This study concentrated on demersal fish, or bottom feeders, that live on or near the ocean floor where they feed on small vertebrates. Hake is an example of a demersal fish. The two main species of importance are shallow-water hake (M. capensis) and deep-water hake (M. paradoxus). Both species occur in South African and Namibian waters; however, the degree of stock separation is indistinct, as they partially overlap at certain depths and commercial landings have not been divided up in terms of species (Arancibia 2015). Figure 1 illustrates the location of M. capensis and M. paradoxus along the Namibian and South African coastlines.

For longline trawling, M. paradoxus is the preferred species as they are larger and mostly firmer than M. capensis because they occur at deeper depths where the water is cooler. Both species increase in size with increasing depth and offshore distance.

Directly after being brought on board the longline trawler, most of the hake is classified as PQs. The term ‘PQ’ is used in the fishing industry and refers to a fish in perfect condition, which can be exported as a value-added product as fresh, whole fish. For this study, ‘PQ’ refers to a fresh, whole hake that has never been frozen. If the PQs are kept at an optimal temperature range of –2 °C to 2 °C, they will retain their quality for between 9 and 15 days from catch to sale. If hake brought on board the trawler is slightly damaged, it is processed into skin-on fillets once the trawler reaches the onshore processing facility. Fresh skin-on fillets will retain edible quality for 14 to 20 days from catch to sale.

According to the head of quality control at Company A, the timeframe for the fresh hake supply chain from catch to sale is:

Fresh fish is characterised by a seaweed to neutral odour, a tight and shiny skin, clear, bulging (convex-shaped) eyes and bright-red gills. When one presses down on the flesh, it will not leave an indent.

Factors that might influence the quality, freshness and shelf life of the fresh product include poor landing techniques, poor handling, temperature breaches and breaks, bacterial and enzymatic spoilage, and delays in distribution and marketing.

Flake ice is preferred to keep the fresh fish at an optimal temperature on board the trawler. Salt is added to the ice, which decreases the freezing point to a temperature below 0 °C to –1.8 °C. The fish is layered between the flake ice in stacked plastic bins in the cargo hold. The fish is randomly packed into plastic bins in no particular manner. The researchers did trials on packing methods and found following a head to tail (dovetail or finger-laid) method when packing the bins showed excellent results for maintaining firmness. Fish that was packed head to tail was firmer on arrival at the on-land processing facility than those randomly packed into the bins. Figure 3 shows the H&G fish selected to be sold as fillets, packed according to the dove to tail (finger-laid) method in a plastic bin. Because of the fact that the factory space on board the trawler is restricted, the fish is usually only gutted on board the trawler and the beheading is done at Company A’s processing factory. The whole fish are also packed dove (head) to tail.

Figure 4 highlights the distinction in the quality of the fresh fish vis-à-vis normal packing compared to the dove to tail (finger-laid) method of packing in terms of the firmness of the fish. Figure 4 shows that the fish are firmer with the dove to tail packing method.

Once on shore, the fresh fish is immediately processed at the processing facility, either as PQs or skin-on fillets, depending on the firmness and quality of the fish. The temperature in the factory is kept constant at 18 °C. Processing takes 12 min to 15 min from receiving the fish to it being boxed and moved into the blast freezer (skin-on fillets) or chiller (PQs). The core temperature of the fresh fish is continually monitored at various stages of processing by probe thermometers. Processing temperatures never exceed 4 °C. Since the PQs maintain their core temperature better than the skin-on fillets, they are not placed in the blast freezer but moved directly to the chiller room, whereas the skin-on fillets are moved into the blast freezer for superchilling.

In the blast freezer, the skin-on fillets are quickly chilled to a temperature ranging from –2 °C to 0 °C. Superchilling prevents large ice crystals from forming in the flesh, which could damage the structure of the flesh when the fish is thawed. This could compromise quality and texture and make the product less attractive and palatable for the consumer (Kaale et al. 2011). After superchilling of the skin-on fillets, both the fillets and the PQs are kept in the chiller room at a temperature ranging between –2 °C and 0 °C until a refrigerated truck transports the fresh hake to the airport.

