Food cold chain management: what we know and what we deserve Open Access

Nowadays, cold chain (CC) management plays a significant part in modern global perishable industries. Although multiple definitions have been offered in the literature, it can be described as “the process of planning, implementing and controlling the flow and storage of perishable goods, related services and information to enhance customer value to ensure low costs” (Singh et al., 2018, p. 532). Perishable products require a precise temperature-controlled atmosphere along the entire supply chain (SC), from production to consumer touchpoints. This requirement, commonly denoted as “CC”, refers to a post-production SC for perishable and temperature-sensitive goods, and is specifically designed to keep these products in a conditioned environment (i.e., within optimal temperature and humidity range) to guarantee product safety, preserve value and maximize commercial potential (Salin and Nayga, 2003; Joshi et al., 2009; Rodrigue, 2014). The perishability of goods is a key point of this definition. In other words, refrigerated transportation and storage are two fundamental aspects to prevent deterioration of the product quality (James and James, 2010). Transport and storage of non-perishable goods requiring controlled and low-temperature conditions (e.g. art objects) are usually not considered a CC (Heap, 2006). Therefore, CC management can be seen as a specific implementation of SC management for perishable goods (Bogataj et al., 2005) adding characteristic features and activities to existing SCs (Kuo and Chen, 2010).

The CC includes a variety of perishable products, namely fresh agricultural products, frozen food, seafood, chemicals and pharmaceutical drugs and photographic film (Allied Market Research, 2019). Regarding end-use, the CC market is classified into five main categories, namely, fruits and vegetables, bakery and confectionary, dairy and frozen desserts, fish and seafood, drug and pharmaceuticals (Liu et al., 2020). This classification integrates two main CC fields, namely, food cold chain (FCC) and pharmaceutical cold chain (PCC) (Herjolfsson, 2019). Although both fields have very diverse packaging requirements (Brenner, 2015), product quality, packaging integrity and punctuality has been worse in FCC as compared to PCC due to better transportation schedules and automatization for pharmaceutical products. In addition, food demand is subject to frequent modifications due to changing customers’ tastes, preferences and lifestyles (Aramyan et al., 2007), which is not the case for pharmaceutical products. Similarly, both the FCC and pharmaceutical CC have different regulatory requirements for handling, sorting, and distributing the related products (Ruiz-Garcia and Lunadei, 2010). In fact, pharmaceutical SC disruptions are limited to deviations from production plants, whereas food SC disruptions refer to lower performance overall (Brenner, 2015).

Within the overarching research area of CC management, the FCC is a rapidly growing research field (Bremer, 2018; Göransson et al., 2018) because it prevents food waste, which has numerous detrimental effects. In fact, food waste has not merely economic implications but also social and environmental ones. The world population is anticipated to reach 8.5 billion by the end of 2030 and 9.5 billion by 2050 (UN DESA, 2015). Currently, even with a world population of 7.7 billion (Patierno et al., 2019), some individuals remain deprived of food. Consequently, above 820 million people sleep with an empty stomach every night worldwide (FAO, 2019a). Hence, meeting future food demand seems impossible and pressures food SC (Shashi et al., 2018). Surprisingly, 30% of the total food produced is either lost or wasted globally every year. This is equivalent to 1.3 billion tonnes of food (FAO, 2017a; Özbük and Coşkun, 2019), US$1tn in economic costs, approximately US$700bn in environmental costs, and approximately US$ 900bn in social costs (FAO, 2017a). Besides, this enormous food wastage is a significant contributor to global greenhouse gas emissions and diminishes the productive output of food systems (FAO, 2017b). The carbon footprint of food wastage is about 3.3 billion tonnes of CO₂ (FAO, 2019b). By controlling this amount of food wastage, it would be possible to feed almost four times the number of hungry people with the food wasted worldwide every year (We Eat Responsibly, 2019). In fact, the deterioration of perishable food products can easily cause adverse effects on consumers’ health, as well as product price and availability. The raw and ready-to-eat products cause the majority of foodborne illness and uncooked products are the source of cross-contamination (Reed, 2005). Hence, reducing global food waste at production, SC, retail and consumer levels is reflected as one of the goals of the 2030 sustainable development agenda to end hunger and negative impact of food wastage (FAO, 2017a, 2017b; FAO, 2019a).

Based on the pivotal role of the FCC in the food sector, theories and practices have evolved to improve the body of knowledge on this topic (Minner and Transchel, 2010; Aung and Chang, 2014a; Qi et al., 2014; Chaudhuri et al., 2018). Consequently, research on FCC has been growing rapidly by incorporating the attributes of a well-defined scientific domain. Two leading journals even published special issues on FCC in 2018 (Carson and East, 2018; Tsai and Pawar, 2018). Moreover, there are many international FCC collaborations and research groups worldwide fostering the growth rate of this rapidly growing research field.

Given the rise in empirical studies on FCC (Joshi et al., 2010, 2012; Zanoni and Zavanella, 2012; Ucar and Ozcelik, 2013; Shabani et al., 2015; Ali et al., 2018; Gligor et al., 2018), some scholars reviewed the FCC literature from different perspectives (James et al., 2006; Raab et al., 2011; Defraeye et al., 2015; Shashi et al., 2016; Mercier et al., 2017; Shashi et al., 2018; Chaudhuri et al., 2018). These works have provided insights into the field through structured review and classification into future research avenues. James et al. (2006) presented an overview of the food transportation system. Raab et al. (2011) identified and compared the already existing and novel temperature monitoring solutions in the meat SC. Defraeye et al. (2015) summarized recent articles on fresh food package solutions to improve CC performance. Mercier et al. (2017) reviewed the current status of commercial CC. Chaudhuri et al. (2018) identified the multiple types of data that can be collected and analyzed by CC practitioners. Shashi et al. (2018) classified FCC literature into four main parts (i.e. factors affecting negatively FCC, popular methods used for FCC performance evaluation, performance measurement indicators and FCC sustainability concerns) and highlighted the associated research gaps.

The literature lacks, therefore, a study that summarizes holistically FCC research advancements and trends for the benefit of multiple stakeholders. Although extant research has provided deep insights into FCC by mobilizing a great variety of disciplines, theoretical frameworks, methodologies, perspectives and research paradigms, this process has led to a gradual fragmentation of FCC scholarship into several sub-areas, as presented above. The lack of a comprehensive overview of the FCC domain has therefore appeared. This is problematic as it prevents researchers, practitioners and policy-makers alike from navigating through the complex domain of FCC and find relevant knowledge to solve key issues, find meaningful answer to theoretical or applied problems or simply understand the evolution of the field. These issues may relate to the absence of past reviews using complementary analyses such as bibliometric and/or network analyses. These types of analyses are very worthwhile in a research field because they identify both established and emerging areas of research (Fahimnia et al., 2015). In addition, they provide both scholars and practitioners alike with a bird’s eye on the state of a research field in terms of authors, countries but also topics and areas of research (Mishra et al., 2018). They also suggest emerging clusters while fostering researchers to collaborate and further expand the current knowledge in the research field (Chen et al., 2010).

Scholars widely acknowledged the imperative role of classifying the research published on an exponentially-growing research field to facilitate researchers and practitioners in attaining a deeper understanding of the field (Merigó et al., 2016; Blanco-Mesa et al., 2017; Nunen et al., 2018). The application of both bibliometric analysis and network analysis may enable to achieve this aim. Previously, researchers also attempted to conduct bibliometric analysis and network analysis in the SC management domain. In this regard, researchers identify and illustrate major research themes associated with information sharing in SCs (Colicchia et al., 2019), green SC (Fahimnia et al., 2015), low carbon SC management (Shaharudin et al., 2019), reverse logistics research (Wang et al. (2017), SC disruption (Xu et al., 2020) and corporate social responsibility for SC management (Feng et al., 2017).

