Green Space Site Suitability in a Dense Urban Setting:
Developing more green space in Somerville, Massachusetts
For the final project, I conducted a site suitability analysis of parcels in Somerville. The analysis was focused on identifying parcels that are appropriate for being developed as green space in Somerville, based on a variety of factors. One underlying assumption involved is that Somerville is in need of more green space, defined in this project as open public spaces covered in vegetation. The Land Use Technical Report  released by the City of Somerville’s Office of Strategic Planning and Community Development determined that the percentage of land dedicated to green space in Somerville is lagging behind that of neighboring communities. The City is furthermore dedicated to increasing the amount of green space available to the public, as detailed in the Somerville Open Space and Recreation Plan 2008–2013 .  The City, along with many others, recognizes that green space provides a host of benefits, touching on physical and mental health, environmental service provision, community cohesion, and more.
Two main factors were involved in the suitability analysis. First was the issue of equity of access to green space. A variety of data was analyzed to determine areas within Somerville that are in need of more green space, since green space is not equitably distributed within the city, nor do all residents have equal access to existing green space within and outside of the community. Second was parcel appropriateness. Somerville is a densely settled city with limited undeveloped areas. Still, green space may be introduced to certain parcels that are either already developed or are vacant. Commercial rooftops and land held by private interests have the potential of being converted to public green space, as they have been in Seattle and San Francisco.   Vacant lots represent other potential green space sites. This project sought to identify parcels that are both structurally appropriate and that address issues of access equity within Somerville.
- Are there currently areas in Somerville that lack access to green space in the immediate vicinity, as compared to other areas?
- Are there currently areas in Somerville that lack ways to access to green space outside of the immediate vicinity (such as public transportation), compared to other areas?
- Where should new green space be placed to create more equitable access to green space within the City of Somerville’s boundaries?
- Which already-developed parcels in Somerville can be converted at least partially to green space use?
- Which undeveloped parcels in Somerville can be converted into green space?
The following sources helped inform this project:
Byrne, Jason and Neil Sipe (2010). Green and open space planning for urban consolidation—A review of the literature and best practice. Brisbane, QLD: Griffith University.
This literature review offers a good conceptual background in green space planning. Specifically, it points out that city residents do not all have the same green space needs, and that the best green spaces are multi-use and are attractive to a variety of people. Thus, the quality of green space (and open space without vegetation) is important, not just the quantity. The authors review many of the benefits access to green space confers on communities, from mental health, to decreasing obesity, to revenue creation. The study also provides a comparison of park spatial standards. In the United States, the most recent standards, which are admittedly old, stipulate that each 10 acres of park should serve 1,000 residents, and should be located within ¼ mile of them. These standards don’t take park quality into account and have not always been achieved (such as in Somerville). The review also covers GIS methodology, and suggests including green space distribution data, accessibility data, and demographic data to determine equity of current and future green space development.
Comber Alexis, Chris Brunsdon, and Edmund Green (2008). “Using a GIS-based network analysis to determine urban greenspace accessibility for different ethnic and religious groups.” Landscape and Urban Planning 86 : 103–114.
This article uses network analysis to answer questions about how access to green space is related to demographics. The authors explore not just the amount of green space in Leicester, England, but examine its distribution in relation to different ethnic and religious groups. They used network analysis to determine the distances between green spaces and the centers of census groups of about 300 people each. The paper further asks which areas in the city need to have enhanced green space provision based on findings of inequitable access and published green space guidelines that suggest parameters for public access to green space (i.e. people should live within 300 meters of green space, etc.). Many of these ideas and methodology can be applied to a Somerville context.
Stahle, Alexander (2010). More green space in a dense city. Urban Design International 15, 47 –67.
This study is useful for thinking about the relationship between a green space’s accessibility and attraction, defined traditionally as distance and surface area. The paper proposes new ways of understanding accessibility and attraction. The latter is especially relevant to this project, since it redefines attraction as use value. GIS can incorporate the number of use values of a green space into analysis, to provide a better picture of the importance of any one space. The authors created a “sociotope” map as part of their research, depicting several use values of green space. They found that combining use values, green space area, and distance to green space was correlated with how people perceived their own access to green space.
Uy, Pham Duc and Nobukazu Nakagoshi (2008). Application of land suitability analysis and landscape ecology to urban greenspace planning in Hanoi, Vietnam. Urban Forestry & Urban Greening 7 , 25 –40.
The authors offer an example of land suitability analysis as applied to urban green space in Vietnam. They detail the criteria used for land suitability, and the method for applying criteria using GIS, including weighted scoring, which is appropriate for this project. The study also goes further and assesses the overall amount of city green space, recommending that the city increase land dedicated to green space and proposing spaces that can be used for this purpose. Similar conclusions may be made for Somerville after a land suitability analysis.
