Food Waste Valorization Pathways Under Context Dependence: A Parameterized Life Cycle Assessment of Composting, Anaerobic Digestion, Animal Feed, and Biobased Chemicals

Abstract

Food waste valorization is central to the circular bioeconomy, yet environmental “best options” remain contingent on local energy systems, nutrient substitution opportunities, regulations, and waste composition. This study develops a parameterized life cycle assessment (LCA) comparing four competing food waste pathways: (i) aerobic composting with land application, (ii) anaerobic digestion (AD) with combined heat and power (CHP) and digestate land application, (iii) conversion to animal feed via hygienization and drying, and (iv) biochemical conversion to platform chemicals (modeled as a lactic-acid-centered biorefinery with residuals sent to AD). Using a functional unit of 1 tonne of source-separated food waste at the treatment facility gate, we implement an attributional LCA with system expansion for avoided products (electricity, heat, mineral fertilizer, conventional feed, and fossil-based chemical production). Impacts are characterized with ReCiPe 2016 midpoints and include climate change, acidification, eutrophication, photochemical ozone formation, particulate matter formation, fossil resource scarcity, and water use (Huijbregts et al., 2017). We evaluate context dependence through scenario analysis (grid carbon intensity, heat utilization, transport distance, fertilizer displacement ratios, feed substitution identity, and waste composition) and quantify uncertainty through Monte Carlo sampling of key parameters.

Across scenarios, no single pathway dominates all impact categories. AD-CHP tends to perform well for climate change where high-carbon electricity or effective heat use exists, while animal feed is frequently the lowest-impact option for land use and climate change when it credibly displaces soy-intensive feed rations. Composting is robustly beneficial for nutrient cycling but can be climate-disadvantaged when process methane and nitrous oxide are not well controlled. Biochemical conversion can outperform AD and composting for fossil resource scarcity when chemical substitution is credible, but it is sensitive to enzyme/utility burdens and co-product handling. We synthesize results into a decision framework identifying “break-even” conditions under which pathway rankings switch, providing a sustainability assessment approach that is explicitly local and waste-type aware.

Introduction

Food waste as an energy-and-resources challenge

Food waste is a persistent socio-technical and environmental problem with implications for climate forcing, nutrient pollution, and land and water appropriation. Global estimates indicate that a substantial share of food produced is not ultimately consumed, translating to avoidable upstream burdens from agriculture, processing, refrigeration, and transport (Gustavsson et al., 2011). Recent synthesis also emphasizes the scale of household food waste and the need for credible measurement and interventions (United Nations Environment Programme [UNEP], 2021). While prevention is typically the highest priority in waste hierarchies, large quantities of unavoidable food waste remain, motivating downstream valorization aligned with circular bioeconomy strategies.

Competing valorization pathways and why “best” is not universal

Food waste can be composted, anaerobically digested to biogas, rendered into animal feed, or converted into biobased chemicals and materials. Each route offers different coproducts—organic amendments, renewable energy carriers, feed nutrients, or chemical intermediates—implying that environmental performance depends on what is displaced in the receiving markets and on local infrastructure (e.g., district heating availability, grid carbon intensity, and nutrient management constraints). These context dependencies are central to interpreting sustainability assessment outcomes and are often the reason why studies reach different conclusions even when analyzing similar technologies.

Methodologically, the question “which option is best?” requires careful definition of function, boundaries, and multi-functionality handling in LCA (ISO, 2006a, 2006b). Food waste management also involves biogenic carbon and nitrogen flows that can drive nontrivial tradeoffs across climate change, acidification, and eutrophication. Composting and digestate spreading can reduce mineral fertilizer production but also lead to ammonia volatilization and nitrous oxide formation; energy recovery can reduce fossil combustion but may be less valuable in a decarbonized grid; and animal feed substitution depends on nutrition, safety regulations, and which feed ingredients are actually displaced.