Packaging also plays an important part in maintaining optimal temperature along the cold chain. Expanded polystyrene boxes are extremely light and comprise 98% air by volume and therefore provide the perfect insulation by reducing the thermal effects of temperature breaches and breaks, especially during transport and transfers from refrigerated trucks to aircraft. To help maintain the temperature inside the EPS box, a single ice pack (552 g) is placed on top of the product before closing the box. A layer of plastic is placed between the fresh fish (for both PQs and skin-on fillets) and the ice pack to prevent freezing damage. After boxing, the product goes through a metal detector that will detect any foreign object, such as a hook left in a whole fish or an odd piece of metal. If metal is detected, the box is opened and searched. The foreign object is removed and the box is resealed. To ensure the safety of the product, it goes through the metal detector once again to make sure that all the metal has been removed.

Table 1 illustrates the processing of fresh hake at the onshore processing facility of Company A. Both categories, namely PQs and fresh skin-on fillets, are shown in the table. Table 1 follows the cold chain of the fresh fish from offloading from the vessel at Company A and moved into the chiller room where the temperature is maintained at –2 °C to 0 °C. From here, the product is prepared for transportation by refrigerated truck to Hosea Kotako International Airport in Windhoek and then to its final destination in Spain.

The processing stages of the PQs and the hake fillets are completed at Company A. The cold chain is meticulously maintained by Company A through continuous monitoring of the internal temperature of the fresh product with temperature probes that measure the core temperature of the product. Once the product is loaded into the refrigerated truck, the data logger amongst each shipment will commence monitoring the ambient temperature of the product (temperature of the air in contact with the product). The refrigerated truck leaves Company A for Hosea Kutako International Airport near Windhoek and from there by aircraft to Frankfurt, Germany, after which it is then transported by refrigerated truck to the retailer in Spain.

For this study, a single Time Temperature Indicator (TTI) was placed in a marked box before loading it into the refrigerated truck. Time temperature indicators are devices used for recording data and subject to requirements, there are numerous types of data loggers. Time temperature indicators are devices equipped with a time temperature-dependent change to reveal the temperature history and related food quality status (Lu et al. 2013). Xsense HiTag wireless data loggers are used by Company A to monitor the fresh, chilled hake export cold chain to Europe. These devices are critical to identify potential temperature deviations in the cold chain of fresh, perishable product exports. Data are gathered at all stages along the supply chain. The data loggers were monitored online on the Xsense portal, where real-time information is available on the status of a shipment. Company A was only willing to cover the costs for the use of one data logger per shipment. After the shipment was prepared for loading into a truck, a single TTI was initiated and placed the EPS box at Company A. The EPS box was clearly marked with a black X (to make it easily identifiable) and was then placed at the top left of the shipment, closest to the door of the truck. The TTI recorded the ambient temperature (not the core temperature) of the fresh fish every 30 min during the cold chain from leaving Company A to its final destination in Spain.

On arrival at Hosea Kutako International Airport in Windhoek, the shipment goes directly from the refrigerated truck through security scanning and is then placed in the on-site chiller facility. The shipment is then repalletised. Figure 5 illustrates the aforementioned process.

The fresh hake is stacked to a maximum of 10 layers. Every aircraft pallet or Unit Loading Device (ULD) is enclosed with a thermal blanket and positioned with stretchy rope to ensure that the fish remains safe during handling. All aircraft ULDs move on automated rollers. The pallets are placed for quick, safe and easy loading of the aircraft. Figure 6 shows how the fish is palletised and enclosed with the thermal blanket, fastened with the stretchy rope and lined for loading.

Coetzee and Dobson (2017) found that ambient temperature is normally 1 °C to 5 °C higher than the core temperature. They concluded in their trials on apples regarding core and ambient temperature that core temperature is influenced less by variations in ambient temperature and will cool down or heat up slower than ambient temperature.

