![FLOODRESILIENCE
Largest cities (2006) ranked by land area
2. URBAN GROWTH FIGURES 0 2000 4000 6000 8000 10000 12000
New York Metro
Tokyo/Yokohama
EXPLORATIONS IN DENSITY: Chicago
Atlanta
Philadelphia
• Large differences between urban area and Boston
Los Angeles
density Dallas/Fort Worth
Houston
SPRAWL
Detroit
Washington
Miami
DEVELOPING COUNTRIES Nagoya
Paris
• 100% growth of urban areas Essen/Düsseldorf
Osaka/Kobe/Kyoto
Seattle
• Annual decline of density of 1.7% (World Johannesburg/East Rand
Minneapolis/St. Paul
Bank, 2005) San Juan
Buenos Aires
• Cities tripled occuplied space Pittsburgh
Moscow
• New inhabitant takes 160m2 (avg)
St. Louis
Melbourne
Tampa//St. Petersburg
Mexico City
Phoenix/Mesa
INDUSTRIALIZED COUNTRIES San Diego
Sao Paulo
Baltimore
• 11% growth of urban areas Cincinnati
Montreal.
• Annual decline of density of 2.2% (World Sydney
Cleveland
Bank, 2005) Toronto
London
Kuala Lumpur
• 2.5x amount of occuplied space Brisbane
Rio de Janeiro
DENSE
• New inhabitant takes 500m2 (avg) Milan
Kansas City
Indianapolis
Manila
San Francisco//Oakland
COMPARE: Virginia Beach
Jakarta
Rotterdam (rank: 101): 2500 ppl/sq Km Providence
Cairo
Mumbai (rank:1): 29650 ppl/sq Km Delhi
Denver
FLOODRESILIENCEGROUP
land area [sqKm] density [people sqKm]
City mayors, 2009
FLOODRESILIENCEGROUP
Page 7](https://image.slidesharecdn.com/lec4urbangrowth022811-100329080630-phpapp01/85/Lecture-on-Urban-Growth-7-320.jpg)
Lecture on Urban Growth
- 1. FLOODRESILIENCE LECTURE 2: URBAN GROWTH William Veerbeek w.veerbeek@floodresiliencegroup.org FLOODRESILIENCEGROUP FLOODRESILIENCEGROUP Page 1
- 2. FLOODRESILIENCE URBAN FLOODING EXPANSION (Asia) VS STASIS (Europe) OECD, 2008 Population exposed to extreme water levels (2005) 30 Ho Chi Min City, 2007 Exposed population 25 20 15 10 5 0 a ia pe ica a a ric ic si As ro er er la Af Eu Am ra Am st N. Au S. Mumbai, 2007 New Orleans, 2005 FLOODRESILIENCEGROUP FLOODRESILIENCEGROUP Page 2
- 3. FLOODRESILIENCE 1. DRIVERS FLOOD VULNERABILITY: HAZARD • Frequency of a flood event • Physicial characteristics of a flood event FLOOD RISK EXPOSURE • Extent of the event • Affected people, assets, items, etc. EXPOSURE CAUSE SENSITIVITY • Consequences of the event • During (coping capacity) and after HAZARD EFFECT (recovery capacity) the event SENSITIVITY FLOODRESILIENCEGROUP Vulnerability Framework FLOODRESILIENCEGROUP Page 3
- 4. FLOODRESILIENCE 1. DRIVERS HOW DOES URBAN DEVELOPMENT AFFECT FLOOD VULNERA- BILITY? HAZARD • Surface runoff (pluvial flooding) • Encroachment (pluvial, fluvial, coastal flooding) VULNERABILITY SUSCEPTIBILITY • Concentration of people, assests EXPOSURE SENSITIVITY CAUSE • Rate of Casualties, injuries, health risks • Damage rate HAZARD • Tangible EFFECT • Intagible CLIMATE • Direct CHANGE • Indirect SENSITIVITY URBAN DEVELOP- MENT FLOODRESILIENCEGROUP Vulnerability Framework FLOODRESILIENCEGROUP Page 4
- 5. FLOODRESILIENCE 2. URBAN GROWTH FIGURES GENERAL FIGURES: • 1800: 3% of the world population lived in cities • 2007: 50% of the world population lived in cities • Different patterns (compare London, Lagos and Tokyo) FLOODRESILIENCEGROUP World bank, 2000 FLOODRESILIENCEGROUP Page 5
