Jeanette Rebecchi
Final Paper
December 17, 2010
City of Marlborough Build-Out Analysis
Project Description
A build out analysis is a useful tool for estimating and illustrating the amount and possible
location of future development under current zoning and development regulations. There are two
phases in this project; the first is to map out the location of all potentially developable land using
GIS. The second analytical phase provides an estimate of the total number of homes,
commercial/industrial square footage, and demographic changes that could result if all buildable
land within a community is developed.
By showing the community this information, planners can receive feedback from the public
about how they want their city to look in future. Planners could then permanently protect desired
open space parcels and change local zoning and subdivision regulation to reflect the
community’s preferences. A build-out analysis is also instrumental for the municipality to
estimate future demands on public infrastructure such as schools, water supply, sewage, utilities,
waste disposal, as well as any future tax revenue that could be gained.
When planners equip themselves with the right tools and information, development can
proceed within a framework of rules and regulations that produce attractive, livable
communities. In order to achieve this, it is necessary to ask these kinds of questions and analyze
these build-out maps to prevent the kinds of developer controlled, sprawling development we
have seen in the past. Not all development has to be undesirable, unsustainable, or suburban
sprawl.
Site Description
The City of Marlborough is located an hour drive west of Boston. It is a growing city
with a dense commercial corridor running along Route 20 that gradually turning more suburban
and rural as one move away from the main highways. New development within the town is
constrained by several large bodies of water and protected open space. This analysis will serve to
inform the City of where development can occur in the future by right (i.e. as allowed by zoning
and not needing a special permit) and how that might influence the future landscape of this
community.
Figure 1 Locator Map
Key Questions
1. How much land area can be developed under existing zoning regulations, and where
will this growth occur?
2. What type of development can occur on the available buildable land?
3. Where will this growth occur?
Data Sources
Data Layer
Source
Source Scale
Year Created
Boundary Arc File
Mass GIS
1:25,000
2009
Boundary Polygon
Mass GIS
1:25,000
2009
Hydrography Polygon M Drive City of Marlborough
unknown
2003
Open Space Polygon
Mass GIS
1:25,000
2010
Wetlands Polygon
Mass GIS
1:12,000
2007
Digital Elevation
Model Raster
Mass GIS
1:5000
2005
Roads Polygon
M Drive City of Marlborough
unknown
2008
Driveway Polygon
M Drive City of Marlborough
unknown
2003
Parking Polygon
M Drive City of Marlborough
unknown
2003
Buildings Polygon
M Drive City of Marlborough
unknown
2003
Easements Polygon
M Drive City of Marlborough
unknown
2007
Zoning Polygon
Marlborough GIS Staff
unknown
2010
Parcels Polygon
Marlborough GIS Staff
unknown
2009
Methodology Phase 1: Build-out Maps
Data Processing
The first step in a build-out analysis is to create a map depicting remaining buildable land
and the zoning designation. After all data layers listed above were added into Arc Maps, I
checked to make sure that all coordinate and measurement systems were the same. Since the
majority of my data layers came from the City of Marlborough and were in feet, I converted the
data frame and the rest of the layers using the Project Tool into the projected coordinate system:
NAD_1983_StatePlane_Massachusetts_Mainland_FIPS_2001_Feet.
To speed up processing time, I then created smaller data sets for all statewide layersopen space, wetlands, DEM elevation layer. To do this I first selected the City of Marlborough
record from the Mass GIS statewide town boundaries polygon. I then exported the selected
record as its own polygon layer creating just a City of Marlborough boundary polygon. Using the
Marlborough Boundary layer I then used the clip tool to create a layer of just the wetlands, open
space, and DEM within the boundary. Afterwards using the intersect tool, I added all the
attributes found in the zoning and land use polygons onto the Marlborough boundary polygon.
The boundary polygon was then used as a base layer to clip or erase all future data layers to.
Figure 2 Creating Marlborough Boundary Polygon
Figure 3 Using Clip Tool to Reduce Statewide Files
Removing Development Constraints
On both residential and commercially zoned parcels you must take into account various
limiting factors such as:






Bodies of water
Protected and recreational open space
DEP designated wetlands
Slopes greater than 25%
Current built features such as roads, parking lots, and driveways
Easements
Using the Marlborough Boundary Polygon as a base layer I removed all constrained land
using the erase tool. One by one I erased the Mass GIS open space and wetlands polygons, as
well as the hydrography, road, driveway, parking, easement polygons from the City, each time
saving a new base map shape file with the most recently removed constraints. The picture below
illustrates having erased the open space and easement polygons, and moving on to erase the
roads polygon.