Wolch, Jennifer, John P. Wilson, and Jed Fehrenbach. Parks and Park Funding in Los Angeles: An Equity-Mapping Analysis. Urban Geography 26(1), 4–35.
This paper lays out a park equity mapping analysis. The authors point out that there are two sources of unequal access to green space—actual unequal distribution, and socioeconomic factors that lead to uneven ability to use the space available (such as being able to afford a membership at a pool, ability to drive to a larger park, etc.). For the analysis, they define access as living within a quarter mile of a park edge. Quarter mile park buffers were created to determine access. The researchers determined park acres per 1,000 population, percent of census tract population living within the buffer zone, and park acres per 1,000 population in the buffer zone. They found, among other things, that lower-income households had inferior access to parks.
• Somerville Parcels (vector), 2009: City of Somerville
• Protected and Recreational Open Space (vector), 2014: MassGIS; http://www.mass.gov/anf/research-and-tech/it-serv-and-support/application-serv/office-of-geographic-information-massgis/datalayers/osp.html
• Bike Trails (vector), 2004: MassGIS; http://www.mass.gov/anf/research-and-tech/it-serv-and-support/application-serv/office-of-geographic-information-massgis/datalayers/bicycle-trails.html
• Building Structures (vector), 2014: MassGIS; http://www.mass.gov/anf/research-and-tech/it-serv-and-support/application-serv/office-of-geographic-information-massgis/datalayers/structures.html
• Poverty (tabular), 2012: American Census Survey 5-year estimate, 2012: American Census Bureau Tiger data, http://factfinder2.census.gov/faces/nav/jsf/pages/index.xhtml
• Census Tracts (vector), 2013: TIGER/Line Shapefiles, https://www.census.gov/geo/maps-data/data/tiger-line.html
• MBTA Nodes (vector), 2006: MassGIS; http://www.mass.gov/anf/research-and-tech/it-serv-and-support/application-serv/office-of-geographic-information-massgis/datalayers/trains.html
• EOT Roads (vector), 2012: MassGIS; http://www.mass.gov/anf/research-and-tech/it-serv-and-support/application-serv/office-of-geographic-information-massgis/datalayers/eotroads.html
• City Boundary (vector), 2005: City of Somerville
• Hydrography (vector), 2010: MassGIS, http://www.mass.gov/anf/research-and-tech/it-serv-and-support/application-serv/office-of-geographic-information-massgis/datalayers/hd100.html
Preparation and Analysis Steps
Three major phases were involved in this project. First, several data sets that informed an assessment of access to green space were analyzed separately and then combined to create a larger picture. The data sets included five different factors: proximity to all public green space, proximity to large green space greater than one acre, proximity to green space with sports facilities, proximity to MBTA train stations (to determine access to green space outside of the immediate area), and poverty by census tract. Once combined, the resulting map provides a visual tool useful in understanding which areas of Somerville have superior access to quality green space and which have lower levels of access, and shows the areas that should be prioritized when considering more green space development.
Second, three different parcel suitability analyses were run. Vacant lots available for green space development were identified. Commercial and industrial rooftops over half an acre made up a second analysis and commercial and industrial parcels equaling more than half an acre without building footprints made up a third. In the third phase, the parcel suitability data was then overlaid on the combined accessibility analysis to identify the best parcels for green space development.
- Import all data layers and project into appropriate coordinate system ( 1983 State Plane Massachusetts Mainland FIPS, feet).
- Create base map of Somerville using town line boundary, hydrography, roads, and open space.
- Join census tract data and poverty data, using tabular Join and Relate tool.
- Calculate all areas (acreage) needed for data sets.
Join open space data layer to bike trail data layer
- Select by attribute for existing and planned bike trails in Somerville.
- Use buffer tool to create a 25-foot span to account for bike trail green space.
- Merge bike trail buffer to open space data.
Proximity analysis for public open space
- Select by attribute for all polygons in open space data layer that are publically accessible and create new data layer.
- Use the raster Euclidean Distance tool to create a raster grid showing distance away from public open space.
- Reclassify proximity to green space:
1- under 1/16 mile
2- 1/16-1/8 mile
3- 1/8-1/4 mile
4- more than ¼ mile
Proximity analysis for large open space.
- Select by attribute in public open space data layer for spaces that are at least an acre in size and create new data layer.
- Use the raster Euclidean Distance tool to create a raster grid showing distance away from large public open space.
- Use same reclassification standards as step 6.
Proximity analysis for open space with sports facilities.
- Use base imagery overlay and the City of Somerville Parks map (http://www.somervillema.gov/sites/default/ files/documents/SomervilleParks7.pdf) to select open space with sports facilities, including basketball courts, baseball diamonds, and soccer fields. Create data layer from these selections.
