**1. Introduction**

Sweden has committed to reducing greenhouse gas (GHG) emissions to net-zero by 2045 and to pursue negative emissions thereafter, in line with its obligations to the Paris agreemen<sup>t</sup> [1,2]. It is clear that the future development over several decades of the economic, social, and technical dynamics that govern demand for energy and materials, and the associated greenhouse gas emissions, are likely to be speculative. Nevertheless, as there is an urgen<sup>t</sup> need to start a transformation towards deep decarbonization, decisions must be made now as to how to best manage the transition, while taking the future into account [3]. This includes starting with the current situation to map mitigation measures to see which measures that can be applied already at present and those which will require longer lead times to be implemented.

Seeing that the energy and climate performance of the user phase of the built environment in Sweden keeps improving, the climate impact of the construction process has increasingly come in

to focus [4]. Emissions arising from manufacturing, transporting, and processing of construction materials to buildings and infrastructure account for approximately one-fifth of Sweden's annual CO2 emissions [5–7]. However, current estimates of the climate impact from building and construction processes in Sweden is associated with a significant degree of uncertainty. Environmentally extended input-output data has provided estimates for the year 2015. These determine territorial emissions associated with building construction to be 6.6 Mt CO2e, increasing to 11.6 Mt CO2e when including imports [7,8]. Territorial emissions linked transport infrastructure construction is similarly estimated at 1.5 Mt CO2e increasing to 1.9 Mt CO2e including imports [7,9]. The imported emissions are associated with greater uncertainty as they are estimated by calculating di fferences in emissions from trading partners compared to emissions in Sweden, giving the limitation of not capturing di fferences between di fferent industries in the importing countries [10]. On the other hand, a process-based bottom-up life cycle analysis (LCA) approach, combining statistics detailing new net area from newbuilds and refurbishments with LCA data per building type, provides a lower estimate of 5.4 Mt CO2e emissions associated with building construction in 2015 [8,11].

Indeed, as demonstrated in literature, there is evidence that life-cycle assessments based on process data and environmental extended input–output (EEIO) tend to lead to very di fferent results, where EEIO LCAs often lead to higher emissions and process LCAs to lower emissions [12]. There are several reasons for these discrepancies, with EEIO LCAs su ffering from inherent homogeneity and linearity assumptions, along with aggregation errors due to several di fferent industries being comprised into one input-output sector [12,13]. The combination multiple economic subsectors with quite di fferent emissions profiles into one sector, along with the assumption that the market price linearly correlates with higher emissions results in systematic overestimations [14]. On the other hand, process LCA su ffers from an inherent 'truncation error' due to indirect impacts (e.g., capital goods) or excluding upstream processes along the supply chain due to the need for a system boundary leading to systemic underestimation [14–16]. Comparative building case studies demonstrate 20–73% higher embodied carbon emissions for EEIO LCA versus process LCAs [12,17–19].

In view of the di fferences in the LCA approaches, several studies regard EEIO methods most useful in assessments of entire economies or industries [13,20,21]. We conclude that, to enable analysis into the ongoing development in the construction sector and the opportunities for the sector to contribute to the national climate targets, better estimates are needed, including the main components making up those emissions, from di fferent materials to transport of those materials and construction processes.

The focus of this study was on the path towards net-zero emissions in 2045, which necessitates not only looking at current emissions and the components therein but also require comprehensive assessments into current, as well as prospective future, abatement options and potentials. In literature, one can find an array of sector-specific or industry level studies focused on future carbon abatement options (see, e.g., Reference [22–26]) for steel, Reference [27–29] for cement/concrete, and Reference [30–33] for heavy transport and construction equipment). A comprehensive review of 40 energy-intensive industry roadmaps was recently performed by Gerres et al. [34]. This review remarked that roadmaps with a focus on subsector specific technology assessments often disregard the cross-sectorial dimensions of the abatement options considered, while top-down approaches tend to provide limited details on technological and economic feasibility. Gerres et al. found little consensus on how deep decarbonizations of industry are to be achieved but could identify a few key areas of importance and agreement, including alternative feedstock and carbon capture in the cement industry, carbon neutral steelmaking, and decarbonization of low temperature heat in the petrochemical industry. The authors finally noted that carbon capture, transport and storage (CCS), the electricity system, and the hydrogen economy, i.e., external system transformations, must be considered when evaluating decarbonization pathways.

