Predicting Urban Growth on the Atlantic Coast Using an Integrative Spatial

Modeling Approach

Jeffery S. Allen and Kang Shou LuClemson University

Strom Thurmond Institute

Coastal Community Workshop, April 20, 2006, Conway, SC

Population density map for North Carolina, South Carolina, and Georgia

# of People Per Square Mile*

> 800

400 - 800

200 - 400

100 - 200

0 - 100

* 1999 population estimates by CACI International, Inc. based on 1990 US Census

Population in the Coastal Counties of South Carolina & Georgia

1950 1960 1970 1980 1990 2000South CarolinaBeaufort 26,993 44,187 51,136 65,364 86,425 120,937Charleston 164,856 216,382 247,650 276,974 295,039 309,969Collecton 28,242 27,816 27,622 31,776 34,377 38,264Georgetown 31,762 34,798 33,500 42,461 46,302 55,797Horry 59,820 68,247 69,992 101,419 144,053 196,629Jasper 10,995 12,237 11,885 14,504 15,487 20,678Total Population 322,668 403,667 441,785 532,498 621,683 742,274

South Carolina 2,117,027 2,382,594 2,590,516 3,121,820 3,486,703 4,012,012USA 151,325,798 179,323,175 203,211,926 226,545,805 248,709,873 281,421,906

GeorgiaBryan 5,965 6,226 6,539 10,175 15,438 23,417Camden 7,322 9,975 11,334 13,371 30,167 43,664Chatham 151,481 188,299 187,767 202,226 216,935 232,048Glynn 29,046 41,954 50,528 54,981 62,496 67,568Liberty 8,444 14,487 17,569 37,583 52,745 61,610McIntosh 6,008 6,364 7,371 8,046 8,634 10,847Total Population 208,266 267,305 281,108 326,382 386,415 439,154

Georgia 3444578 3943116 4589575 5463105 6478216 8,186,453USA 151,325,798 179,323,175 203,211,926 226,545,805 248,709,873 281,421,906

Percent Change in Population in the Coastal Counties of South Carolina & Georgia

1950 1960 1970 1980 1990 2000South CarolinaBeaufort -- 63.70 15.73 27.82 32.22 39.93Charleston -- 31.26 14.45 11.84 6.52 5.06Collecton -- -1.51 -0.70 15.04 8.19 11.31Georgetown -- 9.56 -3.73 26.75 9.05 20.51Horry -- 14.09 2.56 44.90 42.04 36.50Jasper -- 11.30 -2.88 22.04 6.78 33.52Total Percent Change -- 25.10 9.44 20.53 16.75 19.40

South Carolina -- 12.54 8.73 20.51 11.69 15.07USA -- 18.50 13.32 11.48 9.78 13.15

GeorgiaBryan -- 4.38 5.03 55.60 51.72 51.68Camden -- 36.23 13.62 17.97 125.62 44.74Chatham -- 24.31 -0.28 7.70 7.27 6.97Glynn -- 44.44 20.44 8.81 13.67 8.12Liberty -- 71.57 21.27 113.92 40.34 16.81McIntosh -- 5.93 15.82 9.16 7.31 25.63Total Percent Change -- 28.35 5.16 16.11 18.39 13.65

Georgia -- 14.47 16.39 19.03 18.58 26.37USA -- 18.50 13.32 11.48 9.78 13.15

5.3%

30.2%

0%5%

10%15%20%25%30%35%

South Carolina: Comparison of Population Growth to Increase in Developed Land 1992-97

Developed Land

Population

Source: (London and Hill, 2000) -- USDA, US Census Bureau and Jim Self Center on the Future, Clemson University.

Total Acres of Land Conversion by State, 1992-1997 (thousand acres) Rank STATE Acres converted to developed land (1,000 acres)1 Texas 1219.5 2 Pennsylvania 1123.23 Georgia 1053.24 Florida 945.35 North Carolina 781.56 California 694.87 Tennessee 611.68 Michigan 550.89 South Carolina 539.710 Ohio 521.2

Source: (London and Hill, 2000) -- USDA, 1997 National ResourceInventory Summary Report

Purposes and ObjectivesGain a better understanding of urban growth process; Develop a methodology for urban growth prediction; andProvide better information for:

Land use decision-making toward smart growth Impact assessment studies Public education of environmental awareness

Developing an operational urban growth model Calibrating the model using 1990-2000 data Predicting urban extent by year 2030 for the Beaufort-Colleton-Jasper Region

The objectives of this project are:

Urban Growth Models

Lowry’s Model (1957) and Its Variants Cellular Automata (Deltron) Model (San Francisco Bay Area)

--- Clarke (1996) California Urban Future Model (CUF I and II) --- Landis (1994, 1995, and 1997) Land Transformation Model (LTM) (Michigan’s Saginaw Bay Watershed)

---Pijanowski et al (1997)

1. Components or structures of the land use systems:simple vs. complex2. Relationships between components, agents, factors, and processes:

deterministic vs. indeterministic.3. Changes over space (and time): ordered vs. random vs. chaotic4. Spatial distribution or patterns: regularity vs. irregularity (fractal)

