Deconstruction Report
Deconstruction Report
Deconstruction Report
net/publication/336145148
CITATIONS READS
4 1,677
3 authors, including:
Andey Nunes
Portland State University
1 PUBLICATION 4 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Statistical Consulting for USGS Water Quality Monitoring Program View project
All content following this page was uploaded by Andey Nunes on 30 September 2019.
DEQ Materials
Management
700 NE Multnomah St.
Suite 600
Portland, OR 97232
Phone: 503-229-5696
800-452-4011
Fax: 503-229-5850
Contact: Jordan Palmeri
www.oregon.gov/DEQ
DEQ is a leader in
restoring, maintaining and
enhancing the quality of
Oregon’s air, land and
water.
Contact:
Jordan Palmeri
503-229-6766
[email protected]
DEQ can provide documents in an alternate format or in a language other than English upon request. Call
DEQ at 800-452-4011 or email [email protected].
• The sample of 36 homes had an average age of 112 years and average size of 1,177 square feet.
• The average deconstruction of a single-family home in Portland, Oregon, yielded 39,362 pounds
of material (excluding foundation), of which 10,587 pounds (27 percent) was salvaged. The vast
majority of salvaged material by weight was softwood lumber, in the form of framing lumber,
structural beams, and sheathing (shiplap on walls and plank subfloor). This material made up over
85 percent of the total weight of salvaged materials.
• The average deconstructed home showed a carbon benefit of 13.8 MTCO2e while demolition
showed a carbon benefit of 6.2 MTCO2e. Deconstruction yields a net carbon benefit of
approximately 7.6 metric tons of CO2eq per house compared to demolition. The carbon benefits
are mainly attributed to the avoided production of new materials and the continued sequestration
of biogenic carbon in the wood.
• Landfilling wood does result in a benefit for carbon storage but reuse of wood yields a benefit
almost twice as large.
• When considering biogenic carbon as an emissions source and sink, burning wood for energy
emits more carbon than it offsets when replacing natural gas as a fuel in industrial boilers.
• Although the end-of-life fate of recoverable wood greatly influenced the relative carbon benefits
of the deconstruction scenario, sensitivity analyses revealed that deconstruction will always have
a carbon benefit over demolition even with extreme swings in the market for recoverable wood.
• Results are less clear when looking at primary energy demand. The average deconstructed home
showed an energy benefit of 89 GJ, while demolition showed a benefit of 115 GJ, a difference of
26 GJ. Based on DEQ surveys of recovered and disposed materials, much of the clean
recoverable wood (56 percent) is used as a fuel that offsets natural gas use in industrial boilers
regionally. This pathway yields a large energy credit, which is contrasted by this pathway being a
net emitter of carbon.
• For energy impacts, the rate of wood incineration for energy recovery, which offsets the use of
natural gas, highly influenced the results. A decrease in the wood recovery rate from the current
56 percent to 30 percent would make the energy benefits of both the deconstruction and
demolition scenarios approximately equal.
• Material transport, worker transport and equipment use on site was analyzed in detail. Results
indicate that the impacts were inconsequential compared to much larger impacts of material
reuse, recovery and disposal.
During the first full calendar year of the policy, over 100 detached single-family homes were permitted
for deconstruction. This evaluation focuses on the first 36 homes deconstructed, due to timing and data
availability. Using actual material quantity data from these homes, combined with established material
impact factors and regional assumptions, the global warming potential (GWP) and primary energy
demand impacts of each deconstruction project can be evaluated. Comparing this to the ‘business as
usual’ scenario in which all houses are mechanically demolished, an estimation of the net impact of the
ordinance can be produced.
1.1 Background
1.1.1 Portland’s Deconstruction Ordinance
In October 2016, in an effort to reduce waste, support material reuse, and reduce environmental impacts
of demolition, the City of Portland enacted an ordinance requiring manual deconstruction of residential
homes built in 1916 or earlier, or designated as a historic resource regardless of age. Approximately 30-35
percent (by weight) of the landfill bound waste in Portland is construction and demolition debris.
Deconstruction and material reuse provides an opportunity to reduce this waste and help offset the
environmental impacts of producing new materials.
