Copyright is owned by the Author of the thesis. Permission is given for a copy to be downloaded by an individual for the purpose of research and private study only. The thesis may not be reproduced elsewhere without the permission of the Author. Achieving effective traceability systems for the domestic fresh produce industry in New Zealand A thesis presented in partial fulfilment of the requirements for the degree of Master in Food Technology Massey University, Albany Campus, New Zealand Jiaojiao Gao 2019 i Acknowledgements The completion of my Master study in Food Technology at Massey University would have been impossible without the support and the help from several individuals and organisations. Firstly, I am extraordinarily fortunate in having Ying Zhao as my partner whose love and unconditional support has made this project a great success and fulfilled my dream. Words fail me to express my appreciation to my family whose support and persistent confidence in me have got me where I am today. My deepest appreciation goes to my supervisor Dr. Tony Mutukumira at Massey University for his supervision, guidance, encouragement and constant inspiration from the very early stage of this research. I am grateful for his continuous support and help in achieving my goals and fulfilling my dreams. I am also highly indebted to my co-supervisor Dr Pieternel Luning from Wageningen University & Research for her supervision, advice and guidance. Their crucial contributions and supervisions have made them a backbone to this research and so to this report. I would like to acknowledge the financial support given by Ministry for Primary Industries (MPI) through my awarded MPI Postgraduate Science Scholarship to conduct this project and also my MPI mentor Dr Anne-Marie Perchec Merien who guided and encouraged me to complete this project. Above all and most needed she provided me unflinching encouragement and support in various ways without which the completion of this project would have not been possible. I gratefully acknowledge my employer The AgriChain Centre Limited for allowing me access to all company information, data and images and also providing necessary information regarding the project. I would like to thank United Fresh New Zealand Incorporated and GS1 New Zealand for their help and support during this research. I would also like to thank Dr Hans Maurer, Anne-Marie Arts and Melanie Trotman from The AgriChain Centre Limited and the Manager of Quality Services at GS1 New Zealand Owen Dance for sharing their views and ideas in the project and training me during this study. ii Abstract A reliable and effective traceability system is important to the food industry especially when a foodborne illness outbreak occurs. In particular, fresh fruit and vegetables are highly perishable, fragile, seasonal, diverse products with relatively short shelf life, thereby making their value chain complex and fast-paced. Hence, the traceability system in the fresh produce industry becomes critical in the event of a food crisis where products need to be tracked and traced in a timely manner. The objective of this study was to investigate current traceability systems in the fresh produce industry in New Zealand and also to explore potential improvement in the traceability system along the domestic supply chains. There were four different methods applied in this study: observation of traceability information available on fresh produce products, interviews with industry participants using a questionnaire, survey strategy by means of a questionnaire that was sent to growers, and a pilot study using GS1 technology to examine a modelled traceability system in two supply chains of strawberries. There were 336 fresh produce samples observed for traceability information analysis throughout the supply chain. Four growers, three wholesalers and one retailer from the fresh produce industry participated the face to face interviews. The questionnaire developed in the survey was sent to 578 growers with 40 of them responded and answered. Two pallets of strawberries were selected and GS1 (Global Standards One) barcodes and systems were used in the pilot study to track and trace each strawberry punnet throughout the supply chains. Qualitative and quantitative data were collected from produce traceability data samples, interviewed industry stakeholders, surveyed growers, and the pilot study to generate empirical information on traceability systems along fresh produce supply chains in New Zealand. Subsequently, data were analysed using quantitative tools such as frequency distributions, Chi-Square test (X2) and Fisher’s Exact test, and qualitative descriptions in this study. The findings show that fragmentation of product traceability information, lack of standardisation in data format and information asymmetry exist in the domestic fresh produce industry. As only a ‘one-up, one-down’ traceability system for food businesses is required by regulators in New Zealand, industry players intend to solely focus on their own or internal needs without recognising the importance of an industry-wide traceability system in the fresh produce supply chain. The findings pose a question mark as to whether or not the ‘one-up, one-down’ traceability requirement is sufficient for the fresh produce industry. The findings iii also indicate that an effective and efficient external traceability system throughout the fresh produce value chain in New Zealand is feasible to implement by industry-wide cooperation from growers, packers, transporters and receivers/buyers. This study fills the gap found in the literature where few academic papers focused attention on traceability systems in the fresh produce industry in New Zealand. iv Table of Contents Acknowledgements ................................................................................................................. i Abstract .................................................................................................................................. ii Table of Contents .................................................................................................................. iv List of Figures ...................................................................................................................... vii List of Tables ......................................................................................................................... ix List of Appendices ................................................................................................................. x Abbreviations ........................................................................................................................ xi 1.0 Introduction .......................................................................................................................... 1 1.1 Main Objective ................................................................................................................. 2 1.2 Specific Objectives ........................................................................................................... 3 2.0 Literature review .................................................................................................................. 4 2.1 Introduction ...................................................................................................................... 4 2.2 Foodborne illness outbreaks ............................................................................................. 5 2.3 Characteristics of food traceability .................................................................................. 8 2.3.1 Defining food traceability .......................................................................................... 8 2.3.2 Advantages and disadvantages of food traceability ................................................ 10 2.3.3 Internal and external traceability ............................................................................. 13 2.3.4 Components of food traceability systems ................................................................ 14 2.4 Food traceability technology .......................................................................................... 17 2.5 Barriers of the implementation of traceability systems and their current shortcomings 22 2.6 Fresh produce supply chain in New Zealand ................................................................. 26 2.6.1 Characteristics of fresh produce and its supply chain ............................................. 26 2.6.2 Relationship in fresh produce supply chain in New Zealand .................................. 29 2.6.3 Fresh produce supply channels in New Zealand ..................................................... 30 2.6.4 The scale of the fresh produce industry in New Zealand ........................................ 33 2.6.5 Export and domestic markets in New Zealand ........................................................ 33 2.6.6 Current traceability regulations in New Zealand ..................................................... 35 2.7 Some stakeholders of fresh produce supply chain in New Zealand ............................... 36 2.7.1 Role of Ministry for Primary Industries (MPI) ....................................................... 37 2.7.2 Roles of United Fresh New Zealand Incorporated and The AgriChain Centre ....... 37 2.7.3 Role of GS1 ............................................................................................................. 38 3.0 Materials and methods ....................................................................................................... 40 v 3.1 Study design ................................................................................................................... 40 3.2 Limitations and Constraints ........................................................................................... 41 3.3 Current fresh produce traceability observation .............................................................. 41 3.3.1 Fresh produce selection ........................................................................................... 41 3.3.2 Inspection of fresh produce traceability .................................................................. 43 3.3.3 Analysis of data on fresh produce traceability ........................................................ 49 3.4 Industry stakeholder interview ....................................................................................... 50 3.4.1 Development of questionnaire ................................................................................. 50 3.4.2 Conducting interviews ............................................................................................. 50 3.4.3 Sample size .............................................................................................................. 51 3.4.4 Data collection and analysis .................................................................................... 53 3.5 Grower survey ................................................................................................................ 53 3.5.1 Development of questionnaire ................................................................................. 53 3.5.2 Conducting survey ................................................................................................... 54 3.5.3 Data collection and statistical analysis .................................................................... 54 3.6 Pilot study ....................................................................................................................... 55 3.6.1 Selection of pilot sample ......................................................................................... 55 3.6.2 Design of pilot study ................................................................................................ 56 3.6.3 Field trial .................................................................................................................. 