The absence of a universally agreed-upon definition for “food systems” is highly recognized within the literature (Von Braun et al., 2021; Canfield et al., 2021; Caron et al., 2018; Recine et al., 2021; HLPE, 2014). Each definition emerges from specific communities of practice and often manifests in descriptive forms rather than quantitative ones. This leads scholars and practitioners to grapple with the dynamic interplay between conceptual clarity and contextual specificity, resulting in diverse interpretations. This concept was broadly presented by Polly Ericksen using a conceptual framework encompassing drivers-activities-outcomes while acknowledging the intricate connections of environmental and socio-economic factors through feedback mechanisms (Ericksen, 2008). The High-Level Panel of Experts on Food Security and Nutrition (HLPE) presented another conceptual framework with an emphasis on diets and nutrition as vital outcomes “A food system gathers all the elements (environment, people, inputs, processes, infrastructures, institutions, etc.) and activities that relate to the production, processing, distribution, preparation, and consumption of food, and the output of these activities, including socio-economic and environmental outcomes” (HLPE, 2017). The Food and Agricultural Organization of the United Nations (FAO) considers food systems to “encompass the entire range of actors and their interlinked value-adding activities involved in the production, aggregation, processing, distribution, consumption, and disposal of food products that originate from agriculture, forestry or fisheries, and parts of the broader economic, societal and natural environments in which they are embedded. The food system is composed of sub-systems (e.g., farming system, waste management system, input supply system, etc.) and interacts with other key systems (e.g., energy system, trade system, health system, etc.).” The conceptual divergence in defining food systems and their key components may pose challenges in evaluating their performance and in guiding policies toward transformative action.

Navigating a fully operational food system transformation is further hindered by varying conceptualizations of sustainability and resilience, and the interplay among their dimensions (Béné et al., 2019a; Holden et al., 2018; Tendall et al., 2015; Van Wassenaer et al., 2021). Sustainability focuses on meeting the current needs of the food system while ensuring its future viability, whereas resilience enhances the system’s capacity to anticipate, absorb, recover from, and adapt to disruptions and shocks. These differing emphases on sustainability and resilience can complicate the measurement of system performance and the establishment of unified metrics, as each dimension requires distinct criteria and indicators to effectively evaluate and balance their respective goals (Marchese et al., 2018; Soubry and Sherren, 2022; Tendall et al., 2015).

Pisano (2012) and Van Wassenaer et al. (2021) present resilience as “The ability of a system to absorb disturbances and still retain its basic function and structure.” However, Béné and Doyen (2018) argue that the concept of resilience should not be limited to the system’s recovery, but rather extend to include systematic responses. Meanwhile, Tendall et al. (2015) present it as follows: “To build humankind’s capacity to attain today’s goals and meet their needs without compromising the future capacity to deliver the same.” Another definition was presented in New Zealand’s 2019 National Disaster Resilience Strategy that emphasizes long-term adaptation and provides a holistic framework to bridge the gap between reactive and proactive approaches: “The ability to anticipate and resist the effects of a disruptive event, minimize adverse impacts, respond effectively post-event, maintain or recover functionality, and adapt in a way that allows for learning and thriving.”

Similarly, Chrisandina et al. (2022) note that from a process systems engineering perspective, supply chain resilience can be defined as “The capability of a system to anticipate, absorb, adapt and recover in the face of a disruption”. Therefore, resilience can be understood as the ability to deal with unexpected deviations from nominal operation. This characteristic is similarly essential for sustainable system solution generation since these potential uncertain deviations can render process systems inherently environmentally and socially harmful (Di Martino et al., 2023). In essence, if systems are not sufficiently resilient, they cannot be sustainable.

Narrowing down these concepts for food systems, FAO (2018) defines a sustainable food system as “a system that delivers food security and nutrition for all in such a way that the economic, social and environmental bases to generate food security and nutrition for future generations are not compromised.” Tendall et al. (2015) defines a resilient food syst as the “Capacity over time of a food system and its units at multiple levels, to provide sufficient, appropriate and accessible food to all, in the face of various and even unforeseen disturbances.” The resilience of food systems requires a consolidated conceptualization (Van Wassenaer et al., 2021). As a result, grasping the full picture of what a resilient sustainable food system remains challenging and therefore, more efforts should be made to compose it systematically.