1.3 Definitions
In this report, “deconstruction” or “decon” means the systematic dismantling of a structure, typically in
the opposite order it was constructed, in order to maximize the salvage of materials for reuse, in
preference over salvaging materials for recycling, energy recovery, or sending the materials to the
landfill. Typically, material is removed by hand and processed (e.g., de-nailing) on site. Salvaged material
“Demolition” or “demo” refers to the removal of a structure using mechanized equipment (such as a track
hoe). Typically, this process results in much more physical damage to materials, preventing salvage and
reuse. All material from a mechanical demolition is assumed to be removed by dropbox and taken to a
material recovery facility (MRF).
The foundation of the house is excluded from the definitions of deconstruction and demolition in this
analysis, as foundation removal is not typically carried out by deconstruction contractors. The process for
removing and disposing of a concrete foundation is identical between deconstruction and mechanical
demolition.
“Dropbox” refers to a receptacle on a deconstruction or demolition site into which non-salvaged materials
are placed, to be sent to a MRF.
“End of life” (EOL) refers to the processes that occur after the useful life of a component or material,
including deconstruction, reuse, demolition, recycling and disposal. In this study “EOL fate” is used to
refer to the final stages for each respective material.
“Materials recovery facility” (MRF, pronounced “murf”), is a permitted solid waste facility where solid
wastes or recyclable materials are sorted or separated. There are many types of MRFs, and in this study
this may refer to facilities that receive general solid waste, only construction and demolition waste, or
only wood waste.
“Non-Recoverable wood” refers to wood that cannot be reused, recycled or used for energy generation,
such as painted, treated, dirty or rotten wood.
“Recoverable wood” refers to wood that can be reused, recycled OR used for energy generation. Note that
recoverable does not necessarily imply reusable (e.g., a small broken piece of trim may only be suitable
for energy generation).
2.1.1 Goal
The goal of this project is to calculate the carbon and energy impacts of deconstruction and demolition of
single-family houses in Portland, Oregon. The results will allow the City of Portland to measure the
effectiveness of their deconstruction policy in achieving climate and energy goals. The results may be
used by DEQ and others to provide policy guidance and recommendations regarding deconstruction.
2.1.2 Objectives
The objective for this project was to produce a report outlining the findings of this study, including:
1. The average salvage percentage by weight across all projects.
2. The net environmental impacts of 36 residential home deconstruction projects in the City of
Portland from October 2016 to December 2017.
3. The average home net benefit of deconstruction versus mechanical demolition for the 36
deconstruction projects in the City of Portland.
The maximum, minimum, median and average house age and size for the houses included in this study
are shown in Table 1. The newest house in the study was 90 years old – outside of the ordinance year-
built threshold, but required to deconstruct because it was designated a historic resource. The houses were
typical for the Pacific Northwest – timber-framed, with Douglas fir being the predominant species used
for framing, sheathing (shiplap), subfloors and many finished floors.
Material salvage and disposal receipts were collected for each deconstruction project to estimate the total
material quantity by weight (kg). The material weights were then multiplied by their respective carbon
and energy impact factors, which were developed using life-cycle-analysis (LCA) best practices.
Transport of materials, transport of workers, and the use of on-site diesel heavy-machinery equipment
was also included in the development of the project impacts.
For each deconstruction project, two modeling scenarios were conducted: 1) actual salvage scenario, and
2) hypothetical mechanical demolition scenario that assumed the materials would go to their typical EOL
fates based on Metro regional recovery and disposal data. Net impacts for carbon and energy were then
determined on a project level between each scenario modeled.
The assumed dimensions and weights for building components were compiled from literature, weighing
samples of salvaged building materials, and some professional judgement. These assumptions can be
found in Appendix A. Softwood density was sourced from a recent study of lumber properties for
salvaged wood coming from Portland-area deconstructed homes. The Oregon State University study
concluded that the average density of salvaged Doug-fir lumber from Portland-area homes is 530.7 kg per
cubic meter (Raphael Arbelaez, personal communication, October 2018). Hardwood density was assumed
to be 770 kg per cubic meter, the value for American white oak (International Timber, 2018).
Foundations were excluded from this study, as deconstruction typically does not include removal of the
foundation. Even if the foundation were included, there would be no difference between deconstruction
and demolition, as the process for breaking up and removing the foundation is the same in both scenarios.
