
0.424**
0.290*
-0.403**
-0.576***
-0.017
0.085
0.180
0.006
-0.095
0.038
0.273
0.253
0.345**
0.245
-0.206
0.255*
-0.001
-0.314*
0.576***
0.259*
-0.100
0.035
-0.011
0.527***
-0.187
-0.357*
0.615*** 0.572***
-0.058
-0.035
0.011
-0.527***
0.187
0.357*
-0.615*** -0.572***
0.058
Hi,~heat
index;
Alt,~altitude;
LOI,~loss
of
ignition;
DW,
dry
weight;
Mow,mowing;
Gra,
grazing;
Fer,
fertilising;
Plo,
ploughing;
Enc,~enclosures;
Pas,~pastures.
376
G.
Austrheim
et
al./Biologica
Conservation
87
(1999)
369-379
Table
5
Direct
gradient
analyses
of
(1)
all
sites,
(2)
enclosures
and
(3)
pastures
All
sites
Enclosures
Pastures
Variable
Eigenvalue
0/
0
SD
Eigenvalue
0/
0
SD
Eigenvalue
0/
0
SD
Env
0.352
6.9
1.37
Pas
0.352
6.9
1.37
pH
0.320
6.3
2.83
0.080
2.6
1.182
0.136
3.7
1.332
P
0.262
5.1
2.17
0.115
3.7
1.611
0.168
4.5
1.634
Plo
0.245
4.8
2.01
0.146
4.7
1.276
Gra
0.209
4.1
0.97
0.097
3.1
0.698
DW
0.186
3.6
1.93
0.116
3.7
1.554
0.137
3.7
1.553
Mow
0.185
3.6
1.02
0.113
3.6
0.685
LOI
0.165
3.2
1.90
0.094
3.0
1.316
0.125
3.3
1.482
Fer
0.157
3.1
0.98
0.102
3.3
0.638
Alt
0.100
2.0
1.17
0.149
4.8
1.417
0.113
3.0
1.085
Hi
0.099
1.9
1.21
0.108
3.5
1.251
0.144
3.9
1.115
0.091
1.8
1.54
0.133
4.3
1.642
0.141
3.8
1.326
NH
4
K
0.075
1.5
1.10
0.096
3.1
1.156
0.097
2.6
1.297
Eigenvalue,
percent
variation
explained,
and
gradient
length
in
SD
units
is
given
for
canonical
axis
1
of
each
variable
within
each
sample
set.
All
variables
are
significant
(p
<
0.01).
Table
6
Linear
regression
of
species
richness
(mean
per
0.25
m
2
)
and
different
soil
and
topographical
variables
Variable
All
sites
Enclosures
Pastures
r
2
rc
df
r
2
rc
df
r
2
rc
df
P
0.65***
-4.26
39
0.49***
-2.17
25
0.32*
-3.93
12
K
0.42***
-1.19
39
<0.01
-0.11
25
0.20
-0.80
12
pH
0.35***
15.07
39
0.01
2.38
25
0.55**
9.46
12
LOI
0.17**
-0.80
39
0.03
0.23
25
0.06 0.36
12
Hi
0.08
-0.05
39
0.05
-0.02
25
0.55**
-0.08
12
Alt
0.02
0.02
39
0.04
0.02
25
0.08
0.02
12
DW
<0.01
-0.05
39
0.02
-0.11
25
0.17
-0.26
12
NH
4
<0.01
0.01
39
0.14*
-0.80
25
12
r
2
,
goodness
of
fi
t
of
a
linear
model,
rc,
regression
coefficients
and
df,
degrees
of
freedom.
Three
sample
sets
are
given:
(1)
total
sample
set
(n
=
41),
(2)
enclosures
(n
=
27),
(3)
pastures
(n
=
14).
*p
<0.05,
**p<0.01,
***p<0.001.
nificantly
related
to
an
increase
in
species
richness
but
the
fraction
explained
is
low
(r
2
=
0.17).
