challenges for active transport to school in urban and rural new … · 89.1% 11.8% distance(km)...

1

Challenges for Active Transport to School in Urban and Rural New Zealand:

Insights from the BEATS Research Programme

Associate Professor Sandy MandicActive Living Laboratory

University of Otago

Email: [email protected]

TALES Symposium 2019 | Dunedin, New Zealand | 13 Feb 2019 www.otago.ac.nz/active-living

Teamwork | Curiosity | Creativity | Excellence

Areas of research

• Physical Activity• Public Health• Transport• Built Environment• Sustainability

Exercise science

Exp

ert

ise

Geographic information

science(GIS)

Childrenand health

geographies

Behavioural medicine;

Research methods

Māori health

Geographies of transport

andmobilities

Education

Public health; Community

interventions

SandyMandic

AntoniMoore

Christina Ergler

EnriqueGarcía John

Spence

Anna Rolleston

Debbie Hopkins

Susan Sandretto

Kirsten Coppell

Michael Keall

Biostatistics

Physical activity and health

BEATS Team 2018: Multidisciplinary Expertise

Advisory Board Members:

Gavin Kidd, Gordon Wilson (Dunedin Secondary Schools’ Partnership)

Nick Sargent (Dunedin City Council)

Greame Rice (NZ Transport Agency)

Janet Stephenson (Centre for Sustainability)

Frank Edwards (Māori) and FinauTaungapeau(Pacific) community representatives

Physicalactivityand health

promotion

BEATS Research Programme

Report 2013-2018

Now available on the BEATS Study website:

www.otago.ac.nz/beats

Want to know more?

Sign up for Active Living Laboratory Newsletter:

https://goo.gl/jtqdAo

New Zealand Physical Activity Report Card for Children and Adolescents 2018

Indicator Grade

Overall physical activity D-

Organized sport participation B

Active play C+

Active transportation C-

Sedentary behaviours D

Physical fitness INC

Family and peers C-

School B-

Community and environment B

Government B+

Grade indicators: A=81%-100%; B=61%-80%; C=41%-60%; D: 21%-40%; F=0%-20%; N/A=Not applicable; INC=incomplete/insufficient data.

Smith M et al. (2018) Full report.doi: 10.17608/k6.auckland.7295882

Weight Status and Physical Activityin New Zealand Adolescents

National Survey of Children and Young People Physical Activity and Dietary Behaviour in NZ. 2007/08

Age (years)

Mee

ting

PA

gui

delin

es (

%)Underweight

3.2%

Healthy weight69.6%

Overweight20.5%

Obese6.8%

Source: BEATS Study (2014/2015)1,300 Dunedin adolescents

(measured heights and weights)

Mandic et al. Am J Health Behav. 2017;41(3):266-275

2

Physical Activity (PA) in Dunedin Adolescents

Accelerometer assessment of physical activity (n=314)

Sedentary activities: ~9.6 hrs/day (70%)

Light physical activity: ~3.2 hrs/day (25%)

Moderate-to-vigorous physical activity:

~1.0 hrs/day (<10%)

Average daily activity profile:

39.2% met PA guidelines(45.6% boys, 36.0% girls; p=0.101)

Kek CC et al. J Sport Health Sci. (in press)

Guidelines:≥60 min per day

Average: 4.2 ± 2.1 days/week

17.9% met PA guidelines

Mandic et al. Am J Health Behav. 2017;41(3):266-275

Self-reported physical activity

(n=1,300)

Ministry of Transport. (2015). 25 years of New Zealand travel: New Zealand household travel 1989–2014. Wellington: Ministry of Transport.

1989/1990 2010-2014

2.5 million vehicles

3.4 million vehicles

72% car travel 78% car travel

755 deaths 294 deaths

1h/day travel(28 min driving)(10 min walking)

1h/day travel(32 min driving)(8 min walking)

Travel to school:21% driven26% walking19% cycling

Travel to school: 32% driven27% walking3% cycling

Built Environment andTransport-Related Physical Activity

www.designedtomove.org

Built Environment and Transportation

• Walkable community design

– Density

– Connected streets

– Mixed land uses

– Access to transit

• Pedestrian & bicycle facilities

– Access; Connectivity

– Design; Quality; Safety

• Perceived environment: accessibility and convenience

Sallis SF et al. Circulation. 2012;125:729-737http://www.houselogic.com/home-advice/green-living/make-my-neighborhood-more-walkable/

http://switchboard.nrdc.org/blogs/kbenfield/how_communities_can_support_wa.html

• Investigates:

– transport to school habits,

– the neighbourhood environment and

– physical activity habits

in Otago adolescents. www.otago.ac.nz/beatsMandic S et al. BMJ Open. 2016; 6:e011196

BEATS Research Programme at Otago BEATS Research Programme Framework:Ecological Model for Active Transport

Individual

Social/Cultural Environment

BuiltEnvironment

PolicyEnvironment

SociodemographicsBehaviour

Motivations/barriers

Social supportSocial norms

Walkable community design

Pedestrian & bicycle facilities

School policy for ATSSchool’s road safety

procedures

Adapted from Sallis JF et al. Circulation. 2012;125:729-737

Mandic S et al. BMJ Open. 2016; 6:e011196

3

Research Methodology

Survey Maps; GIS Analysis Physical Activity

Focus groups

Adolescents & Parents

Adolescents

Anthropometry

School bag weightAdolescents, Parents, Teachers

Interviews

School Principals


BEATS Research Programme (2013-2022)

URBAN RURAL

BEATS Study

(2014-2017)(Dunedin)

BEATS Rural Study

(2018-2019)(Rural Otago)

BEATS Natural Experiment

(2020-2022)(Dunedin)

11 Schools1014 Adolescents

78 Parents2 Principals


355 Parents14 Teachers12 Principals

Urban versus rural

BEATS Cultural Study (2018-2019; in progress)

Exercise Science

PublicHealth

TransportBuilt

EnvironmentEducation

Partnerships: City Council CommunityAcademia Schools

Disciplines & impact areas:

Significance

Understanding influences of multiple factors will enable the scientific community, policy makers, regional planners, and health promoters to address barriers to active transport to school

Generating important information for key stakeholders for planning future school-, neighbourhood- and city/town-wide built environment changes to encourage active transport to school

If effective, population level initiatives aimed to ↑ physical activity across all groups such as pedestrian and/or cycling infrastructure construction/ improvements will contribute to improving health and wellbeing of communities

Involvement of the key stakeholders facilitates the generation of usable data, relevant to the local contextand generalisable to other areas, and the incorporation of new knowledge into policy and future initiatives

BEATS Research Programme 2013-2018: Overview

BEATS Study(Dunedin; 2014-2017)

BEATS Rural Study(Rural Otago; 2018-2019)

BEATS Natural Experiment(Dunedin;

2020-2022)

Subject to funding


78 Parents2 Principals


355 Parents14 Teachers12 Principals

BEATS Cultural Study

(Dunedin & Bay of Plenty;2018-2019)

In progress

Exercise Science Public Health Transport Built Environment Education

Partnerships: City Council CommunityAcademia Schools

Disciplines & impact areas:

The CatalystProject

(Rural Otago; 2020)

Subject to funding

Transport sector

Members:• 10 Investigators• Advisory Board members• Research staff • Research students• Collaborators

Team’s expertise:• Exercise sciences• Public health• Physical activity promotion• Behavioural medicine• Biostatistics

• Māori Health• Geographies of transport• Children geographies• Geographic Information Science • Education

Research team:

Research data collection:

• Survey• Anthropometry• Accelerometers• Focus group/interview• Mapping

Adolescents

Projects:

Parents Teachers /

Principals

• Built environment analysis (Geographic Information Science)

• Environmental scan of school neighbourhoods

Spin-off projects:• Evaluation of cycle skills training programme (2015-2017)• Examining cycle skills training content and delivery (2017-2018)• BEATS Study Symposium (2014; 2016; 2018)• The Active Living and Environment Symposium (2017; 2019)

Research outputs (as of Feb 2019):15 Journal articles

1 Book chapter103 Conference abstracts

38 Technical reports

25 Seminar presentations for academic audiences

21 Seminar presentations for non-academics

BEATS Research Programme (2014-2018)

Dunedin City (2014-2017)

12 Secondary schools (100% school recruitment rate) 11 Secondary schools

(73% school recruitment rate)

Dunedin

Otago Region(2018-2019)

BEATS Study BEATS Rural Study

Otago Secondary Schools Supporting BEATS (23 out of 27 schools; 85%)

