- Email: [email protected]
Landscape structure, zoning ordinance, and topography in hillside residential neighborhoods: A case study of Morgantown, WV
Landscape and Urban Planning 108 (2012) 28–38
Contents lists available at SciVerse ScienceDirect
Landscape and Urban Planning journal homepage: www.elsevier.com/locate/landurbplan
Research paper
Landscape structure, zoning ordinance, and topography in hillside residential neighborhoods: A case study of Morgantown, WV Jinki Kim ∗ , Xiaolu Zhou University of Illinois, Department of Landscape Architecture, Fine & Applied Arts, 611 Taft Drive, Champaign, IL 61820, United States
h i g h l i g h t s Landscape structures are significantly associated with lot attributes governed by zoning ordinance. Zoning requirements for smaller minimum lot size and lot frontage result in more vegetation but also cause more fragmentation. Geographic slope influences landscape structure in hilly residential neighborhood.
a r t i c l e
i n f o
Article history: Received 23 January 2012 Received in revised form 17 July 2012 Accepted 24 July 2012 Available online 11 August 2012 Keywords: Landscape structure Zoning ordinance Geographic characteristics Single-family parcel Landscape fragmentation
a b s t r a c t The purpose of this study is to investigate the relationship among landscape structure, zoning ordinance, and geographic attributes in Morgantown, West Virginia. Two residential areas zoned by different standards, but with similar development history and socioeconomic status were examined. Aerial photography was digitized into woody and non-woody areas and the landscape structure was quantified with six core landscape metrics related to fragmentation. Computer aided design (CAD) map and digital elevation model (DEM) were used for the calculation of building footprint and geographical characteristics respectively. The findings indicate that differences in zoning requirements result in distinct landscape structure. Zoning requirements for smaller minimum lot size and lot frontage result in more vegetation but also result in more fragmentation than requirements that call for greater minimum lot size and lot frontage. Although building footprint is weakly associated with landscape structure, lot size and lot perimeter are strongly related to vegetation abundance, fragmentation, and dominance. Slope is associated with vegetation patch size, especially in hilly residential neighborhoods. This study provides a better understanding of how human and environmental factors are related to residential landscape structure at different scale. Providing numerous quantitative metrics is useful for understanding social and ecological benefits. The results of this study will help planners, landscape architects, and administrators for planning and designing more ecologically and socially sustainable neighborhoods. © 2012 Elsevier B.V. All rights reserved.
1. Introduction It is widely recognized that urban vegetation provides a variety of important benefits, which can be summarized as follows: (1) social, such as improving social cohesion (Kweon, Sullivan, & Wiley, 1998) and neighborhood vitality (Sullivan, Kuo, & DePooter, 2004); (2) economic, such as improved property values (Tajima, 2003); (3) psychological, such as promoting psychological health and wellbeing (Shin, Yeoun, Yoo, & Shin, 2010); (4) environmental, such as mitigating the urban heat island effect (Yuan & Bauer, 2007), enhancing carbon sequestration (Jo & McPherson, 1995; Nowak
∗ Corresponding author. Tel.: +1 217 244 8658. E-mail addresses: [email protected] (J. Kim), [email protected] (X. Zhou). 0169-2046/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.landurbplan.2012.07.011
& Crane, 2002), providing important habitats for wildlife (Young, 2010), increasing local biodiversity (Kong, Yin, Nakagoshi, & Zong, 2010), and improving air quality (Nowak, Crane, & Stevens, 2006); and (5) physical health, as evidenced by improved health indicators (De Vries, Verheij, Groenewegen, & Spreeuwenberg, 2003). Swanwick, Dunnett, and Woolley (2003) likewise concluded that the presence of green spaces has multiple benefits, such as providing social support and helping maintain urban sustainability and facilitate urban renewal. Urban vegetation structures have been studied in many different ways. Schmid (1975) found that the composition and configuration of vegetation are important in the urban setting. Schmid investigated urban vegetation in Chicago residential areas and found that biomass, species composition, and arrangement of plants varied significantly among neighborhoods with different physical and
J. Kim, X. Zhou / Landscape and Urban Planning 108 (2012) 28–38
social fabrics. Urban vegetation structure has also been proven to be associated with bird fauna distributions (Murgui, 2009), while land cover composition and configuration were shown to affect land surface temperature in urban areas (Zhou, Huang, & Cadenasso, 2011). Within the last two decades, combining ecological and socioeconomic methods has been essential to studies of urban and rural areas (Grove & Burch, 1997; Grove et al., 2006b; Iverson & Cook, 2000; Martin, Warren, & Kinzig, 2004; McDonnell & Pickett, 1990; Troy, Grove, O’Neil-Dunne, Pickett, & Cadenasso, 2007; Zipperer, Sisinni, Pouyat, & Foresman, 1997). A common finding is that household income and level of education are positively correlated with amount of vegetation cover (Grove & Burch, 1997; Grove et al., 2006a; Iverson & Cook, 2000; Troy et al., 2007) and a diversity of urban vegetation (Martin et al., 2004). Iverson and Cook (2000) found wealthy regions have higher tree cover than poor regions. Troy et al. (2007) found a positive relationship between the percentage of African American families and tree canopy cover in an area. Martin et al. (2004) found that vegetation richness in a neighborhood has a strong positive correspondence with socioeconomic status (SES)––that is, neighborhoods with a high SES are more likely than low-SES neighborhoods to have rich vegetation. As urban landscape development policies and codes become more prevalent, studies on the relationship between regulations and vegetation patterns are being conducted. Landry and Pu (2010) found that land development codes were associated with protection of urban forests in residential areas. Robinson and Brown (2009) evaluated land-use development policies on forest cover. Results showed that large lot-size zoning policies led to greater sprawl and increased forest cover. Zhou and Wang (2011) found that the existence of green space policies or regulations had an impact on the process of green space change. Kim and Ellis (2009) found that an ecologically planned neighborhood exhibited a less fragmented forest pattern and that restrictive development guidelines resulted in more ecologically sound environments. Sung (2012) evaluated the efficacy of tree removal permits and found that communities requiring tree removal permits had trees taller than those in communities without such permits. Landscape ecology outlines important principles of green space organization related to the theory of island biogeography (MacArthur & Wilson, 1967). In essence, large patches, high connectivity, and proximity foster species diversity and ecosystem functions. These spatial concepts have been widely adopted in urban landscape architecture and landscape planning projects (Goldstein et al., 1982/1983). Landscapes are distinguished by spatial relations among component parts and can be characterized by both composition and configuration (Turner, 1989). Landscape structure is determined by the flow of materials, animals, energy, and water through the landscape elements of patches, corridors, and matrix. Factors such as patch size and shape, corridor characteristics, and connectivity work together to establish the pattern and process of the landscape (Forman, 1995). Configuration is how the spatial arrangement of patches accounts for their position in relation to the dispersal of individuals through an entire network of patches. Patch geometry describes the attributes of patches such as area, perimeter, and heterogeneity (Borthagaray, Arim, & Marquet, 2012). We use the term “landscape structure” to describe the spatial configuration and geometry of tree canopy in this study. Despite the regulations and socioeconomic factors that shape urban landscape structure, we know relatively little about the relationship between landscape structure and design attributes such as lot size, setback line, ground floor area of buildings, and streets governed by zoning ordinances. Recently, a few studies have considered these aspects, such as street pattern (Conway & Urbani,
29
2007; Stone, 2004), lot size and frontage (Stone, 2004), and zoning class (Wilson, Clay, Martin, Stuckley, & Vedder-Risch, 2003), and found that these attributes are related to urban vegetation conditions. A further dimension that is not well understood but influences landscape structure is the geographic characteristics such as slope, aspect, and elevation. A few studies have been conducted on the relationship among plant communities, landforms, and geomorphic surfaces. Parker and Bendix (1996) examined geomorphic influences on vegetation patterns at the landscape scale and found that the geomorphic processes most strongly linked with vegetation patterns operate. Wondzell, Cunningham, and Bachelet (1996) observed that the correlation of plant communities to landforms and geomorphic surfaces resulted from differences in the redistribution of water and organic matter between landforms. Similarly, Hope et al. (2003) found that elevation is a significant predictor of variation in plant diversity, and Lowry, Baker, and Ramsey (2012) found that neighborhood age is an important covariate that influences how the human and environmental factors relate to the abundance of neighborhood tree canopy. Even though many factors, such as socioeconomic status, geographic condition, built environment, and development regulations, have been used as a means to understand urban vegetation, few comprehensive studies of residential landscape structure have been conducted. This gap in understanding hinders the ability to predict the impacts of future residential development and to design planning strategies that might mitigate negative impacts. We hypothesize that subdivision design attributes governed by zoning and geographic characteristics, however, may exert influence on neighborhood landscape structures along with socioeconomic status and the built environment. Providing numerous quantitative metrics is useful for understanding social and ecological benefits and designing sustainable landscapes, particularly in the urban context, where land resources are limited. Different social groups can derive different social and ecological benefits from the same landscape. The resultant trade-off between benefits of urban vegetation and the issue of safety and management may place a constraint in residential landscape structure. However, the benefits of urban vegetation to residents and to urban-dwelling species are significant; therefore, comprehensively identifying the factors that influence and shape landscape structure in urban areas is thus pivotal. The purpose of this study, therefore, is to investigate a broad range of factors related to residential landscape structure in Morgantown, West Virginia––more specifically, to provide a better understanding of how zoning ordinances and geographic characteristics are related to landscape structure in a hilly residential area at two levels: community and block. The study utilized a core set of landscape metrics that are considered the most useful and relevant for landscape planning and management. A quantitatively robust study to better understand how human and environmental factors are related to residential landscape structure at different scales will provide planners and landscape architects with accurate information they can use to improve neighborhood design and to manage neighborhood landscape more efficiently.
2. Study area West Virginia is located within Appalachian region that follows the ridge of the Appalachian Mountains from southern New York to northern Mississippi. The state is primarily mountainous, with the average elevation at 457 m above sea level, which is the highest of any US state east of the Mississippi River. Most of the smooth surface areas are near large rivers such as the Ohio River or Monongahela River. Extremely irregular in boundary,
30
J. Kim, X. Zhou / Landscape and Urban Planning 108 (2012) 28–38
Fig. 1. (a) Location of Monongalia County and city of Morgantown; (b) zoning map and Morgantown ward boundaries; and (c) the study areas and downtown of Morgantown.