64 3.6.4 Data collection and analysis .................................................................................... 65 4.0 Results and discussion ....................................................................................................... 67 4.1 Current fresh produce traceability .................................................................................. 67 4.2 Industry stakeholder interview ....................................................................................... 73 4.2.1 Objectives and product definitions .......................................................................... 73 4.2.2 Internal traceability, establishment of procedures and flow of materials ................ 74 4.2.3 Information and documentation requirements ......................................................... 80 4.2.4 Structure and responsibilities ................................................................................... 81 4.2.5 Training and external traceability ............................................................................ 81 4.2.6 Internal assessment .................................................................................................. 82 4.2.7 Summary of the industry stakeholder interview ...................................................... 83 4.3 Grower survey ................................................................................................................ 86 4.3.1 Profile analysis of growers ...................................................................................... 86 4.3.2 Perception of growers on traceability ...................................................................... 88 vi 4.3.3 Current practices in relation to traceability ............................................................. 90 4.3.4 Summary of the grower survey ............................................................................. 102 4.4 Pilot study ..................................................................................................................... 103 5.0 General Discussion .......................................................................................................... 109 6.0 Conclusion ....................................................................................................................... 111 7.0 Recommendations ............................................................................................................ 112 8.0 References ........................................................................................................................ 113 9.0 Appendices ....................................................................................................................... 125 vii List of Figures Figure 2.1 Conceptual representation of product and traceability information flow in a food supply chain. .............................................................................................................................. 9 Figure 2.2 Process flow of traceable resource units (TRUs) aggregation and splitting through a typical supply chain in current food traceability systems. .................................................... 16 Figure 2.3 A typical fresh produce supply chain in New Zealand. .......................................... 31 Figure 2.4 Complexity of fresh produce supply channels in New Zealand. ............................ 32 Figure 2.5 Different proportions of export and domestic markets in New Zealand in 2018. .. 34 Figure 2.6 Traceability information needed at different steps in the supply chain. ................. 36 Figure 2.7 Identification in each level of packaging using GS1 standards. ............................. 39 Figure 3.1 Product traceability information on outer packaging from growers. ..................... 44 Figure 3.2 Product traceability information on market packaging from wholesalers. ............. 44 Figure 3.3 Product traceability information on retail packaging (at retailers). ........................ 45 Figure 3.4 A typical product container with traceability information from two sources (left: a card from a packhouse; right: a sticker from a wholesaler). .................................................... 46 Figure 3.5 A typical product traceability information with Product ID and Grower ID (Product ID means product identification; Grower ID means grower identification). ............ 48 Figure 3.6 Selection of supply channels in the pilot study for strawberries at packhouse A. . 58 Figure 3.7 Selection of supply channels in the pilot study for strawberries at packhouse B. .. 59 Figure 3.8 An example of strawberry punnet label used in the pilot study. ............................ 60 Figure 3.9 An example of strawberry crate label used in the pilot study. ............................... 61 Figure 3.10 An example of strawberry pallet label used in the pilot study. ............................ 61 Figure 3.11 Pocket 2-dimensional (2D) barcode scanner. ....................................................... 62 Figure 3.12 Ten process steps involved in the strawberry pilot were modelled into the GEM software system. ....................................................................................................................... 63 Figure 3.13 Screenshot showing GEM software system during process step 1 loading punnets into tray/crate. .......................................................................................................................... 65 Figure 4.1 Prevalence rate of different traceability components on product outer packaging in written and GS1 format. ........................................................................................................... 69 Figure 4.2 Prevalence rate of different traceability components on product market packaging in written and GS1 format. ....................................................................................................... 70 Figure 4.3 Prevalence rate of different traceability components on product retail packaging in written and GS1 format. ........................................................................................................... 71 Figure 4.4 Fresh produce process flow at growers’ premises. ................................................. 76 Figure 4.5 Fresh produce process flow in the three wholesalers. ............................................ 77 Figure 4.6 Fresh produce process flow at the retailer Distribution Centre (DC) ..................... 78 viii Figure 4.7 Choropleth map of responded growers. .................................................................. 86 Figure 4.8 Perceptions (%) of importance on traceability from fresh produce growers. ......... 88 Figure 4.9 Reasons of perception (%) on traceability importance from growers. ................... 89 Figure 4.10 Reasons (%) of traceability recording from growers. .......................................... 90 Figure 4.11 Differences (%) in method of product labelling from growers. ........................... 91 Figure 4.12 Different drivers on product label determination. ................................................ 92 Figure 4.13 Distribution channels of domestic fresh produce. ................................................ 93 Figure 4.14 Product packaging format at selling point from growers. .................................... 93 Figure 4.15 Premises of packing for pre-packed products. ..................................................... 94 Figure 4.16 Traceability information on pre-packed products. ............................................... 95 Figure 4.17 Reasons of using labelling system from growers. ................................................ 96 Figure 4.18 Product labels printing from growers. .................................................................. 96 Figure 4.19 Notification methods on product delivery to wholesaler or distributor from growers. .................................................................................................................................... 97 Figure 4.20 Involvement in product recall from growers. ....................................................... 98 Figure 4.21 Time spent on tracing fresh produce in product recall from growers. ................. 99 Figure 4.22 Extent of difficulty of contacting supply chain partners from growers. ............. 100 Figure 4.23 Product recall plans development from growers. ............................................... 101 Figure 4.24 Challenges in a product recall from growers. ..................................................... 102 ix List of Tables Table 2.1 Examples of foodborne outbreaks due to contaminated produce between 1999 and 2019............................................................................................................................................ 6 Table 2.2 Application of different food traceability technologies in the food industry. ......... 17 Table 2.3 Comparison between barcodes and RFID technology. ............................................ 19 Table 2.4 Characteristics of fresh produce supply chain. ........................................................ 27 Table 3.1 Fresh produce selection. .......................................................................................... 42 Table 3.2 Traceability data collected in the strawberry pilot study. ........................................ 66 Table 4.1 Selected fresh produce sample details. .................................................................... 67 Table 4.2 Estimated (%) costs of growers on product labelling each year. ............................. 87 Table 4.3 Details of strawberry punnets tracked along the supply chain present in this pilot study. ...................................................................................................................................... 103 Table 4.4 Product traceability information extracted from GEM software system for strawberry punnet GTIN 00000940.00611.0502. .................................................................. 105 Table 4.5 A timeline of product movement and relevant traceability details extracted from GEM software system for strawberry pallet SSCC 942101221.30000000. .......................... 106 Table 4.6 Strawberry traceability details extracted from GEM software system for pallet SSCC 942101221.30000000 after the pallet left packhouse A and was depalletised in DC. 107 Table 4.7 Strawberry punnets details extracted from GEM software system for crate (GTIN 942101221.0999.0013) at retail store A1. ............................................................................. 108 Table 4.8 Summary of the challenges encountered in the strawberry pilot study. ................ 108 x List of Appendices Appendix 9.1 Traceability assessment questionnaire. ........................................................... 125 Appendix 9.2 Survey Monkey questionnaire. ........................................................................ 128 Appendix 9.3 Permission from The AgriChain Centre Limited. ........................................... 134 Appendix 9.4 Outputs of statistical analysis of grower survey ............................................. 