Capture-Based Aquaculture of Atlantic Cod (Gadus Morhua L.) in Greenland – Sustainable Distribution of Superchilled, Frozen and Refreshed Products

Downloaded from orbit.dtu.dk on: Oct 07, 2021

Capture-based aquaculture of Atlantic cod (Gadus morhua L.) in Greenland – Sustainable distribution of superchilled, frozen and refreshed products

Sørensen, Jonas Steenholdt

Publication date: 2020

Document Version Publisher's PDF, also known as Version of record

Link back to DTU Orbit

Citation (APA): Sørensen, J. S. (2020). Capture-based aquaculture of Atlantic cod (Gadus morhua L.) in Greenland – Sustainable distribution of superchilled, frozen and refreshed products. Technical University of Denmark.

General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

 Users may download and print one copy of any publication from the public portal for the purpose of private study or research.  You may not further distribute the material or use it for any profit-making activity or commercial gain  You may freely distribute the URL identifying the publication in the public portal

If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Jonas Steenholdt Sørensen Industrial PhD thesis 2020

Capture-based aquaculture of Atlantic cod (Gadus morhua L.) in Greenland

Sustainable distribution of superchilled, frozen and refreshed products

Capture-based aquaculture of Atlantic cod (Gadus morhua L.) in Greenland – Sustainable distribution of superchilled, frozen and refreshed products

Jonas Steenholdt Sørensen, M.Sc.

Industrial PhD Thesis

National Food Institute

Technical University of Denmark

Royal Greenland Seafood A/S

Submitted: 31 March 2020

Datasheet

Title: Capture-based aquaculture of Atlantic cod (Gadus morhua L.) in Greenland – Sustainable distribution of superchilled, frozen and refreshed products

Author: Jonas Steenholdt Sørensen, M.Sc.

Affiliation: Research Group for Food Microbiology and Hygiene National Food Institute (DTU Food) Technical University of Denmark Kemitorvet 202, 2800 Kongens Lyngby, Denmark

Contact: [email protected] https://orcid.org/0000-0002-9578-4930

Assessment Lisbeth Truelstrup Hansen, Professor, DTU Food, Kgs. Lyngby, Denmark committee: Turid Mørkøre, Professor, Norwegian University of Life Sciences, Ås, Norway Morten Sivertsvik, Research Director, Nofima, Stavanger, Norway

Supervisors: Paw Dalgaard, Professor, DTU Food, Kgs. Lyngby, Denmark Flemming Jessen, Senior Researcher, DTU Food, Kgs. Lyngby, Denmark Niels Bøknæs, PhD, Royal Greenland Seafood A/S, Svenstrup, Denmark

Funding: Innovation Fund Denmark Grant no. 5189-00175B

Photos: Front cover: Jonas Steenholdt Sørensen Back cover: Royal Greenland A/S Page i: Royal Greenland A/S

iii

“You did not kill the fish only to keep alive and to sell for food, he thought. You killed him for pride and because you are a fisherman. You loved him when he was alive and you loved him after. If you love him, it is not a sin to kill him. Or is it more?”

Ernest Hemingway, The Old Man and the Sea

Summary

Atlantic cod (Gadus morhua L.) has for centuries been an important species for the Greenlandic fisheries, and traditionally the value-added process has contributed to a large numbers of local jobs. In the period following the last major cod fisheries around 1990 in Greenlandic waters, the value-added processing was reduced and almost exclusively focused on headed and gutted whole cod, which were frozen in blocks and shipped to China for further processing. This was a consequence of cheaper labour in China and low sales prices for cod products in the primary markets in Europe. To optimise product quality and restore local workplaces, Royal Greenland has chosen to rethink the entire catch and value- added processing for cod in Greenland. This has been done with some inspiration from the salmon farming industry in Norway. The new process is referred to as Nutaaq® ("the new" in Greenlandic) and can best be described as capture-based aquaculture, where the cod are caught in the Greenlandic fjords, transferred to net cages where they are stored for two to four weeks without feeding. After this starvation period, the cod are transported by well boats to a processing plant, where the cod are slaughtered, filleted and either frozen or distributed fresh.

The purpose of this PhD project has been to investigate whether the new production method for cod in Greenland can contribute to increased food quality, longer shelf- life and reduced food waste and losses?

The significance of the changed production method was investigated in a study where the difference between traditionally caught and processed cod was compared with capture-based aquaculture produced cod during a frozen storage at -20 ° C. Further, the effect of frozen storage at -20, -40, and -80 °C was investigated for cod produced using capture-based aquaculture. After a freezing period of three, six, nine and twelve months, respectively, cod fillets from each of the two production methods were examined using hyperspectral images to determine colour and blood concentration, texture measurement, water holding capacity, salt soluble protein, and sensory profiling. Two other studies focused on shelf-life of fresh and thawed cod, respectively. The fresh cod was stored under four different storage conditions: (i) iced and packed in atmospheric air, (ii) superchilled and packed in atmospheric air, (ii) iced in modified atmospheric packing (MAP with 40 % CO2 and 60 % N2), and (iv) superchilled in MAP. The thawed cod was stored in ice (+0.4 °C), at +1.4 °C and +3 °C in atmospheric air as well as iced and at +3 °C in MAP. The shelf-life was determined by sensory evaluation and compared with the chemical and microbiological changes during storage.

After frozen storage at -20 ° C for three months, the cod caught and processed using capture- based aquaculture had firmer texture compared to conventionally captured and processed cod. In addition to a better texture, the bleeding was more efficient as it was ensured that all the cod were decapitated in a rested state. For CBA slaughter and processing could be better controlled as live cod was available and pumped into the plant according to the available working capacity. In addition, the factory used for CBA had the possibility to better rinse the cod after decapitation before the bleeding process due to better spacing. As a result of this, the total number of bacteria, and in particular the number of H2S-producing bacteria, was significantly reduced compared to cod bled with gill cuts.

