1 Introduction

Buildings and the construction sector as a whole were responsible for 36% of final energy use and 37% of energy-related carbon dioxide emissions in 2020 (United Nations 2021). Whilst the operation of buildings was the largest emitter (27%), the manufacturing of construction materials such as steel, cement and glass also accounted for approximately 10% of the emissions.

Efficient energy use has been promoted through regulations and other incentives in recent decades (EU 2010, 2018). From 2021, only nearly zero-energy new buildings can be constructed in the European Union and the proposal for the revision of the Energy Performance of Buildings Directive (EPBD) specifies that from 2030 all new buildings shall meet zero-emission standards with the vision for a zero-emission building stock and a climate neutral European Union by 2050 (European Commission 2021).

However, these measures still focus only on the operational energy use (Zöld and Szalay 2007). Higher energy efficiency results in higher material use, which also indicates the importance of the production phase (Sartori and Hestnes 2007). Recent developments have turned the spotlight on the embodied emissions of buildings. The World Green Building Council also urges increased efforts to tackle embodied carbon emissions (WorldGBC 2019). They envision that all new buildings shall be net zero operational carbon and have 40% less embodied carbon by 2030, and that new buildings and renovations will have net-zero embodied and operational carbon by 2050. According to the European Academies Science Advisory Council, regulations on the embodied greenhouse gas emissions are needed for the decarbonisation of buildings (EASAC 2021). In the Monte Verita Declaration, scientists are calling for the introduction of legally binding life cycle greenhouse gas emission benchmarks for new constructions and refurbishments by 2025 (IEA EBC Annex 72 2021).

The whole life cycle environmental impacts of buildings including embodied and operational emissions can be determined with the method of Life Cycle Assessment (LCA) (CEN 2011). Whilst voluntary certification schemes, for example the German DGNB, already prescribe conducting LCA, recently several European countries (e.g. Denmark, Finland and France) have also started introducing LCA as a compulsory analysis. The revised EPBD also includes the mandatory calculation of the life cycle Global Warming Potential for all new buildings from 2030, to be displayed as part of the energy certificate. In the first phase, no maximum impact requirements are introduced, but it is foreseen that after a transition period binding targets will be set (European Commission 2021).

However, in many countries, life cycle-based benchmarks, reference and target values are still missing. Benchmarks are useful to compare the environmental performance of a building with the current average or best practices and they help guide designers towards more sustainable buildings (Trigaux et al. 2021). Benchmarks support the use of LCA early in the design process when there is still a large potential for environmental reductions but when clients may otherwise find it difficult to set environmental performance targets and interpret the results (Hollberg et al. 2019). Benchmarks can also be applied for the ecological labelling and certification of buildings, formulating policies and for setting targets in public procurement (ISO 21678 2020).

Reference benchmark values describing the state of the art can be derived from the assessment of a large set of recent construction projects. Ideally, the sample should be representative, high quality and detailed. However, as the study of embodied carbon has been mainly confined to academic research and voluntary schemes until recently, a large pool of building data is not yet available to derive benchmark values in many countries. As long as a large database does not exist, reference buildings or building typologies could be applied for developing reference values. At the same time, reference buildings — which are usually classified according to size and construction period — have a specific, defined geometry and a specific construction system, floor plan, window ratio, etc. In this paper, an alternative approach is proposed where a large building sample is generated describing the technically feasible buildings. Instead of a few typical buildings, the main parameters describing a building are determined (e.g. floor area, window-to-wall ratio) and the ranges are defined that these parameters typically take. With the variation of these parameters, a large building sample is generated, and the surfaces and built-in material quantities are determined for typical construction solutions. These are then used to assess the environmental performance of the buildings. With the statistical analysis of the sample, reference benchmark values can be derived. Whilst each actual building has a unique design, such a method may provide reference values for use on the level of the entire building stock.

This paper first provides further background on environmental benchmarking, the different types of benchmarks and the state of art in the literature and national policies on building LCA. The “Methodology” section describes how a large building sample is generated and details the LCA method applied. The proposed method is illustrated on developing reference embodied impacts of new residential buildings in Hungary. In the “Results and discussion” section, the derived reference values are presented in detail: the total embodied impact, the contribution of life cycle stages and different construction elements. The main focus is on the Global Warming Potential (GWP) values, but the results for other relevant impact assessment categories are also shown.

2 Background on benchmarking

2.1 Types of benchmarks

According to the standard on sustainability indicators and benchmarks, benchmarking is the “process of collecting, analysing and relating performance data of comparable buildings or other types of construction works” and benchmark is a “reference point against which comparisons can be made” or a “performance level for comparison” (ISO 21678 2020). The standard distinguishes three types of benchmarks:

  • Limit value is the lowest value of acceptable performance, for example the legal minimum.

  • Reference value is the state of the art or business as usual.

  • Target value is an objective going beyond the reference value, the theoretically achievable highest performance. The target value is the result of a target-setting process for the short term, medium term, or long term.

  • In addition, best practice values represent the best available real performance, for example in experimental or demonstration projects.

There are different methods for establishing benchmark values. They can be derived by top-down or bottom-up methods, they can be absolute or relative values and they can differ along the system boundaries considered.

2.1.1 Top-down or bottom-up

Benchmarks can be derived with a top-down or bottom-up process. Top-down benchmarks are usually based on a global political or scientific target, which is then translated into specific targets on the level of nations and sectors. For example keeping the global average temperature increase well below 2 °C and pursuing efforts to limit it to 1.5 °C above pre-industrial levels in the Paris Agreement or the planetary boundaries (Rockström et al. 2009). On the other hand, in a bottom-up process, the benchmarks are derived from technical or economic optimums, usually based on a set of reference buildings or a large pool of buildings (Hollberg et al. 2019). In this case, benchmark values can be linked to statistical indicators; for example limit value can be the third quartile, the reference value the median and the best practice value the first quartile of the data (VTT 2012).