- 6. FLOODRESILIENCE Largest cities (2006) ranked by population size 2. URBAN GROWTH FIGURES 0 5 10 15 20 25 30 35 40 Tokyo Mexico City GENERAL FIGURES 2030 (2000): Mumbai (Bombay) New York São Paulo • 4 billion people live in cities (UN, 2004) Delhi Calcutta Jakarta Buenos Aires DEVELOPING COUNTRIES Dhaka Shanghai Los Angeles • 100% growth of urban areas Karachi Lagos • Annual decline of density of 1.7% (World Bank, 2005) Rio de Janeiro Osaka, Kobe • Cities tripled occuplied space Cairo Beijing • New inhabitant takes 160m2 (avg) Moscow Metro Manila Istanbul Paris Seoul INDUSTRIALIZED COUNTRIES Tianjin Chicago • 11% growth of urban areas Lima Bogotá • Annual decline of density of 2.2% (World Bank, 2005) London Tehran Hong Kong • 2.5x amount of occuplied space Chennai (Madras) Bangalore • New inhabitant takes 500m2 (avg) Bangkok Dortmund, Bochum Lahore Hyderabad Wuhan Baghdad Kinshasa Riyadh Santiago Miami Belo Horizonte Philadelphia St Petersburg Ahmadabad Madrid Toronto Ho Chi Minh City 2020 2006 FLOODRESILIENCEGROUP City mayors, 2009 FLOODRESILIENCEGROUP Page 6
- 7. FLOODRESILIENCE Largest cities (2006) ranked by land area 2. URBAN GROWTH FIGURES 0 2000 4000 6000 8000 10000 12000 New York Metro Tokyo/Yokohama EXPLORATIONS IN DENSITY: Chicago Atlanta Philadelphia • Large differences between urban area and Boston Los Angeles density Dallas/Fort Worth Houston SPRAWL Detroit Washington Miami DEVELOPING COUNTRIES Nagoya Paris • 100% growth of urban areas Essen/Düsseldorf Osaka/Kobe/Kyoto Seattle • Annual decline of density of 1.7% (World Johannesburg/East Rand Minneapolis/St. Paul Bank, 2005) San Juan Buenos Aires • Cities tripled occuplied space Pittsburgh Moscow • New inhabitant takes 160m2 (avg) St. Louis Melbourne Tampa//St. Petersburg Mexico City Phoenix/Mesa INDUSTRIALIZED COUNTRIES San Diego Sao Paulo Baltimore • 11% growth of urban areas Cincinnati Montreal. • Annual decline of density of 2.2% (World Sydney Cleveland Bank, 2005) Toronto London Kuala Lumpur • 2.5x amount of occuplied space Brisbane Rio de Janeiro DENSE • New inhabitant takes 500m2 (avg) Milan Kansas City Indianapolis Manila San Francisco//Oakland COMPARE: Virginia Beach Jakarta Rotterdam (rank: 101): 2500 ppl/sq Km Providence Cairo Mumbai (rank:1): 29650 ppl/sq Km Delhi Denver FLOODRESILIENCEGROUP land area [sqKm] density [people sqKm] City mayors, 2009 FLOODRESILIENCEGROUP Page 7
- 8. FLOODRESILIENCE 3. CAUSES OF URBAN GROWTH 1. AUTONOMOUS POPULATION GROWTH 2. RURAL > CITY MIGRATION 3. CITY > CITY MIGRATION Still marginal compared to other factors FLOODRESILIENCEGROUP FLOODRESILIENCEGROUP Page 8
- 9. FLOODRESILIENCE 3. CAUSES OF URBAN GROWTH 1. AUTONOMOUS POPULATION GROWTH Decline in most Western countries (babyboom), growth in Africa and some other countries FLOODRESILIENCEGROUP FLOODRESILIENCEGROUP Page 9
- 10. FLOODRESILIENCE 3. CAUSES OF URBAN GROWTH 2. Rural to Urban Migration: • Economic progress, opportunity • Macro economic factors (industrialization, technological advancements) Rural-Urban Migration in China 1950-2030 Rural-Urban Migration per Region 1950-2030 FLOODRESILIENCEGROUP FLOODRESILIENCEGROUP Page 10
