Multidimensional Relation of Urban Dwellers and Green Spaces

Gyula Kothencz

Interfaculty Department of Geoinformatics - Z_GIS, Paris Lodron University Salzburg,

Schillerstraße 30, 5020, Salzburg, Austria

Keywords: Geographic Information Systems, Subjective and Spatial Attributes of Green Spaces, Perceived and Actual

Services of Green Spaces, Hot Spots of Spatial Inequalities in Green Space Availability.

Abstract: Green spaces have a positive impact on the daily life of urban communities; however numerous aspects of

the relation between humans and urban green spaces are not widely understood. This PhD thesis is

concerned with this ambiguous relationship and focuses on three of its aspects. First, the thesis explores the

relationship between humans’ subjective evaluations on urban green spaces and objective, spatially explicit

indicators of the same green spaces. Secondly, the work complements green space users’ perception of

services supplied by green spaces, with crowd-sourced data on actual community usage. Finally, the spatial

relation between the geographic distribution of demographic groups of urban population with high green

space demand and the distribution of urban biomass with the highest societal use value is identified. The

expected outcome of the thesis is an improved knowledge on the multidimensional relation between urban

dwellers and green spaces.

1 INTRODUCTION

Urban green spaces (UGS) are essential contributors

to city dwellers’ quality of life (QoL). Their positive

impact on visitors’ physical and mental health is a

well-known fact (Irvine et al., 2013; Lee and

Maheswaran, 2011). Urban green spaces stimulate

social cohesion and provide a place for recreational

activities (Germann-Chiari and Seeland, 2004;

Kaźmierczak, 2013). Other essential services

supplied by UGS are air purification (Tallis et al.,

2011) and microclimate regulation (Bowler et al.,

2010). Shelter and habitat services that they provide

for biodiversity indirectly contribute to QoL

(Fontana et al., 2011). Unambiguously UGS have a

positive impact on the daily life of urban

communities; however numerous aspects of the

relation between humans and urban green spaces are

not widely understood.

This PhD thesis is concerned with the

aforementioned ambiguous relationship of humans

and urban green spaces and focuses on three of its

aspects. The associated research is tackled by the

author of this paper in an article based PhD thesis.

The results will be reported in three ISI ranked

journal articles with first authorship. Each article is

dedicated to one of the three studied aspects of the

human–urban green space relationship and provides

solution for each research problem.

Beyond contributing only to science, another

main concern of the thesis is to strongly support the

work of urban planners and managers with practical

solutions for informed decisions regarding UGS and

QoL. To ensure the fulfilment of this goal, prior to

starting the work, expert interviews were conducted

with urban developers, architects, and urban

environment managers in the City of Szeged,

Hungary, to learn their needs and demands on

perceptual information of urban green spaces. This

initial step eventuated in an active working

collaboration and dialogue with the Szeged

Architect Office and the Szeged Environment

Management Non-profit Ltd. This cooperation

ensures continuous feedback and provides

suggestions from the target user groups to secure the

practical relevance of the research.

2 RESEARCH PROBLEMS

The fragmented knowledge on the relation between

humans and UGS is a multicomponent phenomenon

of which three aspects will be investigated within

this PhD thesis. Accordingly to the aforementioned

Kothencz, G.

Multidimensional Relation of Urban Dwellers and Green Spaces.

In Doctoral Consortium (DCSMARTGREENS 2016), pages 28-34

ambiguities, this thesis is concerned to fill the

subsequent three research gaps.

2.1 The Relation between Human

Perception of Urban Green Spaces

and Their Objective Attributes

The first investigated component is arising from the

human perception of the ambient environment which

differs by individuals. In contrary, objective

environmental indicators foster a relatively

straightforward characterisation of the studied areas.

Therefore the two inherently different data domains

may depict different information on the ambient

environment without featuring any knowledge on

the encompassed discrepancies. An extensive

literature research for evidences of the investigation

of the relation between subjective and spatially

describable objective attributes of UGS resulted in

the definition of the first research problem.

Accordingly, the first research gap which this thesis

is going to fill is articulated as follows. The degree

of relationship between visitors’ subjective

evaluations on urban green spaces and objective,

spatially explicit indicators of the same public green

spaces is unknown.

2.2 Information Gap on Actual Green

Space Use

The second concern of the thesis is described as

follows. The community use and appreciation of

services supplied by UGS can incorporate essential

information to urban planning and development,

although it is not readily measurable. Therefore

current assessments of services derived from UGS

often solely rely on subjective data input, such as

societal benefits perceived and reported by visitors

of UGS. The application of this purely subjective

information for planning and management of UGS

encumbers the objective decision support.