During processing, the core temperature of the fresh fish is measured with probe thermometers, but for the transport journey during trucking and flight, the ambient temperature is measured by a single data logger per shipment.

Xsense HiTag wireless data loggers are utilised by Company A for the continuous monitoring of the fresh, chilled hake export cold chain to Europe. These loggers are real-time loggers but without a Global Positioning System (GPS). They provide transparency, traceability and visibility because temperature monitoring of the fresh product is a requirement for successful delivery to the export destination. The Xsense loggers also help to identify temperature deviations throughout the hake export cold chain.

A truck with authorisation from airport authorities transports the pallets of fresh hake from the chilling facility to the aeroplane for loading at the airport. Figure 7 shows the loading of the shipment from the authorised truck into the hold of the aircraft en route to Frankfurt International Airport. The EPS box market with the X is transferred as part of the shipment and the TTI inside the box continuously records the ambient temperature and will continue to do so, until it is removed by the customer at the end of the shipment’s journey.

As a result of the coronavirus disease 2019 (COVID-19) pandemic and restrictions, the researchers were not permitted to travel internationally and thus could only follow the shipment to this point. After this point, the researchers were solely reliant on the continuous data monitoring done by the TTI on the Xsense online portal.

During each leg of the cold chain journey, a cold chain logistics (CCL) score out of 10 is awarded to evaluate whether an optimal temperature ranging from –2 °C to 2 °C was maintained. A perfect score shows that the optimal temperature was never compromised and that temperature integrity was maintained. The data collected by the data loggers are presented in a graph divided into three legs, which represent the complete cold chain journey of the fresh product being exported, from its point of origin in Namibia to its point of destination in Spain.

After the fish has been processed, it is packed into EPS boxes and chilled at the processing facility of Company A, it is loaded into a refrigerated truck and the first leg of the cold chain journey begins. Recording of ambient temperature commences when the refrigerated truck, managed by a third party, leaves Company A when the logger is initialised. Each shipment’s cold chain journey is divided into three legs on the online Xsense portal. The three legs briefly summarised are:

Initial temperature readings recorded at the start of the cold chain journey by the data loggers were between 8 °C and 10 °C, as this was the temperature of the holding bay where the refrigerated trucks were loaded when the logger was initialised and began recording the ambient temperature of the product. This was before the data loggers were placed inside the EPS boxes that contained the product. One data logger was used per shipment, that is, one data logger for one shipment of PQs and one data logger for one shipment of skin-on fillets. It took the data loggers on average 40 min to reach 2 °C, which is the maximum temperature at which the product may be kept. Table 2 illustrates averages for minimum, maximum and overall temperature for the 60 PQ and 60 skin-on fillet shipments.

Maximum ambient temperatures for the PQs and the skin-on fillets were 6.25 °C and 7.5 °C, respectively.

The lowest minimum ambient temperature registered was –5.8 °C for the PQs, which was well outside the optimal temperature range of –2 °C and 2 °C. The problem with a slow decrease in temperature is the formation of large ice crystals that cause permanent damage to the flesh or muscle of the fish and, after thawing, would show discolouration and tissue softening. A ‘soft quality defect’ occurs, which will affect the quality and shelf life of the product. However, it will not render the product unsafe for human consumption (Tan et al. 2021).

Table 3 illustrates the average CCL score and the average overall CCL score for Legs 1, 2 and 3 of the export cold chain for PQs and skin-on fillets from when the refrigerated truck leaves Company A, up to receiving of the product by the customer in Spain.

Table 3 shows that the average CCL scores recorded for the skin-on fillets were lower than for the PQs. This indicates that it was more difficult to maintain optimal temperature for this product as the fish had been cut into pieces and did not have the bone intact. Investigation concluded that it is easier to maintain temperature when the fish is whole than when it is cut into fillets; however, further research is required to confirm this. The lowest CCL score recorded was for the skin-on fillets at Frankfurt Airport. The fresh hake arrived at the airport outside the optimal temperature range of –2 °C and 2 °C and registered a low CCL score as it took a while to chill the product in the on-site airport cold stores.