In addition to sector-specific abatement studies, we have found recent evidence, particularly in grey literature of synthesis reports, reports which integrate the perspectives from di fferent industries [35–41]. The target of these reports is predominantly either a European or a global level, emphasizing the

cross-sectorial potential of reducing demand for products and services via circular economy, logistic optimization, and material e fficiency measures while highlighting the potential and alternatives contributed by biomass, carbon capture, and electrification, including links to hydrogen.

Thus, we see that roadmaps detailing industry decarbonization on a sector by sector or multinational level are prevalent. However, focusing in on the building and construction sector, there are limited examples in literature of national assessments of future abatement options and potentials and the pathway towards close to zero emissions [42,43], with most studies pertinent to the UK [44–47].

In Sweden, within the government-initiated Fossil Free Sweden (http://fossilfritt-sverige.se/inenglish/) initiative, business industries have drawn up roadmaps towards 2045, describing in varying details technological solutions, investment needs, and obstacles required to be removed. These provide some key information on abatement options within individual industry sectors with the construction sector roadmap capturing a cross-sectorial perspective [48,49]. Some initial assessments have also been made on emissions reductions and energy needs on a cumulative level for the year 2045 [50,51]. However, to explore critical factors on the pathway towards 2045, including impacts from upscaling and the risk of lock-in e ffects, there is a need for studies that take a broader perspective while combining a short and long-term perspective of abatement potential across the supply chain.

In this study, we used material and energy flow analysis combined with an extensive literature review to assess (i) the current status of emissions from the Swedish construction sector and (ii) the extent to which abatement technologies across the construction supply chain could reduce the GHG emissions if combined to its full potential based on implementation timelines linked to their technical maturity and expected readiness for implementation. The ambition was to analyze the current and future GHG emissions reduction potential by considering the development, over time, of emission abatement measures in di fferent parts of the construction supply chain.

With support of scenarios, we created a roadmap exploring di fferent future trajectories of technological developments in the supply chains for buildings and transportation infrastructure. By matching short-term and long-term goals with specific technology solutions, the roadmap made it possible to identify key decision points and potential synergies, competing goals, and lock-in e ffects. While the study is performed in a Swedish setting, and the updated estimate of current emissions are predominantly based on northern European LCAs, the analysis of abatement options, timelines, and pathways are relevant and applicable on a European, if not a global level.

## **2. Materials and Methods**

This work combines quantitative analytical methods, i.e., scenarios and stylized models, with a participatory process involving relevant stakeholders in the assessment process. The participatory process served to identify the main abatement options but also to adjust decisions and assumptions regarding abatement portfolios and timelines to make these as realistic and feasible as possible. Stakeholders have thus provided input and feedback via workshops undertaken during the study development period. Stakeholders include industry representatives and experts along the supply chain: material suppliers, contractors, consultants, clients, and governmental agencies.

Estimates are provided of the magnitude of current and future GHG emissions reduction potential across the building and transport infrastructure construction supply chain by (i) estimating the current emissions, material, and energy flows associated with the sector; (ii) identifying possible GHG abatement options relevant to the construction works and their estimated abatement potentials; (iii) using (i) and (ii) to assess the impact of combining abatement measures along the construction supply chain; and (iv) crafting scenarios to highlight challenges and possibilities up to 2045 given di fferent assumptions regarding future practices and technological development.

Current emissions from the Swedish building and construction industry is analyzed by comparing existing estimates with a mapping of the material and energy flow through the supply chain of building and transport infrastructure construction produced via a literature review of life cycle analyses and

equivalent studies (where literature searches were conducted in Scopus and Web of Science with search string algorithms targeting a combination of LCA OR "life cycle analysis" OR "life cycle assessment" OR "carbon footprint" AND building\* OR construction OR infrastructure with subsequent screening to identify studies of relevance for the scope of this study, e.g., transport infrastructure and buildings of equivalent design and construction techniques, as in Sweden.). In the technology roadmap of this work, we analyze the climate impact linked to construction of buildings and transport infrastructure, i.e., we do not include construction of for example utilities, such as waterworks, wastewater treatment plants, power plants, and power lines. Construction of buildings and transport infrastructure is equivalent to around 80% of construction investments in Sweden [52]. Focus of the analysis is on emissions from materials production and the construction phases (i.e., corresponding to life cycle stage A1–A5 [43]). The latter includes emissions from mass and material transport and the construction process. A schematic of the mapping is shown in Figure 1.