West Virginia has two narrow projections––the northern panhandle, which extends adjacent to Ohio and Pennsylvania; and the eastern panhandle, which is between Maryland and Virginia. These two panhandles are the fastest-growing regions in the state (Fig. 1a). The city of Morgantown is the county seat of Monongalia County. The county has a population of 96,189 (US Census Bureau, 2010) and is located near the Pennsylvania border in the north-central part of West Virginia. This area is one of the major growth areas in the state. Monongalia County is the only north-central West Virginia county to grow in population over the past two decades, and it was one of the fastest-growing counties in the state during the first decade of the 21st century. The current area of Morgantown was settled in the mid-18th century. In 1767, Zachwell Morgan and others made the first permanent settlement at Morgantown. Morgan’s proposed plan in 1785 contained 112 lots, main streets, and smaller alleys. By an act of the legislature, Morgantown was established as town on 50 acres of land belonging to Zachwell Morgan, and he was vested with power to lay out lots for sale and to locate streets (Callahan, 1923). In the late 19th century, a railroad system and the discovery of nearby sand and natural gas stimulated Morgantown’s economic and industrial development. As the city’s population increased, the wealthier citizens moved from the industrialized city center to newly constructed suburbs such as South Park. Suncrest and South Park are adjacent neighborhoods in Morgantown (Fig. 1b and c). They are located approximately 2 km to the north and south, respectively, of downtown Morgantown. South Park, one of the Morgantown’s first suburbs, was developed in early-to-mid 20th century, and part of it is designated as a national
historic district. Prior to development, this area was mostly woods and farmland (Gioulis, 1998). Another area, Suncrest, was incorporated as a town in 1937 and became part of Morgantown in 1949. In late 19th century, the Suncrest area was primarily agricultural land. Like South Park, part of Suncrest is very steep. Biophysical characteristics such as soil, vegetation, and the steep topography of the two areas are very similar. Despite these similarities, Suncrest and South Park have major differences in zoning. The Morgantown city council adopted its first zoning ordinance in 1944. Suncrest and South Park were designated as single-family residential, R-1 and R-1A, respectively (Fig. 1b). R-1 includes most of the territory designated for residential use outside of the city as well as extensive areas within the city. The requirements for the R-1 district with regard to minimum lot size and ground floor area of buildings are greater than those specified for R-1A. The purpose of the R-1 district is to provide attractive, singlefamily neighborhoods for residents who prefer larger lot sizes and are not willing to live in close proximity to other types of land uses. R-1A includes much of the city’s older area, which was platted with lots that do not meet R-1 district development standards. The purpose of the R-1A district is to provide single-family neighborhoods that have smaller lots and are within convenient walking distance to other zoning areas. Both R-1 and R-1A preserve desirable characteristics and protect the areas from change and intrusion that may cause deterioration (City of Morgantown, 2006). Fig. 2 shows the street sections, front yards, and street patterns in the two neighborhoods. In South Park, retaining walls were installed along the street because of the steep slope. Aerial photos of this area also show a dense housing arrangement and a grid street pattern.
J. Kim, X. Zhou / Landscape and Urban Planning 108 (2012) 28–38
31
Fig. 2. Comparison of the study areas: street––the road surface of South Park is paved with various materials, and curb and sidewalk are installed. Front yard––because of the steep slope, retaining walls are installed along the street. Aerial photo––housing density of South Park is high because of its small parcel lot size.
3. Data and methods
3.2. Methods
3.1. Data
As previously described, South Park and Suncrest have similar biophysical characteristics such as soil types, geology, precipitation, and plant species (Fig. 3). Because some blocks in Suncrest and South Park were public open space or existing natural areas and the main focus of this study is the neighborhood environment, those blocks were excluded. This study was carried out at two geographic levels––community level and census block level. Community level was used for general comparison of the two communities at a landscape scale, while census block level was selected for examination of socioeconomic factors and characteristics related to the physical setting of a residential area. The number of selected blocks was 66 in Suncrest and 139 in South Park. For each level, methods were further divided into two components: data preparation and data interpretation (Fig. 4).
Data used in this study included color-infrared ortho-rectified aerial photographs at a resolution of 1 m; computer-aided design (CAD) drawings of existing structures; a digital elevation model (DEM); geographic information system (GIS) files including parcel, census block, and road network; the Morgantown city code; and demographic information. Images were obtained from the West Virginia GIS Technical Center. CAD files obtained from the Morgantown Utility Board were used to extract road networks and building footprints. The derived buildings and road layers were exported to ArcGIS 9.3 for further analysis. The DEM was downloaded from the GIS Technical Center and used to calculate local elevation, slope, and aspect. Zoning ordinances for the two communities were obtained from Morgantown city and planning and zoning codes. Parcel and road shapefiles were obtained from City of Morgantown Planning Services. Demographic information, including ethnic group, educational level, and income, was obtained from the US Census Bureau.
3.2.1. Data preparation at the community level Ortho-rectified aerial photography, DEM, CAD data, and parcel shapefiles were processed to acquire information about landscape
32
J. Kim, X. Zhou / Landscape and Urban Planning 108 (2012) 28–38
Fig. 3. Clipped aerial photo with census block boundary.
structure, parcel attributes, and geographic characteristics at the community level. Ortho-rectified aerial photographs were georeferenced to the Universal Transverse Mercator (UTM) coordinate system using the ERDAS IMAGINE 9.2 software package. Eight control points were carefully selected from Google Earth and used to geometrically correct images using a second-order polynomial transformation and the nearest-neighborhood resampling method. The root mean squared error was controlled within 0.5 pixels. Because computerassisted automatic classification might not attain the desired standard of accuracy, the tree canopy was manually digitized (Fig. 5). The identified tree canopy and the remaining non-woody areas were exported in raster format for the calculation of landscape metrics in FRAGSTATS 3.3 (McGarigal, Ene, & Holmes, 2002). Six landscape metrics were used to quantify landscape structure (Table 1):
1. Percentage of Landscape (PLAND) is a general index that depicts the relative abundance of vegetated areas. 2–3. Patch Density (PD) and Edge Density (ED) are sensitive to landscape fragmentation. 4. Mean Patch Size (MPS) can be used to illustrate landscape dominance and fragmentation. 5. Landscape Shape Index (LSI) is associated with landscape complexity. 6. Euclidean Nearest-Neighbor Distance (ENN) represents the degree of patch isolation, using Euclidean geometry as the shortest crow-fly distance between the same types of patches.
The DEM was clipped in ArcGIS 9.3 and then mosaicked based on the boundary of the two communities. The nearest-neighborhood resampling method was used to resample the created DEM to a
Fig. 4. Flow chart of the analysis.
J. Kim, X. Zhou / Landscape and Urban Planning 108 (2012) 28–38
33
Fig. 5. Digitized woody areas of Suncrest and South Park.
30-m resolution. Spatial analysis was conducted to calculate elevation, slope, and aspect. Building footprints were extracted from the CAD map and exported to ArcGIS 9.3 in shapefile format. The created shapefile was spatially adjusted to the UTM coordinate system and overlapped with digital orthophoto quarter quad images. The parcel and road layers were projected onto the UTM coordinate system. To compare road physical patterns in the two communities, road network connectivity was measured using the network analysis tool in ArcGIS 9.3 (Fig. 6). Intersection and edge numbers were calculated. Beta (Eq. (1)) and gamma (Eq. (2)) network indices were computed based on derived edge numbers and intersections:
ˇ=
=
e
(1)
v e 3(v − 2)
(2)
where e is edge number and v is node number.
3.2.2. Data preparation at the block level Landscape structure, parcel attributes, and geographic characteristics were disaggregated for each block. Geoprocessing, based on Python programming, was conducted to calculate block-level landscape metrics. Slope and elevation were averaged within each block. Aspect for each block was determined from the dominant aspect direction. Building footprints and lot sizes were calculated by averaging lot and building areas within individual blocks. 3.2.3. Data interpretation at the community level At the community level, the difference in landscape structure between the two communities may be attributed to demographic characteristics and road patterns. Therefore, demographic characteristics and road patterns in Suncrest and South Park were compared. Association among landscape structure, parcel lot attributes, and geographic characteristics was also investigated. 3.2.4. Data interpretation at the block level To confirm the association of landscape structure, parcel attributes, and geographic characteristics, statistical analysis at the block level was conducted. A t-test of landscape metrics in both
Table 1 Landscape metrics used in this study. Metric
Description
Unit
Range
Justification
Percentage of Landscape (PLAND) Patch Density (PD)
Proportion of the area of a specific land-use class to the entire landscape area Number of patches in a specific class, divided by the entire landscape area Sum of the lengths of all edge segments involving the corresponding patch type, divided by total landscape area, multiplied by 10,000 Sum of the areas of all patches of the corresponding patch type, divided by the number of patches of the same type, divided by 10,000 Patch perimeter divided by the minimum perimeter possible for a maximally compact patch of the corresponding patch area The distance to the nearest neighboring patch of the same type, based on shortest edge-to-edge distance
Percent
0 < PLAND ≤ 100
General index
Number per 100 ha
PD > 0
Index of fragmentation
Meters per hectare
ED ≥ 0
Index of fragmentation
Hectares
MPS > 0
Index of fragmentation
None
LSI ≥ 1
Index of shape
Meters
ENN > 0
Index of proximity
Edge Density (ED)
Mean Patch Size (MPS)
Landscape Shape Index (LSI) Euclidean Nearest-Neighbor Distance (ENN)
34
J. Kim, X. Zhou / Landscape and Urban Planning 108 (2012) 28–38
Fig. 6. Results of road network analysis for Suncrest and South Park.
communities was carried out. Correlations between landscape metrics and geographic and ordinance variables were analyzed for Suncrest and South Park. 4. Results 4.1. Results at the community level 4.1.1. Demographics and road patterns To examine the potential influences of demographics and road patterns on landscape structure, demographics (ethnic group, educational level, and income) were compared for the two communities (Table 2), and beta and gamma connectivity indices were calculated (Table 3). Results of demographic comparisons showed that people residing in these two communities generally had a similar socioeconomic status. Likewise, the composition of ethnic groups was similar. Beta and gamma indices indicated that road connectivity in South Park was slightly higher than in Suncrest, but these communities generally had similar connectivity levels (Fig. 5). Table 2 Demographic comparison between Suncrest and South Park. Element
Category
Suncrest (%)
South Park (%)
Race
White Black Asian Hispanic K-12 High school College <30K 30–50K 50–75K 75–200K
90.37 1.85 5.26 1.2 7.58 15.30 77.11 31.76 27.05 28.7 12.49
90.25 3.64 2.2 1.45 9.92 12.31 77.77 50.63 23.99 16.56 8.82
Education
Income
Table 4 Landscape metrics of Suncrest and South Park at the community level.
Table 3 Results of road network analysis.
Suncrest South Park
4.1.2. Landscape structure, geographic attributes, and parcel attributes At the community level, PLAND indicated that vegetated areas in South Park were more abundant than in Suncrest. PD and ED were much greater in South Park than Suncrest, while the MPS of South Park was lower than that of Suncrest. These findings provided evidence that vegetated areas in South Park were more fragmented than in Suncrest. Both communities had similar LSI values, indicating that landscape shape did not significantly differ between them (Table 4). Elevation in both communities was similar, while slope differed considerably (Table 5). More than half of Suncrest’s area had a slope percentage of less than 25%, while nearly 70% of land in South Park had a slope percentage of more than 25%. Overall, South Park was much steeper than Suncrest (Fig. 7). As shown in Table 6, minimum lot size and minimum lot frontage of zone R-1 were approximately twice as large as those of zone R-1A. Minimum lot size of R-1 and R-1A was 7200 ft2 (2194 m2 ) and 3500 ft2 (1067 m2 ), respectively; and minimum lot frontage was 70 ft (21 m) in R-1 and 30 ft (9 m) in R-1A. Maximum lot coverages in R-1 and R-1A were 40% and 50%, respectively. Setback lines in R-1 were also wider than those of R-1A. Based on these results, demographics and road patterns were determined not to be likely reasons for the differences of landscape metrics. Differences in geography and ordinances may, however, account for the distinctive landscape structure. The association between landscape structure and geographic and parcel attributes was further investigated at the block level.