Shipping of fresh, capture-based aquaculture cod from Greenland for the European market requires a long shelf-life, as transport is primarily by container ship, which often has a transit time of 8- 12 days. Storing the iced cod in air gave a sensory shelf-life of 15 days and, when replacing the air with MAP, the shelf-life was extended to 22 days. Superchilling of -1.7 °C further increased the shelf-life, and after 32 days of storage, there was no evidence of sensory spoilage of the fish stored in air or MAP. The combination of superchilling and MAP resulted in a low bacteria concentration of 3.9 log CFU/g. Common to all treatments in the study was that time for sensory spoilage correlated with the time to reach a pH value above 7.0, a TVBN concentration of more than 35 mg-N/100g of fish and a TMA concentration of more than 20 mg-N/100 g fish. For superchilled cod fillets stored in air, a bacteria concentration above 7.0 log CFU/g was observed towards the end of the storage trial, but without the cod fillets being sensory spoiled. In order to elucidate how there could be a high concentration of bacteria without the cod being spoiled, the correlation between the growth of Photobacterium spp., Shewanella spp., Pseudomonas spp. and TVBN formation was studied as an indicator of the spoilage activity. This showed that the spoilage activity was significantly higher in Photobacterium spp. than for the other two genera and that high concentrations of Pseudomonas spp. of 7.0 log CFU/g producing very little TVBN. The formation of TVBN in the fresh cod fillets was solely formed by Photobacterium spp. and P. carnosum was identified as the specific spoilage organism for chilled cod from Greenland stored both aerobically and in MAP.

The shelf-life of the frozen cod from capture-based aquaculture depended on the temperature. By lowering the temperature from -20 °C to -40 °C, the high-quality life was extended from 4-6 months to more than 12 months. The shelf-life was determined as the time until the water holding capacity was lower than 65 % and the percentage of the salt-soluble proteins at the same time was less than 70 %. The maximum shelf-life was extended from 7-10 months at -20 °C to a minimum of 12 months by

vi lowering the temperature to -40 °C. The maximum shelf-life was indicated when the levels were <60% for both water holding capacity and salt soluble proteins. A significant sensory difference was observed between capture-based aquaculture produced cod and the conventionally produced cod when storing the cod for 12 months at -20 °C. The difference was particularly evident when assessing the texture as well as the metal taste for the cooked cod, and at the same time, a softer texture was observed in the raw fillet.

The shelf-life of thawed fillets was investigated for cod from capture-based aquaculture production and previously stored at -20 °C for five months. Keeping the cod in in ice and storing in air resulted in sensory shelf-life of 19 days and compared to the fresh cod, the pH level and bacterial count were higher at the time of sensory spoilage. When replacing air with MAP, the shelf-life was significantly extended, and after 32 days of storage the cod was not spoiled at +0.4 or +2.9 ° C. In contrast to cod from several other countries the thawed MAP cod from capture-based aquaculture in Greenland had a low drip loss of <3.6 %.

The research-based results from this PhD project have led Royal Greenland, at the processing plant in Cuxhaven, Germany, to market a new product, "Chilled selection", based on frozen and thawed, capture-based aquaculture iced cod sold with a 10-day shelf-life. The commercial shelf-life of frozen cod has been further been reduced to better match the results obtained from this PhD project.

The implementation of the finding from this PhD project and the choice of optimal distribution and sales channels can help reduce food loss in the retail and consumer stages of the food supply chain. Simulations based on data collected from European retailers showed that by switching from fresh MAP cod to thawed and iced cod packed in air, the food loss would be reduced by up to 80 %.

Important future perspectives include studies to elucidated whether the fishery season has an impact on texture of cod fillets and whether the texture obtained is due to the CBA production form or genetic differences between cod from Greenland and other regions including Iceland and Norway. A better understanding of the texture of the raw material at different season may illuminate the impact of this texture on the frozen shelf-life of the cod. With the frozen raw material, there is a risk of contraction due to thawed rigor mortis; such an event could reduced the fillet sensory quality. Therefore, it would be interesting to understand the optimum thawing procedure, thereby ensuring the most optimal treatment for Royal Greenland's new cod product.

vii

Resumé (Danish summary)

I århundreder har torsk (Gadus morhua L.) været en vigtig art for det grønlandske fiskeri og traditionelt set har forædlingsprocessen bidraget til mange arbejdspladser. I perioden efter det sidste store grønlandske torskefiskeri omkring 1990, overgik forædlingen næsten udelukkende til indfrysning af hele, rensede torsk, der blev sendt til videre forarbejdning i Kina. Dette skete som en konsekvens af billigere arbejdskraft i Kina samt lave salgspriser for torskeprodukter på de primære markeder i Europa. For at optimere produktkvaliteten og genoprette de lokale arbejdspladser valgte Royal Greenland at gentænke hele fangst- og forædlingsprocessen for torsk i Grønland, med inspiration fra lakseopdrætsindustrien i Norge. Denne nye produktionsform omtalt som Nutaaq® (”den nye” på grønlandsk) kan bedst beskrives som fangstbaseret akvakultur, hvor torsk fanges i de grønlandske fjorde, overføres til netbure, hvor de opbevares to til fire uger uden fodring. Efter denne fasteperiode er torsken klar til at blive transporteret med brøndbåde til en fiskefabrik, hvor torsken slagtes, fileteres og enten indfryses eller distribueres fersk.

Formålet med dette ph.d.-projekt har været at undersøge, om den nye produktionsform for torsk i Grønland kan bidrage til øget spisekvalitet, længere holdbarhed og mindske madspild.

Betydningen af den ændrede produktionsform er undersøgt i et studie, hvor forskellen mellem traditionelt fanget og forarbejdet torsk blev sammenlignet med fangstbaseret akvakultur-produceret torsk under fryseopbevaring ved -20 °C. Yderligere blev effekten af fryseopbevaring ved -20, -40 og -80 °C undersøgt for torsk produceret ved hjælp af fangstbaseret akvakultur. Efter en fryseperiode på henholdsvis 3, 6, 9 og 12 måneder, blev torskefileter fra hver af de to produktionsformer undersøgt ved hjælp af hyperspektrale billeder til bestemmelse af farve- og blodkoncentration, teksturmåling, vandbindingsevne, proteinopløselighed i en saltholdig væske, samt sensorisk profilering. To andre studier fokuserede på holdbarheden af henholdsvis fersk og optøet torsk. De ferske torsk var opbevaret ved fire forskellige lagringsbetingelser: (i) iset og pakket i atmosfærisk luft, (ii) superkølet i atmosfærisk luft, (ii) iset i modificeret atmosfære pakning (MAP med 40 % CO2 og 60 % N2) samt (iv) superkølet i MAP. De optøede torsk var opbevaret iset (+ 0,4°C), ved +1,4 °C og ved +3 °C i atmosfærisk luft samt iset og ved +3 °C i MAP. Holdbarheden blev bestemt ved sensorisk bedømmelse og sammenholdt med de målte kemiske og mikrobiologiske forandringer under lagring.

