Use of LCA for assessing the deployment of anaerobic digestion (AD) to treat different food waste streams in the UK

A case study is presented below, in which consequential LCA is applied to evaluate the net environmental change associated with the deployment of anaerobic digestion (AD) to treat different food waste streams, replacing three existing waste management options: (i) landfilling; (ii) in-vessel composting; (iii) animal feeding. More detail on this is provided in Styles et al. (2016). 

Scope and boundary definition

ISO 14040 and ISO 14044 (ISO, 2006a, 2006b) describe the framework for LCA application, according to four main phases:

1. Goal, scope and boundary definition;
2. Inventory compilation;
3. Life-cycle impact assessment;
4. Interpretation and reporting.

Getting the first phase correct is critical and represents a challenge when considering waste management alternatives. In the first instance, the correct LCA approach must be identified. Extensive guidelines produced for product carbon footprinting (e.g. BSI, 2011; Commission Recommendation 2013) or organisation carbon footprinting (WRI, 2004, 2011a; Commission Recommendation 2013) provide a detailed methodological basis and guidance to perform LCA of products and organisations.

Typically, two main LCA approaches are used: an ‘attributional’ one (intended to provide a static representation of average conditions, excluding market-mediated effects) and a ‘consequential’ one, which strives to identify the consequences associated with the change applied to the product/service system (Weidema et al., 2003, 2009).

Figure 1 provides an example of the boundaries and main processes considered for a generic LCA of organic waste treatment by anaerobic digestion (A) or incineration (B). Additionally to fulfilling the ‘main service/function’ (i.e. the treatment of the organic waste in input), the digestion system generates two valuable products: biogas and digestate (organic bio-fertiliser). Similarly, incineration provides energy and aggregate-type material (bottom ash). Owing to these, the system under assessment (both A and B) becomes multifunctional as co-products are generated along with the fulfilment of the main waste system service/function (i.e. the mere treatment of the organic waste input). To address multifunctionality, two options are available (see Commission Recommendation 2013): i) subdivision or system expansion, ii) allocation (mass, energy, or price). Allocation principles, sometimes used in attributional studies, should only be applied when system expansion (or subdivision) is not possible conforming with the best practices described by Commission Recommendation (2013).

Keeping the above in mind, the practitioner, depending upon the scope of the study and following recommendations from the guidelines mentioned earlier, may in this specific case apply: i) system expansion considering average market processes (this is also referred to as an ‘attributional approach’, which applies system expansion using average market data), ii) system expansion considering marginal market processes (also called a ‘consequential approach’), or iii) allocation principles (i.e. no system expansion; this is also referred to as an ‘attributional approach’).

  In the anaerobic digestion example below, expanding the system would mean including energy generation and fertiliser manufacture/application which is displaced by bio-electricity and bio-fertiliser (digestate) application, respectively. In the case of incineration, in addition to the displaced energy, the LCA practitioner should account for the natural aggregates extraction, transport and processing that are avoided when bottom ash is recycled. Then, results may be expressed as comparisons between scenarios (e.g. A versus B below) or as environmental burden changes expected from a particular change of strategy, or the introduction of a new system (e.g. going from A to B or vice versa) – as appropriate to inform waste management strategy from a wider public good perspective. It is important to bear in mind that the magnitude of these changes is highly dependent on the 'displaced processes' (e.g. type of fertilisers, electricity). Consequential LCA should be based on predicted marginal effects, rather than average effects: e.g. the question to be asked could be: “what type of electricity generation is replaced by new bio-electricity fed into the grid from biogas generation”?

Figure 1. An example of the LCA system boundary for the comparison of two alternative management scenarios for wet organic waste: A) anaerobic digestion and B) incineration^[1].

Finally, the environmental scope of LCA may be expanded to consider flows of money (life-cycle costing) and social capital (social life-cycle assessment). The United Nations Environment Programme provides guidelines on how to undertake social LCA (UNEP, 2009).     

