1. Introduction
We are currently facing the threat of climate change, largely driven by anthropogenic environmental impacts [
1]. Steffen et al. [
2] identify 24 socio-economic and Earth system indicators that demonstrate unprecedented human growth coupled with changes to the natural environment that have occurred in the 20th century. They refer to the post-1950 acceleration of Earth system indicators as the ‘Great Acceleration’, demonstrating human impact on the Earth’s systems. The built environment impacts all the Earth system processes, which contributes to the accelerated rate of change we are experiencing. Of particular concern is the built environment’s contribution to greenhouse gas (GHG) emissions, accounting for 36% of global final energy use and 38% of energy-related carbon dioxide (CO
2) emissions in 2019 [
3]. While the life cycle emissions from buildings have been extensively studied, e.g., [
4,
5,
6], the embodied emissions are often underrepresented, shifting the focus away from the construction phase [
6,
7,
8,
9]. Energy-efficient buildings have higher primary energy use for construction and a lower energy requirement for the use phase compared to conventional buildings [
10]. With policy driving an increase of use phase energy efficiency in buildings [
11], embodied energy becomes more significant. It has become increasingly recognized that reducing use phase impacts should be accompanied by equal attention to embodied impacts [
6,
12]. To accurately inform policy decisions, more research into embodied impacts from buildings is necessary.
Currently, the most widely used building materials, such as concrete and steel, have high embodied GHG emissions. To minimize embodied emissions and ensure that the positive benefit from increased operational efficiency is not nullified or postponed too far into the future, a transition away from typical high-carbon building materials is essential [
4,
7,
13,
14]. It is widely accepted that life-cycle assessment (LCA) is the most comprehensive approach for assessing the environmental impacts associated with buildings [
6,
15,
16]. Much of the focus of building LCA studies has been on traditional building materials, such as concrete, steel, and wood [
6], though the positive benefits from substituting conventional building materials with less carbon-intensive alternative materials have been shown to reduce the embodied emissions of new construction projects [
17,
18,
19,
20]. There are still significant knowledge gaps about the embodied emissions associated with buildings constructed with non-conventional materials. Fiberglass is one of these new, and so far, not widely utilized alternatives to conventional materials. To date, no LCA studies assess the environmental impacts of fiberglass panels as the main building components in the context of buildings. In a 2020 study, Işildar, Morsali, and Zar Gari [
21] assess the environmental impacts of fiberglass-reinforced polymer (GFRP) in the construction context when they compare GFRP to traditional steel rebar. When comparing the environmental impacts from the manufacturing of the same quantities of both GFRP and steel, they report higher overall impacts in most impact categories for GFRP, particularly respiratory inorganics and fossil fuels. The authors call for more research into the impacts and lifetime of this material in the construction industry, given the limited understanding we currently have [
21].
Many LCA studies on buildings primarily focus on global warming potential (GWP), representing GHG emissions, e.g., [
22,
23]. However, this is not the only harmful environmental impact associated with buildings [
15]. Other impact categories are less frequently reported, and relatively few studies report those same impact categories. Anand and Amor [
24] point out that the justification for the choice of impact categories is often not clear in many building LCA studies, presenting challenges for comparison. To address the gap in literature regarding reported impact categories other than GHG emissions highlighted by Khasreen et al. [
25], more transparent and harmonized reporting of various environmental impacts from buildings is necessary to establish a baseline for comparison.
Accounting for uncertainty is a key challenge for LCA, despite efforts to classify and define it, as well as to develop estimation methodologies [
26,
27]. Applying uncertainty in LCA studies is not common and is a limitation when interpreting results [
28]. According to Bamber et al. [
28], less than 20% of all LCA studies published in the last 5 years included uncertainty analyses of any kind. Even less reported was propagating uncertainty with a mathematical approach [
28], such as Monte Carlo [
29], though it is the most common method for mathematical uncertainty analysis in LCA e.g., [
30,
31,
32].
This paper aims to contribute to the research gaps identified above. The embodied emissions and environmental impacts from a newly designed fiberglass modular house in Iceland, employing non-traditional building materials and techniques, are presented within the LCA framework. A wide range of environmental impacts are reported utilizing the ReCiPe midpoint and endpoint impact assessment method. Based on the midpoint results, we identify the main hotspots to show where improvements need to take place if this material is to be used to reduce the embodied emissions in comparison to the conventional options. The climate change impact is put into life cycle perspective by presenting use phase energy use scenarios for three different grid conditions (Iceland, Finland, and Poland) and two different energy efficiency levels. Additionally presented is an uncertainty analysis based on Monte Carlo. The results of this study suggest that fiberglass does not provide a viable alternative to conventional building materials if the purpose is to reduce embodied emissions from buildings, although the uncertainty ranges are significant, calling for more research on this material.