A
separate
lin-
ear
regression
within
pastures
shows
a
similar
decrease
in
species
richness
with
increasing
levels
of
P
(r
2
=
0.32)
and
an
increase
with
higher
levels
of
pH
and
heat
index.
Species
richness
in
enclosures
decreases
with
higher
levels
of
P
and
NH
4
.
5.
Discussion
The
semi
-natural,
sub
-alpine
grasslands
in
Budalen
originated
within
the
region
of
potential
Betula
wood-
land.
Human
activities
since
prehistoric
time
have
been
the
major factors
controlling
the
diversity,
composition
and
dynamics
of
these
communities,
and
their
future
maintenance
demands
continued
human
impact
to
pre-
vent
forest
succession.
In
a
cultural
perspective,
these
semi
-natural,
sub
-
alpine
grasslands
have
been
maintained
within
the
mixed
agricultural
system
of
mountain
summer
-farm-
ing,
and
represent
by-products
of
a
pre
-industrial
agri-
cultural
system
trying
to
achieve
high
agrarian
production
under
the
available
methods.
The
conserva-
tion
values
of
such
human
shaped
ecosystems,
and
the
importance
of
considering
both
ecological
and
cultural
factors
have
been
acknowledged
(Birks
et
al.,
1988;
Berglund,
1991;
Pickett
et
al.,
1992;
Heywood,
1995).
The
main
community
variation
of
sub
-alpine,
semi
-
natural
grasslands
in
Budalen
is
related
to
the
land
-use
differences
between
pastures
and
enclosures.
Species
composition,
species
richness
and
values
of
environmental
G.
Austrheim
et
al./Biologica
parameters
all
differ
between
sites
belonging
to
the
two
land
-use
categories.
The
conservation
values
of
pastures
as
compared
to
enclosures
is
mainly
related
to
c.
50%
higher
values
of
species
richness
and
a
higher
frequency
of
vulnerable
species.
High
species
diversity
of
the
semi
-natural
pastures
on
base
-rich
soils
in
Budalen
are
consistent
with
data
from
European
calcareous
grassland
in
general
(van
der
Maarel
and
Titlyanova,
1989;
Willems,
1990).
Further,
it
is
known
that
several
species
within
these
commu-
nities
have
their
origins
and
main
occurrence
in
other
habitats
(Bengtsson-Lindsjo
et
al.,
1991;
Linusson
et
al.,
1998).
A
characteristic
feature
of
sub
-alpine
grassland
composition
is
that
alpine
species
coexist
with
species
from
sub
-alpine
and
lowland
habitats.
Twenty
percent
of
the
species
in
this
study
have
their
main
distribution
in
alpine
areas
which
indicates
an
elevational
dispersion
of
alpine
species
into
human
induced,
sub
-alpine
grass-
lands.
Differences
in
species
composition
between
enclosures
and
pastures
are
related
to
differences
in
fre-
quencies
of
species
with
different
habitat
origins.
The
following
species
groups
namely,
(1)
alpine
species
not
favoured
by
grassland
management
(ALP),
(2)
alpine
species
favoured
by
grassland
management
(ALPGRA),
(3)
general
grassland
species
(GENGRA)
and
(4)
other
sub
-alpine
and
lowland
species
(OSL),
all
have
sig-
nificantly
lower
frequencies
in
enclosures.
This
indicates
that
these
species
have
a
lower
tolerance
to
the
specific
disturbance
that
the
land
-use
in
enclosures
represent.
Four
vulnerable
species,
namely
Primula
scandina-
Gymnadenia
conopsea,
Carex
ornithopoda,
Carex
pulicaris,
are
absent
in
enclosures,
but
they
are
also
very
rare
in
pastures.
Botrychium
lunaria
and
Gentianella
campestris
are
also
rare
but
not
exclusive
for
pastures.
These
vulnerable
species
occur
in
sites
of
high
species
richness.
Thus,
diversity
and
vulnerability
do
not
seem
to
be
independent.
