https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-021-26212-z/MediaObjects/41467_2021_26212_MOESM2_ESM.pdf

Table of Contents

REVIEWER COMMENTS

Reviewer 1 (Remarks to the Author)

This contribution is focused on a crucial GHG mitigation area, GHG emissions embodied in construction materials, and seeks to close a crucial gap in the international analysis, to develop consistent global scenarios for that area. I consider the contribution overall both informative and in many aspects well done. There are the following concerns, which I advise to be taken care of before publication can be considered:

(1) One of the main conclusions of the paper is that – in order to remain within a 1.5 degree target carbon budget - either the strategies mentioned have to be significantly further increased in strictness or other sectors have to mitigate more (both in the abstract and in lines 374-378). This leaves out an actually already pursued third option. This third option is mentioned in the article, although only for the long-term: (Lines 246-248) “Third, substantial GHG emissions can be reduced in steel and cement production through various carbon capture, utilization, and storage (CCUS) technologies, such as chemical absorption, calcium looping, and cement carbonation, among others. These technologies are yet to be fully commercialized and may only play a significant role in reducing emission intensities in the longer term”. (from the context one would assume that longer-term refers to basically post-2060).

However, these technological options are already available at laboratory scale, so could be employed significantly sooner than the article seems to imply. These options thus need to be discussed more inherently and considered in the scenarios, not just to be referred to as “future options”.

(2) The strategies seem to miss out significant further options. E.g. what I am told by building specialists consistently over the last years is a further significant strategy to reduce the material amount of concrete in ceilings (identically weight proof with 40-70% less concrete, and accordingly less cement), by returning to shuttering formworks as used in the early days of concrete, when it was much more expensive than labor. The new “empty segments” won could be used by activated building components, such as for storing heat (new value added chains). My suggestion is to expand for at least mentioning both such technological strategies and their integrated embedding into comprehensive overall strategies.

This is also a question of the approach. When we know that carbon neutrality is our target, such more integrative and comprehensive solutions will be sought, beyond the (somewhat narrower) strategies already mentioned in Table 1.

(3) The argument that non-CO2 emissions are hardly relevant for GHG emissions (of the building material sector) and thus a look at the carbon budget only (CO2 only) suffices, should be discussed sector, and what trends could have it increase (increased relevance of insulation material), and whether we might really neglect it (as we do have a robust emissions budget for CO2 only (given the much shorter lifetime of most non-CO2 GHG emissions). But note, that non-CO2 emissions and their development do have a crucial implication on the temperature targets achieved (Figure SPM.1 of the IPCC SR 1.5 degrees).

(4) Related to that issue: How are building related GHG emissions quantified? Do you rely on a production-based accounting method, or do you – which should be the case here, at least as an alternative and possibly a sensitivity analysis – also take account of the indirect emissions, i.e. base your analysis on a consumption-based accounting approach, e.g. including emissions from electricity generation if electricity is used as an intermediate input in producing construction material. This refers to both, the current status (as given e.g. in the caption of Figure 3 – 8.1% of current GHG emissions), and in the BAU and high efficiency scenario, which both are demanding for a consumption base accounting approach. (I would expect building related emissions to account for a higher share of emissions under the consumption-based approach, which actually is a final demand allocation approach).

(5) Further issues:

a. To avoid any potential misunderstandings you might want to add the explicit information that all your numbers in the abstract in line 22-23 refer to “building material related GHG emissions” only. Similarly, in the results section (lines 121-124).

b. For the “tradeoff between pre-use and in-use emissions whereby highly energy-efficient buildings may require more materials in construction” (lines 47-48)) there is more recent and comprehensive work then cited so far, in particular in the recent special issue of Buildings and Cities
(https://www.buildingsandcities.org/journal-content/special-issues/carbon-metrics.html ), with the trade-off explicitly discussed under a carbon budget setting in https://doi.org/10.5334/bc.32, which also supplies a transparent argument for larger shares of the carbon budget for certain sectors, in particular building material related GHG emissions, given the high share of investment demand in demand for such goods. You might want to consider this argument also for your purposes.