2.1.2 Absolute or relative benchmarks

Regarding the type of benchmark values, absolute or relative values can be distinguished, also called external and internal benchmarks (Ganassali et al. 2016). Absolute benchmarks are fixed values (e.g. absolute primary energy or CO2 values per floor area). Relative benchmarks compare the performance of the building to a baseline. This is a common approach for energy requirements where the building is compared with a reference or notional building, which is a hypothetical building of the same size, shape and orientation with the same activities, zoning and weather but with pre-defined specific properties for building thermal envelope and services. The requirement may be a certain level of reduction, e.g. 10% compared to the baseline. Relative benchmarks depend on the geometry and other parameters of the building, so they are not fixed values.

2.1.3 Whole life cycle or life cycle stages

Benchmarks usually relate to the performance of the whole building for the whole life cycle, which enables holistic optimisation: high values in one domain can be compensated with lower values in another domain. In some schemes individual target values are also defined for the individual life cycle phases; for example SIA (2017) also provides target values for the embodied, operational and building-related mobility impacts (SIA 2017).

2.2 National regulations and schemes on embodied impacts

Several European countries have already introduced or are in the process of introducing whole life cycle considerations in their building regulations and policies, as described in the following short overview. Three countries, the Netherlands, Denmark and France, have already introduced CO2 limits, whilst Sweden and Finland have plans to do so.

The Netherlands was the first European country that introduced environmental calculation as a prerequisite for a building permit in 2013 but originally without any benchmarks. The calculation is necessary for new office buildings and new houses larger than 100 m2 (Frischknecht et al. 2019a). In the applied national MilieuPrestatie Gebouwen (MPG) method, the shadow price of the emissions is used for weighting and combining all impacts into a single indicator (Stichting Bouwkwaliteit 2019). The maximum MPG value allowance of 1 EUR/m2 has been reduced to 0.8 EUR/m2 and will be gradually lowered in the future (BPIE (Buildings Performance Institute Europe) 2021).

In France, in the voluntary E + C − (Énergie Positive & Réduction Carbone) experiment environmental data has been collected for new buildings to prepare the new building regulation (Ministère de la transition écologique 2021). The new regulations were applied to new residential and non-residential projects applying for a building permit. For the whole life cycle, including operational energy use, the limit values are between 12.8 and 14.8 kg CO2-eq/m2/yr (One Click LCA 2021).

In Denmark, the sustainable construction regulation set whole life cycle carbon limits for all new residential and non-residential buildings with a floor area of over 1000 m2 from 2023 (Indenrigs og Boligministeriet 2021). The life cycle threshold values including operation are 12 kg CO2-eq/m2/yr and will get gradually stricter every 2 years. An analysis is also compulsory for buildings with a smaller floor area but there are no limit values yet. The analysis must be performed with the LCAByg tool (Kanafani et al. 2019).

Finland is planning to introduce mandatory carbon footprinting and carbon limits for buildings by 2025 (One Click LCA 2021). In the beginning, this will be applied only to new buildings as part of the energy certification. The preliminary values are 10–14 kg CO2-eq/m2/yr for the whole life cycle over 50 years. In addition, criteria for green public procurement based on Global Warming Potential and climate benefits have been developed (Kuittinen and Häkkinen 2020).

In Sweden, the regulations on climate declaration for new buildings came into force in 2022 (Boverket 2020). The building’s impact is calculated not for the entire life cycle, but only for the construction stage and for certain building components (energy use is assessed separately in energy certification). A national generic database is available, which can be replaced by product-specific data from environmental product declarations. Limit values are expected only after 2027.

In Germany, the German Sustainable Building Council (DGNB) has developed benchmarks for various functions. Benchmarks also exist in the Bewertungssystem Nachhaltiges Bauen (BNB) sustainability assessment system for public buildings, new and renovated office and educational and laboratory buildings (Bundesministerium für Wohnen, Stadtentwicklung und Bauwesen). The benchmarks have been derived with a bottom-up approach based on the statistical analysis of representative building types and building examples, combined with legal requirements for energy performance (König and De Cristofaro 2012). For the construction part, there are benchmarks developed by DGNB, whilst for the operation the reference values are calculated individually for each building according to EnEV. Based on the reference value, limit and target values are calculated. The application of the system is voluntary, except for new and renovated federal government buildings where is it mandatory (Frischknecht et al. 2019a). Ökobau.dat is used as the database (BMI 2023).

2.3 Bottom-up benchmarks in the literature

Whilst top-down methods are helpful for setting target values, bottom-up approaches are more common to derive reference values, which this paper aims to arrive at. Therefore, that part of the literature is analysed in more detail here.

In the literature, the size of the building sample varies widely (Trigaux et al. 2021). Researchers typically analysed between 20 and 40 buildings (Rasmussen et al. 2019; Lasvaux et al. 2017; Lavagna et al. 2018); however, Rasmussen et al. (2019) drew conclusions based on the analysis of only 7 residential buildings in Denmark. Martínez-Rocamora et al. (2021) combined BIM-based LCA tools and machine learning techniques to generate environmental benchmarks. The example of one 11-story Spanish residential building served as a basis to study 240 variations of constructive solutions. Wiik et al. (2020) analysed a sample of 133 Norwegian case studies of different building types to determine life cycle GHG emission benchmark values but concluded that the database may be a too small sample to draw robust conclusions on a national level. Peuportier and Wurtz (2021) combined three typical buildings with parametric variation in materials.

Recently, there have been some more extensive studies. In the USA, the Embodied Carbon Benchmark database contains LCA results of more than 1000 buildings, but only for primary building components (Simonen et al. 2017). In Europe, One Click LCA gathered embodied carbon benchmark data for European buildings to support carbon regulations and policies across Europe (One Click LCA 2021). The calculations have been carried out using a consistent methodology in the framework of the Carbon Heroes Benchmark Program on the One Click LCA platform. The database of 3737 buildings was categorised for three regions (Northern, Western and Eastern Europe) and five main building types (commercial, educational, industrial, office and residential multi-family). The sample is the largest for Northern Europe and residential and office buildings, due to the higher demand for low-carbon buildings and broader use of whole building assessments in sustainability ratings in that region. Another larger sample is the EU-ECB database, which contains LCA data from five European countries (Belgium, Denmark, Finland, France and Netherlands) (Röck et al. 2022). A total of 769 cases of different building use and building structure were statistically analysed with the largest number of cases from France in order to understand the baseline for embodied carbon in buildings. The geographic distribution of the samples is very uneven, with Northern and Western European countries taking the leading role in building LCA.

