Housing starts and the associated wood products carbon storage by county by Shared Socioeconomic Pathway in the United States

Abstract

Harvested wood products found in the built environment are an important carbon sink, helping to mitigate climate change, and their trends in use are determined by economic and demographic factors, which vary spatially. Spatially detailed projections of construction and stored carbon are needed for industry and public decision making, including for appreciating trends in values at risk from catastrophic disturbances. We specify econometric models of single-family and multifamily housing starts by U.S. Census Region, design a method for their spatial downscaling to the county level, and project their quantities and carbon content according to the five Shared Socioeconomic Pathways (SSPs). Starts are projected to decline across all scenarios and potentially drop to below housing replacement levels under SSP3 by mid-century. Wood products carbon stored nationally in structures in use and in landfills is projected to grow across all scenarios but with significant spatial heterogeneity related to disparate trends in construction across counties, ranging from strong growth in the urban counties of the coastal South and West to stagnation in rural counties of the Great Plains and the northern Rockies. The estimated average annual carbon stored in wood products used in and discarded from US residential housing units between 2015–2070 ranged from 51 million t CO2e in SSP3 to 85 million t CO2e in SSP5, representing 47% to 78% of total carbon uptake relative to uptake by all wood products in the United States in 2019.

The potential use of mass timber in mid-to high-rise construction and the associated carbon benefits in the United States

“…The estimated models, along with projections of predictor variables, were then utilized to project new nonresidential floor areas built annually in each of four U.S. regions over the next 50 years (2020–2070) under varying socioeconomic futures described by five SSPs [ 32 ]. We then used the projections of new floor areas built in multifamily residential units that were based on recently published projections of multifamily residential units under the same SSPs [ 22 ]. These individual projections of new floor areas built in nonresidential and multifamily residential buildings were then integrated with a new-product market diffusion model to investigate potential mass timber adoption levels under a range of market adoption scenarios encompassing low, medium, and high adoption to 100% adoption futures.…”

“…The estimates of the avoided emissions benefits of mass timber substitution were derived from a published whole building LCA (WBLCA) study that compared embodied fossil emissions of representative 8-, 12-, and 18-story mass timber and functionally equivalent reinforced concrete buildings in three U.S. regions covering raw material extraction (module A1), raw and intermediary material transportation (module A2), material production (module A3), finished material transportation (module A4), and construction stages (moduleA5) [ 28 ]. Similarly, the estimates of biogenic carbon stored in mass timber materials were modeled using a harvested wood product (HWP) carbon model [ 22 , 33 ] that tracked carbon added in and discarded from buildings. A description of each model and approach follows.…”

“…To account for model uncertainty given assumptions of GDP growth by SSP, we applied bootstrapping methods to project nonresidential floor area built and based our projections of floor area by SSP on the median projection for each region (see Appendix A in the S1 File ). For multifamily residential buildings, statistical models were available only for new structures built by region [ 22 ]. Therefore, to project floor area in new multifamily residential structures, we estimated the linear trends of average historical average floor area per unit of multifamily residential buildings by U.S Census Regions ([ 35 ], S3 Table in S1 File ), which we multiplied by the median projected structures of multifamily residential units under the same SSPs reported in Prestemon et al ([ 22 ], S4 Table in S1 File ).…”

“…For multifamily residential buildings, statistical models were available only for new structures built by region [ 22 ]. Therefore, to project floor area in new multifamily residential structures, we estimated the linear trends of average historical average floor area per unit of multifamily residential buildings by U.S Census Regions ([ 35 ], S3 Table in S1 File ), which we multiplied by the median projected structures of multifamily residential units under the same SSPs reported in Prestemon et al ([ 22 ], S4 Table in S1 File ).…”

“…The estimates of the avoided emissions (S7 Table in S1 File ) were based on the recently published WBLCA results [ 28 ] that provided an estimate of the embodied emissions of 8-, 12-, and 18-story functionally equivalent mass timber and reinforced concrete buildings in three U.S. regions, covering A1 (raw material extraction) to A5 (installation/construction) modules. The estimates of biogenic carbon stored in mass timber materials were based on a HWP carbon model that tracked the amount of biogenic carbon added each year as mass timber entered in new buildings, the amount of biogenic carbon discarded as buildings were demolished after their useful service life, and the biogenic carbon stored in landfills after wood materials were discarded in landfills [ 22 , 33 ]. The estimates of biogenic carbon stored in these pools (in existing buildings and in landfills) were based on an exponential decay function with assumed half-life of 100 years while in use.…”