Deconstruction contractors do not in practice record the relative quantities of each material type going
into dropboxes. The contents of these dropboxes were estimated using the average relative makeup of
construction and demolition (C&D) debris generated in the Portland metro area using 2016 DEQ Waste
Composition Study 1 and 2016 Material Recovery Survey 2 data and adjusted based on professional
judgement. The reason these figures were adjusted is that material coming from the demolition of old
homes is different from the average C&D waste generated in the Portland region, which is what DEQ’s
field surveys measure. For example, the average C&D waste stream contains more metal and cardboard
than was observed in the projects in this study. New construction generates cardboard—not residential
demolitions. Additionally, commercial construction and renovation generates more metal than residential
home demolitions. The average waste stream also contains higher amounts of roofing material than
typically seen on residential demolitions because a lot of roofing material is coming from reroofing
projects as opposed to demolitions. Overall, best professional judgement was used to adjust the dropboxes
to have a material composition that more accurately represents residential demolition sites.
Table 2 shows the total waste generated in 2016 for the material categories used in this study, and how
they were adjusted to better represent a residential demolition projects. The analysis treats the
composition of all drop boxes the same, which is discussed further in Section 6 “Limitations” in this
report.
1
https://www.oregon.gov/deq/mm/Pages/Waste-Composition-Study.aspx
2
https://www.oregon.gov/deq/mm/Pages/survey.aspx
Additionally, dropbox material has been simplified into four categories: ‘recoverable wood,’ ‘non-
recoverable wood,’ ‘metal,’ and ‘other’. The specific components that fall into these categories are shown
in Table 3.
Table 3: Assumed material composition of dropboxes and materials included in each dropbox
category
The impacts and offsets for each material, based on these respective EOL fates, was calculated using EPA
WARM v14, and the Ecoinvent 2.2 and GaBi v8.7 databases. The impact factors used have been included
in the accompanying Excel file to this report. The reason for ‘negative’ EOL carbon and energy impacts
(i.e., benefits) is due to the system expansion employed, in which recycled material and energy generated
from waste replaces virgin material and energy generated from fossil fuels, respectively. This benefit can
be larger than the emissions and energy consumed in the EOL process itself. In Portland, for example,
recovered clean wood is used for heat generation in paper and packaging plants, instead of natural gas. A
summary of these system expansion processes is given in Table 5.
Table 5: End-of-life fates and descriptions
3
Gabi Class 6 truck; Gabi class 8b truck
Roundtrip Total
Distance Number of Distance Truck Classification
Activity Material Transport Leg (miles) Trips (miles (Gross Vehicle Weight)
Deconstruction salvaged items site to retail 22 2 44 6 (19,501 - 26,000lbs)
Deconstruction dropbox site to transfer 10 3 30 8b (33,000lbs+)
Deconstruction recoverable wood transfer to wood end use 114 1 114 8b (33,000lbs+ )
Deconstruction metal transfer to metal end use 120 1 120 8b (33,000lbs+)
Deconstruction disposed transfer to landfill 200 1 200 8b (33,000lbs+)
Demolition dropbox site to transfer 10 4 40 8b (33,000lbs+)
Demolition recoverable wood transfer to wood end use 114 1 114 8b (33,000lbs+)
Demolition metal transfer to metal end use 120 1 120 8b (33,000lbs+)
Demolition disposed transfer to landfill 200 1 200 8b (33,000lbs+)
Figure 4: Relative percentage by weight of material removed by dropbox and salvaged from an
average deconstructed home
Of the salvaged materials, roughly 85% (by weight) was softwood lumber, followed by 3.4 percent
plywood, 1.3% interior wood doors, and 1.25 percent steel products (Figure 5). This indicates that the
vast majority of material salvaged from the deconstruction projects was wood – specifically softwood
lumber. This aligns with the fact that most houses of this generation in Portland use Douglas fir for
framing, plank subfloor, and sheathing (shiplap under the siding). Douglas fir flooring was also common
during this time period.
Using the broader dropbox material categories, the average quantity of each category was calculated on a
per-project basis (Figure 6). The largest of these categories was Recoverable Wood, with 11,798 lbs. This
indicates that there is a large amount of clean wood going to dropboxes with little economic value; this
could include pieces too small to sell, unconventional dimensions, or split/broken wood. This material is
suitable for energy recovery, but not for reuse.
The next category is Other, of which an average of 8,920 lbs per project was removed. This large figure is
due to non-salvageable material such as asphalt roofing, carpet, fiber cement and vinyl siding, any broken
ceramics, and damaged/broken windows. Non-Recoverable Wood (6,618 lbs) was the next largest
category, influenced by the fact that most houses in the study had painted wood siding which is not
suitable for reuse or energy recovery. The final and smallest category is metal (1,439 lbs), which includes
steel and cast iron plumbing, as well as aluminum siding and light fixtures.