**Figure 1.** Schematic figure of material flow mapping for buildings and transport infrastructure construction in Sweden. The height of the category to frame-type boxes represent the approximate relative sizes of the associated emissions. Regarding materials, the dark orange boxes depict materials studied in detail, while the dark orange contours in the primary production column depict material production processes evaluated in detail. The analysis also includes emissions from mass and material transport and construction processes.

The Swedish Transport Administration (STA) provides a breakdown of the emission share from various materials and activities regarding new construction of state-owned transport infrastructure. However, this is not a complete picture of transport infrastructure as around half of the transport infrastructure investments in Sweden are made by regional and local governmen<sup>t</sup> [53]. More detailed analysis has been performed by [9], including both state, municipal, and privately-owned transport infrastructure. The analysis by Liljenström et al. describes the emissions share of material production and on-site activities (transports and construction processes) for both new construction and reinvestments (defined as larger projects intended to restore the infrastructure to its original state by replacing a construction component (for example, the bounded base layer and tunnel lining) with the same, or a similar, type of construction component.) for road and rail infrastructure, ports, and fairways and airports. In this study, we slightly refined the emissions shares given by Liljenström et al. based on additional data [52,54], while excluding airports due to the minor emissions associated with airport construction (0.03 kt CO2e [9]). We further used the total emissions for construction of transport infrastructure provided by the detailed bottom-up analysis performed by Liljenström et al. [9] and national environmentally extended input-output modeling [7], as these provide a coherent result of 1.9 Mt CO2e emissions for the year 2015.

As this coherence does not apply for building construction, an estimate for the national emissions associated with building construction was developed using data on the emissions share from di fferent lifecycle stages and materials sourced from the literature review combined with validated emissions levels of di fferent components. Where available, the literature review was concentrated to LCA studies in a Northern European setting as to account for equivalent design and construction techniques, along with requirements stemming from climatic conditions. While LCA studies of buildings are prevalent, studies that describe and separate material inputs, material transports, and construction processes are more limited, particularly regarding non-residential buildings and refurbishments (see, e.g., reviews in Reference [55–57]). As LCA studies are limited for refurbishments, no detailed breakdown for refurbishments has been developed here. We instead use an adjustment factor to reflect emissions from transports and specific materials considered dominant in refurbishments in the few studies available.

The share of emissions for specific materials for construction of di fferent building types was calculated based on the estimates in literature for these building types and the estimated share of emissions per building type. The total share of emissions for di fferent material/activities for building construction were subsequently calculated using estimates for di fferent life cycle stages for the various building types.

The compilation of material, energy, and emissions flow serves as the baseline when applying identified abatement potentials from the abatement options review. The inventory of GHG abatement options (described in detail in Section 2.2) is established by means of a comprehensive literature review, including industry and governmental agency reports (grey literature), together with input from supply chain stakeholders. (Literature searches were conducted via a combination of academic bibliometric databases (Scopus and Web of Science) and web browser searches was used to enable the sourcing of the relevant grey literature, which is not as evident in academic bibliometric databases. Search string algorithms targeted a combination of the material/activity in question together with "carbon emissions" OR CO2 OR GHG OR "greenhouse gas emissions" AND abatement OR "emission\* reduction" OR mitigation OR decarbonization.), The main types of abatement options considered in the assessment are material e fficiency and optimization measures together with shifts in: material production processes, transport vehicles and construction equipment technologies, and fuel substitutions in both equipment and production plants. The options include certain reuse and recycling measures resulting in emissions reductions, but not for the specific purpose of resource conservation. The inventory comprises both current best available technology and technologies assumed to be available over time to 2045.

A timeline is applied to test the potential implications to the climate impact when constructing the same assets while applying a combination of GHG abatement measures along the supply chain appraised to have reached commercial maturity at di fferent points in time (over 5-year time periods until 2045). From this inventory, portfolios of abatement measures for the respective supply chain activities are constructed with selections of measures applied on a timeline up to the year 2045. The abatement measures are combined in pathways according to strategic choices [58], namely access to biofuels and renewable electricity, as well as enactment of material e fficiency measures (as described in detail in Section 2.3).

The analysis assumes emission factors for electricity and district heating declining in accordance with scenario analysis from the Swedish Energy Agency, implying that GHG emissions related to electricity generation are close to zero in 2045 [59].

