Carbon Footprint Variability in Engineered Wood Products for Timber Buildings: A Systematic Review of Carbon Accounting Methodologies

Abstract

Engineered wood products (EWPs) and timber buildings are increasingly recognised for their potential to reduce greenhouse gas emissions by storing biogenic carbon and replacing emission-intensive materials. This article systematically evaluates the carbon footprint (CF) of EWPs and timber buildings during the production stage (A1–A3), identifies key sources of variability, and extracts quantitative emission reduction metrics. Based on a review of 63 peer-reviewed studies, CF values vary widely, from −40 to 1050 kg CO₂eq m⁻² for buildings and 12 to 759 kg CO₂eq m⁻³ for EWPs, due to inconsistent system boundaries, functional units, and emission factor assumptions. Median CFs were 165.5 kg CO₂eq m⁻² and 169.3 kg CO₂eq m⁻³, respectively. Raw material extraction (50.7%), manufacturing (37.1%), and transport (12.2%) were the dominant contributors. A mitigation matrix was developed, showing potential reductions: 20% via transport optimisation, 24–28% through low-density timber, 76% from renewable energy, 11% via sawmill efficiency, 75% through air drying, and up to 92% with reclaimed timber. The geographic skew toward Europe and North America underscores the need for region-specific data. The findings provide actionable benchmarks and strategies to support carbon accounting, emissions modelling, and climate policy for more sustainable construction.

1. Introduction

1.1. Background

Engineered wood products (EWPs) and mass timber are increasingly recognised as viable substitutes for high energy-intensive materials, such as steel and concrete, due to their lower environmental impact and potential to mitigate climate change [1]. In late 2023, during the 28th Conference of the Parties (COP 28), 17 countries committed to increasing the use of timber in buildings by 2030 as part of the global transition to net zero [2]. The Intergovernmental Panel on Climate Change (IPCC) also emphasises the expanded use of timber in construction as a key climate mitigation strategy due to its ability to reduce greenhouse gas (GHG) emissions and act as a long-term carbon sink [3].

EWPs are widely used in modern construction because of their sustainability and efficient utilisation of wood resources, particularly in response to the declining availability of high-strength timber from old-growth forests [4]. These products can be classified based on their manufacturing processes and the types of wood fibres used (Table 1). EWPs include bonded wood flakes, fibres, or veneer sheets—such as laminated veneer lumber (LVL) and oriented strand board (OSB). Mass timber refers to larger structural panels such as cross-laminated timber (CLT) and glue-laminated timber (Glulam) [5,6,7]. Further classification considers mechanical, thermal, optical, and energy performance properties, which can be enhanced through chemical modifications and structural engineering techniques [8]. The EWP market spans multiple sectors, including furniture manufacturing and construction, with growth influenced by market fluctuations, raw material supply, and international trade competition [9].

There are numerous studies which compared timber with other traditional building materials on environmental impact and structural performance. The most interest is in environmental performance. Timber buildings can reduce carbon emissions by 32.6% and 40.4% compared to brick and reinforced concrete, respectively [10]. This reduction is even more significant for mass timber, which achieves a 14–52% decrease in whole-building embodied carbon compared to concrete and steel [11]. During the product stage, carbon emissions from timber buildings are only 50–60% of those associated with steel or reinforced concrete constructions. Over the full lifecycle, timber buildings emit approximately 90% of the carbon associated with steel or reinforced concrete buildings, further emphasising wood’s potential as a sustainable and carbon-efficient material for construction [12]. As for mechanical properties, timber exhibits a compressive strength ranging from 500 kg cm⁻² to 700 kg cm⁻² and a tensile strength of 500 kg cm⁻² to 2000 kg cm⁻², making it suitable for structural applications [13]. Moreover, Bazli et al. reviewed 100 research studies and found that timber exhibits a high strength-to-weight ratio, making it competitive with materials such as steel and reinforced concrete in structural applications [14]. Its outstanding acoustic and thermal insulation properties further enhance its performance in building systems. Despite these clear benefits, the effective adoption of EWPs and accurate quantification of their carbon impacts face several challenges, particularly in carbon accounting and methodological consistency.

Table 1. EWPs classification based on material used.

Classification	Subtype	Picture	Source
Fibre-based	Wood cement fibreboard (WCF)		[15]
	Wood plastic composite (WPC)		[16]
Lumber-based	Glue laminated timber (Glulam)		[15]
	Nail laminated timber (NLT)		[15]
	Dowel laminated timber (DLT)		[15]
	Cross laminated timber (CLT)		[15]
Strand-based	Oriented strand board (OSB)		[15]
	Laminated strand lumber (LSL)		[15]
	Oriented strand lumber (OSL)		[15]
Veneer-based	Plywood		[15]
	Laminated veneer lumber (LVL)		[16]
	Parallel strand lumber (PSL)		[15]

Table 1. EWPs classification based on material used.

Classification	Subtype	Picture	Source
Fibre-based	Wood cement fibreboard (WCF)		[15]
	Wood plastic composite (WPC)		[16]
Lumber-based	Glue laminated timber (Glulam)		[15]
	Nail laminated timber (NLT)		[15]
	Dowel laminated timber (DLT)		[15]
	Cross laminated timber (CLT)		[15]
Strand-based	Oriented strand board (OSB)		[15]
	Laminated strand lumber (LSL)		[15]
	Oriented strand lumber (OSL)		[15]
Veneer-based	Plywood		[15]
	Laminated veneer lumber (LVL)		[16]
	Parallel strand lumber (PSL)		[15]

1.2. Challenges and Research Gaps in Carbon Accounting of EWPs

Despite these advantages, the adoption of EWPs is hindered by several limitations. The integration of steel connectors, fire-resistant coatings, and energy-intensive treatments such as pressure and chemical applications contributes additional embodied carbon and cost [17]. In regions with limited local timber production, such as Australia and Japan, the reliance on long-distance transportation further exacerbates costs and emissions.

Technical and institutional barriers also hinder the effective use of EWPs in construction. Issues such as susceptibility to water damage, fire risks, and limited expertise in timber applications remain significant concerns [18]. In Germany, insufficient professional knowledge, combined with concerns about fire safety and structural stability, poses additional challenges [17]. A reported shortage of skilled labour across the construction industry complicates the implementation of timber-based solutions, slowing its integration into mainstream building practices.

Several factors are required for the accurate calculation of carbon footprints (CF) for timber products. Variability in carbon accounting methodologies, such as life cycle assessment (LCA), material flow analysis (MFA), and IPCC Tier 3 approaches, often produces diverse results due to differences in underlying assumptions and system boundaries [19]. LCA standardisation for timber buildings is data intensive, requiring detailed building inventory models that introduce significant uncertainties [20]. Accurately modelling forest carbon dynamics is further complicated by regional differences in forest types, harvesting practices, and supply chain emissions. The globalisation of forestry supply chains exacerbates these challenges, as tracing the geographic origins of wood fibres and their associated emissions is often unclear [21].

LCA is widely used to evaluate the CF of products and services across their life cycle, from raw material extraction to end-of-life (EoL) disposal [22]. Applications of LCA to wood-based products, including EWPs and timber buildings, are extensive [23,24]. However, critical gaps remain in building-related LCA research. For example, in a review of 161 studies on building life cycle carbon emissions, Huang et al. found that while 92.5% of studies included the production stage (A1–A3), only 16% of studies provided detailed breakdowns of emissions from processes within each stage, such as raw material extraction, transportation, and manufacturing [25]. This highlights a significant knowledge gap in carbon accounting for wood-based construction materials. Addressing these gaps is essential for improving accuracy, transparency, and comparability in carbon accounting for the broader adoption of EWPs in sustainable building practices.

