Chapter 1: Principles of Green Food Processing (Including Lifecycle Assessment and Carbon Footprint)

The global food and agricultural sectors are facing numerous pressures, including the burgeoning global population, the expanding middle class and the increasing desire of more people to have high-quality, low-cost food.¹  Numerous studies have shown that the main environmental hotspots within the food supply chain are associated with upstream activities (agricultural production, cultivation of crops and animal husbandry) and as a result have received the most attention from the consuming public, governmental organizations and non-governmental organizations (NGOs). These studies suggest that 70–90% of most environmental impacts in a full supply chain assessment can be attributed to the primary production phase; however, many of the same studies point to the food processing and manufacturing stage of the supply chain as being responsible for 10–20% of supply chain impact.^2–4  Also, although it is tempting to focus only on those upstream activities where the majority of impact arises, sustainability cannot be achieved by focusing on those activities alone, but must also identify opportunities and implement improvements at later stages of the supply chain. It is for that reason that this book is an especially strong addition to the literature for its focus on the food processing sector and the technologies and opportunities that exist for improvement of the environmental performance of food supply and improving food security.

The food manufacturing industry has traditionally held the role of ensuring food safety, regulatory compliance (for example, nutritional labeling), marketing and profitability. More recently, an additional layer of providing both information and documenting progress towards a sustainable food supply has been added. It should be clear that concerns over environmental sustainability of the food system will have secondary importance to the sector's traditional functions: unsafe, but environmentally friendly products will never be marketed. Hence the context of this chapter is to define the available operating space and useful techniques for understanding the role that environmental sustainability has in the food processing sector.

There is a consensus that the assessment of sustainability requires a holistic perspective of the system being evaluated. This includes the full supply chain, from cradle to grave, in addition to a full complement of environmental indicators. The cradle-to-grave perspective includes all activities necessary for the production of the item under study, extending back in the supply chain to the original extraction of resources. This means, for example, that coal mining and transport to the power plant to produce electricity for pumping or refrigeration are included. In addition, processes associated with consumption and end-of-life treatment are included. An example of the importance of including the full supply chain is in the evaluation of food packaging. One role of packaging is protection of the product, which reduces loss. Light weighting a package will make the package itself more sustainable, but if it leads to even a slight increase in food loss, the overall effect would very likely be a reduction in the overall sustainability of the system because of the relatively large impacts associated with the production of the food itself. By adopting a system perspective, tradeoffs between supply chain stages can be identified, which helps to avoid unintended consequences. In addition, a range of environmental categories should also be included in the overall assessment. Multiple categories allow the identification of potential tradeoffs between environmental impacts. For example, water use efficiency in a processing facility may be achieved at an additional energy cost and therefore the tradeoff of improved water use comes at the cost of an increased carbon footprint. This highlights the truism that “one size does not fit all.” For example, in water-scarce regions a higher footprint for global warming may be a necessary and acceptable tradeoff.

Sustainability is a complex concept with a deceptively simple definition: to meet the needs of current generations without compromising the ability of future generations to meet their needs. In general, sustainability is considered to have three pillars: social, economic and environmental. The complexity arises in attempting not only to balance environmental tradeoffs as mentioned above, but also to balance these tradeoffs with social and economic values that are deemed important. A major goal of sustainability assessment is therefore to identify the tradeoffs and tensions in the system so that fully informed decisions can be taken in an effort to maintain our collective ability to provide prosperity. Among the tools used for sustainability assessment are lifecycle assessment (LCA), lifecycle costing (LCC), social lifecycle assessment (SLCA), lifecycle sustainability assessment (LCSA), organizational lifecycle assessment (OLCA), environmental risk assessment (ERA) and, in the context of food safety, microbiological risk assessment (MRA). Some of these tools can be used in conjunction with each other or, depending on the needs of the assessment, may be used alone. An emerging paradigm in the context of systems is the so-called circular economy. In this paradigm, there is an explicit and conscious attempt to design products in a manner that makes the utilization of materials at the end of their intended life as raw materials for a subsequent use as streamlined and efficient as possible. Clearly, a fundamental principle of sustainability is resource use efficiency, and in the context of food processing this translates to minimizing energy and water use and food loss while simultaneously producing high-quality, nutritious and safe foods to enhance food security.

The most commonly used tool for system scale assessment of product systems is LCA, which is codified through a series of international standards, including general guidance in addition to specific guides for water footprint and carbon footprint.^5–8  These standards are targeted at providing guidelines for products and services and specifically require a full lifecycle perspective for the reasons outlined above. The International Organization for Standardization (ISO) has not published guidelines at the organizational level; however, the UNEP/SETAC Life Cycle Initiative and World Resource Institute have published guidelines for adapting LCA to the organizational scale.^9,10  LCC is a tool to permit the full cost of a product to be considered using the same system as used in LCA. The goal of LCC is to provide a full cost accounting of the production (including delivery and installation), operation and end-of-life costs (decommissioning and disposal) associated with a product. It may additionally include costs of externalities; for example, where environmental pollution costs that are borne by society can be quantified and verified, these externalities can also be included in the cost assessment.^11,12  Integration of LCC and LCA remains relatively uncommon, yet is an important area because all enterprises must be both economically and environmentally viable. SLCA arose from efforts within corporate social responsibility initiatives to quantify the societal metrics associated with production and consumption. SLCA is the least developed methodology, but recently guidelines have been published¹³  and databases created that allow the assessment of social risks in supply chains.^14–18  SLCA attempts to evaluate and quantify socially relevant indicators, including forced child labor, excessive work time, collective bargaining rights, health and safety and human rights. The European Commission, through the European Platform on Life Cycle Assessment, initiated an effort to extend the framework of LCA to incorporate LCC and SLCA to create LCSA. There remain clear challenges associated with collecting and maintaining up-to-date data in the social databases, and for this reason there are significantly fewer publications related to SLCA and LCSA than environmental LCA. For this reason, the remainder of this chapter will focus on environmental sustainability assessment. It should be noted that despite the growing popularity and utility of LCA, there are also some significant limitations to the methodology. For example, in evaluating agricultural production systems, LCA has limited capabilities with regard to the evaluation of ecosystem services, and also has only nascent capabilities for the inclusion of health effects associated with dietary choices or food-borne pathogens.

ISO requires that the selection of impact categories shall reflect a comprehensive set of environmental issues related to the product system being studied, taking the goal and scope into consideration. In other words, impact categories should be relevant to the product system under study. However, a clear definition of the criteria that define “relevant” remains elusive, but this is beginning to change. In the past decade, Product Category Rules (PCRs) have been developed to set up standardized rules for products that serve the same functions, including choice of metrics used to estimate impacts.¹⁹  However, inconsistency in the choice of impact category has been found in a comparison of five different PCRs developed independently by different organizations. For example, The Sustainability Consortium (TSC) and the Korea Environmental Industry and Technology Institute (KEITI) share five impact categories (climate change, ozone depletion, photochemical ozone formation, acidification and eutrophication), while each requires the inclusion of another impact category: ionizing radiation by TSC and resource depletion by KEITI.¹⁹  To address the inconsistency and duplications of PCRs, The Product Category Rule Guidance was launched in 2013 to provide more specific guidance on developing PCRs.^20–22  However, the selection of impact categories is still up to the individual PCR committee.²⁰ 

In addition to the ISO guidelines for LCA, additional efforts in several countries have led to the development of relative guidelines. Most notable among these are the Environmental Product Declaration (EPD), based on ISO 14025,²³  in Sweden, the Publicly Available Standard (PAS) 2050²⁴  from the UK and the French standard BP X30-323-0,²⁵  each of which provides additional guidance for the performance of LCA. Furthermore, at the level of the European Commission, through the Joint Research Council the ENVIFOOD,²⁶  provision of guidance specific to food and agricultural products is available. The European Commission DG Environment, in an effort to enable companies to market sustainable products without the need to perform assessments specific to each country where they may wish to market the product, established the program Single Market for Green Products Initiative. They have published the Product Environmental Footprint (PEF) and Organizational Environmental Footprint (OEF), guidance documents for developing product and sector-specific category rules defining the way in which LCA can be conducted, with the purpose of communicating to consumers regarding the environmental impact of their purchases.²⁷  PEF pilots were run between 2013 and 2016 to establish methods to ensure a common approach to measuring environmental performance. In doing so, the European Commission provided guidelines for the procedure for choosing impact categories in the pilot phase of conducting a PEF. It says that the identification of the most relevant impact categories shall be based on the normalized and weighted results of a screening study.²⁷  However, pilot studies can also choose impact categories based on the communication purpose,²⁷  which essentially links to the goal and target audience of the LCA.