Therefore, to provide a structured and encompassing overview of the FCC research field, this paper provides:

• A review of the literature of FCC dating back to 1995.

• A robust insight into the research field by using bibliometric and network analyses applied to a total of 1,189 publications, to identify key contributing authors, countries, institutions and journals in the field.

• A comparison of the most influential contributions according to the citations.

• Research clusters on articles and authors within the FCC research field.

Therefore, this paper tends to bridge the aforementioned research gap concerning the lack of a comprehensive overview of FCC using a robust analytical approach, by offering the broad qualitative and quantitative insights of FCC research through bibliometric and network analysis methods. Therefore, this review is expected to make considerable contributions to the extant literature for the following reasons:

• It offers a wide overview of the FCC literature.

• It provides significant information for future researchers summarizing the most significant contributions, the most cited outlets (as represented by journals), the centers of excellence (as represented by institutions) and the most prolific and cited authors in this field.

• It identifies the research scopes, research methodologies and empirical issues adopted by this stream of research.

In addition, this study identifies new research streams and perspectives for further investigation that warrant special attention and offers valuable insights for future research.

The remainder of the paper is structured as follows: after this introduction, Section 2 introduces bibliometric and network analysis methods, material collection and selection. Sections 3 and 4 present and discuss the performance assessment and the science mapping results, respectively. Finally, the conclusions and implications present the contributions to theory, managerial practice, and limitations of the study in Section 5.

Bibliometric analysis is a powerful and effective technique used in many disciplines to systematically capture the growth trends, characteristics and advancements in academic literature related to a specific field or topic (Ertz and Leblanc-Proulx, 2018). It offers a structured macroscopic overview of foundational and impactful literature in terms of prolific authors, impactful publications, as well as leading journals, institutions and countries (Merigó et al., 2016; Gaviria-Marin et al., 2019). In this paper, a bibliometric analysis combines research performance assessment and science mapping approaches (Noyons et al., 1999; Cobo et al., 2011).

Research performance assessment means “counting citations of specific papers, for instance, paper pb1 is cited three times (by pa1, pa2, and pa3)” (van Raan, 2014, p. 18). Therefore, citation analysis for research performance includes a wide range of techniques to analyze bibliometric data including keyword frequency analysis, citation analysis, and counting articles by countries, universities, authors and journals (Thelwall, 2008). Research performance assessment will measure, quantitatively and qualitatively (Alcaide-Muñozet al., 2017), the contribution of FCC, as well as the influence of the particular research themes in the FCC research field, with the aim to grasp the dynamics over multiple years of the evolution of FCC research and identifying the sub-fields that are most productive, prominent and impactful.

As a second procedure of bibliometric analysis, science mapping determines how different authors are related to each other (Small, 1999). This methodology shows specifically the structural and dynamic aspects of the scientific domain under investigation (Börner et al., 2003; Cobo et al., 2012). Science mapping uncovers specific patterns hidden in the mass of published knowledge and assists the researcher in interpreting these patterns (van Raan, 2014). Science mapping will thus provide a novel perspective on FCC by revealing the scientific frontiers and dynamic structures of the research field with visualization methods.

In this study, we used Thomson Reuters’ Web of Science (WoS) database as a scientific search engine to retrieve bibliometric data on FCC. More specifically, this study uses the WoS Core Collection. The choice of WoS over other databases is motivated by several reasons. First, WoS is considered as one of the leading databases worldwide, because it includes a broad range of publications from different disciplines and research areas, with more than 15,000 high-quality journals and 50,000,000 articles, divided into 251 research categories, and 150 thematic research areas (Gaviria-Marin et al., 2019). Second, although other databases (e.g. Scopus) are more comprehensive, WoS is more selective as it includes only journals indexed by the International Scientific Indexing (ISI) (Yong-Hak, 2013). Therefore, the journals indexed in WoS have their impact scores in the Journal Citation Report (JCR), also run by Thomson Reuters, creating synergies between both tools. It is worth mentioning that while WoS is selective with regards to the quality of outlets, the selection does not regard specific topic areas so that even emerging topics are included. Bibliometric researchers consider WoS to be a relevant database because it provides a set of metadata that is essential for this type of analysis a relevant database for conducting interdisciplinary literature review (Fetscherin and Usunier, 2012, p. 735). In fact, it provides a set of metadata that is essential for this type of analysis, including abstracts, references, number of citations, research collaborations, lists of authors, institutions, countries and the journal impact factor (Carvalho et al., 2013; Gaviria-Marin et al., 2019). WoS has complete and consistently formatted citation information for its entries (Trujillo and Long, 2018). Also, it provides a unique feature of citation counts, which allows the relative importance of articles out of a large pool to be qualified through the use of an objective measure of influence. The systematic quantitative literature reviews and bibliometric studies in the social sciences field – and somehow also those related to research streams on food SC – typically use one database: either WoS (Handayati et al., 2015; Jose and Shanmugam, 2019; Óskarsdóttir and Oddsson, 2019; Özbük and Coşkun, 2019) or Scopus (Wahyuni et al., 2019) because data homogenization issues emerge when multiple different databases are deployed (Mariani and Borghi, 2019). Dwivedi et al. (2011, p. 45) also argued that “restricting the search activities to a single publication database removed many of the potential problems of duplication inherent in the use of multiple data sources.” Besides, even authors of literature review and bibliometric studies on research areas that have emerged in the past decade (e.g. blockchain, internet of things, big data analytics) and thus with limited influence, considered WoS and recommended the use of WoS database (Choi et al., 2020; Kamble et al., 2020; Nakhodchi et al., 2020; Prieto-Sandoval et al., 2018). Finally, we perform co-citation analysis using the references cited within sample articles retrieved from WoS. This facilitates the identification of additional relevant literature and scholar communities that may be overlooked in standard approaches to literature search (Trujillo and Long, 2018). Accordingly, in our context, co-citation analysis facilitated the identification of additional relevant literature, not included in WoS database but in other databases (e.g. Scopus, Google Scholar and EBSCO). This enabled us to draw the overall main research topic and trends in the current FCC literature, as well as future research suggestions. Collectively, these reasons suggest that WoS provides more reliable and standardized records than other databases (Falagas et al., 2008). In summary, since WoS represents “the most influential and highest quality journals from a broad variety of disciplines (Tian et al., 2018, p. 150) and constitutes the most suitable database for bibliometric analysis (Gaviria-Marin et al., 2019),” we focused on the WoS database.

A search string of keywords was initially used to search publications between 1985 (first year available in WoS) and 2019. However, through the paper retrieval in WoS, no paper seemed to have been published before 1995 using those keywords. Therefore, we chose 1995 as the starting date instead of 1985, while 2019 as the closing year remained unchanged. This structure accommodates a broad range of search terms to select publications (Tian et al., 2018). The terms (“cold” OR “perishable” OR “fresh” OR “temperature” OR “refrigerated” OR “frozen”) AND (“logistics” OR “supply chain*”) AND (“food*” OR “fish*” OR “meat*” OR “milk*” OR “seafood*” OR “dair*” OR “fruit*” OR “vegetable*” OR “ice cream*” OR “cheese*” OR “butter*” OR “fresh pasta*” OR “egg*” OR “yogurt*”) were used as keywords to retrieve publications that included the terms in the title, abstract or keywords. In addition, both singular and plural relevant terms are combined using asterisk (*) to avoid relevant missing articles.