Originally, I had erased the buildings polygon from the base map, but that did not accurately
represent available buildable land. Because, having removed all building footprints, what was
left was all vacant land plus the remainder of the parcel that did not physically have a building
sitting on it. In many cases these were things like people’s backyards. While I would have
preferred to conduct an analysis considering underdeveloped land, the methodology needed did
not work with the time frame I had to complete this project. In the end, I wound up using the
select by location tool and created a shape file of all parcels that intersected with a building, and
erased it from the base map. What remained was all truly vacant land without any physical
structure on it.
Next, I erased all publically owned land since this land could not be subject to
development without an extensive review process or a change in ownership. I used the select by
attribute tool to select all records in the base map with a land use descriptions that equals
‘municipal’, ‘county government’, ‘electric row’ (utilities), ‘Comm Mass’ (Commonwealth of
Massachusetts- state owned land), ‘US government’ (post office). I exported the selected data
into a new shape file and erased this land from the base map.
Removing areas of land with a slope greater than 25% was the most complicated step in
this process because I first had to create a slope raster file from the DEM file using the Slope
tool. Next, I exported the slope file as a vector file and reclassified the data to read “1 = slope
less than 25%” and “0 = slope greater than 25%”. Using select by attribute = 0, I then exported
the selected records into a new shape file. Using the erase tool, I erased the new shape file from
the base map to remove all land with slope of greater than 25%.
Figure 4: Erasing Constraints from Base Map
Data Cleaning
After all development constraints were removed, I noticed that there were many oddly
shaped parcels and slivers of land left over (see Figure 5 below). After zooming in closely and
using the identify tool, I figured out that these were extra pieces of land that were left over from
the erase process. Most often it was because the Mass GIS open space and wetlands layers did
not coincide with the City’s parcel map, a parcel map that itself had many oddly shaped
unlabeled parcels.
To correct some of these mistakes I first deleted any record less than 2,600 square feet in
size, the average size of a single family home in the New England in 2009.1 I assumed that
because of its small size, this land was not capable of being developed unless steps were taken to
consolidate surrounding parcels. I then selected by location all base map records that intersected
with the roads polygon. I sorted the selected records by parcel owner and deleted all records that
did not have an owner since they often seemed to be landscaped medians or shoulders that were
not erased with the actually road polygon.
Finally, I sorted the attribute table by CNS_TOT_1 which describes the type of building
on the parcel (number of rooms/ bedrooms/bathrooms). I noticed that some parcels that actually
have buildings according to the attribute table were still on the map. This may be because they
were missing from the buildings polygon layer and the land never removed. I deleted these
records using the data editor tool. I also sorted the final land map by land use description, and
deleted using the same tool all records that were labeled parking in the land use description
column since they must have been left off the original parking polygon and never erased.
1
U.S. Census. (2009).” Median and Average Square Feet of Floor Area in New Single-Family Houses Completed by
Location.” http://www.census.gov/const/C25Ann/sftotalmedavgsqft.pdf
Figure 5 Buildable Land Map- Before & After
Methodology Phase 2: Quantitative Analysis
After copying the final base map attribute table into excel, I was able to calculate the
number of parcels and area of land by each zoning type. Due to time constraints and the nature of
my methods I simplified my analysis to just look at the number of available parcels and the
amount of square footage. As shown in the tables below, Marlborough has roughly 9% of its
total dry land area available for future development. Primarily this is the form of residential and
industrial land. Square footage was predicted by removing 10-30% of the parcel to account for
zoning restrictions and site amenities (roads, parking, FAR, etc.). With further breakdown of
zoning classifications as seen in Table 2, limited industrial and single family residential land are
the most common types available (37% and 25% of all buildable land respectively).