Sustainability and resilience emerge in their mutual focus on maintaining the system’s ability to function over the long term. Both concepts consider social, environmental, and economic dimensions and recognize the importance of adapting to changes to support future generations. The integration of metrics assessing social and environmental impacts is common to both, illustrating that they both engage with holistic evaluations of food system performance (Figure 1). Summary of divergence in concepts becomes apparent when examining their core principles and approaches to change and adaptation. Sustainability primarily aims at proactive measures, emphasizing resource conservation, ecological balance, and intergenerational equity. It focuses on preventing environmental degradation and ensuring that current practices do not deplete resources for the future. In contrast, resilience emphasizes the system’s ability to react to disturbances and adapt to changes. It incorporates both proactive and reactive measures, focusing on how systems can recover from and adapt to disruptions, whether anticipated or unforeseen. The metrics for resilience often include adaptive capacity, flexibility, and recovery speed, which contrast with sustainability’s focus on long-term ecological balance and equity. In policy and practical application, these divergences can lead to challenges. While sustainability-oriented strategies may prioritize maintaining ecological integrity and long-term resource use, resilience-oriented policies might focus more on building infrastructure and capabilities that allow for rapid adaptation and recovery. This difference in focus may lead to conflicts or trade-offs when balancing sustainability’s long-term preventive measures with resilience’s immediate responsive needs.

We conducted a query for indexed peer-reviewed papers (articles and reviews), in English language between 2020 and 2022 in the Scopus database. This search targeted specific keywords within the abstract, title, and keywords fields. Scopus was selected as a database because of its extensive collection of journals across fields, although it tends to favor English-language journals with a high coverage of natural science, engineering, and biomedical journals than social sciences, arts, and humanities (Mongeon and Paul-Hus, 2016). It is important to note that this language restriction might introduce an implicit bias in the identified literature toward countries and regions speaking these languages. In tandem, through a snowballing method, the cited work of non-governmental and inter-governmental research and policy organizations in the identified research articles was reviewed and included in the analysis. Through this effort, the authors aimed to ensure that relevant grey literature has been considered in this review. The research team implemented an iterative review process, systematically discussing and refining coding decisions and the selection of grey literature to minimize subjective bias and enhance the methodological rigor of the process.

To carry out this search, the following keywords were used: “food systems,” “sustainability,” “resilience,” “metrics,” and “indicators” and the search process was narrowed down to the areas of Environmental Science and Agricultural and Biological Sciences as sub-areas to focus on listed in the Scopus database. Therefore, the resulting search line in Scopus was:

(TITLE-ABS-KEY (food AND systems, AND sustainability, OR resilience, AND metrics, OR indicators OR sustainable OR resilient OR agrifood AND system) AND PUBYEAR >2019 AND PUBYEAR <2023) AND ((indicators)) AND (food AND system) AND (LIMIT-TO (DOCTYPE, “ar”)) AND (LIMIT-TO (SUBJAREA, “ENVI”) OR LIMIT-TO (SUBJAREA, “AGRI”))

The outcome of this search was 1,038 documents, however, only 112 articles matched the objectives of this study by applying both inclusion criteria:

Articles explicitly providing a conceptual framework for quantifying sustainability and resilience in the food system.

Articles were excluded if they:

Did not explicitly provide a conceptual framework for quantifying sustainability and resilience in food systems.

The data collection approach was inspired by Albrecht et al. (2018) and Mittal et al. (2022). As a first step, the shortlisted papers first screened to extract information related to:

The authors: (Names and lead author’s institution).

The second step of data collection included extracting the following information:

The metrics of sustainability and/or resilience of food systems:

To map the regional and scale-based distribution of case studies that have implemented sustainability and resilience quantification, we compiled the following information:

Case study: Is there a specific case study? Or does the paper only present a conceptual model or approach?