135 xi Abbreviations CDC = Centre of Disease Control and Prevention EAN = European Article Number FCP = Food Control Plan FDA = Food and Drug Administration Global GAP = Global Good Agricultural Practice GS1 = Global Standards One GTIN = Global Trade Item Number LITS = Livestock Identification and Traceability System MPI = Ministry for Primary Industries NFC = Near Field Communication NP = National Programme NZ GAP = New Zealand Good Agricultural Practice PNs = Petri Nets QR Code = Quick Response Code RFID = Radio Frequency Identification SSCC Serial Shipping Container Code TRU = Traceable Resource Unit UCC = Uniform Code Council 1D Barcode = One-Dimensional Barcode 2D Barcode = Two-Dimensional Barcode 1 1.0 Introduction Foodborne disease outbreaks have always been one of the biggest concerns for government agencies, food industry sectors, consumers and medical professionals (Arendt et al., 2013; Haleem, Khan & Khan, 2019; Hussain & Dawson, 2013; World Health Organization, 2008; Bosona & Gebresenbet, 2013; Wognum, Bremmers, Trienekens & Van der Vorst, 2011). In particular, fresh produce contributes significantly to foodborne illness outbreaks as they are often eaten raw, with minimal or no further processing such as heating to kill any potential pathogens (Bennett et al., 2018; Lynch, Tauxe & Hedberg, 2009; Sivapalasingam, Friedman, Cohen & Tauxe, 2004; Julien-Javaux, Gerard, Campagnoli & Zuber, 2019; Mahajan, Caleb, Singh, Watkins & Geyer, 2014; Prakash, 2016). For example, a Yersinia pseudotuberculosis outbreak linked to fresh produce was reported in New Zealand in 2014 and resulted in over 300 illnesses in the country (Ministry for Primary Industries, 2014). Another outbreak caused by Escherichia coli O104:H4 occurred in Germany and France in 2011 was associated with imported fenugreek seeds (De Henriette, 2016). Moreover, a Listeria monocytogenes outbreak caused by contaminated cantaloupe was reported in the United States and was blamed for at least 26 deaths (Claiborn, 2011). In such cases, a reliable traceability system plays a vital role and becomes an effective method when a food recall is required (Dzwolak, 2016; Porter, Baker & Agrawal, 2011; Fan et al., 2019; Mgonia, Luning & Van der Vorst, 2013; Rong & Grunow, 2010; Zhao & Nakano, 2018; Mainetti et al., 2013; Pouliot & Sumner, 2008; Dabbene, Gay & Tortia, 2014). According to Lindh and Olsson (2010), food traceability can be used to improve food safety management systems for food businesses and enhance customer confidence in their products. It was described as the ability to trace food products through the supply chain from farm or production to consumer (Charlebois & Haratifar, 2015). There are various elements and information included in a reliable traceability system throughout the produce value chain, ranging from grower name, grower contact details, product name, variety, colour, size, pack date, harvest date, field code, transport details, to wholesaler and retailer information (Kondo, 2010; Fresh Produce Safety Centre Australia & New Zealand, 2019). This type of information is vital in the event of a product recall (Pouliot & Sumner, 2008). It is well-known that the meat and dairy industries have made huge efforts to ensure that product traceability procedures are implemented and a great deal of research has been undertaken to improve their traceability systems (Donnelly, Karlsen & Olsen, 2009; Mgonia 2 et al., 2013; Mutua et al., 2018; Vulton, 2010). For instance, Mgonia et al. (2013) evaluated the traceability system of a fish processing company with a diagnostic tool. Mutua et al. (2018) conducted a livestock identification and traceability pilot with a beef supply chain between Tanzania and Kenya. On the other hand, there were also a number of studies carried out to investigate current traceability systems for fresh fruit and vegetables value chains worldwide (Manikas et al., 2010; Canavari et al., 2010; Gichure et al., 2017; Kondo, 2010; Kim, Hilton, Burks & Reyes, 2018). Manikas et al. (2010) studied different factors affecting the efficiency of traceability systems in the fresh produce industry in Greece. Canavari et al. (2010) investigated traceability systems as part of information management in the fruit value chain in Italy. Gichure et al. (2017) carried out a case study on product traceability system within the kale supply chain in Kenya and examined a variety of factors influencing the implementation of effective traceability systems. However, very few studies have been conducted to explore the existing traceability systems in the fresh produce industry in New Zealand. Additionally, the challenges and limitations of food traceability system implementation throughout fresh horticultural products supply chain in New Zealand are not well understood. To date, there is no standardized industry-wide method used in New Zealand that allows traceability from the grower, through the supply chain to the consumer. In addition, the fresh produce supply chain is fragmented, and has fragmented information in terms of effective traceability due to the sheer number of products and the intricacies of each supply chain merging at various levels before reaching the consumer (Manos & Manikas, 2010; MacKenzie & Apte, 2017; Manikas et al., 2010). Furthermore, the fresh produce industry is fast-paced, with most products moving from farm to plate in as little as a few days (MacKenzie & Apte, 2017). This all contributes to produce traceability challenges. Therefore, there is a need to make headway in improving traceability systems for the sector. 1.1 Main Objective The main objective of this study was to understand current traceability systems in the fresh produce industry in New Zealand and explore potential improvement in the traceability system along the domestic value chains. 3 1.2 Specific Objectives Specific objectives were: 1. To explore the current traceability practices in the fresh produce industry in New Zealand; 2. To investigate barriers in the design and implementation of tracking and tracing systems; 3. To demonstrate perceptions and attitudes to the uptake of traceability within the fresh produce industry; 4. To propose a reliable traceability framework for the domestic fresh produce industry to achieve an effective traceability system. 4 2.0 Literature review 2.1 Introduction Along with the high consumption of fresh fruit and vegetables due to healthy eating recommendations, the rate of foodborne illness outbreaks associated with fresh produce continues to increase worldwide (Callejon et al., 2015; Hanning, Nutt & Ricke, 2009; Harris et al., 2003; Prakash, 2016; Curtis, Yeager, Black, Drost & Ward, 2014). By implementing a reliable traceability system across the product value chain, food businesses can target the products affected by a food safety problem and therefore minimise any potential public health risks in a timely manner (Bertolini, Bevilacqua & Massini, 2006; Bello, Mirabella & Torrisi, 2005; Mainetti et al., 2013; Sun & Wang, 2019; Kim et al., 2018; Costa et al., 2013). Meanwhile, fresh produce is perishable, seasonal, fragile and has relatively short shelf life, thereby increasing the complexity and difficulty of traceability implementation in the fresh produce supply chain (Gokarn & Kuthambalayan, 2019; Wognum et al., 2011; Priyadarshani & Wickramasinghe, 2018). However, information on the traceability system of the domestic fresh produce supply chain is limited and little research has been conducted in New Zealand. United Fresh NZ Inc., the pan-industry fresh produce trade association, proposed a framework to encourage comprehensive interaction among the produce industry towards establishing a traceability system in New Zealand (Maurer, Dolan & Arts, 2015). The proposed framework was introduced by Maurer et al. (2015) and all fresh produce stakeholders were encouraged to take responsibility and work together as an industry. This proposed framework required a huge involvement and co-operation from all produce stakeholders, including government and regulators (i.e. Ministry for Primary Industries), New Zealand industry bodies, retailer businesses, research bodies, training providers, system verifiers, and so on (Maurer et al., 2015). Maurer et al. (2015) also stated that the framework could be beneficial for the entire produce supply chain as a more structured approach to achieve a robust food safety culture. Nevertheless, details of the implementation of an effective traceability system for the horticultural industry in New Zealand were not well described in this study. Effective traceability systems are important when the product has to move rapidly through the chain especially as a large number of products are consumed raw without further 5 processing (Mainetti et al., 2013). This means that if the product is contaminated with a microbial organism at any point in the supply chain and there is no kill step, it could lead to illness. Outbreaks of Hepatitis A in late 2015 linked to frozen berries and Yersinia pseudotuberculosis in late 2014 potentially associated with fresh carrots and lettuce have been reported in New Zealand (Ministry for Primary Industries, 2014). Therefore, the objective of this literature review was to understand the current traceability systems implemented in the fresh produce industry and identify any outcomes from previous research. Complexity and different roles of stakeholders of fresh produce supply chain in New Zealand were explored and discussed. Food traceability studies from other countries were also examined to assist with understanding of current systems and their application in the New Zealand domestic environment to improve fresh produce traceability system, both internally and externally. This review consists of six sections: foodborne illness outbreaks, characteristics of food traceability, food traceability technology, barriers of the implementation of traceability systems and their current shortcomings, fresh produce supply chain in New Zealand, and some stakeholders of fresh produce supply chain in New Zealand. 2.2 Foodborne illness outbreaks There has been a great number of foodborne illness outbreaks associated with fresh produce consumption over the past decades and many people suffered or even died from this (Julien- Javaux et al., 2019; Ministry for Primary Industries, 2014; Pouliot & Sumner, 2008; Hanning et al., 2009; Kirezieva et al., 2013; Sun & Wang, 2019). Contaminated alfalfa sprouts were linked to two large Salmonella outbreaks in the United States, Finland and Canada between 1995 and 1996 with more than 700 illnesses and one death reported (Hanning et al., 2009; Mahon et al., 1997; Van Beneden et al., 1999). It was concluded that contaminated alfalfa seeds were the main reason causing Salmonella infections after investigations (Mahon et al., 1997; Van Beneden et al., 1999). Some examples of foodborne outbreaks related to contaminated produce within the last two decades were summarised in Table 2.1. 6 Table 2.1 Examples of foodborne outbreaks due to contaminated produce between 1999 and 2019. Produce Pathogen Year No. of illnesses Reference Mung sprouts Salmonella 2000 45 Hanning et al., 2009; Mohle-Boetani et al., 2009 Tomatoes Salmonella 2004 561 Gupta et al., 2007; Hanning et al., 2009 Watermelon Salmonella 2006 20 Hanning et al., 2009 Bagged spinach Escherichia coli O157:H7 2006 204 (3 deaths) Pouliot & Sumner, 2008; Calvin, 2007 Alfalfa sprouts Salmonella 2008 13 Hanning et al., 2009 Cantaloupe Salmonella 2008 51 Hanning et al., 2009 Serrano peppers Salmonella 2008 1442 Hanning et al., 2009 Cantaloupe Listeria monocytogenes 2011 >130 (At least 26 deaths) Claiborn, 2011 Fenugreek seeds Escherichia coli O104:H4 2011 >3,000 (37 deaths) Bilinski et al., 2012; De Henriette, 2016; Sun & Wang, 2019 Fresh produce (likely) Yersinia pseudotuberculosis 2014 217 Ministry for Primary Industries, 2014 Cilantro Cyclosporiasis 2015 546 Julien-Javaux et al., 2019 Strawberry Hepatitis A virus 2016 143 Julien-Javaux et al., 2019 Romaine lettuce Escherichia coli O157:H7 2018 210 (5 deaths) Julien-Javaux et al., 2019 A huge number of studies were performed and investigations were undertaken to analyse the root cause and cross contamination. For instance, foodborne outbreaks associated with fresh fruit and vegetables were studied and reviewed to understand factors contributing to the incidences (Harris et al., 2003). It was indicated that various fresh produce has become a source of Salmonella outbreaks, such as watermelon, iceberg lettuce, orange, cantaloupe and strawberries (Harris et al., 2003). It was also mentioned that a hepatitis A virus outbreak linked to frozen berry fruits was investigated and infected workers and contaminated irrigation water were considered to be the potential causes of cross contamination (Harris et al., 2003). A number of foodborne illness outbreaks associated with fresh tomatoes 7 contaminated by Salmonella were investigated and discussed