- 11. FLOODRESILIENCE 4. CAUSES OF URBAN GROWTH 3. Economic attraction / Globalization • Intra-urban migration Connectivity of Urban Agglomerations: Assumption: The stronger the connectivity and directionality the stronger the urban de- velopment per capita • Connectivity can be subdivided per industrial sector • Connectivity and sectoral diversitiy tell indicate economic resilience Connectivity B Map of global city-firm networks. 100 200 Amsterdam: 8th, Rotterdam: 68th A 50 450 10 100 C 200 D 50 200 100 10 headquarter subsidiary E city 850 Global dataset = 9243 connections 2/3 of global GDP FLOODRESILIENCEGROUP 500 Firms lead to urban patterns Wall & v.d. Knaap, 2007 Wall & v.d. Knaap, 2007 FLOODRESILIENCEGROUP Page 11
- 12. FLOODRESILIENCE 5. SPATIAL URBAN GROWTH PATTERNS EXPANSION (Asia) VS STASIS (Europe) 1990 Urban expansion GANGZHOU, China 1990-2000 YIYANG, China 1990-2000 HYDERABAD, India 1990-2000 LONDON, UK 1990-2000 FLOODRESILIENCEGROUP World Bank, 2005 FLOODRESILIENCEGROUP Page 12
- 13. FLOODRESILIENCE 5. SPATIAL URBAN GROWTH PATTERNS CAIRO 1984-2000 Population growth: 10.1 million (1984) to 13.1 million (2000) Can this expansion be classified into different types? CAIRO 1984-2000 Cairo 1984 Urban expansion Annual Measure 1984 2000 Population 10.1 million 13.1 million 1.58% Built-Up Area (sq Km) 366.50 369.65 2.77% Average Density (persons /sq Km) 27727 22965 -1.16% Built-Up Area per Person (sq m) 36.07 43.54 1.17% Average Slope of Built-Up Area (%) 4.11 4.03 -0.12% Maximum Slope of Built-Up Area (%) 20.65 20.80 0.04% Buildable Perimeter (%) 0.66 0.67 0.06% Contiguity Index 0.62 0.61 -0.9% Compactness Index 0.22 0.22 0% Per Capita GDP USD 2.413 USD 3.281 1.92% FLOODRESILIENCEGROUP World Bank, 2005 FLOODRESILIENCEGROUP Page 13
- 14. FLOODRESILIENCE 5. SPATIAL URBAN GROWTH PATTERNS 1. Infill: • New development within remaining open spaces in already built-up areas. • Infill generally leads to higher levels of density and increases contiguity of the main urban core. CAIRO 1984-2000 Infill FLOODRESILIENCEGROUP World Bank, 2005 FLOODRESILIENCEGROUP Page 14
- 15. FLOODRESILIENCE 5. SPATIAL URBAN GROWTH PATTERNS 1. Infill CHARACTERISTICS: • Compact city • Small footprint • Relatively modest infrastructural needs • Often only a fraction of total development • Not always controlled development Sao Paolo, Brazil Mumbai, India FLOODRESILIENCEGROUP FLOODRESILIENCEGROUP Page 15
- 16. FLOODRESILIENCE 5. SPATIAL URBAN GROWTH PATTERNS 2. Extenstion: • New non-infill development extending the urban footprint in an outward direction. • Extenstion generally leads to an increased ara of contiguity. CAIRO 1984-2000 Extension FLOODRESILIENCEGROUP World Bank, 2005 FLOODRESILIENCEGROUP Page 16
- 17. FLOODRESILIENCE 5. SPATIAL URBAN GROWTH PATTERNS 2. Extension CHARACTERISTICS: • Often low density, sprawl • Large footprint • Relatively high infrastructural needs • Often majority of total development (together with Leapfrog development) • Not always controlled development El Paso, United States Los Angeles, United States FLOODRESILIENCEGROUP FLOODRESILIENCEGROUP Page 17
