University of Campinas
School of Civil Engineering, Architecture and Urban Design
Vanessa Gomes is an awarded Associate Professor at the University of Campinas, UNICAMP. Architecture and Urbanist and Doctor in Civil Engineering from the Universidade de São Paulo (2003), she joined the Editorial board member of PARC Journal (2020) to stimulate research and publications on sustainability research applied to the built environment. Sustainable design and construction champion, she has participated in several international cooperations and exchanges and pioneered studies on the environmental assessment of buildings and life cycle assessment of construction products, elements and whole buildings. More recently, she has been dedicated to specifics of whole building LCA (protocol, uncertainty, inventory collection), assessments at neighborhood and urban scale, and consequential and prospective LCA.

Presenter Of 1 Presentation


Session Type
Academic Sessions
Session Time
01:00 PM - 02:15 PM

Hall B

Lecture Time
01:40 PM - 01:50 PM


Abstract Body

Life Cycle Assessment (LCA) is a data-intensive approach that has proved its value in environmental evaluation by providing decision-makers with high-quality, multicriterial comprehensive results. For its data intensity, LCA’s application at the urban scale is additionally challenged by the scale's intricacies, which amplify significantly the amount of data required. For a complexity compromise, data collection can be reduced to a manageable amount by subdividing the studied area into smaller urban cells such as neighborhoods, by grouping the built stock and infrastructure into a limited set of aggregates with similar characteristics, and by defining archetypes to represent each group. This paper compares four different data modeling strategies bearing different data collection needs to generate neighborhood LCA input. For the building stock, top-down (TD) modeling used different benchmarks based on buildings' use typology (education, health, office, or sports). Mid-down (MD) modeling assumed the characteristics of the most common archetype and pasted its simplified bill of materials for the whole built area. Mid-up (MU) modeling used the simplified bill of materials (only envelope) of nine representative buildings (archetypes) defined by unsupervised machine learning clustering, while the bottom-up (BU) approach used the complete bill of materials of nine archetypes. For the infrastructure modeling, TD used benchmarks for street, energy, and water networks and the BU approach used GIS-supported data. The materials and energy flows were extrapolated accordingly to define the neighborhood life cycle inventory for the six approaches. K-medoids clustering, ArcGIS, and SimaPro v.9 supported the assessments. Findings show that the average variation between MD and BU scenarios was 28.3%. From the thirteen categories considered, abiotic depletion (AD, 68.5%), human toxicity (HT, 46%), and photochemical ozone potential (PCOP, 39%) varied the most, while terrestrial ecotoxicity (TE, 12%), freshwater ecotoxicity (FE, 10%) and renewable primary energy (RPE, 5.5%) were less sensible to the modeling approach. Reinforcing steel and ceramic tiles were the main impact contributors. Despite its detailed flows’ modeling, carrying out detailed BU assessments can be unfeasible in practice, and the MU seemed to best balance modeling refinement and labor intensity.