A
positive
relationship
between
high
species
richness
and
the
occurrence
of
vulnerable
Gen-
tianella
species
is
also
found
for
semi
-natural
grasslands
in
Sweden
(Lennartsson,
1997).
The
enclosures
represents
an
important
production
unit
due
to
grazing
and
haymaking,
and
are
highly
evaluated
from
an
agronomist
point
of
view.
In
addi-
tion,
the
enclosures
are
spatially
connected
to
the
farm
houses
and
thus
perceived
as
part
of
a
cultural
heritage.
This
study
found
no
corresponding
conservation
values
in
those
plant
communities
from
a
biological
point
of
view,
although
species
rich
grasslands
could
develop
in
former
arable
fi
elds
in
the
long
term
perspective
(e.g.
Wells
et
al.,
1976).
Low
levels
of
species
richness
in
enclosures
is
related
to
higher
levels
of
P,
K
and
loss
of
ignition
in
enclosure
sites,
and
higher
pH
levels
in
pas-
tures.
Plant
community
composition
is
dependent
on
both
the
rate
and
ratio
of
supply
of
limiting
soil
resources,
and
fertility
for
a
given
plant
species
com-
monly
depends
on
the
availability
of
N,
P,
K
(Allen
et
Conservation
87
(1999)
369-379
377
al.,
1974;
Tilman
et
al.,
1994).
High
nutrient
levels
are
generally
associated
with
a
decrease
in
species
richness
except
for
very
poor
soils
(Grime,
1973;
Grime,
1979;
Tilman,
1982).
The
amount
of
soil
NH
4
does
not
differ
between
the
two
habitats,
but
soil
phosphorus
is
an
important
variable
in
explaining
vegetation
variation
in
enclosures
and
pastures
separately.
Gough and
Marrs
(1990)
found
that
plant
growth
in
soils
from
different
grasslands,
varying
from
semi
-natural
to
ploughed
fi
elds
was
controlled
by
the
availability
of
P,
and
suggest
that
high
levels
preclude
the
re-establishment
of
species
-rich
grasslands.
The
same
experiment
showed
that
plant
growth
was
not
limited
by
the
availability
of
K.
In
our
sites
P
and
K
levels
are
significantly
correlated,
and
their
isolated
effects
can
not
be
measured.
Nevertheless,
the
direct
gradient
analysis
of
vegetation
environment
relationship
shows
that
P
is
a
much
stronger
explana-
tory
variable
than
K.
An
explanation
of
the
significant
differences
in
nutri-
ent
levels
between
the
two
habitats
could
be
the
former
manure
supply
in
enclosures
in
times
of
traditional
management.
Addition
of manure
also
increases
the
content
of
organic
matter
which
could
explain
a
sig-
nificantly
higher
loss
of
ignition
in
enclosures
while
dry
weight
values
do
not
differ
between
habitats.
Similar
soil
relations
are
found
between
enclosures
and
the
surround-
ings
at
summer
farms
in
western
Norway
(Vanvik,
1995).
The
enclosures
reflect
a
complexity
in
land
-use,
and
a
DCA
ordination
did
not
separate
the
effects
of
mowing,
fertilising
and
ploughing
on
community
patterns,
prob-
ably
due
to
this
complexity.
However,
an
exception
is
the
most
recently
ploughed
and
fertilised
sites
(class
1)
that
differ
from
other
enclosures.
The
positive
relationship
between
species
richness
and
soil
pH
is
consistent
with
data
from
semi
-natural
grass-
lands
in
England
(Grime,
1973;
Tilman
et
al.,
1994)
and
Denmark
(Ejrnes
and
Bruun,
1995).
Soil
pH
differences
in
Budalen
could
be
related
to
land
-use
differences
between
pastures
and
enclosures,
because
fertilisation
has
been
shown
to
lower
pH
levels
in
long-term
experi-
ments
(Tilman
et
al.,
1994).
However,
species
richness
also
increases
with
pH
within
pastures,
which
all
have
the
same
land
-use.