c. Lines 132-134 (and Figure 1, Panel D): that is also a matter of country size /while I understand why the absolute amount is of relevance to you here, reconsider (at least for the SI) a figure with a scaling of vertical axis rather as per capita?

d. Aggregation: Lines 151ff (and Figure 2): Mention how you aggregate these, as potentials clearly overlap.

e. Line 163ff: Note that the three layers exactly match and thus you might want to relate them to the general framework of Avoid - Shift - Improve (Creutzig et al, 2018; https://doi.org/10.1038/s41558-018-0121-1 )

f. Typos: line 201 (infrastructure), line 441 (“that”)

g. Line 234-235: “The fact that several building materials are produced by difficult to decarbonize sectors, such as steel and cement production, presents a serious challenge” – see my main points (1) and (2) above, which need to be discussed in that context here as well, them offering a clear (and
earlier) way out.

Reviewer 2 (Remarks to the Author)

This is an important topic and authors have addressed it well in the paper. I would suggest its publication only after the following major revisions:

Additions

There are indeed tradeoffs between pre-use (embodied) and in-use (operational) energy use. In fact, there are tradeoffs between in-use energy components such as heating vs. lighting or cooling loads. I suggest authors to look at the following recent study that pertains to not just the Introduction but also the Discussion section: “Evaluating the impact of operating energy reduction measures on embodied energy”. I would suggest mentioning explicitly research objectives so that these can be mapped to the mentioned knowledge gaps and results. I would suggest discussing construction automation technologies, particularly additive construction in the Discussion section; such technologies could have profound impact on the future housing construction.

One other barrier could be varying levels of motivation to save energy and non-energy resources between tenants and owners. Owners tend to invest upfront to save later during the building service life, which may not be the case with tenants. Building renovation activities are not only dependent on preventive and corrective maintenance and repairs but also on other uncertain factors such as trends or fashion. These factors must also be discussed along side rezoning and reorganization of urban fabric in the Discussion section.

Clarifications

Lines 190-192. Not a majority of intense steel-concrete structures may be substituted with timber building looking at the life cycle management aspects of timber vs. concrete/steel buildings. One big pitfall of assuming housing alone is the disregard of transportation infrastructure that will accompany any increase in housing demand. Such sectors may not be able to avoid the use of concrete and steel, at least given the current state of horizontal construction.

Lines 192-193. Timber substitution may not be feasible for all regions of the world. Regions such as India have regulations that prohibit over-exploitation of forest. If you consider the increasing housing demands in India and China alone and substitute 10% of these demands with wooden buildings, you could see the grave situation of the forests. Please make a note after this sentence so that readers can understand this aspect.
Lines 187-201. What about the increasing frequency and severity of climate change induced weather and geological disasters? Does your calculation cover post-disaster reconstruction and repairs?

Figure 4. How realistic is the High Efficiency scenario of over 0.9 of the outflow-to-inflow ratios for all material in general and for material such as wood in particular?

Overall methodological issues

There are two major issues that authors need to address either by mentioning them as limitations or describing why these are not really important to consider. First, the time horizon of 40-80 years is too
long and uncertain to predict not just the housing demand but also the type of material use that is changing rapidly with increasing use of virtual technologies and automation (e.g. 3D printed or additively constructed buildings). Does the model utilized by the authors provide any uncertainty assessment to understand these projections? Second, the energy use and energy mix are projected to change profoundly owing to automation, urbanization, and digitalization. Do you consider such
changes in your model? If so, please explain. If not, please describe potential implications concerning you results.

Reviewer 3 (Remarks to the Author)

General comments

This paper estimates greenhouse gas emission associated with the construction and refurbishment of residential and commercial buildings, focusing on eight construction materials, from now until 2060.