Accordingly the second research gap is identified

and explained here. Easily quantifiable data on real

usage and appreciation of services provided by UGS

is not present to decision support processes to

complement perceptual information, such as data

obtained from questionnaire surveys.

2.3 Relation between the Spatial

Distribution of Urban Biomass and

Demographic Groups with High

Green Space Demand

The third studied aspect of the ambiguous human-

UGS relationship is that the distribution of biomass

heterogeneity varies in the urban space as well as

does the distribution of various demographic groups

of urban communities. This eventuates in the third

research gap that this thesis is going to fill: The

degree of relation between the spatial distribution of

urban biomass and spatial locality of those members

of urban communities who most desperately need

easily accessible green space is unknown.

3 OUTLINE OF OBJECTIVES

Concerning the interests explained in Section 2, this

PhD thesis has three main objectives.

3.1 Objective 1

According to Research gap 1, the objective of the

first tier of the thesis is to investigate the degree of

relationship between human perceptions of urban

green spaces and spatial environmental indicators of

these studied green spaces by matching perceived

and reported features of the parks to their spatial

environmental indicators.

3.2 Objective 2

To fill the second research gap the work outlines the

following aim as the second objective of the thesis.

The research aims to complement subjective

information, explicitly green space users’ perception

of services supplied by UGS, with crowd-sourced

data on actual community usage and appreciation of

UGS.

3.3 Objective 3

The third objective of the work is explained here.

The thesis aims to reveal the spatial relation between

the geographic distribution of demographic groups

of urban population with high green space demand

and the distribution of urban biomass with the

highest societal use value to reveal inequalities in

urban green space availability.

4 STATE OF THE ART

To achieve the objectives described in Section 3 the

research introduces the following novelties.

Multidimensional Relation of Urban Dwellers and Green Spaces

4.1 Novelty 1

The thesis delivers the first study that explores the

degree of relationship between perceived, subjective,

properties of UGS and spatially explicit, objective

attributes of the same green spaces.

4.2 Novelty 2

The novelty of the second tier is twofold and

explained here.

The research introduces a methodology that uses

easily quantifiable crowd-sourced, voluntary

information on actual community consumption of

recreational and aesthetic services of UGS. This will

complement questionnaire records on park visitors’

perceptions of services supplied by the study areas.

4.3 Novelty 3

The state of the art of Tier 3 is threefold.

 Spatial location of demographic groups with high

green space demand will be identified.

 The work will apply digital surface model

(DSM) to identify vegetation with the highest

societal usability value.

 Distances between demographic groups with

high green space demand and high societal value

urban vegetation will be identified on the road

network, which allows calculating real distances

instead of following recent practices which use

Euclidean distance to assess green space

availability.

5 METHODOLOGY

This PhD research applies a multi-method approach

to study the composite relationship between urban

green spaces and humans. Firstly, this section

introduces the study areas, then it explains the

applied methodologies that are conducted to solve

the three research problems.

5.1 Study Areas

The research was conducted in the city of Szeged,

Hungary. Arching across the River Tisza, Szeged is

situated in south-east Hungary and serves home for

160,000 urban dwellers. Numerous green spaces of

the city contribute to the improvement of the

citizens’ QoL. The diversity of the studied parks was

a main concern of the research to ensure the

objective and subjective environmental

heterogeneity. Five green spaces have been chosen

accordingly for study areas of the thesis.

Erzsébet liget (ER) located in the vicinity of the

city centre is a relatively large, 21 ha, area with a

high proportion of vegetated surfaces including

mainly lawn and wooded areas. The park is visited

by recreational users of all age groups, from all over

the city. The recently renovated Dugonics tér (DU)

is situated in the city centre, and nearly half of its

size is vegetated. Széchenyi tér (SZ) is the attractive

main square of the city. More than half its size is

covered with lawn, flower beds, and a number of

trees. Both DU and SZ are located in

neighbourhoods consisting of pleasant late-19

and

early-20

century architecture. Whilst SZ mostly

serves as refuge and aesthetic enjoyment for locals

and tourists, DU functions as an important

pedestrian transit area. At the edge of the city, Vér tó

(VE) boasts a large lake and houses a hill at its

eastern edge. VE is surrounded by 5-10 storey

residential housing and suffers from serious noise

pollution due to nearby traffic. Similar to VE, Zápor

tó (ZA) is surrounded by residential blocks and has a

pond located in its centre. An important aspect in

which ZA is rather different from VE is that its quiet

environment serves as a tranquil resting area for its

visitors. Figure 1 pictures the location of the study

areas within Szeged.

Figure 1: Study areas.