Once the LCA and system boundaries have been defined, the impact categories to be considered must be decided – see the section on Life-cycle impact assessment indicators, below.

In the AD case study referred to under “Description”, boundaries were defined to include waste collection and transport, processing through the AD plant, digestate application including fertiliser replacement, biomethane upgrade and replacement of transport diesel, and also avoidance of pre-existing waste management options (landfilling, in-vessel composting and animal feeding – in the latter case avoided cultivation of wheat as an animal feed).  

Inventory compilation

Inventory compilation is the second phase of LCA, in which data on activities and associated inputs, outputs and burdens are compiled for the system of study (e.g. AD system or in-vessel composting system). The International Reference Life-Cycle Data System (ILCD) provides a common basis for consistent, robust and quality-assured life-cycle data, methods and assessments (JRC, 2011), and hosts the European Platform on LCA (http://eplca.jrc.ec.europa.eu/) – an open-access life-cycle inventory database. Various commercial LCA databases also exist, such as Ecoinvent (http://www.ecoinvent.org/), that contain extensive data on common generic processes. Often, it is possible to simply multiply system-specific activity data (e.g. tonne-km of transport) with unit process data from LCA databases (e.g. environmental burdens, such as kg CO₂e, per tonne-km transport in a EURO V compliant 16-32 tonne truck) to generate burdens for particular processes, stages, and ultimately entire systems. In other cases, it may be necessary to use process-specific data to calculate burdens (e.g. measured or calculated methane leakage rates from fermentation, digestate storage and biomethane upgrade). For example, in the case of digestate and compost application to land, Bruun et al. (2006) propose long-term (100-year) soil organic carbon sequestration credit (a CO₂e “credit”) equivalent to 13 % and 14 % of the organic C contained in digestates and composts, respectively. These values were used by Møller et al. (2009) to evaluate the life-cycle environmental performance of anaerobic digestion.

Owing to the number of actors involved in a typical product life cycle, or waste stream flow, it will often be necessary to obtain activity data from other organisations in order to complete an LCA. Care should be taken to evaluate the quality (accuracy and validity of the data) during data collation, so that appropriate uncertainty analyses and sensitivity analyses may be undertaken to facilitate interpretation. Data may be tagged as low, medium, or high uncertainty for example, or statistical distributions (e.g. 95 % confidence intervals) may be recorded.

Inventory data compiled for the AD case study example included:

• diesel consumption for transport of waste to the digester, calculated based on distance transported multiplied by burdens expressed per tonne-km in the Ecoinvent database;
• fugitive emissions of methane from the digester, from digestate storage and from biomethane upgrade, estimated from emission factors of 1 %, 1.5 % and 1.4 % of total biomethane yields, respectively;
• ammonia emissions from digestate storage, estimated from an ammonia-N emission factor of 10 % of ammonium-N in digestate;
• transport diesel fuel replaced calculated based on a biomethane yield of 440 m³ per tonne of dry matter (food waste), a methane lower heating value of 34 MJ per m³, 20 % of biomethane used on site to generate process heat and electricity, and a substitution efficiency of 1 MJ biomethane per 0.75 MJ diesel.

The above list is far from exhaustive, as it excludes, for example, diesel combustion, nutrient losses and fertiliser replacement incurred by digestate application.

Life-cycle impact assessment (LCIA)

Life-cycle impact assessment (LCIA) involves the characterisation of inputs and emissions according to their environmental damage potential, using factors derived from extensive fate and transport modelling (e.g. Huijbregts et al., 2001), thus synthesising inventories of inputs and outputs into a small number of environmental indicators representing key environmental burdens (Pennington et al., 2004).