The process resulted in the removal of those publications that were found to be irrelevant. More specifically, irrelevant publications refer to those articles without any substantial contribution to FCC. An initial sample of 1,779 publications was thus identified and selected. Second, we considered only articles written in English, due to the predominance of that language in scholarly research (Tian et al., 2018). Finally, we chose to consider only journal articles and reviews to foster the reliability of the data, since journal articles undergo a formal double-blind peer-review process. In this regard, conference proceedings, editorial notes, book chapters, book reviews, extended abstracts, technical reports, newspaper articles, consultant reports and reprints were not considered. According to Pittaway et al. (2004), Ertz and Leblanc-Proulx (2018) and Shashi et al. (2020), the selection of the papers was led by two exclusion criteria and one inclusion criterion. The application of the three criteria was progressive. The first exclusion criterion concerns the focus of the abstract and it narrowed the hits to those whose abstract is focused on FCC. The abstracts were read in parallel by two researchers, plus a third one in case of uncertainty. The second exclusion criterion concerns the focus of the entire paper. Starting from the selected abstracts, the entire papers were reviewed by the three researchers to ensure the relevancy and applicability to the study. The non-relevant papers discovered in WoS which were not directly linked to the FCC were retracted from our set. Finally, the inclusion criterion regards the integration of contributions that were not found through the research string and/or comprised in the selected academic database but were cited in the literature on FCC. Therefore, the inclusion criterion is a validation criterion for the choice of search string and academic database. The “Full Record and Cited References” of the final sample comprising 1,189 relevant articles was downloaded using the “save for other file formats” export function with “Tab delimited (window)” and analyzed through VOSviewer.

The total number of publications per year from 1995 to 2019 is shown in Figure 1. The results reveal that the first paper included in our sample dates back to 1995. However, the number of papers published started to grow from 1998 onwards. The findings also reveal that the value of the total number of articles published on FCC increased stably from 1995 to 2008 (i.e. 80 articles were published during this 14-year period). In 2009, the number of papers grew by 75% and, for the following three years – up to 2012 – remained fairly stable. However, only from 2013 there is a significant growth of papers published on FCC topic.

Overall, the FCC literature experienced a significant year-by-year growth, with 1,109 published articles between 2009 and 2019. The results may suggest that globally, hunger and food issues have become rising multi-disciplinary and multi-dimensional topics of interest and this is particularly reflected in the FCC literature.

A particular emphasis on the influence of countries, regions, authors and institutions may grant insights into the development of schools of thought, research hubs and sub-groups of research. Interest at the national or regional level may notably arise among countries or regions that are particularly influenced or dependent on FCC. The influence of authors helps defining relationships between disciplines and research interests. FCC is an inter-disciplinary topic and it is interesting to identify those disciplines that have mostly contributed to its recent development. An examination of institutions will provide insights into their influence and collaboration patterns.

A total number of 85 countries are featured in the 1,189 publications and 22.3% of these countries published only one article. As shown in Figure 2, the USA lead the research on FCC, while China follows closely. The CC infrastructure in the USA is mature and CC transportation facilities availability is approximately 90% for perishable products (Pan et al., 2017). Furthermore, the industry is benefiting from growing online grocery sale which promotes CC developments in the USA (Diego, 2018). In Asia, both China and India are the most productive countries. The extensive population growth in these two countries leads to consider FCC as a critical topic to meet the current food demand and save food for the future. Notably, India is the second most productive country in Asia. However, this is not necessarily well-reflected in terms of number of publications. In contrast, China has approximately 3.5 times more publications than India. This may be explained by the fact that Indian CC industry is still emerging and the CC potential remains untapped (Bharti and Mittal, 2018). The growth in refrigerated infrastructure and distribution systems within the past few years is promising in this regard, but a fully refrigerated SC is still in its infancy (Dharni and Sharma, 2015; Mercier et al., 2017). Besides, China’s CC has been poised for unprecedented growth and benefited from growing penetration of connected devices. Also, the Chinese market has shown potential for growth due to the increasing demand for perishable food items via online channels. Many global CC players are building alliances and collaborative partnerships with China’s local service providers (Market Research Report, 2018).

Interestingly, the contribution of African scholarship (except South Africa) is very limited. This is surprising given that there are issues of acute food shortage in several African countries (Trading Economics, 2019). Lack of sufficient and efficient CC infrastructure creates major instances of food wastage in African countries. Meanwhile, fostering FCC research through collaboration between research institutions, academia and other key stakeholders, developing the knowledge networks among researchers and establishing linkages to professional associations, industry bodies and government ministries, are prime necessities of the present time in the region (FAO, 2015). Accordingly, FCC could prevent price inflation by improving storage conditions, hence stabilizing perishable goods and smoothing demand over longer time spans (Bogataj et al., 2005). Besides, according to the Food and Agriculture Organization (FAO) of the United Nations, in Africa, there remains a dearth of research investments and efforts to spawn the unprecedented benefits of FCC sectors in terms of minimizing food waste (FAO, 2016). Meanwhile, it should be stressed that this low number of publications does not account for the share of studies made by non-African academics in Africa. Thus, this aspect could be justified by the fact that African FCC challenges are being partially addressed by non-African researchers (Maertens et al., 2012; Bekele et al., 2017; Sheahan and Barrett, 2017; Heard and Miller, 2018).

With regard to the number of articles published, the first ten countries including the USA, China, the UK, Italy, Germany, Australia, The Netherlands, Spain, South Africa and India, account collectively for 87.46% of the published articles ( Appendix 1, part 1). As shown in Figure 2, both the USA and China lead FCC research, with 194 and 178 articles, respectively. Yet, Europe dominates the publication in FCC, as five out of the ten top publishing countries are European.

Part 2 of  Appendix 1 shows that the most productive countries are not necessarily the most impactful ones (in terms of the number of citations received). Although the premier publishing nation, the USA ranks second in the number of citations received after the UK. With the slight exception of Canada for North America and South Africa for Africa, all European countries have an equal or better ranking in the number of citations than in the number of publications. Although less productive, European FCC research, appears more influential than FCC research from other regions of the world. The negative gap between the number of articles published and the number of citations received appears larger for Asian countries with the notable exception of Iran. Countries representing other regions of the world (i.e. Australia for Oceania, Brazil for Latin America and South Africa for Africa) follow similar patterns. Likewise, for North America, as represented by the USA.

It is worth mentioning that average values of citations tend to be inflated by a small number of heavily cited papers, which may thus constitute outliers that increase average values. To control this aspect, we present the median value of citations. The results tend to confirm the presence of outliers, as average values tend to be two to three times higher than median values. Besides, the results are consistent with the abovementioned findings suggesting that Europe is more influential than other regions ( Appendix 1, Part 2). More specifically, both Finland and Norway are topping the list with 22 median citations each, followed by Greece, Sweden, The Netherlands, Belgium, Denmark, Spain and the UK, respectively.

The results show that a total of 3,882 authors contributed to the 1,189 sample articles, that is, an average of 3 authors per paper. However, 87.86% of authors have published just one article, suggesting that FCC has essentially been an area of research diversification instead of research specialization (Ertz and Leblanc-Proulx, 2018).  Appendix 2 summarizes the 10 most prolific FCC researchers based on the number of articles published as a first author. Notably, in case of equal publications, citations are considered for ranking. The most productive scholars belong to the most productive and influential countries, namely, European countries, and to a lesser extent, Australia. Xiao Xinqing, from China, is the most prolific author with 8 articles published followed by Stephen Wiedemann with 6 (from Australia), and Badía-Melis Ricardo (from Spain) and Kirezieva Klementina (from The Netherlands), with five articles published for each.

With regard to the average number of citations received per article among the 10 most prolific researchers, Sivakumar Dharini from South Africa, appears to be the researcher with the highest number of average citations received per article (56.75), followed by Badía-Melis Ricardo (32.00) and Kirezieva Klementina (25.80). None of the most prolific authors belongs to North American and Latin American institutions, and only one belongs to Asian institutions.

A total of 1,475 academic institutions have published on FCC from 1995 to 2019, while roughly three quarters (73.35%) contributed a single article, suggesting that FCC has become an area of expertise in a few key institutions.  Appendix 3 reports the top 25 performing research organizations based on both the number of articles published (Part 1) and citations received (Part 2). Based on the number of articles published (Part 1 of  Appendix 3), Wageningen University (44 articles), China Agricultural University (37), Ghent University (19 articles), University of Pretoria (18) and Cranfield University (16) are the top 5 most productive institutions. These institutions also belong to the most productive and impactful countries.