Energy demand for an average home removal for the two scenarios is shown in Table 10. In this case, the
relative difference between the two scenarios is reversed. One of the main reasons for this is that wood is
a material with a relatively low energy intensity (by nature, its main energy input during production is
solar energy). The other main reason is that energy is recovered from much of the clean wood and offsets
energy produced other ways.
Much like the GWP impacts, the energy impacts of the material are by far the largest influence on total
results. Material transport makes a small impact to the totals, while the impact of worker transport and
equipment use is minimal.
Because the largest difference between the two scenarios is in recoverable wood, this material category is
split into the impacts of each EOL pathway in Figure 12. This shows more clearly that in the
deconstruction scenario less material is incinerated, and more material is reused, which has a greater per-
pound carbon benefit than landfilling. The carbon benefit of non-recoverable wood is identical in both
scenarios, as all of this material goes to landfill.
When looking at energy from each material category, both scenarios see a large benefit from recoverable
wood and metal (Figure 13). Metal, while contributing a relatively small amount by weight to the total
materials removed, shows a significant energy benefit, through both reuse and recycling. The ‘other’
category shows the benefit of reusing materials such as lights, doors and cabinets from deconstruction,
while this material uses energy when it is landfilled from demolition.
To investigate recoverable wood energy impacts further, the energy demand by EOL fate is shown in
Figure 14. This shows that while deconstruction and demolition show net benefits, this benefit is achieved
in different ways. Deconstruction achieves this benefit through reuse by offsetting the energy impacts of
new lumber production and gains even larger benefits when non-salvaged clean wood is sent for energy
recovery. The larger quantities of clean wood for energy recovery in the demolition scenario result in a
larger benefit for demolition.
Tables 12 shows that when reducing energy recovery of wood waste to only 30 percent of the waste
stream, demolition and deconstruction provide roughly equal energy benefits. This lower recovery rates
scenario may be more realistic for debris arriving in dropboxes and crunched together from residential
demolitions. From an energy perspective, the relative benefits of deconstruction increase as more
recoverable wood is sent to landfill and decrease when more wood is incinerated for energy recovery.
Table 11 shows that under any EOL fate scenario, deconstruction will always have carbon benefits over
demolition. However, the magnitude of those benefits change according to the EOL fate for recoverable
wood. Unlike the energy scenarios, the relative benefits of deconstruction decrease as more wood is sent
to landfill. This is because landfills store some of the carbon in wood whereas burning for energy
recovery has a net carbon impact because the carbon emissions from burning wood for energy exceed the
carbon savings from offsetting natural gas use.
Features to notice in the scatter plot include no pattern in point sizes and the relative positions on either
axis, for example a couple of smaller points appear to yield more salvage material than some larger
points. Color-based point differentiation by contractor reveals some clustering around the percent salvage
value labels. No clear trends are visible between dropbox quantity and house size. Instead, it is clear that
certain contractors are more able to salvage larger quantities of material per house.
A statistical correlation analysis found that there is no correlation (r = 0.1) between salvage quantity and
house age, and only a weak correlation (r = 0.41) between salvage quantity and house size. Further
statistical analysis can be found in Appendix B.
4
https://www.epa.gov/energy/greenhouse-gas-equivalencies-calculator
One limitation of using region-wide C&D waste data is that it is difficult to apply that data to specific
situations. Particularly of interest is the split between clean wood recovered for energy use and clean
wood landfilled. Transfer stations have hand-sorting lines for all waste, but still the 56 percent/44 percent
split of recovered/landfilled wood waste will come from a mix of clean, wood-only loads, and mixed
loads. It may be that deconstructed waste is easier to separate as it has been removed by hand, while
waste from a mechanical demolition site will have been crushed by heavy machinery. Unfortunately, the
exact amount of wood recovered for energy from a mixed dropbox load from each scenario is unknown.
7. Conclusions
Material quantity data from 36 manual deconstruction projects of single-family homes in the City of
Portland was analyzed to measure the carbon and energy impacts. The analysis resulted in the following
conclusions:
• The sample of 36 homes had an average age of 112 years and average size of 1,177 square feet.