Grasslands
are
in
general
structured
by
processes
acting
on
different
spatial
and
temporal
scales
(Collins
and
Glenn,
1991).
The
heterogeneity
of
these
grassland
communities
is
given
by
a
high
total
number
of
vascular
plant
species
which
corresponds
with
high
values
of
spe-
cies
turnover
among
sites
(i.e.
beta
-diversity).
Processes
at
the
landscape
level
are
indicated
by
the
coexistence
of
species
from
different
distributional
groups.
Regional
dynamics
thus
effect
local
community
patterns
and
have
implications
for
the
maintenance
of
species
-rich
plant
communities
and
vulnerable
species.
In
addition,
all
vulnerable
species
have
low
frequencies
of
local
abun-
dance
and
regional
distribution
in
habitats
considered
378
G.
Austrheim
et
al./Biologica
to
be
their
main
habitats.
According
to
Saunders
et
al.
(1991),
extinctions
are
likely
to
occur
more
rapidly
if
species
are
(1)
restricted
to
vulnerable
habitats,
(2)
have
low
local
densities,
(3)
need
large
areas.
These
char-
acteristics
are
probably
valid
for
all
vulnerable
species
found
in
Budalen.
In
addition,
a
strategy
for
conserva-
tion
of
vulnerable
species
have
to
consider
the
general
dynamic
nature
of
grassland
habitats,
as
well
as
the
fact
that
these
semi
-natural
grasslands
are
exposed
to
a
temporal
shift
in
land
-use.
Pickett
and
Thompson
(1978)
argue for
a
preservation
of
a
"minimum
dynamic
area"
or
an
area
large
enough
to
contain
within
it
mul-
tiple
patches
of
various
disturbance.
Several
issues
con-
cerning
configuration
of
these
patches
such
as
the
critical
size
of
the
grassland
areas,
their
mutual
dis-
tances,
influence
of
neighbouring
areas
etc.
remain
unsolved
for
the
Budalen
area.
A
serious
concern
for
conservation
is
the
high
fre-
quency
of
woody
species
in
these
semi
-natural
habi-
tats,
above
all
in
the
pastures.
Studies
of
semi
-natural
grasslands
in
Norway
(Losvik,
1988;
Norderhaug,
1988,
Norderhaug,
1996)
have
shown
that
trees
and
shrubs
are
invading
former
hay
-making
areas
on
a
longer
time
scale,
in
spite
of
a
moderate
grazing
pressure.
In
general
it
seems
clear
that
semi
-natural
grasslands
should
be
maintained
by
the
processes
that
have
created
them,
i.e.
made
it
possible
for
grasslands
to
develop
by
prevention
of
forest
succession.
Even
if
we
do
not
have
data
avail-
able
for
assessing
changes
within
those
grasslands
over
the
full
period
of
their
existence,
our
present
knowledge
of
the
past
and
present
land
-use
indicates
that
con-
tinuation
of
the
traditional
management
would
max-
imise
the
probability
of
maintaining
biodiversity.
This
implies
that
grazing,
selective
cutting
of
trees
and
shrubs,
possibly
in
combination
with
mowing
is
a
pre-
requisite
for
the
long
term
conservation
of
semi
-natural
subalpine
grasslands
in
Norway.
Acknowledgements
Funding
to
this
research
project
to
E.G.A.O.
was
received
from
the
Norwegian
Research
Council.
We
thank
Drs.
M.E.
Edwards,
H.H.
Blom,
A.
Odland,
B.N.K.
Davis
and
two
anonymous
referees
for
com-
ments
on
earlier
versions
of
this
manuscript,
and
C.
Baggerud
for
computer
drawings
of
maps
and
fi
gures.
References
Allen,
S.E.,
Grimshaw,
H.M.,
Parkinson,
J.A.,
Quarnby,
C.,
1974.
Chemical
Analysis
of
Ecological
Materials.
Blackwell
Scientific,
Oxford.
Aune,
B.,
1993.