Table 1 Buildable Land by Use
% Total
Dry Land2
Acres
Commercial
86
47
2,630,997
3,382,711
7%
0.6%
Residential
527
539
16,074,684
20,667,451
43%
3.9%
Industrial
623
127
19,011,064
24,442,797
50%
4.6%
1,237
713
37,716,746
48,492,959
100%
9.2%
Use Type
Total
Low Sq Ft High Sq Ft
Estimate^ Estimate*
% Total
Buildable
Area
# of
Parcels
^30% of buildable land constrained by zoning and site amenities (parking, road access, etc)
*10% of buildable land constrained by zoning and site amenities (parking, road access, etc)
Table 2 Buildable Land by Zoning Type
Zoning Type
% Total
# of
Area Low Sq Ft High Sq Ft
% Total
Buildable
Parcels (acres) Estimate^ Estimate*
Dry Land3
Area
Business
47
86
2,630,997
3,382,711
7%
0.6%
Industrial
38
160
4,902,726
6,303,504
13%
1.2%
Limited Industrial
89
462
14,108,339
18,139,293
37%
3.4%
1
8
241,893
311,005
1%
0.1%
51
10
316,781
407,290
1%
0.1%
407
311
9,505,406
12,221,236
25%
2.3%
Rural Residence
80
197
6,010,603
77,27,919
16%
1.5%
Total
713
1237
37,716,746
48,492,959
100%
9.2%
Retirement
Community
Multi-Family
Residential (2
units)
Single-Family
Residential
2
U.S. Census. (2000). “GCT-PH1. Population, Housing Units, Area, and Density”.
http://factfinder.census.gov/servlet/GCTTable?_bm=y&-geo_id=04000US25&-_box_head_nbr=GCT-PH1&ds_name=DEC_2000_SF1_U&-format=ST-7
3
U.S. Census. (2000). “GCT-PH1. Population, Housing Units, Area, and Density”.
http://factfinder.census.gov/servlet/GCTTable?_bm=y&-geo_id=04000US25&-_box_head_nbr=GCT-PH1&ds_name=DEC_2000_SF1_U&-format=ST-7
^30% of buildable land constrained by zoning and site amenities (parking, road access, etc)
*10% of buildable land constrained by zoning and site amenities (parking, road access, etc)
To measure the impact if all buildable land was actually developed, I used data from the
US Census and from another build out report conducted for the Town of Hopkinton, MA. Since a
retirement community overlay district was included containing 8 acres of buildable land, I
estimated development impacts by comparing it to another 8 acre retirement village. Using
Keystone Place at Legacy Ridge Retirement Community in Westminster Colorado, I assumed
that since this community contained 163 apartments and that half the population might be living
with a partner the retirement village would create another 245 residents. This number was not
used to calculate new students or additional vehicles on the road. As seen in Table 3, some of the
effects of all this new development such as water usage and additional vehicles are particularly
concerning and something the City of Marlborough should prepare for.
Table 3 Summary Build-Out Statistics (Additional Development & Impacts)
Developable Land Area
Commercial Floor Area
Commercial Water Use
Total Residential Parcels
Dwelling Units
Future Residents
Residential Water Use
Municipal Solid Waste
Students
Additional Vehicles on Road
1,237 acres
27,825,508 sq ft
2,086,913 gallons/day
539
752
1676
125,720 gallons/day
1,719,853 lbs/person/year
385
940
Notes:
1.
2.
3.
4.
5.
4
“Residential Water Use” is based on 75 gallons per person per day4
“Commercial Water Use” is based on75 gallons per 1,000 square feet of floor space 5
“Municipal Solid Waste” is based on 1026lbs per person per year for residential uses only.6
The number of “Students” is based on current average in Marlborough: 26.9% of households7
“Additional Vehicles on Road” is based on the current average in Marlborough in that on average 36% of
household own 1 vehicle, 36% of households own 2, and 17% of households own 3 or more.8
Executive Office of Energy & Environmental Affairs. (2002). “Where Do You Want to Be at Build Out?: The Build
Out Analysis In-depth.” http://commpres.env.state.ma.us/publications/CHAPTER2.PDF?
5
Ibid.
6
Ibid.
7
US Census. 2005-2009 American Community Survey- Five Year Estimates.
8
Ibid.
Conclusion
Overall I felt this analysis was a good way to learn how to use GIS in a municipal setting.