Prior LCA evidence and remaining gaps

Comparative LCAs of food waste management have expanded over the past decade, with recurring themes: (i) AD often outperforms composting for climate change when electricity and/or heat displace fossil energy, but results depend strongly on energy system assumptions and methane leakage (Bernstad & la Cour Jansen, 2012); (ii) composting can be competitive when energy credits are weak and when nutrient substitution is modeled conservatively and emissions are well controlled; and (iii) using food waste as animal feed can provide substantial benefits by avoiding the impacts of conventional feed production, though results depend on processing energy and which feeds are displaced (Salemdeeb et al., 2017). The broader waste hierarchy framing emphasizes that valorization is not purely a technical ranking but a governance and infrastructure question (Papargyropoulou et al., 2014).

Meanwhile, interest in biochemical conversion of food waste into platform chemicals has grown as part of biorefinery development and circular bioeconomy policy (Cherubini, 2010; Lin et al., 2013). However, LCA comparisons that place chemical pathways alongside composting, AD, and animal feed under a consistent functional unit and harmonized multi-functionality assumptions remain relatively limited and, importantly, highly sensitive to substitution choices and process yields.

Research objectives and contributions

This original research develops a parameterized LCA model to compare four food waste valorization pathways under a harmonized scope and to identify the conditions under which pathway rankings change. The study contributes:

• A harmonized, parameterized LCA framework spanning energy, nutrient, feed, and chemical substitution under consistent boundaries.
• Quantified context dependence through scenario and sensitivity analysis for grid carbon intensity, heat utilization, transport, fertilizer displacement, and feed substitution identity.
• A switching-conditions (“break-even”) interpretation that translates LCA results into actionable decision logic for local planners and researchers.

Methodology

Goal and scope definition

The goal is to quantify and compare potential environmental impacts of alternative food waste valorization pathways and to evaluate how results depend on local conditions and waste characteristics. The intended audience is researchers and decision-support practitioners in Energy & Resources, waste management, and sustainability assessment.

Functional unit

The functional unit (FU) is 1 tonne (1,000 kg) of source-separated food waste delivered to the treatment facility gate, expressed on a wet weight basis. A wet-weight FU aligns with municipal decision-making and collection logistics but requires explicit attention to moisture variability in sensitivity analyses.

System boundary and general modeling approach

The system boundary is cradle-to-grave for the treatment system, including:

• Collection from source (modeled as incremental collection and transport to facility)
• Pre-treatment (e.g., depackaging where applicable)
• Core conversion process (composting, AD, feed production, biochemical conversion)
• Management/use of outputs (land application of compost/digestate, energy export, feed use, chemical product substitution)
• Direct air emissions (e.g., methane, nitrous oxide, ammonia)

Upstream burdens for background processes (electricity generation, fuels, enzymes/chemicals, fertilizer production, conventional feed, and conventional chemical production) are represented using generic life cycle inventory data consistent with the ecoinvent system model (Wernet et al., 2016). The analysis is primarily attributional with system expansion to address multi-functionality, consistent with ISO recommendations when avoiding allocation is feasible (ISO, 2006b). A consequential interpretation is discussed qualitatively where market-mediated effects are likely material (e.g., feed displacement and heat export).

Studied pathways

Pathway A: Aerobic composting with land application

Food waste is composted in an enclosed, aerated system with biofiltration. Outputs include compost applied to agricultural land. Credits are given for displaced mineral fertilizers (N, P, K) to the extent that compost nutrients are plant-available and substitute synthetic fertilizers.

Pathway B: Anaerobic digestion with CHP and digestate land application

Food waste is treated in wet, mesophilic AD, producing biogas used in CHP to export electricity and (where feasible) useful heat. Digestate is separated (liquid/solid) and land applied, generating fertilizer substitution credits and field emissions (NH ₃ , N ₂ O, NO ₃ ^- leaching) consistent with standard nutrient management modeling assumptions.

Pathway C: Conversion to animal feed (hygienization + drying)

Food waste is hygienized (thermal treatment) and dried/pelletized into a feed ingredient. The resulting product is assumed to replace a marginal conventional feed ration on a metabolizable energy and protein basis. This approach reflects the central sustainability claim: avoided impacts from conventional feed production (notably soy and cereals) can dominate outcomes (Salemdeeb et al., 2017). Regulatory feasibility and adoption constraints are treated as contextual limitations rather than universal assumptions.