1.3. Research Objectives

This study addresses an important knowledge gap in the carbon accounting for timber buildings and EWPs, with the potential to inform both policy and practice. By improving methodological transparency and consistency, it supports the broader transition toward low-carbon construction and the effective integration of timber into climate mitigation strategies.

Despite extensive research on carbon accounting in the built environment, prior research studies exhibit several shortcomings. Most lack a systematic comparison of how standards and guidelines evolved and interconnected over time. Few efforts recognised this progression or evaluated its implications for methodological consistency. Furthermore, while carbon accounting methods such as LCA or MFA are widely used, their application to timber buildings and EWPs remains under-examined, especially regarding emission variability at the A1–A3 production stage.

This review addresses these knowledge gaps through four key innovations:

(1) it maps and compares major carbon accounting standards, tracing their development and interconnection over time;

(2) it analyses the application of LCA, MFA, and IPCC methods specifically in timber and EWP contexts;

(3) it systematically quantifies CF variability across 63 studies, creating regional and sub-stage emission benchmarks; and

(4) it introduces an original quantitative mitigation matrix, estimating emission reductions for material substitution, energy transition, process optimisation, and reuse strategies.

Additionally, the Discussion section presents new insights on the geographic concentration of research, the underutilisation of dynamic carbon accounting modelling, and the methodological biases limiting broader adoption of EWPs. These findings reinforce the need for region-specific data, harmonised methodologies, and improved data transparency. Overall, the study contributes valuable benchmarks, clarifies methodological uncertainty, and supports evidence-based policymaking for more sustainable construction.

2. Key Standards, Methods, and Applications of Carbon Accounting

Many studies attempted to define carbon accounting. Hespenheide et al. provided a broad definition, describing carbon accounting as the measurement of emissions and removals, with financial implications such as obligations to surrender allowances [26]. Ascui and Lovell highlighted that carbon accounting is interpreted differently by various stakeholders [27]. For scientists, it involves precise and verifiable GHG emission measurements, whereas governments focus on emissions reporting in relation to national commitments. Similarly, Stechemesser and Guenther, after reviewing 129 studies, argued that carbon accounting requires separate definitions across different levels, including national, project, organisational, and product scales [28].

Building on these perspectives, this study focuses on product-scale environmental carbon accounting, particularly in timber-based materials and structures. Timber products, such as EWPs and timber buildings, exemplify this scale and require comprehensive accounting approaches to quantify emissions, carbon storage, and substitution effects throughout their life cycles. An accurate representation of timber’s carbon footprint is essential, given its unique role in carbon sequestration and potential for material substitution for high-emission construction materials. The development of carbon accounting standards is relevant for timber products, as these frameworks provide structured methodologies for quantifying emissions, biogenic carbon storage, and substitution effects. Given timber’s unique role in carbon sequestration, selecting the right accounting approach is crucial for accurately assessing its climate change mitigation potential.

2.1. Carbon Accounting Standards and Guidelines

The structured evolution of carbon accounting standards addresses GHG emissions at different levels, including at the product level, organisational level, and sectoral level. Principle standards, the ISO 14040 and ISO 14044 series [29,30], establish the foundational principles and requirements for LCA, guiding the broader development of both product- and organisation-specific frameworks. Building on these, product-level standards, such as ISO 14067 Carbon Footprint of Products and Publicly Available Specification 2050 (PAS2050) [31,32], focus on assessing CFs at the product level, particularly using the LCA method. These product standards are complemented by EN 15804, which integrates lifecycle principles into environmental product declarations (EPDs) for the construction sector. Similarly, organisational standards, including the GHG Protocol Corporate Standard and ISO 14064-1, target corporate-level emissions, with the GHG Protocol Scope 3 addressing supply chain emissions [33,34,35] and the EU Organisation Environmental Footprint (OEF) advancing environmental reporting practices in Europe [36]. Sector-specific guidance, such as the IPCC Guidelines [37] track the long-term carbon sequestration potential of forests and timber products, emphasising biogenic carbon storage and fluxes. As shown in Figure 1, these standards and guidelines exhibit significant interconnectivity, with later frameworks often building on earlier methodologies. For instance, PAS 2050 draws inspiration from ISO 14044, and the GHG Protocol Product Standard aligns with ISO 14067.

Although these standards provide structured methodologies, differences in assumptions, system boundaries, and biogenic carbon treatment can lead to variability in reported results (summarised in Table S1 and supported by References [38,39,40,41,42,43,44,45,46,47,48,49,50,51]). Garcia and Freire applied ISO 14067, the GHG Protocol Product Standard, and PAS 2050 to calculate the carbon footprint of particleboard, reporting cradle-to-gate emissions ranging from −939 to 188 kg CO₂eq m⁻³ [52]. These variations underscore how methodological choices, particularly regarding biogenic carbon accounting, impact reported outcomes. Similarly, Peter et al. found that the IPCC Tier 1 method overestimated fertiliser-induced emissions by up to 50% compared to Tier 2, highlighting the importance of data resolution and methodological complexity in emission calculations [53,54].

One of the key challenges in timber carbon accounting is the treatment of biogenic carbon emissions and removals. Many existing standards lack explicit guidelines on how to account for carbon storage in timber products over time, leading to inconsistencies in reported values. PAS 2050 (2011) addresses biogenic carbon by calculating credits based on the weighted average time of carbon storage over a 100-year assessment period, with fossil and non-CO₂ biogenic emissions multiplied by their respective globe warming potential (GWP) over the same time limit [55]. ISO 14067 (2018) assumes that carbon uptake in forests constitutes a negative emission (removal) while the release of stored carbon represents a positive emission. The standard provides flexibility for the application of time-weighted metrics, but mandates the inclusion of emissions associated with land use change [32]. This requirement ensures that the biogenic carbon benefits from timber growth are not overestimated where land conversion emissions are significant.

Despite these frameworks, current carbon accounting standards face notable limitations which are relevant for timber products. Dye et al. highlight challenges, such as the complexity of diverse forestry activities, annual variability in emissions, and difficulties in accounting for natural disasters such as wildfires and pest outbreaks [56]. These issues directly impact the accurate estimation of emissions and carbon storage for timber, especially in regions with significant variability in forest conditions. Lenzen and Murray underscore discrepancies between international frameworks, such as the UNFCCC, which attributes CO₂ removals to forestry sectors, enabling ownership and trade of sequestered carbon [57]. In contrast, the System of Environmental and Economic Accounting (SEEA) treats CO₂ sequestration as an ecosystem service without explicit ownership rights. For timber products, these differences create uncertainty in how carbon storage and removals are accounted for and reported, potentially influencing industry claims and market incentives. These limitations highlight the need for more harmonised and robust accounting approaches to address the specific characteristics of timber-based materials and their lifecycle impacts.

Figure 1. Evolution of carbon accounting standards and guidelines for timber buildings and EWPs (italics indicate the interconnectivity) [29,30,31,32,33,34,35,36,37,38,47,49].

Figure 1. Evolution of carbon accounting standards and guidelines for timber buildings and EWPs (italics indicate the interconnectivity) [29,30,31,32,33,34,35,36,37,38,47,49].