The analysis of the number of citations (Part 2 in  Appendix 3) shows that the Wageningen University & Research Center, the Cranfield University, the Katholieke Universiteit Leuven (KU Leuven), the Ghent University, and the Technical University of Denmark are the top 5 most influential institutions. The total number of citations from these publications range from 481 to 1013. Nevertheless, another interesting finding emerged from this analysis.

Among the top 25 in terms of the number of citations, only 7 institutions are extra-European, including 2 institutions from Africa, 2 from Australia, 1 from China, 1 from the USA and 1 from Vietnam. This contrasts with the top 25 most productive institutions, where 11 institutions are extra-European. The results suggest clearly that European scholarship prevails as the most productive and impactful in the FCC research field. When considering the average citations per article, both the British and Spanish institutions, in particular, display the highest scores. Overall, while the average citations per article range between 22.15 and 126.50 in Europe, the variance is lower for institutions outside of Europe with a range of 9.05 to 79.66.

Median citation analysis further confirms that European institutions have published influential FCC research. In this line, University of Birmingham (UK) and Tekniker (Spain) display the highest median values (that are also similar to their respective average values). Other top institutions include the University of York (UK), the Scientific Veterinary Institute Novi Sad (Serbia) and Aalborg University (Denmark). However, the National Institute of Veterinary Research (Vietnam) appeared as an exception as it constitutes a non-European influential institution.

Further refinements taking the QS World University Ranking 2019 [¹] suggested that among the top 25 organizations in number of articles, only twenty were listed in the QS World University Ranking 2019 [²]. More surprisingly among them, only two (i.e. Cornell University and KU Leuven) are listed in the World’s Top 100 Universities ranking of 2019 [³]. These findings reflect that a large part of the FCC research was not conducted by the top research institutions.

We examined the main journals that have published papers on the topic of FCC. Since journals are representative of specific disciplines or multiple disciplines from an interdisciplinary perspective, this examination provides insights into the disciplines that appear to be the most interested in this topic.

The 1,189 articles were published in 458 unique journals.  Appendix 4 summarizes the top 25 contributing outlets in terms of number of papers published on the FCC topic. More than two out of five articles (35.75%) (425 articles) were published in one of the top 25 journals. The remaining 764 journals published 64.25% of the papers. While top-tier journals concentrate major research on FCC, about 35.8% of journals have published more than one article on the topic of relevance, indicating that the FCC literature is relatively scattered across various journals.

The distribution of articles across journals shows great diversity in disciplines and journals.

Food Control emerged as the most popular journal with 42 published articles, followed by Journal of Cleaner Production (34 articles), British Food Journal (33 articles) and International Journal of Production Economics (28 articles). These international journals are very diverse in nature and scope. Food Control is concerned about food process control and food safety. Specific topics such as quality assurance, risk assessment, hazard analysis, food packaging, processing and manufacturing fit well with the scope of FCC. The Journal of Cleaner Production is a trans-disciplinary publication outlet. This journal is directly linked to environmental and environmental sustainability issues such as sustainable products and services, corporate social responsibility, sustainable consumption/development or corporate sustainability. British Food Journal is a long-standing journal that fosters a broad and unique interdisciplinary coverage of food-related topics of research. The journal provides an essential communication link between all the sectors of the food industry while informing on topical issues and emerging trends. This journal covers a breadth of topics including food SC and logistics, food quality/safety and food sustainability and economics (e.g. food and water security). The journal has also a large share of publications on marketing, distribution, retailing and consumer behavior. International Journal of Production Economics is also interdisciplinary and publishes research on the interface between management and engineering. Topics of particular interest include the whole cycles of activities such as product life cycle analysis, from product research and development to product disposal. The journal also addresses the material flow cycle including supply, production and distribution, in which the FCC research field fits well.

According to Shashi et al. (2018), FCC is a crossroad research domain covering different subjects and involving a variety of journals belonging to different countries and publishers. Therefore, it is worth classifying the journals according to their country and publisher to assist researchers in identifying the top publishing countries and publishers of the field (Zolfani et al., 2015). Furthermore, scholars strongly recommended h-index (Mingers and Yang, 2017), SCImago Journal Rank (SJR) (Gonzalez-Pereira et al., 2010) and the impact factor of JCR (Garfield, 2006) as three key measures to assess the journal performance.  Appendix 6 summarizes the publisher, country, coverage, h-index, SJR score, impact factor and subject categories of the top 25 journals. Appendix shows that among 25 journals, Elsevier emerged as a dominant publisher in the FCC research, followed by Emerald.  Appendix 6 also shows that the top 25 journals include ten thematic disciplines that are the most frequently represented.

Regarding ranking of journals in terms of citations, we included in the analysis journals with at least two publications, and consequently, journals with only one published article albeit with a high number of citations did not appeared in the ranking. In this line, the International Journal of Production Economics emerged as the most influential journal (1,089 citations) followed by Journal of Food Engineering (858 citations), Food Control (827 citations) and Trends in Food Science & Technology (588 citations). While the International Journal of Production Economics is highly impactful in terms of number of citations (1,089), OR Spectrum has the highest average number of citations per article (105.50) followed by Industrial Marketing Management (97.00) (see  Appendix 5). Turning to median citation analysis, the findings confirmed OR Spectrum (105.50 median citations) and Industrial Marketing Management (97.00 median citations) as top influential journals, followed by Environment and Planning A-Economy and Space (84.00 median citations) and International Journal of Life Cycle Assessment (43.50 median citations).

The article’s citations are considered as an indicator to evaluate their quality and impact, as well as their contribution to theory building. The impact of frequently cited articles is generally considered higher than the impact of less cited articles (Culnan, 1986; Furrer et al., 2008). Citation analysis thus facilitates the identification of the most impactful articles, journals, organizations and countries (Liu, 1993). A total of 27,544 cited references were featured in the selected articles.  Appendix 7 summarizes the top 50 cited articles among the sample.

The article entitled “Food waste within food SCs: quantification and potential for change to 2050” published in 2010 by Parfitt et al. emerged as the most cited article with 788 citations since 2010 (until 2019), within the data set. This article also displays the highest average number of citations per year (78.8). Furthermore, “Follow the thing: Papaya,” published by Cook in 2004, came out as the second most cited article with 243 citations. Interestingly, these two papers were published in journals that did not emerge in the top 25 contributing journals as per number of articles ( Appendix 4) or the top 25 cited journals as average citations per article ( Appendix 5). This is due to the fact that these journals have contributed a single article to FCC research. Similarly, there are other journals Green Chemistry, Simulation Practice and Theory and the Lancet Infectious Diseases, and that are not specialized in either SC, food studies or agriculture. In fact, the scope of these journals respectively is ecology, mathematics and computer science, as well as medicine. This demonstrates that FCC is inter-disciplinary from its inception and that the development of FCC theory largely benefitted from a variety of disciplines outside of the sole areas of agronomics and logistics.

Keywords summarize the content of an article, but may also include the methods, objectives, purposes and study areas (Tian et al., 2018). According to Keupp et al. (2012) and Feng et al. (2017), keywords analysis is a quantitative approach to scientifically discover linkages among sub-fields. The greater the occurrence frequency of the keywords, the higher the attention paid to the topic. A total of 5,930 different keywords were identified within the 1,189 articles.  Appendix 8 reports the top 20 most frequently occurred keywords. The most frequently cited keywords are relevant to FCC as they include “quality”, “supply chain(s)”, “management”, “model(s)”, “temperature”, “food(s)”, “system(s)”, “food safety” and “cold chain(s)”, with 181, 176, 140, 133, 117, 111, 109, 105 and 101 occurrences, respectively. Keywords such as carbon emission(s), shelf-life, sustainability, traceability and storage, further show that the FCC appears as a major contributor to environmental sustainability and social responsibility.

