Development of a GIS-Based Model for Quantification of Scenic Resources

Doug PierceyESRI International User Conference

August 5, 2008

Presentation Outline

ContextVisual Quality ConceptsModel Development/DescriptionValidationUtilityNext Steps

Context

Increasing demands on the landbase– Importance of

non-traditional usesHolistic management required– Maintain sustainability– Trade-offs often

necessary

Context

Holistic management occurs– across multiple

temporal and spatial scales

– from multiple perspectives

Difficult to manage for non-traditional values– social perspective

(eg. visual quality)

Social

Economic

Ecological

Past

Present

Future

Loca

l

Reg

iona

l

Nat

iona

l

Inte

rnat

iona

l

Context

Why measure visual quality?– Indicator or catalyst for other non-traditional land

uses (eg. tourism, outdoor recreation).– It is what people see.

Visual Quality Concepts

Mapping visual quality

Local scale– 3-D rendered

landscapes– High level of detail over

small areas– Exploring options

Landscape scale– 2-D numerical

representation– Less detail over large

areas– Planning

Visual Quality Concepts

How do we quantify perception of visual quality?

711

5

104 3

1

89

Need to define a conversion between landscape components and public perception of visual quality.Visual Quality Index (VQI)– Manual exercise in the past

(sometimes subjective)– Computer-based model

desired (using GIS/RS data).

2

Model Development

Validation

LiteratureReview VQI Model

DataCollectionSurvey

Model Description

7 Landscape Components– Relative Relief– % Water– Dominant Water Type– Vegetative Variety– Topographic Variety– Degree of Alteration– Type of Alteration

All components GIS-availableHexagonal polygon mapping unit

Model Description

VisualQuality Model

• Relative Relief

•Topographic Variety

• Hexagonal polygon mapping unit

• Water Influence

F eature toR aster

S tra tumGri d

Foc al V arie ty3 X 3

S tra tVa rTemp

Re cla ss ify S t ra tMas k

St ra tumFie l d

F eature toRas te r (2)

P ol ygo nNumb er

Ras te r ToO the r Format

(mul tip le)

Deriv edWork sp ac

InputImageBands

G ri dsWork sp ace

Combi ne TMB and s

Outp ut T MBa nds

Greater T ha n Imag eA re a Mas k

Cons tantVa lue o f 0

Di sturba nc eMas k

D is t Mas k1

D is t Mas k2

D is t Mas k3

Di sturbance G ri ds

photo yr

Imag eY ear

S ing le OutputMap A lgebra A llMas k s

Sa mp le pres ampl es.dbf

CombineV egetated

Cl us ters

O utputVe gClus te rs

Ras te r ToO the r Format(mult ip l e) (2)

D eriv edWo rk s pac

Input V egCl us ters

Greater T ha n(2)

ClusterMask

R ecl ass if y (2) S ampleMask

P oly gon #

Cons ta ntVa lue o f 0

Fores tInvento ry

Ras te r ToO the r Format(mult ip l e) (3)

Deriv edWork sp ac

InputEO SD

EO SD Fore stMa sk

S ing le OutputMa p A lgeb ra

(2)

for_c lusters

OutputForestMask

• Spatial representation of visual quality

• Vegetation

Model Description

% Water Value

0 – 5 0

6 – 30 1

30 2

AlterationLevel

Value

51 – 100 0

26 – 50 1

11 – 25 2

1 – 10 3

0 4

AlterationType

Value

Cutover 0

Agriculture 1

Built Up 2

None 3

Relative Relief

Value

0 – 164 0

165 – 822 1

822 2

Topographic Variety

Value

Uniform 0

2 general topographic

types

1

> 2 topographic

types

2

Vegetative Variety

Value

1 type of vegetation

0

2 types of vegetation

1

> 2 types of vegetation

2

Water Type Value

River 0

Lake 1

Ocean 2

2

Landscape 1 Landscape 2

Flat, primarily agricultural land

(40%) with a small lake. Some

scattered balsam fir stands.

Coastal plain leading inland to

mountain plateau. Minimal

disturbance. Balsam fir and

white birch dominate

13

Model Description

Model includes 2 Python scripts in an ArcToolbox– Hex Generator– Visual Quality Model

Validation Overview

Validation

VQI Model

DataCollectionSurvey

St. John River Valley,New Brunswick

Humber River Basin,Newfoundland

Current ValidationFuture Validation

Validation

Field photographs manually rated using landscape components and associated sub-indices from visual quality model Percent and

Type of Water

Percent and Type of

Alteration

Relative Relief

Topographic

Variety

Vegetative Variety

Validation

Photographs are grouped according to ratings and/or landscape characteristics for public survey

When is a cutover no longer a cutover?

What influence

does development

have?

Validation

Public Survey– Mail routes used to define

survey zones– Intercept at Gros Morne

National Park– ~300 returns

0

2

4

6

8

10

12

14

16

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

ModelSurvey Mean

Survey Mode

Survey Findings – Model Adjustments

Cutovers not seen as negative once regenerating (~5 years) – time since harvest needs to be included in modelStand remnants after harvest lessen negative impact of cutover – not always captured in forest inventory proceduresGreen space within built-up areas have positive impactsOpen field agriculture not as negative an impact as in literature – farms in the study area are integrated in the landscapeGolf courses not considered developedFamiliarity has slight positive impact on perceived values

Model Utility

Time 0 Visual QualityTime 0 Visual Quality

Time 30 Visual Quality – Natural Succession ScenarioTime 30 Visual Quality – Natural Succession Scenario

Time 30 Visual Quality – Forest Management ScenarioTime 30 Visual Quality – Forest Management Scenario

Integrated with other model

outputs in trade-off analysis

Model Utility

Local Scale PlanningE

xplore Options

Value change of

45%

Proposed Agriculture 15% of hex

TCH Viewshed

Next Steps

Adapt model according to survey findings.Finalize ArcGIS toolbox with updated scripts and help files.

Acknowledgements

Thank You!