Torsk fanget og behandlet ved brug af fangstbaseret akvakultur havde en mere fast tekstur end traditionelt fanget og forarbejdet torsk efter fryseopbevaring ved -20 °C i 3 måneder. Udover en bedre

viii tekstur, var udblødningen mere effektiv grundet sikring af, at alle torsk blev hovedkappet i en udhvilet tilstand. Processen for slagtning samt forarbejdning kunne styres bedre på fabrikken, da levende torsk var tilgængelige og kunne pumpes ind på fabrikken svarende til den tilgængelige arbejdskapacitet. Derudover har fabrikken den mulighed, på grund af bedre plads, at de efter hovedkapning af torskene kunne rense dem før udblødning. Herved blev det totale antal bakterier, og i særdeleshed antallet af

H2S-producerende bakterier, reduceret signifikant i forhold til torsk udblødt ved hjælp af kun to gællestik.

Forsendelse af fersk, fangstbaseret akvakultur torsk fra Grønland med henblik på det europæiske marked kræver en lang holdbarhed, da transporten primært foregår med containerskib, hvilket ofte har en varighed på 8-12 dage. Opbevaring af torskene iset i almindelig luft gav en sensorisk holdbarhed på 15 dage og ved udskiftning af luft med MAP blev holdbarheden forlænget til 22 dage. Superkøling ved -1.7 °C øgede holdbarheden yderligere og efter 32 dages lagring var der ikke tegn på sensorisk fordærv af torskene opbevaret i luft eller i MAP. Kombinationen af superkøling og MAP resulterede i et lavt total kimtal på 3.9 log CFU/g. Fælles for alle behandlingerne i forsøget var, at ved sensorisk fordærv blev der målt en pH-værdi på over 7.0, en TVN-koncentration på over 35 mg-N/100g fisk og ligeledes en TMA-koncentration på over 20 mg-N/100 g fisk. For torskefilet opbevaret superkølet i luft, blev der observeret et kimtal på over 7.0 log CFU/g mod slutningen af lagringsperioden, men uden at torskefileterne var sensorisk fordærvet. For at kunne belyse, hvordan der kunne være en høj koncentration af bakterier uden at torsken var fordærvet, blev sammenhængen mellem vækst af henholdsvis Photobacterium spp., Shewanella spp., Pseudomonas spp. og TVN-dannelse undersøgt som mål for fordærvelsesaktiviteten. Dette viste, at fordærvelsesaktiviteten var betydelig højere for Photobacterium spp. end for de to andre slægter, samt at høje koncentrationer af Pseudomonas spp. på 7.0 log CFU/g producerede meget lidt TVN. Udviklingen af TVN i de ferske torskefileter blev udelukkende dannet af Photobacterium spp. og P. carnosum, der blev identificeret som de specifikke fordærvelsesbakterier for kølet grønlandsk torsk lagret både aerobt og i MAP.

Holdbarheden for de frostopbevarede torsk fra fangstbaseret akvakultur afhang af temperaturen. Ved at sænke temperaturen fra -20 °C til -40 °C, blev højkvalitetsholdbarheden forlænget fra 4-6 måneder til mere end 12 måneder. Denne holdbarhed blev bestemt som den tid det tog, før vandbindingsevnen var lavere end 65 % og procentdelen af de saltopløselige proteiner samtidig var mindre end 70 %. Den maksimale holdbarhed blev forlænget fra 7-10 måneder ved -20 °C, til minimum 12 måneder ved at sænke temperaturen til -40 °C. Grænseværdierne for den maksimale holdbarhed var

ix sat til 60 % for både vandbindingsevne og saltopløselige proteiner. Ved frostlagring i 12 måneder ved -20 °C, blev der observeret en sensorisk signifikant forskel mellem fangstbaseret akvakulturelt producerede torsk og de traditionelt producerede torsk. Forskellen var særlig tydelig ved bedømmelse af konsistensen samt for bismagen af metal for tilberedt torsk, og samtidig blev der observeret en blødere tekstur i den rå filet.

I torsk fra fangstbaseret akvakultur produktion, der var opbevaret ved -20 °C i fem måneder, blev holdbarheden af de optøede fileter undersøgt. Isning og opbevaring i luft resulterede i en sensorisk holdbarhed på 19 dage og i forhold til de ferske torsk, var pH og kimtallet højere på tidspunktet for sensorisk fordærv. Ved udskiftning af luft med MAP til opbevaring af torskene, blev holdbarheden forlænget betydeligt og efter 32 dages lagring var de ikke sensorisk fordærvet ved 0,4 eller 2,9 °C. Modsat torsk fra flere andre lande, havde optøet MAP-torsk fra fangstbaseret akvakultur i Grønland et lavt dryptab på < 3.6 %.

De forskningsbaserede resultater fra dette ph.d.-projekt, har blandt andet medført at Royal Greenland på fabrikken i Cuxhaven, Tyskland, har valgt at markedsføre et nyt produkt, "Chilled selection", baseret på frosne og optøede, fangstbaserede akvakultur-torsk, der sælges med en holdbarhed på 10 dage, som isede fileter. Den kommercielle holdbarhed af frosne torsk er yderligere blevet nedsat til at passe bedre med de opnåede resultater fra dette ph.d.-projekt.

Implementeringen og valget af optimale distributions- og salgskanaler kan bidrage til at mindske madspildet i detailhandlen og hos forbrugeren. Ud fra simulationer baseret på data indsamlet fra europæiske detailhandlere vil madspildet, ved at skifte fra fersk MAP torsk til optøet og iset torsk pakket i luft, blive reduceret med op til 80 %.