Challenges Faced in Urban Land Use Modeling

Land

Land Use Systems

Uses

Economic SocialCultural

•Natural resources•Activity settings•Aesthetic sanities •Natural functions

•Functions•Structures•Activities•Ownership•Use status

GeologyGeomorphologyHydrologyClimateSoilVegetation

Human Systems

Physical Systems

•Availability•Suitability•Capacity•Sustainability

Model vs. Reality

Parcel--smallest legal unit

Zone--area demarcated by the major roads

Grid or Cell--square-shaped area

Murrells Inlet

Mount Pleasant

Part of Mount Pleasant

Analysis Units

---200x200 m2 grids (cells) for calibrating models---30x30 m2 grids (cells) for prediction

Georgetown Data

Georegetown County: Predicted population and land use

population

urban land

use(KM2) Ratio 0.48:1

urban land

use(KM2)Ratio 1:1

urban land use(KM2)Ration 2:1

urban land use(KM2)Ration 3:1

1990 46302 22.1148 22.1148 22.1148 22.11481995 51050 23.0328 23.0328 23.0328 23.03282000 55797 23.9508 25.17456 27.31631 29.458072005 60544 24.8688 27.31631 31.59983 35.883342010 65291 25.7868 29.45807 35.88334 42.308612015 70038 26.7048 31.59983 40.16686 48.733892020 74785 27.6228 33.74159 44.45037 55.159162025 79532 28.5408 35.88334 48.73389 61.584432030 84279 29.4588 38.0251 53.0174 68.0097

Horry County Data

Horry County: Predicted population and land use

population

urban land

use(KM2) Ratio 1.02:1

urban land

use(KM2)Ratio 2:1

urban land

use(KM2)Ration 3:1

1990 144053 104.0553 104.0553 104.05531995 170341 123.4497 123.4497 123.44972000 196629 142.844 142.844 142.8442005 222917 162.2384 180.822 199.8112010 249205 181.6327 218.8 256.7782015 275493 201.0271 256.778 313.7452020 301781 220.4214 294.756 370.7122025 328069 239.8158 332.734 427.6792030 354357 259.2101 370.712 484.646

Predictor Variables

• Physical suitability– Land cover, Slope, Soil suitability

• Service accessibility– Transportation, Waterline, Sewer line, CBD, Industrial parks,

Demographic

• Initial conditions– Existing urban, Vacant infill area, Agriculture land, Forest land

• Policy constraints– Protected land, Comprehensive planning, Growth boundary,

Zoning/Ordinance, Natural reserves, Parks, Floodplain, Cultural sites, Land ownership

Data for Deriving Predictor GridsBaseline Years: 1990 and 2000 for Training and TestingProjection Years; 2000-2030

Variable Name Major Data Source

Market factors Population density Census Blockgroup (1990, 2000)

Housing unit value Census Blockgroup (1990, 2000)

Physical suitability Water and wetland NWI (1989) and Land Cover (1990 and 2000)

Slope /relief * DEM (USGS)

Distance to waterfront NWI (1989)

Service accessibility Distance to primary roads Roads (USGS 1990, 2000)

Distance to local roads Roads (USGS 1990, 2000)

Distance to major nodes Roads (USGS 1990, 2000)

Distance to waterline South Carolina Chambers of Commerce (1990)

Distance to sewage

Cost distance to CBD Multiple sources including roads (USGS 1990)

Initial conditions Distance to existing urban Land Cover (1990, 2000)

Policy constraints Protected lands Gap Analysis (USGS, 2003) (GA, NC, and SC) Corporate area Tiger/Lines (1990)

Examples of Predictor Variables

Distance to 2000 Urban Area

Distance to 80 Industry Point

Distance to Roads

Distance to Highway System

Distance to Water Lines

Distance to Sewage system

US Hwys

Waterfront

Pop. Density 2000

Water Lines

Probabilities

(dark is higher)

Horry 2010 r 3:1

Horry 2020 r 3:1

Horry 2030 r 3:1

Predicted Urban Growth in the Myrtle Beach Region, South Carolina, 2000-2030

115 sq. mi. 164 sq. mi. 213 sq. mi.

1995 - 56 sq. mi.

1992

2001

2010 3:1

2020 3:1

2030 3:1

Simulated Growth

Urban Sprawl Problems

Urban growth is necessary and unavoidable. But uncontrolled growth - urban sprawl results in many problems such as:

Increased cost of living Rising taxes and pressure on infrastructure and urban services Traffic congestion and increased (travel) time Environmental pollution Loss of farm/forest land, habitats and rural (natural) landscape Downtown declines and community segregation

Benefits of Urban Growth

Increased standard of living Generation of wealth Increase in amenities Production of affordable housing Increase in tax base New business opportunities New job opportunities Increased “freedom” with the automobile It is what we desire - “Freedom of Choice”

Urban Growth Trends

The pattern follows paths of subsidy.

•Undervalued infrastructure

•Discounted resources

•Reductions for individual risk

•Unintended consequences of past policies

What do we do now?

Growth is coming whether we want it or not Determine where we do not want to grow Increase communication among SPD’s, etc. Be inclusive in planning Provide incentives for growth in “growth areas” Provide “dis-incentives” for areas to protect Make users pay the freight for new growth It is always easier said than done!!!