Dunedin (2014/15)(12 out of 12 school)

Rural Otago (2018)(11 out of 15 school)

Total sample of adolescents (n=2,656)

4

Spin OffProjects

Evaluations of Cycle Skills

Training

(2015-2017)

Cycle Skills Training Content

and Delivery

(2018-2019)

BEATS Study Symposia

2014 | 2016 | 2018

2nd 2019

1st 2017

(2020)

Comprehensive Dissemination of Research Findings

Last updated: Feb 2019

Journal articles

Conference abstracts

Technical Reports SymposiaPresentations

BEATS Research Programme Outputs to Date

38 Published(including 4

progress reports)

2 Keynotes3 Invited

25 Academic21 Non-academic

3 Local(2014; 2016; 2018)

2 International (2017; 2019)

15 Published2 in review

6 in preparation1 Book chapter

28 International21 National

54 Local

BEATS Research Programme Findings (2014-2018)

Transport to School Habits across OtagoTotal sample

(n=2,656)

15.2%

55.0%

29.8%

Active Transport

Motorised + Active

Transport

Motorised Transport

No significant difference across urbanisation settings

Mandic S et al. (abstract). ISBNPA 2019 (Accepted)

Among adolescents ineligible for subsidised school bus

(living within 4.8 km from school)

38.8%47.9%

58.5%

0%

20%

40%

60%

80%

100%

Main urbancentre

Semi urbanareas

Ruralsettlements

Active transport to school

p<0.001

89.9% liked how they travel to school

78.9% had a bicycle at home

75.8% had 2+ vehicles at home

Threshold distance for walking to school:

≤2.25 kmSensitivity: 85%Specificity: 86%

AUC: 93%

Threshold Distance for Walking and Cycling to School in Adolescents

1.3-2.5 km

≤4.0 km

Pocock T et al. Health & Place. 2019;55:1-8

≤2.3 km ≤4.0km Beyond4.0km

n=1,475(boarders excluded)

31.7%

53.4%(31.7% + 21.7%)

46.6%

Walkable and Cycleable Distance to Secondary School in Dunedin

Average distanceto secondary school

in Dunedin:

6.2 ± 7.4 km

5

Within walkable distance: 32%

(range: 6% to 64%)

Walking to school rates: 26%

(range: 12% to 47%)

Walkable and Cycling Distancesto Dunedin Secondary Schools

Within cycling distance: 53%

(range: 37% to 79%)

Cycling to school rates: 1.5%

(range: 0% to 7%)

Transport to School Habits across Otago

Source: BEATS Study and BEATS Rural Study (n=2,656)

Optimal distance for walking to school

≤ 2,25 km

Sensitivity: 85%Specificity: 86%

AUC: 93%

89.1% 11.8%

Dis

tan

ce(k

m)

3.7 3.26.2

0

2

4

6

8

10

Main urbancentre

Semi urbanarea

Ruralsettlements

Median distance to school (km)

p=0.037

Significant difference across urbanisation settings

Distance to schoolPro

babi

lity

of w

alki

ng

to s

cho

ol

Pocock et al. Health and Place (in press)

(≤4.8 km)


39%55% 50% 46% 45% 41% 39% 38% 35% 30% 30% 26%

17%

0%

25%

50%

75%

100%

URB Bay Que Log Tai JMc Col Kin OB StH Kai Kav OG

Rates of Active Transport to School(Living ≤4.8 from school; boarders excluded)

*Most of the time / All of the time

Urban (n=897)

48%62%

51% 45% 42% 41% 35%

0%

25%

50%

75%

100%

S-URB Cro Tok MtA WG SO StK

Semi-urban (n=457)

59% 65% 62% 60% 56% 50%

0%

25%

50%

75%

100%

RUR BMC Law Cai Rox Man

Rural (n=81)

1% 5% 3% 3% 3% 1% 1% 0% 0% 0% 0% 0% 0%0%

25%

50%

75%

100%

URB Tai Log Kai Kin OB Bay Col JMc Kav OG Que StH

Rates of Cycling to School(living ≤4.8 from school; boarders and mixed modes excluded)

Urban (n=897)

16%32% 22%

6% 3% 2% 0%0%

25%

50%

75%

100%

S-URB MtA Cro Tok StK SO WG

Semi-urban (n=457)