by some researchers (Dev Kumar et al., 2018; Hanning et al., 2009). In addition, the involvement of Salmonella in tomato associated outbreaks and its growth in soil and water were examined and studied (Dev Kumar et al., 2018). Another Salmonella outbreak associated with serrano peppers from Mexico was confirmed by FDA in 2008 with over 1000 illnesses reported (Hanning et al., 2009). An investigation was undertaken to trace products back through their distribution channels to the origins of products and contaminated irrigation water from a farm in Mexico was identified as the source of causing this outbreak in the United States (“FDA extends consumer,” 2008). Nevertheless, the root cause of contamination from a foodborne illness outbreak could not always be confirmed in spite of numerous time and labour being spent on the investigation. For example, a Yersinia pseudotuberculosis outbreak likely associated with fresh produce was reported in New Zealand in 2014 and resulted in a total of 217 cases linked to infection by Yersinia pseudotuberculosis from 1 September to 7 October (Ministry for Primary Industries, 2014). However, the source of contamination was not clearly identified even though huge efforts have been made to undertake investigations throughout the product supply and distribution chains (Ministry for Primary Industries, 2014). A hepatitis A virus outbreak associated with frozen strawberries was reported in 2016 and the Food and Drug Administration (FDA) carried out investigations to trace back the contaminated strawberry products (Julien-Javaux et al., 2019). It was found out that the strawberries were imported from Egypt (Julien-Javaux et al., 2019). However, it was still unknown of the root cause of the contamination on the Egyptian strawberries from investigations (Julien-Javaux et al., 2019). Furthermore, industry stakeholders such as produce growers could greatly be affected due to unexpected outbreaks linked to their fresh fruit and vegetables. For instance, the Escherichia coli O104:H4 outbreak in 2011 which caused over 3,000 infections and 37 deaths was initially thought to be linked to cucumbers produced in Spain (Abend, 2011). The reputation of Spanish producers was therefore greatly damaged and it was estimated that Spanish agriculture sales lost 200 million euros in the first week of the crisis (Abend, 2011). According to Pouliot and Sumner (2008), a Escherichia coli outbreak in September 2006 was linked to bagged fresh spinach and caused at least 200 illnesses in the United States. The contaminated spinach was traced back to the original packer Natural Selection Foods and eventually tracked back to one of four farms in California from the investigation (Pouliot & 8 Sumner, 2008). After six months of the E. coli outbreak, the consumption of bagged spinach still kept decreasing and below the level from previous year whilst the consumption of bunched spinach has rebounded (Calvin, 2007). 2.3 Characteristics of food traceability 2.3.1 Defining food traceability There were various definitions with regards to food traceability rather than a standardised definition applied in literatures, even though some of them were presenting very similar information (Bosona & Gebresenbet, 2013). For example, traceability was defined as the ability to trace and follow a food product through all stages of production, processing and distribution by European Parliament and Council (2002), whilst Souza-Monteiro and Caswell (2010) explained that traceability indicated an information flow amongst food companies and also involved interfirm coordination. The Codex Alimentarius Commission (1999) specified traceability as the ability to trace the history, application or location of an entity with recorded identifications. According to Olivier, Fourie and Evans (2006), food traceability was considered as a voluntary and regulative framework to bridge the information gaps among industry participants, including grower or farmer, food producer, retailer and consumer. An effective traceability system in the fresh produce industry indicates that any movement of products could be traced backwards and forwards at any point within the supply chain (Haleem et al., 2019). It is the ability to track the history, application or location of fresh fruit and vegetables through all stages of growing, processing and distribution, ranging from growers, packhouses, wholesalers, distribution centres to retailers (Canavari et al., 2010). Similarly, Mainetti et al. (2013) indicated that traceability was the ability to capture, collect and store data associated with all process operations within the supply chain in terms of the origin, location and life history of a product, thereby providing guarantee to both industry participants and the consumer. Charlebois and Haratifar (2015) described traceability as the ability to trace products back throughout the supply chain from farm or production to the end user. Manos and Manikas (2010) explained that fresh produce traceability in the primary industry was referred to the ability to trace the produce product history through the entire supply chain, ranging from product identification to operations undertaken during the product 9 life cycle. Overall, three key elements of food traceability were summarised by Bosona and Gebresenbet (2013) as tracing or backward follow-up of products, tracking or forward follow-up of products, and the history information of product along the supply chain. Bosona and Gebresenbet (2013) also presented the concept of food traceability with a schematic representation in Figure 2.1. Figure 2.1 Conceptual representation of product and traceability information flow in a food supply chain. (Bosona & Gebresenbet, 2013). In addition, according to ISO 22005:2007 standards, all food businesses were required to establish a one-up, one-down traceability system and to know both their customers and suppliers (International Organization for Standardization, 2006). The concept of one-up, one- down approach was further illustrated by Bosona and Gebresenbet (2013), including two questions that food businesses should ask themselves: 1. Who is the supplier/s of ingredients and/or partially processed food? 2. Who is the receiver/s of food items? Moreover, the European Parliament and the Council established the principles and requirements of general food law through regulation (EC) 178/2002 and created the European Food Safety Authority in 2002 (Souza-Monteiro & Caswell, 2010; Wognum et al., 2011). 10 General procedures in terms of food safety were provided and the implementation of traceability systems in the food supply chains in Europe was required by the European Parliament and the Council (Stranieri, Cavaliere & Banterle, 2017). 2.3.2 Advantages and disadvantages of food traceability Food traceability has always been a focus by food businesses to improve their food safety management systems and also gain customer confidence on their food products (Lindh & Olsson, 2010; Olivier et al., 2006; Haleem et al., 2019; Costa et al., 2013). In addition, Souza-Monteiro and Caswell (2010) demonstrated that traceability could be utilised as a tool to obtain product information and manage food safety risks linked to these products. Traceability system was also considered as an important tool to control and optimise food production, obtain better industrial data and make better decisions, as well as profile desirable product characteristics (Storoy, Thakur & Olsen, 2013). Furthermore, food traceability systems have also been widely applied in the food industry for meat and dairy products (Bennet, 2010; International Organization for Standardization, 2011; Mania, Delgado, Barone & Parisi, 2018). Traceability has emerged as a risk management tool that enables food companies or authorities to recall or withdrawal unsafe food products promptly (Mgonia et al., 2013). A traceability system was considered as an effective and efficient tool to manage product information flow between different parties within a supply chain (Souza-Monteiro & Caswell, 2010; Dai, Ge & Zhou, 2015). Additionally, it could enhance logistics processes and minimise the impact of food safety risks (Hobbs, 2004: Meuwissen, Velthuis, Hogeveen & Huirne, 2003). Furthermore, Cozzolino (2012) highlighted that food adulteration could be reduced and minimised by tracing and authenticating agricultural products. Bosona and Gebresenbet (2013) identified and explained the benefits of food traceability implementation in their study, including increased customer satisfaction, enhanced food safety incidents management, improved food supply chain management, competence development for food businesses, technological and scientific contribution, and contribution to agricultural sustainability. 11 Traceability systems enable food businesses to track products effectively and efficiently in the event of a product recall due to food safety issues, thereby preventing consumers from any potential illness or injury caused by this product purchase (Bai & Li, 2014; Olivier et al., 2006). In addition, as traceability systems identify specific products to be recalled, the other products within the same category could potentially be avoided suffering the same destiny and disruption of trading will be reduced (Olivier et al., 2006). For instance, when a batch of fresh orange associated with foodborne illness is provided by a grower and needs to be recalled from a retailer, effective traceability system enables this retailer to target the batch of orange and track it back to a specific grower, or even a specific field block. Corrective and preventative actions will be taken to investigate the food safety issue with the affected orange grower and all affected orange products will be isolated from other products. Therefore, all other orange growers with the same variety of orange can still supply their produce to the retailer and disruption of trading is minimised. Olivier et al. (2006) stated that the advantage of product tracing was the ability to react quickly and demonstrated the procedures of tracing product in the event of a food safety crisis. Typically, the affected product and issue were identified at first, then the origin of the product and its problem were identified through the supply chain. Subsequently, other products at risk and their locations within the supply chain were detected. Finally, appropriate actions were taken and if necessary product withdrawal or recall was carried out to minimise their risks to the public (Olivier et al., 2006). Costa et al. (2013) implied that food traceability in the agri-food sector was an efficient tool to share information and was potentially capable of providing more information for industry actors through the production and distribution chain. In addition, the time needed to intervene in the event of a food crisis was reduced in terms of recalling a whole stock from the market and individuating the real origin of the problem by applying an effective traceability system (Costa et al., 2013). Olivier et al. (2006) considered that it was important to maintain effective traceability in the food chain not only due to food safety risks but also because it may result in painful consequences from food scandals for both industry participants and those in positions of authority such as government. In addition, it was stated that a better fresh produce traceability system was needed to avoid loss of sales through food safety recalls due to illness outbreaks (Food & Drug Administration, 2007; Olivier et al., 2006). More importantly, a robust traceability system could help the Food and Drug Administration (FDA) to narrow its search 12 quickly and the costs caused by foodborne illness could be reduced to a minimum (Food & Drug Administration, 2007). It was stated that traceability could be used for food crisis management such as distinguishing product identity and food fraud prevention (Badia-Melis, Mishra & Ruiz-Garcia, 2015). Lower cost distribution systems and reduced product recall expenses could also be achieved