Life Cycle Assessment in the Livestock and Derived Edible Products Sector

Abstract

The livestock production sector represents more than 40 % of the economic value of EU primary productions. This sector consists of a huge diversity of processes and techniques depending on the animal species and the final products. Because of these differences, livestock productions are associated with several adverse effects on the environment, especially in the breeding phases and feeding composition and management; moreover, in terms of raising awareness of the environmental implications of livestock productions, LCA applications are of increasing importance for systematic assessment of the environmental burdens connected with this sector. After an overview of the structural and economic characteristics of the most significant livestock supply chain and its main environmental problems, we provide a description of the international state of the art of LCA implementations for livestock. Methodological problems connected with the application of LCA are investigated, starting with the critical analysis of international papers and the few Italian papers in the scientific literature. Finally, the best practices regarding LCA methodology implementation are proposed, in order to improve results and manage the methodological problems identified.

Conclusions and Lessons Learned

This environmental impact assessment of the livestock sector presents some critical issues that may occur regardless of the methodologies used, but that in the case of LCA or LCA-based methodologies make the evaluation extremely complex. The complexity of approaches, data requirements, and model specifications has become so high that some standardisation is necessary to make things more credible and comparable. The variability in the results of all the livestock supply chains is caused by the difference in the production systems and methodological choices (functional unit, system boundaries, allocation method, etc.). For instance, in beef production systems a very wide variety, ranging from very intensive to very extensive (Nijdam et al. 2012), was observed; in dairy farms, which generally produce more than one product, the whole impact of dairy activities should be shared and allocated between all of them and the environmental performance of their processed products also depends on the use of milk supplied by different farms with different rearing systems. Of special interest in LCA analysis of the livestock production systems was the definition of the functional unit (FU), or, rather, the unit respecting which the environmental impacts are defined. The choice of a “corrected functional unit”, such as fat and protein or energy, could be an efficient approach which takes into account the nutritional value of livestock products and allows the comparison of the results of different studies. Livestock products differ in terms of production techniques and economic values, protein content and live weight (Nguyen et al. 2012a). Thus, the use of more complex FUs is mandatory for studies focussed on the evaluation of the environmental and economic impacts of the whole beef supply chain (Weidema et al. 2009) or comparison of different livestock sectors (beef, pigs, chickens, sheep and goats, etc.) (de Vries and de Boer 2010). The choice of FU is critical, as pointed out in Sect. 1.4, for studies addressing environmental impacts and load allocation of milk and meat in beef production (Cederberg and Stadig 2003). A common characteristic of all the analysed studies is the heterogeneity in system boundaries’ (SB) definition. Besides the variety and the complexity of livestock transformation processes, a relevant critical methodological point for LCA analysis, the inclusion of crop production (fodder especially) in rearing systems’ impact assessment is a critical and debated question. Meat and milk production systems are characterised, moreover, by a high number of co-products and by-products, let alone the production of both meat and milk. Almost all the studies reviewed consider the cradle to farm gate life cycle and exclude capital goods from the analysis. However, the environmental impact from capital goods has been included in some recent publications which found that capital goods contribute significantly to the total impact of agricultural production systems (Blengini and Busto 2009; Frischknecht et al. 2007). As regards different methods of impact assessment and classification there are several approaches (often IPCC 2007; EDIP; CML; CED; Impact 2000 + and Eco-indicator 99) that are choosen according to the goal and scope of the studies and their effectiveness in showing results. The phase with the greatest impact, in all the studies, is animal rearing; enteric CH₄, NH₃ and N from animal excreta are the major culprits responsible for environmental loads. The land use impact category is particularly relevant for beef and dairy production, which has the highest impact compared with other meat production systems (pigs and chickens). Data availability remains a long-standing problem and is hard to solve, as witnessed by scientific studies dedicated to system definition and inventory construction. A more complete picture of the environmental impacts (some of which have not been adequately addressed so far) and of the phases along the whole chain should be included as improvements for future research. A strong interaction between research experts and economic organisations (e.g. farmer’s associations) could make the LCA methodology useful in the decision-making process connected to the definition of an environmental chain strategy. This interaction is useful for many reasons: to support LCA data requirements, improving and expanding databases; to support the standardisation process and levels; to stress the main gaps in current knowledge on which future research and developments should be focussed.

From a methodological perspective, there are many studies oriented to the evaluation of environmental impacts; only a few of them (Weidema et al. 2009; Basarab et al. 2010; Van Middelaar et al. 2011; Oishi et al. 2013) combine LCA with the evaluation of economic impacts (e.g. life cycle costing, net present value, value added, etc.). A few studies combine LCA with farm simulation models (Beauchemin et al. 2010; Leip et al. 2010; Clarke et al. 2012), since they can give useful information for the improvement of rearing systems and related environmental impacts.