The science mapping was conducted using network analysis. In recent years, several software tools with different characteristics were developed to conduct network analysis to map science. In this paper, we use VOSviewer to conduct science mapping analyses and more specifically, to perform co-authorship and co-citation analyses. VOSviewer [⁴] is a freely accessible program widely used for developing and visualizing network maps using bibliometric data (van Eck and Waltman, 2010). VOSviewer provides “a low-dimensional visualization in which objects are located in such a way that the distance between any pair of objects reflects their similarity as accurately as possible” (van Eck and Waltman, 2007, p. 1).

Overall, the two-dimensional visualization map created by VOSviewer includes items and links. The items are the unit of analysis of the map and may refer to articles, authors, countries, journals or organizations (Rizzi et al., 2014). The size of each item and the font of its label reflect the frequency of occurrence. The higher the item size and the label form, the higher the frequency of occurrence. Items connect via links and, if present, they indicate that a relationship between items does exist. For instance, co-citation links suggest links between articles and co-authorship links suggest links between authors. Each link has also certain strength, a positive value that may vary depending on the degree to which the relationship between items is strong. The larger the strength of the link, the stronger the co-authorship/co-occurrence/co-citation will be. In VOSviewer, both the x- and the y-axes of the network visualization do not entail intrinsic significance and therefore the visualized maps can be freely interpreted (Nunen et al., 2018). However, items located close to each other are strongly related, whereas items located far away from each other are weakly related. The items of the network can be grouped into clusters. Items in the same cluster reveal their relatedness and similar characteristics. As for the layout and clustering of the map, the association strength normalization method is used for normalizing the strength of the links between items and for visualizing maps (Van Eck and Waltman, 2009).

In this study, we use co-citation analysis with VOSviewer to perform science mapping. Since scientific documents are created by citing earlier scholarly work, the network of citations provides evidence of the intellectual base of a knowledge domain (Liu et al., 2015). More specifically, co-citation frequency and patterns provide clues of knowledge domains as larger co-citation frequencies between articles indicate stronger relationships and groups of highly co-cited articles represent collective knowledge (Feng et al., 2017; Liu et al., 2015). Co-citation appears when both A and B (considering that A and B may be articles, authors or journals) are together cited by C (where C may be an article, an author, or a journal) (Ertz and Leblanc-Proulx, 2018). High (low) co-citations demonstrate similar (different) research themes and interests (Benckendorff and Zehrer, 2013). Co-citation analysis can either be used with authors and/or publications to identify and study the links between authors, articles, journals and countries (Pilkington and Liston-Heyes, 1999). This paper examines two types of co-citation trends: co-citations of cited references and co-citations of cited authors. According to White and Griffith (1981, p. 163), the authors’ co-citation analysis may contribute to a better understanding of the intellectual structure in the sciences and “in other areas to the extent that those areas rely on serial publications.” In addition, highlighting the importance of the authors’ co-citation analysis, Mishra et al. (2018) reported that it shows the structural configuration of associations between authors. Conversely, co-citation analysis of cited references connects specific published documents (McCain, 1990). Therefore, the number of identical citing items defines the strength of co-citation between the two cited papers (Small, 1973). Accordingly, cited articles create the intellectual structure of a research field (Ding et al., 1999; McCain, 1986; Dzikowski, 2018), its structure, dynamics and evolution (Pilkington and Meredith, 2009; Liu et al., 2015). Koseoglu (2016) advocated that the visualization of the co-citation networks can help researchers to clarify the strength of the ties within the entire network and the positioning of a given citation within the field (Koseoglu, 2016). Therefore, co-citation analysis of authors and articles is implemented to identify groups of topics and authors and investigate how they might be related (Chen, 2006).

We provide a detailed explanation of the co-citation analysis methodology for both cited authors and cited references, in Sections 4.1 and 4.2, respectively.

After processing the cited references’ data retrieved from 1,189 articles belonging to the data set, we obtained a pool of 31,423 cited authors. This pool was further trimmed down to authors with at least 35 citations, resulting in 74 authors cited 3,881 times. As for articles, a data clustering placed together sets of authors sharing similar characteristics (Radicchi et al., 2004) (Table 1). Figure 3 reveals the four main clusters identified and that “Ruiz-Garcia I” and “Jedermann R” are the most highly co-cited authors (120 co-citations) followed by “Jedermann R” and “Defraeye T” (108 co-citations); “Henson S” and “Reardon T” (91 co-citations); “Jacxsens I” and “Luning PA” (91 co-citations); and “Jedermann R” and “Badia-Melis R” (80 co-citations).

Each extracted cluster comprises a few leading FCC researchers whose inputs have had a profound role in the growth of FCC research. For example, in Cluster 2, Van der Vorst’s work specializes in food SC performance improvement and in simulation modeling for food SC redesign; integrated decision-making on product quality, environmental sustainability and logistics, had a substantial influence on authors working on the durability of FCC such as Zanoni or Govindan, both visually close to van der Vorst. Van der Vorst also furthered the application of statistical tools such as planning models, used by Ahumada – another influential author - or optimization tools used by Rong – who specializes in managing fresh food quality. Both are topically close and leading authors in Cluster 2. Cluster 1’s influential authors are Jedermann who specializes in intelligent food logistics and to a lesser extent Ruiz-Garcia whose work is also tangential to technology use in agricultural SC. In contrast, Cluster 3 counts mainly supra-national (e.g. FAO, European Commission and WRAP) or national (e.g. USDA) organizations, as influential contributors. Both FAO and WRAP specialize in food waste. Finally, Cluster 4 counts the seminal author Christopher on agile SC, whose visually central position in the cluster shows how deeply connected his work is to those of other authors in the cluster.

To develop a better understanding pertaining to the theoretical roots of the sampled articles, we use a co-citation analysis in which the quoted references constitute the key element of analysis. Within the original sample of 1,189 articles, a total of 45,753 cited references were found and further reduced to references with a minimum of 18 citations, resulting in 61 articles cited 1,728 times. The co-citation analysis was performed on this reduced sample. Figure 4 shows how the articles that are the most frequently co-cited are connected together within a single cluster, and the size of the nodes indicates the frequency of citations of the given article by other articles in the reduced data set.

The “Application of planning models in the agri-food SC: A review” published by Ahumada and Villalobas (2009) and “An optimization approach for managing fresh food quality throughout the SC” published by Rong et al. (2011) came out as top co-cited articles (both co-cited 36 times). Besides, “An optimization approach for managing fresh food quality throughout the SC” published by Rong et al. (2011) and “SC strategies for perishable products: The case of fresh produce” published by Blackburn and Scudder (2009) appeared as second top co-cited articles (both co-cited 31 times). One of these three top co-cited articles is review-based and the other two contributions propose an optimization-based method and strategies to improve food freshness in FCC. Besides, they were all published during the stable period ranging from 2009 to 2011, and that makes the juncture between a phase of low publication and a phase of heavy publication (Figure 2). These results show that review and this new technological innovation inspired scholars to reshape their views on the domain, considering under-theorized themes, topics and perspectives, which altogether propelled the domain in an exponentially-growing field from 2012 onward.

According to Hjørland (2013), articles that are often cited in conjunction have a higher probability of sharing the same area of interest. Consequently, a thorough analysis of the articles that are part of a cluster informs on the research area of that cluster. Figure 4 shows four different clusters while Table 2 reports the details of the four clusters. The research areas and the contents of the leading articles were in-depth analyzed to identify the research focus area of each cluster.