by the application of a reliable traceability system (Golan, Krissoff & Kuchler, 2004). An effective traceability system could also promote food product sales as part of decision- making factors from buyers (Sun & Wang, 2019). Sun and Wang (2019) studied the preference of buyers such as retails in terms of sourcing products from suppliers such as food producers in a food supply chain. It was indicated that the buyer would prefer to source food products from the supplier who had a high-level traceability (Sun & Wang, 2019). In addition, Hobbs, Bailey, Dickinson and Haghiri (2005) used experimental auctions to examine consumer willingness in terms of paying for traceability in beef and pork products in the United States and Canada in their study. It was pointed out that American and Canadian consumers were willing to pay a premium for beef and pork traceability (Hobbs et al., 2005). However, Wognum et al. (2011) discussed challenges in food supply chains and explored factors influencing the buying decisions of consumers. It was concluded that consumers appeared not to be willing to pay more for better traceability (Wognum et al., 2011). However, there were also disadvantages of food traceability identified and discussed in previous research (Bosona & Gebresenbet, 2013; Souza-Monteiro & Caswell, 2010). Barriers and limitations to achieve effective food traceability implementation were discussed by Bosona and Gebresenbet (2013). There were five key limitations identified during their study from different directions, covering resource, information, standard, capacity and awareness (Bosona & Gebresenbet, 2013). Firstly, extra effort and cost were required for food businesses to develop and implement traceability systems and could potentially lead to financial issues. Secondly, traceability in the agricultural industry was linked to inherent uncertainty and sometimes it was difficult to obtain detailed information in a timely manner throughout all steps within the food supply chain. Thirdly, lack of sufficient and standardised information and methods of data exchange also became one of the main obstacles in terms of food traceability improvement and implementation. Fourthly, there was no adequately skilled employees that able to develop, implement and manage the traceability due to the complex nature of food traceability system in the supply chain. Finally, traceability implementation 13 was considered as extra burden by industry partners and there was a lack of clarity information to understand traceability and its advantages, leading to concerns of the investment cost over its benefits and causing lack of awareness as well as less willingness to achieve traceability implementation (Bosona & Gebresenbet, 2013). On the other hand, Souza-Monteiro and Caswell (2010) stated that food traceability was costly because it required obtaining, storing and sharing information. Furthermore, the benefits of a food traceability system may not be evenly distributed throughout the supply chain (Souza-Monteiro & Caswell, 2010). Bosona and Gebresenbet (2013) undertook a comprehensive literature review on food traceability systems along different supply chains and indicated five main factors or concerns influencing the development and implementation of food traceability: regulatory concern, concerns of food safety and quality, social concern, economic and technological concerns. 2.3.3 Internal and external traceability According to GS1 (2010), there are two different types of traceability: internal and external. Whilst, food traceability could also be categorised into internal traceability and chain traceability (Moe, 1998; Souza-Monteiro & Caswell, 2010). Similarly, Lindh and Olsson (2010) named external traceability as chain traceability and explained that participation and co-ordination of all actors from the supply chain were important to chain traceability. Good co-operation within all actors of the supply chain was required for an effective chain traceability system to ensure that it could be implemented successfully (Lindh & Olsson 2010). However, Olivier et al. (2006) emphasized that the supply chain co-operation was very difficult to achieve since it involved traceability information synchronisation through the supply chain and required a high level of transparency and collaboration between industry players. Both internal and external traceability systems play important roles in the fresh fruit and vegetables supply chain. Internal traceability was considered as the proprietary data that a company used within its operations, but external traceability was the information exchange between trading partners (GS1, 2010; Souza-Monteiro & Caswell, 2010). Internal traceability allows people to track fresh produce through all steps of process flow within the business, 14 while external traceability is associated with tracking fruit and vegetables outside of the business but within the supply chain (Rong & Grunow, 2010). According to Mgonia et al. (2013), internal traceability focuses on product activities within one company and it relates to traceability information of raw materials and processes to the final product prior to delivery. On the other hand, external or chain traceability is associated with data from one level to the next across the supply chain (Rong & Grunow, 2010). Souza-Monteiro and Caswell (2010) explained that chain traceability tracked the history of a batch within the value chain, including production, transport, storage, processing, distribution and sales. For example, a produce wholesaler is able to track its produce products from product arrival, quality checks, temporary storage, dispatch to delivery within the wholesaler through its internal traceability system. However, external traceability becomes vital when a product recall occurs and fresh produce purchased by consumers needs to be traced back to where it was grown and when it was packed. A slightly different description was stated by Bello et al. (2005) in terms of internal and external traceability but still with fundamentally the same principles. Bello et al. (2005) indicated that two monitoring systems were required to achieve a full traceability: one system operated at production level and recorded product and process data from each stage of production; the other system operated at factory level and coordinated traceability information received from the first one for logistics (Bello et al., 2005). 2.3.4 Components of food traceability systems Recently, Olsen and Borit (2018) defined Traceable Resource Unit (TRU) and described different components of a traceability system. According to Olsen and Borit (2018), a TRU in a traceability system was considered as a traceable object or product along a supply chain. The same concept of TRU was also proposed by Fan et al. (2019) and used to test a food traceability system and equipment in their study. In addition, Dabbene and Gay (2011) considered TRU as the foundation to implement an effective traceability system and stated that TRUs must be defined in the traceability system. A TRU was a uniquely identifiable unit and it needed to be consistent with traceability information recorded along the supply chain (Fan et al., 2019; Olsen & Borit, 2013). 15 There were three different types of TRUs identified in a traceability system (Olsen & Borit, 2018; Aung & Chang, 2014). A TRU could be a trade unit such as a box or a crate, a logistic unit such as a pallet or a production unit such as a batch (Olsen & Borit, 2018). Nevertheless, Aung and Chang (2014) proposed three types of TRUs which were slightly different compared to Olsen and Borit (2018). Batch, trade unit and logistical unit were considered as three types of TRUs by Aung and Chang (2014). A batch was a quantity or a number of products that went through the same processes. A trade unit was described as a unit delivered from one company to the next within a supply chain while a logistical unit was referred to a type of trade unit created as a group by a company prior to its transport or storage (Karlsen, Donnelly & Olsen, 2011). Olsen and Borit (2018) analysed and discussed different components of a food traceability system. It was concluded that there were three main components of a traceability system, including a mechanism for identifying each TRU throughout its entire life cycle, the way of recording TRU relationships (e.g. connections between TRUs as they move through the supply chain), and the way of documenting all information relating to TRU attributes (Olsen & Borit, 2018). Firstly, a traceability system has to be able to identify the unit traced and therefore each unit must be unique and identifiable from the others. Secondly, the joining and splitting of units in the entire supply chain need to be recorded and documented in the traceability system to clearly show the product movements. Finally, the traceability system should also capture data of unit attributes such as product weight, location and temperature (Olsen & Borit, 2018). The transformation of TRUs was commonly manifested by package splitting and aggregation (Fan et al., 2019). Fan et al. (2019) illustrated the joining and splitting of TRUs within a typical supply chain in current food traceability systems in Figure 2.2. 16 Figure 2.2 Process flow of traceable resource units (TRUs) aggregation and splitting through a typical supply chain in current food traceability systems. (Fan et al., 2019). However, Opara (2003) considered that an integrated agricultural and food value chain traceability system consisted of six important elements: product traceability, process traceability, genetic traceability, inputs traceability, disease and pest traceability, and measurement traceability. Product traceability determines physical locations of products at any stage within the supply chain while process traceability ascertains the type and sequence of activities applied on products. Genetic traceability covers genetic formation of products and inputs traceability determines type and origin of inputs such as fertilizer, chemical sprays and irrigation water. Disease and pest traceability traces the epidemiology of pests and biological hazards. Measurement traceability relates individual measurement results to accepted reference standards (Opara, 2003). Olivier et al. (2006) pointed out that there were three main types of traceability for fresh produce. Firstly, the origin of a fresh produce was identified and subsequently the relevant product records was linked to the farm. Secondly, a product was tracked within a facility or internal processes such as packing or palletising. Thirdly, fresh produce products were tracked and traced through the value chain, commonly with a ‘one step forward and one step back’ approach (Olivier et al., 2006). 