SPATIAL ANALYSIS TOOLS

1
SPATIAL ANALYSIS TOOLS TERRITORIAL COHESION
Claude GRASLAND University Paris 7

• Espon Conference on European Territorial Research
• Luxembourg, 13-14 Oct. 2005

2
INTRODUCTION
3
I. DOES SPACE MATTER ?

• Does space offer an interesting problem to
  society ?
• What is the difference between territorial
  cohesion and economic or social cohesion ?
• How to can we formalize the spatial dimension ?

4
I.1 Is spatial dimension interesting ?

• For J. Levy, Space does not necessary offer an
  interesting problem to societies . We can indeed
  theoretically imagine
• Pre-geographic societies where location are fully
  determined by natural constraint and where
  therefore distance does not matter
• Geographic societies where the cost of relation
  is variable according to distance and where
  spatial organisation does matter.
• Post-geographic societies where generalised
  accessibility between place is achieved and where
  therefore distance does not matter.

5
I.1 Is spatial dimension interesting ?

• TEST Do you consider that the 3 situations
  presented below are equivalent ?

• Answer YES Thank you Mr Sapir
• Answer NO OK but why ? And how do you
  prove it ?

6
I.2 What is territorial cohesion ?
The Hypercube of territorial cohesion (simplified)
7
I.2 What is territorial cohesion ?
Dimension 1 TERRITORY is a combination of
space and society which means that both social
and spatial dimension should be combined when
analysing territorial cohesion. Dimension 2
COHESION can be defined in a structural way as a
level of homogeneity (similarity of social and
spatial units) or in a systemic way as level of
integration (flows and networks). Dimension 3
MULTILEVEL ANALYSIS is necessary in every case
because of scale conflicts (cohesion at one level
can be related to dis-integration at another
one). Dimension 4 DYNAMICS reflects the fact
that cohesion is more a process than a state.
Actual situation are related to past trends
(inheritages) but also to future (anticipations).
8
II. TERRITORIAL ANALYSIS METHODS

• Territorial analysis methods are based on a
  hierarchy of territorial division (NUTS0, NUTS1,
  NUTS2, NUTS3 ) which are considered a priori as
  relevant and should not be modified or removed by
  spatial analysis tools.
• They introduced typically two kind of distances
  which are discrete (discontinuous)
• Territorial Belonging
• Territorial Contiguity

9
Territorial belonging
10
Signification of territorial belonging
Theoretical assumption Regions belonging to the
same unit of upper level are more likely to
interact than regions separated by a border at
upper level Political signification Spatial
planning depends from various levels of political
decision (EU, States, ) which are hierarchically
organised.
11
Territorial neighbourhood
12
Signification of territorial neighbourhood
Theoretical assumption The regions which share a
common border developped specific relations that
are not only related to distance. Political
signification A common border offers opportunity
of interaction which can be encouraged (INTERREG)
or discouraged (Ceuta Melilla).
13
EXAMPLE OF MULTISCALAR TERRITORIAL
ANALYSIStarget variable Unemployement 1999
14
EXAMPLE OF MULTISCALAR TERRITORIAL
ANALYSISGLOBAL DEVIATION 100 EU25
15
EXAMPLE OF MULTISCALAR TERRITORIAL
ANALYSISMEDIUM DEVIATION 100 National Mean
16
EXAMPLE OF MULTISCALAR TERRITORIAL ANALYSISLOCAL
DEVIATION 100 mean of contiguous regions
17
EXAMPLE OF MULTISCALAR TERRITORIAL
ANALYSISMULTISCALAR SYNTESIS High
unemployement (gt 120)
18
EXAMPLE OF MULTISCALAR TERRITORIAL
ANALYSISMULTISCALAR SYNTESIS Low unemployement
(lt 80)
19
III. SPATIAL ANALYSIS METHODS

• Spatial analysis methods are based on a various
  forms of distance (euclidean, cost, time, )
  which are generally quantitative and continuous.
  The official territorial divisions (NUTS) are not
  considered a priori as relevant and can be
  eventually modified or removed.
• They introduced typically two kind of distances
• Euclidean Distance (isotropy, homogenity)
• Network accessibility (discontinuity, anisotropy)

20
Spatial accessibility
21
Spatial accessibility
Theoretical assumption The intensity of
interactions between regions decrease regularly
according to continuous measures of distance.
Political signification Euclidean distance
indicate potential interactions between
territories which could be developped if (1)
borders effects are removed and (2)
transportation system is homogeneized
22
Network accessibility
23
Network accessibility
Theoretical assumption The anisotropy of space
implies that relations are polarised by a limited
number of nodes. Political signification Developme
nt of a polycentric urban and transport system
which limit the concentration of population and
activity around major nodes.
24
EXAMPLE OF MULTISCALAR SPATIAL ANALYSIStarget
variable Peaks of population density
A given location i is characterised by two levels
of neighbourhood V1 and V2 The first
neighbourhood define the local situation (V1),
The second neighbourhood define the global
situation (V2) the neighbourhood V1 is included
in neighbourhood V2 .
25
EXAMPLE OF MULTISCALAR SPATIAL ANALYSIStarget
variable Peaks of population density

• A neighbourhood can be defined as a circle (place
  located at a distance lower than R) but it can
  also be based on various spatial interaction
  function decreasing with distance, like power or
  exponential functions.
• In the present case, we have used gaussian
  functions of neighbourhood based on euclidean
  distance.

26
EXAMPLE OF MULTISCALAR SPATIAL ANALYSIStarget
variable Peaks of population density
A peak of population density appears when the
density is higher in local neighbourhood V1 than
in global neighbourhood V2 In this case, it is
possible to define the spatial concentration as
the quantity of population P which should move
from V1 to V2 in order to obtain an equilibrium
of population density in V1 and V2
27
EXAMPLE OF MULTISCALAR SPATIAL ANALYSIS Peaks of
population for neighbourhoods of 50-100 km
28
EXAMPLE OF MULTISCALAR SPATIAL ANALYSIS Peaks of
population for neighbourhoods of 50-100 km (zoom)
29
EXAMPLE OF MULTISCALAR SPATIAL ANALYSIS Peaks of
population for neighbourhoods of 100-200 km
30
EXAMPLE OF MULTISCALAR SPATIAL ANALYSIS Peaks of
population for neighbourhoods of 100-200 km (zoom)
31
EXAMPLE OF MULTISCALAR SPATIAL ANALYSIS
Delimitation of polycentric area of population
concentration
32
CONCLUSION
33

• THANK YOU
• FOR YOUR ATTENTION !