In the USA, a 2010 study reported the estimated annual food waste at the retail and consumer level to be 638.7 kg per person. This food waste was valued to be $165.6 billion.²⁸  Similarly, another study reported that 90 million tons of food waste was generated in the European Union in 2010.²⁹  However, researchers looking at the methodology to calculate food waste highlighted the importance of data integrity and data collection. They suggested that government officials need to create policies that define and quantify what constitutes “food waste” due to variations among the definitions of different authors or agencies.³⁰  Sustainability is widely viewed as consisting of three equal pillars: social, environment and economic. However, in the Handbook of Sustainability for the Food Sciences, Morawicki argues that the environment is the foundational support for both social and economic sustainability.³¹  In the global context, it is certainly true that without the full suite of functioning ecosystem services, society and the economy are threatened. Nonetheless, it is important that sustainability assessment considers the three pillars equally, acknowledges unavoidable tradeoffs and provides a lens for informed decision making – these two perspectives are shown in Figure 1.1. Environmental and social impacts are often external costs and are difficult to assess because there may not be an immediate and direct or obvious effect on profits.³²  Similarly to the sulfur dioxide cap and trade that played a major role in mitigating acid rain, initiatives to implement a carbon emission tax or excessive waste dump taxes are actions by the government intended to put a price on these externalities. Another role of government policies is to foster communication between federal, state and local levels, guiding the discussion to improve sustainable practices. These policies have the potential not only to improve food safety of the industry, but also to improve sustainable practices among food plants.³³  The impact of law and policy practices on the food industry will be further explored in a later chapter.

In principle, LCA is simply an accounting of material and energy flows that result from each of the activities in the supply chain of a product or service. These flows are ultimately characterized and combined to provide a picture of the impact, across several environmental dimensions, of the system under study. LCA is a science-based, robust and standardized methodology for assessing the potential environmental impacts of products, services, or organizations. As described in ISO 14044, LCA consists of four stages, which will be outlined in the following subsections: Goal and Scope Definition; Life Cycle Inventory Collection; Life Cycle Impact Assessment; and Interpretation. Frequently in the process of performing an LCA it will be necessary to revise certain aspects based on new information, and thus the process becomes one of iterative refinement.

There are numerous reasons for performing an LCA, including the following: define opportunities for improvement through identifying activities that are major contributors to impacts (hotspots); develop performance benchmarks allowing the documentation of continual improvement; differentiate markets permitting the targeting of products towards consumers concerned about the impacts of products they purchase; design for environment, that is, considering the full lifecycle of a product at the initial development stages, enabling significant improvements to be made in overall lifecycle performance. With regard to marketing sustainable products, LCA is the basis for type III Environmental Product Declarations (EPDs) in which the environmental performance of a product can be communicated to consumers, following international standards.²³  An example of design for environment would be the consideration of the biodegradability or recyclability of packaging in the early design that resulted in reductions in landfill burdens for the end-of-life phase.

Broadly, there exist two modeling paradigms for performing an LCA: attributional and consequential.³⁴  Attributional LCA is an approach in which an average system is retrospectively evaluated and the emissions associated with the system are apportioned or allocated between multiple functions based on normative rules. Consequential LCA, on the other hand, does not consider the average system, but evaluates the anticipated change resulting from increased demand for an additional unit of product. Thus, the lifecycle inventories for the two systems are different, the first relying on averages and the second on data from the margin. Specifically, marginal processes are those processes anticipated to respond as a consequence of the additional demand. Attributional LCAs are typically used for identifying hotspots and benchmarking supply chain performance, whereas consequential LCAs are more relevant for decision support.

A relatively more recent approach for performing LCA, primarily in a business context, is known as organizational LCA.⁹  Most food manufacturing businesses have multiple facilities and each facility may have multiple products produced from a potentially wide range of input materials. Frequently, attributional and consequential LCA are used to evaluate individual products in a company's portfolio. This raises several technical and methodological issues needed to perform the accounting of inputs and outputs and assign these to the individual products. This multifunctionality can be managed in different ways, all of which have advantages and disadvantages. In some cases, however, in the evaluation of an organization as an entity in which its entire product portfolio is the functional output of the system of study, the methodological issues of multifunctionality are generally less important. An organizational, or even facility, level LCA can be performed based on bills of materials, utility consumption, transportation and waste management information collected at either the organization or facility scale. The scale of analysis is valuable for benchmarking performance and documenting continual improvement when there are not specific marketing requirements for individual products.

All LCAs involve four phases as described in the international standard:⁶  Goal and Scope Definition, Life Cycle Inventory, Life Cycle Impact Assessment, and Interpretation.

Because there are a range of potential uses for the results of an LCA, a critical first step is to define the purpose of the study, as that also has important implications for the data collection requirements. In addition, the ISO standards require different levels of data quality and external review for studies that are intended to make claims of superiority among products. The Goal and Scope Definition requires the establishment of the system boundary, functional unit, impact assessment categories, data quality requirements, and other methodological choices.

By definition, LCA addresses environmental impacts through the whole lifecycle of products and services, that is, from raw material extraction all the way to final recycling/disposal of the products. This is called the “cradle-to-grave” approach in the system boundary, which defines the unit processes that are included in the system under study. Covering the whole lifecycle provides a comprehensive understanding about all the potential environmental impacts through all lifecycle stages. However, the amount of data to be collected to fulfill the cradle-to-grave requirement is substantially higher. Depending on the goal of the LCA, truncated system boundaries such as cradle-to-gate or gate-to-gate can also be used. ISO defines the “gate” as the production site gate.⁸  Cradle-to-gate therefore means that impacts are assessed when the product leaves the production site, and further use and disposal stages are not included. Gate-to-gate therefore means that only processes at one or more single processing facilities are included. In food LCAs, cradle-to-gate often refers to cradle-to-farm or processor gate, whereas gate-to-gate often means from farm to processor or retailer, or down to a single processing stage, excluding use (e.g. cooking) and disposal of packaging and food waste. Note that cradle-to-gate LCA is only a partial LCA since the full lifecycle is not included and important contributors may be missed out owing to the truncation, while gate-to-gate is arguably not an LCA despite often being applied for various purposes.

System boundaries are tailored to the goal of the LCA study, which reflects the stakeholder's concerns. If the goal is to understand the impacts throughout the whole lifecycle, then obviously cradle-to-grave would be a rational choice. If the goal is to facilitate communications between a supplier and a processor, then cradle-to-gate may be sufficient for the processor to understand the “embedded” burden with its purchased products from its supplier. If the goal is to understand how consumers use the products and how to reduce food waste, then the use and disposal stages need to be considered. If the goal is for internal benchmarking (e.g. compare different drying methods), then gate-to-gate can also be deployed. When reading the literature, it is important to note differences in the boundaries of studies. Figure 1.2 shows an example system description for a cradle-to-grave assessment of a ready-to-eat meal.

LCA is intended to allow comparisons between products or production scenarios. A fair comparison can be achieved only if the functions delivered by each of the systems are identical. Hence proper definition of the system function is imperative. In principle, the definition of the functional unit should describe what function is fulfilled with which obligatory properties and for what length of time it will fulfill the function. The time characteristic is primarily used for goods that can be repeatedly used for the same purpose, e.g. a shirt. In food processing, this would be manifest in a specified shelf life. Because food safety is paramount, all food processing functional units must have, either explicitly or tacitly, the obligatory characteristic of meeting all food safety requirements.

The nutritional differences between foods and the associated health impacts from the consumption of different diets complicate the definition of a functional unit for food. Only recently have nutritional health aspects (non-communicable disease burden) of food been incorporated into LCA.^35–40  In the context of this book, however, the nutritional quality of foods is expected to be of relatively minor importance, and the principal use of LCA in the food processing sector will be focused on continual improvement either for individual product lines or for overall facilities. For these situations, a functional unit of a specified quantity of packaged food produced is generally adequate. However, depending on the specific goal and scope of a study, if it is known that the processing has a significant effect on the nutritional quality either through enrichment or fortification (e.g. vitamins added to cereals or milk, respectively),⁴¹  the inclusion of nutritional qualities as obligatory characteristics of the functional unit may be important. An alternative approach, based on assessing the health impacts of nutrition in the diet, is discussed in Section 1.10.2.

In this phase of an LCA, a supply chain model is constructed from linked unit processes. Each unit process is a complete description of one stage or activity in the supply chain. The unit process has inputs from other unit processes or directly from nature (resource extraction) and outputs both to the environment (emissions) and to other unit processes [the product(s) of the particular activity]. Practitioners will typically try to collect primary data for processes closely related to the product under study (foreground processes) and use information from existing databases for the processes in the background system. Background processes might include electricity generation or the production of refrigerants or fuels. Frequently, the unit processes are linked using software that allows rapid calculation and analysis of the results. An important consideration in this phase of an LCA is to realize that when converting inventory into impacts, there can be mismatch of the inventory flow and impact category. This can occur, for example, if ecotoxicity is chosen as an impact category while some relevant chemicals are not included in the inventory, or a characterization factor is missing for a specific inventory flow, or even results from something as simple as a different naming convention. Some LCA software provides a tool to identify each inventory flow for which a corresponding characterization factor is missing. This highlights an important characteristic of performing LCA, namely an iterative process where information gleaned from subsequent steps may require modification of earlier steps.