Vertex
Edge
Beta
Gamma
378 480
436 618
1.153 1.287
0.386 0.430
Metric
Suncrest
South Park
PD ED MPS LSI ENN PLAND
209.62 383.52 0.093 1.467 18.06 19.67
318.14 478.49 0.081 1.337 16.53 25.77
PD, Patch Density; ED, Edge Density; MPS, Mean Patch Size; LSI, Landscape Shape Index; ENN, Euclidian Nearest Neighbor; PLAND, Percentage of Landscape.
J. Kim, X. Zhou / Landscape and Urban Planning 108 (2012) 28–38
35
Fig. 7. Slope and aspect analysis for Suncrest and South Park.
Table 5 Geographic characteristics of Suncrest and South Park.
4.2. Results at the block level
Element
Category
Suncrest
South Park
Elevation
Low High 0–2% 2–5% 5–12% 12–25% >25% N NE E SE S SW W NW
939 ft 1198 ft 2.27% 8.28% 19.54% 25.94% 43.97% 12.35% 10.04% 5.93% 2.95% 16.16% 21.25% 16.22% 15.09%
816 ft 1259 ft 0.87% 2.90% 7.65% 18.97% 69.58% 27.67% 9.60% 4.98% 5.59% 9.30% 7.58% 8.26% 26.13%
Slope
Aspect
4.2.1. Differences in landscape structure At the block level, PD and PLAND for South Park were 7.75% and 4.83% higher, respectively, than those for Suncrest area, while ED and ENN for South Park were lower than for Suncrest (Table 7). A t-test indicated that these metrics significantly differed between the two communities. Results confirmed that, although vegetated areas in South Park were more abundant than in Suncrest, they tended to be more fragmented and isolated compared to Suncrest at the block level. MPS and LSI did not significantly differ between the two places.
4.2.2. Differences in geographic and parcel attributes As for the geographic and parcel attributes at the block level, all values showed significant differences between the two sites.
36
J. Kim, X. Zhou / Landscape and Urban Planning 108 (2012) 28–38
Table 6 Zoning ordinances of Suncrest and South Park. Zoning
Parcel lot
Suncrest (R-1) South Park (R-1A)
Setback (ft)
Min. lot size (sf)
Min. lot frontage (ft)
Max. lot cover (%)
Min. front
Max. front
Min. rear
Min. side
7200 3500
70 30
40 50
25 8
30 20
25 20
10 5
Lot sizes and perimeters in South Park were smaller than those in Suncrest, and building footprints in South Park were about 26.45% smaller than those in Suncrest. The average slope of South Park was 32.45% steeper than that of Suncrest; however, average elevation of South Park was lower (Table 8). 4.2.3. Association between landscape structure and geographic and parcel attributes Associations between landscape structure and geographic and parcel attributes were analyzed in Suncrest and South Park. Parcel attributes in Suncrest area revealed a strong relationship with PD, MPS, and PLAND. Results indicated that ordinances were associated with landscape structure in terms of spatial heterogeneity and fragmentation. Parcel lot size was positively correlated with PLAND, indicating that larger lot sizes provide more space to develop vegetation cover. On the other hand, parcel lot size was negatively correlated with PD, suggesting that fragmentation is less likely to happen where parcel sizes are larger. However, geographic variables did not show an obvious relationship with landscape metrics in Suncrest (Table 9). Lot sizes and perimeters in South Park were also significantly correlated with PLAND and negatively correlated with PD. However, the association between PD and lot size and perimeter in South Park was less significant than the association between those factors in the Suncrest area. Geographic attributes were associated with MPS, indicating that steeper and higher locations had more large areas of vegetation and might have more beneficial microclimate condition for plants. LSI was correlated with slope, suggesting that irregularly shaped vegetated patches were more distributed in blocks with steeper slopes (Table 10). Table 7 Landscape class metrics of Suncrest and South Park at the block level, with p values for statistical difference by t-test. Metric
PD ED MPS LSI ENN PLAND
Suncrest
South Park
p
Mean
SD
Mean
SD
792.27 633.03 .040 1.444 8.47 25.19
317.50 183.84 .034 .175 3.91 9.85
858.88 593.28 .044 1.443 8.31 26.47
541.62 232.63 .045 .182 5.99 13.12
.003 .044 .125 .585 .001 .007
PD, Patch Density; ED, Edge Density; MPS, Mean Patch Size; LSI, Landscape Shape Index; ENN, Euclidian Nearest Neighbor; PLAND, Percentage of Landscape.
Table 8 Difference between zoning ordinances and geographic characteristics in Suncrest and South Park. Metric
AREA (sf) PERI (ft) BLDG (sf) SLOPE (%) ELEV (ft)
Suncrest
South Park
p
Mean
SD
Mean
SD
13563.40 465.77 1877.43 30.34 1003.83
5493.07 85.95 503.50 17.95 35.69
4220.90 268.90 1484.66 44.92 974.28
1815.38 51.91 267.51 27.53 106.08
.000 .000 .000 .004 .000
AREA, parcel lot size; PERI, parcel lot perimeter; BLDG, building footprint; SLOPE, slope; ELEV, elevation.
5. Discussion The results of this study suggest a direct link between landscape structure and land development regulations that govern parcel attributes in residential areas. Higher-density patterns of single-family residential development are associated with greater vegetation abundance, but those areas also have a more fragmented landscape structure than similar developments with lower-density patterns. Geographic slope also influences landscape structure in hilly areas. 5.1. Influences of ordinances and geographic attributes on landscape structure Community-level analyses of landscape structure reveal the different landscape structures between South Park and Suncrest. South Park has more vegetated areas, but those areas tend to be more fragmented than in Suncrest. The Mean Patch Size of vegetation is smaller in South Park. Zoning ordinances in South Park are less restrictive than those in Suncrest: minimum lot size and minimum lot frontage requirements are half those of Suncrest. Moreover, the maximum lot coverage in South Park is 10% more than that in Suncrest. Results indicate that South Park, the higher-density community, which has smaller lot sizes and shorter lot frontages, has more vegetated areas. This result is related to the findings of Stone (2004) that higher density patterns of single-family development are associated with a smaller area of Table 9 Correlation coefficients of landscape metrics with geographic and ordinance attributes in Suncrest. Metric
PD
ED
AREA PERI BLDG SLOPE ELEV
−.430 −.416** −.250* −.023 .023 **
.048 .074 .152 .118 .230
MPS **
.516 .527** .277* .042 .058
LSI
ENN
PLAND
.214 .215 .234 −.046 .048
.001 −.012 −.079 .096 −.194
.383** .394** .213 .194 .164
PD, Patch Density; ED, Edge Density; MPS, Mean Patch Size; LSI, Landscape Shape Index; ENN, Euclidian Nearest Neighbor; PLAND, Percentage of Landscape; AREA, parcel lot size; PERI, parcel lot perimeter; BLDG, building footprint; SLOPE, slope; ELEV, elevation. * p < .05. ** p < .01.
Table 10 Correlation coefficients of landscape metrics with geographic and ordinance attributes in South Park. Metric
PD
ED
MPS
LSI
ENN
PLAND
AREA PERI BLDG SLOPE ELEV
−.192* −.197* −.075 .024 −.139
.085 .070 −.037 −.050 −.134
.386** .380** .166 .178* .198*
.145 .165 .068 .174* .050
−.046 −.065 .046 .058 −.102
.454** .419** .165 −.016 .038
PD, Patch Density; ED, Edge Density; MPS, Mean Patch Size; LSI, Landscape Shape Index; ENN, Euclidian Nearest Neighbor; PLAND, Percentage of Landscape; AREA, parcel lot size; PERI, parcel lot perimeter; BLDG, building footprint; SLOPE, slope; ELEV, elevation. * p < .05. ** p < .01.
J. Kim, X. Zhou / Landscape and Urban Planning 108 (2012) 28–38
impervious cover per unit of occupancy. With respect to geographic attributes, South Park is hillier than Suncrest. This attribute might be related to microclimate that is beneficial to plants and could provide more opportunities for vegetation to grow. This result conforms to the findings of Lowry et al. (2012) that slope and elevation are positively associated with mean annual precipitation, which would affect tree canopy abundance. Block-level analyses further confirm the association of ordinances and geographic attributes on landscape structure. Vegetation abundance, distribution, and fragmentation are significantly different in the two communities, while the attributes of landscape shape and complexity do not have an evident impact. In conformance with zoning ordinances, lot sizes and building footprints at the block level are smaller in South Park. This result indicates that land use related ordinances play a critical role in parcel layout and vegetation patterns in residential areas. Correlation analyses in Suncrest and South Park show a positive association between PLAND and parcel size, suggesting that larger parcel size is related to higher levels of vegetation. Moreover, parcel size and perimeter are positively and negatively associated with MPS and PD, respectively, indicating that larger parcels provide more opportunities to maintain vegetation dominance. The stronger association between parcel size and MPS in Suncrest suggests that larger lots, as required by zoning ordinances, play a greater role in vegetation dominance. Additionally, although building footprints differ between the two study areas, that factor does not appear to be related to vegetation patterns. As for geographic attributes, slope does not markedly affect landscape structure in Suncrest, but it is related to MPS in South Park. This finding suggests that geographic attributes may affect vegetation fragmentation, but do so more directly in hilly than in less hilly places. 5.2. Application of landscape metrics in residential vegetation studies The presence of vegetation in a residential neighborhood is important to the people who live there, but spatial composition and configuration of that vegetation are also crucial. A well-designed neighborhood landscape structure can increase the benefits of green space to humans and––equally important––to urban wildlife. Less-fragmented areas of vegetation can provide viable wildlife habitat, while a more-complex landscape structure may promote urban biodiversity. The importance of ecologically sound neighborhood planning is becoming increasingly acknowledged by local planners. Accordingly, quantifying ecological indicators to guide the process of neighborhood planning is beneficial. Landscape metrics are useful in several planning phases, analysis, prognosis, and synthesis (Leitão & Ahern, 2002). They can provide useful directions for planning and also provide comparative measurements and information to inform insight about landscape structure–function relationships. In this context, landscape ecology-based metrics provide a means to measure and assess the spatial structures of landscape elements in terms of size, number, shape, and spatial arrangement of land elements (Turner, Gardner, & O’Neill, 2001). Characteristics of land form can be quantified by landscape metrics with respect to dominance, fragmentation, connectivity, diversity, and isolation (McGarigal et al., 2002). Our study investigated landscape structures in two communities and demonstrated the usefulness of landscape metrics in capturing the neighborhood landscape composition and configurations. 5.3. Geographic scale This study analyzed landscape structure at two scales: community and block. At the community scale, qualitative comparisons
37
between synoptic landscape structure and geographic and parcel attributes were conducted. An overall understanding was achieved about the associations between landscape structure and the potential factors influencing those associations. At the block scale, indicators related to landscape metrics, ordinances, and geographic attributes were disaggregated to finer units. Analyses at the block level strengthened the understanding at the community level; the block-level analyses also showed the degree of association and the representation of zoning ordinances in each block. This study also demonstrated that landscape metrics could capture vegetation patterns at scales as small as the block level.