The information was initially coded into categories to analyze the selected papers by journal and year, identify the institutions focusing on this topic, determine the proportion of papers addressing sustainability and/or resilience, and examine the common objectives stated by the authors. The identified quantification criteria were coded under sustainability or resilience categories, then compared to reveal similarities, differences, and gaps. Additionally, they were assessed to determine if they were cross-sectional or longitudinal. Finally, the assessment of quantifying sustainability and resilience in food systems considered the implementation scale and region. For example, research focused on local or regional scales often emphasizes community-driven initiatives, while studies at national or global scales tend to focus on policy frameworks, institutional coordination, and large-scale interventions.

The data analysis process involved an initial categorization of the collected data, with a focus on key aspects such as:

Distribution of papers across journals and years,

Next, the identified quantification criteria were coded under either sustainability or resilience. This categorization also included an assessment of whether the studies were cross-sectional or longitudinal. In the final phase, case studies quantifying sustainability and resilience in food systems were evaluated by recognizing trends and patterns, including frequency of studies across regions and scales. The results are presented in the following section, emphasizing opportunities for future research on food systems transformation and providing actionable recommendations based on the analysis. Given the inherent context-specific nature of studies and their associated food supply chains at assessments at different scales, divergence among quantification metrics for food system sustainability and resilience is expected. While this observation was no surprise, this work makes an important contribution by mapping and cataloguing key quantification methodologies and metrics across diverse research focuses and scales, a gap in the current literature. By providing a structured glossary of these approaches, this work serves as a resource for researchers seeking to align their localized approaches with established metrics, enabling comparability and facilitating the transfer of insights between studies.

To address the first research question of this paper—To what extent do the definitions of “sustainability” and “resilience” in food systems literature align, and what quantification methods exist to evaluate them?—this section examines the key features of the literature. Around 10% of the screened literature (112 of 1,038 articles) included metrics for quantifying sustainability and/or resilience in food systems. Our review reveals that most efforts are aimed at conceptually defining sustainability and/or resilience within the food system, rather than specifically providing quantifiable metrics. These efforts focus on offering descriptive visualization of sustainable and resilient food systems, emphasizing themes related to the development and governance of existing food systems. Thus, a key insight from the reviewed literature is that the development of explicit and reproducible metrics for quantifying sustainability and resilience in food systems remains in its early stages, with a notable absence of a comprehensive systems perspective.

Figure 3 shows that the overall 112 studies are divided as follows:

The number of articles exclusively addressing sustainability is 102.

The reviewed literature also highlights a dispersed global interest in this research topic, as evidenced by the distribution of lead author institutions. Italy notably stands out with 11 institutions engaged in quantification, indicating a significant interest within the research landscape. Figure 4 illustrates the worldwide distribution of lead author institutions for further insight.

In addressing the inherent complexity of food systems, we explored the thematic breadth of the literature and scrutinized each article’s objectives and their alignment to our overarching theme. Based on the author’s analysis of the 112 screened studies, seven categories emerged, with two key recurring objectives under each. Notably, these objectives span across various facets of the food system, tightly linked with stressors such as climate change resources like biodiversity and land. While the majority of articles focused on enhancing policy and governance frameworks within the current landscape, there was a noticeable absence of literature reflecting multi-sectoral collaboration and a systemic perspective. Figure 5 visually presents the key themes and objectives elucidated in the reviewed literature.

To address the first part of the second research question—What types of metrics for evaluating sustainability and resilience in food systems exist in the literature —this section explores the methods and metrics identified in the literature. It examines how these metrics are applied to quantify sustainability and resilience in food systems. Table 1 presents the categories of metrics reported by the authors of the 112 screened papers. The “example metrics” listed in the second column were selected to illustrate representative examples, along with their corresponding references. These categorizations (e.g., environmental, economic, social, governance, health, cultural, farming practices, circularity, flexibility, and security) reflect the multi-dimensional nature of food systems and align with the definitions and dimensions of sustainability and resilience identified in the literature. They aim to bridge conceptual gaps and provide a structured framework for quantifying food system performance, consistent with the broader definitions and frameworks discussed in Section 2. Each category addresses a specific dimension critical to assessing sustainability and resilience:

Environmental: Metrics including emissions, biodiversity, and land use quantify the ecological footprint of food systems, aligning with sustainability’s focus on reducing environmental impacts.