#### *2.1. Pathway Generation and Quantification Approach*

Total emissions from buildings and infrastructure construction in each time period t is calculated as:

$$E\_{\text{tot},t} = E\_{b,t} + E\_{ti,t} \tag{1}$$

where *Eb*,*<sup>t</sup>* is the emissions resulting from building construction, and *Eti*,*<sup>t</sup>* is the emissions resulting from transport infrastructure construction. The analysis includes emissions from materials production and the construction phase (i.e., corresponding to life cycle stage A1–A5 according to EN 15978 [60]), with the latter comprising emissions from mass and material transport, and the construction process (A4 and A5, respectively).

#### 2.1.1. Emissions from Transport Infrastructure Construction

The transport infrastructure construction emissions, *Eti,t*, are calculated as the sum of emissions from the material production stage and the construction activities as:

$$E\_{\rm ti,t} = \sum\_{m} (E\_{\rm ti,m,t}) + \sum\_{\rm ttc} (E\_{\rm ti,tc,t})\_{\rm t} \tag{2}$$

where *Eti*,*m*,*<sup>t</sup>* is the emissions associated with material production of material *m* in timestep *t*; and *Eti*,*tc*,*<sup>t</sup>* is the emissions for construction activities *tc* in timestep *t*. Construction activities *tc* comprise mass and material transport and the construction process. Five material categories (*m*) are included in the analysis: concrete, reinforcement steel, construction steel, asphalt, and others.

The share of emissions from transport infrastructure construction coming from materials production and the construction process activities, respectively, in the base year, year 2015, is based on data from the Swedish Road Administration [53,61] and Liljenström et al. [9]. The emissions *Etc*,<sup>2015</sup> from the construction activities, *tc*, in the base year, year 2015, is calculated as:

$$E\_{\rm ti,tc,2015} = E\_{\rm ti,2015} \* \sum\_{i,c,tc} (\varepsilon\_{i,c} \* \varepsilon\_{i,c,tc})\_{,} \tag{3}$$

where *Eti*,<sup>2015</sup> is the total emissions from transport infrastructure construction in 2015; *ei*,*<sup>c</sup>* is the share of emissions from transport infrastructure type *i* (i.e., road, railway, ports, and fairways) and construction type *c* (i.e., new construction and reinvestment); *ei*,*c*,*tc* is the share of emissions from transport infrastructure type *i*, construction type *c* and construction activities *tc*.

Correspondingly, emissions from material production are calculated as:

$$E\_{m,2015} = E\_{ti,2015} \* \sum\_{i,c,m} (e\_{i,c} \* e\_{i,c,m}) ,\tag{4}$$

where *Em*,<sup>2015</sup> is the emissions from material production in 2015 for the specific material *m*; *Eti*,<sup>2015</sup> is the total emissions from transport infrastructure construction in 2015; *ei*,*<sup>c</sup>* is the share of emissions from transport infrastructure type *i* and construction type *c*; *ei*,*c*,*<sup>m</sup>* is the share of emissions from transport infrastructure type *i*, construction type *c* and material *m*.

#### 2.1.2. Emissions from Building Construction

The building construction emissions are also calculated as the sum of emissions from the material production stage and the construction activities:

$$E\_{b,t} = \sum\_{\text{lm}} \left( E\_{b,m,t} \right) + \sum\_{\text{tc}} \left( E\_{b,\text{tc},t} \right), \tag{5}$$

where *Eb*,*m*,*<sup>t</sup>* is the emissions associated with material production of material *m* in timestep *t*; and *Eb*,*tc*,*<sup>t</sup>* is the emissions for construction activities *tc* in timestep *t*. The analysis covers seven material categories, *m*, including: concrete, reinforcement steel, construction steel, insulation, gypsum and plaster, plastics and paint, and others (glass, aluminium, and wood).

For the base year of 2015, validated emissions for construction equipment (as per data from the national EEIO data reported in [7]) was used to extrapolate total building construction emissions:

$$E\_{b,2015} = \frac{E\_{cp,2015}}{\varepsilon\_{cp}},\tag{6}$$

where *Eb*,<sup>2015</sup> is the total annual emissions associated with building construction and refurbishment in 2015; *Ecp*,<sup>2015</sup> is the emissions estimate for construction equipment in 2015 according to the national EEIO data; and *ecp* is the emissions share estimated for construction processes.

The construction equipment data from the national EEIO data is considered reliable as construction equipment contribute to domestic emissions only and is used in construction and refurbishments (and not in operation of buildings). Once the total emissions estimate is produced, it is validated by means of comparing the resulting emissions for specific materials with available data to confirm its feasibility.