5.4. Future research Our study compared two communities that differ in terms of zoning requirements but otherwise are very similar. We recommend future work comparing a greater number of communities to corroborate the current findings. In addition, future research should divide land use into finer categories than the two types (woody and non-woody) examined in this study, classifying vegetated areas into types such as grasslands, shrubs, and tree canopies in order to examine the effects of ordinances and geographic attributes on landscape structure on each vegetation type. Finally, because this study examined only zoning types R-1 and R-1A, we recommend that future research include more zoning categories.
6. Conclusion This study investigated the relationship among landscape structure, zoning ordinances, and geographic attributes in hilly residential areas. Its specific purpose was to better understand how zoning ordinances and geographic characteristics are related to landscape structure in a hilly residential area at two levels: community level and block level. Two residential areas with different residential zoning ordinances, but with similar development history and socioeconomic status, were examined. Six core landscape metrics were selected for calculation of landscape structure. The findings indicate that differences in zoning requirements result in distinct landscape structure––specifically, that zoning requirements for smaller minimum lot size and lot frontage result in more vegetation but also result in more fragmentation than requirements that call for greater minimum lot size and lot frontage. Although building footprint is weakly associated with landscape structure, lot size and perimeter are strongly related to vegetation abundance, fragmentation, and dominance. As for geographic characteristics, slope is associated with patch size, especially in hilly communities. The findings here suggest that landscape ecology-based metrics are efficient ecological indicators to guide the process of residential neighborhood planning. These metrics are beneficial for understanding how elements of local development regulations and geographic characteristics affect landscape structure. This understanding is important for landscape architects, planners, and administrators because it can lead to better strategies for planning and designing socially and ecologically healthy environments. Appropriate zoning ordinances and geographic attributes jointly contribute to sound neighborhood landscape structures. Applying the elements of local ordinances and geographic characteristics that affect landscape structure will likely play an important role for landscape architects, planners, and administrators in developing and amending community land-use regulations.
38
J. Kim, X. Zhou / Landscape and Urban Planning 108 (2012) 28–38
Acknowledgements The research described in this article has been funded in part by the United States Department of Agriculture with Cooperative State Research, Education and Extension Service WVA0049 2H-Hatch. The authors greatly appreciate the thorough review and valuable comments of professor William Sullivan. The authors also thank the efforts of three anonymous reviewers and editor for assisting in the publication of our research. References Borthagaray, A. I., Arim, M., & Marquet, P. A. (2012). Connecting landscape structure and patterns in body size distributions. Oikos, 121, 697–710. Callahan, J. M. (1923). History of West Virginia old and new. Chicago and New York: The American Historical Society, Inc. City of Morgantown, WV. (2006). Morgantown city code. Retrieved from http://www.morgantown.com/planning-docs/PlanZonCode as-of-09-012007.pdf Conway, T. M., & Urbani, L. (2007). Variations in municipal urban forestry policies: A case study of Toronto, Canada. Urban Forestry Urban Greening, 6, 181–192. De Vries, S., Verheij, R., Groenewegen, P., & Spreeuwenberg, P. (2003). Natural environments––Healthy environments? An exploratory analysis of the relationship between greenspace and health. Environment and Planning, 35(A), 1717–1731. Forman, R. T. T. (1995). Land mosaics––The ecology of landscapes and regions. New York: Cambridge University Press. Gioulis, M. (1998). Historic resource survey of the Morgantown Riverfront of Monongalia County West Virginia. Morgantown, WV: Morgantown Historic Landmarks Commission. Goldstein, E. L., Gross, M., & Degraaf, R. M. (1982/1983). Wildlife and greenspace planning in medium-scale residential developments. Urban Ecology, 7, 201–214. Grove, J. M., & Burch, W. R., Jr. (1997). A social ecology approach to urban ecosystem and landscape analyses. Journal of Urban Ecosystems, 1(4), 259–275. Grove, J. M., Cadenasso, M. L., Burch, W. R., Jr., Pickett, S. T. A., Schwarz, K., O’NeilDunne, J., et al. (2006). Data and methods comparing social structure and vegetation structure of urban neighborhoods in Baltimore, Maryland. Society and Natural, 19, 117–136. Grove, J. M., Troy, A. R., O’Neil-Dunne, J. P. M., Burch, W. R., Jr., Cadenasso, M. L., & Pickett, S. T. A. (2006). Characterization of households and its implications for vegetation of urban ecosystems. Ecosystems, 9, 578–597. Hope, D., Gries, C., Zhu, W., Fagan, W. F., Redman, C. L., Grimm, N. B., et al. (2003). Socioeconomics drive urban plan diversity. Proceedings of the National Academy of Sciences of the United States of America, 100(15), 8788–8792. Iverson, L. R., & Cook, E. A. (2000). Urban forest cover of the Chicago region and its relation to household density and income. Urban Ecosystems, 4, 105–124. Jo, H. K., & McPherson, E. G. (1995). Carbon storage and flux in urban residential greenspace. Journal of Environmental Management, 45, 109–133. Kim, J., & Ellis, C. D. (2009). Determining the effects of local development regulations on landscape structure: Comparison of The Woodlands and North Houston, TX. Landscape and Urban Planning, 92, 293–303. Kong, F., Yin, H., Nakagoshi, N., & Zong, Y. (2010). Urban green space network development for biodiversity conservation: Identification based on graph theory and gravity modeling. Landscape and Urban Planning, 95(1–2), 16–27. Kweon, B.-S, Sullivan, W. C., & Wiley, A. R. (1998). Green common spaces and the social interaction of inner-city older adults. Environment and Behavior, 30(6), 832–858. Landry, S., & Pu, R. (2010). The impact of land development regulation on residential tree cover: An empirical evaluation using high-resolution IKONOS imagery. Landscape and Urban Planning, 94, 94–104. Leitão, A. B., & Ahern, J. (2002). Applying landscape ecological concepts and metrics in sustainable landscape planning. Landscape and Urban Planning, 59, 65–93. Lowry, J. H., Jr., Baker, M. E., & Ramsey, D. (2012). Determinants of urban tree canopy in residential neighborhoods: Household characteristics, urban form, and the geophysical landscape. Urban Ecosystems, 15(1), 247–266. MacArthur, R. H., & Wilson, E. O. (1967). The theory of island biogeography. Princeton, NJ: Princeton University Press.
Martin, C. A., Warren, P. S., & Kinzig, A. P. (2004). Neighborhood socioeconomic status is a useful predictor of perennial landscape vegetation in residential neighborhoods and embedded small parks of Phoenix, AZ. Landscape and Urban Planning, 69, 355–368. McDonnell, M. J., & Pickett, S. T. A. (1990). Ecosystem structure and function along urban rural gradients: An unexploited opportunity for ecology. Ecology, 71(4), 1232–1237. McGarigal, K., Ene, E., & Holmes, C. (2002). FRAGSTATS (Version 3): FRAGSTATSMetrics. University of Massachusetts-Produced Program. Available at the following website: http://www.umass.edu/landeco/research/fragstats/documents/ fragstats documents.html Murgui, E. (2009). Influence of urban landscape structure on bird fauna: A case study across seasons in the city of Valencia (Spain). Urban Ecosystems, 12(3), 249–263. Nowak, D. J., & Crane, D. E. (2002). Carbon storage and sequestration by urban trees in the USA. Environmental Pollution, 116, 381–389. Nowak, D. J., Crane, D. E., & Stevens, J. C. (2006). Air pollution removal by urban trees and shrubs in the United States. Urban Forestry Urban Greening, 4, 115–123. Parker, K. C., & Bendix, J. (1996). Landscape-scale geomorphic influences on vegetation patterns in four selected environments. Physical Geography, 17, 113–141. Robinson, D. T., & Brown, D. G. (2009). Evaluating the effects of land-use development policies on ex-urban forest cover: An integrated agent-based GIS approach. International Journal of Geographic Information Science, 23(9), 1211–1232. Schmid, J. A. (1975). Urban vegetation: A review and Chicago case study Department of Geography Research Paper No. 161. Chicago: University of Chicago. Shin, W., Yeoun, P., Yoo, R., & Shin, C. (2010). Forest experience and psychological health benefits: The state of the art and future prospect in Korea. Environmental Health and Preventive Medicine, 15(1), 38–47. Stone, B., Jr. (2004). Paving over paradise: How land use regulations promote residential imperviousness. Landscape and Urban Planning, 69, 101–113. Sullivan, W. C., Kuo, F. E., & DePooter, S. F. (2004). The fruit of urban nature––Vital neighborhood spaces. Environment and Behavior, 36(5), 678–700. Sung, C. Y. (2012). Evaluating the efficacy of a local tree protection policy using LiDAR remote sensing data. Landscape and Urban Planning, 104, 19–25. Swanwick, C., Dunnett, N., & Woolley, H. (2003). Nature, role and value of green space in towns and cities: An overview. Built Environment, 29(2), 94–106. Tajima, K. (2003). New estimates of the demand for urban green space: Implications for valuing the environmental benefits of Boston’s Big Dig project. Journal of Urban Affairs, 25, 641–655. Troy, A. R., Grove, J. M., O’Neil-Dunne, J. P. M., Pickett, S. T. A., & Cadenasso, M. L. (2007). Predicting opportunities for greening and patterns of vegetation on private urban lands. Journal of Environmental Management, 40, 394–412. Turner, M. G. (1989). Landscape ecology: The effect of pattern on process. Annual Review of Ecology and Systematics, 20, 171–197. Turner, M. G., Gardner, R. H., & O’Neill, R. V. (2001). Landscape ecology in theory and practice: Pattern and process. New York: Springer. US Census Bureau. (2010). http://be.wvu.edu/demographics/census.htm (accessed on 10.04.11). West Virginia GIS. West Virginia GIS Technical Center. http://wvgis.wvu.edu/ data/dataset.php (accessed on 15.10.10). Wilson, J. S., Clay, M., Martin, E., Stuckley, D., & Vedder-Risch, K. (2003). Evaluating environmental influences of zoning in urban ecosystems with remote sensing. Remote Sensing of Environment, 86, 303–321. Wondzell, S. M., Cunningham, G. L., & Bachelet, D. (1996). Relationships between landforms, geomorphic processes, and plant communities on a watershed in northern Chihuahuan Desert. Landscape Ecology, 11(6), 351–362. Young, R. F. (2010). Managing municipal green space for ecosystem services. Urban Forestry and Urban Greening, 9(4), 313–321. Yuan, F., & Bauer, M. E. (2007). Comparison of impervious surface area and normalized difference vegetation index as indicators of surface urban heat island effects in Landsat imagery. Remote Sensing of Environment, 106, 375–386. Zhou, W., Huang, G., & Cadenasso, M. L. (2011). Does spatial configuration matter? Understanding the effects of land cover pattern on land surface temperature in urban landscapes. Landscape and Urban Planning, 102, 54–63. Zhou, X., & Wang, Y. (2011). Spatial–temporal dynamics of urban green space in response to rapid urbanization and greening policies. Landscape and Urban Planning, 100, 268–277. Zipperer, W. C., Sisinni, S. M., Pouyat, R. V., & Foresman, T. W. (1997). Urban tree cover: An ecological perspective. Urban Ecosystems, 1, 229–246.