In Table 1, we can see that Mouratiadou et al. (2021) for example utilize 1,128 metrics, including those related to biodiversity (76), productivity (70), food and nutrition (70), and soil characterization (61). While this level of detail allows for a nuanced analysis, it also hinders the application of the proposed methodology to other fields due to its inherent complexity. It is not uncommon among the identified studies that more than 100 indicators are utilized (Eggert et al., 2021; Heredia-R et al., 2022; Mackenzie and Davies, 2022). Some of these indicators are highly specialized which further hinders the applicability and transfer to other scenarios of performed studies. This also renders the performed analysis of this work challenging.

The categories included in the table were directly derived from the papers, not based on the authors’ interpretation or judgment. This diversity can be attributed to the varying concepts used in food system transformation, the diverse backgrounds of researchers, and the complex nature of food systems. Despite this, there is a noticeable lack of metrics specifically designed to assess food system vulnerability and associated risks. Water is a prominently recognized category, underscoring its critical role in transformation; however, related metrics could further include aspects of recycling and treatment. We can also appreciate a divergence between theoretical concepts or definitions and practical implementation. The open literature converged that sustainability consists of the three pillars of environmental, economic and social considerations. However, all three aspects are only considered in 42 out of 112 studies, all exclusively in the sustainability study category. This underlines that not only resilience can be a nebulous and vague concept, but sustainability as well, in the sense that not all dimensions of it are commonly explicitly investigated.

Furthermore, the plethora of utilized metrics and their relative occurrence can be visualized in Figure 6. It is not surprising that environmental and economic metrics are the most often incorporated metrics. However, after these two metrics, a drop in utilization frequency can be observed to around 39 to 48% for social, water and energy metrics. The next section constitutes around 20 to 30% of studies which include health, management, security, cultural and governance metrics. Next, farming practice, biological, circularity, and flexibility metrics are classified as contributors in the 10 to 20% range. Lastly, chemical and physical metrics are only used scarcely, i.e., in 2 to 3% of studies. Further, it is surprising that a lot of studies analyze environmental implications, but do not consider extensions to energy and water systems. This means that the full environmental impact can inherently not be analyzed by these approaches (Di Martino et al., 2023). Security and flexibility together can be understood as a prerequisite of a resilient process system (Chrisandina et al., 2022). Interestingly, security and flexibility are only evaluated together in 5 studies (Sustainability and Resilience: 1, Sustainabillity: 1, Resilience: 3). Security alone is analyzed in 24, and Flexibility alone in 4 sustainability studies. This indicates that indeed preliminary aspects of resilience are considered in 27 sustainability studies without explicitly stating it, which increases the initially observed penetration of resilience considerations in sustainability studies from 3 to 26%. However, in all studies on resilience and resilience combined with sustainability, no coherent metric or concept of resilience has been introduced.

Figures 7, 8 evaluate the utilized metrics category per location and scale to address the second part of the second research question—How do these metrics vary across regions (Asia, Africa, Europe, etc.) and across scales (household, farm, urban, regional, global)?.No clear convergence of used metrics per scale or location can be observed. This is a reflection of the absence of best practices or uniform framework for selecting quantification metrics per scale, location and research focus. It is observed that metrics related to the environmental category are present in all locations. Additionally, the metrics in Europe cover all the categories, although they do not have equal representation. We can recognize a more uniform representation of metrics linked to environmental, economy, water and energy in North America.