LCA addresses environmental impacts through multiple metrics, i.e. impact categories. This step is referred as lifecycle impact assessment (LCIA). In this way, emissions and resources consumptions collected through the lifecycle inventory stage are classified into impact categories and quantified through impact indicators. This phase of an LCA is typically performed using existing frameworks and is characterized by the process in which a large number (typically thousands) of emissions to the environment are characterized for their relative contributions to a limited number of midpoint impact categories, typically between 12 and 20, depending upon the specific impact assessment framework chosen for the study. This also makes the inventory easier to interpret, as it “translates” inventory into environmental concerns. The LCIA phase of an LCA constitutes three mandatory steps:⁶ 

• Select impact categories and characterization models.

• Classification: assign LCI results to impact categories.

• Characterization: calculate category indicator results.

Two classifications of impacts have been defined: midpoint and endpoint categories. ISO defines an endpoint as “attribute or aspect of natural environment, human health, or resources, identifying an environmental issue giving cause for concern,” and midpoints are located between the inventory and the endpoints in an environmental cause–effect chain. In other words, endpoints are the fundamental environmental damage about which we are concerned, and this currently covers three areas, namely human health, ecosystem health and resource use or depletion. Endpoint impact categories therefore signify a damage-oriented approach. Midpoint impact categories, on the other hand, reflect a problem-oriented approach, because they describe the specific problems resulting from the inventory flows (e.g. soil acidification). Endpoint results have a higher level of uncertainty than midpoint results owing to additional modeling down the environmental cause–effect chain, but they provide direct indicators of the environmental concerns and are therefore easier to communicate.

One of the first attempts to define best practice for impact category selection was made in 1999 during the SETAC–Europe second working group on LCIA, where 10 midpoint impact categories were proposed:⁴²  abiotic resource use, biotic resource use, land use, climate change, stratospheric ozone depletion, human toxicity, ecotoxicity, photooxidant formation, acidification and eutrophication. Since then, various methodologies have been developed with different packages of impact categories, but most of the baseline categories are still included as summarized in Table 1.1.

LCA researchers have developed characterization factors for a wide variety of emissions. As one example, the Intergovernmental Panel on Climate Change (IPCC) has established that methane has a radiative forcing factor that is 28 times greater than that of carbon dioxide over a 100 year time frame; thus the characterization factor for methane is 28 kg CO₂e kg⁻¹ methane.⁴³  The carbon footprint of a product is obtained by multiplying all of the greenhouse gas (GHG) emissions (across the entire supply chain) by their respective characterization factor and then summing to report the cumulative emissions as an equivalent quantity of carbon dioxide (CO₂e). Characterization factors are developed by following an environmental cause–effect chain that accounts for the cascade of events beginning with the emission of substance into the environment until it interacts with either a receptor or the environment itself to cause an impact.⁶  As another example, nitrate emissions in wastewater will travel downstream, ultimately reaching the marine environment where it may cause eutrophication.

Similar, although less certain, cause–effect chains are used to develop characterization factors that combine midpoint categories into endpoint categories. For example, carcinogens, non-carcinogens and particulate matter and also ionizing radiation and climate change have direct or indirect impacts on human health. In this example, the endpoint impact is reported in units of disability-adjusted life-years (DALYs). DALYs are a measure of the decrease in life expectancy as a result of the emissions leading to ill health, disability or early death. These are the same units as used in estimating the impacts of food-borne illness from microbial risk assessment; the connection will be discussed in more detail below.

In this phase of an LCA, the implications of the results from the impact assessment are discussed and placed into context of the system under study. Hotspots and tradeoffs will be identified, and opportunities for improvement should be discussed, including the potential benefits that may be realized through their implementation. Because LCA studies complex systems and requires large amounts of data, it is also important in the interpretation phase that limitations of the conclusions be clearly articulated. Any study that reports on a comparison of two systems should also include an assessment of uncertainties that can arise from differences in data quality and potentially from methodological assumptions. Some additional, optional, steps of LCIA can also be taken and can be useful for interpretation, as follows:⁶ 

• Normalization: calculate the magnitude of indicator results relative to reference values.

• Grouping: sort and rank the impact categories.

• Weighting: convert and aggregate indicator results across impact categories using weighting factors based on value choices.

• Data quality analysis: understand the robustness of the results.

When multiple impact categories are used, normalization can reveal the relative importance of each impact category within a certain geographical context. Normalized impact results are compared with a reference value, e.g. the average environmental impact per capita in one year within a certain country. Thus, normalized results are expressed as dimensionless fractions and therefore can be compared across impact categories that were previously incomparable owing to the use of different units. This gives some idea about the relative magnitude and therefore relevance of each impact category compared with the others.

By addressing multiple impact categories, LCA avoids the narrow focus of a single impact and therefore leads to a more comprehensive understanding about the overall impacts. However, LCA addresses only the environmental issues that are specified in the goal and scope. Therefore, LCA, despite accounting for multiple metrics, may not provide a complete assessment of all environmental issues regarding the product system under study.⁵  For example, noise and odor are not yet accounted for in LCIA. In addition, many impacts that are accounted for do not fully include the effect of the location of the emission in the characterization. Extending LCIA to account for geospatially explicit inventory and impact is a current area of intense investigation.^44–47 

Tradeoffs are often found between different impact categories. For example, UHT milk has a longer shelf life and potentially less spoilage, but the energy consumption for thermal treatment is higher than for pasteurized milk. The nutrient content of UHT is lower than that of pasteurized milk, if vitamins are taken into consideration. A study comparing environmental impacts of primary livestock products found similar tradeoffs (Figure 1.3). For example, pork has a relatively small land occupation but a comparatively high energy intensity.

Since food processing is downstream of agricultural production, the two are often assessed together in LCA studies to meet the ISO standard requirement of extending the system boundary to include extraction from nature. Common midpoint impact categories for agricultural production are therefore often applied to food processing, including global warming, acidification, eutrophication, water consumption, (agricultural) land occupation, abiotic resource depletion (e.g. use of phosphorus) and biodiversity. A recent study has found that LCAs focusing on climate-related impact categories have proportionally increased over the last 20 years.⁴⁸  This reflects the growing concerns about climate change but also narrows the environmental focus of LCA studies.

Generally, processed food has higher environmental impacts than less processed food over the lifecycle. This can be seen from the energy intensity of dairy products (Figure 1.4).^49–51  For example, liquid milk has a far lower energy intensity than powder products, which require thermal energy to remove moisture. The trend is likely to remain when making a comparison on a dry mass basis (e.g. protein), because further processing tends to increase the concentration of solids. For example, agricultural production tends to dominate the climate change impact for livestock-derived food products. This is because the farming stage tends to produce CH₄ and N₂O, which are GHGs with a high global warming potential (GWP) of 28 and 265 kg kg⁻¹ CO₂e, respectively,⁴³  whereas inputs to industrial processing, generally related to combustion of fossil fuels, tend to be produce CO₂, which has a GWP of 1 kg kg⁻¹ CO₂e. This may not be the case for other impact categories. An interesting study on the environmental impacts of a Spanish diet found that when human excretion is included in the whole lifecycle of food consumption, it became a significant contributor to the eutrophication owing to the emissions of nutrients in treated sewage (Figure 1.5).⁵² 

Even though the majority of environmental impacts occur in the agricultural production stage, there have been a number of studies that were focused on the food processing sector. One of the first environmental LCAs to be conducted, in 1969, by the Coca-Cola Company evaluated the environmental impact of their containers; hence the food processing industry has a long history of endeavoring to understand the environmental consequences of their products. More recently, a significant number of LCAs related to dairy processing have been performed,^{15,50,53–65}  whereas some have taken a broader view,^66–71  and others have focused on different sectors such as ready-made meals,^72–74  vegetable oils,⁷⁵  eggs,⁷⁶  vegetables,^77,78  meat processing^79,80  and packaging.^69,81–83 

Applying the cradle-to-grave approach to food LCA provides a holistic picture of the impacts throughout the system. For example, a study on phosphorus efficiency through the US food system found that only 15% of the total phosphorus extracted from nature for the provision of food was eventually ingested by humans, the remaining 75% being lost to the environment at different stages of the lifecycle (Figure 1.6).⁸⁴  The top three contributors along the food chain were identified as livestock, meat and dairy production and crop cultivation, and household food waste, mining waste and fertilizer manufacturing waste were also significant; however, the food manufacturing sector was a very small contributor.⁸⁴  This type of analysis supports targeted efforts on those stages in the supply chain where the greatest benefits can be realized, and may provide decision support for national or international policies, where a whole-system view is necessary to manage tradeoffs. It should be noted that the 46% of emissions to soil may not be necessarily completely lost as they represent stages in phosphorus cycling, thus highlighting the importance of defining system boundaries when performing LCA.

