Work in Progress towards WUDAPT 1&2 and linked GCP:

Preliminary considerations

GCP-WUDAPT Workshop. Thun Switzerland, June 29-30

Slides courtesy of Chao Ren, Valery Masson, Julia Hildalgo, Xuemei Wang

Jason Ching and Linda See

The World Urban Database and Access Portal Tools (WUDAPT)

The most recent report from the Intergovernmental Panel on Climate Change (IPCC) notes the dearth of information on urban areas. The WUDAPT initiative is designed to fill this gap.

FAST URBANIZATION

http://www.wpclipart.com/geography/Country_Maps/C/China/China_populatio

Overview and Considerations towards GCP-WUDAPT

• WUDAPT –Brief review of progress –Focus on developments towards Level

1 and 2 • Initial exploration and suggestions

towards goal of linkage to Global Carbon Project

Outline of presentation • Background and Context • Design Elements of WUDAPT 1&2 (W12)

– Introducing Building Typology (BT) paradigm – Spatial representation consistent paradigms (LCZ, Blocks and Grids) – Potential approaches Wikis, Satellite (SARs) – Algorithm developments for generating the products (Development of

specialized APPS – Test Beds ( Based on activities of MAPUCE, in CHINA, etc)

• Way forward (Since ICUC9 (Toulouse, Fr) – Planning workshops

• WUDAPT Hong Kong 2015 • GCP Tokyo 2015 • WUDAPT-MAPUCE Toulouse 2016 • GCP-WUDAPT Thun 2016 • GURME Land Cover development workshop- Montreal Fall 2016 • WUDAPT rollout Workshop/Meetings-Hong Kong, Dec 2016 • AMS Annual Meeting w/BUE) Jan 2017, Seattle, WA,USA

Context • WUDAPT LEVEL 0 (W0) provides city maps of Local Climate Zones

(LCZ) and associated ranges of critical parameter values for conducting geo-urban analyses and modeling.

• WUDAPT design requires use of common methodology platform and freely available input information for all cities in the world.

• Current testing and implementation of W0 is proving to be very successful, the quality of the methodology is being assured, and the number of cities being generated exceeds 100 and is increasing rapidly. WRF and Surface Energy Budget modeling now demonstably possible with inputs from WO.

• W12 objective is being devised to (1) augment the products associated with WO and (2) provide spatial accuracy and precision to the parameter products for each and all LCZ classes, with the paradigm requiring common methodology and freely available inputs for its implementation.

• Acquire and make accessible coherent and consistent descriptions and information in the aspects of FORM, Material composition and FUNCTION of cities relevant to climate studies on WORLDWIDE basis.

Objective 1

• Build Portal (tools) that will EXTRACT relevant urban parameters and properties for models and for model APPLICATIONS at appropriate scales for various climate, weather, Urban Planning purposes.

Objective 2

WUDAPT Design Objectives

3-Levels of WUDAPT Product Development

Level 0 • Cities are mapped using the Local Climate Zone (LCZ)

scheme (Stewart & Oke, 2012). Each LCZ type is described in terms of the typical appearance of each in ground-based and aerial photographs and is linked to some urban parametersparameter values.

Level 1 • The LCZ maps are used to sample urban

landscapes to provide more information on the aspects of form and function in greater details.

Level 2 • This is the highest level and it refers to urban

data gathered at a specified spatial scale (e.g. 250 m) across the entire urban area (‘wall-to-wall’ coverage).

(Ching, et al., 2016)

Local Climate ZoneParadigm

LCZ 1: Compact high-rise LCZ 3: Compact low-rise

10

Example: CLASSIFYING LCZs

Sendai, JAPAN

Visual Clues Few if any trees

Little or no green space Tightly packed buildings

10+ stories tall

Diurnal temperature range: small medium large

Medellin, COLOMBIA

Visual Clues Few if any trees

Little or no green space Tightly packed buildings

1 – 3 stories tall

Diurnal temperature range: small medium large

Source: Iain Stewart.

Beijing Chicago

Guangzhou Sao Paolo

Kolkata

LCZ Classes

Rich diversity and complex distributions of climate zones observed within each city LCZ signature palate different and unique to each city Distributions of “Form and Function” modeling parameters produce unique climate and meteorological responses to each urban area

LCZ Classes- Each urban area is unique

16 cities WUDAPT Level “0” generated at 2014 Workshop in Dublin Ireland

WUDAPT: Major Cities in China- Progress by Ren and Ng team

Beijing Shanghai Wuhan Guangzhou Changsha Dalian Guangzhou Hangzhou Jinan Nanjing Qingdao Shanghai Shenyang Tianjing Xi’an Xiamen ………