Pathway D: Biochemical conversion to platform chemicals (lactic acid-centered biorefinery)

Food waste is processed through sorting/pulping, hydrolysis, and fermentation to produce lactic acid as a representative platform chemical for the bioeconomy (Lin et al., 2013). Residual solids (non-fermentable fraction) are sent to AD for biogas production; wastewater is treated. Credits are given for displaced conventional lactic acid production (modeled as a generic industrial product) and for exported energy from residual AD.

Life cycle inventory (LCI): key parameters

Because real facilities vary widely, we use a parameterized foreground inventory . Point values define a base case, and probability distributions support uncertainty analysis. Parameters are drawn from peer-reviewed syntheses where possible, with additional engineering assumptions for the biochemical pathway reflecting typical fermentation/hydrolysis energy requirements (Lin et al., 2013) and generic industrial practice. Table 1 summarizes the main parameters used to generate comparative results.

Multi-functionality and avoided burdens

All pathways are multifunctional. We apply system expansion (avoided burden) as the base-case approach, crediting the system for displacing marginal production of substituted goods. The net impact in category c is computed as:

(1)

where includes direct process and field emissions, includes upstream burdens (electricity, fuels, chemicals), is the quantity of coproduct j exported/used, and is the avoided impact factor per unit of displaced product j . Equation (1) highlights why local context—particularly for electricity and heat—can reverse rankings.

Energy substitution

Exported electricity displaces grid electricity with an intensity set by scenario. Exported heat displaces marginal heat (modeled as natural gas boiler heat in the base case). The avoided climate impact from electricity export is approximated by:

(2)

where is net exported electricity (kWh/FU) and is the grid emission factor (kg CO ₂ e/kWh). We explicitly vary to represent decarbonizing systems.

Nutrient substitution

Compost/digestate nutrient credits depend on nutrient content and plant-availability. We model effective mineral fertilizer substitution as:

(3)

where is the nitrogen substitution efficiency (Table 1). We apply analogous credit logic for P and K with higher substitution efficiencies, while also including field emissions (NH ₃ volatilization, N ₂ O formation, and leaching) because organic amendments can shift emission pathways rather than eliminate them.

Feed substitution

For the feed pathway, we credit avoided production of conventional feed ingredients. Because actual substitution is uncertain, we report scenario results for soy-intensive and cereal-intensive displacement. The importance of this assumption is widely noted: impacts of feed production can be substantial and heterogeneous across ingredients and geographies (Poore & Nemecek, 2018).

Chemical substitution

The biochemical pathway is credited with avoiding production of a functionally equivalent quantity of conventional lactic acid. Because industrial production routes vary (e.g., fermentation using refined sugars vs alternative feedstocks), we treat this avoided process as a parameter with sensitivity bounds rather than a single universal value.

Impact assessment method and categories

We use ReCiPe 2016 midpoint indicators (hierarchist perspective) for the following categories: climate change, terrestrial acidification, freshwater eutrophication, marine eutrophication, photochemical ozone formation, particulate matter formation, fossil resource scarcity, and water consumption (Huijbregts et al., 2017). These categories capture salient tradeoffs for food waste management: greenhouse gas emissions, nitrogen-related impacts, air pollution precursors, and fossil energy substitution.

Scenario design: capturing local energy and infrastructure dependence

We evaluate four scenario families:

1. Electricity system carbon intensity : low-carbon (50 g CO ₂ e/kWh), medium (300 g), high (800 g).
2. Heat utilization (AD and biochemical residual AD): 0%, 50%, 90% useful heat export.
3. Transport/infrastructure : short-haul (10 km), typical (40 km), long-haul (120 km).
4. Substitution identities : soy-intensive vs cereal-intensive feed displacement; high vs low fertilizer substitution efficiency.