Contents lists available at SciVerse ScienceDirect
Landscape and Urban Planning journal homepage: www.elsevier.com/locate/landurbplan
Research paper
Landscape structure, zoning ordinance, and topography in hillside residential neighborhoods: A case study of Morgantown, WV Jinki Kim ∗ , Xiaolu Zhou University of Illinois, Department of Landscape Architecture, Fine & Applied Arts, 611 Taft Drive, Champaign, IL 61820, United States
h i g h l i g h t s Landscape structures are significantly associated with lot attributes governed by zoning ordinance. Zoning requirements for smaller minimum lot size and lot frontage result in more vegetation but also cause more fragmentation. Geographic slope influences landscape structure in hilly residential neighborhood.
a r t i c l e
i n f o
Article history: Received 23 January 2012 Received in revised form 17 July 2012 Accepted 24 July 2012 Available online 11 August 2012 Keywords: Landscape structure Zoning ordinance Geographic characteristics Single-family parcel Landscape fragmentation
a b s t r a c t The purpose of this study is to investigate the relationship among landscape structure, zoning ordinance, and geographic attributes in Morgantown, West Virginia. Two residential areas zoned by different standards, but with similar development history and socioeconomic status were examined. Aerial photography was digitized into woody and non-woody areas and the landscape structure was quantified with six core landscape metrics related to fragmentation. Computer aided design (CAD) map and digital elevation model (DEM) were used for the calculation of building footprint and geographical characteristics respectively. The findings indicate that differences in zoning requirements result in distinct landscape structure. Zoning requirements for smaller minimum lot size and lot frontage result in more vegetation but also result in more fragmentation than requirements that call for greater minimum lot size and lot frontage. Although building footprint is weakly associated with landscape structure, lot size and lot perimeter are strongly related to vegetation abundance, fragmentation, and dominance. Slope is associated with vegetation patch size, especially in hilly residential neighborhoods. This study provides a better understanding of how human and environmental factors are related to residential landscape structure at different scale. Providing numerous quantitative metrics is useful for understanding social and ecological benefits. The results of this study will help planners, landscape architects, and administrators for planning and designing more ecologically and socially sustainable neighborhoods. © 2012 Elsevier B.V. All rights reserved.
1. Introduction It is widely recognized that urban vegetation provides a variety of important benefits, which can be summarized as follows: (1) social, such as improving social cohesion (Kweon, Sullivan, & Wiley, 1998) and neighborhood vitality (Sullivan, Kuo, & DePooter, 2004); (2) economic, such as improved property values (Tajima, 2003); (3) psychological, such as promoting psychological health and wellbeing (Shin, Yeoun, Yoo, & Shin, 2010); (4) environmental, such as mitigating the urban heat island effect (Yuan & Bauer, 2007), enhancing carbon sequestration (Jo & McPherson, 1995; Nowak
∗ Corresponding author. Tel.: +1 217 244 8658. E-mail addresses: [email protected] (J. Kim), [email protected] (X. Zhou). 0169-2046/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.landurbplan.2012.07.011
& Crane, 2002), providing important habitats for wildlife (Young, 2010), increasing local biodiversity (Kong, Yin, Nakagoshi, & Zong, 2010), and improving air quality (Nowak, Crane, & Stevens, 2006); and (5) physical health, as evidenced by improved health indicators (De Vries, Verheij, Groenewegen, & Spreeuwenberg, 2003). Swanwick, Dunnett, and Woolley (2003) likewise concluded that the presence of green spaces has multiple benefits, such as providing social support and helping maintain urban sustainability and facilitate urban renewal. Urban vegetation structures have been studied in many different ways. Schmid (1975) found that the composition and configuration of vegetation are important in the urban setting. Schmid investigated urban vegetation in Chicago residential areas and found that biomass, species composition, and arrangement of plants varied significantly among neighborhoods with different physical and
J. Kim, X. Zhou / Landscape and Urban Planning 108 (2012) 28–38
social fabrics. Urban vegetation structure has also been proven to be associated with bird fauna distributions (Murgui, 2009), while land cover composition and configuration were shown to affect land surface temperature in urban areas (Zhou, Huang, & Cadenasso, 2011). Within the last two decades, combining ecological and socioeconomic methods has been essential to studies of urban and rural areas (Grove & Burch, 1997; Grove et al., 2006b; Iverson & Cook, 2000; Martin, Warren, & Kinzig, 2004; McDonnell & Pickett, 1990; Troy, Grove, O’Neil-Dunne, Pickett, & Cadenasso, 2007; Zipperer, Sisinni, Pouyat, & Foresman, 1997). A common finding is that household income and level of education are positively correlated with amount of vegetation cover (Grove & Burch, 1997; Grove et al., 2006a; Iverson & Cook, 2000; Troy et al., 2007) and a diversity of urban vegetation (Martin et al., 2004). Iverson and Cook (2000) found wealthy regions have higher tree cover than poor regions. Troy et al. (2007) found a positive relationship between the percentage of African American families and tree canopy cover in an area. Martin et al. (2004) found that vegetation richness in a neighborhood has a strong positive correspondence with socioeconomic status (SES)––that is, neighborhoods with a high SES are more likely than low-SES neighborhoods to have rich vegetation. As urban landscape development policies and codes become more prevalent, studies on the relationship between regulations and vegetation patterns are being conducted. Landry and Pu (2010) found that land development codes were associated with protection of urban forests in residential areas. Robinson and Brown (2009) evaluated land-use development policies on forest cover. Results showed that large lot-size zoning policies led to greater sprawl and increased forest cover. Zhou and Wang (2011) found that the existence of green space policies or regulations had an impact on the process of green space change. Kim and Ellis (2009) found that an ecologically planned neighborhood exhibited a less fragmented forest pattern and that restrictive development guidelines resulted in more ecologically sound environments. Sung (2012) evaluated the efficacy of tree removal permits and found that communities requiring tree removal permits had trees taller than those in communities without such permits. Landscape ecology outlines important principles of green space organization related to the theory of island biogeography (MacArthur & Wilson, 1967). In essence, large patches, high connectivity, and proximity foster species diversity and ecosystem functions. These spatial concepts have been widely adopted in urban landscape architecture and landscape planning projects (Goldstein et al., 1982/1983). Landscapes are distinguished by spatial relations among component parts and can be characterized by both composition and configuration (Turner, 1989). Landscape structure is determined by the flow of materials, animals, energy, and water through the landscape elements of patches, corridors, and matrix. Factors such as patch size and shape, corridor characteristics, and connectivity work together to establish the pattern and process of the landscape (Forman, 1995). Configuration is how the spatial arrangement of patches accounts for their position in relation to the dispersal of individuals through an entire network of patches. Patch geometry describes the attributes of patches such as area, perimeter, and heterogeneity (Borthagaray, Arim, & Marquet, 2012). We use the term “landscape structure” to describe the spatial configuration and geometry of tree canopy in this study. Despite the regulations and socioeconomic factors that shape urban landscape structure, we know relatively little about the relationship between landscape structure and design attributes such as lot size, setback line, ground floor area of buildings, and streets governed by zoning ordinances. Recently, a few studies have considered these aspects, such as street pattern (Conway & Urbani,
29
2007; Stone, 2004), lot size and frontage (Stone, 2004), and zoning class (Wilson, Clay, Martin, Stuckley, & Vedder-Risch, 2003), and found that these attributes are related to urban vegetation conditions. A further dimension that is not well understood but influences landscape structure is the geographic characteristics such as slope, aspect, and elevation. A few studies have been conducted on the relationship among plant communities, landforms, and geomorphic surfaces. Parker and Bendix (1996) examined geomorphic influences on vegetation patterns at the landscape scale and found that the geomorphic processes most strongly linked with vegetation patterns operate. Wondzell, Cunningham, and Bachelet (1996) observed that the correlation of plant communities to landforms and geomorphic surfaces resulted from differences in the redistribution of water and organic matter between landforms. Similarly, Hope et al. (2003) found that elevation is a significant predictor of variation in plant diversity, and Lowry, Baker, and Ramsey (2012) found that neighborhood age is an important covariate that influences how the human and environmental factors relate to the abundance of neighborhood tree canopy. Even though many factors, such as socioeconomic status, geographic condition, built environment, and development regulations, have been used as a means to understand urban vegetation, few comprehensive studies of residential landscape structure have been conducted. This gap in understanding hinders the ability to predict the impacts of future residential development and to design planning strategies that might mitigate negative impacts. We hypothesize that subdivision design attributes governed by zoning and geographic characteristics, however, may exert influence on neighborhood landscape structures along with socioeconomic status and the built environment. Providing numerous quantitative metrics is useful for understanding social and ecological benefits and designing sustainable landscapes, particularly in the urban context, where land resources are limited. Different social groups can derive different social and ecological benefits from the same landscape. The resultant trade-off between benefits of urban vegetation and the issue of safety and management may place a constraint in residential landscape structure. However, the benefits of urban vegetation to residents and to urban-dwelling species are significant; therefore, comprehensively identifying the factors that influence and shape landscape structure in urban areas is thus pivotal. The purpose of this study, therefore, is to investigate a broad range of factors related to residential landscape structure in Morgantown, West Virginia––more specifically, to provide a better understanding of how zoning ordinances and geographic characteristics are related to landscape structure in a hilly residential area at two levels: community and block. The study utilized a core set of landscape metrics that are considered the most useful and relevant for landscape planning and management. A quantitatively robust study to better understand how human and environmental factors are related to residential landscape structure at different scales will provide planners and landscape architects with accurate information they can use to improve neighborhood design and to manage neighborhood landscape more efficiently.
2. Study area West Virginia is located within Appalachian region that follows the ridge of the Appalachian Mountains from southern New York to northern Mississippi. The state is primarily mountainous, with the average elevation at 457 m above sea level, which is the highest of any US state east of the Mississippi River. Most of the smooth surface areas are near large rivers such as the Ohio River or Monongahela River. Extremely irregular in boundary,
30
J. Kim, X. Zhou / Landscape and Urban Planning 108 (2012) 28–38
Fig. 1. (a) Location of Monongalia County and city of Morgantown; (b) zoning map and Morgantown ward boundaries; and (c) the study areas and downtown of Morgantown.