5.2 Methodology 1

As described in Section 3.1, this study seeks for the

relationship between human perception of urban

green spaces and their spatially explicit

environmental indicators. To achieve this aim the

following study was conducted.

DCSMARTGREENS 2016 - Doctoral Consortium on Smart Cities and Green ICT Systems

5.2.1 Used Datasets

Perceptual (subjective, qualitative) data:

 Records of a questionnaire survey campaign in

the study areas.

Spatial (objective, quantitative) data:

 Raster data: 2 m ground resolution Multispectral

bands of Pléiades satellite imagery of Szeged

(Astrium et al., 2014) to derive Normalized

Difference Vegetation Index (NDVI); 10 cm

ground resolution colour aerial image of Szeged

(DPGG, 2011) to delineate vegetated areas and

water surfaces of the green spaces.

 Vector data: Building height layer of Szeged

(Sümeghy et al., 2011); Cadastral layer of

Szeged.

5.2.2 Description of Methodology

A questionnaire survey campaign was conducted in

the five UGS of Szeged, in April and May 2014, to

explore visitors’ subjective evaluations on objective,

spatially describable attributes of the studied areas.

The topics of the survey were the “Personal

judgement of the area”; the “Perceived greenness of

the UGS”; the “Accessibility”; and finally the

“Functions of the area”. Two hundred and fifty five

questionnaires were collected. After validation 227

answers supplied the subjective, perceptual data for

the research.

Area of the park grounds, water surfaces and

vegetated areas of the green spaces were digitised

from the aerial image to vector layers. An NDVI

image was generated form the multispectral bands of

the Pléiades satellite scene.

Applying geographical information systems

(GIS) and remote sensing technology (RS) the

following spatial indicators were calculated for each

green space based on the available objective data:

Area of the parks (ha); Area of vegetated surfaces

(ha); Percentage of vegetated surfaces (%); Area

weighted NDVI for the parks; Percentage of water

surfaces (%); Number of building units in a 50 m

buffer zone around the parks; Percentage of built up

area in a 50 m buffer zone around the parks (%);

Average building height in a 50 m buffer zone

around the parks; Standard deviation of building

heights in a 50 m buffer zone around the parks.

During a principal component analysis (PCA),

performed on the questionnaire data, three subjective

assessment dimensions were identified: Impression

of green, healthy and recreational environment;

Impression of the state of the park and Assessment

of accessibility of the park by private or public

transport. The subjective assessment dimensions,

representing visitor’s perception on the studied

UGS, and the objective data were tested against each

other with multiple regression analysis to seek for a

correlation and the degree of the relation between

the two inherently different information domains.

Figure 2 demonstrates the workflow of the first tier.

Figure 2: The workflow of Tier 1.

The results of the analysis are explained in

Section 6.1.

5.3 Methodology 2

5.3.1 Used Datasets

Perceptual (subjective, qualitative) data:

 Questionnaire survey on perceived services of

the studied five urban green spaces of Szeged.

Crowd-sourced voluntary information on actual

usage and aesthetic appreciation of the study areas:

 Crossing and tangential running paths uploaded

to online recreational applications, Futótérkép

and Runtastic (www.futoterkep.hu;

www.runtastic.com);

 Photos from Panoramio, 360cities, Flickr and

Instagram picturing aesthetic appreciation and

recreational use of the five green spaces

(www.panoramio.com, www.360cities.net,

www.flickr.com and www.instagram.com).

5.3.2 Description of Methodology

The questionnaire data, described in Section 5.2.1

and Section 5.2.2, also incorporated information on

the perceived services provided by the five studied

parks. The answers for the questionnaires were

Multidimensional Relation of Urban Dwellers and Green Spaces

collected through a 1 to 5 Likert scale. The scores

reflect on green space users’ assessment of the

quality of recreational and infrastructural services of

the parks, as well as the subjective (perceived)

capacity of the five urban green spaces to mitigate

natural and human generated environmental

nuisances. As this information reveals individually

different opinions of survey participants it is rather

subjective. Therefore it reflects perceived, though

still crucially important, properties of the study

areas. Since the actual park usage and aesthetic

appreciation cannot be revealed by questionnaire

surveys the addition of this information is crucial for

urban planning and public space management.

To complement perceived information with data

on real park usage, the running paths crossing or

touching the park grounds were collected from the

Futótérkép and the Runtastic recreational crowd

services. For the assessment of recreational and

aesthetic use of the parks geo-located photos taken

and uploaded by users of the five public spaces were

downloaded from Panoramio, 360cities, Flickr and

Instagram photo sharing services.

Figure 3: The workflow of Tier 2.