17 2.4 Food traceability technology In order to establish an efficient and effective traceability system along the supply chain, innovative technologies were required for product identification, process characterization, system integration and dealing with relevant data such as capture, analysis, storage and transmission (Mainetti et al., 2013; Borrero, 2019; Dabbene & Gay, 2011). These technologies involved in both hardware and software solutions (Mainetti et al., 2013). The hardware solutions included identification tags and labels for products, logistic units and locations while software solutions covered computer programs and information systems (Mainetti et al., 2013). In particular, a wide range of technologies were applied in food traceability to identify and trace products along their value chains (Abdullah, Nurilmala & Sitaresmi, 2019; Fan et al., 2019; Foster, Buskirk, Schweihofer, Grooms & Clarke, 2018; Institute of Food Technologists, 2019; Ishiyama, Kudo & Takahashi, 2016; Qiao, Wei & Yang, 2013; Costa et al., 2013; Dandage, Badia-Melis & Ruiz-Garcia, 2017). Specifically, barcodes, radio frequency identification (RFID), near field communication (NFC) and other technologies such as blockchain were examined and discussed in many studies in the past years (Abdullah et al., 2019; Alfian et al., 2017; Bai et al., 2017; Bello et al., 2005; Feng, Fu, Wang, Xu & Zhang, 2013; Ghaani, Cozzolino, Castelli & Farris, 2016; Luvisi et al., 2012; Mainetti et al., 2013; Sermpinis & Sermpinis, 2018; Kim et al., 2018; Bibi, Guillaume, Gontard & Sorli, 2017; Galvez, Mejuto & Simal-Gandara, 2018). The different technologies applied in the food industry and their benefits were summarised in Table 2.2 below. Table 2.2 Application of different food traceability technologies in the food industry. Technology Advantage Application Reference DNA mini- barcodes As a molecular marker to prevent food fraud and to improve traceability system Hairtail fish products authentication Abdullah et al., 2019 Barcodes Cost-effective; Ease of use Animal identification Bai et al., 2017; Ghaani et al., 2016 RFID Real-time; Accurate data acquisition and transmission; Environmental monitoring in a kimchi supply chain; Alfian et al., 2017; Feng et al., 2013; Luvisi et al., 2012; 18 High efficiency Product tracking in the beef industry; Grapevines Bibi et al., 2017 NFC Derived from the RFID family; Inherit basic features of RFID but could also share data across active devices Piloted in a vegetable supply chain for location identification Mainetti et al., 2013 Blockchain Tamperproof; Transparent; Real-time; High efficiency and security Theoretical study and pilot only so far Sermpinis & Sermpinis, 2018; Kim et al., 2018; Galvez et al., 2018 Barcodes have been widely applied in retail over the last three decades to facilitate their inventory control and stock recording as a low cost and easy to use technology (Ghaani et al., 2016; Bibi et al., 2017). Barcodes were classified as one-dimensional (1D) barcode and two- dimensional (2D) barcode (Fan et al., 2019). 1D barcode was also named as linear barcode (Bello et al., 2005). It was a pattern of parallel spaces and bars arranged to represent a series of digits and its encoded information was read by an optical barcode scanner (Bibi et al., 2017; Fan et al., 2019). Subsequently, the relevant information was sent to a system to be stored and processed during the scanning stage (Fan et al., 2019). 2D barcode was associated with composite codes and contained multi-row bars in a matrix (Bello et al., 2005). Compared to 1D barcodes, 2D barcodes were able to store more information by combining dots and spaces arranged in an array or a matrix rather than bars and spaces (Fan et al., 2019; Liang et al., 2013). In particular, the quick response (QR) code was a typical 2D barcode used widely in product identification and traceability on labels (Fan et al., 2019). QR code technology was very similar to 1D barcode but it could only become readable by a QR code reader or scanner (Seino et al., 2004). RFID has drawn attention from the public since it was implemented by Wal-Mart in its supply chain in 2003 with publicly announced initiative (Singh, Singh, Desautels, Saha & Olsen, 2010). Recently, RFID has increasingly been adopted in the food industry for product traceability systems due to its superiority of automated and highly precise data recording (Bai et al., 2017; Bello et al., 2005; Mainetti et al., 2013; Bibi et al., 2017). It was a flow control technology which enabled traceability of products along all stages of the supply chain to be captured (Costa et al., 2013; Kelepouris, Pramatari & Doukidis, 2007; Ngai, Moon, Riggins & Yi, 2008). Additionally, RFID became an important technology applied in traceability due 19 to its advantages such as eliminating the manual control, improving versatility of operational and logistical contexts, long service life and miniaturisation of equipment and devices (Sarac, Absi & Dauzere-Peres, 2010). RFID was composed by three different elements: an RFID tag, an interrogator and a database system (Costa et al., 2013; Bibi et al., 2017). The tag was generally applied on the displacing product directly and the interrogator was a device such as a reader to capture data from the tag, while the database system or a host computer was to store the data obtained through the interrogation process (Costa et al., 2013; Bibi et al., 2017). The RFID tag normally had an identification code and a microchip unit containing memory storage. The tags were classified as read-only and read-write tags (Costa et al., 2013). The former could be read in multiple times but could not be modified, while the latter could be both read and modified for several times (Costa et al., 2013). RFID technology allowed contactless identification of products and effective information sharing with customisation and handling (Bosona & Gebresenbet, 2013; Fan et al., 2019). Olivier et al. (2006) implied that RFID enabled food products to be tracked and monitored across the supply chain without human intervention. RFID was also considered as a technology that could be used to enhance the information flow management through the food supply chain and improve security in the agri-food industry (Costa et al., 2013). RFID technology was more convenient to identify products compared to barcode systems because visual contact was not required by RFID tags and therefore they could be placed into product containers, injected into animals or embedded into an object (Bibi et al., 2017). However, barcodes were still widely used and dominated on the market because the cost of barcode systems was less than RFID technology (Bibi et al., 2017). The differences between barcodes and RFID technology discussed in literatures were summarised in Table 2.3. Table 2.3 Comparison between barcodes and RFID technology. (Fan et al., 2019; Ghaani et al., 2016). Attribute/feature 1D barcode 2D barcode RFID Technology Optical (laser) Optical (laser) RF (radio frequency) Environment condition Sensitive to environment, dirt, scratches Sensitive to environment, dirt, scratches Durable, waterproof Security Low Higher Highest 20 (non-encryption) (simple encryption) (deep encryption) Reading distance Near Near Far Storage capacity Very small (only express numbers) Small (can express alphabet or other characters) Big (can store data around 32-128Bit) Reading/writing Cannot be updated Cannot be updated New information can be over-written Price Cheap Cheap Expensive A RFID based traceability system applied in beef production was examined and explored (Feng et al., 2013). In particular, it was concluded that real-time and accurate data acquisition and transmission were achieved by applying the RFID-enabled system. Furthermore, it was highly efficient to track traceability information throughout the beef production supply chain with the application of RFID based traceability system (Feng et al., 2013). Recently, a barcode-RFID bidirectional transformation equipment was used by Fan et al. (2019) to enhance continuous traceability systems in the food industry. Apart from this, NFC was demonstrated as a short-range wireless technology derived from RFID but it was distinguished by the ability to share information through active and powered devices (Mainetti et al., 2013). It was illustrated that blockchain technology could have great applications across a number of areas in the food industry especially for traceability (Institute of Food Technologists, 2019). Borrero (2019) also considered that the implementation of blockchain technology and cooperation of industry participants could be beneficial to the horticultural industry by providing more transparency and reducing the risk of food fraud and adulteration. Blockchain technology used hash-based cryptology to ensure security and trust and it had three essential parts of data: transaction details, the transaction timestamp and a new hash connecting the hash and details from previous transaction (Institute of Food Technologists, 2019; Pierro, 2017). A hash was described as an encrypted type of a string or sequence of characters and it was considered as computationally impossible to derive the original without a key (Institute of Food Technologists, 2019). Subsequently, each transaction was distributed throughout the network and therefore a continuous encrypted record of the transaction was kept and became immutable once added to the blockchain (Institute of Food Technologists, 2019; Pierro, 2017). Transactional and distributed ledger functionality were provided via 21 blockchain to allow different industry parties to operate without the requirement of a centralised and trusted authority (Galvez et al., 2018). Blockchain technology enabled a more transparent and decentralised system across companies within the supply chain and allowed businesses to add information into the system with a level of anonymity and control (Institute of Food Technologists, 2019). Every computer or node within the network stored a copy of the blockchain and the nodes were periodically synchronised to ensure that all were sharing the same database (Galvez et al., 2018). In addition, different companies throughout the food supply chain could potentially add data into a shared ledger and the shared ledger could reach both ends of the value chain from grower to consumer (Institute of Food Technologists, 2019; Tian, 2017; Galvez et al., 2018). Digital product data such as grower details, batch numbers and logistical information were connected to food products and relevant information was added into the blockchain at each stage of the process along the supply chain with the application of blockchain technology (Galvez et al., 2018). Furthermore, industry participants could add traceability data while keeping important proprietary information hidden (Institute of Food Technologists, 2019). However, Galvez et al. (2018) pointed out some challenges in implementing a blockchain technology on food traceability within the value chains. Its complexity made it difficult to be adopted and all industry participants in the supply chain had to collaborate to implement (Iansiti & Lakhani, 2017). Additionally, there was a lack of understanding and standards for the implementation of blockchain since it was still at an early stage of development (Galvez et al., 2018). Moreover, all involved parties needed to operate on an agreed type of blockchain, which made them under pressure (Galvez et al., 2018). Noticeably, costs of technologies related to food traceability systems was still considered as high and therefore blocked their wide adoption in the fresh produce industry especially for low cost products (Mainetti et al., 2013). 