Tracing the whole lifecycle can also reveal the locations where the impacts occur and how that affects the overall environmental impacts. A study on environmental impacts associated with Dutch private consumption showed that food production and consumption were dominant in the environmental load for pesticide use, fish extraction, water use, land use, eutrophication and acidification, most of which took place outside The Netherlands.⁸⁵  Contrary to some common beliefs, local food products may not always be an effective way to reduce the climate change impacts associated with food, since transportation contribution tends to be much smaller on a total lifecycle basis than the production stage.^86,87  This suggests that using the most efficient production technologies is frequently more important than the production location.

Full LCA can reveal the importance of food packaging and waste management beyond the farm gate. In a study of packaging and food losses of five food items, it was found that packaging was an important tool in reducing food waste, and therefore reducing the total environmental impact, even if there was an increase in impact from the packaging itself.⁸⁸  For example, it was found that reduction of losses from milk, bread, beef and cheese by 2% or more decreased the total energy use despite the increase in energy use in the packaging phase.⁸⁸  This was because all impacts are scaled against one unit of product, hence the less product is lost the lower are the overall impacts. Another study of cradle-to-grave LCA on a canned ready meal found that solid waste management was the dominant contributor to marine ecotoxicity and fresh water ecotoxicity, whereas the food ingredients (cooked pulses and pork meat cut) were the dominant contributor to land use, carcinogens, acidification/eutrophication, global warming and terrestrial ecotoxicity.⁷⁴  A further study comparing ready-made meals with homemade meals suggested that homemade meals have a lower impact, largely owing to less food waste, and also that frozen meals have a greater impact than chilled ready-made meals owing to the additional energy requirements in the supply chain. Interestingly, ingredients sourced from Brazil or Spain had a lower impact than similar ingredients sourced locally in the UK.⁷²  This shows the different environmental aspects to which food ingredients and solid waste contribute. A cradle-to-grave perspective and broad impact categories can ensure a comprehensive understanding of the impacts and provide insight into sourcing and design for waste minimization.

The primary production stage of food products is usually found to be a significant contributor to the overall impacts. For example, a cradle-to-grave LCA of yogurt found that dairy farming was the main contributor to acidification (91%), eutrophication (92%) and climate change (62%).⁸⁹  A cradle-to-gate LCA of quick-service restaurant chicken meat found that the poultry grow-out stage was a major contributor to climate change (41%) and non-renewable energy depletion (55%), and feed production (upstream to poultry grow-out) was a significant contributor to climate change (34%) and water depletion (66%).⁹⁰ 

It is therefore not surprising to find that food LCAs are still heavily focused on the agricultural “gate.” Searching for “lifecycle assessment dairy” in Thomson Reuters Web of Science, of the 45 journal articles published in 2016 (up to 10 December) that were focused on dairy or dairy-related systems (e.g. beef), 24 articles (53%) set the system boundary from the cradle-to-farm gate, meaning that environmental impact is assessed when raw milk is ready for delivery from the dairy farm.

At the farm gate, differences were found between animal- and plant-based food products. A study on 84 common foods found that animal-based foods were associated with higher energy use and GHG emissions than plant-based foods, except vegetables produced in heated greenhouses or transported via air freight (Figure 1.7).⁹¹  This is mainly due to relatively low feed conversion rate (FCR) (kg feed consumed per kg product) in animal production, a measurement of an animal's efficiency in converting feed into useful products (milk, meat, eggs, etc.). A study on livestock systems in Europe found that FCR ranged from 19.8 : 1 for beef, 4.1 : 1 for pork, 3.3 : 1 for poultry, 2.8 : 1 for eggs, to 1.2 : 1 for cow's milk.⁹² 

Food-preserving technology during processing was also found to play an important role in the overall impacts of food. A study compared four thermal and non-thermal techniques (autoclave pasteurization, microwaves, high hydrostatic pressure and modified atmosphere packaging) and their effects across six impact categories (acidification, eutrophication, global warming potential, photochemical oxidation, water depletion and cumulative energy demand) and found that they each had strengths in different circumstances.⁷⁰  Microwaves and modified atmosphere packaging were found to reduce energy demand and GHG emissions, non-thermal technologies (modified atmosphere packaging, high hydrostatic pressures) were found to have lower water requirements, and modified atmosphere packaging was found to be the most sustainable option when a shelf life shorter than 30 days is required (Figure 1.8).⁷⁰ 

With all the merits of cradle-to-gate and cradle-to-grave LCAs, gate-to-gate studies can still provide a detailed and focused investigation on specific parts of the supply chain such as transportation. For example, a study on 50 fluid milk processing plants across the USA analyzed the farm gate to retailer loading dock GHG emissions of fluid milk processing, and found that truck fleet tailpipe emissions were the largest contributor to GHG (29%), followed by plant electricity usage (27%) (Figure 1.9).⁵⁸  The significant contribution of transport was due to the long distance from farms to processing plants, averaging 850 km per round trip.⁹³  The plant's annual volume of milk production (<80, 80–120, >120 million kg) was not found to be a strong indicator of the GHG total emission intensity (kg CO₂e per kg of packaged milk), despite the fact that larger plants tended to have higher distribution-related emissions. This was because the overall GHG intensity was also dependent on factors such as regional electricity grid mix and whether or not there was onsite wastewater treatment and multiple coproducts.⁵⁸ 

A gate-to-gate approach has also been applied to study specific technologies for food processing. An LCA study comparing four cleaning-in-place (CIP) methods in dairy processing found that those with small volumes and low temperatures (i.e. enzyme-based cleaning and one-phase alkaline cleaning) had lower impacts on energy use, global warming, acidification, eutrophication and photooxidant formation than the conventional methods (i.e. alkaline/acid cleaning with hot water disinfection, and with disinfection using cold nitric acid at pH 2), and phosphorus and nitrogen in the detergents influenced the eutrophication in addition to milk residue in the rinsing phase.⁶³  Another LCA study compared two drying methods for apple powder and found that drum drying, despite being simpler and cheaper, had higher impacts than multistage drying (ultrafiltration, crystallization and spray drying) in all selected impact categories.⁹⁴ 

It may be argued that gate-to-gate studies, despite being narrowly focused on individual stages and processes, provide more insights at the operational level. It is often difficult to collect detailed data through the whole lifecycle and many approximations and assumptions must be made without possibly conducting a full validation. This is arguably an inherent weakness of the LCA methodology, namely that it may be spread too thin to reveal the details. Gate-to-gate studies provide useful supplementary data to the whole lifecycle and therefore can make valuable contributions to understanding the bigger picture. When one or more stages of the lifecycle have been found to dominate the whole lifecycle impacts from a hot spot analysis (e.g. industrial stages dominate the impacts of apple powder),⁹⁴  gate-to-gate studies are particularly useful to provide in-depth knowledge by dissecting the impacts of individual stages.

This study was a LCA intended to compute the lifecycle inventories and evaluate the environmental impact in terms of global warming potential (CO₂e) for the aggregate fluid milk processing sector. In mid-February 2008, an extensive survey was completed by 50 individual processing plants reporting on their production during calendar year 2007. Information in each survey included plant major energy consumption; water consumption; truck fleet fuel consumption; refrigerant purchases for both the plant and truck fleet; on-site milk packaging production; packaged milk type and sizes; and annual production for total plant fluid, fluid milk and packaged milk. Information provided in the surveys supplemented by correspondence for clarification and verification were the primary data sources for the study.

The functional unit was defined as 1 tonne of packaged milk. The environmental impact metric was chosen as kg CO₂e emitted to the atmosphere. The gate-to-gate system boundaries began with raw milk entering the plant's refrigerated storage silos and ended with the delivery of packaged milk to the customer via the plant's distribution truck fleet. Incidental effects such as employees’ commutes and business travel for industry executives were not included. Where allocations of inputs were required, the allocation procedures followed the ISO allocation hierarchy. Primary allocations occurred for processing plant electricity and plant heating fuel; allocation was based on the fraction of total packaged milk with respect to all other total processed plant fluids. Packaging and distribution were not allocated, but were based on totals reported on the survey based on packaged milk volume. For example, diesel for the distribution of milk was specifically requested on the survey, and therefore not corrected to account for delivery of any other products.

Three general types of fluid milk containers [high-density polyethylene (HDPE), polyethylene terephthalate (PET) and paperboard carton] of various sizes (gallon, half gallon, quart, pint, half pint, single serve, and other) were reported by each plant. Most plants purchased HDPE resin to blow mold their own gallon and half-gallon containers on-site. Total emissions associated with packaging included container material (and material raw-material extraction and manufacture), bottle production and caps. The majority of fuel energy in milk processing was used to produce steam for thermal processing, equipment cleaning and other plant processes.

Each facility reported overall consumption of energy in terms of the primary energy carrier (e.g. fuel oil or natural gas). When the electricity consumption location was known, this study used emission factors (in kg CO₂e kWh⁻¹) for the three US regional interconnection grids. National average emission factors were used when the location was unknown, specifically to account for containers that are blow molded offsite and transported preformed. The three regions were Eastern Interconnection, Western Interconnection and the Electric Reliability Council of Texas (ERCOT) Interconnection.