Researchers belonging to Cluster 1 have provided a thorough conceptual, theoretical and empirical perspective on RFID, a seemingly crucial technology in FCC. The main advantage of RFID for improving FCC is its traceability capabilities (Kelepouris et al., 2007; Regattieri et al., 2007; Abad et al., 2009), temperature fluctuations management (Jedermann et al., 2009; Montanari, 2008; Bogataj et al., 2005; James et al., 2006; Tijskens and Polderdijk, 1996) and shelf-life management (Qi et al., 2014; Jedermann et al., 2014). Other contributions reviewed the potential applications of RFID and demonstrated the various economic and technical challenges delaying the widespread use of RFID in the agricultural and food industry (Ruiz-Garcia and Lunadei, 2011). Importantly, most of these studies take a holistic SC perspective. For example, Aiello et al. (2012) framed CC as a pipeline of stocking and transportation functions starting from the farm and ending ultimately at consumer touchpoints, each of these functions being characterized by a deterministic temperature and a stochastic interval in which RFID could provide useful solutions to critical issues. Further with regard to RFID, Kerry et al. (2006) evaluated the potential of RFID use for meat and CC of meat products. By contrast, Kuo and Chen (2010) proposed a logistics service model based on multi-temperature joint distribution system for thermal protection in FCC. On another note, Bosona and Gebresenbet (2013) shed light on definition, drivers, hurdles in designing and implementing food tractability systems, paybacks, traceability technologies and related improvements, as well as their performances.

Research in Cluster 2 essentially revolved around the application of production and operations planning models in the context of FCC to resolve the problems of SC and distribution. There are two broad approaches presented in the literature. An important stream of studies focuses on the programming, simulation, optimization and statistical tools to solve FCC-related problems (e.g. stochastic vehicle routing, distribution planning). These include heuristic algorithms (Osvald and Stirn, 2008), simulation procedures (Van der Vorst et al., 2009), Tabu search algorithm (Zhang et al., 2003), time-windows model (Hsu et al., 2007) or linear programming (Rong et al., 2011). Studies in this cluster also analyzed how management and administration tools solve technical issues while maintaining broader profitability and business-oriented objectives. These studies cover agri-food business models on procurement and harvesting planning (Lowe and Preckel, 2004; Ahumada and Villalobos, 2009), cost control frameworks (Blackburn and Scudder, 2009), food distribution management (Akkerman et al., 2010) and flowering-harvesting (Widodo et al., 2006). The contributions by Ahumada and Villalobos (2011a, 2011b) focused on studying FCC from an integrated perspective. In fact, Ahumada and Villalobos (2011a) developed an operational model considering various factors (labor management cost, preservation value of fresh foods, transportations modes and products’ quality) for short-period planning decisions. Likewise, Ahumada and Villalobos (2011b) proposed an integrated tactical planning model for production and distribution decisions based on traditional factors (price estimation, availability of resource, price dynamics, product decay, costs of transportation and inventory). Both studies pioneered a subsequent stream of research concentrating on developing multi-objective integrated frameworks (Amorim et al., 2012) some even starting to incorporate sustainability concerns (Govindan et al., 2014). However, such research remained scarce since Soto-Silva et al.’s (2016) review of operational research models applied to FCC, confirmed the unavailability of holistic approaches for the design and management of FCC.

Cluster 3 mainly focused on postharvest waste, causes of postharvest wastage and perishable inventory ordering polices and models. Bakker et al. (2012) summarized the inventory models with deteriorating items reported since 2001 by revisiting the Goyal and Giri's (2001) review. Ferguson and Toktay (2006) designed frameworks to support a manufacturer’s recovery strategy. Mena et al. (2011) categorized the food waste roots into three groups: mega-trends in marketplace, general causes and management root causes. In this context, several studies started to frame food waste in the FCC as a sustainability issue. Parfitt et al. (2010) – also the most cited article – was one of the first to link food wastage with social responsibility and environmental sustainability aspects. They stressed that food waste reduction can accelerate both economic and social sustainability. Drawing on Parfitt et al.’s (2010) foundational paper, researchers subsequently studied the value of food wastage at retail and consumer levels and covered the economic dimension of FCC sustainability (Buzby and Hyman, 2012). For instance, Papargyropoulou et al. (2014) examined the factors of food wastage throughout the SC to propose a hierarchy to prioritize the practices to prevent and manage food waste from a triple bottom line perspective (i.e. environmental, social and economic). Albeit not focused directly on sustainability, another stream of research contributed to it indirectly by putting an emphasis on policies to improve food product quality. These include policies to ensure fresh products (Goyal and Giri, 2001), price policies according to shelf life (Wang and Li, 2012), with the overarching purpose of maximizing profits, as well as ordering policies (Nahmias, 1982). In the past decade, research in this cluster also started to harness the power of technology to improve FCC performance. For example, Kaipia et al. (2013) argued that efficient information sharing system and prompt deliveries can improve the FCC performance. Similarly, van Donselaar et al. (2006) concluded on how intelligence in automated store ordering systems in supermarkets can be enhanced to curb food perishability. Finally, Buzby and Hyman (2012) introduced the discussion of food waste value at the retail and consumer levels as well as economic incentives to reduce food waste in developed nations.

Studies in Cluster 4 were concerned with critical issues in FCC. A critical aspect is the identification of major operational issues (Shukla and Jharkharia, 2013; Lynch et al., 2009), critical success factors of FCC performance (Dolan and Humphrey, 2000), and the subsequent proposition of key indicators to measure, evaluate and trace the evolution of problematic areas (Aramyan et al., 2007). This cluster, therefore, discusses private standards (Henson and Reardon, 2005) and best practices (Cai et al., 2010) to ensure food traceability and safety. In this regard, Trienekens and Zuurbier (2008) discussed the challenges and development of quality and safety standards and underscored that quality assurance will constitute an important aspect of FCC. Quality assurance and technological innovation are key topics to ensure FCC performance. Due to the inherent objective of combating wastefulness, the research in this topic presents some connections with the environmental sustainability research stream. For instance, Van der Vorst et al. (2009) provided a simulation tool for designing food SC by taking food quality change and environmental sustainability issues of different scenarios into the consideration. Tangentially close to sustainability, Aung and Chang (2014b) demonstrated the methods implemented for setting optimal target temperature for multi-commodity cold storage and supported that sensor-based methods are superior to traditional visual assessment method. Other contributions argued that the use of refrigerators in CC negatively impacts the environment as they require additional energy (James and James, 2010), while Zanoni and Zavanella (2012) covered the costs pertaining to additional energy requirement to run FCC operations. The debate on the sustainable nature of FCC is ongoing in this cluster but offers promising research avenues in this regard.

The four clusters previously identified are analyzed more in-depth in Table 3 in terms of current research and suggestions for future research. The clusters need to be considered in relation to each other. More specifically, as an emerging theme, environmental sustainability needs to be considered as a key objective (Clusters 3 and 4). Yet, to build FCC for environmental sustainability, Clusters 1 and 2 have to be analyzed within the framework of environmental sustainability and to pursue environmental sustainability objectives.

Publications in Cluster 1 revolve around two major axes, namely, the benefits of RFID technologies in FCC and the limitations and challenges related to the use of RFID. Benefits consist of traceability capabilities and temperature fluctuations management. With further technological advancements, especially internet of things (IoT), which is part of the broader Web 4.0 comprising semantic Web, 5 G and cyber-physical systems, many more objects can now be traced and monitored efficiently along the FCC. Future research may therefore put a greater emphasis on the benefits of other sensor-based or wireless technologies in an industry 4.0 context. Besides, autonomous objects self-monitor themselves and implement retroactive actions according to the data sensed in the environment (Mahmud et al., 2018). These intelligent items maybe machinery or transportation means and could thus remotely self-diagnose, self-coordinate or operate autonomously (Borgia, 2014) while providing real-time data about the food, crop or livestock they carry or interact with. This may also contribute to the standardization challenge identified as a limitation in past research (Ruiz-Garcia and Lunadei, 2011) since all connected objects would operate on the standard Web protocol. In fact, Web 4.0, and its corollary the semantic Web and 5 G may contribute to seamless data exchange through multiple devices and networks with the standard Web protocol at its core. Machine learning tools may then be used to analyze the large volume of data generated throughout the SC with those technologies. Besides, remote control of connected objects also means that “firms and users control product functionalities and personalize their experience remotely” (Alcayaga et al., 2019, p. 628). This means greater opportunities for cooperation, trust, information sharing and even coopetition (cooperation between competitors) that may rise to solve the information sharing issues often associated with RFID (Ruiz-Garcia and Lunadei, 2011). Future research should therefore:

• Assess the benefits of technologies related to Web 4.0 and Industry 4.0 for the FCC.