22 2.5 Barriers of the implementation of traceability systems and their current shortcomings A variety of studies have explored the implementation of traceability systems and discussed its barriers in the food industry (Donnelly et al., 2009; Feng et al., 2013; Mania et al., 2018; Mgonia et al., 2013; Mutua et al., 2018; Manikas et al., 2010; Olivier et al., 2006). For example, in the meat, dairy and seafood industry, Mgonia et al. (2013) introduced a diagnostic tool to assess company traceability systems internally for fish processing businesses. They pointed out that traceability systems were generally developed at company level and therefore only limited traceability data were provided. In addition, these data were fragmented and uncoordinated in approach through the entire supply chain (Mgonia et al., 2013). A RFID based traceability system in beef production was developed and assessed by Feng et al. (2013). The process flow for beef production along the supply chain and key information from the beef products traceability system were identified via a survey (Feng et al., 2013). Recently, a livestock identification and traceability system (LITS) was designed and piloted by Mutua et al. (2018) within the northern Tanzania-Narok-Nairobi beef value chain. Traceability data from identified animals along with the value chain were collected and added into an online database and meat samples were also analysed for tetracycline and diminazene residues. In addition, a questionnaire survey was carried out and stakeholders from the beef value chain such as traders, producers and transporters were interviewed at the end of the pilot to understand their perceptions and level of acceptance on the LITS. Cui et al. (2018) introduced a systematic modelling approach to capture relevant critical traceability information from the sheep meat supply chain in China. A process modelling tool named petri nets (PNs) was proposed and used in this study. Advantages of PNs based information traceability model were also explained and illustrated by Cui et al. (2018). Some studies were also undertaken to determine the implementation of traceability systems in the fresh produce industry worldwide. However, there was limited research carried out to explore the implementation of traceability system in the fresh produce industry in New Zealand and challenges of its implementation were not well understood in detail. A case study was conducted in Greece to examine a variety of factors affecting the efficiency of traceability with regards to fresh produce supply chain. Semi-structured in-depth interviews 23 with key personnel from each company were employed and multi fragmentation and lack of vertical integration were observed in this fresh produce supply chain (Manikas et al., 2010). Traceability as part of information management in the fruit supply chain was also explored by Canavari et al. (2010) through semi-structured in-depth interviews with key personnel from the Italian fresh produce supply chain. There were 17 key informants interviewed from the fruit supply chain, including producers, co-operatives, wholesalers, major and small retailers, and a catering company. Furthermore, interviews were carried out at decision-making level with six topics covered in the interview questions, including information about the company, product management, information management, purchasing needs versus company capabilities, co-ordination issues, and compliance with other management systems and voluntary certifications (Canavari et al., 2010). It was considered that data and information were managed and transferred in different ways by fruit businesses within the Italian supply chain, depending on their positions in the fruit supply chain and their mission as well as resources (Canavari et al., 2010). Additionally, Canavari et al. (2010) stated that poor co-ordination with food business operators among different stages of the fruit supply chain caused transaction costs and resource wastage, especially at the control stages. Specifically, control costs were created by quality controllers in the field, warehouse, or retailers to control produce quality and verify the correspondence to their signed agreements. However, control of the same product in various ways by different produce companies caused inefficiency and unnecessary costs in the entire supply chain system. The costs for physical samples, laboratory tests, labour cost, instruments and transport costs could potentially be minimised with a better co-ordination system. Olivier et al. (2006) carried out a traceability study in the fruit export industry in South Africa and semi-structured interviews were conducted with 27 key stakeholders and experts from the industry. It was highlighted that information fragmentation and the demand of managing costs carefully in a highly competitive market resulted in the need for effective traceability system within the entire supply chain (Olivier et al., 2006). A specific study with regards to fresh produce traceability implementation was undertaken and its drivers and constraints were examined by Manos and Manikas (2010) in the Greek supply chain. Interviews of key representatives from the Greek agricultural businesses were 24 carried out and qualitative data were collected using a questionnaire in this study. It was concluded that tight profit margins from the agricultural businesses and inadequate knowledge of understanding potential benefits from a reliable traceability system have become challenges in terms of the traceability implementation (Manos & Manikas, 2010). Gichure et al. (2017) performed a case study on traceability systems along organic kale supply chain in Kenya to understand the factors influencing its implementation and maintenance. Interviews were carried out with organic kale farmers, traders and farmers’ market officials using semi-structured questionnaires in this study. According to Gichure et al. (2017), it was suggested that the there was a need to increase the awareness of industry stakeholders on traceability through organisational activities and enhanced information flow to achieve safer products. Mainetti et al. (2013) investigated food traceability system through the fresh vegetables supply chain using RFID technology and implemented a pilot project in one of the largest fresh vegetables producers in Italy. A gapless traceability system using RFID technology in both greenhouses and the processing factory was proposed by Mainetti et al. (2013). Kim et al. (2018) introduced an integrated ‘farm to fork’ food traceability system using blockchain technology and illustrated this theoretical system in detail. The proposed traceability system for the agricultural value chain applied blockchain distributed ledgers to streamline data sharing across the network and allowed all industry participants to have access to a trustless source of data (Kim et al., 2018; Miller, 2018). The blockchain ledger was updated accordingly with process information and delivery trucks were tracked while in transit as products moved through facilities and loaded onto trucks. Product traceability information such as temperature and other logistic data were transmitted into the blockchain ledger at the same time. When products arrived at the destination and accepted by the buyer, contractual payment for products was executed through triggering blockchain smart contract by electronic notification. Therefore, the growers were able to monitor their products for when it reached the consumer through the blockchain enabled supply chain. Meanwhile, the consumer could scan a product at a retailer store and find relevant information related to the product, including where the product was from, how long it was in the store and other logistic details (Kim et al., 2018). Borrero (2019) developed a proof of concept for agri-food supply chain traceability to explore the possible implication of blockchain application in the berries value chain in Spain 25 recently. It showed that the adoption of blockchain technology in terms of traceability provided added value in the agri-food supply chain. In addition, it was feasible to combine all data from the field, processing factory, packaging and transporting into a chain of blocks using an authorised ledger and a smart contract (Borrero, 2019). Compared to traditional methods with centralised databases in different companies knowing different stages of product information through its value chain, the same level of relevant information from industry participants could be shared using blockchain technology within the whole industry (Borrero, 2019). According to Wognum et al. (2011), an international benchmark study was carried out in Australia, The Netherlands, Germany, Spain, Sweden and the United Kingdom as well as the United States of America. It was indicated that traceability performance levels between food supply chains differed while the differences among the countries were not very large (Wognum et al., 2011). Legislation was considered as an important motivation for food businesses to meet requirements of traceability standards (Wognum et al., 2011). However, there was no clear requirement or standard identified from current legislation to provide details for food companies to implement full traceability and most legislations focused on in- company traceability instead of full supply chain traceability (Wognum et al., 2011). One of the consequences was that most food companies placed emphasis on their own business rather than the entire supply chain in terms of traceability implementation and the majority had developed certain traceability across company borders (Wognum et al., 2011; Luning, Devlieghee & Verhe, 2006; Van der Vorst, 2006). In addition, most food businesses were focusing on the prevention of product recall rather than traceability implementation because they could not always obtain benefits from traceability especially when the traceability outcomes were not very precise. It was common that retailers removed all batches of the food products from the shelf rather than only the specific batch concerned in the event of a food safety crisis (Wognum et al., 2011; Luning et al., 2006; Van der Vorst, 2006). Remarkably, vulnerabilities of produce supply chain were discussed intensively and traceability was identified as a key issue by the government and industry due to the food tampering incidents which occurred in Australia and New Zealand, involving sewing needles inserted into fresh strawberries in September 2018 (Food Standards Australia New Zealand, 2018). It was concluded that some factors affecting an effective traceability system included: • Lack of regulatory requirements in the horticulture sector; 26 • The fragmented nature of the sector, with many small companies and no regulatory or industry oversight; • The current practice of mixing and combining produce from more than one grower or supplier; • The complexity of the supply chain; • Loosely sold produce without any packaging (where product traceability information could be present); • The seasonal nature of labour hiring practices resulted in difficulties in monitoring workers. Furthermore, it was indicated that the ‘one step forward, one step backward’ approach to traceability was not sufficient and traceability along the supply chain was suggested to be better understood to improve supply chain integrity (Food Standards Australia New Zealand, 2018). 2.6 Fresh produce supply chain in New Zealand 2.6.1 Characteristics of fresh produce and its supply chain Fresh fruit and vegetables are considered as foods from plant origin with limited shelf life (Gokarn & Kuthambalayan, 2019; Hospido et al., 2009; Clements, Lazo & Martin, 2008). In addition, there are many uncertainties in the fresh produce supply chain due to their intrinsic properties, including seasonality, perishability and quality variation (Gokarn & Kuthambalayan, 2019). The high level of perishability and fragility of fresh produce can result in high wastage levels and therefore make it very difficult for retailers to manage when consumers require fresh and high-quality produce (Clements et al., 2008). Horticultural products are highly perishable due to their nature that fresh fruit and vegetables continue metabolic processes after harvesting, which lead to their ripening or senescence and eventually unmarketable state (Falagan & Terry, 2018; Mahajan et al., 2017; Kramer, Wunderlich & Muranyi, 2018; Song, He & Xu, 2019; Mahajan et al., 2014). The perishable nature of fresh produce and variability in quantity and quality due to weather conditions, seasonality and biological variation increases the difficulty and complexity of traceability 27 implementation in food supply chains (Wognum et al., 2011; Priyadarshani & Wickramasinghe, 2018). Clements et al. (2008) also indicated that the biological nature of horticultural products posed further challenges to maintain product quality and continuous supply, including seasonal production, unpredictable weather, pest of disease outbreaks. Hence, in order to maintain the quality of fresh produce and reduce product losses caused by its perishability, industry coordination from all participants such as growers, storage operators, processors, shippers and retailers are required (Mahajan et al., 2017; Ali, 2016). Furthermore, fresh fruit and vegetables have short deterioration time and are easily contaminated, making the management of their value chains especially important (Gokarn & Kuthambalayan, 2019). For instance, Codron et al. (2014) explained that fresh produce growers in the supply chain were considered as one of the sources of