• The average deconstruction of a single-family home in Portland, Oregon yielded 39,362 pounds
of material (excluding foundation), of which 10,587 pounds (27 percent) was salvaged. The vast
majority of salvaged material by weight was softwood lumber, in the form of framing lumber,
structural beams, and sheathing (shiplap on walls and plank subfloor). This material made up over
85 percent of the total weight of salvaged materials.
• The average deconstructed home has net carbon benefit of approximately 7.6 metric tons of
CO2eq per house compared to demolition. The carbon benefits are mainly attributed to the
avoided production of new materials and the continued sequestration of biogenic carbon in the
wood.
• Landfilling wood does result in a benefit for carbon storage but reuse of wood yields a benefit
almost twice as large.
• When considering biogenic carbon as an emissions source and sink, burning wood for energy
emits more carbon than it offsets when replacing natural gas as a fuel in industrial boilers.
• Although the end-of-life fate of recoverable wood greatly influenced the relative carbon benefits
of the deconstruction scenario, sensitivity analyses revealed that deconstruction will always have
a carbon benefit over demolition even with extreme swings in the market for recoverable wood.
• Results are less clear when looking at primary energy demand. The average deconstructed home
showed an energy benefit of 89 GJ, while demolition showed a benefit of 115 GJ, a difference of
26 GJ. Based on DEQ surveys of recovered and disposed materials, much of the clean
recoverable wood (56 percent) is used as a fuel that offsets natural gas use in industrial boilers
regionally. This pathway yields a large energy credit, which is contrasted by this pathway being a
net emitter of carbon.
• For energy impacts, the rate of wood incineration for energy recovery, which offsets the use of
natural gas, highly influenced the results. A decrease in the wood recovery rate from the current
56 percent to 30 percent would make the energy benefits of both the deconstruction and
demolition scenarios approximately equal.
• Material transport, worker transport and equipment use on site were analyzed in detail. Results
indicate that the impacts were inconsequential compared to much larger impacts of material
reuse, recovery, and disposal.
• There was no correlation between salvage quantity and house age. Correlation with house age
could become more pronounced if deconstruction is applied to newer houses that contain lower
quality/value material and are physically more challenging to deconstruct due to adhesives.
• Although there was little correlation between the quantity of material salvaged and house size the
study did find that salvage quantities were more closely tied with specific contractors, indicating
37
16 Appliances (including furnaces) and furniture (which includes medicine cabinets and ‘built-in
shelving’ not specified as wood in linear feet, square feet, board feed or lbs) are not included in
the raw data summary
38
flooring (sq_ft) hardwood flooring 0.083333 ft thickness
steps hardwood flooring 0.25 cu_ft per 12x36x1_inch step
doors outer door (solid wood) 46.83 kg per door
doors inner door (solid wood) 23.26 kg per door
doors inner door (hollow wood) 18.29 kg per door
windows window (single wood) 19.518 kg per (3'x4' = 12sq_ft)
window
windows window (double wood) 26.2764 kg per (3'x4' = 12sq_ft)
window
windows window (double vinyl) 29.3652 kg per (3'x4' = 12sq_ft)
window
cabinets cabinets (lower) 15.27 kg per lineal ft
cabinets cabinets (upper short) 8.21 kg per lineal ft
cabinets cabinets (upper long) 11.08 kg per lineal ft
light fixtures light fixture 5 lbs per fixture
softwood lumber softwood lumber 530.7 kg per cubic meter
hardwood lumber hardwood lumber 770 kg per cubic meter
39
Appendix B: Additional Figures
and Tables
Figure 16: Heat map diagram of material quantities salvaged from each project
Detailed breakout of the total material weight by the specific material types on each project, showing
larger yield with increasing color intensity. The greatest single material quantity from any project was one
project that reported ~1700 kg of plywood, while the greatest yield of any material from each project was
in the Softwood lumber category (which is mostly framing wood and some fir flooring).
40
Figure 17: Correlation of material salvaged (lbs) by house size (square feet)
Slight positive trend in kg of salvage materials with increasing house size, but still only a modest
correlation (0.4).
41
Figure 18: Correlation of materials removed by dropbox (lbs) by house size in square feet
Nearly flat (indicating no correlation) trend in disposal quantity over house size. Note, in both salvage and
disposal quantity, the confidence band is nearly 3x larger than the width of the band nearer to the median
values
Figure 19: Correlations of salvage and dropbox quantities by house size and age
42