The emissions profile from each of the 50 individual facilities is presented and the profile of one individual plant (the same plant in each figure) is highlighted. These results show that there is significant variability, even among modern processing facilities, with respect to environmental impacts. Furthermore, individual plants may be high performing or low performing in one or more of the operational stages. With data such as these, individual plant managers can see how they are performing with regard to a representative industry average and, if their performance is worse than average, they can target specific efforts to reduce impacts (energy and therefore costs) associated with that operational phase that is higher than average. Figure 1.9 provides a contribution analysis for each major operational component of the average emissions aggregated over all 50 plants.

Processing GHG emissions are associated with the thermal processing and standardization of raw milk, including container filling. There are three components to processing GHG emissions: plant electricity (minus any onsite blow molding, included in packaging), plant heating fuel (generally natural gas) and refrigerant leakage. The emissions for the processing operations at each of the 50 facilities are presented in Figure 1.10.

Packaging emissions are associated with the packaging materials for processed milk containers. There are two components to packaging GHG emissions: packaging materials and electricity consumption for package manufacturing. Emissions associated with packaging operations are presented in Figure 1.11.

Distribution emissions are associated with the transportation of processed and packaged fluid milk from the plant to the point of delivery to the retailer. There are two components to GHG emissions: truck tailpipe emissions and truck refrigeration system refrigerant leakage. Total emissions for distribution are given in Figure 1.12.

Figure 1.13 presents the total gate-to-gate emissions profile of the 50 facilities participating in the study. The black column in each of the previous charts represents the same facility and demonstrated the variability even among relatively modern and well-managed facilities. This suggests that there is significant potential across the food processing industry to collectively improve sustainable practices through the identification of the fundamental sources of this variability.

As mentioned above, a major function of the food processing sector is to provide nutritious and safe foods with a sufficiently long shelf life to minimize losses due to spoilage. Significant energy resources are expended in food processing for the control of pathogenic or toxin-producing microbes. In the following sections, we briefly introduce several emerging technologies that have potential for significant energy savings while maintaining the same food safety standards as conventional technologies. Subsequent chapters provide significant details for each of these alternative technologies.

High hydrostatic pressure processing (HHP), also known as high-pressure processing (HPP), improves shelf life and decreases the need for chemical preservatives. In addition, the energy requirement of HPP is expected to be lower than that of thermal processing, but more research analyzing the energy requirements of HPP needs to be carried out to validate the reduced energy claims.⁹⁵  HPP is a promising technology to combat viruses in high-risk foods because the treatment of pressure applies an equal force to all areas of the foods at the same time.⁹⁶  The pressures and temperatures generally used in industry are around 600 MPa and <45 °C. Specifically, the products are placed in an HPP chamber and the vessel is closed, filled with a pressure-transmitting medium and pressurized either by pumping into the vessel or by reducing the pressure chamber volume, for example, by using a piston. The product is then maintained at high pressure for a specified time to inactivate microbes or change the composition of the product. Several pressure-transmitting fluids are used to displace space and protect the inner vessel surface from corrosion, including water and food-grade solutions such as castor oil, silicone oil, sodium benzoate, ethanol and glycol.⁹⁵  Pressure applied to foods being processed is transmitted isostatically and instantaneously; hence the process is not dependent on the shape, size or composition of the food. Lou et al. reported, in a review, that HHP has a superior ability to inactive viruses compared with other novel technologies, and specifically a significant reduction (≥5 log) was shown by HPP in fresh foods and shellfish, but not by ionizing irradiation, high-intensity ultrasound or ultraviolet light processing.^96,97  They also reported that HPP has potential to inactivate both enveloped and non-enveloped virus particles, and stated that HPP is one of the most promising technologies to ensure food safety and quality.

Ohmic (OH) heating of foods has been used by the food industry in applications such as blanching, drying, evaporation, dehydration and fermentation. OH is an electric resistance heating method. Electrodes are used to pass current through the food, which results in the production of heat within the food (Joule heating, due to the electrical resistance of the food) and leads to thermal inactivation of microbes. Thus, conductivity of the food is a critical factor in the effectiveness of OH processing. Other key factors are heat generation in the system, electrical field strength, residence time and how the food flows through the system.⁹⁸  In a review, Pereira and Vicente reported that OH can reduce the time required to reach temperatures for high-temperature, short-time pasteurization.⁹⁹  Moreover, OH provides uniform heating of a liquid with little or no difference in external and internal thermal penetration. Consequently, OH has a high energy efficacy with minimal heat loss during processing. Similarly to current thermal processes, OH can be used for continuous processing, but it has fewer problems due to surface overheating and fouling in comparison with conventional thermal processing. Hence OH systems are associated with low maintenance costs. Research sponsored by the US Environmental Protection Agency (EPA) highlighted potential cost savings for food manufacturers and suggested that OH can potentially reduce energy demand in the food industry. A study using sweet potatoes as the model food found that OH reduced freeze-drying time by 25%. Exploring uses of OH for food by-products may identify opportunities to reduce the environmental impact of food processing. For example, OH has recently been investigated as a means to increase the extraction yield of antioxidant bioactive compounds from winery waste products. The OH treatment increased the polyphenol extraction of grape pomace by 36%. This prevents food by-products from simply being thrown away by providing added value and reduces the overall impact of the food processing on the environment.

Pulsed electric field (PEF) processing exposes the food to an intermittent, high-intensity electric field for short periods of time. The high field intensities are achieved through storing a large amount of energy in a capacitor bank (a series of capacitors) from a DC power supply, which is then discharged in the form of high-voltage pulses.¹⁰⁰  Studies on energy requirements have concluded that PEF is an energy-efficient process compared with thermal pasteurization, particularly when a continuous system is used.¹⁰⁰  The energy savings associated with PEF result from the decreased operating temperatures, which reduce the energy needed for cooling the processed product. The potential emission savings estimated for the not-from-concentrate orange juice industry are in the range 33–66 kt CO₂e per year.¹⁰¹  Laboratory- and pilot-scale treatment chambers have been designed and tested for PEF treatment; however, continuous operation is more appropriate for industry-scale applications. One limiting factor of PEF technology is that it does not have the capability to inactivate bacterial spores, most likely owing to penetration limitation through the spore cytoplasmic membrane. In a recent review of the effectiveness of PEF processing on the deactivation of microorganisms, Saldaña et al. showed that PEF resistance varies greatly across different strains of bacterial species and food types (e.g. composition, pH).¹⁰²  Hence the potential sustainability benefits of this technology have yet to be fully demonstrated.

Plasma technology has been studied since the mid-1990s for food processing. It uses a mixture of electrons, ions, atomic species, UV photons, and charged particles to kill microorganisms.¹⁰³  The advantages of plasma processing, as reported by Thirumdas et al., are (1) high efficiency of microbial inactivation treatment at low temperatures, (2) precise generation of plasmas tailored for processing of specific foods and (3) “just-in-time” production of the disinfection agent, (4) low impact on internal food matrix, (5) no water or solvents used and (6) no resides.¹⁰⁴  Plasma technology can be divided into three categories: (1) remote-treatment cold plasma systems, (2) direct-treatment cold plasma systems and (3) electrode contact systems. Niemira stressed that cold plasma antimicrobial inactivation rates require continued validation, and concurrent research should also consider the impact of cold plasma treatment on sensory attributes, i.e. taste, smell and flavor, because sustainable food processing must also provide desirable products.¹⁰⁵  As cold plasma surface and gas-phase antimicrobial treatment research continues, the economics will help guide the food industry. Li and Farid's review described recent developments in atmospheric cold plasma (ACP) technology and found that it involves lower water consumption compared with current post-wash treatment of produce.¹⁰⁶  Moreover, they reported that post-packaging treatment can prevent contamination associated with non-sterile packaging, which leads to improved food safety.¹⁰⁶  A study by Ziuzina et al. highlighted the effectiveness of ACP in the disinfection of Escherichia coli, Salmonella enterica serovar Typhimurium and Listeria monocytogenes on fresh produce.¹⁰⁷  They concluded that ACP treatment of food in a sealed package can significantly lower microbial activity 24 h post-packaging.¹⁰⁷  By reducing the microbial load on fresh produce, shelf life can be increased and thereby reduce wasted food, which has been estimated, in 2008, to be 123 kg per capita in the USA.²⁸  Reducing or eliminating food waste is a very effective tool for enhancing food supply sustainability because, as mentioned previously, a large share of the full supply chain environmental burden typically occurs in the primary production phase.

Microwave heating is founded on the principle of converting electromagnetic field energy into thermal energy by oscillating the polarity of molecules in materials. A frequency range of 300 MHz–300 GHz (2450 MHz is most common) is used for microwave food processing. Microwaves cause internal heating based on the food's dielectric properties. Moreover, microwaves can easily be combined with other technologies such as microwave/convection drying, microwave/vacuum drying, microwave freeze-drying, microwave-assisted fluidized bed drying, microwave drying and osmotic dehydration.¹⁰⁸ 

A review on microwave processing in the disinfestation of cereals and pulses estimated that insects, mites, rodents and microbes in stored grains cause an estimated 10% of food waste. In developing nations, the loss is as high as 30%. Owing to differences in the dielectric properties of the grain and the insects, selective heating of the insects is possible. Thus, the dielectric based heating kills insects, but leaves the grains unaffected, reducing food waste, which represents an important sustainability gain.