Vigtige fremtidige spørgsmål inkluderer hvorvidt fangstsæsonen har en indvirkning på teksturen og hvorvidt den opnåede tekstur skyldes genetiske forskelle mellem torsk fra Grønland og Island/Norge eller produktionsformen. En bedre forståelse af teksturen af råvaren kan belyse indvirkningen af denne tekstur på den frosne holdbarhed af torsken. For den frosne råvare er der en risiko for at fileten trækker sig sammen som følge af et optønings rigor mortis-forløb. Derfor ville det være interessant at forstå den optimale optøning, og derved også sikre den mest optimale behandling for Royal Greenlands nye torskeprodukt.

Imaqarniliaq (Greenlandic summary)

Pissuseqatigiiaat saarulliit (Gadus morhua L.) ukiuni untritillini Kalaallit Nunaanni aalisarnermut pingaaruteqarluinnartuunikuugamik tunisassiarineri suliffippassuarnik pilersitsinikuupput. Kalaallit Nunaanni saarulleqarluarnerata kingulliup 1990-ikkunni pisup kingorna tunisassiorneq saarullinnik ilivitsunik erlaviikkanik qerisunik kingorna Kinami tunisassiareqqitassanngorlugit nassiunneqartartunik tunisassiornermut nuutinneqarpoq. Tamanna Kinami sulisunut aningaasartuutit annikinnerunerannik Europamilu niuerfinni pingaarnerni saarulliit akikinnerunerannik tunngaveqarpoq. Royal Greenlandip tunisassiap pitsaassusaa qaffanniarlugu najukkamilu suliffinnik pilersitseqqinniarluni Norgemi kapisilinnik tukertitsiveqarneq aallaavigalugu Kalaallit Nunaanni saarullinniarnerup tunisassiornerullu ingerlarnga eqqarsaatersuutigeqqissallugu aalajangerpoq. Tunisassioriaaseq Nutaaq®-tut taaneqartoq pisat naapertorlugit immami katersinertut nassuiarneqarluarsinnaaneruvoq, tassami saarulliit Kalaallit Nunaata kangerluini pisareriarlugit immami ungalunut nuullugit sapaatip akunnerinilu marlunni sisamaniluunniit nerlersornagit ungaluniitinneqartarput. Saarulliit nerlersorneqarnatik ungaluneereeraangamik umiatsiaq uumatitsivik atorlugu tunisassiorfimmut assartorneqartarput, toqorarneqariarlutillu nerpiliarineqanngikkunik, qeritinneqanngikkunik nutaajutillugit nassiussorneqartarput.

Kalaallit Nunaanni saarullinnik tunisassioriaatsip nutaap inuussutissatut pitsaassusaata qaffanneqarnissaanut, attartussusaata sivitsornissaanut nerisassiarereeraannilu igiinnarneqannginnissaata annikillisarnissaanut ilapittuutaasinnaanersoq misissornissaa Ph.d.-tut inaarutaasumik allaaserisap matuma siunertaraa

Tunisassioriaatsip allanngortinnerata sunniutaa ilisimatusarnikkut misissorneqarnikuuvoq, tassanilu pisarneq malillugu saarullinniariaaseq tunisassioriaaserlu pisat naapertorlugit immami katersat tunisassiarineqartut -20 °C-milu qeritinneqartut sanillersuunneqarput. Taamatuttaaq saarullinnik immami katersanik tunisassiornikkut -20, -40 aamma -80 °C-mut qerititsilluni uninngatitsineq misissorneqarpoq. Saarulliit nerpii tunisassioriaatsit marlut taaneqartut atorlugit tunisassiarineqarnikut qaammatini 3, 6, 9 aamma 12-ni qeriteriarlugit qalipaataat aagiaqassusaallu, sannaat imermik akoqassusaallu, imermi tarajulimmi proteiniisa arrorsinnaassusaat malugisaatillu atorlugit misissorlugit paasiniarlugillu hyperspektralimik, tassa inuup isaanik takuneqarsinnaanngitsut paasiniarniarlugit assilineqarput. Ilisimatusarnerni allani marlunni aalisakkat nutaat aatsitallu attartussusaat ukkatarineqarsimapput. Saarulliit nutaat uningatitseriaatsit sisamat atorlugit uninngatinneqarsimapput: (i) sikulerlugit silaannarmi nalinginnaasumi poortukkat, (ii) silaannalimmi qerinasuartitat, (ii) sikulerlugit

xi silaannarmi naleqqussakkami poortukkat (MAP 40 % CO2 og 60 % N2-jutillugu) kiisalu (iv) MAP-imi qerinasuartillugit. Saarulliit aatsitat sikulerlugit (+ 0,4°C-imi) uninngatinneqarput, +1.4 °C-imi aamma +3 °C-imi silaannarmiitinneqarput kiisalu MAP-imi sikulimmi +3 °C-imiitinneqarput. Attartussusaat malugisaatit atorlugit misissorneqarlunilu kemi atorlugu uninngatinnerinilu tappiorannartuisa allanngoriartornerinut naleqqersuunneqarpoq.

Saarulliit immami katersoriarlugit suliaralugillu pisat sannaat qaammatini pingasuni -20 °C-mi qeritinneqareerlutik saarullinnit nalinginnaasumik suliarineqarlutillu pisarineqartuniit suikkaanerupput. Pitsaanerusumik sannaqarnermik saniatigut saarulliit qasuersimanerat qulakkeeqqaarlugu niaquerneqartarnerat pissutigalugu aaviarluarneqarnerusinnaapput. Saarulliit uumasut imaaliallaannaq tunisassiorfimmi sulisoqarneq naapertorlugu annertussusilerlugit tunisassiorfimmut milluaatikkut ikaartinneqarsinnaammata toqoraaneq tunisassiornerlu aqukkuminarnerupput. Tamatuma saniatigut tunisassiorfik inissaqarnerummat saarullinnik niaquiaareernerup kingorna aaviaatinnani erlaviiaasinnaanissaminut periarfissaqarpoq. Taamaalilluni bakteeriaqassuseq tamakkiisoq, ingammik bakteerissat H2S-inik pilersitsisartut annertussusaat saarulliit masiitigut illuttut kapillugit niaquikkat bakteeriaqassusaannut sanilliullugu malunnartumik annikillivoq.

