modelling urbanization processes in europe · 9 what is luisa? (2) linked to several upstream...
TRANSCRIPT
1st EC-JRC Urbanization Workshop
Modelling urbanization
processes in Europe
Filipe Batista et al.
Ispra, 27th May 2015
EC-JRC-IES, Sustainability Assessment Unit https://ec.europa.eu/jrc/en/luisa/
2
Cities…
Where a plethora of good and bad things happen:
Environmental and social issues: soil imperviousness,
pollution, local climate change, congestion, crime, spread
of diseases...
Socio-economic and environmental advantages due to
scaling: economies of scale (more efficient use of
resources), concentration of critical mass and economic
opportunities, productivity, innovation, economic growth…
Cities…
Bettencourt et al. (2007) PNASOECD (2014)
Regions with big cities experience more economic growth
Higher wages and more
innovation in bigger cities
4
5
Cities in Europe…
Have expanded greatly in the last two centuries due to
economic growth and urbanization.
But what is the future of cities in Europe in a period of
demographic decline? Which cities will win and which will
loose?
Either way, the future of Europe is tied to the future of its
cities…
6
LUISA: a holistic approach for land use and activity modelling
How LUISA works
Baseline trajectory of future urbanization
Discussion
Contents
7
LUISA: a holistic approach for land use and activity modelling
How LUISA works
Baseline trajectory of future urbanization
Discussion
Contents
8
What is LUISA? (1)
Conceived to contribute to Territorial Impact Assessment of EU policies in an
integrated manner;
Co-funded by ENV (initiator in 2007), REGIO, CLIMA;
More than one stand-alone model. It is a platform of inter-linked data,
processes, models and indicators;
Accommodates multi-policy scenarios to represent interacting dimensions of
the EU policies (e.g.: regional and investment policies + CAP + transport +
environment + …), ensuring coherence and consistency of assumptions;
9
What is LUISA? (2)
Linked to several upstream models to capture sector dynamics and policies
(emissions, energy, economy, demography, agriculture/forestry, water, …);
A core, state-of-the-art, ‘spatial allocation module’ simulates future land use
changes given a set of scenario specifications, and resolving competition for
land by the different sectors;
Beyond conventional land use/cover modelling: new approach towards
land use functions and activity-based modelling:
Endogenous dynamic allocation of population (implemented), activities, and services (foreseen);
Endogenous modelling of effects of infrastructure improvements and other spatial planning policies.
10
Macro and exogenous dynamicsDemography, Economy
(define demand for land)
Local conditionsGeography, neighborhood effects, accessibility, soil
characteristics.(define local suitabilities)
Land allocation
Policies with territorial impact
Current Land Uses;Conversion rules,
transition costs, etc..
Eg. Zoning, spatial planning
regulations
Eg. Policies influencing macro dynamics (usually through scenarios)
What is LUISA? (3)
LUISA merges top-down and bottom-up dynamics:
11
Main land use categories in LUISA
Urban areas Industrial/commercial areas
Forested areasAgricultural areas
12
What is LUISA? (4)
Main technical characteristics
Originally based on the CLUE model family (EUClueScanner)
Supported by the GeoDMS Data Model Server (GeoDMS)
Spatial extent: EU-28
Spatial resolution: 100 meters
Thematic resolution: 8 fully simulated land use classes (+ agricultural breakdown +
‘abandoned’ land uses)
Temporal resolution / time span: yearly / up to 2050
Base-map: CORINE Land Cover 2006 ‘refined’
Primary outputs: Maps of land use, population, and accessibility
Secondary outputs: Spatially explicit thematic indicators
13
LUISA: a holistic approach for land use and activity modelling
How LUISA works
Baseline trajectory of future urbanization
Discussion
Contents
Land requirements from different sectors are
estimated given exogenous macro-drivers:
Economy and Demography.
Computes land function indicators.
Resolves the spatial arrangement of land uses at
fine pixel scale, taking into consideration
competition for land, land suitability, and
neighborhood.
14
Demand
module
Allocation
module
Indicator
module
Accessib
ilityNr. of
people
3 tier structure
1521 May 2015
100 m pixels
Biophysical & economic suitability
Regions (NUTS2)
Regional land demands
Allow rules + discrete allocationNeighbourhood
dependencies
How does LUISA work?