Another review reported the development of “smart drying” using microwaves coupled with real-time sensors that monitor various material qualities. These sensors can measure and report parameters such as moisture content, color, shape and conditions inside the drying equipment such as pressure, velocity, temperature and humidity. If impurities or defects are detected, then automated adjustments are made to the processing line through a process control feedback loop. Thus, an optimal level of microwave power is applied during processing, which will improve energy conservation.¹⁰⁹ 

Microwave processing has attracted renewed interest as a method to extract the essential oils and bioactive components of spices and herbs, without the use of solvents or water. In place of the solvents or water, the process uses a combination of microwave heating and dry distillation. Again, the sustainability gains associated with reduction in water and chemical solvent consumption may be noteworthy, but have yet to be fully studied.

High-intensity (20 000 times more intense than sunlight) pulsed-light (HIPL) processing is a surface decontamination technology.¹¹⁰  In addition, HIPL can be used to decontaminate food contact surfaces in food production facilities. The US Food and Drug Administration (FDA) approved pulsed UV light in 1996 for the production, processing and handling of foods. Synonyms for pulsed-light technology include pulsed UV light (PUV), intense pulsed light (IPL), high-intensity broad-spectrum pulsed light (BSPL), intense light pulses (IPL) and pulsed white light (PWL).¹¹¹  HIPL uses a xenon flash lamp, converting electric pulses into short-duration (1 μs–0.1 s) and high-power pulses of radiation with a broad emission spectrum, ranging from ultraviolet (200 nm) to infrared (1100 nm). This technology targets microbes’ DNA by causing irreversible damage to the DNA, leading to death because of the impaired replication. In addition, an increase in temperature from HIPL processing induces photothermal microbial inactivation, which increases the HIPL lethality.¹¹²  The HIPL process does not produce any chemicals, carcinogenic residues or toxic by-products.¹¹³  When HIPL is evaluated on an energy cost per microbial reduction basis, HIPL is more energy efficient than other emerging technologies; however, HIPL demands a high start-up energy input and may require cooling units to prevent overheating.¹¹⁴  A study compared HIPL with near-UV–visible light (395±5 nm) and continuous UV light treatment and found HIPL to inactivate more efficiently E. coli and Listeria innocua in a bench-top liquid suspension experiment.¹¹⁵  Other tests have shown it to be successful in application to various food products such as produce, fruits, meats, dairy products and honey.^111,113,114  The main sustainability gains achievable using HIPL appear to be related to energy conservation.

Infrared electromagnetic waves cause radiative heating. Infrared heating can be sub-categorized into near-infrared (0.75–3 µm), mid-infrared (3–25 µm) and far-infrared (25–1000 µm). It can be used in drying, baking, roasting, grilling and reheating of food products. Owing to the variations in how infrared heating can be created, it can function either like hot air heating or like microwave heating.¹¹⁶  A review summarized the advantages of infrared radiation's high thermal efficiency compared with convection as (1) high thermal and energy efficacy, (2) faster heating rate, (3) shorter response time to reach operating temperatures, (4) uniform drying temperature, (5) high degree process control, (6) cleaner working environment and (7) the possibility of selective heating. The disadvantages include (1) low penetration power depending on the electromagnetic wave source, (2) prolonged exposure may cause undesired effects on the food's structure and (3) the heating is not sensitive to the reflective properties of coatings.¹¹⁷  Two compelling industry applications are in tomato peeling and corn drying operations. It was found that compared with conventional lye peeling, infrared dry peeling required a heating time of only 30 s compared with 75 s, showed a lower peeling loss (8.3–13.2% versus 12.9–15.8%) and reduced the energy to tear tomatoes from two different tomato cultivars.¹¹⁸  When infrared heating was applied to shelled corn drying, it was found that infrared heating was most efficient for corn with a high initial moisture content (the highest used in the study was 28%).¹¹⁹  Hence infrared drying could be used as an initial drying step to reduce the overall time and energy of the process.¹¹⁹  Infrared food processing therefore touches on two important sustainability points: reducing food loss and energy conservation. Additional specific applications in the food industry can be found in Rastogi's review.¹²⁰ 

Microbial inactivation occurs in ultrasonically irradiated liquids in the frequency range 20 kHz–10 MHz. High-power ultrasound, also known as power ultrasound, in the frequency range 20–10 kHz can cause cavitation. Cavitation, the collapse of microbubbles formed as a result of the ultrasonic sound waves, creates small regions of sufficiently high temperature and pressure to cause the formation of hydroxyl radicals. These strong oxidants provide a mechanism of microbial inactivation during ultrasonic processing. Because these temperature and pressure changes are very localized within the food matrix, there are minimal overall changes to the food.¹²¹  One study reported that ultrasonic cleaning was superior to a cleaning wash with 0.5% KOH or water only. Ultrasonic cleaning was tested against bacterial pathogens known to be recalcitrant due to biofilm formation and also yeasts and molds commonly found in food processing plants. In addition, a case study evaluating ultrasonic application on two cheese molds and milk transportation crates showed a faster turnaround for sanitation and a reduction in the temperature needed for sanitation (below 60 °C).¹²²  Likewise, allergens, such as milk and wheat, were removed more efficiently from the processing line conveyor belts using ultrasonic cleaning compared with a water spray. The improved cleaning technique could reduce the amounts of energy, chemicals and water needed to clean processing lines.¹²³ 

Moreover, ultrasound processing has been explored to create added-value products from food waste. One study assessed the ability of ultrasound-assisted extraction (UAE) to increase the high added-value compound extraction levels from olive oil waste and olive oil production by-products. Polyphenol (and antioxidant) extraction from olive mill wastewater using UAE demonstrated higher yields than conventional filtration methods.¹²⁴  Another study reviewed the potential of ultrasonic treatment for absorbent regeneration, drying and dehydration in foods. Although still inconclusive, ultrasonic treatment for adsorbent regeneration (i.e. active carbon, polymeric resin, etc.) may be more efficient than thermal methods. Pretreatment of products by ultrasonic processing resulted in greatly reduced drying and dehydration times.¹²⁵  Technologies that lead to more efficient processing or cleaning or that permit more effective creation of value-added products all support a more sustainable processing industry by reducing input requirements and potentially providing additional revenue streams.

The most common supercritical fluid used in food processing is carbon dioxide (CO₂). CO₂ has been proven to have a lower environmental impact than solvents such as chlorofluorocarbons. CO₂ is recognized as safe by the FDA and European Food Safety Authority (EFSA) for food processing. The main operating parameters of supercritical fluid extraction (SFE) are temperature, pressure, time and type and percentage of polarity modifiers. SFE of coffee grounds, a by-product of coffee production, has been studied. The oil extracted from the coffee grounds contained 44.5% linoleic acid and 37.5% palmitic acid, and it was estimated, through an economic analysis, that the extracted oil could yield a net income of €21.9 million per year.¹²⁶  Another study evaluated SFE as an alternative process for palm oil extraction to reduce the energy and water required for the intense processing of palm oil in Malaysia. Malaysia produced an estimated 17.7 million tons of palm oil in 2008 and is the second largest producer of palm oil, responsible for more than half of the world production, so even very small energy savings will have a large cumulative effect. While highlighting the benefit of SFE in the palm oil industry, the reviewers noted that there are still problems with the scalability of SFE owing to the current lack of industrial-scale SFE equipment.¹²⁷  SFE research also continues in the field of natural functional ingredients, exploring the potential of adding SFE garlic extracts to sunflower oil. The results showed that the garlic extract allicin, when evaluated against synthetic butylated hydroxytoluene (BHT), performed similarly to BHT in preventing lipid oxidation.¹²⁸ 

Another use of supercritical fluids (SCFs) is as an alternative non-thermal method for pasteurization. Supercritical carbon dioxide (scCO₂) causes microbial inactivation through (1) cell wall rupturing caused by the strong interaction of the fluid with lipids, (2) enzyme deactivation, (3) disruption of the metabolic chain caused by inhibition of decarboxylases by excess CO₂ surrounding the cell and (4) intercellular electrolyte alterations. However, spore inactivation with SCFs is not fully understood, although it appears to be caused by perforation of the spores’ outer layer.¹²⁹  A recent study reported 100% inactivation of both Aspergillus niger and Penicillium simplicissimum spores.¹³⁰  Parameters used in supercritical pasteurization are temperature, pressure, moisture and exposure time. Just as with thermal pasteurization, an increase in temperature results in a reduced processing time.¹³¹  One study looked at the effectives of scCO₂ pasteurization for post-harvest processing (PHP) of oysters. In addition to validating that scCO₂ was equivalent to other FDA-approval PHP treatments of oysters, the results showed that scCO₂ processing was also able to remove or reduce the load of toxic metals, biotoxins and chemical environmental contaminants.¹³²  These contaminants are a concern in oysters along with food-borne pathogens. Other research investigated scCO₂ as a non-thermal alternative to pasteurizing apple cider and apple juice and found that a 5 log reduction in E. coli could be achieved without altering either the flavor or color.¹³³  Hence scCO₂ is a compelling non-thermal alternative.