The paper is global in coverage and considers 26 regions around the world. The authors model seven different climate change mitigation strategies to estimate embodied greenhouse gas emissions reductions associated with construction materials. They use a process-based life cycle inventory (namely ecoinvent 3.6) to calculate embodied greenhouse gas emissions from ‘cradle-to-gate’ (stages A1-A3 in EN15978). The main findings demonstrate the critical need to abate embodied greenhouse gas emissions associated with construction materials, highlighting that under the best scenario, their allocated carbon budget would not be enough. A cross-sectorial approach is needed to tackled embodied greenhouse gas emissions of construction materials.

Overall, I agree with the authors on the need for this research, and I laud its global coverage and investigation of scenarios. I do have a few suggestions to further improve the paper. These are listed below.

Firstly, the authors use ecoinvent 3.6, with some adapted versions to represent variations in energy supply chains, to calculate embodied greenhouse gas emissions. While this could be fine for a global study aiming at providing trends and comparing scenarios, the use of process data will systematically result in a truncation error in the life cycle inventory, as demonstrated by various authors in the last 20+ years (Crawford, 2008; Crawford, Bontinck, Stephan, Wiedmann, & Yu, 2018; Lenzen, 2000; Majeau-Bettez, Strømman, & Hertwich, 2011; Treloar, 1997). That means that the absolute embodied GHG figures obtained and discussed in lines 151-252, are actually even higher when using comprehensive life cycle inventory approaches such as hybrid analysis.

In the only available database of hybrid embodied environmental flow coefficients for Australian construction materials (Crawford, Stephan, & Prideaux, 2019), the truncation error¹ of ecoinvent data is more than 50% on average. I would invite the authors to at least comment on the fact that their embodied GHG estimation is truly a lower boundary in their text, as this would have even more stringent implications about ‘decarbonisation’ and a cross-sectorial approach.

Secondly, the authors use 8 different building types to model hundreds of millions of buildings around the world. Per m² of building modelled, I would tend to think that the ‘archetypal resolution’ used is rather low and this truly provides very broad-brush estimates. Recent bottom-up GIS-based studies ((Augiseau & Barles, 2017; Stephan & Athanassiadis, 2017)) have demonstrated that simply using material intensity per m² and low archetypal resolutions can lead to potentially significant errors in the quantification of the built stock, from which GHG emissions are derived. A comment on that from the authors, directly in the manuscript, would help the reader better gauge the reliability of the results.

In a similar way, the authors focus on seven main materials, but previous research shows that materials outside this list can contribute significantly towards the life cycle embodied energy or greenhouse gas emissions of a building (e.g. carpet, paint, ceramic tiles, etc.)

Thirdly, in light of the above, it would be great to have some estimation of uncertainty in the final graphs produced. Only Figure 3 is visually explicit about the fact that we have bands, rather than lines. All other figures are not representing that uncertainty, which I find potentially misleading. This is a shame given that in the supplementary information, the authors have done a good sensitivity analysis, which could be represented using shaded areas, whiskers, etc. on the final graphs. I think that being very upfront about the uncertainty in this exercise would be to the credit of the authors.

Fourthly, it would be great to have some additional information on the method. The methods section does a good job at describing the procedural approach, but it might be stronger if the authors could include some further justification about their choices and their inherent limitations. For instance, choosing ecoinvent has its advantages like the authors mention, but it also results in an underestimation of the total GHG emissions. The latter is not mentioned.