Saarullinnik Kalaallit Nunaata imaani katersanik nutaanik Europamut nassiussissagaanni attartussusaat sivisusariaqarpoq, nassiussinermi pingaarnertut containerinut ilioraalluni umiarsuakkut ulluni 8-niit 12-inut sivisutigisumik assartornerisigut ingerlanneqartarpoq. Saarulliit sikulerlugit silaannarmi nalinginnaasumiitillugit malugisaatit atorlugit misissorneratigut ullunik 15-inik attartussuseqartut paasineqarpoq, silaannarli MAP-imik taarserneqarmat attartussusaat ullunut 22-nut sivitsorpoq. -1.7 °C-mut qerinasuartitsineq attartussutsimut sivitsueqqippoq ullullu 32-it uninngatinneqareermata malugisaatit atorlugit misissorneratigut saarulliit silaannarmiisitat MAP- imiisitalluunniit asiujartorsimaneranik takussutissaqanngilaq. Qerinasuartitsinerup MAP-illu ataatsikkoortinneratigut tappiorannartut amerlassusaat 3.9 log CFU/g-jugami appasippoq. Malugisaatit atorlugit asiujartorneq pillugu misilittaanerni tamani seernassutsip pH-p kisitsisaa 7.0-mit qaffasinnerusoq, TVN-eqassuseq aalisakkat 100g-iugaangata/35 mg-N-imit qaffasinnerusoq aammalu TMA-qassuseq aalisakkat 100g-iugaangata/20mg-N-imit qaffasinnerusoq paasineqarpoq. Saarulliit nerpii silaannarmi nillaarissupilussuarmi uninngatinneqartut uninngatinneqarnerisa qaangiukkiartorneranni tappiorannartuisa amerlassusaat 7.0 log CFU/g-imit qaffasinneruvoq saarulliilli nerpii malugisaatit atorlugit misissoraanni asiujartorsimanerinik takussutissaqanngilaq. Bakteeriaqassuseq qaffasikkaluartoq saarulliup asiujartunnginnera pillugu asiujartornerup ingerlarnga paasiniarlugu

xii

Photobacterium spp.-it, Shewanella spp.-it, Pseudomonas spp.-it aamma TVN-init peqalersarnerisa ataqatigiinnerat misissorneqarpoq. Bakteerissanut allanut marlunnut sanilliullugu Photobacterium spp.- eqartillugu asiujartorneq sukkanerusartoq aamma Pseudomonas spp. 7.0 log CFU/g-iikkaangami TVN- imik annikittuinnarmik pilersitsisartoq misissuinerup takutippaa. Saarullinni nutaani TVN-ip ineriartornera Photobacterium spp.-imit aamma P. carnosum-imit pilersinneqartartoq, kalaallillu saarulliini aerobt MAP-ilu atorlugit uninngatinneqartuni bakteerissat asiutitsilersartut taakkorpiaasut uppernarsineqarpoq.

Saarulliit immami katersoriarlugit pisat qerisitat attartussusaat nillissutsimut attuumassuteqarpoq. Nillissuseq -20 °C -imiit -40 °C -imut appartinneratigut pitsaassutsip qaffasissup attartussusaa qaammatinit 4-6-iniit qaammatinut 12-init sivisunerusumut tallineqarpoq. Imeqassutsip 65% ataatsinnagu taamatuttaarlu proteinit tarajumit arrortinneqartartut 70% suli ataakkaangagit attartussuseq taamak sivisutigisartoq paasineqarpoq. Attartussuseq sivisunerpaaq -20 °C-mi qaammatini 7-10-nik sivisusseqartoq nillissutsip -40 °C-imut appartinneratigut minnerpaamik qaamatinut 12-inut sivitsorneqarpoq. Attartussutsip sivisunerpaaffissaa imeqassutsimut proteininullu taratsumik arrortinneqartartunut 60%-imut inissinneqarpoq. Qaammatini 12-ini -20 °C-imi qerisitsinikkut saarulliit immami katersoriarlugit tunisassiat saarulliillu nalinginnaasumik tunisassiat malugisaatit atorlugit misissorneratigut assigiinngissusaat ersarissoq maluginiarneqarpoq. Assigiinngissusaat ingammik issorsimassusaatigut saarulliillu nerisassiassatut uutat saviminersunninneratigut maluginiarneqarpoq, saarulliullu aapasup nerpiata aqinnerunera aammattaaq maluginiarneqarpoq.

Saarulliit immami katersoriarlugit pisat -20 °C-mi qaammatini tallimani uninngatinneqartut nerpii aatsitat attartussusaat misissorneqarpoq. Saarulliit sikulersukkat silaannarmilu uninngatitat malugisaatit atorlugit ullunik 19-inik attartussuseqartut paasineqarpoq, saarullinnullu nutaanut sanilliullugit malugisaatit atorlugit misissuinikkut seernartoqassusaat, tassa pH-qassusaat tappiorannartoqassusaallu asiujartornerata nalaani qaffasinneruvoq. Saarulliit uninngatinnerini silaannaap MAP-imik taarserneqarneratigut attartussusaat malunnartumik tallivoq ullullu 32-it 0.4 imaluunniit 2.9 °C-meereerlutik malugisaatit atorlugit misissorneqarmata asiujartornerannik malunnartoqanngilaq. Muminganik, saarulliit nunani allaneersut Kalaallit Nunaanni saarullinnut immami katersukkanut MAP-imilu aatsitanut sanilliullugit isseqassusaat < 3.6 %-imik annikinneruvoq.