4. Discussion and Conclusions
This study aimed to quantify the embodied environmental impacts of a fiberglass house in Iceland and evaluate the environmental potential of fiberglass as a replacement for conventional high-impact materials. As identified in the literature review, there is a gap in the LCA literature addressing the impacts of unconventional building materials, specifically fiberglass. The pre-use phase has also been identified as an emissions hotspot. This is often underexplored in building LCA literature [
6] but actually causes a high emissions “spike” during a short time period. Comparatively, the use phase emissions accumulate over a long time and might easily be overestimated when calculated as if they occurred at the same time as the pre-use phase emissions [
7]. Furthermore, while GHG emissions have been the focal point of the climate discussion, other environmental impacts are important to consider, especially in light of the fact that we have already crossed, or are fast approaching, critical planetary boundaries [
61]. More widespread reporting of midpoint and endpoint impact categories, as shown in this analysis, is necessary to develop baseline results for comparison.
This study was a process LCA, modeled in the OpenLCA software using the Ecoinvent 3.7.1 inventory database. The ReCiPe impact assessment method was used to calculate the environmental impacts of a new building solution for eighteen midpoint categories and three endpoint categories. Several key findings were made based on this assessment.
First, midpoint results showed the GWP for the pre-use phase was 311 kg CO
2eq/m
2. This is similar to typical concrete buildings [
6] and well above natural materials-based buildings and houses, particularly if the carbon storage is accounted for [
6,
13]. The case example by Dabaieh et al. [
19] shows how the carbon storage potential of natural materials has the potential to significantly reduce the associated GHG emissions though the potential benefit is determined by the end-of-life assumptions. The carbon storage potential of natural materials and treatment of end-of-life is a developing field [
16], therefore, more research is necessary to accurately account for this benefit; however, their superiority in terms of GHG emissions seems to be a relatively conclusive finding.
Table 8 and
Figure 9 present a comparison to selected previous detached house LCA studies.
The results for the fiberglass panels as frame and wall elements were also compared to a selection of previous studies in which the same information was available.
Table 9 lists the studies and the materials of the compared structures and
Figure 10 shows the results comparison. As can be seen, the GWP impact from the frame/wall structures of the fiberglass house falls to the higher end of values reported from the other studies.
The use phase energy analysis was conducted as a simplified comparison of two different energy efficiency levels in three different locations to put the embodied GWP into context. This analysis showed that the embodied emissions are extremely high and dominate the life cycle emissions over the 50-year period, unless the grid emissions are very high, as shown by the Poland case. Even in the case of the Poland location, the embodied component is significant. We did not try to estimate the increase of the embodied emissions along with higher energy efficiency, but Säynäjoki et al. [
7] provided an estimate of a 5% increase of the embodied emissions when increasing the energy efficiency from business as usual to a passive house standard level. User influence on the use phase energy consumption is also significant, and inversely connected to the energy efficiency level: the worse the energy efficiency, the higher the fiscal incentive for energy-saving behavior, and vice versa [
69]. The magnitude of this impact is also tightly connected to the energy prices and the affluence of the users.
Other impact categories were reported to demonstrate the wide range of environmental impacts from the fiberglass house, beyond just GHG emissions; though it is difficult to draw conclusions from these results because a comprehensive baseline in the context of buildings does not exist. Previously, toxicity impacts were associated with high uncertainty due to limited understanding [
70,
71]. Harmonization in the assessment of toxicity-related impact categories is ongoing [
72,
73,
74]; therefore, accuracy in the context of LCA is improving, though there is still uncertainty with these impact categories. Other common environmental impacts, such as global warming, terrestrial acidification, freshwater eutrophication, fine particulate matter formation, mineral resource scarcity, and fossil resource scarcity, are associated with less uncertainty due to a better understanding and more reporting; therefore, these results are considered to be more reliable [
75], and baselines for buildings can be developed along with more reporting.
The endpoint results were shown to be 1.07 × 10−03 DALY/m2 for damage to human health, 1.44 × 10−06 species.yr/m2 for damage to ecosystems, and 3.1 × 10 USD2013/m2 for damage to resource availability. Endpoint results are rarely reported or discussed, so it is impossible to position these results; however, with more reporting, baselines can be developed.
Next, the fiberglass material was identified as an environmental hotspot in most impact categories. Fiberglass had the highest contribution to fossil resource scarcity, stratospheric ozone depletion, fine particulate matter formation, freshwater eutrophication, climate change, ionizing radiation, land use, ozone formation—human health, ozone formation—terrestrial ecosystems, terrestrial acidification, and terrestrial ecotoxicity. Improvements to the fiberglass material quantity and composition could result in a better-performing material in the context of buildings. More research into solutions for reducing the environmental impact of fiberglass is necessary to guide the development of fiberglass as an alternative building material.
Steel and copper were also found to be hotspots. Steel contributed most of the impact to human carcinogenic toxicity and had the highest contribution to marine eutrophication and mineral resource scarcity. Copper contributed the most to freshwater ecotoxicity, human non-carcinogenic toxicity, and marine ecotoxicity. It also had a significant contribution to freshwater eutrophication, mineral resource scarcity, and terrestrial ecotoxicity. Based on the hotspot results, the use of these materials, along with the fiberglass, should be the target for improvement in future development of this alternative housing option.