10% 15% 12% 11%0% 0%

0%

25%

50%

75%

100%

RUR Law BMC Rox Man Cai

Rural (n=81)

WanakaCromwell

Transport to School and Physical Activity in Dunedin Adolescents

Chiew Ching Kek

Motorized Transport (n=185)

Active and Motorized Transport

(n=56)

Met physical activity guidelines

Physical activity differences by transport mode were observed:

• In girls but not boys

• On school days but not weekend days

• Only during school commute time

Active Transport

(n=73)

47.9% 46.4% 33.5%

Kek CC et al. J Sport Health Sci. (in press)

Both active transport and combined active and motorized transport to/from school are potential avenues to increase daily physical

activity in adolescents, particularly in adolescent girls.

Perceptions of Distance to School(among adolescents living ≤4.8 km from school)


28%

8%0%

0%

20%

40%

60%

80%

100%

Main urbancentre

Semi urbanareas

Ruralsettlements

It is too far to walk to school.

37%

20%12%

0%

20%

40%

60%

80%

100%

Main urbancentre

Semi urbanareas

Ruralsettlements

It is too far to cycle to school.

*p<0.001 *p<0.001

6

Adolescents’ Perspective

Mandic S et al. Journal of Transportation and Health. 2017: 4:294-304.

BEATS Student Survey (n=764)(adolescents living ≤4 km from school)

The Tale of Two New Zealand CitiesDunedin

(n=430)

38.4%

47.2%

1.9%

Christchurch (n=373)

16.9%

50.9%

17.7%

Living ≤4 km from school

Age (years)15.0 ± 1.6

44.9% boys55.1% girls

Age (years)14.3 ± 1.6

62.7% boys37.3% girls

Frater et al. Transp Res F: Traff Psych Behav. 2017;49:205-214

Dominant influence

Attitude Subjective norm

Future interventions should consider interpersonal factors, the social needs of adolescents and adolescents’ cycle skills

Dr Jillian Frater

(Canterbury)

Adolescents’ Perceptions of Route to Schoolfor Walking or Cycling

n=753(non-boarders; living ≤4km from school)

Convenienceof being driven to

school by someoneon the way to

something else

Too much traffic

35.6%

Too many hills

32.9%

Dangerous crossing(s)

32.4% 52.9%

Boring route

33.3%


Parental Barriers to Active Transport to School

Convenienceof trip chaining

Fewer barriers for walking compared to cycling

Future interventions should address parental barriers for active transport to school (especially for cycling).

Mandic S et al. (Abstract); OERC Symposium 2016 and ISBNPA 2017.

68.8%68.1% 65.5% 60.8%

57.3% 37.2%

66.4%

41.6%

Perceptions of Safety(among adolescents living ≤4.8 km from school)


It is unsafeto walk to school.

12% 9% 0%0%

25%

50%

75%

100%

Urban Semi-urban

Rural

It is unsafe to cycle to school.

*p<0.006

Adolescents’concerns

Parentalconcerns

(reported by adolescents)13% 8% 3%

0%

25%

50%

75%

100%

Urban Semi-urban

Rural

40%23%

5%0%

25%

50%

75%

100%

Urban Semi-urban

Rural

31%15% 4%

0%

25%

50%

75%

100%

Urban Semi-urban

Rural

*p<0.006 *p<0.001

*p<0.001

Perceptions of Cycling to School(From Student and Parental Focus Groups)

• Perceived safety:– A complex range of factors including:

• Features and perceptions of the built environment

• Traffic safety (including behaviours of other road users)

• Previous cycling experiences (including accidents)

• Adolescents’ cycling skills and on-road experiences

• Implicit messages

• Social normsHopkins D and Mandic S. International Journal of

Sustainable Transportation 2017;11(5):342-356

Dr Debbie Hopkins(Oxford)

7

Cycle Skills Training: Perceptions and Evaluation

Mandic S et al. Transport Res F Psychol Behav. 2016;42:217-226

39% of adolescentsCycle skills training could make adolescents

safer in traffic

Mrs Charlotte Flaherty (DCC)

73% of parentsInterventions should capitalize on this interest

Mandic S et al. J Transp Health. 2017;6:411-419

Evaluation of Cycle Skills Training Programme in Dunedin (2015-2017)

https://www.odt.co.nz/news/dunedin/action-urged-get-more-cycling-school

Mandic et al. J Trans Health. 2018;8:271-282 Mandic et al. J Trans Health. 2018;9:253-263

(BEATS Spin-off project)

Cycle Skills Training (CST) Intervention

Mandic S et al. J Trans Health. 2018; 8: 271-282Mandic et al. J Trans Health. 2018; 9: 253-263

Effects of Cycle Skills Training (CST) in Children

Traffic-Free CST

(n=164)

Traffic-Free + On-Road CST

(n=265)

To achieve behavioural change, additional interventions may be necessary.

