spatial metrics and landsat data for urban landuse change detection in addis ababa, ethiopia

89 Pages·2009·3.44 MB·English

SPATIAL METRICS AND LANDSAT DATA FOR URBAN LANDUSE
CHANGE DETECTION IN ADDIS ABABA, ETHIOPIA

MESFIN TADESSE BEKALO
itle
i
SPATIAL METRICS AND LANDSAT DATA FOR URBAN LANDUSE

CHANGE DETECTION IN ADDIS ABABA, ETHIOPIA

SUPERVISOR:
FILIBERTO PLA (PhD)- Department of Information systems, Universitat Jaume I,
Castellon, Spain

CO-SUPERVISORS:

PEDRO CABRAL (PhD)- Instituto Superior de Estatística e Gestão de Informação
Universidade Nova de Lisboa, portugal
WERNEN KUHN (PhD)- Institute for geoinformatics (ifgi),University of Munster,
Germany
PEDRO LATORRE (PhD)- Department of Information systems, Universitat Jaume I,
Castellon, Spain

March, 2009
ii
SPATIAL METRICS AND LANDSAT DATA FOR URBAN LANDUSE
CHANGE DETECTION IN ADDIS ABABA, ETHIOPIA

MESFIN TADESSE BEKALO
iii
SPATIAL METRICS AND LANDSAT DATA FOR URBAN

LANDUSE CHANGE DETECTION IN ADDIS ABABA,
ETHIOPIA

SUPERVISOR:
FILIBERTO PLA (PhD)- Department of Information systems, Universitat Jaume I,
Castellon, Spain

CO-SUPERVISORS:

PEDRO CABRAL (PhD)- Instituto Superior de Estatística e Gestão de Informação
Universidade Nova de Lisboa, portugal
WERNEN KUHN (PhD)- Institute for geoinformatics (ifgi),University of Munster,
Germany
PEDRO LATORRE (PhD)- Department of Information systems, Universitat Jaume I,
Castellon, Spain

March, 2009
iv
ACKNOWLEDGMENTS

There are many people that deserve heartfelt thanks for their precious contributions to
this study. First, I me thankful for European Commission for prestigious Erasmus
Mundus scholarship for financing the whole phase of the program. I have many
thanks for diligent and dedicated coordinators of the consortium from UJI
(Department of Information Systems), University of Munster (IFGI) and New
University of Lisbon (ISEGI) and all staff members who are the backbone of the
program. I am most grateful to Professor Dr. Filiberto Pla for his tremendous
professional support and moral guidance. I will always remain indebted for his
valuable comments to shape this work.

Thanks are due to Prof. Dr. Pedro Cabral for introducing me to the multiple
application area of GIS and continuous academic support, and comments for this
work. My gratitude also goes to Professor Dr. Pedro Latorre for his suggestions and
his valuable time to evaluate the work and to Professor Dr. Michael Gould, Prof Dr.
Werner Kuhn, Prof. Dr. Hebert Perez-Roses and Prof. Dr. Lubia Vinhas for spending
their valuable time on the evaluation of this study.

I am grateful to my friends and colleagues in the same program: Addis Getnet, Yikalo
Araya for their friendship and constructive suggestions and academic helps. My
friends in Munster, Lisbon and Castellon deserve my sincere gratitude for their deep
and honest friendship that has been connecting us for so many years now. All the
support I received from you is sincerely acknowledged. For her comments on
language I owe Dolores C. Apanewicz a big thanks.

v
SPATIAL METRICS AND LANDSAT DATA FOR URBAN LANDUSE

CHANGE DETECTION IN ADDIS ABABA, ETHIOPIA

ABSTRACT

The rapid development of urbanization coupled with fast demographic change and
high demand for land resource requires landuse information for management and
planning activities of urban regions. The advent of geospatial tools has great potential
to long term monitoring and assessment of urban growth and its associated problems
in surrounding landcover. This study analyzes urban landuse/landcover change of
Addis Ababa, Ethiopia, using landsat TM and ETM+ acquired, respectively, in 1986
and 2000. The landcover maps with four classes were generated using the Maximum
Likelihood Algorithm of Supervised Classification. Overall classification accuracy
was tested by Confusion Metrics and Kappa Coefficient. The landcover dynamics in
pattern and quantities were analyzed using selected Spatial Metrics units. It has been
found that tremendous changes in landcover occurred over the study period. The
results indicated that the built-up area expanded to 49% with annual growth rate
around 3.5% with significant fragmentation and contagion of small and isolated urban
patches. Managers and planners could use this data as a decision-support tool for
urban and environment management.