Past research emphasized a great variety of challenges and limitations related to RFID application (Ruiz-Garcia and Lunadei, 2011; Bosona and Gebresenbet, 2013). Future research might therefore focus on building IoT using RFID. This might be done by linking firm’s physical capital resources (e.g. hardware, sensors, tags), human resources (e.g. training, education) and organizational resources (i.e. values, culture, processes) – of resource-based theory (Barney, 1991) – to big data capabilities (i.e. analytics capabilities), education and creative intensity, as suggested by Erevelles et al. (2016). Therefore, future studies might:

• -Building IoT using RFID by linking company resources to big data analytics capabilities and creative thinking.

• - Solving multi-granularity data issues by applying machine learning tools to analyze large quantities of FCC data and at multiple levels (e.g. pallet-level tagging, item-level tagging), to extract meaningful information and intelligence for managerial decision-making.

• - Assessing the value of autonomous objects and internet of automated things for the FCC.

Studies in Cluster 2 focus on FCC-specific programming methods, on the one hand, and management tools, on the other. Therefore, future research might grow further along these two axes. First, with regards to statistical tools, past research focused on a wide array of methodological approaches such as linear programming (Rong et al., 2011), simulation models (Van der Vorst et al., 2009), heuristics algorithms (Zhang et al., 2003; Osvald and Stirn, 2008) or metaheuristics (Hsu et al., 2007). Hence, most studies focus on a single methodological approach. Yet, methods and techniques tend to have advantages and disadvantages, thus the focus on any single one of them may not compensate for their limitations (Tufféry, 2011). A wider array of tools may also constitute a fruitful research avenue, to explore problem-solving from different perspectives and with different assumptions. Hence, the research avenues may include:

• Combining different methodological approaches and/or comparing them to solve specific FCC-related issues.

• Using a broader array of methodological approaches including non-linear programming, stochastic optimization, dynamic programming or hybrid models with different assumptions and postulates to enrich problem-solving capabilities.

Second, most methodological approaches focus on specific stages or decision variables in the FCC process. Several focus on two or more, such as procurement and harvesting (Lowe and Preckel, 2004; Ahumada and Villalobos, 2009), production and distribution (Ahumada and Villalobos, 2011b; Amorim et al., 2012), flowering and harvesting (Widodo et al., 2006) and some focus only on one stage such as distribution (Akkerman et al., 2010). Studies adopting an integrated and holistic approach including the planting, harvesting, production, distribution and inventory variables, remain scarce. Besides, the proposed modeling approaches focus predominantly on the tactical and operational decision levels. However, focus on the strategic decision level appears not as well-addressed despite the importance of higher-order decision-making. Therefore, future research might consider the following avenues:

• Developing methodological approaches that adopt an integrated and holistic framework for FCC-related problem-solving.

• Developing methodological approaches that focus on the strategic levels of decision-making or alternatively, integrate multiple decision levels including operational, tactical and strategic.

Research in Cluster 3 focus on specific aspects related to the identification of major operational issues in FCC. One major facet involves the postharvest wastage and perishable inventory management. Currently, research tackles this issue from a great variety of perspectives which has produced a rich body of knowledge on ways to prevent, handle and measure food loss. For example, food losses and waste in the SC are considered from an integrated perspective including agriculture, food processing and manufacturing (Parfitt et al., 2010; Lundqvist et al., 2008). Since Parfitt et al. (2010), the study of food waste has started to cover both the retail and consumer levels as well (Buzby and Hyman, 2012). However, as emphasized by Papargyropoulou et al. (2014), the adoption of integrated approaches considering multiple stages simultaneously remains infrequent. Besides, the distinction between the retail-consumption nexus remains yet to be more thoroughly made. In fact, retail and consumption are often considered together (Parfitt et al., 2010; Buzby and Hyman, 2012). However, food loss differs in means and magnitude across both stages. For instance, in retail, food loss may arise due to damage during transport, spoilage, poor handling, losses caused by lack of cooling or cold storage (Parfitt et al., 2010). Some retailers also perform secondary processing functions (e.g. mixing, cooking, frying, molding, cutting extrusion), product evaluation (quality control) and packaging (weighing, labeling, sealing), when issues such as process losses, contamination, product discarding, destructive testing, or inappropriate packaging may arise (Parfitt et al., 2010; Papargyropoulou et al., 2014). In contrast, food loss at the consumption stage arises due to over-purchase or inappropriate purchasing, food loss during storage, preparation, portioning and cooking but also confusion over “best before” and “use by” dates (Parfitt et al., 2010; Papargyropoulou et al., 2014). Researchers may adopt a more consumer-centric perspective by considering consumer behavior models such as the model of goal-directed behavior (Perugini and Bagozzi, 2001). Further emphasis should be equally placed on the analysis of a variety of waste handling strategies especially prevention which is the most desirable option (European Parliament Council, 2008; Ali et al., 2019) but also food preparation for re-use, recycling, recovery and disposal, whether with food that is fit or unfit for human consumption. Future studies might therefore consider:

• Exploring operational issues from an integrated perspective involving all stages of the food SC (i.e. agriculture, processing, retail, consumption).

• Establishing a clearer distinction between food loss at the retail stage and at the consumer stage and adopt the consumer perspective with consumer behavior models.

• Focusing on the optimal strategy of preventing food waste without neglecting less optimal solutions including prepare for re-use, recycling/composting, recovery or disposal.

Research on waste management has been more specifically studied from the angle of inventory policies and models. A broad array of policies has been considered in this regard, including ordering policies (Nahmias, 1982; Goyal and Giri, 2001), pricing policies (Wang and Li, 2012), transportation policies (Kaipia et al., 2013). This approach can be qualified as “ex ante” since it seeks to reduce waste at ordering and planning level before food is actually purchased and/or delivered. Some research also investigated manufacturers’ recovery strategies (Ferguson and Toktay, 2006). In this vein, a promising avenue of research relates to food waste management from an “ex-post” vantage point, that is, after items have been purchased and/or delivered. Recent technological advances that were not so prevalent over the past 20 years might constitute powerful enablers in this regard. Past research has shown that efficient information sharing systems are powerful means to manage food waste (Kaipia et al., 2013). Likewise, recent studies showed for example how digital platforms acting as “circularity brokers” endorse matching roles to prevent food waste (Ciulli et al., 2019, p. 1). Those typically involve the pairing of providers with obtainers (Ertz et al., 2019). Such initiatives might be further investigated for applicability throughout the FCC:

• Examining “ex post” opportunities offered by technology in general, and digital platforms (circular brokers), in particular, to spur food redistribution and waste prevention.

Research in Cluster 4 is devoted to the identification of major operational issues. Yet, one of the major aspects to identify those issues and inform adequate decision-making relates to appropriate measurement (Wesana et al., 2019). This involves the identification of key indicators to measure, evaluate and trace problematic areas (Aramyan et al., 2007), standards and best practices (Henson and Reardon, 2005; Cai et al., 2010), as well as quality measurement tools (Trienekens and Zuurbier, 2008). Measurement capabilities should be further developed along these lines with a specific focus on standardization of measurement norms, standards and best practices for measurement. More accurate measurement tools will also enable a more precise identification of factors of problematic issues and thus improve the forecasts by identifying the right predictors:

• Improved measurement capabilities (standardization of norms, standards, best practices for measurement, evaluation and controlling of problematic areas).