Cycling-related knowledge

… in parks/reservesSelf-perceived confidence to cycle

… on the road… to school

Pre-training Post-training

Cycling behaviours and preferences

Improved

(1-6weeks)

Rates of cycling to school

Cycling as a preferred modeof transport to school

Improved

Improved

No change

Improved

Improved

Improved

No change

No change(3.1% to 2.4%)

Increased(10.6% to 12.5%)




Mandic et al. J Trans Health. 2018; 8: 271-282

Cycle Skills Training (CST) in Adolescent Girls

Traffic-Free CST

(n=43)

Traffic-Free + On-Road CST

(n=74)

Future CST programs should be tailored to adolescents’ needs and preferences.

Cycling-related knowledge

… in parks/reservesSelf-perceived confidence to cycle

… on the road… to school


Cycling behaviours and preferences

Improved

Rates of cycling to school

Cycling as a preferred modeof transport to school

Improved

No change

Improved

Improved

No change

No change(0% to 0%)

No change(0% to 0%)



(1-10 weeks)

No change

No change


Mandic et al. J Trans Health. 2018;9:253-263

Enablers of Cycling to School:Adolescents’ Perspective

Cycle-friendlyuniform

41.4% 40.1%

Safer bike storageat school

Slower traffic

36.4%

Bus bike racksfree of charge

26.2% 32.7%

Bike ownership Cycling withouta helmet

22.1%

n=764(non-boarders;

within ≤4km)


Cycle Helmet Legislation as a Barrier to Cycling to School

774 Adolescents living ≥4 km from school (BEATS Study; Dunedin) Molina-García et al. (2018) J Transp Health. 11, 64-72

Adolescents would cycle to school more if helmet use was not mandatory

Recommendations:

Design educational interventions to influence adolescents’ attitudes towards the use of helmets. • Could be offered

as part of cycle skills training

22%

Significant factors:

• Distance to school • Māori and other ethnicities (vs. NZ European)

• Cycling to school is ‘not cool’• Cycling often with friends• Boring route to school• Cycling as a great way to exercise (+)

8

Active transport users were less likely to report heavy

school bags

School Bag Weight as a Barrier

682 Adolescents and 331 parentsBEATS Study (Dunedin) Mandic S et al. (2018) Children. 5:129

Adolescents:School bag perceived as a barrier to active transport to school

68% of parents58% for walking66% for cycling

Full bag weight:

5.6 kg (± 2.1 kg)

9.3% (± 3.9%) of adolescents’

body weight

37.9% Above10%

of body weight

Actual school bag weightsdid not differ by mode of

transport to school.

Being Driven to School: Adolescents’ Perspective

Olivia Eyles

Dr Christina Ergler

(Otago)

Positives Negatives

Eyles O et al. 2018. (Abstract) BEATS Study Symposium.

Bus to School – or Not?Isobelle

LaneDr Christina

Ergler (Otago)

Lane I et al. 2018. (Abstract) BEATS Study Symposium.

New Zealand’s Systematic Meta Analysis

Ikeda E et al. J Transp Health. 2018;9:117-131

Higher rates of active transport to school:

• When distance to school is <1.3 km compared to greater distances• In boys versus girls• In medium decile compared to high decile schools• In home neighbourhoods with higher versus lower street connectivity

Pocock T et al. Health & Place. 2019;55:1-8

• School locations:

• Small-to-medium urban areas (6 schools)

• Rural settlements (5 schools)

• Adolescents’ perceptions of the school neighbourhood environment

• Subsample: Modified NEWS-Y Questionnaire for school neighbourhoods

School Neighbourhoods and Active Transport to School

• Environmental scan of school neighbourhoods

• GIS analysis of school neighbourhood built environment

• Adolescents’ perceptions of the route to school

Tessa Pocock

Schools in an urban centre (Dunedin)