West Virginia has two narrow projections––the northern panhandle, which extends adjacent to Ohio and Pennsylvania; and the eastern panhandle, which is between Maryland and Virginia. These two panhandles are the fastest-growing regions in the state (Fig. 1a). The city of Morgantown is the county seat of Monongalia County. The county has a population of 96,189 (US Census Bureau, 2010) and is located near the Pennsylvania border in the north-central part of West Virginia. This area is one of the major growth areas in the state. Monongalia County is the only north-central West Virginia county to grow in population over the past two decades, and it was one of the fastest-growing counties in the state during the first decade of the 21st century. The current area of Morgantown was settled in the mid-18th century. In 1767, Zachwell Morgan and others made the first permanent settlement at Morgantown. Morgan’s proposed plan in 1785 contained 112 lots, main streets, and smaller alleys. By an act of the legislature, Morgantown was established as town on 50 acres of land belonging to Zachwell Morgan, and he was vested with power to lay out lots for sale and to locate streets (Callahan, 1923). In the late 19th century, a railroad system and the discovery of nearby sand and natural gas stimulated Morgantown’s economic and industrial development. As the city’s population increased, the wealthier citizens moved from the industrialized city center to newly constructed suburbs such as South Park. Suncrest and South Park are adjacent neighborhoods in Morgantown (Fig. 1b and c). They are located approximately 2 km to the north and south, respectively, of downtown Morgantown. South Park, one of the Morgantown’s first suburbs, was developed in early-to-mid 20th century, and part of it is designated as a national
historic district. Prior to development, this area was mostly woods and farmland (Gioulis, 1998). Another area, Suncrest, was incorporated as a town in 1937 and became part of Morgantown in 1949. In late 19th century, the Suncrest area was primarily agricultural land. Like South Park, part of Suncrest is very steep. Biophysical characteristics such as soil, vegetation, and the steep topography of the two areas are very similar. Despite these similarities, Suncrest and South Park have major differences in zoning. The Morgantown city council adopted its first zoning ordinance in 1944. Suncrest and South Park were designated as single-family residential, R-1 and R-1A, respectively (Fig. 1b). R-1 includes most of the territory designated for residential use outside of the city as well as extensive areas within the city. The requirements for the R-1 district with regard to minimum lot size and ground floor area of buildings are greater than those specified for R-1A. The purpose of the R-1 district is to provide attractive, singlefamily neighborhoods for residents who prefer larger lot sizes and are not willing to live in close proximity to other types of land uses. R-1A includes much of the city’s older area, which was platted with lots that do not meet R-1 district development standards. The purpose of the R-1A district is to provide single-family neighborhoods that have smaller lots and are within convenient walking distance to other zoning areas. Both R-1 and R-1A preserve desirable characteristics and protect the areas from change and intrusion that may cause deterioration (City of Morgantown, 2006). Fig. 2 shows the street sections, front yards, and street patterns in the two neighborhoods. In South Park, retaining walls were installed along the street because of the steep slope. Aerial photos of this area also show a dense housing arrangement and a grid street pattern.
J. Kim, X. Zhou / Landscape and Urban Planning 108 (2012) 28–38
31
Fig. 2. Comparison of the study areas: street––the road surface of South Park is paved with various materials, and curb and sidewalk are installed. Front yard––because of the steep slope, retaining walls are installed along the street. Aerial photo––housing density of South Park is high because of its small parcel lot size.
3. Data and methods
3.2. Methods
3.1. Data
As previously described, South Park and Suncrest have similar biophysical characteristics such as soil types, geology, precipitation, and plant species (Fig. 3). Because some blocks in Suncrest and South Park were public open space or existing natural areas and the main focus of this study is the neighborhood environment, those blocks were excluded. This study was carried out at two geographic levels––community level and census block level. Community level was used for general comparison of the two communities at a landscape scale, while census block level was selected for examination of socioeconomic factors and characteristics related to the physical setting of a residential area. The number of selected blocks was 66 in Suncrest and 139 in South Park. For each level, methods were further divided into two components: data preparation and data interpretation (Fig. 4).
Data used in this study included color-infrared ortho-rectified aerial photographs at a resolution of 1 m; computer-aided design (CAD) drawings of existing structures; a digital elevation model (DEM); geographic information system (GIS) files including parcel, census block, and road network; the Morgantown city code; and demographic information. Images were obtained from the West Virginia GIS Technical Center. CAD files obtained from the Morgantown Utility Board were used to extract road networks and building footprints. The derived buildings and road layers were exported to ArcGIS 9.3 for further analysis. The DEM was downloaded from the GIS Technical Center and used to calculate local elevation, slope, and aspect. Zoning ordinances for the two communities were obtained from Morgantown city and planning and zoning codes. Parcel and road shapefiles were obtained from City of Morgantown Planning Services. Demographic information, including ethnic group, educational level, and income, was obtained from the US Census Bureau.
3.2.1. Data preparation at the community level Ortho-rectified aerial photography, DEM, CAD data, and parcel shapefiles were processed to acquire information about landscape
32
J. Kim, X. Zhou / Landscape and Urban Planning 108 (2012) 28–38
Fig. 3. Clipped aerial photo with census block boundary.
structure, parcel attributes, and geographic characteristics at the community level. Ortho-rectified aerial photographs were georeferenced to the Universal Transverse Mercator (UTM) coordinate system using the ERDAS IMAGINE 9.2 software package. Eight control points were carefully selected from Google Earth and used to geometrically correct images using a second-order polynomial transformation and the nearest-neighborhood resampling method. The root mean squared error was controlled within 0.5 pixels. Because computerassisted automatic classification might not attain the desired standard of accuracy, the tree canopy was manually digitized (Fig. 5). The identified tree canopy and the remaining non-woody areas were exported in raster format for the calculation of landscape metrics in FRAGSTATS 3.3 (McGarigal, Ene, & Holmes, 2002). Six landscape metrics were used to quantify landscape structure (Table 1):
1. Percentage of Landscape (PLAND) is a general index that depicts the relative abundance of vegetated areas. 2–3. Patch Density (PD) and Edge Density (ED) are sensitive to landscape fragmentation. 4. Mean Patch Size (MPS) can be used to illustrate landscape dominance and fragmentation. 5. Landscape Shape Index (LSI) is associated with landscape complexity. 6. Euclidean Nearest-Neighbor Distance (ENN) represents the degree of patch isolation, using Euclidean geometry as the shortest crow-fly distance between the same types of patches.
The DEM was clipped in ArcGIS 9.3 and then mosaicked based on the boundary of the two communities. The nearest-neighborhood resampling method was used to resample the created DEM to a
Fig. 4. Flow chart of the analysis.
J. Kim, X. Zhou / Landscape and Urban Planning 108 (2012) 28–38
33
Fig. 5. Digitized woody areas of Suncrest and South Park.
30-m resolution. Spatial analysis was conducted to calculate elevation, slope, and aspect. Building footprints were extracted from the CAD map and exported to ArcGIS 9.3 in shapefile format. The created shapefile was spatially adjusted to the UTM coordinate system and overlapped with digital orthophoto quarter quad images. The parcel and road layers were projected onto the UTM coordinate system. To compare road physical patterns in the two communities, road network connectivity was measured using the network analysis tool in ArcGIS 9.3 (Fig. 6). Intersection and edge numbers were calculated. Beta (Eq. (1)) and gamma (Eq. (2)) network indices were computed based on derived edge numbers and intersections:
ˇ=
=
e
(1)
v e 3(v − 2)
(2)
where e is edge number and v is node number.
3.2.2. Data preparation at the block level Landscape structure, parcel attributes, and geographic characteristics were disaggregated for each block. Geoprocessing, based on Python programming, was conducted to calculate block-level landscape metrics. Slope and elevation were averaged within each block. Aspect for each block was determined from the dominant aspect direction. Building footprints and lot sizes were calculated by averaging lot and building areas within individual blocks. 3.2.3. Data interpretation at the community level At the community level, the difference in landscape structure between the two communities may be attributed to demographic characteristics and road patterns. Therefore, demographic characteristics and road patterns in Suncrest and South Park were compared. Association among landscape structure, parcel lot attributes, and geographic characteristics was also investigated. 3.2.4. Data interpretation at the block level To confirm the association of landscape structure, parcel attributes, and geographic characteristics, statistical analysis at the block level was conducted. A t-test of landscape metrics in both
Table 1 Landscape metrics used in this study. Metric
Description
Unit
Range
Justification
Percentage of Landscape (PLAND) Patch Density (PD)
Proportion of the area of a specific land-use class to the entire landscape area Number of patches in a specific class, divided by the entire landscape area Sum of the lengths of all edge segments involving the corresponding patch type, divided by total landscape area, multiplied by 10,000 Sum of the areas of all patches of the corresponding patch type, divided by the number of patches of the same type, divided by 10,000 Patch perimeter divided by the minimum perimeter possible for a maximally compact patch of the corresponding patch area The distance to the nearest neighboring patch of the same type, based on shortest edge-to-edge distance
Percent
0 < PLAND ≤ 100
General index
Number per 100 ha
PD > 0
Index of fragmentation
Meters per hectare
ED ≥ 0
Index of fragmentation
Hectares
MPS > 0
Index of fragmentation
None
LSI ≥ 1
Index of shape
Meters
ENN > 0
Index of proximity
Edge Density (ED)
Mean Patch Size (MPS)
Landscape Shape Index (LSI) Euclidean Nearest-Neighbor Distance (ENN)
34
J. Kim, X. Zhou / Landscape and Urban Planning 108 (2012) 28–38
Fig. 6. Results of road network analysis for Suncrest and South Park.
communities was carried out. Correlations between landscape metrics and geographic and ordinance variables were analyzed for Suncrest and South Park. 4. Results 4.1. Results at the community level 4.1.1. Demographics and road patterns To examine the potential influences of demographics and road patterns on landscape structure, demographics (ethnic group, educational level, and income) were compared for the two communities (Table 2), and beta and gamma connectivity indices were calculated (Table 3). Results of demographic comparisons showed that people residing in these two communities generally had a similar socioeconomic status. Likewise, the composition of ethnic groups was similar. Beta and gamma indices indicated that road connectivity in South Park was slightly higher than in Suncrest, but these communities generally had similar connectivity levels (Fig. 5). Table 2 Demographic comparison between Suncrest and South Park. Element
Category
Suncrest (%)
South Park (%)
Race
White Black Asian Hispanic K-12 High school College <30K 30–50K 50–75K 75–200K
90.37 1.85 5.26 1.2 7.58 15.30 77.11 31.76 27.05 28.7 12.49
90.25 3.64 2.2 1.45 9.92 12.31 77.77 50.63 23.99 16.56 8.82
Education
Income
Table 4 Landscape metrics of Suncrest and South Park at the community level.
Table 3 Results of road network analysis.
Suncrest South Park
4.1.2. Landscape structure, geographic attributes, and parcel attributes At the community level, PLAND indicated that vegetated areas in South Park were more abundant than in Suncrest. PD and ED were much greater in South Park than Suncrest, while the MPS of South Park was lower than that of Suncrest. These findings provided evidence that vegetated areas in South Park were more fragmented than in Suncrest. Both communities had similar LSI values, indicating that landscape shape did not significantly differ between them (Table 4). Elevation in both communities was similar, while slope differed considerably (Table 5). More than half of Suncrest’s area had a slope percentage of less than 25%, while nearly 70% of land in South Park had a slope percentage of more than 25%. Overall, South Park was much steeper than Suncrest (Fig. 7). As shown in Table 6, minimum lot size and minimum lot frontage of zone R-1 were approximately twice as large as those of zone R-1A. Minimum lot size of R-1 and R-1A was 7200 ft2 (2194 m2 ) and 3500 ft2 (1067 m2 ), respectively; and minimum lot frontage was 70 ft (21 m) in R-1 and 30 ft (9 m) in R-1A. Maximum lot coverages in R-1 and R-1A were 40% and 50%, respectively. Setback lines in R-1 were also wider than those of R-1A. Based on these results, demographics and road patterns were determined not to be likely reasons for the differences of landscape metrics. Differences in geography and ordinances may, however, account for the distinctive landscape structure. The association between landscape structure and geographic and parcel attributes was further investigated at the block level.