To further highlight this discussion, the employed quantification methods of all studies are summarized in Table 2 and analyzed next. Similar to the quantification metrics, the quantification methodologies are categorized based on their analytical framework and methodological approach. In this instance, we could summarize all methods under the following: sustainability impact assessment, life cycle analysis (LCA), Food-Energy-Water Nexus (FEWN) approach, Emergy synthesis, resilience assessment, or resilience and sustainability assessment. Most of the identified studies utilize historic data and tailored indices for various applications. Additionally, resource balancing and accounting across system boundaries are regularly employed as well. Less often, optimization approaches of identified indicators are leveraged. According to the selected topic, these approaches have been categorized as the respective assessment strategy. However, in 92 out of 112 studies, the presented theoretical concepts have been translated into a real-world impact which is an outstanding result within an FEWN context (Di Martino et al., 2022).

Figure 9 further evaluates the selected quantification methods according to their employed scale and location under investigation, respectively. Interestingly, LCA and FEWN approaches are regularly employed besides sustainability impact assessment metrics followed by Energy Synthesis, resilience assessment and lastly resilience and sustainability assessment methods. Sustainability impact assessment is utilized in the majority of studies due to the strong potential of this approach to tailor it to specific case studies based on the introduction of specialized metrics or balances. In turn, it directly follows that one sustainability impact assessment strategy is potentially completely different from another one on the basis of the selection of fundamentally different performance metrics. This ties back to the identified absence of good practices for sustainability and resiliency assessment in food systems. This is further illustrated by the fact that there does not appear to be a clear utilization of a single quantification strategy for a single scale or location. In fact, the opposite is the case: each quantification methodology is utilized for a variety of scales and locations.

Figure 10 illustrates the metrics utilized for each quantification methodology. Overall, this figure underlines that across the quantification methods there is no clear convergence in utilized metrics, the only exception being Emergy synthesis which focuses mostly on environmental, economic, and energy metrics. Flexibility and security metrics are most often significant contributing metrics for FEWN, resilience assessment and resilience and sustainability assessment strategies. These two metrics can be understood as preliminary resiliency quantification metrics, which indicates that the FEWN can be leveraged as a powerful strategy to assess resiliency, besides the respective assessment strategies. On the other hand, LCA together with sustainability impact assessment employ regularly environmental, economic, and social metrics, i.e., all three pillars of sustainability, putting these approaches at the forefront of sustainability evaluation. One shortcoming of the FEWN and Emergy synthesis strategies is the missing incorporation of social aspects into the methodology. This metric, however, is regularly utilized in resiliency and resiliency and sustainability strategies. Interestingly, no energy or water metrics are used in resilience and sustainability assessment strategies. Similarly, no implications of system solutions to the water system are incorporated in Energy Synthesis. This highlights the missing interconnection of food systems to other resource supply systems for a holistic sustainability and resiliency evaluation outside FEWN strategies. Overall, Figure 10 underlines the complexity and the potential multiplicity of food system sustainability and resilience. In addition, a divergence, i.e., missing consistency, among sustainability and resiliency assessment strategies could be identified. Therefore, once again, we can identify the absence of defined good practices for the assessment of resiliency and sustainability of food systems which can facilitate to close these identified gaps.

To address the third research question — What is the regional and scale-based distribution of case studies that have adopted sustainability and resilience quantification in food systems, and what unique features do these case studies exhibit? —this section highlights the global scope of the analyzed literature. The case studies within the examined literature spanned over 121 countries worldwide, underscoring the global significance and collective efforts dedicated to transformation initiatives. As previously mentioned, 92 publications present adoption strategies of suggested frameworks, i.e., translate their work into real-world action. In turn, the remaining 20 publications focus on pushing the boundaries of theoretical frameworks, such as developing novel agri-food supply chain mapping strategies (Ivo de Carvalho et al., 2022) or a multi-objective programming approach for the agricultural FEWN in a random environment (Li et al., 2019). The following section focuses on all publications presenting adoption strategies to analyze research advances of sustainability and resilience in food system practices.