1621 May 2015
Allow rules + discrete allocation
AccessibilityNr of people
100 m
Land-use pattern
Outputs from LUISA
at pixel level
1721 May 2015
AccessibilityNr of people
100 m
Land-use pattern
Indicators obtained through combination and/or
aggregation
}50+ indicators
- Land use- Resources- Productivity- Air quality- Ecosystems- Hydrology- Accessibility
100 m
Outputs of LUISA
1921 May 2015
Outputs of LUISA
Population Urban sprawl
Potential accessibility
20
Defined as:
The expected land surface required to support future societal
and economic activities.
In LUISA there demand is specified for the ‘active land use classes’:
Urban (residential, touristic, multifunctional built-up)
ICS (industry + commerce + services built-up)
Agriculture (arable + pasture + permanent crops)
Forest
Determined within LUISA exogenously;
The output of LUISA is very sensitive to the ‘demand’ definition.
The concept of ‘demand’ in the LUISA framework
21
Population
projection
Nr.
households
Urban land
demand
Parameter:Average
household size
*
*
*
* Subject to scenario / policy option specification
Parameter:Urban
density
Other drivers of urban land (e.g. tourism)
Demand for urban land use
22
Definition of urban demand
Urban demand
Residential Touristic
Main driver Nr. of households Nr. of beds
Key parameters
Average household size (persons/household)
assumed to converge to 1.8
Relationship between nr. of arrivals and nr. of beds
statistically derived
Land use intensity (households / ha of urban land)
follows historical trend
Land use intensity(nr. of beds / ha of urban
land use)
Share of touristic buildings nationally
based on survey (avg. 3.5%)
SourcesDemographic projections (ESTAT,
NUTS0)Touristic projections (UNWTO,
Macro-regions)
23
‘Industry-commerce’ comprises :
Industrial areas, commercial areas,
public facilities
Future demand for ICS is calculated according
to an approach developed and described by
Batista e Silva et al. (2014), which relies on a
key parameter:
Land Use Intensity
Defined for each sector and for each
region.
Demand for Industry-commerce land use
24
The GVA per sector is ‘translated’ into demand for industrial, commercial
and services (ICS) land use, by means of an ”Intensity approach”.
Accounts for gains in land use productivity (w), and changes in production
structure.
Variables:
L = Land use intensity
GVA = Gross Value Added
A = Land area
w = Productivity increase
Demand for Industry-commerce land use
Indices:
s = sector
t = time
n = number of years
j = number of sectors
25
Demand for Industry-commerce land use
2621 May 2015
2 step allocation workflow – population and land uses
Regional land use demand
External models:Economy
DemographyAgriculture
(…)
Suitability for
population
Population allocation
Overall suitability for n land uses
Discrete allocation of n land uses
Accessibility Proximity to roads Slope Neighborhood Land uses (t-1)
Transport network investments
Spatial planning policies
Population Neighborhood Allow rules Transition costs Physical suitability (terrain, climate…)
Investments Spatial policies
1
27
Allocation mechanism - POPULATION
Continuous allocation method;
Dynamically allocates inhabitants (given population projections) within
each modelling region (NUTS) with a resolution of 100 m;
Based on spatial econometric methods (see Anselin 2001), with the
factors: Potential accessibility, Proximity to road, Slope, Land uses;
Estimated at country level for t=0, based on high resolution pop map.
Const Accessibility ln(m. to roads)
Urban Industrial Oth. Arable Perm. crops
Pastures Forests λ
Germany -5,904** 0,791** -0,357** 16,619** -5,312** -4,796** -5,755** -3,169 -6,053** 0,303**
Netherlands -7,099** 0,877** -0,909** 15,344** -14,443** -7,877** -4,252 -7,570** -7,605** 0,435**
Portugal -10,591** 2,092** -0,332** 18,345** 1,106 1,986** 3,659** 6,222** 1,960** 0,323**
2821 May 2015
LUISA Population potentialsRed: highestYellow: mediumGreen: lowest
29
Population allocated in year t1 includes population increase + 10%
internal migrants = (Pt1 – Pt0) + (0.1 * Pt0)
Housing supply approximated -> penalizes/incites population changes in
case of shortage/surplus;
People allocated in discrete numbers.
Allocation mechanism - POPULATION
3021 May 2015
2 step allocation workflow – population and land uses
Regional land use demand
External models:Economy
DemographyAgriculture
(…)
Suitability for
population
Population allocation
Overall suitability for n land uses
Discrete allocation of n land uses
Accessibility Proximity to roads Slope Neighborhood Land uses (t-1)
Transport network investments
Spatial planning policies
Population Neighborhood Allow rules Transition costs Physical suitability (terrain, climate…)
Investments Spatial policies
2
31
Allocation mechanism – LAND USE
Discrete allocation method;
Based on multinomial logistic regression – addressing competition for
land
The highest bidder takes land (i.e. the land use type with the highest
overall suitability level at each location).