Vertex
Edge
Beta
Gamma
378 480
436 618
1.153 1.287
0.386 0.430
Metric
Suncrest
South Park
PD ED MPS LSI ENN PLAND
209.62 383.52 0.093 1.467 18.06 19.67
318.14 478.49 0.081 1.337 16.53 25.77
PD, Patch Density; ED, Edge Density; MPS, Mean Patch Size; LSI, Landscape Shape Index; ENN, Euclidian Nearest Neighbor; PLAND, Percentage of Landscape.
J. Kim, X. Zhou / Landscape and Urban Planning 108 (2012) 28–38
35
Fig. 7. Slope and aspect analysis for Suncrest and South Park.
Table 5 Geographic characteristics of Suncrest and South Park.
4.2. Results at the block level
Element
Category
Suncrest
South Park
Elevation
Low High 0–2% 2–5% 5–12% 12–25% >25% N NE E SE S SW W NW
939 ft 1198 ft 2.27% 8.28% 19.54% 25.94% 43.97% 12.35% 10.04% 5.93% 2.95% 16.16% 21.25% 16.22% 15.09%
816 ft 1259 ft 0.87% 2.90% 7.65% 18.97% 69.58% 27.67% 9.60% 4.98% 5.59% 9.30% 7.58% 8.26% 26.13%
Slope
Aspect
4.2.1. Differences in landscape structure At the block level, PD and PLAND for South Park were 7.75% and 4.83% higher, respectively, than those for Suncrest area, while ED and ENN for South Park were lower than for Suncrest (Table 7). A t-test indicated that these metrics significantly differed between the two communities. Results confirmed that, although vegetated areas in South Park were more abundant than in Suncrest, they tended to be more fragmented and isolated compared to Suncrest at the block level. MPS and LSI did not significantly differ between the two places.
4.2.2. Differences in geographic and parcel attributes As for the geographic and parcel attributes at the block level, all values showed significant differences between the two sites.
36
J. Kim, X. Zhou / Landscape and Urban Planning 108 (2012) 28–38
Table 6 Zoning ordinances of Suncrest and South Park. Zoning
Parcel lot
Suncrest (R-1) South Park (R-1A)
Setback (ft)
Min. lot size (sf)
Min. lot frontage (ft)
Max. lot cover (%)
Min. front
Max. front
Min. rear
Min. side
7200 3500
70 30
40 50
25 8
30 20
25 20
10 5
Lot sizes and perimeters in South Park were smaller than those in Suncrest, and building footprints in South Park were about 26.45% smaller than those in Suncrest. The average slope of South Park was 32.45% steeper than that of Suncrest; however, average elevation of South Park was lower (Table 8). 4.2.3. Association between landscape structure and geographic and parcel attributes Associations between landscape structure and geographic and parcel attributes were analyzed in Suncrest and South Park. Parcel attributes in Suncrest area revealed a strong relationship with PD, MPS, and PLAND. Results indicated that ordinances were associated with landscape structure in terms of spatial heterogeneity and fragmentation. Parcel lot size was positively correlated with PLAND, indicating that larger lot sizes provide more space to develop vegetation cover. On the other hand, parcel lot size was negatively correlated with PD, suggesting that fragmentation is less likely to happen where parcel sizes are larger. However, geographic variables did not show an obvious relationship with landscape metrics in Suncrest (Table 9). Lot sizes and perimeters in South Park were also significantly correlated with PLAND and negatively correlated with PD. However, the association between PD and lot size and perimeter in South Park was less significant than the association between those factors in the Suncrest area. Geographic attributes were associated with MPS, indicating that steeper and higher locations had more large areas of vegetation and might have more beneficial microclimate condition for plants. LSI was correlated with slope, suggesting that irregularly shaped vegetated patches were more distributed in blocks with steeper slopes (Table 10). Table 7 Landscape class metrics of Suncrest and South Park at the block level, with p values for statistical difference by t-test. Metric
PD ED MPS LSI ENN PLAND
Suncrest
South Park
p
Mean
SD
Mean
SD
792.27 633.03 .040 1.444 8.47 25.19
317.50 183.84 .034 .175 3.91 9.85
858.88 593.28 .044 1.443 8.31 26.47
541.62 232.63 .045 .182 5.99 13.12
.003 .044 .125 .585 .001 .007
PD, Patch Density; ED, Edge Density; MPS, Mean Patch Size; LSI, Landscape Shape Index; ENN, Euclidian Nearest Neighbor; PLAND, Percentage of Landscape.
Table 8 Difference between zoning ordinances and geographic characteristics in Suncrest and South Park. Metric
AREA (sf) PERI (ft) BLDG (sf) SLOPE (%) ELEV (ft)
Suncrest
South Park
p
Mean
SD
Mean
SD
13563.40 465.77 1877.43 30.34 1003.83
5493.07 85.95 503.50 17.95 35.69
4220.90 268.90 1484.66 44.92 974.28
1815.38 51.91 267.51 27.53 106.08
.000 .000 .000 .004 .000
AREA, parcel lot size; PERI, parcel lot perimeter; BLDG, building footprint; SLOPE, slope; ELEV, elevation.
5. Discussion The results of this study suggest a direct link between landscape structure and land development regulations that govern parcel attributes in residential areas. Higher-density patterns of single-family residential development are associated with greater vegetation abundance, but those areas also have a more fragmented landscape structure than similar developments with lower-density patterns. Geographic slope also influences landscape structure in hilly areas. 5.1. Influences of ordinances and geographic attributes on landscape structure Community-level analyses of landscape structure reveal the different landscape structures between South Park and Suncrest. South Park has more vegetated areas, but those areas tend to be more fragmented than in Suncrest. The Mean Patch Size of vegetation is smaller in South Park. Zoning ordinances in South Park are less restrictive than those in Suncrest: minimum lot size and minimum lot frontage requirements are half those of Suncrest. Moreover, the maximum lot coverage in South Park is 10% more than that in Suncrest. Results indicate that South Park, the higher-density community, which has smaller lot sizes and shorter lot frontages, has more vegetated areas. This result is related to the findings of Stone (2004) that higher density patterns of single-family development are associated with a smaller area of Table 9 Correlation coefficients of landscape metrics with geographic and ordinance attributes in Suncrest. Metric
PD
ED
AREA PERI BLDG SLOPE ELEV
−.430 −.416** −.250* −.023 .023 **
.048 .074 .152 .118 .230
MPS **
.516 .527** .277* .042 .058
LSI
ENN
PLAND
.214 .215 .234 −.046 .048
.001 −.012 −.079 .096 −.194
.383** .394** .213 .194 .164
PD, Patch Density; ED, Edge Density; MPS, Mean Patch Size; LSI, Landscape Shape Index; ENN, Euclidian Nearest Neighbor; PLAND, Percentage of Landscape; AREA, parcel lot size; PERI, parcel lot perimeter; BLDG, building footprint; SLOPE, slope; ELEV, elevation. * p < .05. ** p < .01.
Table 10 Correlation coefficients of landscape metrics with geographic and ordinance attributes in South Park. Metric
PD
ED
MPS
LSI
ENN
PLAND
AREA PERI BLDG SLOPE ELEV
−.192* −.197* −.075 .024 −.139
.085 .070 −.037 −.050 −.134
.386** .380** .166 .178* .198*
.145 .165 .068 .174* .050
−.046 −.065 .046 .058 −.102
.454** .419** .165 −.016 .038
PD, Patch Density; ED, Edge Density; MPS, Mean Patch Size; LSI, Landscape Shape Index; ENN, Euclidian Nearest Neighbor; PLAND, Percentage of Landscape; AREA, parcel lot size; PERI, parcel lot perimeter; BLDG, building footprint; SLOPE, slope; ELEV, elevation. * p < .05. ** p < .01.
J. Kim, X. Zhou / Landscape and Urban Planning 108 (2012) 28–38
impervious cover per unit of occupancy. With respect to geographic attributes, South Park is hillier than Suncrest. This attribute might be related to microclimate that is beneficial to plants and could provide more opportunities for vegetation to grow. This result conforms to the findings of Lowry et al. (2012) that slope and elevation are positively associated with mean annual precipitation, which would affect tree canopy abundance. Block-level analyses further confirm the association of ordinances and geographic attributes on landscape structure. Vegetation abundance, distribution, and fragmentation are significantly different in the two communities, while the attributes of landscape shape and complexity do not have an evident impact. In conformance with zoning ordinances, lot sizes and building footprints at the block level are smaller in South Park. This result indicates that land use related ordinances play a critical role in parcel layout and vegetation patterns in residential areas. Correlation analyses in Suncrest and South Park show a positive association between PLAND and parcel size, suggesting that larger parcel size is related to higher levels of vegetation. Moreover, parcel size and perimeter are positively and negatively associated with MPS and PD, respectively, indicating that larger parcels provide more opportunities to maintain vegetation dominance. The stronger association between parcel size and MPS in Suncrest suggests that larger lots, as required by zoning ordinances, play a greater role in vegetation dominance. Additionally, although building footprints differ between the two study areas, that factor does not appear to be related to vegetation patterns. As for geographic attributes, slope does not markedly affect landscape structure in Suncrest, but it is related to MPS in South Park. This finding suggests that geographic attributes may affect vegetation fragmentation, but do so more directly in hilly than in less hilly places. 5.2. Application of landscape metrics in residential vegetation studies The presence of vegetation in a residential neighborhood is important to the people who live there, but spatial composition and configuration of that vegetation are also crucial. A well-designed neighborhood landscape structure can increase the benefits of green space to humans and––equally important––to urban wildlife. Less-fragmented areas of vegetation can provide viable wildlife habitat, while a more-complex landscape structure may promote urban biodiversity. The importance of ecologically sound neighborhood planning is becoming increasingly acknowledged by local planners. Accordingly, quantifying ecological indicators to guide the process of neighborhood planning is beneficial. Landscape metrics are useful in several planning phases, analysis, prognosis, and synthesis (Leitão & Ahern, 2002). They can provide useful directions for planning and also provide comparative measurements and information to inform insight about landscape structure–function relationships. In this context, landscape ecology-based metrics provide a means to measure and assess the spatial structures of landscape elements in terms of size, number, shape, and spatial arrangement of land elements (Turner, Gardner, & O’Neill, 2001). Characteristics of land form can be quantified by landscape metrics with respect to dominance, fragmentation, connectivity, diversity, and isolation (McGarigal et al., 2002). Our study investigated landscape structures in two communities and demonstrated the usefulness of landscape metrics in capturing the neighborhood landscape composition and configurations. 5.3. Geographic scale This study analyzed landscape structure at two scales: community and block. At the community scale, qualitative comparisons
37
between synoptic landscape structure and geographic and parcel attributes were conducted. An overall understanding was achieved about the associations between landscape structure and the potential factors influencing those associations. At the block scale, indicators related to landscape metrics, ordinances, and geographic attributes were disaggregated to finer units. Analyses at the block level strengthened the understanding at the community level; the block-level analyses also showed the degree of association and the representation of zoning ordinances in each block. This study also demonstrated that landscape metrics could capture vegetation patterns at scales as small as the block level.