Membrane separation has been very successful for waste water management and continues to provide innovative separation solutions.^134–136  Membranes have a wide range of applications based on the size range of the material filtered from the product. Membranes are classified based on their application as microfiltration (MF), ultrafiltration (UF), dialysis (D), nanofiltration (NF), reverse osmosis (RO), pervaporation (PV), gas separation (GS) and ion-exchange membranes (IEM). Membranes do not require additives and can perform separations at low temperatures.¹³⁷  In one study, RO separation was shown to result in approximately one-third of the total operating cost compared with evaporative separation.¹³⁸  One example of membranes supporting sustainable food processing can be found in the dairy industry. Whey, once a waste by-product, is now a functional ingredient in high demand. This process not only reduced a waste with a high biological and chemical oxygen demand produced by dairy plants, but also provided another product line for dairy processors. The filtration step also increased the food safety of these products owing to removal of bacteria and spores. Both whey protein concentrate and isolate can be used as functional ingredients, i.e. as emulsifying, gelling and foaming agents.¹³⁹  Other valuable individual whey proteins (immunoglobulins, lactoferrin, lactoperoxidase, bovine serum albumin, α-lactalbumin and β-lactoglobulin) and lactose can be fractionated using membrane technology.

The basis of food-borne illness surveillance systems follows the Food and Agricultural Organization–World Health Organization (FAO–WHO) guidelines, which describe how a surveillance system must first set a baseline of a current illness rates caused by an infection and then map how to reduce the infection rate and impact over time, as shown in Figure 1.14.

Cassini et al. argued that surveillance systems cannot accurately capture and measure the true impact of infectious and food-borne diseases because of two fundamental issues: (1) not everyone who becomes ill will seek care and (2) a failure of communication between healthcare providers and government agencies.¹⁴⁰  Therefore, risk managers must continue to find new ways to identify better both the sources and rates of infection and improve the collection and processing of the information. To address and prevent continuing food-borne illness issues effectively, epidemiological data collection and local risk assessment actives should be combined.¹⁴⁰  Two example programs that are improving the management of epidemiologic data are the Burden of Communicable Disease (BCoDE) database used in the European Union (EU) and the FDA-iRisk database used in the USA. The EU has developed a toolkit for the BCoDE to map the health outcomes (symptomatic levels versus asymptomatic) through representations of the transitions from different infection states. From a probability standpoint, each infection has a certain chance of a particular outcome which is influenced by a person's demographic risk factors. The user's specific input allows an estimation of the impact of food-borne pathogen infections as DALYs.^141,142 

The FDA-iRisk is the equivalent to EU's BCoDE food safety surveillance system. The FDA-iRisk creates models to assess, compare and rank the risks brought by multiple food hazard interventions across the entire food supply chain from “farm to fork.” It addresses risk at (1) the primary production location, (2) processing, (3) retail markets and (4) the consumer. The structural parallel to LCA is obvious (see Section 1.10.2). The model is based on seven components: “the food, the hazard, the population of consumers, a process model (i.e. food production, processing and handling practices that follow the fate of the hazard through the supply chain), consumption patterns in the population, dose–response relationships and burden of disease measures associated with health effects.”³³  The process models use four inputs: (1) the introduction of the hazard, (2) patters of consumption, (3) dose–response relationships and (4) health effects. Chen et al. reported a case study examining L. monocytogenes and Salmonella as microbial hazards using the FDA-iRisk system.³³  Both the BCoDE and FDA-iRisk are modeling programs that can effectively estimate the linked MRA and DALY impact for companies and government officials to prevent, correct and respond to food-borne illness better. This not only improves public heath safety, but also encourages management decisions based on quantitative data.¹⁴³ 

Food safety surveillance systems directly impact our ability to estimate accurately the impact of food-borne illness. It is known that seven pathogens cause 90% of all food-borne illness in the USA: Campylobacter, Clostridium perfringens, E. coli 0157, L. monocytogenes, non-typhoidal Salmonella (NTS), Norovirus and Toxoplasma gondii.¹⁴⁴  The estimated total annual DALYs from these seven food-borne pathogens was 120 000 based on the health impact calculated from years of life loss (YLL), years lost due to disability (YLD) sequelae (a subsequent illness resulting from the infection) and YLD acute. Without the communication and reporting efforts orchestrated between different organizations, it would be impossible to make an accurate estimation of the impacts.

In addition to the surveillance programs growing in the USA and EU, China is continuing to improve its public health awareness and prevention of food-borne illness. In 2009, China created and enacted the Food Safety Law of the People's Republic of China, which was revised in 2015.¹⁴⁵  The regulations led to the National Heath and Family Planning Commission (NHFPC), which works with various government agencies within China's government. Open and fluid communication between agencies is imperative to keep the cases of food-borne illness low within a country of approximately 1.35 billion people. The NHFPC uses food surveillance to provide a foundation and technical support for MRA planning, standard setting and a promulgation of food safety laws. China has reported success of its new food surveillance system by decreasing the number of cases of Coronobacter spp. from 1.1 to 0.35% in infant powdered formula.¹⁴⁶ 

Advances in LCA, particularly as it is implemented in food production and consumption, are ongoing. There are efforts under way to incorporate directly the epidemiology of diet on health and also the beginning of work to permit an integrated accounting of food-borne pathogen impacts on health. In addition to the continued developments in food processing technology, which are the subject of much of the remainder of this book, other technological advances involving food quality hold promise for the continual improvement of the sustainability of the global food supply chain.

The effort to evaluate the entire impact of DALYs of food-borne illness globally was launched in 2006 with the WHO initiative entitled the “Initiative to Estimate the Global Burden of Foodborne Diseases.” Furthermore, the WHO created the Foodborne Disease Burden Epidemiology Reference Group (FERG) to support future DALY studies. They concluded that there are 31 significant international foodborne diseases, that is, diseases with food as the primary vector for infection. DALY is the sum of the number of healthy years of life lost (YLL) due to premature mortality and the number of years lost due to disability (YLD). YLL is calculated by multiplying the number of deaths (D) by the remaining life expectancy at the age at which death occurs in years (E). YLD is calculated by gathering the number of incidents cases, symptom duration and symptom severity, or “disability weight,” ranging from 0=perfect health to 1=death.^147–149  This initiative built upon previous work that the FAO and WHO had conducted by helping organizations define, assess and mitigate food-borne illness risk. One method is through an MRA, which was first defined by the Codex Alimentarius Commission CAC/GL 30-1999, with amendments in 2012 and 2014, as “a scientifically based process consisting of the following steps: (i) hazard identification; (ii) hazard characterization; (iii) exposure assessment; and (iv) risk characterization.”^140,150  Not only can MRA help the food industry continue to improve public heath safety from food-borne illness and disease, it also can allow the industry to allocate limited resources more effectively through the identification of hotspots of pathogenic risk in the supply chain.¹⁴³ 

Figure 1.15 highlights the importance of adequately identifying the risk and shows how risk characterization, risk management and risk communication, while having separate definitions, share overlapping boundaries that must be executed well by management to mitigate a microbiological risk and reduce outbreaks of food-borne illness.¹⁵¹  See Romero-Barrios et al. for an example of modeling quantitative MRA.¹⁵² 

The application of LCA in the context of the food system must be predicated on the understanding that the functional unit for the assessment must always meet certain obligatory characteristics: all legal and regulatory requirements (e.g. nutrient label requirements) and food safety requirements. Food safety is frequently modeled using quantitative MRA, where the probability of illness from exposure to a food-borne pathogen is expressed in DALYs. The observation that LCIA and quantitative MRA both report impacts in terms of DALYs suggests the possibility of combining the two assessment frameworks. Many papers have been published discussing the basis for combining risk assessment (in general) and LCA;^{151,153–158}  however, despite these discussions, there have been relatively few applications in which the two methods have been combined.

There are some broad similarities between risk assessment (and MRA specifically) and LCA: both share the goal of system assessment through quantitative modeling in an effort to support decision makers. Yet there are also some differences that may partially explain the relatively few applications of the combined assessments. Risk assessment is site and scale specific and employs a conservative approach to evaluate safety in the context of an exposure or acceptable threshold level. Lifecycle assessment, although evolving, is frequently site generic, using an approach that estimates the average potential impacts frequently with the goal of identifying hotspots or comparing alternatives. Nonetheless, there have been some efforts to combine the approaches, as outlined in the following sections.

As mentioned previously, with regard to food processing, the first concerns that arise will be food safety and microbiological risk. However, it may be surprising to find that this is rarely addressed in LCA. Failure of food safety protocols ultimately leads to impacts on human health, which is an endpoint impact category in lifecycle impact assessment. For food LCAs that include accounting of human health, it is generally based on exposure to chemicals, including pesticide residues, rather than microbiological risks in the food supply chain.