2.2. Carbon Accounting Methods

Following from the foundational understanding of carbon accounting standards and their relevance to timber products, it is essential to investigate the methodologies that underpin these frameworks. Jasinevičius et al. reviewed 89 carbon accounting models on wood products and found the most used three methods are the IPCC tiered approach, MFA, and LCA [58]. Each method differs in scope, principles, and applications, and together they provide complementary perspectives on carbon accounting.

2.2.1. IPCC Guidelines for Timber Buildings and Products

The IPCC Guidelines for National Greenhouse Gas Inventories (2006; 2019 Refinement) focus on estimations and removals of emissions at a national or sectoral level. Timber products are accounted for as harvested wood products (HWPs), with an emphasis on biogenic carbon fluxes such as carbon sequestration, emissions from decay, and land use change impacts [59]. The IPCC uses the tiered approaches (Tier 1 to Tier 3) based on the data availability. Tier 1 assumes no carbon stock in EWPs. Tier 2 applies default emission factors (EFs) when country-specific data are unavailable. Tier 3 requires advanced models (e.g., EFs are adjusted with region-specific data to reflect local characteristics, including decay dynamics) and high-resolution datasets to provide more accurate estimates.

When it applies to wooden buildings and EWPs, Tier 2 is commonly used because it strikes a balance between accuracy and feasibility. Tier 1 is often too simplistic to capture emission variability, while Tier 3, despite offering greater accuracy, is rarely used due to its high data demands, model calibration requirements, and time-intensive processes [60]. Peter et al. (2016) noted that Tier 2 is regarded as a medium-effort approach that provides a practical compromise: it enhances accuracy over Tier 1 without the complexity of Tier 3 [53]. Jasinevičius et al.’s [60] study applied Tier 3 methods using forest data from the Czech Republic and compared the results to Tier 2 estimates based on the FAOSTAT database. Their findings indicate that forest carbon stock estimates were 16% higher with the Tier 3 approach. Conversely, Jang and Youn conducted a similar study in South Korea, estimating carbon storage from 1970 to 2080, and found that Tier 2 methods calculated values approximately twice as high as those from Tier 3 [61].

2.2.2. LCA Methods for Timber Buildings and Products

LCA is a widely adopted method for carbon accounting, assessing environmental impacts across a product’s lifecycle, from raw material extraction to disposal or reuse. It is grounded in the ISO 14040/44 standards and relies on detailed life cycle inventory (LCI) data for each stage. It applies EFs to quantify emissions at each stage, allowing for a cradle-to-grave or cradle-to-gate system boundary [62]. Various LCA methods are employed based on complexity and data availability. For instance, process LCA requires detailed lifecycle inventories, forming the backbone of EPDs. As shown in Figure 2, EPD standards evolved to enhance accuracy, progressing from basic lifecycle data to modular frameworks such as EN 15804, which include biogenic carbon indicators for timber products and buildings. Input–output LCA (IOA-LCA) is less work-intensive than process LCA to conduct, as it applies national IO tables with EFs to estimate impacts across the broader supply chain (e.g., how outputs from forestry contribute to construction) [63]. Liu et al. analysed the CLT manufacturing sector IO table in Japan and found that the total GHG emissions in 2020 from the CLT manufacturing sector were approximately 17% those of the cement sector and 15% of emissions from the steel sector [64].

Hybrid LCA combines IOA and process-based LCA for a more comprehensive system boundary and more accurate results [65]. Lausselet et al. used hybrid LCA to calculate coefficients for embodied GHG emissions in Norway’s building sector, comparing it to process-based LCA [66]. The results show that timber products accounted for 24% of total emissions in process LCA, but this increased to 40% in hybrid LCA, while concrete dropped from 26% to 14%. The increased share of timber in hybrid LCA reflects the significant influence of IO data on hybrid coefficients. However, Yang et al. argue that while hybrid or IOA LCA covers a complete system boundary, it is not necessarily more accurate than process LCA due to sectoral aggregation, which may misrepresent products with diverse technological and environmental profiles, leading to inaccuracies despite avoiding double counting [67]. Lastly, dynamic LCA introduces temporal dimensions, which are especially relevant for EWPs used in construction, where long-term carbon storage and delayed emissions significantly affect the environmental footprint [68]. Peñaloza et al. assessed a four-story CLT building’s lifecycle emissions over 50 years using both static and dynamic LCAs [69]. The study found that dynamic LCA resulted in approximately 10% and 22% reductions in GHG emissions compared to static LCA when using 100-year and 300-year time horizons for climate impact assessment, respectively.

Figure 2. EPD framework and lifecycle stages for timber products and buildings, adopted from EN 15804:2012 + A2:2019 + AC:2021 (green: mandatory modules, blue: optional modules) [70].

Figure 2. EPD framework and lifecycle stages for timber products and buildings, adopted from EN 15804:2012 + A2:2019 + AC:2021 (green: mandatory modules, blue: optional modules) [70].

2.2.3. MFA Methods for Timber Buildings and Products

MFA is a quantitative method used to track and analyse physical material flows across defined systems, applying carbon content or EFs to each flow. It treats the system as a mass balance, where inputs, outputs, accumulation, and losses are explicitly modelled [22]. In wood products, MFA calculates carbon emissions by applying carbon EFs to each lifecycle stage (e.g., harvesting, processing, transportation, and disposal), while also accounting for the carbon stored in the wood [71]. MFA is commonly used to estimate wood flows within the forestry sector [72]. Static MFA focuses on material flows within a system at a specific lifecycle stage or a fixed time limit. For instance, Bergeron simulated the metabolism of wood and waste wood in Switzerland over a century, developing 32 different combinations of forest harvesting and waste wood treatment [73]. Their results indicate that energy recovery from waste wood could generate 2110 GWh per annum and reduce 364 t of CO_2-eq annually. Dynamic MFA, on the other hand, accounts for material flows that change over time, incorporating variations in stocks, uses, and waste over a period. Wang and Haller analysed wood flows in Germany from 1991 to 2020, including consumption, production, imports, exports, waste, recycling, and stocks [74]. They found that over 80% of wood was used as material, with carbon emissions peaking in 2007 due to storms.

2.2.4. Comparison of Carbon Accounting Methods

EFs are essential to carbon accounting methods, translating activity or material flow data into greenhouse gas emissions. Different carbon accounting methods apply EFs in distinct ways (see Equations (1)–(3) for IPCC, LCA, and MFA methods, respectively).

The IPCC method (Equation (1)) uses activity data (e.g., volume of harvested wood) multiplied by an EF (e.g., kg CO₂eq per m³) to estimate emissions. This approach provides a consistent reporting framework for national inventories under international protocols such as the Paris Agreement [75]. Tiers 1–3 vary in data specificity, with Tier 3 offering more accurate, localised estimates by incorporating forest-specific parameters such as species growth, decay rates, and land use dynamics [76].

In LCA (Equation (2)), emissions are calculated by summing stage-specific emissions across the timber product’s life cycle—from raw material extraction to end-of-life. Each stage’s mass or volume is multiplied by its EF (e.g., manufacturing energy intensity or transport distance), providing a comprehensive view of cradle-to-gate or cradle-to-grave impacts. These estimates help identify carbon hotspots and guide material optimisation strategies [77]. In contrast, MFA (Equation (3)) focuses on carbon flows through a defined system, applying EFs to track inputs, outputs, and accumulation. While static MFA captures system balances at a given time, dynamic MFA models long-term stock changes and emissions trajectories. This is particularly valuable in forestry systems where delayed emissions and long-term carbon storage are critical [78].