Temperaturnormaler,
normalperiode
1961-1990.
Norwegian
Metrological
Institute
2,
1-63.
Conservation
87
(1999)
369-379
Bengtsson-Lindsjo,
S.,
Ihse,
M.,
Olsson,
G.A.,
1991.
Landscape
pat-
terns
and
grassland
plant
species
diversity
in
the
20th
century.
In:
Berglund,
B.
E.
(Ed.),
The
Cultural
Landscape
during
6000
Years.
Ecological
Bulletins,
vol.
41.
Munksgaard,
Copenhagen,
pp.
388-
396.
Berglund,
B.E.
(Ed),
1991.
The
Cultural
Landscape
During
6000
Years.
Ecological
Bulletins,
Vol.
41.
Munksgaard,
Copenhagen.
Birks,
H.H.,
Birks,
H.J.B.,
Kaland,
P.E.,
Moe,
D.
(Eds),
1988.
The
Cultural
Landscape,
Past,
Present
and
Future.
Cambridge
Uni-
versity
Press,
Cambridge.
Collins,
S.L.,
Glenn,
S.M.,
1991.
Importance
of
spatial
and
temporal
dynamics
in
species
regional
abundance
and
distribution.
Ecology
72,
654-664.
Conover,
W.J.,
1980.
Practical
Nonparametric
Statistics.
Wiley,
New
York.
Ejrnxs,
R.,
Bruun,
H.H.,
1995.
Prediction
of
grassland
quality
for
environmental
management.
Journal
of
Environmental
Manage-
ment
41,
171-183.
Forland,
E.J.,
1993.
Nedborsnormaler,
normalperiode
1961-1990.
Norwegian
Metrological
Institute
39,
1-63.
Fremstad,
E.,
Elven,
R.
(Eds),
1991.
Enheter
for
vegetasjonskartleg-
ging
i
Norge.
NINA
utredning,
28,
1-196.
Gough,
M.W.,
Marrs,
R.H.,
1990.
A
comparison
of
soil
fertility
between
semi
-natural
and
agricultural
plant
communities:
implica-
tions
for
the
creation
of
species
-rich
grasslands
on
abandoned
agri-
cultural
land.
Biological
Conservation
51,
83-96.
Grime,
J.P.,
1973.
Control
of
species
density
in
herbaceous
vegetation.
Journal
of
Environmental
Management
1,
151-167.
Grime,
J.P.,
1979.
Plant
Strategies
and
Vegetation
Processes.
Wiley,
New
York.
Heywood,
W.H.
(Ed.),
1995.
Global
Biodiversity
Assesment.
Cam-
bridge
University
Press
for
UNEP,
Cambridge.
Hill,
M.O.,
1979.
DECORANA-a
FORTRAN
program
for
detren-
ded
correspondence
analysis
and
reciprocal
averaging.
Cornell
University
Press,
Ithaca,
NY.
Hill,
M.O.,
Gauch,
H.G.,
1980.
Detrended
correspondence
analysis:
an
improved
ordination
technique.
Vegetatio
42,
47-58.
Kielland-Lund,
J.,
1992.
Viktige
vegetasjonstyper
i
kulturlandskapet,
Ost-Norge.
NINA,
As.
Krogstad,
T.,
1992.
Metoder
for
jordanalyser.
Institutt
for
jordfag.
Rapport
6,
1-32.
Kvamme,
M.,
1988.
Pollen
analytical
studies
of
mountain
summer
farming
in
Western
Norway.
In:
Birks,
H.H.,
Birks,
H.
J.
B.,
Kaland,
P.E.,
Moe,
D.
(Eds.),
The
Cultural
landscape,
past,
present
and
future.
Cambridge
University
Press,
Cambridge,
pp.
349-367.
Lennartsson,
T.,
1997.
Demography,
reproductive
biology
and
adap-
tive
traits
in
Gentianella
campestris
and
G.
amarella-Evaluating
grassland
management
by
using
indicator
plant
species.