Royal Greenland tunisassiorfimmi Cuxhavenimi Tysklandimiittumi tunisassiaq nutaaq ”Chilled selection”, tassa saarullinnik immami katersoriarlugit pisat nerpiannik qeriteriarlugit aatseqqeriarlugillu sikulersukkanik ullunillu qulinik attartussusilinnik tuniniaalerneranik Ph.D.-nngorniarluni suliami

xiii ilisimatusarnermik tunngaveqartumi matumani inernerit kinguneqarput. Ph.d.-nngorniarluni suliami matumani ilisimatusarnermit inernerit naapertorlugit saarulliit qerisut attartussusaasa inissinneqarnerat nioqqutigineqarnerinut naapertuunnerulerportaaq.

Atuutsitsilerneq siammarterisussanik tuniniaasussanillu eqqortunik toqqaanikkut niuertarfinni tuniniaasut pisisartullu igitsiinnartarnerannut annikillisaanermut ilapittuutaalluarsinnaavoq. Saarulliit nutaat MAP-imiittut saarullinnik aatsitanik sikulerlugillu silaannalikkanik taarsernerisigut igitsiinnartarneq 80% angullugu annikillisinneqarsinnaasoq Europami niutarfinni tuniniaanermit paasisat naapertorlugit pisuusaartitsinerup takutippaa.

Pisamaffik ilumut sannaanut sunniuteqartarnersoq saarulliillu Kalaallit Nunaaneersut Island/Norgemeersullu sannai sananeqaatikkut assigiinngissuteqarnersut, aammalu tunisassioriaatsit apeqqutaanersut pillugit apeqqutit pingaarutillit siunissami akineqartariaqarput. Tunisassiassap sannaa pillugu ilisimasaqarnerulernikkut saarulliup qerilluni attartussusaata sannaanut sunniuteqartarnera paasisaqarfigineqarsinnaavoq. Tunisassiassap qerisup aatsinneratigut nerpiata qerattarneranit nukiisa sukannerat nerpimut ajoqutaaratarsinnaavoq. Taamaammat qanoq eqqortumik aatsinnissaa, taamaalillunilu Royal Greenlandip saarullinnik tunisassiaata pitsaanerpaamik passunneqarnissaa paasisaqarfigissallugu soqutiginarpoq.

xiv

Preface

The PhD project was funded by the Innovation Fund Denmark in the research programme called Industrial Researcher – Industrial PhD Within this research programme the PhD student is employed by a company. The present PhD project “Capture-based aquaculture of Atlantic cod (Gadus morhua L.) in Greenland – Sustainable distribution of superchilled, frozen and refreshed products” was carried out in collaboration between Royal Greenland Seafood A/S, the Technical University of Denmark – National Food Institute (DTU Food) and Jonas Steenholdt Sørensen, M.Sc. in Parasitology.

Before the PhD project, Jonas Steenholdt Sørensen was employed by Royal Greenland, with the task of establishing knowledge of a capture-based aquaculture process in Maniitsoq, Greenland. This included seasonality of the raw material, rigor mortis profile related to ante mortem handling and good laboratory practice (GLP) in a remote areas, such as the processing plant. This employment was vital to gain a smooth transition into the PhD project and coordination of fieldwork.

During the first year of the project, seasonal changes of the cod were determined by the fillet properties (dry matter, protein, trimethylamine oxide (TMAO), drip loss, water holding capacity (WHC), cooking loss, fatty acids, salt, pH level and lactic acid content). This information was used together with Fulton’s condition factor (k), hepatosomatic index and gonadosomatic index (GSI) to characterise the overall condition of the cod. Results concerning seasonality are not included in the present PhD thesis. However, data on the total length of the fish and raw liver samples were included in an accepted research paper in the journal Parasitology Research (Severin, N.L., Yurchenko, M., Sørensen, J.S., Zuo, S., Karami, A.M., Kania, P. W., Buchmann, K., 2020. Anisakis nematode larvae (Anisakis simplex, Pseudoterranova decipiens and Contracaecum osculatum) in liver of Atlantic cod Gadus morhua L. from West Greenland.)

The cod protein quality measured by WHC, glucose and lactate in the blood was related to the handling of the cod, with focus on the resting period after the cod was pumped from a well-boat to a holding tank. These data formed the basis for a poster presentation at the 47th West Atlantic European Fishery Technologies Associations conference in 2017. After 2017, Royal Greenland changed the setup of the processing plant in Maniitsoq, Greenland and future work related to the resting of fish after pumping was regarded as less important. Based on the new design of the processing plant, three ways of distributing cod fillets were investigated (i) fresh and superchilled, (ii) frozen and (iii) frozen and thawed (refreshed). Capture-based aquaculture production and processing of cod as well as these three distribution channels form the basis of the present PhD thesis.

Background of the company

The Danish Crown established Royal Greenlandic Trade in 1774 and in 1985 the ownership was given to the Greenlandic state and Royal Greenland was formed as a stock-based corporation, with the Greenlandic state owning all the shares. By 2019, the company was a vertically integrated seafood company which owned fishing rights, vessels, processing plants and distribution and sales channels. The main species and products sold globally were northern shrimp (Pandalus borealis), Greenland halibut (Reinhardtius hippoglossoides), Atlantic cod (Gadus morhua L.), snow crab (Chionoecetes opilio) and lumpfish (Cyclopterus lumpus L.) roe.

In 2012, Royal Greenland announced a new business strategy, “The North Atlantic Champion”, where the mission of the strategy was “We sustainably maximise the value of the North Atlantic marine resources for the benefit of Greenland”. In 2012, Royal Greenland operated 20 land-based production facilities in Greenland and two sites in Denmark, two in Germany, one in Poland and one in Canada. In addition to the land-based facilities, Royal Greenland had three shrimp freezer trawlers and two fish freezer trawlers for Greenland halibut and Atlantic cod. Employing in total 2,057 people, with only 910 in Greenland corresponding to 44 % of the total employees (Royal Greenland A/S, 2013).

In 2018, the third version of the North Atlantic Champion strategy was initiated. Six years after the start of this strategy, the revenue of the core species from Greenland, including shrimp, Greenland halibut, Atlantic cod and snow crab, contributes with 75 % of the company revenue, and this represented an increase of 41 % since 2012. In the same period, the production facilities on land in Greenland have increased to 37 locations and outside of Greenland, the number of facilities has been reduced to one location in Europe and seven in Canada. The switch of focus also had an impact on the number of employees, with 1,487 (66 %) in Greenland out of 2,228 (Royal Greenland A/S, 2018).