Uncertainty analysis

We implement Monte Carlo simulation by sampling uncertain parameters (Table 1) and recomputing net impacts via Equation (1). The analysis reports medians and interquartile ranges (IQR) and estimates the probability that each pathway is the lowest-impact option per category. This approach is consistent with established LCA practice when foreground parameters are uncertain and when comparative claims depend on overlapping distributions (ISO, 2006b).

Data quality, transparency, and limitations of inference

This study is a modeling analysis rather than an empirical measurement campaign. Foreground parameter ranges represent plausible operational variability; they are not a substitute for plant-specific monitoring data. Background system datasets depend on the chosen database and system model (Wernet et al., 2016). We therefore emphasize switching conditions and qualitative robustness rather than claiming universal numerical rankings.

Results

Base-case comparison (medium-carbon grid; moderate heat utilization)

Table 2 summarizes indicative midpoint results per FU for the base-case context: grid intensity 300 g CO ₂ e/kWh and 50% CHP heat utilization for AD and residual AD in the biochemical pathway. Values are reported as net impacts (burdens minus avoided credits). Because this is a parameterized model, results are presented as rounded medians with uncertainty ranges referenced in subsequent subsections.

Three patterns emerge. First, climate change benefits are largest for animal feed and AD in the base case, driven by avoided conventional feed production and avoided grid electricity plus fossil heat, respectively. Second, acidification and eutrophication remain positive (burdens) for all pathways , reflecting the difficulty of fully avoiding nitrogen losses during biological processing and land application; credits for avoided mineral fertilizers only partially offset these emissions. Third, the biochemical pathway produces moderate climate benefits but shows higher water and upstream utility sensitivity, consistent with the additional processing steps typical of biorefineries (Cherubini, 2010; Lin et al., 2013).

Contribution analysis: what drives each pathway

To interpret Table 2, we decompose climate change impacts into dominant contributions:

• Composting : net climate impacts are dominated by direct CH ₄ and N ₂ O emissions during composting and, secondarily, by field N ₂ O . Fertilizer credits are material but often insufficient to overcome process emissions when control is suboptimal.
• AD-CHP : the largest credit comes from electricity export , followed by useful heat and fertilizer substitution via digestate. Methane slip and digestate emissions are the main countervailing burdens.
• Animal feed : benefits are dominated by avoided feed ingredient production . The largest burdens come from drying energy and, depending on energy source, associated air emissions. This result structure matches the central conclusion of prior assessments: if feed substitution is real, it can outweigh processing energy (Salemdeeb et al., 2017).
• Biochemical conversion : benefits depend on avoided conventional lactic acid and residual energy recovery , while burdens are driven by electricity/steam demands , enzymes/chemicals (where used), and wastewater treatment.

Scenario results: grid decarbonization and heat utilization

Electricity and heat substitution is the key mechanism favoring AD and residual AD in the biochemical pathway. Figure 3 summarizes how AD’s climate benefit changes with grid carbon intensity and heat utilization. As expected from Eq. (2), lower reduces avoided electricity credits, narrowing the gap between energy-focused pathways and non-energy pathways.

In high-carbon grids, AD-CHP remains strongly beneficial even with low heat utilization. In low-carbon grids, AD climate performance becomes much more sensitive to whether heat can be exported and to methane slip controls. This finding aligns with broader energy system logic: as electricity decarbonizes, the relative value of electricity-generating waste-to-energy declines, and nutrient or material substitution pathways may become comparatively more attractive (Bernstad & la Cour Jansen, 2012).

Scenario results: transport and infrastructure

Transport affects all pathways similarly but can become decisive when facilities are distant or when collection requires dedicated trips. Long-haul transport shifts results most for composting and AD because their credits are smaller per tonne than soy-intensive feed credits in many contexts. Under a 120 km transport scenario, climate benefits for AD and biochemical conversion decrease, while animal feed often remains net-beneficial unless drying energy is also high-carbon. This highlights a practical planning insight: infrastructure siting and route optimization can matter as much as process choice for marginal decisions, particularly in regions with dispersed waste generation.