LCIA involves the multiplication of inputs and outputs by relevant characterisation factors to represent contributions towards environmental burdens or impacts. LCIA is typically performed across three areas of protection: human health, natural environment, and natural resource use, and may include the following impact categories (JRC, 2011): climate change, ozone depletion, eutrophication, acidification, human toxicity (cancer- and non-cancer-related), respiratory inorganics, ionising radiation, ecotoxicity, photochemical ozone formation, land use, and resource depletion (materials, energy, water).

Table 1 summarises LCIA methods recommended for the International Reference Life-Cycle Data System (JRC, 2011).

Table 1. Midpoint life-cycle impact assessment methods proposed by JRC (2011) for the harmonisation of methods in the International Reference Life-Cycle Data System

Method	Flow property	Reference unit
Global warming potential, GWP100	Mass CO2 equivalents	Units of mass (kg)
Ozone depletion potential, ODP	Mass CFC-11 equivalents	Units of mass (kg)
Cancer human health effects, CTUh	Comparative Toxic Unit for humans (CTUh)	Units of items (cases)
Non-cancer human health effects, CTUh	Comparative Toxic Unit for humans (CTUh)	Units of items (cases)
Respiratory inorganics, PM2.5 equivalents	Mass PM2.5 equivalents	Units of mass (kg)
Ionising radiation, ionising radiation potential	Mass U235 equivalents	Units of mass (kg)
Photochemical ozone formation potential, POCP	Mass C2H4 equivalents	Units of mass (kg)
Acidification, accumulated exceedance	Mole H+ equivalents	Units of mole
Eutrophication terrestrial, accumulated exceedance	Mole N equivalents	Units of mole
Eutrophication fresh water, P equivalents	Mass P equivalents	Units of mass (kg)
Eutrophication marine, N equivalents	Mass N equivalents	Units of mass (kg)
Ecotoxicity fresh water, CTUe	Comparative Toxic Unit for ecosystems (CTUe) * volume * time	Units of volume*time (m3*a)
Land use, soil organic matter	Mass deficit of soil organic carbon	Units of mass (kg)
Resource depletion – water, fresh water scarcity	Water consumption equivalent	Units of volume (m3)
Resource depletion – mineral, fossils and renewables, abiotic resource depletion	Mass Sb equivalents	Units of mass (kg)
Source: JRC (2011).

 

Indicator results may be normalised (divided by “total” environmental loadings at a specified scale) to enable comparison of relative contributions across environmental impact categories. For example, Andersen et al. (2012) present LCIA indicator results normalised as milli-person equivalents (contributions to annual per capita loadings, divided by 1 000).

In Figure 2, burden data for a partial-expanded-boundary LCA of one tonne of organic waste treated by decentralised composting are presented after normalisation against average European citizen per capita loadings. Positive values indicate an adverse impact on the environment, whilst negative values indicate environmental savings compared with the alternative of separate waste collection (though the alternative waste management option is not accounted for in this particular partial LCA). Emissions of nitrous oxide and methane during composting give rise to a significant GWP burden, soil emissions of ammonia following application give rise to a significant AP effect, and replacement of fertilisers with organic nutrients following field application leads to significant EP, AP and FRDP savings (Figure 2).

Figure 2. Results for global warming potential (GWP), eutrophication potential (EP), acidification potential (AP) and fossil resource depletion potential (FRDP) for decentralised composting of household organic waste  

 

A full consequential LCA would account for burdens and savings associated with alternative (replaced) waste management option(s), such as centralised composting, anaerobic digestion or MSW incineration. Results for the consequential LCA of the AD case study are displayed in the next section, expressed using the same four environmental indicators used in Figure 2.

Following on from the characterisation of input and output data to generate environmental indicators, ISO 14040 (ISO, 2006a) defines three optional steps:

• Normalisation: Indicator values (e.g. kg PO₄e) are converted into environmental loadings relative to a reference value – often “total” loading at national, EU or global scale, or for example per capita.
• Grouping: The impact categories are sorted and possibly ranked.
• Weighting: The different environmental impacts are weighted relative to each other so that they can then be summed to get a single number for the total environmental impact.