product loss caused by chemical contamination from pesticides. Additionally, it was stated that produce spoilage and postharvest loss caused by deterioration and contamination could occur at any stage across the fresh produce supply chain (Mahajan et al., 2017). Blackburn and Scudder (2009) also implied that fresh produce deteriorated rapidly and their value decreased significantly over time along the supply chain at rates which were highly temperature and humidity dependent. The characteristics of fresh produce supply chain were summarised by Gokarn and Kuthambalayan (2019) in Table 2.4 below: Table 2.4 Characteristics of fresh produce supply chain. (Gokarn & Kuthambalayan, 2019). No. Element Characteristics 1 Nature of product Perishable 2 Nature of supply chain Complex, inefficient 3 Nature of demand Fluctuating 4 Wastage High 5 Cost pressure High 6 Product range Diverse 7 Dependencies over climate High 8 Sector Agriculture, unorganised 28 According to Mahajan et al. (2017), there were many factors influencing the perishability and deterioration of fresh fruit and vegetables since each individual produce had inherent physiologies and biochemistries. For example, different skins of fresh produce are associated with different gas exchange, water loss and rates of metabolism especially respiration, which are linked to various storage potentials (Mahajan et al., 2017). Additionally, the harvested part from plants could vary depending on different developmental stages of products, including sprouts, stems, leaves, inflorescences, partially developed fruit, fully developed fruit, roots and tubers (Mahajan et al., 2017). Furthermore, even for the same fresh produce, the harvest time and physiological characteristics such as shape, size, colour, and sugar level vary depending on different varieties, which extends the diversity of the fresh produce products (Mehdi, Ahmad, Yaseen, Adeel & Sayyed, 2016). There have always been fluctuations in the supply of agricultural produce as fresh fruit and vegetables are seasonal products and they are only available in a short period of time during the year (Van Walbeek, 1996; Mehdi et al., 2016; Ge, Canning, Goetz & Perez, 2018). A typical example is that mango produced in Pakistan is only available during summer as a type of seasonal fruit (Mehdi et al., 2016). The period for apples harvesting in Hawkes Bay in New Zealand was estimated to be from February to May depending on the cultivar (Goossens, 2019). Van Walbeek (1996) stated that climatic and weather conditions largely influenced the production of fresh produce, thereby leading to the seasonal variation in the quantity and quality of produce. For instance, Hospido et al. (2009) pointed out that some north European countries were not able to grow fresh produce all year round because of their climatic constraints. Furthermore, Ge et al. (2018) explained that each variety of fresh produce was harvested in a given range of weeks within the harvest period and the products had to be consolidated quickly and carefully. Noticeably, the fluctuation and seasonality of fresh produce supply could also be influenced by extreme climate events such as heavy rainfalls and floods. For example, contamination of irrigation water sources due to heavy rainfalls or floods would potentially result in food safety issues and product loss in the agricultural produce industry (Kirezieva, Jacxsens, Van Boekel & Luning, 2015). 29 2.6.2 Relationship in fresh produce supply chain in New Zealand Fresh produce supply chain involved a great deal of different industry actors, including farmers or growers, manufacturers or processors, wholesalers, logistics, retailers, supermarkets, and others (Borrero, 2019; Codron et al., 2014; Jadav, Leua & Darji, 2011; Givens & Dunning, 2019). Hardesty et al. (2015) categorised produce suppliers into different groups by the type of retail outlet and the number of stores they own: farmers market, independent vendor representing single store, neighbourhood store meaning 2 to 6 stores, and chain store for retailers owning more than 7 stores. Fouayzi, Caswell and Hooker (2006) illustrated the produce supply chain and considered that it covered growers, packers, processors, shippers, wholesalers, retailers, food service operations and consumers. According to Worinu (2007), industry actors along the fresh produce supply chain included supermarkets, hotels, restaurants, fast food outlets, urban and roadside markets and others. These could be generally categorised into two main systems: formal and informal ones (Worinu, 2007). Formal supply channel or market is also named as direct market and refers to when fresh produce is delivered from growers or packhouses through wholesalers to retailers such as supermarkets, the relationships between different parties within the supply channel are generally characterised by company agreements and products are supplied on an ongoing basis (Trienekens, Van Velzen, Lees, Saunders & Pascucci, 2017). There are some relationships built up between the suppliers and buyers (Worinu, 2007). Whilst, informal supply channel or market is commonly referred to open market and is related to purely trust and commitment developed between industry actors such as farmers and buyers and products are provided occasionally or periodically without agreed contracts (Givens & Dunning, 2019; Timsina & Shivakoti, 2018). For example, fresh fruit and vegetables sold via roadside stalls and farmers markets by growers are considered as informal supply channel and there is no long-term relationship between farmers and buyers as products are supplied occasionally (Worinu, 2007). Curtis et al. (2014) interpreted that farmers markets provided local growers a great opportunity to eliminate costs of the middle man and it was easier for small producers to participate in with relatively lower vendor fees and limited contractual obligations. It was indicated that the number of industry participants with formal and informal relationships across a food supply chain could be large (Wognum et al., 2011). Relationships in fresh produce supply chains have been changing over the past few decades due to demand conditions, new technologies, private supply chain requirements and public 30 regulatory standards (Fouayzi et al., 2006). An adversarial relationship with multiple suppliers has gradually become a closer and on-going relationship with a few selected suppliers or even exclusive supplier (Clements et al., 2008; Kalwani & Narayandas, 1995). This trend could be driven by various reasons (White, 2000; Fearne & Hughes, 2000; Hingley & Lindgreen, 2002). White (2000) indicated that on-going relationship allowed retailers to have access to adequate fresh produce products and maintain the continuity of supply. In addition, a competitive advantage over other retailers was obtained through exclusive access to the best raw material producers (Fearne & Hughes, 2000; Hingley & Lindgreen, 2002). Furthermore, search time and cost were reduced and therefore productivity was improved through the integration of retailer and supplier systems (Hingley & Lindgreen, 2002; Brookes, 1995). Meanwhile, this on-going relationship would also benefit producers from selected suppliers through increased security and reduced risk (Hingley, 2001). Lee and Nuthall (2015) carried out a case study on supplier commitment to agri-food supply chains in New Zealand and factors attracting suppliers to be committed to long-term relationships were examined. According to Lee and Nuthall (2015), increased price certainty, premium prices and related quality were considered as three main factors. Clements et al. (2008) investigated fresh produce supply chains in New Zealand focusing on relationship connectors, supply chain functions and product characteristics. In addition, two case studies of fresh produce supply chains in the South Island of New Zealand were undertaken under this research. Clements et al. (2008) stated that many factors in the fresh produce supply chain could impact on relationships within a supply chain for both between chain partners and within companies themselves, such as difficulties of guaranteeing continuity of supply due to seasonal and unexpected shortage or oversupply, logistics and quality management, information exchange and process alignment along the chain. 2.6.3 Fresh produce supply channels in New Zealand There are several studies undertaken to understand the fresh produce supply channels in New Zealand over the past two decades (Maurer, 1999; Clements et al., 2008; Ministry for Primary Industries, 2014). Maurer (1999) introduced a framework to describe a typical produce supply chain in New Zealand in Figure 2.3 as below. 31 Figure 2.3 A typical fresh produce supply chain in New Zealand. (Maurer, 1999). Fresh produce was initially harvested and transported to packhouse or storage facility, through distribution centre or market and subsequently delivered to stores for consumer purchasing (Maurer, 1999). Noticeably, Maurer (1999) explained that the all possible supply chain variations for New Zealand produce were incorporated in the framework, however, not all produce products moving along the supply chain necessarily passed through all stages. However, the fresh produce supply chain is not always linear (Ministry for Primary Industries, 2014). Products could reach the consumer through a number of different routes (Ministry for Primary Industries, 2014). For instance, fresh produce may be purchased by consumers via farmers markets or online sales from the grower directly. In addition, products could also be supplied by growers and transported to retail stores nationally (Ministry for 32 Primary Industries, 2014). Givens and Dunning (2019) demonstrated that fresh fruit and vegetables could be delivered by farmers directly to restaurants as a short supply channel to avoid extra expense via intermediary distributors. Moreover, a food supply chain was also considered as a network since industry participants in the supply chain commonly had several or many suppliers and customers which made the whole chain more complex (Wognum et al., 2011). The network of fresh produce supply channels in New Zealand was illustrated in Figure 2.4 (Ministry for Primary Industries, 2014; Clements et al., 2008; Maurer, 1999). Figure 2.4 Complexity of fresh produce supply channels in New Zealand. (Ministry for Primary Industries, 2014; Clements et al., 2008; Maurer, 1999). According to Clements et al. (2008), some retail stores throughout New Zealand were individually owned but managed by one organisation who had a cooperative buying structure to source most of fresh fruit and vegetables for stores, whereas the other retail outlets were owned by one company who had a corporate business structure around the country. Products Prepacked Market Wholesalers Packhouses Growers Retail stores Loose packed Often the same company C o n s u m e r s Farmers markets or online sales Product Movement (Note that information paths are excluded and this only shows product movement along the fresh produce supply chain in New Zealand). 33 sourced by organisations for retail stores could be from market wholesalers or directly from their own preferred growers (Clements et al., 2008). In addition, market wholesalers also had their own preferred and non-preferred growers where they purchased fresh produce from (Clements et al., 2008). 2.6.4 The scale of the fresh produce industry in New Zealand It was indicated that New Zealand was a major agricultural producer as an island nation and free from many diseases and pests around the world (Webb, Strutt & Rae, 2016). The geographical restriction impeded both importing and exporting of fresh fruit and vegetables to certain extent in the country (Webb et al., 2016). As a relatively small country in the world, the total population of New Zealand was 4.762 million in 2017 and was expected to grow steadily (Horticulture New Zealand, 2017). It was reported that there were approximately 5,000 growers in total in New Zealand and around 121,413 hectares la