• Improve measurement to strengthen forecasts by identifying the right issues predictors.

Overall, the food waste issue has close connexions with ethics (Ciulli et al., 2019), social responsibility and environmental sustainability (Parfitt et al., 2010), and should be further framed by involving theoretical frameworks, tools and theories from cleaner production, sustainability, circular economy and waste management research areas. Recent studies have started this endeavor (Plazzotta et al., 2020), such as by connecting FCC to the closed-loop SC literature (Russo et al., 2019). Consequently, future research might consider the following:

• Establishing closer connexions between the study of food waste management and social responsibility, as well as environmental sustainability areas.

This paper provides a unique contribution to the literature on FCC while extending previous reviews (James et al., 2006; Raab et al., 2011; Defraeye et al., 2015; Mercier et al., 2017; Chaudhuri et al., 2018; Shashi et al., 2018) in five original ways. First, the study goes beyond a systematic literature review of the FCC research field by applying bibliometric analysis (i.e. performance assessment and network analysis) to identify the most influential works and authors according to citations and co-citations.

Second, drawing on co-citation analysis of articles, this study identifies four clusters of articles (“application of RFID technologies,” “production and operation planning models,” “postharvest waste, causes of postharvest wastage and perishable inventory ordering polices and models” and “critical issues in FCC”) focusing on specific areas of FCC. These research areas range from SC and distribution (Cluster 2) through performance evaluation (Cluster 3), while also considering technology (Cluster 1) and environmental sustainability as well as social responsibility (Cluster 4).

Third, an author co-citation analysis also identifies four clusters of authors that partially match the aforementioned clusters of articles. Through its emphasis on quality standards, risk assessment, consumers or performance measurement, Cluster 1 seemingly conflates with Cluster 3 (i.e. postharvest waste, causes of postharvest wastage and perishable inventory ordering polices and models) identified in the co-citation analysis of cited references. Based on co-citation analysis of cited authors, Cluster 2’s emphasis on operations management and logistics, as well as quantitative methods (e.g. optimization, quality management), relates to Cluster 2 on production and operation planning models, in the co-citation analysis of cited references. Studies focused on the RFID technology (i.e. Cluster 1 in the co-citation analysis of cited references) are to be found in both Clusters 2 and 3, but more specifically in cluster 3 in the communication research stream where Ruiz-Garcia and Jedermann published extensively on this topic (Jedermannet al., 2009; Ruiz-Garcia and Lunadei, 2011). The focus on post-harvest issues analyzed in Cluster 4 matches almost completely Cluster 4 in co-citation analysis of cited references. Interestingly, in this cluster, the influence of institutions (e.g. FAO, European Commission) and organizations (e.g. WRAP) in FCC research, is comparatively higher than that of scholars. This may be due to the fact that environmental sustainability in FCC, although being addressed partially in the durability stream of Cluster 2, remains an emerging theme. Past bibliometric analysis showed that emerging themes are often broadly initiated outside of academia (e.g. consultants) and then strengthened theoretically and empirically within academia. This was the case for the collaborative economy (Ertz and Leblanc-Proulx, 2018) and for smart cities (Mora et al., 2017), and it appears that the comprehensive and explicit study of environmental sustainability within FCC is following a similar pattern. This delineation of the literature coupled with advanced information on the most prolific and impactful FCC authors worldwide, which allows researchers to identify core areas of research interests and consider the development of conjoint research projects with other researchers, institutions and centers of researchers based on common research goals and objectives.

Fourth, this study related current research themes to emerging or under-explored ones. The approach undertaken to identify these “dead spots” and “research gaps” was a conventional process in which we analyzed each theme in-depth and related its content to the extant literature. The literature consisted of the most up to date corpus of publications on a given subject. In some cases, when available, the unaddressed research avenues proposed by authors at the end of their papers were also considered as potential candidates for the theory-based research agenda.

Fifth, the research reveals the relationships between the clusters of articles and authors. It also argues that an emerging research stream of environmental sustainability, and to a lesser extent, social responsibility, in the FCC is currently emerging. In sum, the current study argues that better integration of technological advances, namely, RFID technology (Cluster 1), and use of both management and quantitative tools and techniques (Cluster 2) to solve FCC-related issues that can be effectively evaluated (Cluster 3) will result in more environmental sustainability in the FCC (Cluster 4). Therefore, future research may include all these clusters starting from aiming at environmental sustainability and social responsibility in the FCC, considering technological tools and techniques, as well as the evaluation of performance.

This paper is expected to make significant theoretical contributions for several reasons. First, the present study applies bibliometric and network approaches to uncover the most influential articles, scholars, institutions and countries as per the number of articles published and citations. Second, scholars working on FCC may easily recognize the researchers, research institutions and countries conducting research on specific research areas and topics. Therefore, interested scholars can develop joint research projects, share their ideas and discuss their results with the leading authors. Third, the findings can assist the industries and governments to identify the main academic institutions and research centers working in the field of FCC for research projects. Finally, editors organizing special and regular issues on FCC topics can invite leading authors and institutions.

This research offers multiple opportunities to the public authorities, organizations and practitioners that are engaged in leveraging the advantages of SCs by using FCC. This paper equips managers with different perspectives and schools of thought that allow them to harness the advantages from the FCC in their work. Such knowledge is important for managers because it allows them to locate precisely FCC expertise worldwide. Access to FCC insights and potential applied or even fundamental research projects could then be developed subsequently based on this information.

Besides, the four-cluster classification of articles enables managers to:

• assess the current state of FCC in terms of technology, tools and techniques, evaluation, as well as environmental sustainability; and

• reveal their future needs in the appropriate clusters to take relevant decisions on whether to leverage current technologies, tools and techniques, improve the evaluation of FCC, as well as re-think the implications for environmental sustainability strategies through FCC.

The result of this study may have important implications for both regional and national development, regarding the impact on global logistics, shipping activities of foods and environmental policies. Moreover, the results may inspire firms to promote the integration of the entire FCC partners for a common goal and improve the overall performance of the network. In fact, the population growth and the scarcity of resources needed to meet the increasing needs of people require a great attention from institutions and stakeholders. Based on the previous discussions, emerges the need for political action and feasible guidance by the government to enact policies able to guarantee high-quality standard in the management of CC. CC infrastructure and, consequently, the supply network integration, the partners' performance and the stakeholders’ interests are deeply influenced by the policies implemented by the central governments since they are responsible for resources allocation. Therefore, they have to establish policy enforcement and offer incentives and favorable measures to regulate and promote the CCM.

Although great care has been taken to ensure the validity of the study procedure and its results, a number of limitations need to be mentioned. First, we used 1,189 articles published in the past 25 years. Yet, these articles have been selected according to selection criteria. Although we defined a selection criterion to validate the choice of search string and academic database, we only considered papers published in the WoS database in the initial search and excluded other databases such as Scopus, ABI/Inform or Business Source Complete, for example. Besides, only peer-reviewed journal articles were considered while other formats of publications such as conference proceedings, books, book chapters and reports, although possibly influential, have been excluded. Besides, a common issue in the bibliometric analysis is that we used specific keywords for this particular research, yet the use of other keywords and combinations thereof might have led to divergent results. Second, we performed the co-citation analysis using network analysis and VOSviewer software. However, other methods and software may be used, such as Gephi for example. Thirdly, the co-citation analyses of both articles and authors results in four research clusters each. Nevertheless, other methods may lead to other types of classifications. Finally, our study pertains to the FCC research area. However, future review research can be carried out considering CC management in the field of medical, chemical and pharmaceutical industries to underline the research advancements and highlight similarities and differences with FCC domain.

Declaration of conflict of interest: The authors declared no conflict of interest.

Statement of funding sources: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Authors’ contribution: Authors are listed in surname alphabetical order and have equally contributed to the article.

Table A1 

Table A2 

Table A3 

Table A4 

Table A5 

Table A6 

Table A7 

Table A8