Only Weidema et al. (2009), using a hybrid methodology combining macroeconomic data (input/output tables) with the emissions generated by the analysed productive processes (Suh et al. 2004), evaluate the whole life cycle from cradle to fork. They include the transformation, marketing and use phases in the SB.

In LCA analysis of the livestock sector, economic allocation is the most frequently used method although the ISO standards recommend physical or biological criteria (carcass weight, protein content, etc.) in preference (Yan et al. 2011). Indeed, De Vries and de Boer (2010) and Nguyen et al. (2012a) used these allocation criteria to identify studies to include in their review of the environmental impacts of different livestock productions and to assess the environmental impacts of four beef farming systems. Moreover, economic allocation methodology does not account for the environmental benefit produced by the milk system with the reduction of biological methane and ammonia emissions (Cederberg and Stadig 2003). Thus, the application of system expansion is preferable. When a system expansion is applied, for example to dairy and beef production systems, it is assumed that the meat from both the culled dairy cow and the raised dairy calf replaces beef meat produced in a cow-calf system. The choice of meat and also the production system used to obtain this by-product is crucial because the amount of environmental impact from beef production depends on it (Flysjö et al. 2012). However, these allocation methods could be avoided, according to Weis and Leip (2012), who suggest allocating input and output flows of the processes for raising and fattening young animals for meat, and dividing the activities of dairy and suckling cows for milk production into the raising of young animals during pregnancy (which is allocated for meat) and the production of milk. The management of by-products, in LCA analysis, is another critical point. For livestock production, this is the case of manure because of its dual simultaneous effect (Garnett 2009): manure increases the nutrients in the soil and also the soil’s carbon sequestration potential (FAO 2001). It was estimated that, globally, almost 22 % of the total nitrogen and 38 % of the total phosphate used for agriculture productions derive from animal excreta, of which half come from beef production (UNEP, n.d.). At the same time, manure, according to the report “Livestock’s Long Shadow—Environmental Issues and Options” (FAO 2006), is responsible for N₂O and CH₄ emissions, contributing 5 % to global GHG emissions. In the majority of the studies analysed, manure management is considered as a means for the production of organic fertiliser that, depending on its relative nitrogen content,⁴ is used as a substitute for chemical fertiliser (e.g. Casey and Holden 2006; Beauchemin et al. 2010; Nguyen et al. 2010; Basarab et al. 2010; Beauchemin et al. 2011). This methodological approach allows the authors to assess the environmental impacts allocated to livestock production better without leaving out the manure use impacts (linked to acidification, eutrophication and GHG emissions); but Garnett (2009) considers it incomplete, because as a natural source of nitrogen, manure reduces the need for chemical fertiliser production and transport. According to Weidema et al. (2008) and Leip et al. (2010), manure must be considered in any system expansion approach aspiring to perform good LCA analysis. This means considering the impacts of manure fertilisation on the entire chemical N fertiliser supply chain, defining the level of substitution of chemical nitrogen fertilisers by manure (it varies from 20 to 60 % depending whether manure is spread during grazing or collected from stables) (Nguyen et al. 2010). Many LCA studies on livestock production consider the land use impact categories only in terms of the m² required annually for livestock production (Cederberg and Stadig 2003; Cederberg et al. 2009; Doreau et al. 2011; Ridoutt et al. 2011; Flysjö et al. 2012). This category is particularly relevant for beef production, which has higher impact than certain other meat production systems (pigs or chickens). In particular, as regards meat and dairy production, to obtain 1 kg of beef meat some 27–49 m² of land are needed (de Vries and de Boer 2010). The high values of land required for beef production are related to two factors: the low efficiency of feed conversion rate and the lowest number of annual progeny compared with pigs and chickens. Within the beef production system, the impact on the land use category is higher for suckler-cows than the system in which the herds are a co-product of milk (de Vries and de Boer 2010). However, despite the importance of including GHG emissions because of land use changes (Flysjö et al. 2012), there is no consensus on how to include those emissions in environmental impact estimates. Three methods which include land use and land use changes in LCA of milk production were analysed by Flysjö et al. (2012), who clearly showed how GHG emission estimates differed depending on the methodology used. Many LCA analyses deal with this problem by identifying the required land to produce a specific amount of output in a given period of time (Lindeijer 2000). This type of information is useful for evaluating land use efficiency, but, according to Nguyen et al. (2010), there are other issues that need to be considered: “the opportunity cost of land” (Garnett 2009, pp. 493), or rather the cost of the land if it was used for other purposes, and the potential land use change that derives from an increase in demand for land and land products. The opportunity cost of land use has been estimated in terms of emitted CO₂ as between 2.8 and 2.2 kg CO₂/m² year depending on whether it was converted to crop production or grassland (Nguyen et al. 2010) In another study, Nguyen et al. (2012b), using this indicator, found a potential impact reduction on GW between 20 and 48 % if grasslands were converted to forests rather than to annual crops. Roer et al. (2013) considering four sub-processes in the life cycle of bovine meat and milk production (concentrate, forage, cattle rearing and others) found that the GWP of meat production varies from 17.7 to 18.4 CO₂-eq/kg of carcass weight. This is also the only impact category which depends on cattle rearing and accounts for 45–48 % of the total GWP of the whole life cycle.