Ph.D.
the-
sis,
Swedish
University
of
Agricultural
Sciences,
Uppsala.
Lid,
J.,
1985.
Norsk,
svensk
og
fi
nsk
fl
ora.
Det
Norske
Samlaget,
Oslo.
Linusson,
A.G.,
Berlin,
G.A.I.,
Olsson,
E.G.A.
Reduced
community
diversity
in
semi
-natural
meadows
in
Southern
Sweden,
1965-1990.
Plant
Ecology,
136,
77-94.
Losvik,
M.H.,
1988.
Phytosociology
and
ecology
of
old
hay
meadows
in
Hordaland,
Western
Norway
in
relation
to
management.
Vege-
tatio
78,
157-187.
Losvik,
M.H.,
1993.
Total
species
number
as
a
criterion
for
conserva-
tion
of
hay
meadows.
In:
Bunce,
R.G.H.,
Rysckowski,
L.,
Paoletti,
M.G.
(Eds.),
Landscape
ecology
and
agroecosystems.
Lewis,
Boca
Raton,
FL,
pp.
105-111.
Marrs,
R.H.,
1993.
Soil
fertility
and
nature
conservation
in
Europe:
theoretical
considerations
and
practical
management
solutions.
Advances
in
Ecological
Research
24,
241-300.
Moe,
D.,
Indrelid,
S.,
Fasteland,
A.
(1988)
The
Halne
area,
Hard-
angervidda.
Use
of
a
high
mountain
area
during
5000
years
-
an
G.
Austrheim
et
al./Biologica
interdisciplinary
case
study.
In:
Birks,
H.H.,
Birks,
H.J.B.,
Kaland,
P.E.,
Moe,
D.
(Eds.),
The
Cultural
Landscape,
Past,
Present
and
Future.
Cambridge
University
Press,
Cambridge,
pp.
429-444.
Moen,
A.,
1990.
The
plant
cover
of
the
boreal
uplands
of
Central
Norway.
I.
Vegetation
ecology
of
Solendet
nature
reserve,
haymak-
ing
fens
and
birch
woodlands.
Gunneria,
vol.
63.
1-451.
Tapir,
Trondheim.
Nilsen,
0.,
1991.
The
bedrock
geology
of
the
Budalen
area.
In:
Espe-
lund,
A.
(Ed.),
Bloomery
ironmaking
during
2000
years.
Seminar
in
Budalen,
Sor-Trondelag,
Norway.
26-30
August
1991,
vol.
I.
Ancient
Ironmaking
in
a
Local
and
General
Norwegian
context.
Tapir,
Trondheim,
pp.
8-21.
Nordhagen,
R.,
1943.
Sikilsdalen
og
Norges
fjellbeiter.
En
planteso-
siologisk
monografi.
Bergens
Museums
skrifter
22,
1-607.
Norderhaug,
A.,
1988.
Urterike
slatteenger
i
Norge,
rapport
fra
for-
prosjekt.
Okoforsk
utredning
3,
1-92.
Norderhaug,
A.,
1996.
Hay
meadows:
biodiversity
and
conservation.
Ph.D.
thesis,
University
of
Goteborg,
Sweden.
Olsson,
E.G.A.,
1991.
Agro-ecosystems
from
Neolithic
time
to
the
present.
In:
Berglund,
B.
E.
(Ed.),
The
Cultural
Landscape
During
6000
Years.
Ecological
Bulletins,
vol.
41.
Munksgaard,
Copenha-
gen,
pp.
293-314.
Olsson,
G.A.,
Austrheim,
G.,
Bele,
B.,
Grontvedt,
E.,
1995.
Seter-
landskapet
i
Budalen
og
Endalen,
Midtre
Gauldal,
Midt
-
Norge.
Kulturhistoriske
og
okologiske
forhold
i
fjellets
kulturlandskap.
Fylkesmannen
i
Sea-
Trondelag,
Miljovernavdelingen.
Rapport
2,
1-89.
