Existing green space
and vacant parcel opportunity sites
City-wide building-level population dispersal
Current level of access to green space
Opportunity sites that would most efficiently increase access
Selecting Optimal Vacant Parcels to Increase Green Space Access
4-person team in UC Berkeley's environmental planning studio
6 weeks (1 in Portugal)
Aerial Imagery Processing
Public Access and Capacity Assessment
Maximize Capacitated Coverage Analysis
All maps and graphics by Allan Kapoor
Barreiro, Portugal has experienced population decline and economic stagnation due to deindustrialization over the last few decades. Many parcels in quaint historic neighborhoods are vacant or unmaintained and young people have fled the city.
But rents in nearby Lisbon are rising and a bridge across the Tagus Bay is under construction. Barreiro is poised for revitalization if it can attract young families back to the city.
My graduate environmental planning studio travelled to Portugal to work with officials from the City of Barreiro and faculty at the Technical University of Lisbon before returning to UC Berkeley to complete the project.
Conversations with officials and residents revealed an interest in increasing green space as a strategy to spur revitalization. Vacant parcels throughout the city could be inexpensively acquired and turned into municipal parks or informal open spaces.
Our studio team set out to assess the current level of access to green space and to determine the optimal opportunity sites for new parks that would most efficiently increase public access.
Our analysis showed that only 35.7% of Barreiro's population has sufficient access to green space.
The model recommended 10 currently vacant parcels that, if developed into parks, would increase access to 47.8% of the population.
GREEN SPACE & VACANT PARCELS INVENTORY
Besides CAD files of building footprints and streets, the City of Barriero did not possess any city-wide GIS data.
As a first step in our process, we processed multi-band satellite imagery to extract different categories of green space and then reviewed the results with municipal staff to ensure accuracy.
Existing Green Space
Organized Parks within the urban fabric
Non urban open space with a majority covered by tree canopy
Non Urban Open Spaces
Sparsely vegetated open space outside the urban fabric
Urban Open Space
Urban Open Space
APPROXIMATING BUILDING LEVEL POPULATION DISPERSAL
To accurately model green space access in a small city, we needed very granular population data. Barreiro's only population data was from the Portuguese Census at the subdivision level - comparable to US census blocks.
My team developed a method for approximating building-level population by allocating subdivision populations to the residential buildings within them based on the size of each building.
ASSESSING CURRENT ACCESS TO PUBLIC GREEN SPACE
Using ArcGIS Network Analyst we built a location-allocation model that assigned population points to green spaces within a 1/4 mile walking distance until they reach capacity, based on a World Health Organization recommendation for per capita urban green space.
Our results showed that even though 73.8% of Barreiro's population is within 1/4 mile waking distance of green space, if park capacity is factored in, only 35.7% have sufficient access.
SELECTING OPPORTUNITY SITES
To determine which vacant parcels the City should attempt to acquire for green space, we updated our model to perform a maximize capacitated coverage calculation that selects a combination of opportunity sites from vacant or underutilized parcels to increase access to the greatest number of people.
The model selected 10 sites from over 50 candidate parcels. The selected sites would increase public green space access to 47.8% of the total population.