These procedures may facilitate an understanding of the relative importance of nominal indicator values across impact categories, but weighting is not recommended in ISO 14040 owing to the introduction of value judgements. In converting nominal indicator units into comparable burden fractions, normalisation facilitates the comparison of contributions to different environmental problems and relative trade-offs.

 

Interpretation and reporting

According to ISO 14044 (ISO, 2006b), the interpretation phase of an LCA study comprises the following elements:

• identification of significant issues based on the findings (life-cycle inventory (LCI) and life-cycle impact assessment (LCIA) phases);
• an evaluation that considers completeness, sensitivity and consistency;
• conclusions, limitations, and recommendations.

It is useful to structure results from the LCI and LCIA phases according to life-cycle stages and processes to underpin contribution analysis that in turn facilitates presentation, interpretation, validation and anomaly assessment (ISO, 2006b).

Mass or energy balance analysis of all input and output data may also be applied to check for anomalies, according to the law of conservation of mass and energy. The influence of uncertainty on final results can be tested using sensitivity analysis (e.g. Clavreul et al., 2013). Uncertainties for individual process interventions can be aggregated up to the system level based on error propagation methods.

Where results of comparative studies are intended for public disclosure they should be critically evaluated by an appropriate expert or panel of interested parties, and the results of the evaluation disclosed, according to ISO 14044 (ISO, 2006b). The critical review process shall ensure that:

• methods used to carry out LCA are consistent with the ISO standard;
• methods used to carry out LCA are scientifically and technically valid;
• data used are appropriate and reasonable in relation to the goal of the study;
• interpretations reflect the limitations identified and the goal of the study;
• the study report is transparent and consistent.

With respect to reporting LCA results, the goal, scope and boundaries applied should be clearly reported.

Table 2 and Figure 3 below summarise the environmental changes that arise, expressed as credits (negative values) and burdens (positive values) across avoided and incurred processes (Figure 3), and expressed as net environmental burden change (Table 2), in relation to one tonne of food waste dry matter – from the AD consequential LCA case study. Avoided waste management and avoided fossil energy (transport diesel) give rise to substantial environmental credits (negative values) in most cases, indicating that AD performs better than avoided waste management options – apart from in the case of animal feed. In this respect, it should be noted that Styles et al. (2016) considered animal feeding as a particular food waste management option.

Where food factory waste can be used as animal feed for example, this avoids cultivation of wheat as an animal feed, and therefore generates significant environmental credits. These credits are no longer realised if waste is sent to AD rather than animal feed, and so become represented as a burden for AD (in Figure 6 see red "waste management" for the animal feed scenario). 

These results are unique to the precise scenarios and underlying operational assumptions for typical UK conditions defined in Styles et al. (2016). Undertaking consequential LCA is associated with a high degree of specificity in relation to the transitions considered (from which baseline to which option), and a high degree of uncertainty. Results should include a robust sensitivity analysis on the scenario uncertainties (e.g. choice of displaced 'marginal' technologies/processes) and on the parameters' uncertainties (e.g. efficiencies, transport distances). On this basis, results should therefore be interpreted cautiously and always in relation to the specific scenarios and context considered.

Figure 3. Net environmental burden changes, expressed per tonne of dry matter organic waste processed, when anaerobic digestion replaces landfilling, in-vessel composting or use of hygienic organic waste for animal feed  

 

Table 2. Net environmental burden changes, expressed per tonne of dry matter organic waste processed, when anaerobic digestion replaces landfilling, in-vessel composting or use of hygienic organic waste for animal feed  

	Landfill	Compost	Animal feed
Global warming (kg CO2e)	-2 640	-1 306	-74
Eutrophication (kg PO4e)	0.8	-1.8	8.4
Acidification (kg SO2e)	2.7	-2.7	8.4
Fossil resource depletion (MJe)	-6 516	-14 449	-9 492