• IPCC methods:

  $$\begin{array}{c}E=AD·EF\end{array}$$

  (1)

  where

E: Emissions, kg CO_2eq, (e.g., GHG emissions from timber products);

AD: Activity data, kg or m⁻³, the quantity of timber or wood product at each lifecycle stage (e.g., harvested timber, wood processed for construction);

EF: Emission factor, kg CO₂eq kg⁻¹ or kg CO₂eq m⁻³, represents the GHG emissions per unit of activity (e.g., per cubic meter of timber). It can be default (Tier 1 and Tier 2) or adjusted (Tier 3).

• LCA method (process-based):

  $$\begin{array}{c}E={\displaystyle \sum _{i=1}^n}\left(M_i·E{F}_i\right)\end{array}$$

  (2)

  where

n: Number of life cycle stages;

E: Emissions, kg CO_2eq, represents the total emissions from the timber product’s life cycle;

M_i: Mass or quantity of timber used in the ‘i’ life cycle stage, kg (e.g., processing, transport, and disposal);

EF_i: Emission factor for the ‘i’ life cycle stage, kg CO₂eq kg⁻¹ or kg CO₂eq m⁻³ (e.g., processing, transport, and disposal).

• MFA method (static MFA):

  $$\begin{array}{c}E={\displaystyle \sum _{i=1}^n}\left({MF}_i·E{F}_i\right)\end{array}$$

  (3)

  where

E: Emissions, kg CO_2eq, represents the total emissions from the timber product’s life cycle;

MF_i: Material flow for each stage (e.g., harvested timber, processed wood, and waste wood) through the system;

EF_i: Emission factor associated with the material flow at each stage, kg CO₂eq kg⁻¹ or kg CO₂eq m⁻³, which can include carbon emission, carbon sequestration, waste wood treatment impacts, etc.

The methods are interconnected. For instance, the IPCC tiered approach leverages LCA and MFA to improve accuracy. MFA often enhances carbon storage estimates, as shown in a Czech study where Tier 3 methods using MFA estimated carbon inflow in EWPs 15.8% higher than Tier 2 [60].

Table 2. Comparisons of IPCC, LCA, and MFA methods.

	IPCC	LCA	MFA
Principle	Tiered approach (Tier 1 to Tier 3) using EFs to estimate GHG emissions and carbon stocks	Tracks carbon emissions and sequestration across the lifecycle stages	Tracks material flows and carbon stocks through production, use, and end-of-life phases
Scale of analysis	National to regional level	Product or project level	System-level (e.g., material flows within an economy sector)
EFs application	Tiers 1 and 2 rely on standardised EFs; Tier 3 adjusts EFs using region-specific data	Phase-specific EFs for each lifecycle stage (e.g., production, use, disposal)	Links EFs to material flows for tracking direct emissions and carbon storage over time
Strengths	Standardised approach for global comparability Supports national-level GHG reporting Links to policy frameworks (e.g., Paris Agreement)	Comprehensive lifecycle analysis, covering all phases from raw material to disposal Suitable for product-specific CF Well-established databases for various industries	Integration of diverse data sources from different stages of the supply chain Useful for long-term carbon sequestration modelling. Tracks reuse, recycling, and decay dynamics
Limitations	Regional approach can incentivise local emission reductions at the expense of global optimisation Lack of a unified global framework leads to challenges in transparency and comparability across regions and sectors	Requires extensive data for all lifecycle stages Different model choices can lead to variations in outcomes (e.g., especially for building material, with the large variety of materials, different life cycle stages, and long service lives)	Regional biases (the regional focus can also result in a lack of climate justice) Complex and integration challenges with other methods Limited applicability for short-term analyses
Key tools	FAOSTAT database, national inventory datasets, default, and region-specific EFs	Databases such as Ecoinvent and EPDs	Material flow datasets, stock-flow models, and decay dynamic models
References	[54,79,80]	[81,82]	[54,83,84]

summarises and compares IPCC, LCA, and MFA methods, highlighting their principles, applications, strengths, and limitations.

Overall, the selection of carbon accounting methods reflects a trade-off between analytical precision, data intensity, and applicability. IPCC guidelines are widely used in national reporting due to their standardised and internationally accepted framework, which facilitates policy making and cross-country comparisons [85]. LCA is valued for its detailed tracking of emissions across all life cycle stages [86]. Hybrid and dynamic LCA methods offer improved modelling precision but remain challenging to apply consistently due to data intensity and lack of standardisation. MFA is increasingly used to model long-term carbon flows and stock dynamics in timber systems, contributing valuable insights for circular economy strategies [87]. However, MFA often lacks resolution on indirect emissions and is limited by data gaps and simplified assumptions around decay, reuse, and carbon substitution [83]. These methodological limitations can lead to divergent emission estimates for the same system [54,85], hindering comparability, weakening policy alignment, and reducing trust in carbon reporting. Addressing these challenges requires greater methodological transparency and a shift toward harmonised or hybridised approaches that integrate the complementary strengths of IPCC, LCA, and MFA. Such integration can enhance the consistency, reliability, and global applicability of carbon accounting for timber-based systems [88].

3. Methodology

This study employs a systematic review method to identify and analyse carbon accounting studies on timber buildings and EWPs, with a focus on the production stage [89]. This review was conducted in accordance with the PRISMA 2020 guidelines [90]. Peer-reviewed journal articles published between 2014 and 2025 were retrieved from ScienceDirect and Web of Science Core Collection. The keyword strategy was designed to capture studies focused on carbon accounting methods and their application to timber products. It combined terms such as “life cycle assessment”, “carbon footprint”, “cradle-to-gate”, “engineered wood products”, and “timber buildings” to ensure comprehensive coverage of relevant literature. As shown in Figure 3, an initial pool of 1205 records was identified, with 188 duplicates removed. The remaining studies were screened in two steps based on title, abstract, and a full-text review, using predefined exclusion criteria (e.g., lack of quantitative carbon, failed to specify life cycle stages relevant to the production stage). In total, 51 studies met the inclusion criteria. Additionally, 12 EPD reports were manually included, as they provided valuable and comparable insights not captured through the initial search terms. In total, 63 studies were included for analysis, with a summary of the reviewed study characteristics and methodological details provided in Supplementary Materials I Tables S2–S6, and the completed PRISMA 2020 Checklist is available as Supplementary Materials II. All 63 included studies are cited here to ensure proper attribution and are listed in Supplementary Tables S2–S6 [69,91,92,93,94,95,96,97,98,99,100,101,102,103,104,105,106,107,108,109,110,111,112,113,114,115,116,117,118,119,120,121,122,123,124,125,126,127,128,129,130,131,132,133,134,135,136,137,138,139,140,141,142,143,144,145,146,147,148,149,150,151].

4. Results

4.1. Bibliometric Analysis

A total of 63 journal articles and reports on the carbon footprint (CF) of timber buildings and engineered wood products (EWPs) were reviewed, covering the period from 2014 to 2025 (Figure 3). Figure 4 presents a bibliometric overview of these studies, illustrating the annual publication trends and regional distribution. The number of publications increased steadily over time, with a notable peak between 2021 and 2024, reflecting the growing attention to decarbonising the building sector. Europe contributed the highest share of publications, followed by North America and Asia, highlighting the concentration of timber and LCA research in these regions. Although Oceania and South America contributed fewer studies, their recent upward trend indicates a rising interest in timber-based sustainability research across emerging markets.