LCZ Pearl River Delta Megacity study Climate Zone Map Landsat Entity ID date

GPRD

LC81210442013278LGN00 5-Oct-13

LC81220452013365LGN00 31-Dec-13

LC81210452015220LGN00 8-Aug-15

LC81220442015291LGN00 18-Oct-2015

Locatio

n in China

Area Climatic Region Terrain Annual rainfall

Ave. temp

Extreme high temp

Ave. high temp

GPRD Northeast

41,600 km²

Tropical monsoon climate

plain 1,801.4mm

22.4℃ 39.1℃ 26.7℃

1993

2002 (USGS Landuse Database)

WUDPAT: Yangtze River Delta Region LCZ Map

Location in China Area Climatic Region Terrain Annual

rainfall Ave. temp

YRD Region East coast of China 100,000km² Subtropical monsoon

climate

Most of the elevation is lower

than 10m with some hilly area

1000～1400mm 15～16℃

http://www.mdpi.com/2071-1050/3/7/1074 http://www.china.org.cn/environment/report_review/2008-10/17/content_16625642.htm

LCZ Bei Chi Col Dub Kol Kua Mil Sao Van

Compact high-rise 18.2 7.4 1.5 2.3 4.5 8.2 0.0 9.0 3.6 Compact mid-rise 5.7 2.4 28.7 8.5 14.1 2.2 20.2 1.2 0.8 Compact low-rise 2.8 3.9 13.3 3.6 14.6 18.6 0.2 11.3 9.5

Open high-rise 17.9 6.0 6.2 0.1 7.9 15.7 5.6 6.0 10.4 Open mid-rise 14.4 3.5 9.8 5.3 7.6 10.0 18.8 4.3 5.9 Open low-rise 12.4 30.9 28.1 31.5 12.4 14.4 13.2 25.3 22.2

Lightweight low-rise 6.0 0.0 0.9 0.0 0.6 0.6 0.0 4.3 0.0 Large low-rise 14.9 13.0 11.5 44.7 9.2 10.6 19.9 18.8 14.8 Sparsely built 4.1 19.7 0.0 0.0 29.1 13.7 22.1 16.7 32.8

Heavy industry 3.8 13.3 0.0 4.0 0.0 6.0 0.0 3.1 0.0 Kappa 0.90 0.91 0.64 0.82 0.62 0.73 0.84 0.82 0.89

Total area 3406 3479 338 2396 622 1406 1630 4141 1408

A preliminary comparison of the LCZ make-up of 12 cities (Bei – Beijing, Chin; Chi – Chicago, US; Col – Colombo, Sri Lanka; Dub – Dublin, Ireland; Kol – Kolkata, India; Kua – Kuala Lumpur, Malaysia; Sao – Sao Paolo, Brazil and; Van – Vancouver, Canada. The kappa value is a measure of accuracy and Total Area is expressed in terms of satellite cell number (each cell is 120m on a side).

Each city displays a different (UNIQUE) characteristic distribution of LCZ classes

LCZ Level 1 data experiment Stewart & Oke 2012

λV λb λI λV λb λI

Compact high-rise 10.5 42.4 47.1 <10 40-60 40-60

Compact mid-rise 11.3 43.9 43.7 <20 40-70 30-50

Compact low-rise 17.6 36 45.1 <30 40-70 20-50

Open high-rise 25.9 24.3 48.9 30-40 20-40 30-40

Open mid-rise 39.1 19.8 36.8 20-40 20-40 30-50

Open low-rise 39.4 22.2 38.1 30-60 20-40 20-50

Sparsely built 62.3 11.5 24.9 60-80 10-20 <20

The plan fraction (%) of vegetation (λV), buildings (λb) and impervious (λI ) surface for the Dublin urban area based on a Geowiki application compared to the ranges from Stewart and Oke (2012). Results are reasonable.

Range of values for various UCPs for each LCZ class compare. Dublin example: Level 0 (LCZ) vs preliminary version of W1 (Geowiki)

Other comparisons WUDAPT Table Look-up (Range in red) vs Guangzhou using Satellite based methodology (Blue dots)

Guidance from preliminary studies • Range of urban parameter values associated with LCZ classes

are reasonable, however, indications of spatial variability of such parameters for each LCZ are noted throughout each city. Thus:

• Regarding canopy geometry parameters, goal of WUDAPT Level 1 and 2 are to add precision and allocate values of each parameter spatially across each and all the LCZ classes throughout each and all cities.

• Augmentation to include details of material content and exterior of buildings, the usage in different energy sectors, for each and all LCZs and/or other spatially allocated parameters.

• Building height delimiters for tall building classes need to be reviewed and revised to address more realistic values of tall buildings, especially noted in rapidly developing cities such as in Hong Kong and elsewhere.