Scenario results: feed displacement identity

Animal feed results are highly contingent on what ingredients are displaced. Displacing soy meal tends to provide larger avoided burdens than displacing cereals because soy supply chains can have higher land-use and associated emissions burdens, although these vary by geography and production practices (Poore & Nemecek, 2018). When the displaced ration is cereal-dominant and processing energy is fossil-intensive, the feed pathway’s advantage can shrink or, in extreme cases, reverse in climate change or particulate matter categories. This finding reiterates a core caution from the literature: feed valorization is environmentally promising, but only when substitution is nutritionally and economically credible and when processing is efficient (Salemdeeb et al., 2017).

Scenario results: waste composition effects (carbohydrate-rich vs mixed)

Waste composition affects:

• AD yields (higher readily degradable carbohydrates typically increase methane yield)
• Composting emissions (moisture and structure influence anaerobic pockets and CH ₄ )
• Feed suitability (salt, contamination risk, and variability affect feasible inclusion rates)
• Biochemical yields (fermentable carbohydrate fraction determines lactic acid yield)

When food waste is modeled as carbohydrate-rich (e.g., bakery/produce streams), the biochemical pathway improves due to higher lactic acid yield and lower residual fraction, while AD also improves due to higher methane yield. For protein- and lipid-rich streams (e.g., kitchen waste with meat/dairy), AD tends to improve relative to biochemical conversion (if the biochemical system is carbohydrate-targeted), whereas feed suitability may be more constrained by regulations and food safety considerations—constraints not captured by LCA but essential for feasibility.

Uncertainty analysis: probability of being best by impact category

Monte Carlo results reinforce that “best” depends on the impact category and local assumptions. In the base-case context, animal feed has the highest probability of being lowest-impact for climate change and land-related indicators (not fully reported in Table 2 but consistent with Poore and Nemecek’s (2018) finding that animal product supply chains drive large land burdens). AD-CHP often has the highest probability of being best for photochemical ozone formation and fossil resource scarcity when the grid is medium-to-high carbon and heat use is feasible. Composting is rarely lowest-impact for climate change under median assumptions but becomes more competitive under low methane/N ₂ O emissions and high fertilizer substitution efficiency.

Discussion

Interpretation: why rankings switch across contexts

The central explanatory mechanism is the relative magnitude of credits versus direct emissions in Equation (1). Pathways differ in (i) the size and certainty of their coproduct credits, and (ii) the extent to which they generate difficult-to-control biogenic emissions (CH ₄ , N ₂ O, NH ₃ ). These differences produce switching behavior:

• AD-CHP switches from strongly beneficial to modestly beneficial (or even near-neutral) for climate change as the grid decarbonizes and if heat cannot be utilized. This is a predictable outcome of declining in Eq. (2).
• Composting switches from beneficial to burdensome for climate change primarily as a function of methane and nitrous oxide control. Even small amounts of CH ₄ and N ₂ O can overwhelm fertilizer credits due to high global warming potentials (IPCC, 2021).
• Animal feed switches depending on whether high-impact ingredients (e.g., soy meal) are displaced and whether drying/hygienization energy is low-carbon. This is a substitution credibility issue, not merely a process issue.
• Biochemical conversion switches based on chemical yield and on whether the avoided chemical production is genuinely displaced in the market. Its additional processing steps create higher exposure to electricity/steam assumptions, making it more sensitive to decarbonization trajectories than composting or feed in some categories.

Alignment and tension with the food waste hierarchy

The food waste hierarchy emphasizes prevention first, then redistribution, then animal feed, then materials/energy recovery, and finally disposal (Papargyropoulou et al., 2014). Our results do not contradict this logic but refine it: feed often performs best environmentally under credible substitution , yet its feasibility is constrained by regulation, contamination risk, and logistical variability. Where feed is not feasible, AD and biochemical pathways can be strong climate options depending on energy context. Composting remains an important route for soil amendment and nutrient cycling but requires stringent emission control to be climate-competitive.

