Presentation by Jaime Díaz Pacheco from Complutense University of Madrid on Esri European User Conference 2011.

1. Exploring the limitations of CORINE Land
Cover for modeling urban land use change.
Jaime Díaz Pacheco
jdiazpac@ghis.ucm.es
Universidad Complutense. Departamento de Geografía Humana.

2. PRESENTATION INDEX
1. Brief description of the whole project where this work is included.
-Objective.
-Framework: Dynamic models of land use change in urban regions.
- Methodology: Building urban CA based model on a GIS platform (ArcGIS 10).
2. Exploring the weakness of CORINE Land Cover for modeling urban land use
change.
-Objective.
- Methodology: Comparison of geodatabases CLC-MLU at different scales.
3. Conclusions.
Jaime Díaz Pacheco
jdiazpac@ghis.ucm.es
Dpto. de Geografía Humana
Universidad Complutense

3. 1. Brief description of the whole project where this work is included. - Objective:
PROJECT : RELATIONSHIP AMONG LAND USE AND URBAN MOBILITY.
MARS CA
Metropolitan Activity Relocation Simulator Cellular Automata
Regional Transport and Regional Land Use
Land Use Dynamic Dynamic Model
Model 25x25 m. cell (Scale)
Municipality (Scale)
TRANSyT - Transport Research Centre - Human Geography Department
UPM UCM
Acknowledgements: Financial support from Ministerio de Ciencia e Innovación (project TRA2008-06682) is gratefully acknowledged.
Jaime Díaz Pacheco
jdiazpac@ghis.ucm.es
Dpto. de Geografía Humana
Universidad Complutense

4. 1. Brief description of the whole project where this work is included. - Framework:
THE URBANIZATION PHENOMENA AND THE URBAN STRUCTURES DEVELOPMENT…
…continues intensively since the first industrial cities bloomed in Europe and U.S.
…It is occurring in every large urban
world region but in a different way
(proportions, rates of growth,
shapes…)
Despite this evolutional diversity, the
large active urban spaces share
common characteristics. A high level
of dynamism and high index of
growth.
Fuente: GawC Inventory of World Cities.
ENDOGENOUS AND
Can we approach the urban phenomenon through the land
EXOGENOUS FACTORS
use dynamics? COMPLEXITY OF THE HUMAN
Are there factors which are in SOCIO-SPATIAL PHENOMENA.
GEOGRAPHICAL MODELS,
relationship with urban
PATTERNS GENERATION
IMPLEMENTED TECHNIQUES
transformations? How are these CITIES ARE A DYNAMIC SELF
IN GIS . CA, ANN…
factors contributing on the land ORGANIZING SYSTEM
Fuente: http://arsberlin.wordpress.com/ use dynamic in the cities? LAND USE DYNAMICS

5. 1. Brief description of the whole project where this work is included. - Methodology:
CELLULAR AUTOMATON ELEMENTS
Jaime Díaz Pacheco
jdiazpac@ghis.ucm.es
Dpto. de Geografía Humana
Universidad Complutense

6. LAND USE CHANGE DYNAMIC MODEL
LAND USE
FUTURE True
Period t0 (2000)
SCENARIOS LAND USE
Period t9
(2009)
Changing rules VALIDATION
LAND USE
Changing demand
CHANGE
BEHAVIOUR Simulation
LAND USE
LAND USE
Period t6 (2006)
Period t9
Where? How much?
(2009)
POTENTIAL LAND USE
TRANSITION DEMAND
(RULES) MODEL
CALIBRATION
Simulation True
LAND USE LAND USE
Period t6 (2006) Period t6 (2006)
MAP COMPARISON
Jaime Díaz Pacheco
jdiazpac@ghis.ucm.es
Dpto. de Geografía Humana
Universidad Complutense