The reviewed publications were distributed across 36 journals, with Sustainability contributing the highest number of articles (7), followed by Journal of Cleaner Production and Energy and Buildings (5 articles each). This reflects the interdisciplinary nature of the field, encompassing sustainability science, industrial ecology, and construction engineering.

The keyword analysis across 51 journal articles identified 164 unique terms, with 14.6% occurring more than once. After grouping similar terms, such as “life cycle assessment” and “life cycle analysis,” the most frequent keyword was “life cycle assessment,” appearing 24 times, followed by “cross-laminated timber” and biogenic carbon”. The co-occurrence network map (Figure 5) reveals significant clusters, with “life cycle assessment” at the centre, highlighting its pivotal role in timber building and EWP research. Key associated terms include “embodied carbon,” “environmental impacts,” and “construction,” reflecting a focus on sustainability and CF reduction over the past decade.

4.2. System Boundary and Functional Unit

System boundary selection influences LCA outcomes significantly [152]. Figure 6 illustrates the dominance of cradle-to-grave system boundaries, representing 42.9% of the reviewed studies. For timber buildings, 56% of the 34 studies used cradle-to-grave, with B6 (operational energy) considered in 60% of cases. Interestingly, the EoL stages (C1–C4) and post-EoL stage (D) were more often assessed than the building use stage (B1–B7). In contrast, 53% of EWP studies applied cradle-to-gate, followed by 24% using cradle-to-grave (details shown in Tables S3 and S5). Notably, inconsistencies were observed in boundary definitions. For example, cradle-to-site (A1–A5) overlaps with production and construction (P&C) in some studies (e.g., [153,154]). Additionally, Moncaster et al. conduct a cradle-to-site and EoL study for a wooden building in Cambridge, UK [117]. Allan and Phillips assessed similar stages but used cradle-to-grave to define the system boundary [94]. These variations highlight the need for greater methodological standardisation in LCA boundary definitions.

The functional unit varied across timber building studies, as illustrated in Figure 7 and Table S2, 1 m² of floor area being the most used (35.3%), followed by the entire building (32.4%) and heated floor area (8.8%). Other functional units, such as net floor area per annum and 1 m² of portal frame, accounted for smaller proportions but demonstrated the diversity in reporting practices. In contrast, nearly all LCA studies on EWPs adopted 1 m³ of product to use as the functional unit in most cases. For wood flooring studies, 1 m² was often adopted to align with usage-specific impacts.

4.3. Life Cycle Inventory (LCI) Database

The choice of LCI databases affects LCA results. As shown in Figure 8, Ecoinvent was the most frequently used database (24.7%), followed by SimaPro (15.1%) and USLCI (13.7%). Other databases, such as GaBi, Athena IE4B, and CORRIM, collectively accounted for 31.4%. Notably, some studies used software tools such as SimaPro (https://simapro.com/) without specifying the integrated LCI database (e.g., Ecoinvent), leading to inconsistencies in reporting. Approximately 36.5% of studies relied on multiple data sources, combining factory-collected data, localised databases, and expert interviews (e.g., [127,135]). While GaBi was used in 50% of EPD reports, it was not explicitly cited in academic publications, reflecting differing preferences between industry and academic research.

4.4. Engineered Wood Product (EWP) Categories

From the 63 articles reviewed, a total of 66 different cases were found, as some studies covered more than one type of EWP. Articles which did not specify a particular wood product were excluded from the analysis. As shown in Figure 9, CLT dominated, accounting for 31.25% of the cases. Glulam was the second most studied product, accounting for 12.5% of the cases, followed by wood panels, including plywood and OSB, at 15.6%. A significant portion of the studies (7.8%) fell into the miscellaneous category, which included niche or emerging products. This result demonstrates the prominence of mass timber products, which together accounted for over 40% of the cases analysed. However, the diversity of other EWPs suggests that the research focus is increasingly expanding to alternative and specialised materials, emphasising efforts to optimise CF and performance in different applications.

The results reveal that CLT is the most applied EWP in timber buildings, with approximately 80% of timber building studies featuring CLT or hybrid CLT (e.g., CLT with rock wool or wood fibre). In contrast, the LCA studies of other EWPs exhibit greater diversity, including glulam, lightweight timber (LWT), OSB, MDF, DHF, fibreboard, and LVL. Notably, most studies did not specify the tree species used, with the exceptions primarily from the USA. As detailed in Table S4, commonly referenced species for EWPs include Douglas-fir (Pseudotsuga menziesii), loblolly pine (Pinus taeda), western hemlock (Tsuga heterophylla), red oak (Quercus rubra), and white oak (Quercus alba).

4.5. Carbon Emission Analysis from the Production Stage

4.5.1. Regional Analysis of Carbon Footprint

The CF of timber-based buildings and EWPs was analysed across regions based on the A1–A3 production stage. To reflect data variability and distribution, the CF results were summarised using quartile statistics—median (CF_med), first quartile (CF_25%), and third quartile (CF_75%)—along with the interquartile range (IQR) (

Table 3. Carbon emissions calculation table for timber-based buildings.

	Asia	Europe	North America	Oceania
CFmed (kg CO2eq m−2)	190.5	163.9	160.4	319.2
CF25% (kg CO2eq m−2)	138.3	110.6	145.5	289.4
CF75% (kg CO2eq m−2)	287.0	330.3	289.4	349.0

CF_med, CF_25%, CF_75%: the median, first quartile, and third quartile of carbon emissions.

and

Table 4. Carbon emissions calculation table for engineered wood panels.

	Asia	Europe	North America	Oceania
CFmed (kg CO2eq m−3)	139.0	25.2	184.5	118.0
CF25% (kg CO2eq m−3)	124.5	15.1	162.8	97.8
CF75% (kg CO2eq m−3)	168.5	38.8	329.8	149.0

CF_med, CF_25%, CF_75%: the median, first quartile, and third quartile of carbon emissions.

). Outliers were identified using the 1.5 × IQR method, which classifies a value as an outlier if it lies more than 1.5 times the IQR above the third quartile or below the first quartile, following Tukey’s rule [155].

For timber buildings, the CF_med (CF_25%–CF_75%) was 190.5 (138.3–287) kg CO₂eq m⁻² in Asia, 163.9 (110.6–330.3) kg CO₂eq m⁻² in Europe, 160.4 (145.5–289.4) kg CO₂eq m⁻² in North America, and 319.2 (289.4–349) kg CO₂eq m⁻² in Oceania. Comparatively, for EWPs, the CF_med (CF_25%–CF_75%) was 139 (124.5–168.5) kg CO₂eq m⁻³ in Asia, 25.2 (15.1–38.8) kg CO₂eq m⁻³ in Europe, 184.5 (162.8–329.8) kg CO₂eq m⁻³ in North America, and 118 (97.8–149) kg CO₂eq m⁻³ in Oceania.

To enhance quantitative analysis, we calculated the standard deviation (SD) of CF values for each region (

Table 5. Standard deviation of carbon footprint values for timber buildings and EWPs during the production stage (A1–A3).