The share of emissions for the construction processes, material transports and material production were calculated using estimates for different life cycle stages for various building types

$$c\_{\rm lc} = \sum\_{i=0}^{n} \left( c\_{i} \* c\_{\rm lc,i} \right) \tag{7}$$

where *elc* is the emissions share associated with the different life cycle stages *lc* (equivalent to A1–A3 for material production, A4 for material transport, and A5 for the construction process according to the European standard for "Sustainability of construction works - Assessment of environmental performance of buildings" (EN 15978)); *ei* is the emission share for building type *i*; and *elc*,*<sup>i</sup>* is the emissions share for life cycle stage *lc* and building type *i*. The analysis covers three building types, *i*, including: multi-family dwellings, single-family dwellings, and non-residential buildings.

The share of emissions for different materials for construction of different building types were calculated based on the estimates in literature for these building types and the estimated share of emissions per building type. Where available and most applicable (i.e., for multi-family dwellings), the building type was also divided into building typology and frame type, namely concrete frame and wood frame. The emissions share *em* associated with material production of the material m is thus calculated as:

$$c\_{\mathfrak{M}} = \sum\_{i=0}^{n} (c\_i \* c\_{m,i})\_{,} \tag{8}$$

where *ei* is the emission share for building type *i*; and *em*,*<sup>i</sup>* is the emissions share for material *m* and building type *i*.

The initial estimated shares for both life cycle stages and materials were subsequently amended based on validated data for certain components in combination with adjustments for materials commonly used in refurbishments.

#### 2.1.3. Material and Energy Demand

Emissions and energy intensity factors for materials, activities, and fuels were combined with the emissions figures to estimate material and energy demand. The emission intensity factors for materials, activities, and fuels, along with data for associated quantity and source of energy used for material production, were sourced in a literature review. Table 1 lists the details for the reference energy carriers, materials or material combinations used in the calculation of material and energy demand for the construction of buildings and transport infrastructure in the year 2015. Details on specific materials, material production processes, and energy sources can be found in Tables A1 and A2 in Appendix A.

**Table 1.** Emissions and energy intensity factors along with energy mix in the production of reference materials and energy carriers used in the construction of buildings and transport infrastructure in the base year of 2015.


The specific emission intensity figures were combined with emission shares to calculate the resulting material and energy demand. Accordingly, the material demand *Mm* for each material *m* for the base year of 2015 is calculated as:

$$M\_m = \frac{\left(E\_{b,2015} + E\_{ti,2015}\right) \* c\_m}{E f\_m} \,\tag{9}$$

where *Eb*,<sup>2015</sup> is the total annual emissions associated with building construction and refurbishment; *Eti*,<sup>2015</sup> is the total annual emissions associated with construction of transport infrastructure; *em* is the emissions share associated with material *m*; and *E fm* is the emission intensity factor associated with material production of the material *m*.

The energy demand for material transports and construction processes for the base year, year 2015, is calculated as:

$$Q\_{tc} = \sum\_{ts} \frac{\left(E\_{b,2015} + E\_{ti,2015}\right) \* c\_{tc,s}}{E f\_s} \tag{10}$$

where *Qtc* is the energy demand for construction activities *tc*; *etc*,*<sup>s</sup>* is the emissions share associated with energy source *s* for construction activity *tc*; and *E fs* is the emission intensity factor for energy source *s*.

The total energy demand per energy source is calculated by using energy intensity factors combined with the energy mix data for production material processes and fuels used in transport and construction processes:

$$Q\_{\rm lot,s} = \sum\_{m} (M\_m \ast Qf\_m \ast q\_{\rm m,s}) + \sum\_{\rm tc} Q\_{\rm tc} \ast q\_{\rm tc,s} \tag{11}$$

where *Qtot*,*<sup>s</sup>* is the total energy demand associated with energy source *s*; *Mm* is the material demand of each specific material *m*; *Q fm* is the energy intensity associated with the production of material *m*; *qm*,*<sup>s</sup>* is the share of energy in the material production of material *m* of the energy source *s*; *Qtc*,*<sup>s</sup>* is the energy demand for construction activity *tc* and energy source *s*; and *qtc*,*<sup>s</sup>* is the energy share in the reference fuel used for construction activity *tc* of the energy source *s*. Three energy sources are detailed in the analysis: fossil fuels (coal, gas, oil, and fossil waste), biofuels, and electricity.
