"coefficient of 
50 & 1968) as a 
a whole and for 
:able that the 
nethodology for 
3 also becomes 
însing applica- 
:s 
.n map form is a 
l management of 
¡search on the 
:es. Maps can be 
ation between 
ers. The amount 
in map form is 
projections can 
way. 
. (GIS) is a 
analysing and 
f data that are 
dated, analysed 
ed, either in 
ifferent scales 
provides a base 
ng information 
graphy and GIS 
the data was 
. The technical 
3 lead to an 
analysis, the 
e map, in the 
stored. Both 
drawbacks, see 
;ures of raster 
lie data. 
data 
conveniently 
¡d and area 
i (e.g. over- 
e efficient 
egrated with 
: data in 
mat 
ices (screens 
rs) are 
y used 
uming 
1 and scale 
are time- 
:or and ret- 
Dbjects is 
of different 
iical. Some of 
tore or less 
:r time (e.g. 
lent pattern) . 
lynamic, they 
:tors may be 
sensitive to changes, like agricultural production, 
or more or less independent, like rainfall. The 
static factors of the landscape are important to 
know, they often determine the vulnerability of the 
landscape. A good example is soil erosion, where 
topography, geology and pedology, determines the 
vulnerability to erosion. Dynamic factors like 
vegetation, land use and rainfall intensity are the 
agents causing changes in the landscape, where the 
degree of change is determined by interaction between 
dynamic and static factors. To study this interaction 
of different environmental factors, the geographical 
information system combined with remote sensing is an 
ideal tool. Examples of some different applications 
are described below. 
a) Improve remote sensing analysis. The aim of most 
remote sensing applications is to transform image 
data into thematic information on e.g. vegetation 
(type and quantity), soil and land use. If ancil 
lary data can be included in the analysis of image 
data, the result can be improved significantly. 
Therefore it is important to integrate remote 
sensing with the GIS. 
b) Generation and presentation of spatial infor 
mation. It is important to present scientific 
information in a way that non-specialists can make 
proper use of it. In a GIS there are possibilities 
for different kinds of communications media. The 
most common way is to present spatial data as 
maps. It is then important to supply relevant 
background data for each case, e.g. maps showing 
grazing resources could benefit from additional 
information showing water supply and veterinary 
facilities. An alternative way is to present 
result as tabular data in numerical of graphical 
form. 
c) Integrated research on the interaction between 
climate, environment and land use. Although 
drought, crop failure and famine have been common 
features of arid and semi-arid lands in Africa the 
last decades, our knowledge of causes and con 
sequences is still very limited. The research on 
these complex interactions have to be inter 
disciplinary and must be carried out at different 
scales. In order to integrate information from 
several disciplines and at different scales, and 
to carry out multivariate analysis of the 
relationships between variables, the GIS-approach 
will be an important tool. 
7. GEOGRAPHICAL MODELLING FOR INTEGRATED MONITORING 
Generally speaking, remote sensing methods are 
frequently used as tools for construction of maps 
describing the supply of natural resources, e.g. 
biomass or agricultural yield. The ultimate goal of 
resource monitoring must however be to go one step 
further, to analyse SUPPLY, but also DEMAND and 
ACCESSIBILITY of the resources. 
Data which are not in raster format, e.g. areal 
statistics are difficult to treat in a conventional 
remote sensing system. Different methods of interpo 
lation can be used to transform sampled continuous 
data (in a spatial sense) e.g. topographical or 
meteorological data, into a regular grid net. Data 
that are not spatially continuous, e.g. village 
populations and capacity of wells, can not be treated 
in this way. The data are confined to a certain point 
or valid only for a certain region, but excert an 
influence on the surroundings. Spatial models (Olsson 
1985) can serve as a tool for representing this kind 
of data, and to simulate the spatial influences, in a 
raster format. It may be a way towards the ultimate 
monitoring system, where supply, demand and access 
ibility of natural resources can be analysed in an 
integrated fashion. It can also be seen as a means 
for more efficient use of the ancillary information 
stored in a GIS. I will outline five different kinds 
of spatial modelling attempts that are useful in 
combination with an integrated GIS/remote sensing 
system. 
7.1 Equidistance models 
Many problems in connection with resource utilisation 
are related to a question of distance to different 
landscape features, e.g. water, forest, communication 
links and central places. Examples of application 
could be studies of range resources in relation to 
water supply (Olsson 1984). Different animal species 
have different drinking requirements. 
The equidistance model can be used for generation 
of thematic maps/information layers in a GIS, descri 
bing the distance to the nearest point where a 
certain facility can be found, e.g. well, market or 
road. Two different methods can be applied, depending 
on whether a full-covering map or not is to be 
generated. 
7.2 Potential models 
This is a fairly well known concept in geography, 
what may be new is the application of potential 
models in combination with remote sensing on integ 
rated resource monitoring. The idea behind the poten 
tial model is that the degree of interaction between 
two points, or the influence from a point on its 
surrounding, decreases with distance, in analogy with 
e.g. electrical or gravitational potential in 
physics. 
Different kinds of land use excert different types 
of influence on the surrounding environment. Grazing 
of domestic animals is an example of a land use type 
that is spatially very flexible, while cultivation is 
a type of land use that is spatially much more 
restricted. A potential model can be modified to best 
fit a certain type of activities in two ways. 
1) The number of points (e.g. population centres) 
that are "allowed" to have influence on the same 
piece of land 
2) The distance weight can be varied in order to 
control the balance between proximity of a point 
and the magnitude (e.g. number of inhabitants) of 
it 
7.3 Population density mapping 
Population density is a commonly used parameter when 
dealing with planning and management of resources. 
The conventional measure of population density is 
calculated over administrative, or other regions, 
resulting in very rough figures, insensitive to local 
variations inside the region. An alternative method 
to this, is to use spatial filtering, in analogy with 
filtering of image data, of the population data. 
The population centres, represented as x-, y- and 
z- (=number of inhabitants) coordinates, are trans 
ferred to a regular grid net, corresponding to a map 
projection. Spatial filtering can then be used to 
calculate the number of inhabitants over a certain 
area unit. The sum of all village populations within 
a filter is divided by the filter size, to give the 
inhabitants per area unit. The advantages of this 
type of population density are mainly two: 
1) the population density is measured continuously 
over the area, and sensitive to variations within 
the region; 
2) the concept of inhabitants per area unit is con 
ceptually logical.