	Asia	Europe	North America	Oceania
SDBuildings (kg CO2eq m−2)	110.15	162.07	106.52	59.26
SDEWPs (kg CO2eq m−3)	32.59	17.52	123.85	37.33

), providing a measure of statistical dispersion beyond the quartile summaries already presented. For timber buildings, SDs were highest in North America (237.0 kg CO₂eq m⁻²) and Europe (166.4 kg CO₂eq m⁻²), reflecting a broader spread in reported CF values likely driven by variations in data sources, construction typologies, and system boundaries. For EWPs, North America again showed the greatest variability (SD: 258.2 kg CO₂eq m⁻³), followed by Asia (66.9 kg CO₂eq m⁻³). These high SDs reinforce the fact that methodological heterogeneity, such as LCI database selection or EF assumptions, significantly affects reported CF outcomes.

Regional variations highlight distinct trends. As seen in Figure 10 and Figure 11, North America exhibited the widest CF range, spanning from 90.66 to 1049.95 kg CO₂eq m⁻² for buildings and 36 to 759.15 kg CO₂eq m⁻³ for EWPs, reflecting diverse manufacturing practices and energy sources. Europe showed the lowest median CF values, particularly for EWPs, mainly because of its reliance on low-carbon energy grids and comprehensive LCA standards [156]. In contrast, Oceania had the highest CF_med for buildings (319.2 kg CO₂eq m⁻²), influenced by the materials used, such as steel and concrete.

4.5.2. Contributions of Substages A1–A3 to Carbon Footprint

Figure 12 highlights the distribution of CFs for the A1–A3 sub-phases. The median contributions for A1, A2, and A3 are 50.7%, 12.2%, and 37.1%, respectively (detailed data are shown in Table S7), with significant differences between studies. A1 raw material extraction dominates in most cases, with contributions ranging from 16.5% to 75%, reflecting the energy-intensive nature of forestry operations [157]. A2 has a median of 12.2%, due to the shorter transport distances in the studied cases and thus a smaller contribution, while A3 manufacturing emissions range widely (8.84% to 64.5%) due to different production processes and energy sources.

A comprehensive regional analysis shows that North America and Oceania have higher A1 contributions due to energy-intensive forestry practices, while Europe has higher A3 emissions, reflecting detailed reporting and production techniques. Notably, only 9% of wooden construction cases and 37.5% of EWP cases provided phased data, highlighting the need for standardised reporting.

4.5.3. Temporal Trend Analysis

Figure 13 represents the temporal trend of CF. For timber buildings, CF values range widely from nearly 0 to 760 kg CO₂eq m⁻² due to the differences in methodologies, boundary definitions, and material compositions across studies. A marked increase in data availability is observed after 2019, with a significant cluster of studies between 2022 and 2024 coinciding with heightened research interest in decarbonisation strategies. The variability in CF values is substantial, with lower values (<100 kg CO₂eq m⁻²) linked to buildings using low-carbon materials or energy-efficient designs, while higher values (>500 kg CO₂eq m⁻²) suggest intensive use of hybrid structures involving non-wood materials. For EWPs, CF values range from 11.9 kg CO₂eq m⁻³ to 759.15 kg CO₂eq m⁻³ with consistent reporting emerging after 2016, reflecting an increasing focus on material-level assessments. The highest CF values (>700 kg CO₂eq m⁻³) are associated with energy-intensive manufacturing processes or specific product types, such as laminated veneers. Comparing the two categories, the wider range of CF values for wood construction highlights the greater variability in building practices and system boundaries. While there is also variation in EWPs, it is more tightly clustered around the median, especially in Europe, reflecting the potential for more standardised reporting.

4.5.4. Biogenic Carbon and Dynamic Factor Analysis

Biogenic carbon, representing carbon absorbed and stored by bio-based materials such as wood, is considered in 55.6% of the reviewed studies. The calculation methodologies vary, reflecting differences in standards, assumptions, and analytical approaches. The majority of studies (e.g., [128]) employed calculations based on EN 16449, which estimates carbon stock in wood products using wood density and moisture content. The equation commonly used is as follows:

$$\begin{array}{c}{C}_s=V·BD·C_F(1-MC)F_S\end{array}$$

(4)

where

C_S: Carbon sequestered (kg CO_2eq m⁻³);

V: volume of the timber product (m⁻³);

BD: bulk density of the timber (kg m⁻³);

C_F: carbon fraction of dry wood (typically 0.5);

MC: moisture content (fraction, dry basis);

F_S: the molecular weight ratio of CO₂ to carbon, typically 3.67, with the molar mass of CO₂ being 44 g mol⁻¹ and carbon being 12 g mol⁻¹.

Several studies used software tools to compute the GWP_biogenic index without disclosing detailed calculations. For instance, Heidari et al. reported biogenic carbon emissions using default parameters embedded in SimaPro TRACI methodology, underscoring a lack of transparencies in calculation methods [132]. Some studies followed ISO 21930 [158] guidelines (e.g., [133]), which integrate biogenic carbon flows into the LCA framework. This method treats carbon absorption as a negative emission (−1 kg CO_2eq) during biomass growth and carbon release as a positive emission (+1 kg CO_2eq) during combustion or material decay. This balance highlights the dynamic nature of carbon cycling in bio-based materials. Lao et al. included the delayed biogenic carbon in the system [135], using the following equation:

$$\begin{array}{c}DE={\displaystyle \frac{100-0.76t}{100}}BCS\end{array}$$

(5)

where

DE: delayed biogenic emissions, kg CO_2eq m⁻³;

t: the service time of EWPs, annual;

BCS: biogenic carbon storage of EWPs, kg CO_2eq m⁻³.

A few studies applied custom dynamic LCA models to assess biogenic carbon. For instance, Lan et al. integrated a forest residue decay rate with Markewitz’s simulation to estimate carbon sequestration dynamics over a 100-year period [134,159]. Similarly, Doraisami et al. used field and remote sensing data to calculate the aboveground biomass, linking it with carbon content models to predict storage and emissions over time [160].

Only four cases included dynamic analysis [69,95,134,141]. These studies combined dynamic factors such as forest management practices, forest transformation (changes in carbon stocks), and emissions from EoL scenarios. For example, Andersen et al. evaluated the climate impacts of delayed carbon emissions over a 100-year building lifespan [95]. These studies highlighted the importance of wood rotation periods and temporary storage of carbon in building materials. Wang and Lan extended the dynamic analysis to EoL scenarios, evaluating delayed emissions from landfilled wood and bioenergy recovery [141].

5. Discussion and Implications

5.1. Impact of System Boundary and Functional Unit Selection

The cradle-to-grave boundary was the most commonly used framework observed in this study, aligning with findings from Liang et al. [114]. However, 27% of the reviewed cases adopted a cradle-to-gate boundary, considerably higher than the 8.6% reported by Huang et al. [25]. This variation is attributable to the focus on EWPs and their production stage (A1–A3), which excluded studies lacking quantified A1-A3 data. Similarly, 8.8% of timber building studies used a cradle-to-site boundary, while 53% of EWP studies employed cradle-to-gate boundaries, reflecting distinct priorities between building- and product-level assessments. These methodological differences pose challenges in aligning system boundaries for comparative analysis.