Preliminary Recommendations under consideration for Level 1&2

– Adopt Building Typology paradigm – Aggregate building stocks into LCZ (and other spatial

polygons) using their footprints. – Incorporate freely worldwide coverage of satellite SAR

& Stereo inputs to generate the 3-D building information

– Consider city block typology – Scale considerations model grid-scale – Principle of Common Methodology applicable to all

cities is required.

Considerations towards expanding the Scope to link with Carbon and Energy

• Strong synergies – Energy and GHG Emissions related – GC map can be extended in stages,

• W0 (LCZ) • W12 Building Typology provides common platform for Building

sector inputs • Transportation sector needs thought

• Features to consider – Spatial downscaling Carbon maps to LCZ – Introduce the elements of Satellite and BTs into the

linkage

Path Forward Method Developments

• Develop W12 Methodology • Carbon Mapping Downscaling with W0 • Fall Workshop with GURME/WMO Montreal • Rollout of WUDAPT Level 0; December WUDAPT

Workshop Meetings –Hong Kong

http://www.ced.berkeley.edu/research/metropolitanlandscapes/?page_id=150

Design Bases of WUDAPT Level 1 and 2:

From LCZ Classes to Urban Forms &

Building Types

WUDAPT Level 1 and 2 (W12): Consideration of various methodologies per guidance deliberated

at June 2016 Workshop in Toulouse, France Objective of W12: Develop methodology and paradigm augmenting W(0) on Form, Materials and Function (FMF) details

Approaches: (1) Collecting relevant Building Typology data as means towards deriving representative FMF

Outputs (2) - Incorporate Building Typology paradigm as powerful and convenient means to satisfy

requirements for form, material and function requirements of Level 1&2 - Using Geowiki and crowdsourcing methodologies that captures the relevant FMF

information to LCZ, Blocks and Grids - Information gathering to include critical matrix of questions, - Learning from TABULA for building energy use - Learning from and using MAPUCE to inform the methods developments and providing a

baseline testbed

(3) Consideration for Incorporating and utilizing advance satellite information such as from SAR as they are becoming free and available. Allows for the creation of DEMS and derivable form parameters

(4) Consider the added dimension of socio-economic aspect in database.

W12 CONCEPTUAL DESIGN GUIDANCE • OBJECTIVE: To determine the most representative

indicator value of Form, Material and Function (FMFs) parameters or variable to each LCZ class, Block and Grid (LBGs over a city.

• GOAL: To develop a consistent, reliable and common methodology that most likely represents these FMFs for each LBGs with greatest accuracy and precision.

• APPROACH: Incorporate Building Typologies, freely available data sources and Crowdsourcing Tools.

Approach: Initial suggestions • Suggestion to list target FMFs products

represented at LCZ, Block and grids levels • Incorporate the Building Typologies into

aggregation methodology to LCZ, Blocks and Grids displays

• Step1: Inputs data obtained based on crowd sourcing Apps (geowikis) and Advanced satellite data such as SARs

• Step 2: Develop algorithms (converting data to desired set of FMFs (key testbed cities as baseline)

Representing Urban areas into its Building Typological morphological forms: Example (LSE-Cities, 2014): Focus on Energy use

Representing various LCZ into its morphological BT counterpart Part 1: Aerial view Example from LSE-Cities/Rode et al

Representing various LCZ into its morphological BT counterpart Part 2: Building footprints Example from LSE-Cities/Rode et al

London

Paris

Berlin

Istanbul

(Rode et al., 2014)

Examples: Categories of morphology samples for heat energy demand modeling

Real Comparable Idealized

London

Paris

Berlin

Istanbul

Detached housing

Detached housing

Detached housing

Detached housing

High Rise

High Rise

High Rise

Apartments Slab housing

Slab housing

Slab housing

Gecekondu

Compact urban block

Compact urban block

Compact urban block

Compact urban block

Terraced

Regular block

Row housing

Modern apartment

Represents wide and diverse range of building densities, heights, packing density, sfc/volume, open space ratios, as well as cultural and age differences. Heat-Energy modeling shows factor of 6 variations due to morphology; compact urban blocks best, detached housing worst

Cities and Energy, Rode et al, 2014

W12 Methodology • Data Collection Options:

– For each LCZs derived from WUDAPT ”0” maps, City blocks and Grids – Crowd-sourcing Geowiki (Geopedia, GeoODK, WUDAPT Collect..) – Individual building and morphological sampling – Incorporate available SAR satellite – Mobilizing collaborating teams of local expert

• Processing APPS – Create and apply algorithms that generates parameters from data collected

using geowiki tools. Outputs checked and certified by local experts – Geo Referenced (WGS84)

• To LCZ • To Blocks • To structured grids

• Generate the Database and Application portals

Data processing Apps to produce sets of model specific high priority UCPs

• H (Building height) height areal distribution, variability)