BEATS Rural Study (2018)

White B et al. (abstract) TALES Symposium 2019

Brittany White

Maps with route to schooln=740

1,462 digitized segments347 (24.3%) marked as ‘unsafe’

Adolescents perceptions of ‘unsafe’ segments:

50.2% Built environment characteristics (roads, intersections, lack of footpaths)

39.6% Traffic safety (vehicles, traffic)

25.5% Personal safety (people, dogs, street lighting)

6.3% Other (weather-dependent, winds, glare)

Rodda J et al. BEATS Symposium 2016 (Abstract)

A/Prof Antoni Moore

Dr Judith Rodda

9

School Choice and Transport to School

Mr Gordon Wilson (DSSP)

Dr Susan Sandretto

Mandic S et al. Journal of School Choice. 2018;12(1):98-122

Comments from students and

parents

Social connections

Programmes and

facilities

Co-ed status

Positive feedback about school:

51.9% Students 51.3% Parents

Reasons for school

choice

50.7% Friends enrolled

35.1% Sibling(s) went or enrolled

27.0% Parents attended

68.6% Preferred co-ed

25.3% Preferred single-sex school

52.3% Facilities

46.3% Sports programmes

33.0% Cultural programmes

n=1,465

(boarders excluded)

Other

7.3% Other reasons

Proximity

36.5% Closest school to home

Adolescents’ perspective Negative feedback

from people at closest school:

8.0% Students 8.4% Parents

Mandic S et al. Journal of School Choice. 2018;12(1):98-122

Implications of School Choice Decisions on Active Transport to School

Mandic et al. Journal of Transport and Health.

2017; 6:347-357

Adolescents (n=797)(Special character (integrated) schools and schools with zoning excluded)(Students from 6 out of 12 Dunedin secondary schools (50%) included in the analysis)

Without school zoning,51.3% of adolescents

enrolled in the closest school.

Distance

Importance of school proximity to home

Co-educational school status

Peer feedback

Closest school

Distant school

46.5% 8.8%

School choice has implications for education, health, transport and

environment

40.3% 68.8%

Baseline data (BEATS) (2014/2015) (12 schools)

Follow-up data (2020/2021) (≥10 schools)

Exposure areas:

Cycling and pedestrian

infrastructure construction 2014-2018

(6 ‘exposure’ schools)

Control areas:

No change (no new

infrastructure construction)

(6 ‘control’ schools)

J

A

F

HIE

C

B K

GL

‘Exposure’ schools: Pedestrian infrastructure‘Exposure’ schools: Cycling infrastructure

‘Control’ schools

D

Cycling infrastructure constructed since 2014

Next Challenge:BEATS Natural Experiment (2020-2022)

Online survey

GISAnalysis

Physical activity

(40-50 studentsper school)

School bag weights

Height &weight

Mapping route& school

neighbourhood(30 students per school)

Focus group

(6-8 studentsper school

Environmental scan ofschool

neighbourhoods

Additional activities

Data collection at schools(120-150 students per school; 1 school period)

Built environment analyses

BEATS Natural Experiment: AssessmentsSummary: Factors Influencing Active

Transport to School in Otago Adolescents

Individual

Social/Cultural Environment

BuiltEnvironment

PolicyEnvironment

Sociodemographic characteristics

Preferences, IntentionPerceptionsCycle skills

Social supportSocial norms

School decile/SES

Distance to schoolBuilt environment features

Perceptions of safety

School choice policySchool uniforms

Cycling helmet legislation

Adapted from Sallis JF et al. Circulation. 2012;125:729-737


Future research:

• Urban versus rural settings

• Walking versus cycling• Understanding local

context

10

Policy Implications

• Different interventions and approaches are required for increasing rates of walking versus cycling to school

• Different interventions are required in urban versus rural settings

• Minimise adolescents’ and their parents’ safety concerns related to walking and cycling to school

• Promote mixed modes of transport to school when relying solely on active transport is not feasible

• Encourage and support evaluations of the effectiveness of active transport-related infrastructure changes in short-, medium- and long-term

• Consider implications of other relevant policies such as education policies related to school choice, school uniforms and school bag requirements

challenges for active transport to school in urban and rural new … · 89.1% 11.8% distance(km)...

Documents