Overall suitability (for each land use in each location) is a combination of
bio-physical suitability, allow rules, transition costs, neighborhood and
land demand.
32
LUISA: a holistic approach for land use and activity modelling
How LUISA works
Baseline trajectory of future urbanization
Discussion
Contents
A European baseline scenario
A baseline scenario…
assumes the most likely socio-economic trends (often business-as-usual
approach) and the expected effect of established policies;
serves as benchmark for comparison of policy alternatives, allowing their
ex-ante evaluation.
The “Energy-Climate Reference Scenario 2013” is used as BASELINE:
Macro assumptions (demographic and economic);
Cohesion policy (e.g. investments in infrastructure (TEN-T));
Agriculture (CAP);
Biodiversity and habitat protection;
Renewables Directive; (…)
Sector/Theme Model Current implementation
Agriculture CAPRI RES_2_04XX_PRIMESCOR (driven by PRIMES)
Industry GEM-E3 Update 07/2013 (ECFIN Ageing Report 2012)
Cohesion Policy RHOMOLO MFF (Preliminary allocations)
Residential areas EUROPOP EUROPOP 2010
Forestry UNFCC Historical Data
Tourism UNWTO
Transport TEN-T TRANSTOOL Approved network upgrades
Climate Change Various ENSEMBLE (A1B, E1)
Crop suitability BIOMA AVEMAC
Thematic Indicators ESTIMAP, CBM, LISFLOOD/QUAL, .. Ecosystem Services, Forest, Water ..
The ‘reference’ scenario in LUISA
35
On aggregate… further losses of land use intensity
36
1990-2010
2010-2050
But… large spectrum of behaviors among NUTS3 regions…
37
38
39
40
LAND TAKE / YEAR LAND TAKE / CAPITA / YEAR BUILT-UP PER MN INHABITANTS
Focus on Large Urban Zones…
Population density - 2010
Population density - 2050
Expansion
2010 2050
Stagnation / de-growth
2010 2050
Strong concentration
2010 20502050
Absolute changes (population)
47
RC= Vd x Conc x GA
Pollutant concentration maps (current & future) Pollutant removal capacity
Focus on environmental/quality of life indicators (e.g. Air Quality)
48
Urban population exposed to pollutant concentration exceeding the daily
limit value for more than 35 days in a year.
Focus on environmental/quality of life indicators (e.g. Air Quality)
49
LUISA: a holistic approach for land use and activity modelling
How LUISA works
Baseline trajectory of future urbanization
Discussion
Contents
Up-date of reference Scenario (Dec 2015 – depending on upstream models runs)
Pan-European coverage (2015-16 – gradual extension)
Update of base year.
More economic rationale in LUISA:
spatial allocation with net present value considerations
modelling of services and employment distribution
economic-demographic feedbacks
exploit regional opportunities -> attract investments/create jobs
Urban efficiency indicators (land use, and energy/transport efficiency);
Validation activities and cross model comparison exercises.
On-going and planned activities / improvements
Demography Economy
Land demand
Allocation mechanism
Population
LU
IS
AEX
OG
.
LU Accessibility
Internalmigrations
Land use functions
Demography Economy
t+1
t
t+1
Economic gradients influence migration flows between regions.
> Urban patterns influence regional economic growth;
> ESS and natural assets can foster regional opportunities;
> Different regional growths vs. regional Profiles.
Regionalization
Under development
Implemented
52
Lavalle C, et al. 2011. A High Resolution Land use/cover Modelling Framework for
Europe: introducing the EU-ClueScanner100 model. In Computational Science and
Its Applications - ICCSA 2011, Part I, Lecture Notes in Computer Science vol. 6782,
edited by B. Murgante et al. Apduhan. Berlin: Springer-Verlag.
Batista e Silva F, et al. 2014. Estimating demand for industrial and commercial
land use given economic forecasts. PLOS ONE 9 (3):e91991.
Jacobs-Crisioni C, et al. 2014. Accessibility and territorial cohesion in a case of
transport infrastructure improvements with endogenous population
distributions. Forthcoming.
Lavalle C, et al. et al. Land use and scenario modelling for Integrated
Sustainability Assessment. Forthcoming.
More at: http://sa.jrc.ec.europa.eu/?page_id=763
Some references
Thank you