7. POTENTIAL
TRANSITION
(RULES)
Neighborhood
Adapted from: Engelen, G.; White, R.; Uljee, I.; Drazan, P. (1995)

8. 1. Brief description of the whole project where is included this work. - Methodology:
BUILDING AN URBAN CELLULAR AUTOMATON ON MODEL BUILDER (ArcGIS 10)
MODEL BUILDER (ARCGIS 10)
Land Use
LAND USE CHANGE DYNAMIC MODEL BASED ON A CELLULAR AUTOMATON
Potential Transition
Allocation
loop
Ranking cells
Building simulation
Initial Land USe LU Simulation
t0 Demand Model t1

9. 2. Exploring the weakness of CORINE Land Cover for monitoring urban land use
changes.
- Objective:
- Measure to what degree CLC is applicable for an
urban CA-BASED land use modeling at particular
scales.
- Methodology:
- Compute possible underestimations and
overestimations of CLC on artificial (urban) land use, by
comparison with a higher precision regional land use
geodatabase (MLU) at particular scales of analysis.
TOTAL AREA OF According CLC
ARTIFICIAL LAND Regional Scale
(Madrid
2000 & 2006
Community)
LAND USE Comparison
CHANGE TO Local Scale
ARTIFICIAL LAND (municipalities)
According MLU
2000-2006

10. 2. Exploring the weakness of CORINE Land Cover for monitoring urban land use dynamics.
- Methodology: Comparison of geodatabases CLC-MLU at different scales.
DETAIL OF CORINE LAND COVER, 2000. LOCALIZATION CORINE Land Cover.
- EU. EEA.
- Coordinate,
homogenize, LULC
information in EU.
- 1999-2000-2006.
Land Uses - 25 ha. m.m.u.
Urban Residential - Remote
Industrial/Commercial/Services
DETAIL OF MADRID LAND USE, 2000.
Airport
Sensing/Automatic
Infrastructures and visual. (Computer-
Sport and Leisure assisted classification)
Urban Green
Non-Urban Land
MADRID Land Use.
³ - UCM (H. Geography).
- LU and urban mobility
Km
0 50 100 200 (Madrid C.).
- 2000-2006-2009.
- 0.25 ha. m.m.u.
Jaime Díaz Pacheco
jdiazpac@ghis.ucm.es - Photo interpretation,
Dpto. de Geografía Humana
Universidad Complutense
cartography support.

11. 2. Exploring the weakness of CORINE Land Cover for monitoring urban land use dynamics.
- Methodology: Comparison of geodatabases CLC-MLU at different scales.
MADRID Land Use (2006) CORINE Land Cover (2006)
Airports
Sport and Leisure
Industrial/Commercial/Service
Road Network
Urban Residential
Urban Green
³
0 10 20 40
Km
A NEW SET OF CATEGORIES TO ASSIMILATE CLC and MLU CLASSES.

12. COMPARISON OF ARTIFICIAL LAND USE COMPUTED BY CLC AND MLU AT REGIONAL SCALE.
Jaime Díaz Pacheco
jdiazpac@ghis.ucm.es
Dpto. de Geografía Humana
Universidad Complutense

13. 2. Exploring the weakness of CORINE Land Cover for monitoring urban land use dynamics.
- Methodology: Comparison of geodatabases CLC-MLU at different scales.
ARTIFICIAL LAND USE EVOLUTION, 2000-2006. OVERESTIMATIONS AND UNDERSTIMATIONS, CLC.
Jaime Díaz Pacheco
jdiazpac@ghis.ucm.es
Dpto. de Geografía Humana
Universidad Complutense

14. 2. Exploring the weakness of CORINE Land Cover for monitoring urban land use dynamics.
- Methodology: Comparison of geodatabases CLC-MLU at different scales.
Overestimations / Underestimations
Low ( < 5%)
Medium (5% - 25%)
High (25%-50%)
Low (< 5%)
Medium (5%-25%)
High (25%-50%)
³
0 12.5 25 50 Km
Very high (> 50%) Very high (>50%)
Jaime Díaz Pacheco
jdiazpac@ghis.ucm.es
Dpto. de Geografía Humana
Universidad Complutense