Villamonte et al. reported that there is reduced human health impact from prepared meals treated with HPP than traditional methods and, although the DALYs from Listeria remain below the cumulative carcinogenic effects of the supply chain, the impact increases with the storage time.¹⁵⁹  This is the only report directly connecting food-borne pathogen MRA with LCA, and it provides a reasonable methodological approach. Other work connecting MRA with LCA has been published, again providing a methodological framework for future improvements.^160,161  The authors evaluated the combined environmental and pathogenic impacts on human health associated with different options for the treatment of sewage sludge and they showed that pathogen exposure can contribute up to 20% of the total human health impact measured in DALYs, with the remaining impact arising from exposure to heavy metals and other chemicals.

There is growing belief that dietary guidelines need to be based on both nutritional and environmental science.^162,163  This is likely driven by the observation that food production consumption is a significant driver of global environmental impact, and the desire to provide sound nutritional recommendations with reduced environmental impact to improve global sustainability. A growing number of studies are starting to consider the benefits and costs of differences in the nutritional qualities of food in LCA. Although the epidemiology of non-communicable diseases associated with dietary choices is relatively well established, these mechanisms have not yet been incorporated into LCA. This is not surprising, as LCA arose from industrial ecology, and only more recently has it expanded to include food production and processing and therefore generally includes only negative impacts. Thus, current LCIA categories, both midpoint and endpoint, quantify negative rather than positive impacts. To include dietary effects in LCA, a new mechanism needs to be established to connect the inventory with the impact/benefit category. A recent study proposed a combined nutritional and environmental LCA framework for diet using published epidemiologic data relating food consumption to reported nutritional health effects (both positive and negative) expressed in DALYs.³⁵  Dietary health impacts, such as microbiological risk assessment and chemical exposure, are all commonly quantified in units of DALY. Since DALY is the impact category metric used in LCA to assess (negative) impacts on human health, in principle it should also be applicable to account for both burdens and benefits associated with the nutritional content of foods. It can be argued that all endpoint impact categories (resource and ecosystem) could be “expanded” in the same way to include benefits derived from human activities. Benefits presented at the midpoint level may be expressed as negative values for the indicator.

As an example, a 2013 study conducted in France, Sweden and The Netherlands modeled the potential outcome of additional calcium intake from dairy foods to reduce the risk of osteoporotic fractures. The economic impact of supplemental calcium was quantified based the cost of additional dairy intake and the cost saving through the prevention of osteoporotic fractures. The study found that low calcium intake (<600 mg) led to a reduction of 6263 DALYs in France, 1246 DALYs in Sweden and 374 DALYs in The Netherlands. The associated total cost that could potentially be saved by the prevention of osteoporotic hip fractures was €100 311 274 for women in France, €23 912 460 for women in Sweden and €5 121 041 for women in The Netherlands.¹⁶⁴  Another LCA study conducted in 2015 evaluated the impact of milk on both the environment and health also measured in DALYs. The study did not focus on food-borne illness but, similarly to the previous study, focused on the health impact of non-communicable disease associated with dietary intake. The study is an excellent model for how a quantitative epidemiology-based study can estimate the synergies and tradeoffs between nutrition and environmental human health impact expressed in DALYs. The new method that the research developed is called Combined Nutritional and Environmental Life Cycle Assessment (CONE-LCA); CONE-LCA measures and compares in parallel the environmental and nutritional effects of food or diets.³⁵ 

As mentioned previously, reduction of food waste is a significant opportunity for improving the sustainability of the food supply chain. Global food waste is estimated at 1.3 MT per year.¹⁶⁵  There are many causes for this loss. In the developing world, much of it is preharvest associated with pest damage or the inability to harvest in time, whereas in the developed world, it is more commonly caused by postproduction losses, and significantly driven by consumer behavior.^{28,29,166–171}  Therefore, although the efforts of the food processing sector to provide new technology that reduces avoidable food losses are important, education to change consumer behavior is equally important.

Figure 1.16 outlines the goals of the US Environmental Protection Agency (EPA) to reduce food waste and improve food recovery. Some source reduction is easily achieved through education because there is documented confusion over date labeling on foods.^29,166,172  Although the one of the goals of the EPA is to use food waste as compost material, studies must be conducted to ensure that the composted food waste does not pose a pathogen risk. For example, a study investigating the recycling of sludge and food waste to reduce gas emissions and improve phosphorus recovery highlighted the difficulty of waste recycling programs. Food waste to be reused in the food system (e.g. as compost) introduces the potential risk of food-borne pathogen contamination to the food supply. All waste recycle programs must ensure that no pathogens enter the food supply during the recycling process.¹⁷³ 

Another by-product of a strong food safety plan is the avoidance of large and costly food recalls, which also contribute to food waste. Table 1.2 shows the number of recalls and the associated pounds of meat recalled in the USA in 2015.

A recall of meat products is especially problematic for the environment because animal products require more water and energy input than protein pulses. At the highest end, beef requires 20 times more water than a pulse protein and accounts for one third of the world's water supply dedicated to animal production.¹⁷⁴  A food recall is ultimately the failure of a food safety plan being either implemented or executed. One of the easiest steps to make the food supply chain more sustainable is to reduce number of recalls due to mislabeling and microbial contamination. One innovative technology that is just emerging is blockchain technology, which is a distributed ledger that records every transaction in a completely traceable and transparent manner. Walmart is currently testing the technology, which should permit, for example, very selective rather than mass removal of contaminated food when a contamination event is identified.^175,176 

Another way to avoid food waste is through improved communication of microbial risks in food. Consumers throw away food that is still safe to eat from a microbial viewpoint over concern for food safety because of a misunderstanding between the meaning of “expiry” and “best by” date information on packaging, which is mandated by food labeling laws. An expiry date is used to describe a date when the packaged food product may begin to become unsafe by microbiological standards and should not be consumed. A “best by” date is a guarantee by the food manufacturer that all nutritive and quality aspects of the product will be maintained until at least that date. However, properly stored food that has passed its “best by” date does not automatically have an increased food-borne illness risk from a microbiological viewpoint. The food industry tries to use expiry dates primarily for more perishable goods whereas less perishable foods normally have “best by” dates. Notwithstanding the food industry efforts, consumers still discard food that is safe owing to confusion about the meaning of these two terms.^29,166,172 

Sensors and monitoring tools are also generating synergy for sustainability when integrated with continuing improvements in technology. A few of the benefits and potential outcomes are as follows:

• Less waste resulting from food safety inspections.

• Improved sampling techniques.

• More monitoring and measurement to improve quality and reduce waste: coupling radiofrequency identification (RFID) and blockchain technology.¹⁷⁶ 

• Wireless sensing/remote monitoring through the Internet of things (IoT).

• Automatic lighting and temperature controls throughout the food supply chain.

• Enabling “big data” analysis to highlight risk allows a sharper focus on food safety.¹⁷⁷ 

• Quicker response times.

• Safety and environmental best practices can be shared quickly across the industry through improvements in communication technologies.

• Focuses of sourcing energy for food processing through renewable energy sources.

• Management focus on water neutrality, in turn lessening the load on water sanitation.

• Multiple hurdle approach to extend the shelf life of foods in lieu of focusing on just one type of extended shelf life technique (e.g. aseptic packing).³¹ 

An example of how technology integration has led to the reduction of food waste and improved food safety can be seen in an experiment conducted in 2013 that mapped and identified high-risk areas for food-borne pathogens on a farm by combining GIS (geographical information systems technology) information and pathogen-specific growth patterns to predict preharvest food safety hazards. Using these predictions, the farmer has a better opportunity to mitigate the hazard before harvest, in turn reducing the risk of illness from food-borne pathogens across the entire food supply chain.¹⁷⁸  In addition, some databases are now using unstructured data collection methods, e.g. social media sites, to help track a food-borne illness.¹⁷⁹  Continuing innovation will improve these tools by empowering more people to report food-borne illness efficiently and ease communication across domestic and international government agencies.

We must encourage students to study interdisciplinary food-related programs, such as food engineering coupled with lifecycle management principles, which can provide new minds to help implement more sustainable technology and management in the food industry. In addition, it is imperative that education includes a dimension for socioeconomic innovation.¹⁸⁰  As disciplinary boundaries blur, our collective ability to communicate and solve more and more complex problems improves. It is important to remember that sustainability, because of its multidimensional nature and myriad tradeoffs, does not always result in immediate economic benefit to an organization. As mentioned earlier, the definition of sustainability adopted by each organization should be understood and communicated across all levels within an organization.³¹ 

Given the global imperative for food security under the combined pressures of climate change, resource scarcity (especially water), growing population and affluence, the opportunities and responsibilities falling on the food processing sector are significant. However, our ability to benchmark and document continuous improvement and adapt new technologies gives reason for optimism in the face of our collective challenges.