5.4. Future research Our study compared two communities that differ in terms of zoning requirements but otherwise are very similar. We recommend future work comparing a greater number of communities to corroborate the current findings. In addition, future research should divide land use into finer categories than the two types (woody and non-woody) examined in this study, classifying vegetated areas into types such as grasslands, shrubs, and tree canopies in order to examine the effects of ordinances and geographic attributes on landscape structure on each vegetation type. Finally, because this study examined only zoning types R-1 and R-1A, we recommend that future research include more zoning categories.
6. Conclusion This study investigated the relationship among landscape structure, zoning ordinances, and geographic attributes in hilly residential areas. Its specific purpose was to better understand how zoning ordinances and geographic characteristics are related to landscape structure in a hilly residential area at two levels: community level and block level. Two residential areas with different residential zoning ordinances, but with similar development history and socioeconomic status, were examined. Six core landscape metrics were selected for calculation of landscape structure. The findings indicate that differences in zoning requirements result in distinct landscape structure––specifically, that zoning requirements for smaller minimum lot size and lot frontage result in more vegetation but also result in more fragmentation than requirements that call for greater minimum lot size and lot frontage. Although building footprint is weakly associated with landscape structure, lot size and perimeter are strongly related to vegetation abundance, fragmentation, and dominance. As for geographic characteristics, slope is associated with patch size, especially in hilly communities. The findings here suggest that landscape ecology-based metrics are efficient ecological indicators to guide the process of residential neighborhood planning. These metrics are beneficial for understanding how elements of local development regulations and geographic characteristics affect landscape structure. This understanding is important for landscape architects, planners, and administrators because it can lead to better strategies for planning and designing socially and ecologically healthy environments. Appropriate zoning ordinances and geographic attributes jointly contribute to sound neighborhood landscape structures. Applying the elements of local ordinances and geographic characteristics that affect landscape structure will likely play an important role for landscape architects, planners, and administrators in developing and amending community land-use regulations.
38
J. Kim, X. Zhou / Landscape and Urban Planning 108 (2012) 28–38
Acknowledgements The research described in this article has been funded in part by the United States Department of Agriculture with Cooperative State Research, Education and Extension Service WVA0049 2H-Hatch. The authors greatly appreciate the thorough review and valuable comments of professor William Sullivan. The authors also thank the efforts of three anonymous reviewers and editor for assisting in the publication of our research. References Borthagaray, A. I., Arim, M., & Marquet, P. A. (2012). Connecting landscape structure and patterns in body size distributions. Oikos, 121, 697–710. Callahan, J. M. (1923). History of West Virginia old and new. Chicago and New York: The American Historical Society, Inc. City of Morgantown, WV. (2006). Morgantown city code. Retrieved from http://www.morgantown.com/planning-docs/PlanZonCode as-of-09-012007.pdf Conway, T. M., & Urbani, L. (2007). Variations in municipal urban forestry policies: A case study of Toronto, Canada. Urban Forestry Urban Greening, 6, 181–192. De Vries, S., Verheij, R., Groenewegen, P., & Spreeuwenberg, P. (2003). Natural environments––Healthy environments? An exploratory analysis of the relationship between greenspace and health. Environment and Planning, 35(A), 1717–1731. Forman, R. T. T. (1995). Land mosaics––The ecology of landscapes and regions. New York: Cambridge University Press. Gioulis, M. (1998). Historic resource survey of the Morgantown Riverfront of Monongalia County West Virginia. Morgantown, WV: Morgantown Historic Landmarks Commission. Goldstein, E. L., Gross, M., & Degraaf, R. M. (1982/1983). Wildlife and greenspace planning in medium-scale residential developments. Urban Ecology, 7, 201–214. Grove, J. M., & Burch, W. R., Jr. (1997). A social ecology approach to urban ecosystem and landscape analyses. Journal of Urban Ecosystems, 1(4), 259–275. Grove, J. M., Cadenasso, M. L., Burch, W. R., Jr., Pickett, S. T. A., Schwarz, K., O’NeilDunne, J., et al. (2006). Data and methods comparing social structure and vegetation structure of urban neighborhoods in Baltimore, Maryland. Society and Natural, 19, 117–136. Grove, J. M., Troy, A. R., O’Neil-Dunne, J. P. M., Burch, W. R., Jr., Cadenasso, M. L., & Pickett, S. T. A. (2006). Characterization of households and its implications for vegetation of urban ecosystems. Ecosystems, 9, 578–597. Hope, D., Gries, C., Zhu, W., Fagan, W. F., Redman, C. L., Grimm, N. B., et al. (2003). Socioeconomics drive urban plan diversity. Proceedings of the National Academy of Sciences of the United States of America, 100(15), 8788–8792. Iverson, L. R., & Cook, E. A. (2000). Urban forest cover of the Chicago region and its relation to household density and income. Urban Ecosystems, 4, 105–124. Jo, H. K., & McPherson, E. G. (1995). Carbon storage and flux in urban residential greenspace. Journal of Environmental Management, 45, 109–133. Kim, J., & Ellis, C. D. (2009). Determining the effects of local development regulations on landscape structure: Comparison of The Woodlands and North Houston, TX. Landscape and Urban Planning, 92, 293–303. Kong, F., Yin, H., Nakagoshi, N., & Zong, Y. (2010). Urban green space network development for biodiversity conservation: Identification based on graph theory and gravity modeling. Landscape and Urban Planning, 95(1–2), 16–27. Kweon, B.-S, Sullivan, W. C., & Wiley, A. R. (1998). Green common spaces and the social interaction of inner-city older adults. Environment and Behavior, 30(6), 832–858. Landry, S., & Pu, R. (2010). The impact of land development regulation on residential tree cover: An empirical evaluation using high-resolution IKONOS imagery. Landscape and Urban Planning, 94, 94–104. Leitão, A. B., & Ahern, J. (2002). Applying landscape ecological concepts and metrics in sustainable landscape planning. Landscape and Urban Planning, 59, 65–93. Lowry, J. H., Jr., Baker, M. E., & Ramsey, D. (2012). Determinants of urban tree canopy in residential neighborhoods: Household characteristics, urban form, and the geophysical landscape. Urban Ecosystems, 15(1), 247–266. MacArthur, R. H., & Wilson, E. O. (1967). The theory of island biogeography. Princeton, NJ: Princeton University Press.
Martin, C. A., Warren, P. S., & Kinzig, A. P. (2004). Neighborhood socioeconomic status is a useful predictor of perennial landscape vegetation in residential neighborhoods and embedded small parks of Phoenix, AZ. Landscape and Urban Planning, 69, 355–368. McDonnell, M. J., & Pickett, S. T. A. (1990). Ecosystem structure and function along urban rural gradients: An unexploited opportunity for ecology. Ecology, 71(4), 1232–1237. McGarigal, K., Ene, E., & Holmes, C. (2002). FRAGSTATS (Version 3): FRAGSTATSMetrics. University of Massachusetts-Produced Program. Available at the following website: http://www.umass.edu/landeco/research/fragstats/documents/ fragstats documents.html Murgui, E. (2009). Influence of urban landscape structure on bird fauna: A case study across seasons in the city of Valencia (Spain). Urban Ecosystems, 12(3), 249–263. Nowak, D. J., & Crane, D. E. (2002). Carbon storage and sequestration by urban trees in the USA. Environmental Pollution, 116, 381–389. Nowak, D. J., Crane, D. E., & Stevens, J. C. (2006). Air pollution removal by urban trees and shrubs in the United States. Urban Forestry Urban Greening, 4, 115–123. Parker, K. C., & Bendix, J. (1996). Landscape-scale geomorphic influences on vegetation patterns in four selected environments. Physical Geography, 17, 113–141. Robinson, D. T., & Brown, D. G. (2009). Evaluating the effects of land-use development policies on ex-urban forest cover: An integrated agent-based GIS approach. International Journal of Geographic Information Science, 23(9), 1211–1232. Schmid, J. A. (1975). Urban vegetation: A review and Chicago case study Department of Geography Research Paper No. 161. Chicago: University of Chicago. Shin, W., Yeoun, P., Yoo, R., & Shin, C. (2010). Forest experience and psychological health benefits: The state of the art and future prospect in Korea. Environmental Health and Preventive Medicine, 15(1), 38–47. Stone, B., Jr. (2004). Paving over paradise: How land use regulations promote residential imperviousness. Landscape and Urban Planning, 69, 101–113. Sullivan, W. C., Kuo, F. E., & DePooter, S. F. (2004). The fruit of urban nature––Vital neighborhood spaces. Environment and Behavior, 36(5), 678–700. Sung, C. Y. (2012). Evaluating the efficacy of a local tree protection policy using LiDAR remote sensing data. Landscape and Urban Planning, 104, 19–25. Swanwick, C., Dunnett, N., & Woolley, H. (2003). Nature, role and value of green space in towns and cities: An overview. Built Environment, 29(2), 94–106. Tajima, K. (2003). New estimates of the demand for urban green space: Implications for valuing the environmental benefits of Boston’s Big Dig project. Journal of Urban Affairs, 25, 641–655. Troy, A. R., Grove, J. M., O’Neil-Dunne, J. P. M., Pickett, S. T. A., & Cadenasso, M. L. (2007). Predicting opportunities for greening and patterns of vegetation on private urban lands. Journal of Environmental Management, 40, 394–412. Turner, M. G. (1989). Landscape ecology: The effect of pattern on process. Annual Review of Ecology and Systematics, 20, 171–197. Turner, M. G., Gardner, R. H., & O’Neill, R. V. (2001). Landscape ecology in theory and practice: Pattern and process. New York: Springer. US Census Bureau. (2010). http://be.wvu.edu/demographics/census.htm (accessed on 10.04.11). West Virginia GIS. West Virginia GIS Technical Center. http://wvgis.wvu.edu/ data/dataset.php (accessed on 15.10.10). Wilson, J. S., Clay, M., Martin, E., Stuckley, D., & Vedder-Risch, K. (2003). Evaluating environmental influences of zoning in urban ecosystems with remote sensing. Remote Sensing of Environment, 86, 303–321. Wondzell, S. M., Cunningham, G. L., & Bachelet, D. (1996). Relationships between landforms, geomorphic processes, and plant communities on a watershed in northern Chihuahuan Desert. Landscape Ecology, 11(6), 351–362. Young, R. F. (2010). Managing municipal green space for ecosystem services. Urban Forestry and Urban Greening, 9(4), 313–321. Yuan, F., & Bauer, M. E. (2007). Comparison of impervious surface area and normalized difference vegetation index as indicators of surface urban heat island effects in Landsat imagery. Remote Sensing of Environment, 106, 375–386. Zhou, W., Huang, G., & Cadenasso, M. L. (2011). Does spatial configuration matter? Understanding the effects of land cover pattern on land surface temperature in urban landscapes. Landscape and Urban Planning, 102, 54–63. Zhou, X., & Wang, Y. (2011). Spatial–temporal dynamics of urban green space in response to rapid urbanization and greening policies. Landscape and Urban Planning, 100, 268–277. Zipperer, W. C., Sisinni, S. M., Pouyat, R. V., & Foresman, T. W. (1997). Urban tree cover: An ecological perspective. Urban Ecosystems, 1, 229–246.