There are significant uncertainties in this study that should be taken into consideration when interpreting the results. The fiberglass material is associated with high uncertainty, largely due to a limited understanding of the material in the context of buildings. This contributed to the overall high uncertainty of the results. The uncertainty analysis conducted in the study focused on the midpoint results. For some impact categories, such as climate change, human non-carcinogenic toxicity, and stratospheric ozone depletion, the wide range between the 5th and 95th percentile represent highly uncertain data. However, the bulk of the estimated values are much nearer to one end in the skew distributions, indicating that the practical uncertainty is generally lower than the total variability shows. This can be attributed to the uncertainty factors assigned to the foreground processes, which did not closely represent the LCI. While processes were modified to reflect the actual production data as close as possible, for example energy mix in production countries, location-specific data were not available in the Ecoinvent database for many processes. Often, the available data are based on average or single case values [
15], limiting the representativeness of the results presented. Additionally, the Ecoinvent database did not include data for specific materials, such as the acrylic resin used in the fiberglass; therefore, the closest alternative within the constraints of the database was used: polyester resin. The quantities used for modeling 1 kg of fiberglass in Ecoinvent are shown in
Appendix A Table A1, and the modeling details in
Appendix A Table A2. There are no LCA studies on acrylic resin, so it is not known how the impacts from acrylic resin compare to the polyester resin reflected in this study, resulting in high uncertainty. Polyester resin is the most common resin used in fiberglass production [
76] and used in the general process for fiberglass in Ecoinvent ‘glass fiber reinforced plastic production,’ so it is justified in this case. Based on these uncertainties, the results do not represent the exact environmental impacts of the case study, but are more intended to report the potential impact of fiberglass panels as a new building solution. More research on fiberglass material is necessary to minimize uncertainty.
It has also been suggested that the results are strongly connected to the selected LCA database. Emami et al. [
56] assessed the emissions related to the construction of a wooden detached house and a concrete structure apartment building with both SimaPro/Ecoinvent and with GaBi, and found the estimates by the two to be highly different for both.
The next limitation relates to the LCI. Available data for the LCI were provided by the producer of the fiberglass house, but it was not comprehensive, which limited the system boundary, though this is a common limitation with LCA. In this study, the system boundary was selected based on processes that have significant mass. Materials with insignificant mass (<0.002%) were excluded for the pre-use phase. Suh et al. [
77] highlight that there are limitations when only using mass to determine the system boundary because it is not guaranteed that a small mass will result in minimal environmental impacts. This is demonstrated by Heinonen et al. [
15], who found that mass is not a good cutoff criterion for demonstrating the correlation between mass and impacts to 18 midpoint categories for an apartment building in Finland. Further, they report the cutoffs from omitting building systems and materials can be significant, even for finishes and mechanical works. Mass was used in this study following the main LCA tradition in the building sector [
78], though it does result in uncertainty. This is demonstrated by the copper process in this study, which accounted for only 0.03% of the total mass but had a large contribution to many midpoint impact categories. Other phases were excluded from the system boundary due to a lack of data, including the impacts from construction work (A5) in the pre-use phase, the use phase (B1-B7), the end-of-life phase (C1-C4), and benefits outside the system boundary (D), as defined by EN15804 in
Figure 3. Based on the durability of the fiberglass material and the potential for reuse is high, and therefore, a design for disassembly approach could significantly reduce end-of-life impacts following Joensuu et al. [
79].
Finally, truncation error is a problem present in all process LCA studies, meaning process LCA studies always cut out higher-level tiers [
77]. Heinonen et al. [
15] and Säynäjoki et al. [
6] also refer to “first-tier truncation,” meaning that some emissions are left out of the assessment when only the materials that make up certain building components are assessed, leaving out the final processing and assembly steps in converting the materials to the actual building components. The first-tier truncation problem does not apply to all materials, but examples of such building components are windows and doors, electrical systems, or pipes. This problem is similar to the truncation error that concerns the higher-level tiers, but first-tier truncation cuts off the first tier. To address the first-tier truncation in this study, the processes for aluminum, steel, HDPE pipe and PVC pipe, and copper wire include transformation processes equal to the raw material input. For many components, the information was not available, which leads to a downwards bias, although this is not different from other building LCA case studies.
In conclusion, this study presented a wide range of both midpoint and endpoint results for a newly designed fiberglass house in Iceland within the LCA framework. As this was the first study conducted on a modular house utilizing fiberglass as the main building component, this study contributes to existing research by comprehensively reporting the environmental impacts of an alternative building solution. Based on the results, the fiberglass material, as assessed in this study, does not demonstrate environmental benefits compared to traditional non-renewable building materials. Therefore, it does not provide a suitable replacement when striving to reach environmental targets, unless improvements to the material quantity and/or composition are made to reduce environmental impacts. Based on the uncertainties and limitations discussed above, the results from this study should be interpreted with care, though the comprehensive reporting of midpoint and endpoint impacts can contribute to developing a baseline for comparison for building LCA studies. Replacing traditional building materials with low-impact alternatives has the potential to significantly reduce environmental impacts from the built environment, though future development should be guided by a thorough understanding of the lifetime environmental impacts from buildings.