In the literature, the quality and representativeness of the sample used to establish benchmarks also vary widely. In France, a sample of 40 low-energy individual houses was selected according to the market shares for the load-bearing construction (reinforced concrete, concrete block, wooden houses, brick and steel frame) and the climatic region (Lasvaux et al. 2017). In Norway, all cases were exemplary projects built to very high energy standards, such as zero-emission building (Wiik et al. 2020). The One Click LCA database contains mostly projects pursuing a green certification scheme (One Click LCA 2021). As can be seen, the selection of buildings is often biased, as LCA is predominantly applied to high-performance buildings or innovative research concepts, which may deliver a skewed picture.

Most of the case studies are planned or already realised new projects or, to a lesser extent, renovations. In some cases, real or virtual reference buildings are used that are representative of the building stock (Ballarini et al. 2014). Lavagna et al. (2018) calculated benchmarks for the EU housing stock to quantify the average life cycle environmental impact of housing and provide reference values for policy development. They considered 24 models altogether, clustered according to the building type (multi-family house and single-family house), year of construction (4 periods) and three climate zones. The study used only one basic geometry for single-family houses and one for multi-family houses, but complemented them with the characteristics of each building type (floor area, window-to-wall ratio, typical construction technologies, etc.). With this method, average values could be determined, but not the variation in the environmental impact because the archetype buildings were based on average geometric data. Moschetti et al. (2015) analysed four Italian residential buildings representative of recent constructions to determine environmental impacts, energy and life cycle costs. The building categories were derived from the TABULA project: single-family house, terraced house, multi-family building and apartment block. The analysis considered one typical massive envelope construction adapted to three climatic zones by changing the insulation thickness, three energy systems and two methods for thermal bridge modelling. Also, TABULA archetypes were modelled in Belgium to determine the embodied impacts of buildings (Röck et al. 2022).

When setting benchmarks, it is important that the calculation rules are clear and the methodology is transparent (Rasmussen et al. 2019; Schlegl et al. 2019). Differences in methodology make it difficult to derive benchmarks. Wiik et al. (2020) mentioned that it was difficult to establish benchmarks because the case studies were calculated with different tools that were not entirely harmonised. Simonen et al. (2017) could not establish benchmarks due to the methodological differences. Different assessment methodologies, databases and scopes were challenging the data processing in an international study (Röck et al. 2022). The results are very sensitive to the system boundaries (Lasvaux et al. 2017), e.g. the calculation period, the life cycle stages considered (e.g. treatment of stages C and D) or the completeness of the physical model of the building. For example Gervasio et al. (2018) calculated only embodied impacts for the structural system of residential buildings of the IMPRO-Building project and defined preliminary benchmarks based on this.

As seen from the literature, large datasets are still missing in many regions and further extensive samples are needed to understand the environmental impact of buildings and draw generic conclusions. In the absence of data from real projects, building typologies may be applied. However, these may be representative of the existing building stock and historic construction trends, but may not adequately represent future buildings, which will be built with other geometric and construction features. There is often a lack of transparency and consistency in the development of benchmarks. These are very strongly dependent on the applied methodology: large differences are caused by different databases and different system boundaries, for example service lifespans, the life cycle stages considered or the completeness of the physical model. Benchmark values should not be taken from another country or system without carefully checking their validity, which underlines the need for transparent methodologies.

3 Methodology

This paper presents a methodology for developing reference benchmark values for the environmental impact of buildings. In this method, a large population of buildings is automatically generated based on some basic building parameters. The building sample covers the population of technically feasible buildings, with rational shapes and spans. This approach has been used earlier to set the requirements for nearly zero-energy buildings (Szalay and Zöld 2014) and now it is extended to calculate the embodied impacts of the building.

In the description of the methodology, the general considerations are illustrated with a specific application to the national case of Hungary. It is preferable to set whole life cycle benchmark values, but this paper only focuses on the embodied impacts, as data sources are scarcer in the latter field. However, the methodology is suitable for combining the embodied and operational energy impacts and defining full life cycle benchmark values as well.

3.1 Building types and geometry parameters

First, building categories need to be defined. For Hungary, four building types were selected: (1) one- and (2) two-storey single-family houses (SF), (3) low-rise and (4) medium high-rise multi-family apartment buildings (MF). Semi-detached, terraced dwellings and high-rise residential buildings are not common in Hungary, so these were not included in the analysis. This paper focuses on residential buildings.

As a next step, geometric parameters and their typical ranges need to be defined for each building category. The main geometrical parameters selected are the floor area, the number of storeys, ceiling height, the compactness of the building, the window-to-wall ratio, the ratio of internal walls in the layout and the roof slope. Based on the geometric parameters, the area of the building elements and the volume of the building can be calculated and also the built-in mass and the embodied impacts can be derived.

Finally, a large building sample is generated with a random variation of the parameters in the given ranges with a Monte Carlo simulation.

3.1.1 Building parameters

The parameters describing the building geometry and the realistic ranges of these parameters were determined based on functional and architectural considerations and statistics. It is expected that real new buildings will fall into these ranges in the future. It must be emphasised that the goal was to determine the technically and functionally feasible ranges and not a statistically representative distribution. Typical or average existing houses are only characteristic of the building trend of a certain period. The composition of the building stock is changing with time. For example in Hungary in the 1960s and 1970s approximately 800,000 one-storey detached houses were built with a standardised, square-shape floor plan, small windows and a hip roof (Csoknyai 2013). These represent approximately 20% of the existing dwellings, which is statistically significant. However, these buildings were built to a minimum standard regarding size, functionality and daylighting, which later constructions have surpassed.

The considered building parameters are briefly described in the following. The parameters and the assumed ranges are summarised in Table 1.

Table 1 Classification of residential buildings and typical ranges of the geometric parameters
Full size table
3.1.1.1 Number of storeys

Detached houses were divided into one- and two-storey houses, multi-family houses into low (3–4 storeys) and medium high-rise buildings (5–10 storeys above ground). This division is according to the Hungarian building regulations (medium high-rise buildings are defined to have a topmost floor level between 13.65 and 30 m) and also refers to the higher structural requirements in taller buildings. Residential buildings above 10 storeys are still rare in Hungary.