- 18. FLOODRESILIENCE 5. SPATIAL URBAN GROWTH PATTERNS 3. Leapfrog development: • New development not intersecting the urban footprint leading to scattered development. • Leapfrog generally leads to an increased level of fragmentation. CAIRO 1984-2000 Extension FLOODRESILIENCEGROUP World Bank, 2005 FLOODRESILIENCEGROUP Page 18
- 19. FLOODRESILIENCE 5. SPATIAL URBAN GROWTH PATTERNS 3. Leapfrog development CHARACTERISTICS: • Often low density, sprawl • Largest footprint (since often indepent from morpholical constrains) • Highest infrastructural needs (far away from centers) • Often majority of total development (together with Leapfrog development) • Often planned new residential areas • (Can become foundation for network cities) Las Vegas, United States Newman & Kenworthy, 1989 Relation between densitity and petrol consumption 80000 Houston 70000 Petroleum use p/a (average per capita) United States 60000 Los Angeles of America Washington 50000 New York 40000 Melbourne Australia and 30000 Toronto Canada Sydney 20000 Paris Europe Vienna London 10000 Far East Singapore Tokyo Hong Kong and Russia Moscow 0 0 150 200 FLOODRESILIENCEGROUP 250 300 50 100 Density (persons per hectare) FLOODRESILIENCEGROUP Page 19
- 20. FLOODRESILIENCE 5. SPATIAL URBAN GROWTH PATTERNS Classification of urban areas • Main Core (Central Business District) • Secondary Core (Neighborhood centers) BUILT-UP AREA • Fringe (Suburbs) • Ribbon (Suburbs along main infrastructure) • Scatter (Secondary towns) 30 TO 50% >50% URBAN URBAN Extension, Leapfrog <30% URBAN Infill, Extension Extension, Leapfrog Leapfrog Infill, Extension LARGEST LINEAR SEMI- CONTIGUOUS ALL OTHER CONTIGUOUS ALL OTHER DEVELOPMENT DEVELOPMENT DEVELOPMENT DEVELOPMENT (100M WIDE) MAIN CORE SECONDARY CORE FRINGE RIBBON SCATTERFLOODRESILIENCEGROUP FLOODRESILIENCEGROUP Page 20
- 21. FLOODRESILIENCE 5. SPATIAL URBAN GROWTH PATTERNS Classification of urban areas • Main Core (Central Business District) • Secondary Core (Neighborhood centers) • Fringe (Suburbs) • Ribbon (Suburbs along main infrastructure) • Scatter (Secondary towns) Example: Chengdu, China, 1991-2002(!) Boston University, 2000 FLOODRESILIENCEGROUP FLOODRESILIENCEGROUP Page 21
- 22. FLOODRESILIENCE 6. CONSEQUENCES Increase of impervious areas > surface runoff • Strong relationship between land-use and level of imperviousness. • Urbanized areas result in large runoff coefficients. LAS VEGAS 2001 Extension FLOODRESILIENCEGROUP Veerbeek, 2008 FLOODRESILIENCEGROUP Page 22
- 23. FLOODRESILIENCE 6. CONSEQUENCES Relating urbanization to imperviousness • Relation is not always straightforward • Local differences resulting from urban typologies Is SEATTLE the GREENEST CITY? PHOENIX 2001 SEATTLE 2001 LAS VEGAS 2001 FLOODRESILIENCEGROUP Veerbeek, 2008 Veerbeek, 2008 Veerbeek, 2008 FLOODRESILIENCEGROUP Page 23
- 24. FLOODRESILIENCE 6. CONSEQUENCES Causes IMPERVIOUSNESS: • Building footprint • Paving private gardens • Roads, parking Unknown Moscow, Russia FLOODRESILIENCEGROUP FLOODRESILIENCEGROUP Page 24
- 25. FLOODRESILIENCE 6. CONSEQUENCES Causes IMPERVIOUSNESS: • Paving private gardens Halton (Leeds suburb) 1971-2004 13% increase of impervious areas 12% increase in runoff 75% due to paving of residential front gardens! Perry & Nawaz, 2008 FLOODRESILIENCEGROUP FLOODRESILIENCEGROUP Page 25