Inconsistencies in functional unit selection further complicate cross-study comparisons. A significant proportion (32.4%) of timber building studies used the entire building as the functional unit, which hinders standardisation due to differences in design, size, and usage. Research shows that larger dwellings exhibit lower impacts per m² due to resource distribution, while smaller homes display higher impacts due to greater resource intensity [161,162]. Standardising functional units to incorporate spatial, temporal, and service-related dimensions is essential to improve consistency, enabling more reliable insights for sustainable construction decision-making.

5.2. Analysis of Databases and Panel Types

The choice of databases significantly influenced the reported lifecycle impacts. Ecoinvent was the most widely used (24.7%), followed by SimaPro (15.1%), though the latter integrates various background databases, such as USLCI. GaBi, despite its global prominence, was absent from the reviewed studies, a notable finding given its methodological differences from Ecoinvent [163]. This disparity underscores the importance of transparency in database selection to ensure accurate interpretation of LCA results.

EPDs are increasingly prevalent due to their standardisation of environmental data. However, many EPDs lack traceability and omit critical lifecycle stages, particularly beyond cradle-to-gate, where discrepancies can exceed 90% [164]. Further, the limited detail provided by databases such as ÖKOBAUDAT restricts their applicability in assessing biogenic carbon storage and EoL impacts [165]. Harmonised data reporting and improved traceability are essential to enhance the reliability of timber-related LCAs.

Among EWPs, CLT dominated research due to its popularity, along with its growing market presence. Veneer-based mass panels (VBMPs) have been identified as promising alternatives due to their potential for more efficient resource utilisation [166]. However, further research is needed to validate the technical and commercial feasibility and overall sustainability of VBMPs, particularly in markets such as Australia, to expand sustainable timber options.

5.3. Large Discrepancies in CF Results

Significant variability in LCA results for timber-based buildings and EWPs is observed across studies, particularly in the production stage. For instance, CF values for similar timber building cases can differ by over 500%. This inconsistency reflects a combination of methodological differences, data quality issues, and regional variations. These variations limit the comparability and reproducibility of LCA results and hinder the development of coherent carbon accounting frameworks.

One major source of variability lies in the selection and quality of LCI data. Researchers may use global databases such as Ecoinvent or GaBi, country-specific datasets, or direct on-site measurements. These sources vary significantly in accuracy, temporal relevance, geographic specificity, and transparency. For instance, values derived from national databases often reflect local forestry and manufacturing practices, whereas international databases may rely on averaged or outdated data [167]. In addition, a recent evaluation by Guo et al. (2025) found that only 40–60% of sampled processes in major LCI databases had identifiable sources, and fewer than 5% were fully accessible [168]. This lack of harmonisation and documentation constrains reproducibility and undermines trust in reported CF values. Improving transparency in system boundaries, EFs, and biogenic carbon assumptions is essential for enhancing the credibility and comparability of timber LCA results.

Differences in how biogenic carbon is modelled also contribute to variability. Several studies report negative CF values by integrating biogenic carbon sequestration with production-stage emissions. However, the methods used to allocate or calculate sequestration are often poorly described, leading to inconsistent assumptions about system boundaries, carbon pools, and decay rates. For example, studies such as [124,169] present biogenic carbon benefits without clearly disclosing how material flows or carbon storage dynamics were calculated. This lack of transparency undermines the interpretability and credibility of reported CF values.

Timber’s unique supply chain introduces further challenges in data collection. Unlike more standardised materials, such as steel or concrete, timber products vary by species, region, treatment process, and intended application. Many LCA datasets omit critical parameters such as tree species, harvesting methods, or sawing efficiency, particularly for EWPs. This absence of detailed, species-specific data leads to significant uncertainty in CF outcomes. Moreover, inconsistent applications of LCA principles—such as differing allocation methods, truncation of forestry impacts, or omission of end-of-life pathways—exacerbate inter-study variability.

Cradle-to-gate studies on EWPs originate from the USA, which frequently reports higher CF values compared to other regions. This trend may be attributed to the availability of comprehensive forest supply chain databases in the USA, which provide detailed upstream emissions data often omitted in other regions. While these databases enrich the accuracy of individual analyses, they also contribute to variability in CF outcomes due to regional differences in energy sources, manufacturing practices, and forest management techniques.

5.4. Comparison with Benchmark Studies and Sensitivity Analysis

The red dashed lines in Figure 14 represent quantitative benchmarks derived from our systematic review (mean CF values: 247 kg CO₂eq m⁻² for buildings and 208 kg CO₂eq m⁻³ for EWPs). These serve as model-based baselines for future comparison and decarbonisation target setting. Benchmarking these results against prior reviews and industry reports reveals regional trends and areas for improvement [24,25,98,103,170,171,172,173,174]. For example, the mean CF of timber buildings identified in this review aligns with the findings of Hemmati et al. and reflects regional differences. North America (shown as NA in Figure 14) shows the highest emissions, likely due to comprehensive research and localised databases providing detailed inventory and EFs, whereas Europe (shown as EU in Figure 14) reports lower emissions, attributed to its strong sustainability orientation and high market demand for timber buildings [25,175]. For EWPs, our mean CF estimate is consistent with previous studies, though Australia reports the highest emissions, largely due to upstream activities such as log haulage and harvesting, accounting for 37% and 21% of emissions in softwood plantations and native hardwood forests, respectively [176].

While these benchmarks provide valuable insights, they are shaped by a literature base disproportionately focused on Europe and North America. This geographic concentration is not unique to carbon accounting; similar patterns have been identified in climate, ecology, and conservation science, where research tends to originate in wealthier nations with cooler climates and higher historical emissions [177,178]. In CF studies, the availability of detailed inventory data in North America and strong policy support in Europe enhance methodological depth, but also skew benchmarks toward conditions specific to high-income regions. As a result, current CF methodologies may not be directly transferable to regions with differing forestry systems, energy infrastructures, or socio-economic conditions.

Underrepresented regions such as Africa, South America, and Southeast Asia, despite their growing roles in timber supply chains, are rarely featured in CF studies. This mirrors geographic and taxonomic biases seen in other disciplines, where under-studied regions are constrained by limited funding, weak research infrastructure, and exclusion from high-impact journals [179,180]. The resulting narrow evidence base may lead to misinformed policy and ineffective mitigation strategies in data-poor regions. Addressing these disparities requires expanding geographic coverage, strengthening local research capacity, and ensuring more equitable participation in CF data development, modelling efforts, and policy discourse [181,182].

The CF of timber buildings and EWPs during the production stage shows significant variability across studies, driven by methodological differences, regional practices, and data assumptions. On average, the CF of timber buildings is notably higher than that of EWPs, primarily due to the additional materials required for structural connections, fire protection, and acoustic insulation. These components can increase upfront carbon emissions by up to 60% in residential developments and 30% in commercial buildings [183]. Transportation distances and energy supply systems also contribute significantly to CF variability. In regions such as Oceania and North America, where local production capacity is limited, long transportation distances elevate CF substantially. Optimising transportation logistics can reduce GWP by 20%, while utilising low-density timber species can further lower GWP by up to 24% [184]. Renewable energy systems, such as a mix of 70% wind and 30% biomass, have been shown to reduce emissions by up to 76% compared to fossil fuel-based systems [121]. These findings highlight the critical importance of addressing both material and logistical factors to lower CF in timber buildings.