15. 2. Exploring the weakness of CORINE Land Cover for monitoring urban land use dynamics.
- Methodology: Comparison of geodatabases CLC-MLU at different scales.
Underestimations Overestimations
n: 128 n: 51
X: 55.25 X: 36.42
s : 83.46 s : 35.79
Max: 495.84 Max: 168.33
Min: 0.18 Min: 2.02
Underestimations Overestimations
Jaime Díaz Pacheco
jdiazpac@ghis.ucm.es
Dpto. de Geografía Humana
Universidad Complutense

16. 2. Exploring the weakness of CORINE Land Cover for monitoring urban land use dynamics.
- Methodology: Comparison of geodatabases CLC-MLU at different scales
EXAMPLES: AREAS CLASSIFIED AS SPORT AND LEISURE CATHEGORY. IT HAVE OTHER USES!
Jaime Díaz Pacheco
jdiazpac@ghis.ucm.es
Dpto. de Geografía Humana
Universidad Complutense

17. EXAMPLES: AREAS MAPPED IN CLC 2000, NOT MAPPED IN 2006
Detail of mapped elements. CLC, 2000 and 2006.
Location of negative
growth areas.
Hectareas
Nega tive G rowth
< 25 0
250 - 500
500 - 1.50 0
1.50 0 - 2.50 0
2.50 0 - 30 .0 00
Location of negative
growth areas.
³
Artificial Land
200 0 Km
200 6
0 50 100 200
Jaime Díaz Pacheco
jdiazpac@ghis.ucm.es
Dpto. de Geografía Humana
Universidad Complutense

18. 3. CONCLUSIONS.
- This work has tried to explore the coherence of CORINE land cover and its capacity to
serve as input land use geodatabase for modeling urban land use change.
- CORINE land cover may be useful to work at small scales (coarse resolution) for
assessing urban land use change, but it is not so suitable to work at scales greater (finer
resolution) than 1:100,000. It is necessary to be aware of this limitation of CLC for urban
modeling.
- Recently in the EU a new LULC geodatabase at larger scale (0.25 ha. m.m.u.) is available
for large cities (urban regions), nonetheless there is still only one time period geotabase
(2010 for Madrid).
- The measurements of artificial land use do not show a high difference in quantity of land
use computed between CLC and MLU for 2000 and 2006 (static method), nonetheless if
the evolution of artificial land uses between 2000-2006 is computed, there is a marked
increase in the difference between CLC and MLU.
- The overestimations and underestimations of CLC over MLU increase as we increase the
scale of analysis (regional-municipal). In other words the error degree tends to increase
with the scale.
- Certainly the land use classes behavior is not the same for every category. There are
extreme cases as the evolution of road network according CLC. In fact if we analyze the
data we find decrease rather than increase for these categories between 2000-2006. This
kind of problem makes results from CLC incoherent.
Jaime Díaz Pacheco
jdiazpac@ghis.ucm.es
Dpto. de Geografía Humana
Universidad Complutense

19. Thank you!

20. Exploring the weakness of CORINE Land
Cover for modeling urban land use change.
Jaime Díaz Pacheco
jdiazpac@ghis.ucm.es
Universidad Complutense. Departamento de Geografía Humana.
Urban land change modeling has become an important tool to support
territorial planning and management in a given region. Land use models
based on cellular automata have been successful because of their
simplicity. A set of simple rules allows to simulate the complexity of the
land use patterns in a given region.
To achieve a successful model based on urban cellular automata, it is
necessary to have a land use geodatabase for several time periods. In
that sense, CORINE land cover (1990, 2000, 2006) a European initiative
for monitoring land use change within EU member states and adjacent
areas, is useful. This paper aims to measure to what degree this land use
geodatabase is applicable for an urban CA-BASED land use modeling at
particular scales.
Jaime Díaz Pacheco
jdiazpac@ghis.ucm.es
Dpto. de Geografía Humana
Universidad Complutense