3.1.1.2 Floor area

For the area, the area of one floor (AF) was used and multiplied with the number of storeys to calculate the total floor area (AN). Different regulations in different countries use different concepts (e.g. gross floor area, gross heated floor area), which may influence the final results, so it is vital to be transparent about the geometrical concept selected. In Hungary, the energy regulations apply the net heated floor area, so this was used in this research too, for consistency. Very small buildings with a floor area under 60 m2 are not typical and are exempt from the energy regulations. The upper limit is harder to set; however, detached houses with a total floor area above 240 m2 are rare. The size of detached houses has increased in the last decades: whilst in 2004 the average total floor area was 123.8 m2, this increased to 150.7 m2 in 2021 (KSH (Hungarian Central Statistical Office) 2022). A floor area of 60–180 m2 was assumed for one-storey buildings, 60–120 m2 for two-storey buildings, 120–400 m2 for low and 150–1400 m2 for medium high-rise buildings. In an apartment building, the area of different floors may differ, with the top floor usually being smaller, but the values refer to the average floor.

3.1.1.3 Basement and roof

In the reference case, detached houses are built with a slab-on-ground without a basement, whilst multi-family houses are made with unheated basement below ground, as this is the most prevailing solution in Hungary (Csoknyai 2013). For low-rise multi-family buildings, one floor below ground was assumed, whilst for medium–high buildings it was two storeys. The number of floors below ground is influenced by the number of necessary parking spaces (the regulations require at least one parking space per flat) and by soil conditions.

The topmost floor was assumed to be a flat roof in apartment buildings, an unheated attic in one-storey detached houses and a heated attic in two-storey detached houses, according to the dominant shares of roofs (Csoknyai 2013). For pitched roofs, gable end and an average slope of 40° were assumed, typical for Central European climate. This was considered as a fixed parameter, but it would also be possible to define typical ranges for this. In a heated attic, the roof was assumed to start from a 1-m-high parapet wall.

In reality, buildings exist with a partial cellar and also buildings where only a part of the attic was converted to a heated living space. This option was not considered here.

3.1.1.4 Ceiling height

The ceiling height refers to the useful height of the rooms. The minimum ceiling height allowed in the regulations is 2.5 m and the maximum seldom exceeds 3 m in modern buildings. In the below-ground storeys, lower height is typical.

3.1.1.5 Building shape

The shape of the building floor plan can be described by the perimeter-to-floor area ratio, which is influenced by the absolute dimensions and the compactness of the plan. Excluding atypical circular buildings, a square floor plan was considered to be the most compact. To determine the highest possible perimeter-to-floor area ratio describing irregularly shaped or very narrow buildings, the concept of the “equivalent rectangle” was introduced (Szalay 2008). The equivalent rectangle is a rectangle having the same perimeter and area as the actual floor shape. The depth of the equivalent rectangle mirrors the average building depth on the one hand, and the complexity of the plan on the other hand. The question is the minimum depth of the equivalent rectangle. In case of small dwellings, a minimum depth of 5 m was assumed as very low spans are not economical and functional. In apartment buildings, the minimum economical depth was assumed to be 8 m in low-rise and 9 m in medium high-rise apartment buildings. The maximum building depth was the square root of the floor area, but in apartment buildings it was limited to 20 m to allow for sufficient daylight penetration. This would be feasible with a middle corridor (2 m), 2 × 6-m daylit rooms and 2 × 3-m subordinate rooms (e.g. bathroom) without daylight.

3.1.1.6 Window-to-wall ratio

The ratio of windows can be expressed in relation to the floor or the façade. Here, the window-to-wall ratio was used. The minimum ratio was assumed to be 10% to comply with daylighting requirements, whilst the maximum was assumed to be 30% in detached houses and 35–40% in apartment buildings. The frame factor was assumed to be 20–30%, typical for moderns PVC or wooden frames of average sizes. The orientation of windows is not relevant for the purposes of this study because only the embodied impacts are evaluated and not the operational energy use.

3.1.1.7 Ratio of internal walls in the floor plan

The assumptions for the length of internal walls per floor area were based on functional considerations derived from the average room sizes and position of the rooms (Szalay 2008). These values include non-load-bearing and load-bearing walls. In apartment buildings, a party wall between flats was added, assuming an average flat size of 60 m2 and an average party wall length of 9 m per unit. The surface of internal doors was assumed to be approximately 15% of the internal wall surface.

3.1.1.8 Difference between net and gross area

For the calculation of the transmission losses, the net area of the building thermal envelope should be considered according to the Hungarian energy regulation, but for the calculation of the mass and embodied impacts, the gross area including internal wall junctions, wall corners, etc. was taken into account. For the external wall, the area difference was estimated to be 10%. For the floor, the thickness of internal walls was added and for roofs, the width of the ring beam and a 0.5-m roof overhang.

3.1.2 Validation of building parameters

The parameters were determined based on architectural and functional considerations, but they have been validated with statistics and other sources to check whether real buildings are in line with the determined ranges.

First, the geometry data from the National Building Typology database was analysed (Csoknyai 2013). This database is based on an on-site survey of 2000 residential buildings, which were clustered into 23 building types. The main purpose was to compile data on the energy performance of a representative building sample with more details than in usual energy performance certificates. The survey data included the geometric characteristics required for building energy certification, i.e. the area of the building thermal envelope and the basic geometric parameters, such as the heated useful floor area and heated volume of the building. However, they did not include other characteristics that have no effect on the energy demand of a building, such as the length of internal walls. The building types analysed for this paper were the four most recent categories (small and large detached houses and small and large apartment buildings, construction period 2006–2015).

In addition to the National Building Typology database, some real buildings have been selected for detailed analysis. Twenty detached houses were studied from the National House Catalogue of Hungary (NMTK) (Lechner Nonprofit Kft.), and in addition, 15 apartment buildings were selected as case studies. The catalogue was created in 2021 on the basis of an architectural design competition to support citizens in the realisation of their single-family house and improve the quality of the built environment. The database contains new constructions and refurbishment plans with different floor areas and number of rooms, which are ideal for different types of families or suitable for different building types on different plots. In the case of apartment buildings, the detailed design plans of 6 small and 9 large buildings built recently were analysed. For the main geometrical features, the average, median and the 10th and 90th percentiles were determined (Table 2).