• Plan Area Density λp (Packing density) – Alt: Building coverage(footprint area)

ratio – Alt: Open space ratio – Alt Building density

• Frontal Area Density • H/W or Aspect ratio • SVF or Sky View Factors • FS (Floor space ratio) • SV (Surface to volume ratio) • Zo (Roughness length) • D (Displacement height) • …………. Others, TBD) •

• Building Use and %, age • Wall frame/exterior,

construction • Roof type, shape composition,

insulation • Window coverage, of

residential, commercial, industrial building typologies

• Brightness, Color • Street orientation (patterns) • Energy usage (central, window) • Parking, cars/household • Building height

Towards WUDAPT Level 1& 2 Products

New Technology for 3-D Urban Morphology (Form) Extraction

Based on Multi-source Satellite Imagery Initial Exploratory Investigation by collaborators Ren and Ng

Applications of Extracted 3-D Urban Morphology Data

Level 1& 2 Products

• WUDAPT level 1 & 2 applications

• Urban climatic analysis and Weather/Climatic Modelling, like WRF;

• Energy Balance study; • Climatic-Spatial

Analysis and Planning, like Urban Climatic Map;

• Early warning of high-risk urban areas with adverse urban climate issues.

Xu, Y., et al., A Satellite-based Analysis of Urban Morphological Parameters and Their Impact on Urban Climatic Variation in High-density Urban Areas Landscape and Urban Planning, 2016. (under review)

Stereo images

SAR images

Procedure of the proposed approach

Stage 1

Adopting Multi-source Satellite images

Urban climate studies using satellite data (Overall Accuracy : ~80%)

Xu, Y., et al., A Satellite-based Analysis of Urban Morphological Parameters and Their Impact on Urban Climatic Variation in High-density Urban Areas Landscape and Urban Planning, 2016. (under review).

Level 1& 2 Products

NEIGHBOURHOOD The city’s urban morphology & roughness

≈ 2 KM x 2 KM

Ways to obtain Levels 1 and 2 data

• Urban databases • Remote sensing • Crowdsourcing of urban experts

– Urban climatologists – Urban planners – Architects – Students (university / GLOBE) – Citizens

Approaches from Toulouse • Expertsourcing with architects:

– Focus on identifying building typologies and transform to parameters needed

• Expertsourcing with urban experts and students – Simple ground-based questionnaire app – Online app

• Crowdsourcing via a map-based app • Crowdsourcing via a game

Approach 1: Building Typologies • Ask architects to fill in a matrix of building types at a

city level

• Need to add location to this – Sample across an LCZ

• Retrieve geotagged photographs • Use a simple app to classify photographs

Type Building Use Date of construction

Houses

Buildings

High rise buildings

Industrial and commercial buildings

Building types

App to Gather Information from Photos

Classify the Picture by Building Material

Tile

Shingle

Thatch

Roof

Building Brick

Wood

Glass

Modified App for Architects

Classify the Picture by Building Type

Type 4

Type 5

Type 6

Type 1

Type 2

Type 3

Approach 2: Expert sourcing with Urban Experts + Students

• Data collection on the ground using the LCZ as a sampling framework – Geotagged photograph – Questionnaire app – Random selection of small number of questions

from a pool of questions

• Data collection from an online application

WUDAPT Collect

WUDAPT Collect

Online Collection of Level 1 Data

Collecting Data on Individual Buildings

Near Term Developments for W12

• Devise and test option using Wikis and available data sets such as Landscan, Streetview, populations incorporating Building Typologies to generate FMF for LCZ, block and grid allocations.

• Further explore utility of SAR data for generating form (building geometry) for LCZ, block and grid allocations.

• Test options against baselines ( MAPUCE, Hong Kong and other comprehensive “standards”

• Test deployment implemetation strategies.

Considering Goals of linking WUDAPT and GCP programs.

• Buildings constitute significantly to global energy demands (40% in developed world, 20% elsewhere)

• Globally: (2004): 8.6 gtons CO2 = 33% energy related GHG emission, 17.6% all GHG

• Specific cities: – London, 70% of total energy demands, same GHG – NYC, somewhat less than London

• Intervention potentials (for sustainability) – Behavioral adjustment, – Technology advancements,

• Design and – Urban geometry (Canopy Ventilation)

Regarding Total CO2 Annual emissions for Indianapolis, (Gurney et al., ES&T 2012)

WUDAPT-GCP Considerations • Link Global Carbon Maps with LCZ maps as framework

(aggregating sectorized energy emission to LCZ/Blocks/grids)

• Synergistic developments using building typology and energy approach of WUDAPT (1&2)

• Introduce, add Socio- Economic layers • Test and revise prototypes based on one or more target

of opportunity cities: – European city (Dublin, Toulouse, Hamburg, Madrid…) – USA city (Indianapolis, Phoenix,…, – Chinese city (Hong Kong, Guangzhou, Shanghai….., – Indian City (Kolkuta,….

Thanks!