Okland,
R.H.,
Eilertsen,
0.,
1994.
Canonical
correspondence
analysis
with
variation
partioning:
some
comments
and
application.
Journal
of
Vegetation
Science
5,
117-126.
Pahlson,
L.
(Ed.),
1984.
Naturgeografisk
regionindelning
av
Norden.
Nordiska
Ministerradet.
Stockholm.
Pickett,
S.T.A.,
Thompson,
J.N.,
1978.
Patch
dynamics
and
the
design
of
nature
reserves.
Biological
Conservation
13,
27-37.
Pickett,
S.T.A.,
Parker,
T.,
Fiedler,
P.L.,
1992.
The
new
paradigm
in
ecology:
Implications
for
conservation
biology
above
the
species
level.
In:
Fiedler,
P.
L.,
Jain,
S.
K.
(Eds.),
Conservation
Biology.
Conservation
87
(1999)
369-379
379
The
Theory
and
Practice
of
Nature
Conservation,
Preservation
and
Management.
Chapman
and
Hall,
New
York,
pp.
65-88.
Price,
L.W.,
1981.
Mountains
and
Man.
A
Study
of
Process
and
Environment.
University
of
California
Press,
Los
Angeles,
CA.
Reinton,
L.,
1955.
Seterbruket
i
Noreg
I,
setertypar
og
driftsformer.
Institutt
for
sammenlignende
kulturforskning.
H.
Aschehoug
and
Co.,
Oslo.
Saunders,
D.A.,
Hobbs,
R.J.,
Margules,
C.R.,
1991.
Biological
con-
sequences
of
ecosystem
fragmentation:
a
review.
Conservation
Biology
5,
18-32.
ter
Braak,
C.J.F.,
1987.
CANOCO-a
FORTRAN
program
for
canonical
community
ordination
by
(partial)
(detrended)
(canoni-
cal)
correspondence
analysis,
principal
component
analysis
and
redundancy
analysis,
version
2.1.
TNO
Inst.
Appl.
Comp.
Sci.,
Stat.
Dept.
Wageningen, Wageningen,
The
Netherlands.
ter
Braak,
C.J.F.,
1990.
Update
notes:
CANOCO,
version
3.10.
Agri
-
cult.
Math.
Group,
Wageningen.
Wageningen,
The
Netherlands.
Tilman,
D.,
1982.
Resource
Competition
and
Community
Structure.
Princeton
University
Press,
Princeton,
NJ.
Tilman,
D,
Dodd,
M.E.,
Silvertown,
J.,
Poulton,
P.R.,
Johnston,
A.E.,
Crawley,
M.J.,
1994.
The
Park
Grass
Experiment:
Insights
from
the
most
long-term
ecological
study.
In:
Leigh,
R.
A.,
Johnston,
A.
E.
(Eds.),
Long-term
Experiments
in
Agricultural
and
Ecological
Sci-
ences.
Cambridge
University
Press,
Cambridge,
pp.
287-304.
van
der
Maarel,
E.,
Titlyanova,
A.,
1989.
Above
-ground
biomass
relations
in
steppes
under
different
grazing
conditions.
Oikos
56,
364-370.
Vanvik,
V.,
1995.
Mountain
summer
farms
in
Roldal,
western
Nor-
way
-vegetation,
soils
and
ecology.
Master
thesis,
University
of
Bergen,
Norway.
Wells,
T.C.E.,
Sheail,
J.,
Ball,
D.F.,
Ward,
L.K.,
1976.
Ecological
studies
on
the
Porton
Ranges:
Relationships
between
vegetation,
soils
and
land
-use
history.
Journal
of
Ecology
64,
589-626.
Willems,
J.H.,
1990.
Calcareous
grasslands
in
continental
Europe.
In:
Hillier,
S.H.,
Walton,
D.W.H.,
Wells,
D.A.
(Eds.),
Calcareous
Grasslands
-Ecology
and
management.
Bluntisham
Books,
Hun-
tington,
UK,
pp.
3-10.