Table 2 Statistical indices of the floor area, ceiling height and window-to-wall ratio in the National Building Typology database, the National House Catalogue of Hungary and selected apartment buildings
Full size table

The statistical indices show that the floor areas and the window-to-wall ratio match those in Table 1. The ceiling heights in the National Building Typology database show lower minimum values but these may be somewhat biased because buildings with heated attics are also included in the sample where the average ceiling height is lower due to the tilted surfaces.

3.1.3 Calculation of surfaces

The main geometric indicators of the buildings, such as the volume, floor area, envelope surface and beams, were calculated based on the parameters described above and given in Table 1. Most of the parameters were defined with ranges of minimum and maximum values and a uniform distribution. In some cases, another type of distribution would be more adequate, but in the absence of sufficient data, the distribution type could not be defined. For the random generation, the built-in function of MS Excel was used (Data table/What-if Analysis).

3.2 Building construction solutions

As the goal was to set the baseline, the most typical modern construction systems were selected, and no innovative solutions were included. In Hungary, heavyweight constructions are prevailing, and the share of lightweight buildings is still very low. The exact market shares of construction solutions are not available, but building construction handbooks, the National Building Typology database and the National House Catalogue of Hungary were consulted, and also, expert opinions were taken into account. As new buildings are the focus, all envelope constructions were designed to comply with the current regulations on the thermal transmittance (U-values), which are necessary for nearly zero-energy buildings.

The following main building elements are considered in the evaluation:

  • External wall;

  • Ground floor in contact with a heated or unheated space;

  • Pitched roof in contact with a heated or unheated space;

  • Floor above basement;

  • Attic floor;

  • Flat roof;

  • Internal non-load-bearing and load-bearing wall;

  • Cellar/basement wall;

  • Internal floor;

  • Window and door and window blind;

  • Building structures: column, shear wall, ring beam, lintel beam, foundation, perimeter beam and balcony; and

  • Technical systems.

The most typical construction of external walls is insulating bricks with additional external insulation. In detached houses, these also act as a load-bearing wall, but in multi-family buildings they are infill walls between load-bearing columns. The most widespread insulation material is expanded polystyrene. For floor slab, a precast beam and block structure was assumed for detached houses and a reinforced concrete slab in multi-family houses. For floor covering, it was assumed that 75% of the floor area is covered with wooden parquet and 25% with ceramic tiles. In detached houses, a pitched roof with a wooden roof structure is typical, covered with ceramic roof tiles. In heated attics, insulation is placed between and below the rafters.

3.2.1 Structural elements

The two main types of load-bearing systems are load-bearing walls and frame type structures consisting of columns, beams and shear walls. Detached houses are usually built with load-bearing walls in Hungary. In this case, if the depth of the house exceeded the usual span of 6 m, transversal internal load-bearing walls were assumed at every 6 m. For stiffening, reinforced concrete columns were assumed in the four corners. In detached houses with a heated attic, short columns were assumed every 2 m along the longer side of the building for taking lateral loads.

Apartment buildings are typically built with frames. An average column spacing of 6 m was assumed both on the external façade and internally, which can be regarded as typical based on the analysis of the apartment buildings. This means that if the depth of the equivalent rectangle exceeded 8 m, one row of internal columns was assumed, and two rows above 16 m. The stiffness of frame type multi-storey buildings is ensured via shear walls usually arranged around a circulation core.

A ring beam was assumed along the perimeter of the building and lintel beams above the windows, with the length calculated from the window area and 1.5-m average window height. The surface area of external columns and beams was subtracted from the external envelope area to avoid double counting of surfaces. Instead of calculating the area of the stairs, we calculated the mass as if the floor was continuous. In multi-family buildings, a balcony of 3 m2 was assumed for each flat.

For the load-bearing columns, reinforced concrete internal columns of 30 × 30 were assumed in low-rise and 30 × 80 cm in medium high-rise multi-family buildings and shear walls with an average thickness of 20 cm. For lintel beams, 25 × 30-cm reinforced concrete beams were assumed, and for ring beams 25 × 20 cm.

In detached houses, a strip foundation of 50 × 80 cm and a perimeter beam of 30 × 30 cm were calculated under the perimeter and under the load-bearing walls where applicable. In multi-family buildings, slab foundation with 40-cm thickness takes the loads of the buildings. Depending on the soil quality and other factors, in some buildings deep foundation may be necessary, but these special cases were excluded from the analysis.

Pitched roofs in detached houses have a wooden structure. For the wooden structure, a timber usage of 0.03 m3/projected m2 was assumed based on the average values of the National House Catalogue, in addition to the rafters, which were calculated with an average spacing of 1 m.

The exact composition of the assumed building elements is provided in Annex 2.

3.3 LCA methodology

In benchmark setting, the database, the applied methodology and the benchmarks are inseparable from each other and should be interpreted together. However, the approach for generating the building sample recommended in this paper can also be applied using another database and another LCA methodology. In this work, the general LCA framework is applied (ISO 14040 2006; ISO 14044 2006) and the European standards for building LCA (CEN 2011), with the specific considerations detailed in the following sections.

3.3.1 Goal of the assessment

The goal of the assessment is to determine the environmental impact of a large sample of residential buildings in order to set reference benchmark values, which characterise the current business-as-usual state.

3.3.2 Functional equivalent and reference study period

The functional equivalent is a new residential building in Hungary. The main building categories are 1- and 2-storey detached houses, and small and large multi-family houses. The reference study period (RSP) is 50 years, which is the design working life recommended in EN 1990 for residential and office buildings, and also the most widespread value in the national methods (Frischknecht et al. 2019b). The calculation is carried out for the RSP, but the lifetime of the building may be longer than this. This may lead to an overestimation of the embodied emissions per year. The results will be provided on the level of the building with values normalised to the net heated floor area and per year, as this is the reference unit that is used in energy certifications in Hungary.