The strategy for Atlantic cod was based on the fact that Royal Greenland in 2012 had a production of cod at 17 different land-based locations in Greenland. A large number of land-based locations represented a challenge for Royal Greenland in relation to profitable production (Fig. 1A). The strategy was to establish three epicentres for cod production and to focus the upgrading of facilities around these epicentres (Fig. 1B). To supply the three processing plants with a large quantity of cod, capture-based aquaculture (CBA) was chosen to expand the catching area for each plant and utilise the possibility of transporting live cod from distant fjords. Maniitsoq was chosen as a case processing plant and in 2014, the first fishing season for CBA was launched.

xvi

Before CBA, cod was processed as headed and gutted (H&G) post-rigor mortis and due to the low capacity of the workforce and freezers, the fish could often be up to six-day-old by the time of freezing. The freezing method was mainly vertical contract freezers and the combination of six-day-old fish and a rough physical freezing method resulted in the cod becoming very soft (Himelbloom et al., 1997). To obtain acceptable fillet yield, the H&G cod was shipped by sea to China and the semi-thawed cod was hand-filleted and frozen again for distribution to the European market. The texture quality determined the price level of the final product. CBA production aimed for the premium market, with wholesale prices for H&G of approximated 4 €/kg, frozen fillets of approximate 8 €/kg and refreshed fillets with a price of approximated 10 €/kg (Royal Greenland wholesale prices February 2020). Furthermore, the latest Royal Greenland strategy for Atlantic cod focused on increasing local employment in Greenland and it is the aim that results from the present PhD project indirectly contribute to this goal.

xvii

Acknowledgements

The present PhD project would not have been possible without support from the processing plant in Maniitsoq, Greenland. I thank Niels Bøknæs and Sune Mejer for giving me the opportunity to work with Royal Greenland during the PhD project.

Throughout the project, I had all the help I could ask for at the process plant in Maniitsoq, Greenland and I would like to thank all employees for helping me with small and more extensive problems alike. This made the fieldwork an enjoyable experience. A special thanks to the plant manager, Susanne Marie Knudsen and quality coordinator Heidi Haraldsen, for helping with the logistic of collecting and shipping raw material for the experiments and to new product developer Jan Zoutenbier, for being responsible for the sensory evaluation of the frozen cod.

I would like to thank my supervisors, Paw Dalgaard, Flemming Jessen, Niels Bøknæs and Ole Mejlholm, for the help of forming the project and all the supervision, discussions and guidance they have given into the seafood research area. I would to thank bachelor student Oliver Ørnfeld-Jensen, laboratory technicians, Govand Babaee, Mia Laursen, Rannvá Høgnadóttir Houmann, Margrethe Carlsen, Heidi Olander Petersen, Rie Sørensen and Hanne Lilian Stampe-Villadsen for hours of work and help with practical questions. Lastly, thanks to all my colleagues at DTU for making a good working environment.

Jonas Steenholdt Sørensen, March 2020

xviii

List of publications

The following publications were written as part of the PhD project and included in the thesis. For the sake of simplicity, these manuscripts are referred to as papers within the thesis. The printed paper is included in the pre-printed format due to copyright agreements.

Publications: Paper I Jonas Steenholdt Sørensen, Niels Bøknæs, Ole Mejlholm and Paw Dalgaard. Superchilling in combination with modified atmosphere packaging resulted in long shelf-life and limited microbial growth in Atlantic cod (Gadus morhua L.) from capture-based-aquaculture in Greenland. Food Microbiology, 88, 2020. https://doi.org/10.1016/j.fm.2019.103405

Paper II Jonas Steenholdt Sørensen, Niels Bøknæs, Ole Mejlholm, Karsten Heia, Paw Dalgaard and Flemming Jessen. Short-term capture-based aquaculture of Atlantic cod (Gadus morhua L.) generates good physicochemical properties and high sensory quality during frozen storage. Submitted to Innovative Food Science & Emerging Technologies.

Paper III Jonas Steenholdt Sørensen, Oliver Ørnfeld-Jensen, Niels Bøknæs, Ole Mejlholm, Flemming Jessen and Paw Dalgaard. Thawed and chilled Atlantic cod (Gadus morhua L.) from Greenland - Options for improved distribution. Submitted to LWT- Food Science and Technology.

xix

List of dissemination activities

The minimum resting period for Atlantic cod (Gadus morhua L.) to regain pre-stressor status after pumping in a capture-based aquaculture operation

Jonas Steenholdt Sørensen, Paw Dalgaard, Niels Bøknæs, Ole Mejlholm and Flemming Jessen

Poster presentation at the 47th West European Fish Technologists’ Association (WEFTA) conference

9 Oct – 12 Oct 2017

Dublin, Ireland

Fødevareinnovation

Jonas Steenholdt Sørensen, Nikoline Ziemer, Monica Mathiassen and Jan Petersen

Oral panel debate at NORA region conference 2018

2 Jun 2018

Nuuk, Greenland

Atlantic cod (Gadus morhua) from capture-based aquaculture has better colour and cooking properties than traditionally caught cod

Jonas Steenholdt Sørensen, Ole Mejlholm, Niels Bøknæs and Flemming Jessen

Oral presentation at the 70th Pacific Fisheries Technologists (PFT) Meeting

24 Feb – 27 Feb 2019

San Carlos, Mexico

Superchilling of Atlantic cod from Greenland extent shelf-life to more than 32 days and MAP (40% CO2

/60% N2) in combination with superchilling prevent microbial spoilage

Jonas Steenholdt Sørensen, Niels Bøknæs, Ole Mejlholm and Paw Dalgaard

Oral presentation at the 49th West European Fish Technologists’ Association (WEFTA) conference

14 Oct – 18 Oct 2019

Tórshavn, Faroe Islands

List of abbreviations

ASL Available shelf-life ATP Adenosine triphosphate CBA Capture-Based aquaculture CFU Colony-forming unit DMA Dimethylamine EF Efficient frontier FAO Food and Agriculture Organization FCC Fresh case cover FIFO First-in-first-out