For EWPs, CF variability is strongly influenced by forest management practices, the type of wood product, and transportation. Sustainable forestry practices, such as reduced impact logging in Southeast Asia, have been shown to reduce emissions from logging operations by 36.7% compared to conventional methods [185]. Selecting low-density timber species has also been shown to decrease CF by up to 28.3% [186], emphasising the importance of material choices. Kaulen et al., based on case studies in Germany, found that the net carbon storage of supplied timber is reduced by 1.5% to 5% due to emissions from harvesting and transport activities, with timber logistics identified as the largest contributor to these emissions [187].

Additionally, the sawmill efficiency significantly impacts timber carbon emissions. Efficient sawmill operations further minimise CF by lowering energy consumption during lumber processing. For instance, using advanced head saws, energy-efficient lighting, and optimised air compressors can significantly cut emissions [188]. The drying process for sawn timber also plays a critical role: Martínez-Alonso and Berdasco found that air-dried timber has a CF four times lower than kiln-dried timber due to its reduced energy requirements [189]. In the Northeastern USA, sawmill size and energy sources also impact emissions, with efficient management reducing emissions by approximately 10.96% [190]. These factors underscore the need for optimising forest management and production processes to achieve lower emissions in EWP production.

5.5. Synthesis of Emission Reduction Metrics and Implications

To address the significant variability in CF values identified across the reviewed studies, this section presents a synthesis of original findings derived from our systematic review. The red dashed lines in Figure 14, representing mean CF values of 247 kg CO₂eq m⁻² for timber buildings and 208 kg CO₂eq m⁻³ for EWPs, serve as benchmark baselines for comparison. These values offer a model-based foundation for setting decarbonisation targets and tracking improvements over time.

Building on these benchmarks, we developed a quantitative mitigation matrix that consolidates emission reduction strategies across life cycle stages (

Table 6. Quantitative emission reduction metrics for timber products and buildings.

Intervention Type	Estimated Emission Reduction Potential	Affected Life Cycle Stage(s)	Reference
Air-drying versus kiln-drying	Reduces manufacturing emissions by approximately 75%	A3—Manufacturing	[121]
Transportation optimisation	Decreases transport-related emissions by up to 20%	A2—Transport	[184]
Selection of low-density species	Lowers emissions by 24–28.3% during extraction and transport stages	A1—Raw material extraction, A2—Transport	[186]
Sawmill operational efficiency	Achieves approximately 11% reduction in processing emissions	A3—Manufacturing	[189]
Renewable energy substitution	Replaces fossil fuels, reducing emissions by up to 76%	A3—Manufacturing	[190]
Use of reclaimed timber	Reduces embodied emissions by up to 92% through material reuse	A1–A3 (avoided)	[191]

). The strategies (sourced from peer-reviewed studies synthesised in Section 4.5 and 5.4) were grouped into four categories: (1) material substitution, (2) energy system transition, (3) process optimisation, and (4) reuse and circular economy. Each strategy was selected for its documented reduction potential and relevance to timber production systems. The matrix does not merely collate prior findings; rather, it transforms them into an actionable tool to guide future emissions modelling, procurement specifications, and carbon disclosure practices.

Importantly, this matrix allows scenario-based reductions to be explored. For instance, combining material and energy interventions, such as switching to air-dried low-density timber processed in efficient sawmills powered by renewable energy, could potentially reduce the CF of EWPs from 208 kg CO₂eq m⁻³ to below 130 kg CO₂eq m⁻³. These values provide not only sectoral targets, but also policy-relevant indicators for investment prioritisation, especially in regions such as Oceania and North America where CF values remain high. Furthermore, mapping each intervention to its corresponding life cycle stage offers a structured basis for designing interventions that are both stage-specific and measurable.

5.6. Limitations and Future Research Directions

This review identified several gaps and limitations in current carbon accounting practices for timber buildings and EWPs:

(1) Methodological inconsistencies: Biogenic carbon accounting varies widely across studies, often omitting land use change emissions and carbon storage credits. There is also a risk of double counting, where both forestry and construction sectors claim the same sequestration benefits. Harmonising accounting rules and clarifying carbon credit allocation remain key challenges [21,191].

(2) Limited use of dynamic modelling: Most studies rely on static LCA methods that cannot capture time-dependent sequestration, delayed emissions at end-of-life, or regional differences in forestry dynamics. While recent research modelled time-adjusted GWP factors [192], end-of-life carbon release [193], and international carbon flows [194], dynamic models remain underutilised. A broader application of dynamic LCA is needed to enhance long-term assessments.

(3) Lack of transparent EPDs: Many timber EPDs selectively report biogenic storage while omitting fossil-based emissions, leading to incomplete carbon footprint profiles. All EPDs reviewed in this study were static and varied in scope and data quality. Future work should develop dynamic EPDs and standardised reporting requirements that transparently disclose both fossil and biogenic emissions, especially for internationally traded products [191].

(4) Regional bias and supply chain gaps: The reviewed literature is heavily concentrated in Europe and North America. Many studies lack the ability to capture geographic differences in forestry practices, energy sources, and transport logistics, despite their strong influence on CF results [21]. Region-specific emission factors and improved tracking of carbon flows along global timber supply chains are needed.

(5) Engineering applicability: The benchmark CF values and mitigation matrix developed in this study provide practical guidance for carbon-informed procurement and early-stage design. These tools are particularly useful for architects, engineers, and sustainability consultants applying LCA-based decision-making. However, they are not suitable for evaluating structural performance or material durability. Future research should integrate environmental modelling with structural analysis tools.

(6) Underexplored reuse and circular strategies: Reusing reclaimed timber can reduce global warming potential by up to 92% compared to virgin materials [191], yet most LCAs still focus on first-use timber. Regulatory, financial, and informational barriers continue to limit adoption. Research should assess the carbon performance of reuse scenarios, explore enabling policy mechanisms, and evaluate modular design strategies that promote material recovery and disassembly [195].

Finally, a formal correlation analysis (e.g., Pearson or Spearman) could not be performed due to insufficient metadata across the reviewed studies. This limits the ability to quantify statistical relationships between carbon outcomes and methodological choices. Future studies should improve metadata reporting to support quantitative analysis and enhance cross-study comparability.

6. Conclusions

This study systematically reviewed 63 studies on carbon accounting for timber buildings and EWPs, focusing on emissions during the production stage (A1–A3). Based on this synthesis, the following key conclusions are drawn:

(1) Significant variability exists in CF results, with values ranging from −40 to 1050 kg CO₂eq m⁻² for buildings and 12 to 759 kg CO₂eq m⁻³ for EWPs. These discrepancies are driven by inconsistent system boundaries, functional units, emission factors, and regional data sources.

(2) We establish benchmark CF baselines—247 kg CO₂eq m⁻² for timber buildings and 208 kg CO₂eq m⁻³ for EWPs—derived from our systematic review. These benchmarks serve as reference models for future comparisons, procurement criteria, and decarbonisation targets.

(3) An innovative mitigation matrix was developed, identifying six intervention strategies with quantified emission reduction potentials (e.g., air drying timber reduces emissions by up to 75%, renewable energy use up to 76%, and reclaimed timber up to 92%). Scenario combinations could reduce EWPs CF from 208 kg to below 130 kg CO₂eq m⁻³.

(4): Key research gaps such as limited adoption of dynamic models, inconsistent treatment of biogenic carbon, and poor regional coverage remain. Existing benchmarks are biased toward Europe and North America. Future studies should prioritise standardised, transparent approaches and regionally inclusive data to enable globally relevant and equitable decarbonisation strategies.