The reason for a low CCL score during Leg 2 was investigated. After observation of the cargo space in the passenger aircraft, the researchers discovered that the cargo hold of the aircraft is divided into two sections, namely a bulk cargo hold, which is set at 14 °C for less temperature-sensitive products, and a forward cargo hold, which is set at 5 °C for perishables such as chilled fresh fish. Company A has no control over where the product is placed because agents handle the cargo; therefore, temperature increases during flights could depend on where in the cargo hold the product was loaded. Other reasons for increases in the product temperature could also be that passenger cargo gets preference above optional freight; the product could have been left on the tarmac, either waiting for the aircraft to arrive or to be transferred back to the on-site airport cold stores, as there was no cargo space left on the aircraft.

The first leg of the skin-on fillets had a lower CCL score than the PQs. It was determined that the temperature of the refrigerated truck on the Namibian side was set too low at –8 °C, which caused the fresh product to chill too much, causing its partial freezing. The researchers recommend that the truck’s temperature be set to 0 °C in future, which is the same setting as the refrigerated trucks on the Spanish side.

The question also arose whether the single ice pack placed on top of the product in the EPS box was sufficient to maintain the optimal temperature range of –2 °C to 2 °C. Valtysdóttir (2010) suggested that two ice packs, one on the top and one at the bottom of the product, would be optimal. However, as the ice pack weighs 550 g, this would further increase the weight of the EPS box, resulting in higher airfreight costs, which would be less cost-effective according to Company A.

On a few occasions, the researchers observed partial freezing of the product around the outer extremities of the EPS box and where fish was in direct contact with the ice pack. The researchers had two explanations for this phenomenon. The partial freezing could be caused by the ice pack; however, this did not answer the question why the outer extremities were being affected. After investigation, the researchers determined that the refrigerated trucks were set at too low a temperature, which caused the partial freezing of the product along the box edges.

The researchers suggest that the single layer of plastic placed between the fresh fish and the ice pack could be replaced by a layer of bubble wrap, which would protect the product better and prevent partial freezing and improved insulation. Adding a layer of bubble wrap would not affect the weight of the EPS box.

The data loggers allowed the researchers to determine during which leg of the cold chain temperature breaches and breaks occurred, but it remained difficult to determine the exact location where the temperature had been compromised. A real-time data logger with GPS would be more suitable, as it would show the exact location in the cold chain where temperature problems occurred.

When temperatures rise above 5 °C for cumulative periods of less than 120 min, the product should immediately be refrigerated to 0 °C, as the fresh fish would still be safe to sell and consume.

After examining the 60 PQ and 60 skin-on fillet shipments, the researchers identified several temperature spikes, breaches and breaks during the export cold chain. Figure 8a and 8b shows the breakdown of temperature spikes, breaches and breaks per shipment for the respective shipments.

It should be noted that COVID-19 negatively affected transportation during the pandemic, as the availability of trucks were limited and flights were irregular and restricted. This is reflected in Figure 8a and 8b as temperature spikes, breaches and breaks. The researchers could not accompany the product from Windhoek to Frankfurt and to its final destination by refrigerated truck because of the travel ban. However, the researchers were able to follow each shipment’s journey through the continuous recording of the ambient temperature of the fresh product by the data logger placed among each shipment. The data recorded were available on the online Xsense portal. After identifying temperature spikes, breaches and breaks, the study identified during which leg of the export cold chain temperature breaches had occurred. As spikes did not play a significant role, they were left out of the discussion. Figure 9a and 9b illustrates the number of temperature breaches during each leg of the export cold chain for the 60 PQ and 60 skin-on fillet shipments.