However, what seemed like a relatively straight forward project did take a lot of time and effort
particularly when there were many ‘buildable’ land parcels that clearly were not buildable.
Without going through and examining each parcel individually it’s hard to ensure complete
accuracy. Furthermore, I used a generic estimate to calculate square footage of development,
when a more thorough analysis would have taken into account all zoning requirements such as
minimum lot size or FAR. Additionally, further study could have also examined the
development potential of underdeveloped parcels. This could have been done by looking at each
property individually to determine if the land can be subdivided or if a home can be converted to
multi-family residence or commercial uses. If the property is commercial, one can examine if the
business can be expanded or changed to residential or if parcels can be consolidated to form
superstores. By including underdeveloped land, the final impact estimates would have been more
accurate.
Resources
1. Montgomery County Planning Commission.(1996) Shaping Future Development: The Role of
Current Zoning. http://www.epa.gov/greenkit/pdfs/futgrow0.pdf
This manual gives an informative background on how build-out analyses can help
planners make informed decisions about their city’s infrastructure and growth controls. It
also highlights the effects that new development has on transportation, infrastructure,
population, open space, etc.
2. Theobald, D., Hobbs, Thompson. (1998) “Forecasting Rural Land-use Change: A Comparison
of Regression and Spatial Transition-based Models”. Geographical & Environmental Modeling.
Vol 2. No. 1. pp. 65-82. http://warnercnr.colosctate.edu/~davet/theobald_hobbs_1998_gem.pdf
This article compares two different types of land use changes models. When comparing
accuracy regarding predicted land use change, the spatial model worked better.
3. US Environmental Protection Agency. How to Do a Build-Out Analysis
http://www.epa.gov/greenkit/build_out.htm
This guide from the EPA provides basic, “how to” information. This was a good starting
point to uncover what data I will need and what questions to ask.
4. The Official Website of the Office of Geographic Information (MassGIS)
“Scope of Services for Build-Out Analysis” http://www.mass.gov/mgis/buildout.htm
This vital document summarizes the methodology and deliverables to be performed by
the contractors when preparing a build-out analysis for the State. It goes in depth into the
formulas and data layers used in the quantitative part of the analysis.
5. The Executive Office of Energy and Environmental Affairs. (2002) “Where Do You Want To
Be at Build-out?; The Build-Out Analysis in Depth”.
http://commpres.env.state.ma.us/publications/CHAPTER2.PDF?
This publication goes into depth into the build-out analysis performed for Hopkinton
Massachusetts.
6. Godschalk, David. (2006). “Build-Out Analysis: A Valuable Planning and Hazard Mitigation
Tool”. Zoning Practice. American Planning Association. Issue 3.
http://www.planning.org/zoningpractice/2006/pdf/mar.pdf
This guide from the APA while providing basic info, it describes the methodology that
the State of Massachusetts used to conduct a build out analysis for every city and town
across the State as part of the Community Preservation Initiative of Executive Office of
Environmental Affairs. The article described how undeveloped land in each zoning
district was identified through the interpretation of orthophotos without regard to parcel
boundaries. Using overlay and spreadsheet tools, future residential units and
industrial/commercial areas were then estimated. The final analysis included the number
of housing units, population and number of school children, square feet of commercial
land and industrial space, gallons of water demanded, and miles of roads.
This article also offered great examples of further resources available on this topic.
7. Lacy, Jeffrey. (1990) Manual of a Build-Out Analysis. University of Massachusetts- Amherst.
http://scholarworks.umass.edu/cgi/viewcontent.cgi?article=1023&context=larp_ms_projects
This manual introduces the planning tool of build-out analysis, and describes the data
requirements, materials, and techniques necessary to complete both the GIS and
quantitative portions of the analysis.
8. Amengual, Matthew. (2001). “Charlestown at Build-out: Modeling Development and
Conservation.”
http://envstudies.brown.edu/oldsite/Thesis/2001/amengual/buildout/methodology.htm
This thesis is another example of how to perform a build-out analysis using GIS.
9. Polimeni, John. “Simulating Agricultural Conversion to Residential Use in the Hudson River
Valley: Scenario Analyses and Case Studies”. Agriculture and Human Values. Vol 22. No 4.
pp.377-393
Examples of using build out analysis for agricultural land conversion in the Hudson
Valley.