3.3.3 System boundary

The system boundary in this study is formed by the embodied impacts. Whilst operational impacts are relevant for buildings, much larger datasets are already available for the operational impacts. The goal of this study is to set reference benchmark values for the embodied impacts where large datasets are still missing. The life cycle stages considered in the study are highlighted in Fig. 1. The product and construction process stages are included (A1–A5), but the pre-construction stage A0 (e.g. preliminary and design study) and maintenance (B2) are excluded due to the limited data availability. Use (B1) is connected to the building use; hence, it is out of the scope of this paper. Repair (B3) and replacement of building components (B4) are considered. In the end-of-life stage, all modules are taken into account (C1–C4). Module D is not considered in the study.

Fig. 1
figure 1

Building life cycle stages (included stages with bordering; figure adapted from CEN 2021)

Full size image

3.3.4 Building model

The geometry of the building model is developed according to the geometry parameters as described earlier. The entire building is considered with all the envelope elements and internal structures. For technical systems, the heat generation (air–water heat pump) and heat emitters (underfloor heating) are included in the assessment, but the heat distribution system as well as the water-sewage and electrical systems are excluded due to the absence of reliable data. Fittings and furnishing are excluded from the analysis as they are not directly linked to building design. External works (e.g. pavement, fencing, external parking areas) are not included.

3.3.5 Database

Environmental datasets can be classified into manufacturer-specific data, market average datasets and generic datasets. Whilst Environmental Product Declarations (EPD) for specific products are increasingly available for different products, their availability shows large spatial variations. For example, in Hungary, there is still a low number of building products with an EPD. This was one of the reasons for choosing a generic database for the assessment. The other reason is that generic datasets are more suitable for representing the reference scenario in a country. For the assessment, the ecoinvent 3.8 cut-off database was applied (Ecoinvent Centre 2023). Ecoinvent publishes consistent and transparent life cycle inventory data for a variety of sectors, including the construction sector. Currently, it includes a total of more than 18,000 datasets altogether (Wernet et al. 2016).

3.3.6 Scenarios and assumptions

Based on the building geometry and composition of the building elements, the built-in weight of the materials could be calculated. The product stage (A1–A3) was calculated from the assigned ecoinvent datasets and the quantities. For the heat generator, the impacts related to the use phase were removed in order to avoid double counting.

To account for transport from the factory to the construction site (A4), the building materials were classified into five transport categories depending on whether the product is imported or not and how many production factories exist in the country and standard distances were applied (Table 3). Building materials primarily produced in Hungary belong to the categories 1 and 2. For products in the first category, there are numerous production sites scattered around the country, whilst the products in the second category are produced in a few factories. Products belonging to categories 3 and 4 are mainly imported from neighbouring countries.

Table 3 Assumed transport distances and transport systems
Full size table

Stage A5 includes the cutting waste estimated based on Kellenberger and Althaus (2009), which was 10% for wooden products, 5% for mortars and membranes, 4% for steel and 2–3% for concrete, bricks, insulation materials, finishes and claddings. The construction of the building itself also requires energy and results in emissions, noise and solid waste. For this phase, very little information is available. Many authors neglect it completely, saying that its influence on the total life cycle is negligible. Ecoinvent estimates that the energy demand for excavation is approximately 5.5 MJ per m3 excavation volume or 5 MJ diesel per m3 building volume above ground and 10 MJ diesel per building m3 above ground for demolition (Kellenberger et al. 2007). For an average ceiling height of 3 m, this would be 15 MJ/m2 and 30 MJ/m2, respectively. The electricity need for construction, refurbishment and demolition is about 0.3 kWh/m3 (0.9 kWh/m2). According to the actual data, the construction of a detached house in Italy required 2556 MJ electricity and 1040 m3 of soil excavation (Blengini and Di Carlo 2010). This would mean about 3.7 kWh/m2 electricity and 30 MJ/m2 diesel when normalised for the heated floor area of 192 m2. According to another source, 1–5 kWh/gross floor area of electricity may be considered for residential houses and 5–20 kWh/gross floor area for multi-storey residential buildings (Gervasio et al. 2018). Lasvaux et al. (2017) assume 1.3 l/m2 of diesel consumption on the construction site, which is about 50 MJ/m2. In this study, 30 MJ/m2 of diesel and 5 kWh/m2 of electricity were assumed for construction, 1 kWh/m2 electricity for replacement and 30 MJ/m2 of diesel for demolition.

The periodic replacement of building elements (B4) was considered based on their expected lifespan. This involves a high level of uncertainty as besides technical deterioration, functional and aesthetic obsolescence are also major reasons for changing a building element. Various literature sources were consulted to establish the life times for this study (British Standards Institution 2015; Kompetenzzentrum “Kostengünstig qualitatsbewusst Bauen” 2006; Goulouti et al. 2021). It was assumed that the load-bearing structure of the building (wall, floor and roof structure) remains unchanged during the building life. The life times were adjusted based on the position of materials in the building element (e.g. it was assumed that acoustic insulation below cement screed has the same service life as cement screed). Another basic assumption was that the embodied impacts of the production of building elements stay constant over the building life and the production technology does not change significantly. For small repairs, 10% of the total replacement costs was added (B3).

In the end-of-life stage, a transport distance of 20 km was assumed, and waste processing or disposal datasets were assigned to each material based on the ecoinvent database. The specific assumptions on cutting waste, transport scenario, life time and end of life are provided in Annex 2 for the construction elements.

3.3.7 Environmental indicators

The environmental impacts are assessed with the core indicators listed in prEN 15978 (CEN 2021). The EF v3.0 impact assessment method was applied, as provided by the openLCA 2.0.4 software (GreenDelta 2023).

4 Results and discussions

For the reference system, 500 geometries per building category were generated and the sample was analysed with statistical methods. The reference system is the most widespread construction system, i.e. insulating bricks with external insulation (see Annex 1).

Whilst all environmental problems are of great importance, carbon footprint expressed as GWP is analysed in most detail. The greatest attention is paid to this indicator due to the severe climate change impacts and it is also easier to communicate the results to a non-specialised audience (Martínez-Rocamora et al. 2021). Also, the building sector is proven to contribute significantly to GWP and the depletion of abiotic resources (ADP).