vi
KEYWORDS

Remote Sensing
Spatial Metrics
Landuse/landcover map
Image classification

vii
ACRONYMS

CSA- Central Statistical Authority.
CORINE- Co-ordination of Information on the Environment
EEA- European Environment Agency
ETM- Enhanced Thematic Mapper
GIS-Geographic Information System(Science)
LC- Landcover
LULCC- Landuse/Landcover Change
LU- Landuse
TM- Thematic Mapper

viii
Contents
Abstract………………………………………………………………………………..ii
List of tables…………………………………………………………………………..vi
List of figures………………………………………………………………………....ix
Acknowledgements…………………………………………………………………....x
CHAPTER 1. INTRODUCTION…...…………………………………………………1
1. Introduction…………………………………………………………………………1
1.1. Study Background…...…..………..………………………...…………………..1
1.1.1. LULCC and geospatial tools……………………………………………2
1.2. Statement of the problem.……………………………………………………...3
1.3. Objective of the study…………………………………………………………...3
1.4. Research Hypothesis and questions…………………………………………....5
1.5. Significance of the study………………………………………………………..6
1.6. Study Area………………………………………………………………...…….6
1.6.1. Physical expansion trend in Addis Ababa……………………………….9
1.6.2. Physiographic nature of the study area…..……………………………..10
1.7 . Organization of the paper.……………………………………………….……11
CHAPTER 2 .REVIEW OF THE LITERATURES...……………………………….13
2. Introduction……………………………………………………………………..…13
2.1: History and application of remote sensing………………….……………...…13
2.2. Landuse and landcover………..………….…….…………..............................17
2.3 Application of Remote Sensing on LULCC……………..…………………….17
2.4. Urban landuse landcover dynamics……...…...……………………………….19
2.5. GIS and remote sensing for urban land change detection and monitoring…...20
2.6. Remote Sensing of Urban areas…...………………………………………….21
2.7. Image classification techniques………………………………………………23
2.7.1. Pixel Based image classification……………………………………....24
2.7.2. Object Oriented image classification…………………………………..25
2.7.3. Advanced classification approaches…...………………………………25
2.8. Change detection and analysis techniques……….………………..………….26
2.8.1: Change detection overview………. …………………………………..26
2.8.2: Applications and approaches of change detection……...……………..26
2.8.3. Landuse change detection techniques…………………………………27
ix
2.8.4.: Landcover detection and analysis techniques……………………… ..28
2.9. Spatial Metrics for urban LULCC analysis…………………………………….28
CHAPTER3. RESEARCH METHODOLOGY ……………….…………………….30
3. Introduction………………………………………………………………………..31
3.1. Data source and type………………………………..……………………….31
3.1.1: data ………………………………………………………………..31
3.2: Landcover Nomencluture……….…………………………………………...34
3.3. Image classification……………...……………….….………………………36
3.4: Training site collection……………...………..……………………………..37
3.5: Urban landcover change analysis techniqies…..…………………………….37
3.6: Accuracy Assessment…………………………..……………………………38
CHAPTER 4. RESULTS AND ANALYSIS ……….………………………………40
4. Introduction……………………………..…………………………………………40
4.1: Landcover maps…………………………………………………………….41
4.2: Change detection and reclassification comparision………………………...42
4.3: Urban LULC analysis using Spatial Metrics……………………………….47
4.3.1: Class area (CA) AND Largest……………………………………...48
4.3.2. Number of patches……………………………………….………….49
4.3.3. Largest Patch Index (LPI)……….…..…….…………………...…...49
4.3.4: Edge Density (ED)………………………………………………….50
4.3.5: Area Weighted Mean Patch Fractal Dimension (AWMPFD)………50
4.3.6: Euclidean Mean Nearest Neighbour (ENN_MN)…………………..51
4.4. Accuracy Assessment using descriptive analysis…………………………..52
4.4.1. Overall accuracy…………………………….……………………...53
4.4.2.: Producer’s accuracy……………………….………………………54
4.4.3: User’s accuracy……………………………….……………………54
4.5. Kappa coefficient…………………… ……….……………………..55
4.6: Demographic, urban expansion and development trends …………..….…56
x