The downloaded photos were individually

studied and classified into seventeen categories each

representing different aspects of the aesthetic

appreciation and recreational use of the green

spaces. The number of running paths will be

summed for each park in the next phase of the

second tier. The number of running paths and the

number of images per classes will provide real park

usage information and demonstrates the popularity

and suitability of the studied urban green spaces for

recreational and aesthetic purposes.

The survey records, the number of running paths

and the number of images per classes will be input

for an ordinal logistic regression analysis. The

analysis is expected to reveal the different

“measurement” capabilities of the perceived services

of the parks collected through the survey, and the

green space visitor generated actual usage of

aesthetic and recreational benefits from the study

areas. Figure 3 demonstrates the workflow of the

second tier of the thesis.

5.4 Methodology 3

5.4.1 Used Datasets

 Stereo scene of the panchromatic band of the

Pléiades satellite imagery;

 Normalized Difference Vegetation Index derived

from the multispectral product of the Pléiades

satellite imagery;

 Enumeration district level, tabular format census

data of Szeged sourced from the 2011 Census of

Hungary;

 Vector layer of census enumeration districts of

Szeged;

 OpenStreetMap road network of Szeged.

5.4.2 Description of Methodology

Digital surface model (DSM) of the city of Szeged

will be generated from the stereo scene of the

Pléiades satellite imagery. The DSM will be used to

detect object heights within the satellite scene. The

vegetation within the scene will be indicated by the

NDVI product.

Using object heights and NDVI values, the

vegetation will be identified and classified into three

levels of vegetation height classes using object based

image analysis (OBIA) methods:

 Low vegetation (e.g. lawn);

 Middle height vegetation (e.g. bushes);

 High vegetation (tree canopy).

The High vegetation class will indicate tree cover.

As green spaces with tree canopies have the highest

societal value, this class will be used for the further

analysis. Families with children and elderly citizens

have an increased demand on green space

availability. The method, being elaborated in Tier 3,

will provide a solution for calculating green space

deprivation based on demographic data and real

distances measured on the road network.

Records of the tabular format census data will be

joined to the corresponding enumeration polygons of

the Szeged enumeration district vector layer.

Number of children, their family members and

elderly citizens in each enumeration district will be

calculated based on the demography sensitive

variables of the census records (e.g. “Households by

DCSMARTGREENS 2016 - Doctoral Consortium on Smart Cities and Green ICT Systems

household composition and by the age composition

of household members, 2011”; “Families by family

composition and by the number of children living in

the family, 2011”; etc.). By applying this method

census blocks with high proportion of children

and/or elderly citizens will be geographically

identified.

In the next step, routable road network of Szeged

will be generated from the OpenStreetMap data.

Using the road network, walking distances to the

closest vegetation patch with “high societal value”

will be calculated for polygons of each enumeration

district. By using the walking distances to high

societal value vegetation patches, and the number of

people with high green space demand within the

enumeration districts, hot spots of spatial

inequalities in personal availability of urban green

spaces of Szeged will be identified. Figure 4

demonstrates the workflow of the second tier of the

thesis.

Figure 4: The workflow of Tier 3.

6 EXPECTED OUTCOME

The general expected outcome of the thesis is an

improved knowledge on the multidimensional

relation between urban dwellers and green spaces.

This outcome is expected to be achieved through the

results from the three tiers of the work. The expected

outcomes of the three studies are described in this

section.

6.1 Outcome 1

The multiple regression analysis, applied in the first

tier, found two minor and two mediocre correlations

between the perceived properties and the ten

objective spatial attributes of the studied green

spaces. The study concluded that there is a weak

relation between subjective and objective, spatially

explicit, attributes of urban green spaces. Therefore

the work proposes a combined use of perceptual

information and spatially explicit data for green

space planning and management.

6.2 Outcome 2

First, the results from Tier 2 are expected to unveil

the potentials of park visitor generated crowd-

sourced information to provide quantifiable input

data on actual use of aesthetic and recreational

services of urban green spaces. Secondly, the

statistical test is supposed to prove that questionnaire

surveys, as the most common sources of perceptual

information, and actual park usage reveal different

aspects of societal benefits visitors derive from

urban parks. Therefore, the tier will propose the

addition of crowd-sourced park usage information to

the results of questionnaire surveys to provide a

more comprehensive input for decision making

processes. The research warns: only the application

of both information domains can ensure thorough

and informed urban development and management

practices.

As questionnaire surveys are often conducted by

urban governments and crowd-sourced data is

readily available online, the introduced methodology

has a high transferability to further study areas

regardless of environmental settings.