The methodological innovations emerging from this review seek to limit this huge variability by focusing on a combination of models from representative livestock farms and related emissions assessment with the LCA analysis. A thorough review of these methodological approaches can be found in Crosson et al. (2011), who summarise the GHG emissions per kg of product in 35 whole farm modelling studies (from 31 published papers) of beef and dairy cattle production systems. Beauchemin et al. (2010, 2011) use the HOLOS model, which is based on IPCC methodology, to assess CH₄, N₂O and CO₂ emissions at the farm level. This model accounts for all the emissions linked to the beef production supply chain (fertiliser and herbicide production and transport, feed production, etc.). Modelled results were used in LCA to assess the environmental impacts of beef production in Canada. Bonesmo et al. (2013) used the HOLOS model adapted to the Norwegian situation (HOLOSNor) to assess the GHG emission intensity of Norwegian dairy and beef production systems. They found that the main culprits responsible for GHG emissions per kg of carcass weight were, in order of relevance: soil’s nitrous oxide emissions, indirect energy use, soil C loss and enteric methane which was not significantly correlated to the variation in total GHG emissions per kg of carcass weight (Bonesmo et al. 2013). The same approach was used in a GGELS project (Leip et al. 2010), in which the CAPRI model was used to define the six livestock systems of EU−27 (Loudjani et al. 2010). The same method has been used to assess the GHG emissions and removals in the whole EU livestock sector (at regional scale) including: methane, nitrous oxide, carbon dioxide and also land use and land use changes (Weiss and Leip 2012) and CAPRI model results were used as input and output of the LCI. Oishi et al. (2013) evaluated the economic and environmental impacts resulting from changes in the age of animals at slaughter and in diet in the cow-calf system (race Japanese Black) in Japan, using as indicators the actualised net income and the environmental impacts from an LCA analysis. The input for economic evaluation was based on the continuous coupling system that was found (Oishi et al. 2011) to be the most economic and efficient one. The LCA analysis was built on the system presented by Ogino et al. (2007); a cradle to farm gate system with 1 kg of total live weight as FU. Then, the LCA results were normalised and the relative contribution of each category to the environmental impact of the whole system aggregated into a single dimensionless indicator, as suggested by Brentrup et al. (2004), for comparison with the economic indicator. Finally, a multi-criteria analysis was used to aggregate global warming (GWP), acidification potential (AP) and eutrophication potential (EP) impact categories following the approach suggested by Hermann et al. (2007). The results of this complex study showed that increasing culling parity to an economically efficient level can reduce the total environmental impact; changes in diet have no effect on environmental and economic impact (Oishi et al. 2011). Capper (2011) also pointed out that reducing time-to-slaughter can represent an option for decreasing CO₂-eq emissions per unit of beef because of the lifetime dilution of maintenance energy costs.

The combination of a bioeconomic model for livestock management with partial LCA (Carbon Footprint, Ecological Footprint, etc.) in order to assess the environmental impacts of livestock production systems represents an innovative approach to the environmental impact assessment for this sector. One example is the use of GBSM (Grange Beef System Model) with a partial LCA analysis in order to assess the GHG emissions of beef production systems The integration of farm management models with LCA analysis has also been suggested (e.g. Beauchemin et al. 2011; Oishi et al. 2011; Foley et al. 2011; Clarke et al. 2012), and some studies use other impact assessment methodologies to quantify the environmental loads produced by livestock production. All these attempts, which are in line with Place and Mitloehner (2012), represent efforts to account for the complex biogeochemical processes that occur within the rumen of cattle fed on different diets and also to account for varying management strategies such as age-to-slaughter, which can meaningfully alter the environmental load per unit of beef.