First, the results of the total embodied GWP are presented. Second, the distribution of the impacts in the different life cycle stages is studied. Third, the GWP impacts are decomposed into main construction groups to analyse the contribution of substructure, superstructure, external fabric, internal elements and services to the impacts. Also, the share of the elements part of the building’s thermal envelope and other elements are determined. Fourth, the reference system is compared with other conventional construction systems that are typically applied, although with a lower share. Finally, the trends in other environmental impact categories are compared with GWP.

4.1 Total embodied GWP impact

The embodied GWP, including stages A1–A5, B3–B4 and C1–C4, of the reference system for 1 m2 of net heated floor area and expressed per year is shown in Figs. 2 and 3.

Fig. 2
figure 2

Box plot diagram of the embodied Global Warming Potential of the building sample (reference construction system)

Full size image
Fig. 3
figure 3

Embodied Global Warming Potential of the building sample as a function of the surface-to-volume ratio (reference construction system)

Full size image

The specific embodied GWP is the highest for single-storey detached houses and the lowest for high apartment houses. For detached houses, the embodied values range between 9.5 and 15.5 kg CO2-eq/m2/yr. Two-storey detached houses have lower impact values per floor area than single-storey buildings. This can be explained partly by the more compact shape of the buildings, and also because the heated pitched roof is typically a lightweight assembly where the rafters are filled with insulation. This way the amount of built-in materials does not increase significantly but the floor area nearly doubles because of the heated attic. For apartment buildings, the values are 9.1–14.3 kg CO2-eq/m2/yr. Apartment buildings have a higher material usage than detached houses because of the basement storeys and the higher structural need. On the other hand, these buildings are very compact and have a lower surface area of external walls per floor area.

In our analysis, if the yearly values are multiplied by the reference study period of 50 years, the values range from 475 to 775 kg CO2-eq/m2 for detached houses and 455 to 715 kg CO2-eq/m2 for apartment buildings. Fifty years is shorter than the actual life span of most buildings, which leads to an overestimation of the embodied impacts per year.

In France, the average life cycle greenhouse gas emissions were 590 kg CO2-eq/m2 for detached houses, of which building products and technical equipment were responsible for 420 kg CO2-eq/m2 and operational energy for 155 kg CO2-eq/m2 (Lasvaux et al. 2017). In Italy, the results from 28 residential buildings showed 380-kg CO2-eq/m2 embodied emissions (Rasmussen et al. 2019). In Norway, the embodied emission results showed an interquartile range of 240–492 kg CO2-eq/m2 with a mean of 324 kg CO2-eq/m2 for all building typologies in the as-built phase (Wiik et al. 2020). Our results are somewhat higher than these results. The reference system is a heavy-weight construction system, which is typical for today’s construction practices in Hungary and explains the higher values.

However, the two extensive international studies also found higher values. For residential buildings, Röck et al. (2022) found that the embodied carbon values range from around 400 to 800 kg CO2-eq/m2 with a mean value of 600 kg CO2-eq/m2. One Click LCA found 580–700 kg CO2-eq/m2 values for Eastern Europe (One Click LCA 2021) and lower values for Western European buildings (510–600 kg CO2-eq/m2) and Northern European buildings (310–350 kg CO2-eq/m2). According to their findings, the difference can be explained not only by the construction practices but also by the sample distribution as Eastern European buildings in the sample were large, complex projects. Our values are in line with these values.

Figure 3 shows the GWP as a function of the surface-to-volume (A/V) ratio of the building. The surface-to-volume ratio is often applied in energy performance regulations because the energy demand correlates with the ratio of the external building thermal envelope to the heated building volume. For example, in the Hungarian energy regulations, the requirements related to the maximum heat loss coefficient and total primary energy demand depend on the surface-to-volume ratio. The embodied GWP also correlates with A/V but the correlation is weaker, especially for detached houses. The R2 values for the fitted lines are 0.12–0.22 for detached houses and 0.52–0.66 for multi-storey buildings, which implies a very weak correlation for the detached houses.

4.2 Life cycle stages

The division of the embodied GWP (Figs. 4 and 5) according to life cycle stages shows the dominancy of the product stage (A1–A3). This accounts for about 63–69% of the total embodied impacts in one-storey buildings, followed by the replacements (B3–B4) in the use stage (15–23%). The share of the construction process and the end-of-life stage is low, about 6–9% each. The dispersion of these stages is also low as shown by the box plots on the figure, as many of the assumptions are linked to the floor area. The contributions are similar for every building category, with only the two-storey buildings showing a somewhat higher contribution for replacements (see Annex 3).

Fig. 4
figure 4

Embodied Global Warming Potential of the one-storey detached building sample according to life cycle stages (reference construction system)

Full size image
Fig. 5
figure 5

Contribution of life cycle stages in the average embodied Global Warming Potential of the building sample (reference construction system)

Full size image

4.3 Construction groups

In LCA, building elements and materials are often organised and classified with systematic decomposition. Different standards and countries apply different classification systems (Soust-Verdaguer et al. 2023). The embodied impacts were decomposed into the following five main categories (Fig. 6):

  • Substructure: foundations, basement walls and slab-on-ground;

  • Superstructure: load-bearing walls, columns, beams, floor slabs and roof structure;

  • External fabric: external infill walls, insulations, external finishes, windows and doors;

  • Internal elements: internal floor finishes, internal non-load-bearing walls and internal doors; and

  • Building services: heat generator and installations.

Fig. 6
figure 6

Embodied Global Warming Potential of the one-storey detached building sample according to construction groups (reference construction system)

Full size image

In one-storey single-family houses, superstructure and the external fabric have the largest share of the impacts with 28% and 32% on average, respectively (Fig. 7). Even though the one-storey buildings are calculated without a basement, foundation and the slab-on-ground account for a considerable share of the impacts with 22% on average. The share of the internal elements and services are 11 and 6%, respectively. The inclusion of the heat distribution (pipes) and other elements (valves, pumps, etc.) would slightly increase the GWP of the technical system by about 10% (Decorte et al. 2024). The external fabric has the highest dispersion as this depends on the compactness of the building shape. In multi-family houses, the share of the internal elements increases whilst the role of the external fabric is reduced. The total share of substructure plus superstructure equals to about 50% for all building types except the 2-storey buildings.

