LANDSAT TM DATA FOR DOCUMENTATION AND SIMULATION OF LANDSCAPE CHANGE 
IN SWEDEN 
ISPRS Commission IV 
K. Hall Kónyves 
A-M. Berggren Bärring 
University of Agriculture, 
Department of Landscape Planning, 
Box 58, 
S-230 53 Alnarp, 
Sweden 
ABSTRACT 
Extensive structural changes are in progress in the Swedish landscape and for landscape planning 
purposes methods for documentation and simulation of landscape change are required. The aim of this 
preliminary work is to develop a model that can be used to simulate different landscape scenarios. This 
model is supposed to assist evaluation of the effects of various alternatives. A framework based on 
digital classification of LANDSAT TM data was developed. The data were manipulated to- simulate 
changes and the hypothetical scenarios were discussed with reference to vegetation, animal and human 
requirements. 
Key words: Landscape change, LANDSAT TM, simulation. 
1. INTRODUCTION 
Governmental proposals and subsequent parliamentary 
resolutions as response to European agricultural policy will 
bring change to the Swedish landscape. During a ten year 
period approximately 2096 of the area under cultivation will 
be excluded from high yielding production to reduce a 
surplus of agricultural products. An environmental goal is 
included in the resolution, where a rich and complex 
agricultural landscape is to be preserved and environmental 
pollution by agricultural production to be minimised. The 
conversion is directed by economic support and economic 
help is provided for farmers who adapt to the new policy. 
We are facing an enormous structural change of the Swedish 
landscape if the goal of 20% less arable land at the year 2000 
is to be reached and it will cause both ecological and visual 
change. This is an opportunity as well as a threat. 
No decisions have been taken to guide farmers as to which 
fields to ‘abandon’ and actions must be preceded by numerous 
considerations. Each area has an economic value for the 
owner, an ecological value for the region and a value as a 
visual landscape. With the specified environmental goal in 
mind each decision must create opportunities for preservation 
or development of a rich and complex environment. 
The chances to create something of considerable value 
depend on the possibilities and abilities to make forecasts and 
also on the possibilities to exert influence on decision- 
makers. The mission for landscape ecologists is to provide 
information about effects and to give recommendations in 
specified regions. 
94 
The aim of this preliminary study is to develop a model that 
can be used to simulate different landscape scenarios. The 
purpose of the model is to assist recommendations 
concerning various landscape-change alternatives. 
2. BACKGROUND 
Several articles have been written about deforestation and 
habitat fragmentation is, according to many authors, the most 
serious threat to biological diversity (Wilcox and Murphy, 
1985). A forest consists of edge and interior core. Species 
have different requirements; some prefer edge conditions, 
some core conditions. 
Deforestation creates new patterns; edge length in an area 
may increase or even decrease but the amount of interior core 
is always reduced (Zipperer, 1989). In order to describe 
patches and to estimate proportions between edge and core 
different methods have been used. A simple area (A) to 
perimeter (P) ratio can describe patch form. The quota 
reveals were the patch lies on a scale were maximum is given 
by calculating the A/P ratio for a disc with the same area. 
Minimum is given by A/(A x 2 + 2), which holds for areas 
without islands. Even fractal dimensions have been used to 
characterise landscape patches. Here the fractal dimension 
could be described as the relationship between a quantity (Q), 
and the length scale (L) over which Q is measured. Hence the 
relationship between area and patch length could be 
described as A=BLPa and the relationship between perimeter 
and patch length as P=tLDp. From this we can derive a 
relationship between area and perimeter as A= (P/t)(Da/Dp) 
(Milne, 1991). 
The . 
differ 
A/P r 
ecolo, 
cluste 
data 
as we 
to res 
What 
to a 
plant: 
locali 
conne 
effec! 
Ways 
netw« 
The 
envir 
docui 
for i 
studi 
1991 
Struc 
imag 
SPO 
illust 
an al 
the e 
aerial 
centr: 
manu 
appli 
quant 
withi 
conti 
planr 
geogi 
for di 
was 
SPO' 
and à 
was 
block 
block 
ecolc 
were 
infra 
infor 
studi 
habit 
data 
habit 
be ac 
The 
Soutl 
altho 
cover