6.3 Outcome 3

The methodology, applied in Tier 3, will identify the

spatial location of demographic groups of urban

population with high green space demand. More

importantly, the tier reveals real distances between

the identified groups of Szeged’s population and

urban green spaces with high societal value. By

achieving this, hot spots of spatial inequalities in

personal availability of high societal value urban

green spaces of Szeged will be identified.

Although the method introduced in this research

is highly transferable to other cities, the

demographic development of the society is constant

in time and space. To overcome this difficulty, the

more frequently available micro-censuses or mid-

year population estimates can also be used for the

described analysis rather than relying solely on less

frequent census data. Thereby, the adaptability of the

Multidimensional Relation of Urban Dwellers and Green Spaces

methodology to other urban environments can be

ensured.

7 STAGE OF THE RESEARCH

This PhD research is planned to last for four years.

The work was started in October 2013 and is

planned to be finished by September 2017.

The first tier of the work has been fully

elaborated and written up for publication in an ISI

ranked journal, Urban Forestry & Urban Greening.

The manuscript was accepted with major revisions.

The revisions have already been carried out. The

revised draft will be read by co-authors and proof-

readers, and then it will be resubmitted by April

2016.

Research associated to the second tier of the

thesis is being conducted at the moment. Green

space user generated, voluntary information on

aesthetic and recreational use of services generated

by the study areas have been collected from crowd-

sourced recreational and content sharing

applications. At the moment, February 2016, images

picturing aesthetics of the five green spaces are

being evaluated. The analysis of running paths will

be carried out in March and April 2016. The

statistical analysis will take place in May, then the

course of the work and the results will be written up

for an ISI ranked journal article. The draft is

expected to be submitted in the third quarter of

2016.

The work related to the third tier has already

been planned. Practical work accomplished so far is

that point cloud data has been retrieved from the

stereo pair of Pléiades satellite imagery for DSM

extraction. Secondly, routable road network of

Szeged has been generated form OpenStreetMap

data. At the moment, access to census data is being

negotiated with the Szeged Branch of the Hungarian

Central Statistical Office. The research is expected

to being accomplished in the third and the fourth

quarters of 2016. The work and its results will be

written up for an ISI ranked journal article parallel to

the research. The manuscript will be submitted in the

first quarter of 2017.

The recent state of the entire PhD thesis is

approximately thirty percent of completeness.

ACKNOWLEDGEMENTS

This research was jointly funded by the Austrian

Science Fund FWF through the Doctoral College

GIScience (DK W 1237-N23) and the University of

Salzburg.

REFERENCES

Astrium, CNES & Airbus Defence and Space 2014.

Pléiades Satallite Imagery. Toulouse, France: Astrium,

and Airbus Defence and Space.

Bowler, D. E., Buyung-ali, L., Knight, t. M. & Pullin, A.

S. 2010. Urban greening to cool towns and cities: a

systematic review of the empirical evidence.

Landscape and urban planning, 97, 147-155.

DPGG 2011. Szeged Ortofotó 2011. Szeged, Hungary:

University of Szeged, Department of Physical

Geography and Geoinformatics (DPGG).

Fontana, S., Sattler, T., Bontadina, F. & Moretti, M. 2011.

How to manage the urban green to improve bird

diversity and community structure. Landscape and

Urban Planning, 101, 278-285.

Germann-chiari, c. & seeland, k. 2004. Are urban green

spaces optimally distributed to act as places for social

integration? Results of a geographical information

system (gis) approach for urban forestry research.

Forest policy and economics, 6, 3-13.

Irvine, K., Warber, S., Devine-Wright, P. & Gaston, K.

2013. Understanding urban green space as a health

resource: a qualitative comparison of visit motivation

and derived effects among park users in Sheffield, UK.

International Journal of Environmental Research and

Public Health, 10, 417-442.

Kaźmierczak, A. 2013. The contribution of local parks to

neighbourhood social ties. Landscape and Urban

Planning, 109, 31-44.

Lee, A. C. K. & Maheswaran, R. 2011. The health benefits

of urban green spaces: a review of the evidence.

Journal of Public Health, 33, 212-222.

Sümeghy, Z., Gál, T. & Unger, J. 2011. Szeged 3d

Városmodellje. Geodézia és Kartográfia, 63 (1), 13-

16.

Tallis, M., Taylor, G., Sinnett, D. & Freer-Smith, P. 2011.

Estimating the removal of atmospheric particulate

pollution by the urban tree canopy of london, under

current and future environments. Landscape and

Urban Planning, 103, 129-138.

DCSMARTGREENS 2016 - Doctoral Consortium on Smart Cities and Green ICT Systems