Fig. 7
figure 7

Contribution of construction groups in the average embodied Global Warming Potential of the one-storey building sample (reference construction system)

Full size image

Another division of building elements is possible where the elements are classified into two groups: envelope and other elements. This classification may be interesting because it is foreseen that the revised EPBD regulations will include a requirement to calculate the life cycle GWP of new buildings as part of the energy performance calculations and requirements. Currently, energy performance calculations include only the building thermal envelope, i.e. the building surface that is in contact with the external environment or unheated spaces. Other internal elements are usually excluded from the analysis because they do not have a direct effect on the energy balance. More detailed energy models for dynamic simulations also include internal walls and floors because they act as thermal mass in the building but these are usually not considered in the simplified calculations needed for a building permit or energy performance certificate. Other elements, such as foundation and lintel beams, are not even included in detailed energy calculations.

The exact rules for calculating the life cycle GWP are not defined yet in the EPBD. The question arises whether it is sufficient to calculate the embodied GWP of only the building thermal envelope. As visible in Fig. 8, the envelope accounts for about 70% of detached houses and only about 30–40% of apartment houses. This means that a high share of the impacts is embedded in other elements beyond the building thermal envelope and only a fraction of the total impacts would be captured if the system boundary was chosen to be the building thermal envelope only, especially in multi-storey buildings. The calculation of the total embodied impact will require more input data and increase the workload of the energy assessor, but the consideration of the “other” elements is important.

Fig. 8
figure 8

Embodied Global Warming Potential of the one-storey detached building sample according to envelope and other elements (reference construction system)

Full size image

4.4 Other construction systems

The reference construction system is the most widespread in Hungary but others are also used, although with a lower market share (the exact statistics are not available). The reference system (insulating bricks with external insulation) is compared with three construction systems:

  • Heavy bricks with external insulation,

  • Autoclaved aerated concrete system (AAC) and

  • Timber frame system.

Insulating bricks are commonly used special hollow bricks with a thermally optimised structure and a density of about 740 kg/m3. Heavy bricks have a higher density of around 1400 kg/m3 and a high thermal mass. In apartment buildings, all the systems are assumed to be made with a reinforced concrete slab. Timber frame systems are currently only applied in detached houses because of fire regulations. The exact construction build-ups are shown in Annex 1. These systems are all well-established systems on the market; no innovative systems are considered. The box plot for one-storey buildings shows that the embodied emissions of insulating bricks and AAC systems are in the same range, whilst heavy bricks have higher impacts and timber frame systems have lower impacts (Fig. 9).

Fig. 9
figure 9

Box plot diagram of the embodied Global Warming Potential of the one-storey building sample with different construction systems

Full size image

4.5 Other environmental impact categories

Annex 4 and Annex 5 show the embodied impacts for the other environmental impact categories besides GWP for the building sample with the reference construction system and for one-storey buildings with different construction systems. The environmental impacts are assessed for the core indicators of prEN 15978. Additional environmental impact indicators are not evaluated, which is a limitation of this study. Benchmarks are currently developed only for GWP in most national and international certifications, but for other purposes, information on other impact categories may also be relevant.

The depletion of abiotic resources — fossil fuels (ADP-f) and the photochemical ozone formation (POCP) categories show very similar tendencies to GWP. The order of the building types and the construction systems is the same. For acidification (AP), two-storey buildings are in the same range as one-storey buildings, whilst timber frame buildings show relatively higher values than in GWP and are in the same range as heavy brick constructions. In the ozone depletion (ODP) category, building services play a significant role; they account for about 40% of the total impacts. In the depletion of abiotic resources — mineral and metal category (ADP-mm), timber buildings show higher values than the other construction systems. In the eutrophication (EP) categories, terrestrial and marine eutrophication show similar ratios as GWP concerning the building types and construction systems. In the freshwater eutrophication category, the autoclaved aerated concrete technology shows significantly higher values than the other systems, but altogether the values are small.

4.6 Benchmark setting

Researchers recommend setting benchmark values to some statistical indices of a building sample. Business as usual can be the mean or median value of the environmental performance of a building sample, whilst best practice may be defined by the best performance of the building sample, for example the first quartile (VTT 2012).

The statistical approach described in this paper is suitable for setting reference or baseline values. Our sample is a virtual sample, describing the technically feasible buildings with typical construction systems. The distribution of the sample will be different from the real-world sample. Hence, for a limit value used in the context of regulations, the 3rd quartile or 90th percentile of the sample could be recommended, which means that 75% or 90% of the technically feasible new buildings will have a lower environmental performance than this value. For GWP, the 90th percentile would mean 14.5 kg CO2-eq/m2/yr for detached houses and 13.2 kg CO2-eq/m2/yr for apartment buildings. Such values would be suitable in a building regulation that can be realistically fulfilled by most new buildings. With time, the baseline values can be lowered, as technology develops and building materials with lower embodied impacts become available.

5 Conclusions

This paper proposed a method for setting reference benchmark values for the embodied impacts of buildings. In the method, building categories and basic building parameters are defined (e.g. floor area, window-to-wall ratio) together with the realistic ranges of the parameters. A large building sample is then artificially generated based on the parameters, and the environmental performance is calculated. Limit values can be set based on the statistical indices of the sample; for example the 90th percentile will be fulfilled by most future new buildings. With this approach, extreme buildings with untypical geometries and solutions would not be captured. However, these are less important when setting benchmarks on a building stock level.

This method is suitable for estimating the environmental impact of typical new buildings in countries where a large pool of real building data is not yet available. With this method, preliminary reference values can be developed that can be later refined as LCA methodology becomes more widespread and the pool of real buildings grows. The method is also suitable for calculating whole life cycle benchmark values, which are preferable in order to capture the full environmental impact of buildings and consider the trade-offs between embodied and operational impacts.

The development of benchmark values is important as benchmark values help orient designers amongst environmental performance values. Benchmark values can be used in national legislation or voluntary sustainable rating systems. In this paper, values were determined with a bottom-up approach for setting the baseline. In the future, these can be combined with top-down target values based on the planetary boundaries or the 2-degree target to determine the environmental performance gap.