- Use the raster Euclidean Distance tool to create a raster grid showing distance away from open space with sports facilities.
- Use same reclassification standards as step 6.
Proximity analysis for MBTA nodes
- Use the raster Euclidean Distance tool to create a raster grid showing distance away from T stations.
- Use same reclassification standards as step 6.
Poverty by census tract data
- Convert poverty by census tract data to raster form using the Polygon to Raster tool.
- Reclassify poverty by census tract (1 is highest income, 5 is lowest income).
Use the Raster Calculator tool to combine all reclassified data sets and change color ramp for clarification. All criteria are weighted equally, and the range stretches from 6 to 21.
- Use Select by Attribute tool to identify all parcels without buildings (first by selecting parcels with buildings and then switching the selection). This is an approximation of vacant lots.
- Create a new data layer.
Rooftop parcel suitability
- Convert buildings to points using the Feature to Point tool.
- Use Spatial Join tool to join building points to parcels, in order to assign CLS codes to building structures.
- Use Select by Attribute tool to identify building structures with commercial or institutional CLS codes.
- Use Select by Attribute tool again to identify buildings with rooftop footprints over 1/2 acre and create new data layer.
Commercial building land
- Use Erase tool to get rid of building structures from parcels.
- Use Select by Attribute tool to identify commercial and institutional parcels.
- Use Select by Attribute tool again to identify parcels that are still at least 1/2 acre and create new data layer.
- Select out all land dedicated to schools (with the assumption that schools offer open space such as playgrounds to the public) and already-established parkland.
The transparent open space data layer shows some of the redundancy of the identified commercial/industrial land available for green space. In this case, parcels that were already green space were selected out, such as the parcels in the extreme western portion of Somerville.
- Do a visual overlay to compare all three layers (vacant lots, rooftops, and commercial/institutional land) to the accessibility map to determine which sites should be prioritized for green space development.
Difficulties and Limitations
The accessibility factors were by no means exhausted, and only provide a partial picture of the inequities of access to green space. More socioeconomic data would be useful as well as a more thorough assessment of access to quality green space, including factors such as proximity to playgrounds, picnic areas, etc. The accessibility analysis could even be grouped into two raster calculated maps: underlying socioeconomic factors such as poverty, immigrant status, female-headed households, and health issues on the one hand, and proximity analyses on the other.
In the second part of this project, the vacant lot selection method was particularly flawed, but was the most straightforward way to approximate vacant lots. Some identified parcels did not have buildings, but should not be considered vacant. Some parcels identified as vacant were in fact already green space.
Finally, the parcel data used in this analysis was from 2009 and out of date, but should improve with the addition of more attributes to the recently released Somerville parcel data for 2014. Some parcels have been developed already, in some cases into green space. For example, Draw 7 park in East Somerville has been created since the last parcel assessment in 2009, which this analysis identified as a prime vacant land spot.
The darker turquoise parcels (such as the square in the middle of this image) show parcels that were identified as vacant lots, but are in fact already green space parcels without a building on them. Additionally, some of the vacant land above is no longer vacant, such as the large blue parcel in the northwest corner of this image.
Conclusion and Next Steps
The combined proximity and poverty data reveal inequities in access to green space in Som erville and the need to develop more, particularly in the outer western, eastern, and southern portions of the city. Fortunately, these methodologies have produced many parcels that could be suitable for public green space conversion, although they are limited. The results produced a large number of potential sites, meaning that further investigation including analysis of parcel ownership, current surface type, and groundtruthing would be necessary to determine which of the identified parcels could actually be developed into green space.
Furthermore, this project does not exhaust the possibilities for determining the best sites for green space development in Somerville. The methodologies can be used in conjunction with additional data to provide a more thorough analysis of current green space accessibility as well as additional innovative land use possibilities, to hone in on more parcels of different types that could be converted to public green space.
 Office of Strategic Planning and Community Development (2011). Trends in Somerville: Land Use Technical Report. http://www.somervillema.gov/sites/default/files /LandUseTrendsReportFinalMay2011.pdf.
 City of Somerville (2009). Somerville Open Space and Recreation Plan 2008–2013 . http://www.somervillema.gov/sites/default/files/documents/2008-2013-OSRP-NarrativeAndAppendicesFINAL.pdf.
 “Privately-Owned Public Open Space and Public Art.” San Francisco Planning Department. Last updated October 6, 2013. http://www.sf-planning.org/index.aspx?page=3339
 “Privately Owned Public Open Spaces.” Settle City Council. Accessed April 13, 2014. http://www.seattle.gov/council/licata/public_space.htm