- 26. FLOODRESILIENCE 7. URBAN GROWTH MODELING Quantitative vs Spatial QUANTITATIVE GROWTH MODELING: • Statistical regression and extrapolation to future SPATIAL GROWTH MODELING: Clarke et al, 1997 • Spatial representation of urban growth (past, future) FIRST MODELS BASED ON REGIONAL ECONOMY: • Central place hierarch (Weber, 1909) • Power distribution of settlements (Allen, 1954) • Equlibrium states (Alonso,1964) Theoretical models describing ‘ideal cities’ in equilibrium MODELS HAVE DIFFICULTY DESCRIBING REAL URBAN GROWTH FLOODRESILIENCEGROUP FLOODRESILIENCEGROUP Page 26
- 27. FLOODRESILIENCE 7. URBAN GROWTH MODELING Dynamic urban growth models • Diffuse Limited Aggregation (fractal) • Markov models (conditional probability) • GEOGRAPHIC AUTOMATA CELLULAR AUTOMATA ‘A regular array of identical finite state automata whose next state is determined solely by their current state and the state of their neighbours.’ • Cells • Cell states • Cell space (n-dimensional, n > 0) • Transition rules • Neighborhood 0 1 • Iteration 2 3 • Starting position 4 5 6 7 8 9 10 11 12 13 14 15 FLOODRESILIENCEG FLOODRESILIENCEG R FLOODRESILIENCEGROUP LOO RESILIENCEGRO O ESI ENC GR ENCE 1-d CA with rule 30, Wolfram, 2005 FLOODRESILIENCEGROUP Page 27
- 28. FLOODRESILIENCE 7. URBAN GROWTH MODELING CELLULAR AUTOMATA • Deterministic yet intractable • Capable of simulating complex behavior • Simplicity E.g. GAME OF LIFE (Gardner, 1970) • Remarkably complex behavior generated by 4 simple rules LONELINESS A cell with less than 2 adjoning cells dies OVERCROWDING A cell with less more than 3 adjoning cells dies REPRODUCTION A cell with more than 3 adjoining cells comes alive STASIS A cell with exactly 2 adjoning cells remains the same FLOODRESILIENCEGROUP Game of Life, Gardner, 1970 FLOODRESILIENCEGROUP Page 28
- 29. FLOODRESILIENCE 7. URBAN GROWTH MODELING FROM CELLULAR AUTOMATA to URBAN GROWTH MODELING Geographic automata (Benenson & Torrens, 2004) Berlin actual data Berlin simulated • Cell states > Land cover/use classes • Cell space > Region 1875 • Transition rules > Rules for urban development • Neighborhood > Influence of current urban extent • Iteration > Time • Starting position > Urban extent at some point in time 1920 IS URBAN GROWTH DETERMINED BY UNIVERSAL LAWS? 1945 Maybe, but at least local conditions differ • Extending cell states by properties (GIS Data) Maxe et al, 1998 • Definining more complex transition rules John Holland, 1995: (...)”A city is a pattern in time. No single constituent remains in place.” “The mystery (of urban economical balance) deepens when we observe the kaleidoscopic nature of large cities. Buyers, sellers, administrators, streets, bridges, and buildings are always changing, so that a city’s coherence is FLOODRESILIENCEGROUP somehow imposed on a perpetual flux of people and structures.” FLOODRESILIENCEGROUP Page 29
- 30. FLOODRESILIENCE 7. URBAN GROWTH MODELING WHY COULD THERE BE UNIVERSAL GROWTH LAWS? CITIES SHOW A HIGH LEVEL OF SELF-ORGANISATION • Spontaneous order • robust • adaptive PROPERTIES • organisation based on local interactions (decentralised) • high level of redundancy • system state is emergent Flocking of birds, NASA, 2005 ALLIGNMENT COHESION SEPERATION FLOODRESILIENCEGROUP FLOODRESILIENCEGROUP Page 30