Implications for the circular bioeconomy

From a bioeconomy perspective, biochemical conversion to platform chemicals is attractive because it targets higher-value outputs than energy alone (Cherubini, 2010; Lin et al., 2013). However, “higher value” does not automatically mean “lower impact.” The biochemical pathway’s environmental performance depends on achieving high yields, minimizing utility demand, and ensuring that the chemical product meaningfully displaces incumbent production. In a future where electricity is low-carbon, biochemical pathways may become relatively more attractive compared to electricity-exporting AD, but only if upstream burdens (enzymes, nutrients, purification energy) are also decarbonized and controlled.

A decision framework: identifying break-even conditions

To translate results into actionable guidance, we propose a break-even decision framework based on three dominant contextual axes:

• Energy context : grid carbon intensity and availability of heat sinks (district heating, industrial users)
• Nutrient context : credible land application outlets, fertilizer displacement efficiency, and nutrient management constraints
• Substitution context : credibility of feed or chemical displacement (market uptake, regulations, quality control)

This framework is not a replacement for site-specific LCA. Instead, it clarifies which variables most strongly determine outcomes and therefore what data should be prioritized for local sustainability assessment.

Methodological reflections: attributional vs consequential interpretation

Although this study uses attributional LCA with system expansion, some questions are inherently consequential. For example, “What feed is displaced?” depends on market response, regulatory constraints, and farm ration formulation; similarly, “What electricity is displaced?” depends on marginal generation. ISO standards acknowledge that the choice of modeling approach should follow the decision context (ISO, 2006a, 2006b). We therefore recommend that studies intended to inform policy instruments (e.g., credits, mandates) complement attributional results with consequential sensitivity analyses, especially where markets are elastic or substitution claims are central.

Limitations and sources of uncertainty

• Foreground emission factors : Methane and nitrous oxide emissions from composting and digestate/compost land application are uncertain and site-dependent; improved monitoring would materially improve LCAs (Bernstad & la Cour Jansen, 2012).
• Substitution assumptions : Feed and chemical substitution dominate results when credited. Empirical evidence on displacement ratios is often limited, and adoption constraints can be binding (Salemdeeb et al., 2017).
• Database dependence : Background results depend on LCI database choices and system models (Wernet et al., 2016).
• Scope exclusions : This study does not model food waste prevention or redistribution, which can be higher priority options in the hierarchy (Papargyropoulou et al., 2014). Nor does it include potential soil carbon sequestration benefits from compost/digestate due to high uncertainty and site specificity.

Research needs

Three research directions would substantially strengthen comparative sustainability assessment of food waste valorization:

• Empirical substitution evidence for feed and biochemical products, including market-mediated displacement and quality constraints.
• Improved emissions measurement for composting and digestate/compost land application under diverse management conditions.
• Integrated energy–nutrient modeling that aligns waste infrastructure planning with grid decarbonization pathways and agricultural nutrient management.

Conclusion

This study compared four food waste valorization pathways—composting, anaerobic digestion with CHP, animal feed production, and biochemical conversion to platform chemicals—using a harmonized, parameterized LCA with system expansion. The results support four conclusions relevant to Energy & Resources researchers and practitioners:

• No universal best option exists. Rankings shift across impact categories and local contexts, particularly with changes in grid carbon intensity and heat utilization.
• Animal feed is often environmentally favorable when it credibly displaces high-impact conventional feed ingredients and when processing energy is efficient and low-carbon; however, feasibility constraints are decisive.
• AD-CHP is most attractive in high-carbon energy systems and where heat export is feasible; its advantage narrows as grids decarbonize or if methane slip and digestate emissions are poorly controlled.
• Composting is robust for nutrient cycling but climate-sensitive to CH ₄ /N ₂ O control and realistic fertilizer substitution efficiencies; biochemical conversion is promising for the bioeconomy but highly sensitive to yields and utility burdens.

As a practical outcome, we propose interpreting food waste LCA through “switching conditions” that identify when a pathway becomes preferable rather than relying on a single static ranking. This approach better matches the reality of heterogeneous waste streams, evolving energy systems, and infrastructure constraints, and it provides a more decision-relevant sustainability assessment for the circular bioeconomy.

References