The researchers observed that most temperature breaches occurred during Leg 1. Company A uses third-party trucking companies to obtain refrigerated trucks. After investigation, the researchers determined that the trucks had been set at –8 °C for the inland transport leg in Namibia. This caused partial freezing of the fresh product and meant that the temperature had fallen outside the optimal range of –2 °C to 2 °C. Leg 2 experienced temperature breaches because the product often arrived at Frankfurt Airport at temperatures outside the optimal range of –2 °C to 2 °C after being in the aircraft’s cargo hold where it was kept at either 5 °C or 14 °C, depending on where it was placed in the hold of the aircraft. Maintaining the product’s temperature during flights depended on the single ice pack placed on top of the product in the EPS box. A thermal blanket covering each ULD provided further insulation. The least number of temperature breaches occurred during Leg 3, as the product was loaded into a refrigerated truck set at 0 °C and transported to its retail destination.

After the researchers identified the temperature breaches, Figure 10a and 10b was compiled to illustrate the duration of temperature breaches (T/Bs) for each leg of the export cold chain for PQs (60 shipments) and skin-on fillets (60 shipments). The respective 60 shipments are arranged from the oldest at the top to the most recent at the bottom, in descending order.

After investigation, problems such as wrong temperature settings in the refrigerated trucks, loading delays onto the aircraft and flight delays were identified. Preventative and corrective measures were put in place by Company A, which resulted in shorter temperature breaches and breaks. Company A sent their recommendations through to the third parties handling transportation. However, as Company A is dealing with third parties, they do not have control over both the trucking and airport activities and can only trust that their suggestions will be followed through.

From Figure 10a and 10b, the researchers concluded that the duration of temperature breaches had improved over time for Legs 1 and 2; however, for some shipments, when transportation began for Leg 3 in the refrigerated truck on the Spanish side, the temperature decreased too much and fell outside the optimal range of –2 °C to 2 °C. This would need further investigation as the researchers could not be there in person because of the COVID-19 travel ban.

After identifying the duration of the temperature breaches for each shipment, the period (Leg 1 + Leg 2 + Leg 3) for each shipment was totalled and assessed as a percentage of the duration of the whole export cold chain from origin to destination (in minutes) for all 60 PQ and 60 skin-on fillet shipments. For the 60 PQ shipments, the researchers concluded that 9% of the total duration of the export cold chain had experienced ambient temperature breaches. For the 60 skin-on fillet shipments, 13% of the total duration of the export cold chain had experienced ambient temperature breaches. There were more ambient temperature breaches during the skin-on fillet shipments, which confirms the fact mentioned earlier that a whole fish with the bone intact maintains temperature more efficiently than cut skin-on fillets.

After the temperature breaches had been identified and graphically illustrated, the maximum ambient temperature reached for each shipment during the export cold chain was identified. The maximum temperature reached for each of the 60 PQ and 60 skin-on fillet shipments is illustrated in Figure 11a and 11b.

Shipments with a maximum ambient temperature of 5 °C for a duration of longer than 120 min were identified. As mentioned previously, a temperature break in this research is defined as the cumulative period in minutes when the ambient temperature rises above 5 °C for longer than 120 min. Three temperature breaks (2h 51m, 3h 50m and 2h 09m) with maximum ambient temperatures of 2.75 °C, 6.25 °C and 5.75 °C, respectively, were identified for the 60 PQ shipments. However, keeping Coetzee and Dobson’s (2017) research on ambient and core temperature in mind, which states that there could be a difference of 1 °C to 5 °C between ambient and core temperature, these shipments were not rejected but immediately cooled to 0 °C. One temperature break (10h 16m) with a maximum ambient temperature of 7.5 °C for the 60 skin-on fillet shipments was identified. This temperature break was longer than the 2h to 4h rule and the product was rejected. The long period during which a higher-than-recommended ambient temperature occurred and not the core temperature itself played a major role in this decision. However, no revenue losses were incurred as the exporter had insurance. Figure 12 is a combined summary illustrating the temperature breaks that occurred for both the 60 PQ and the 60 skin-on fillet shipments.