Specific improvements

1. Throughout: please don’t use ‘energy consumption’. Try ‘energy use’ instead. Energy cannot be consumed according to the first principle of thermodynamics.
   ...
2. In all figures, please change CO2 to CO»subscript 2«
3. In all figure captions, please expand GHG and other abbreviations so that the capital can stand alone. If you use abbreviations such as GDP, please define these in the caption.
4. Line 194, the authors might want to cite, in addition to 33, the following paper: Arehart, J. H., Hart, J., Pomponi, F., & D’Amico, B. (2021). Carbon sequestration and storage in the built environment. Sustainable Production and Consumption, 27, 1047-1063.
   doi:https://doi.org/10.1016/j.spc.2021.02.028
5. Lines 295-298: this need for more detailed information for urban mining has been called for by papers in bottom-up material stock modelling. I would suggest that the authors bring in the voices of Tanikawa and Hashimoto (2009) and Stephan and Athanassiadis (2018) to further reinforce this argument.
6. L312, please avoid the short form, e.g. don’t, in academic writing
7. L313, the authors might want to corroborate their findings on the potential of reducing floor area per capita as measure that improves environmental performance with those of previous studies on the matter, e.g. Wilson and Boehland (2005) and Stephan and Crawford (2016).
8. In the supplementary information, please check your reference entry 45, which seems to be incomplete.
9. I appreciate that the authors have shared the code and added comments across the steps. This being said, the code is extremely long and repetitive. This is just a comment and not something the authors need to act on. For the next time, it might be worth investing some time in using either vector calculations using pandas, for loops or generator expressions, to streamline the code, e.g. iterative over the list of materials and calculating relevant quantities for each.

References

Augiseau, V., & Barles, S. (2017). Studying construction materials flows and stock: A review. Resources, Conservation and Recycling, 123, 153-164. doi:10.1016/j.resconrec.2016.09.002
Crawford, R. H. (2008). Validation of a hybrid life-cycle inventory analysis method. Journal of Environmental Management, 88(3), 496-506.
doi:https://www.doi.org/10.1016/j.jenvman.2007.03.024
Crawford, R. H., Bontinck, P.-A., Stephan, A., Wiedmann, T., & Yu, M. (2018). Hybrid life cycle inventory methods – a review. Journal of Cleaner Production, 172, 1273-1288. doi:https://doi.org/10.1016j.jclepro.2017.10.176
Crawford, R. H., Stephan, A., & Prideaux, F. (2019). Environmental Performance in Construction (EPiC) database. Melbourne: The University of Melbourne. Lenzen, M. (2000). Errors in Conventional and Input-Output-based Life-Cycle Inventories. Journal of Industrial Ecology, 4(4), 127-148. doi:https://www.doi.org/10.1162/10881980052541981
Majeau-Bettez, G., Strømman, A. H., & Hertwich, E. G. (2011). Evaluation of process- and inputoutput-based life cycle inventory data with regard to truncation and aggregation issues. Environmental
Science & Technology, 45(23), 10170-10177. doi:https://www.doi.org/10.1021/es201308x
Stephan, A., & Athanassiadis, A. (2017). Quantifying and mapping embodied environmental requirements of urban building stocks. Building and Environment, 114, 187-202. doi:http://dx.doi.org/10.1016/j.buildenv.2016.11.043
Stephan, A., & Athanassiadis, A. (2018). Towards a more circular construction sector: Estimating and spatialising current and future non-structural material replacement flows to maintain urban building
stocks. Resources, Conservation and Recycling, 129, 248-262. doi:https://doi.org/10.1016/j.resconrec.2017.09.022
Stephan, A., & Crawford, R. H. (2016). The relationship between house size and life cycle energy demand: Implications for energy efficiency regulations for buildings. Energy, 116, Part 1, 1158-1171. doi:http://dx.doi.org/10.1016/j.energy.2016.10.038
Tanikawa, H., & Hashimoto, S. (2009). Urban stock over time: spatial material stock analysis using 4d-GIS. Building Research & Information, 37(5-6), 483-502. doi:https://www.doi.org/10.1080/09613210903169394
Treloar, G. J. (1997). Extracting embodied energy paths from input-output tables: towards an inputoutput-based hybrid energy analysis method. Economic Systems Research, 9(4), 375-391. doi:https://doi.org/10.1080/09535319700000032
Wilson, A., & Boehland, J. (2005). Small is Beautiful U.S. House Size, Resource Use, and the Environment. Journal of Industrial Ecology, 9(1-2), 277-287. doi:https://www.doi.org/10.1162/1088198054084680

Overall, I think that the authors have done some great work and with some additional nuancing, discussion and more transparency about uncertainty and underestimation of GHG emissions, the paper would be a significant contribution to our climate emergency.

Kind regards,
Prof. André Stephan