- 31. FLOODRESILIENCE 7. URBAN GROWTH MODELING Clarke et al, 1997 URBAN GROWTH MODELING SLEUTH MODEL SLOPE • GIS information as additional input data • Thus: spatially heterotropic • Influence of transition rules determined by weights • Control over growth rate NASA, 2005 LAND COVER EXCLUSION URBAN Simulation of Washington DC, 2005 TRANSPORTATION What is a good prediction? NEED FOR EVALUATION CRITERIA HILLSHADE FLOODRESILIENCEGROUP FLOODRESILIENCEGROUP Page 31
- 32. FLOODRESILIENCE 7. URBAN GROWTH MODELING EVALUATION CRITERIA COMPARING SIMULATED DATA TO ACTUAL DATA Yang et al, 2008 Shenzhen actual data Shenzhen simulated • X2 Criteria (classification errors) • Fractal dimension (amount of space filled by shape) • Human interpretation ACCURACY CURRENTLY AROUND 80% (X2 Criteria) Parameters • Neighborhood (computational load) • Cell states/properties (complexity) • Global rules • Transition rules (bottom-up vs top-down) FLOODRESILIENCEGROUP FLOODRESILIENCEGROUP Page 32
- 33. FLOODRESILIENCE 7. URBAN GROWTH MODELING STATE-OF-THE-ART 1. Capping growth rate using a Constrained CA • Mixing quantitative growth and spatial growth • Rank list of candidate cells Von Neuman Moore Von Neuman r=2 2. Neighborhood size variation • size • using n-hood hierarchy 3. Regression of transition rules instead of definition • machine learning (e.g. neural network) adjustment transition rules growth model (cells, application of actual data t0 neighborhoods, output evaluation transition rules transition rules) actual data t1 FLOODRESILIENCEGROUP FLOODRESILIENCEGROUP Page 33
- 34. FLOODRESILIENCE 8. URBAN GROWTH MODELING FROM CELLULAR AUTOMATA to URBAN GROWTH MODELING Geographic automata (Benenson & Torrens, 2004) • Cell states > Land cover/use classes • Cell space > Region • Transition rules > Rules for urban development • Neighborhood > Influence of current urban extent • Iteration > Time • Starting position > Urban extent at some point in time IS URBAN GROWTH DETERMINED BY UNIVERSAL LAWS? Maybe, but at least local conditions differ • Extending cell states by properties (GIS Data) • Definining more complex transition rules John Holland, 1995: (...)”A city is a pattern in time. No single constituent remains in place.” “The mystery (of urban economical balance) deepens when we observe the kaleidoscopic nature of large cities. Buyers, sellers, administrators, streets, bridges, and buildings are always changing, so that a city’s coherence is FLOODRESILIENCEGROUP somehow imposed on a perpetual flux of people and structures.” FLOODRESILIENCEGROUP Page 34
- 35. FLOODRESILIENCE 8. CONCLUSIONS URBAN GROWTH IS A MAJOR DRIVER IN FLOOD VULNERABILITY 1. Increased number of people/assets 2. Influence on runoff behavior NOT EVERY TYPE OF URBAN GROWTH IS SIMILAR 1.Infull, extension, leapfrogging 2. Main Core, Secondary Core, Fringe, Ribbon, Scatter SPATIAL URBAN GROWTH MIDELING IS VITAL TOOL 1.Providing insights in future vulnerability 2. Difficult since growth characteristics are locally defined FLOODRESILIENCEGROUP FLOODRESILIENCEGROUP Page 35