The distribution of resilience and sustainability practices of food supply systems is visualized in Figure 11 (left). Zooming into the Mediterranean Basin, a notable concentration of case studies emerges in Spain, France, Italy, Greece, and Turkey. These countries are considered a part of Europe’s breadbasket and thus, these endeavors are interconnected with the continuous initiatives led by the European Commission to transform food systems and achieve food security, notably as part of the Green Deal and Farm to Fork Strategy. However, a scarcity of studies is prominent in the rest of the Mediterranean Basin, even though it is a recognized hotspot for climate change, confronting significant water scarcity challenges and agriculture is considered a major contribution to the GDP of each country.

These observations can be further analyzed by examining the frequency of case studies across different continents, as illustrated in Figure 10 (right). This distribution highlights the locations investigated, with Australia and Antarctica being the only ones not covered. Interestingly, global analyses are relatively rare, which is surprising given the prominence of global supply chains and their disruption during the COVID-19 pandemic. Europe is the most studied continent in terms of the transformation fronts, which correlates with the location of the institutions researching these topics the most (Italy). A multi-continental investigation, i.e., Asia and Europe together (Eurasia), has only been studied in two publications. Indeed, only North and South America separately and not together have been studied, underlining the scarcity of multi- or inter-continental studies. Therefore, we can postulate that a multi-continental analysis, i.e., North and South America together, Africa and Asia, or around the Mediterranean Sea (Asia, Africa, and Europe) should be investigated in more detail. Indeed, water bodies such as the Mediterranean Sea, Red Sea (Africa – Asia), or Gulf of Mexico and Caribbean Sea (North America – South America) can potentially facilitate these future investigations as a potential basin for food systems sustainability and resiliency across continents.

These previous observations regarding the location under investigation are mirrored in the scales of the identified studies, as shown in Figure 12. The farm level is the most employed scale, whereas the household level is the least utilized scale. Intuitively, this is conclusive since the focus of the identified literature is food system resilience and sustainability. On the contrary the scarce investigation of supply chains and the utilization of global scales raises concerns post COVID-19 and can be identified as a significant research gap.

Figure 13 analyzes sustainability studies based on the identified scale and location of investigation. For each location, various scales have been utilized: multi-continent studies primarily compare urban food supply systems, while global investigations focus on supply chains, farm-level comparisons, and general global scales. Notably, in Europe and South America, the farm level is the most investigated scale, while in Asia, the regional scale is predominant. Africa shows an equal distribution across national, regional, local, and farm levels, and North America features a balanced focus on supply chain, farm, and crop levels. This variability underscores a shifting focus depending on the location of case studies, potentially influenced by the cultural backgrounds of research institutions and the evolving research fields addressing these challenges. This variability could signify differing regional priorities and research needs. It might reflect varying levels of infrastructure, resources, or policy focus in different regions. For instance, regions with advanced agricultural practices might focus more on farm-level studies, while those with significant supply chain issues might prioritize supply chain analyses. Additionally, it could indicate diverse responses to local challenges and opportunities in food system sustainability.

Similar to Figures 13, 14 (left) depicts the identified scales and locations for resiliency studies. One key observation is the reduction in number of studies and diversity of scales from changing study focus from sustainability to resilience. For Asia an equal share of farm and crop level studies can be seen. In turn, Europe has been analyzed more from a farm level than a national perspective, as was the case for sustainability studies, and North America has only been studied in terms of regional boundaries. It is surprising that for resilience studies of food systems no global, multi-continental or supply chain scales have been used. Lastly, Figure 14 (right) summarizes the scales and locations of resiliency and sustainability studies. As before, the number of studies and diversity of scales further reduces by combining a resilience and sustainability focus. Both concepts together have only been investigated based on a national or farm level scale. The reduction in scale diversity and the focus on national or farm-level studies for resilience and combined resilience-sustainability research may reveal underlying methodological constraints or data limitations. Additionally, the notable absence of global and supply chain analyses highlights a significant research gap that could impact the development of comprehensive strategies for enhancing food system resilience and sustainability on a broader scale. Addressing these gaps through interdisciplinary approaches and improved data collection could provide a more holistic understanding and better inform policy and practice.