FMSY Maximum sustainable yield

FPA Fishing pressure FSC Food supply chain GDP Gross domestic product GHG Greenhouse gas GLP Good Laboratory Practice GSI Gonadosomaic index H&G Headed and gutted HQL High quality life IA Iron agar ICES International Council for Exploration of the Sea IQF Individual quick frozen JND Just noticeable difference k Fulton’s condition factor LH Long and Hammer MA Marine agar MAP Modified atmosphere packaging MCQI Multi-compound quality index MSC Marine Stewardship Council NOK Norwegian kroner OSA On-shelf availability

xxi

PCA Plate count agar PSL Practical storage life QDA Quantify descriptive analyse QIM Quality index method Refreshed Frozen and thawed RRS Relative rate of spoilage SCQI Single-compound quality index SDG Sustainable development goal SSO Specific spoilage organism TAC Total allowable catch TMA Trimethylamine TMAO Trimethylamine oxide TVBN Total volatile basic nitrogen TVC Total viable count USD United States Dollar WHC Water holding capacity

xxii

Table of contents

Datasheet ...... iii Summary ...... v Resumé (Danish summary) ...... viii Imaqarniliaq (Greenlandic summary) ...... xi Preface ...... xv Background of the company ...... xvi Acknowledgements ...... xviii List of publications ...... xix List of dissemination activities ...... xx List of abbreviations ...... xxi Table of contents ...... xxiii Table of Figures ...... xxv 1. Introduction ...... 1 1.1 Cod fishery in Greenland and globally ...... 1 1.2 Aquaculture production of Atlantic cod ...... 4 1.3 Capture-based aquaculture (CBA) ...... 6 2. Aim of the study ...... 12 3. Shelf-life and indices of spoilage ...... 15 3.1 Freshness and quality ...... 15 3.2 Sensory quality ...... 15 3.3 Indices of spoilage ...... 17 3.3.1 Indices of spoilage for fresh and superchilled cod (Paper I) ...... 17 3.3.2 Indices of spoilage for frozen cod (Paper II) ...... 24 3.3.3 Indices of spoilage for refreshed cod (Paper III) ...... 29 3.4 Best practice for CBA cod production, processing and distribution ...... 33 4 Food waste and loss ...... 33 4.1 Food waste for cod fishing and processing ...... 34 5.2 Food losses for cod during distribution ...... 35 6 Sustainable Development Goals ...... 43 6.1 Positively impacted SDG ...... 43 6.2 Negatively impacted SDG ...... 45

xxiii

6. Conclusions ...... 46 7. Perspectives ...... 47 References ...... 49 Paper I ...... 70 Paper II ...... 104 Paper III ...... 135

xxiv

Table of Figures

Figure 1 Atlantic cod processing plants in Greenland. A) 17 active processing plants in 2012 B) Vision with 3 processing plants, with a large catchment area for each plant and based on capture-based aquaculture (source: internal document at Royal Greenland)...... xvii Figure 2 World capture fisheries and aquaculture production (FAO, 2018)...... 1 Figure 3 Annual global production of farmed Atlantic cod in the period 1989 to 2015 (FAO, 2020b)...... 5 Figure 4 Catchment of Atlantic cod in Greenland from 2013 to 2019; purple bars indicate the tonnes produce with the CBA concept (Statistics Greenland, 2019, personal communication with Royal Greenland)...... 9 Figure 5 Fishermen in the preparation of gathering the cod inside the Royal Greenland developed mobile net cages. Photo: Jonas Steenholdt Sørensen ...... 10 Figure 6 A) Poundnet fishery in the Greenlandic fjords, in the conventional method, cod would be gutted and rinsed in the small boat. B) In the CBA method, the cod is moved to small mobile net cages and starved for two to four weeks. C) The live starved fish are collected and transported by well-boat. D) At the process plant, the cod are released to a net enclosure and allowed 12 hours to rest. E) At the capacity of the process line, the cod is pumped into the plant and anesthetise by electricity, decapitated, rinsed and bled in circulating refrigerated water. The headed and gutted cod were machine filleted, hand-trimmed and individual quick frozen in a gyrofreezer...... 11 Figure 7 Graphical summary of papers (Next page) ...... 13 Figure 8 Quality factors contributing to the overall quality and emphasis of the studied indices of spoilage, yellow markers indicate quality factors studied in papers I, II and III. Modified from (Oehlenschläger and Sørensen, 1997)...... 15 Figure 9 Concentrations of microorganisms in iced cod in MAP, enumerated by ■ Long and Hammer and ♦ Iron agar, modified from Paper I...... 19 Figure 10 The spoilage activity of ● Photobacterium spp., ■ Shewanella spp. and ▲ Pseudomonas spp., icons symbolised measurement and lines are the calculated formation of TVBN. The blue line is the calculated TVBN formation of Shewanella spp. with the spoilage activity obtained from (Dalgaard, 1995)...... 20 Figure 11 Relation between storage temperature (°C) and the square root transformed maximum growth rates (1/days). Line shows regression of all data from literature (Table 2), ●) growth rate of product in air, ■) growth rate of product in MAP (Table 2), ▲) growth rate from paper I in air, ▼) growth rate from paper I in MAP...... 20 Figure 12 Atlantic cod fished and produced with the conventional method (Fig. 5A), stored in boxes and covered by iced, source: Royal Greenland...... 35 Figure 13 Food Supply chain (FSC) of Atlantic cod from Greenland to Europe of fresh cod, the shelf-life determined in Paper I starts at stage A and the temperature in A-C is equal to those in Paper I...... 37 Figure 14 Food Supply chain (FSC) of Atlantic cod from Greenland to Europe of frozen and thawed cod, the shelf- life determined in Paper III starts at stage D...... 37 Figure 15 Efficient frontiers for the food supply chains studied in Paper I and Paper III, ● fresh iced cod in MAP, ■ Superchilled fresh cod in air or MAP, ▼ refreshed cod in air at 1.4 °C, ♦ refreshed cod in air at 0.4 °C, ▲ refreshed cod in air at 2.9 °C and ● refreshed cod in MAP...... 41 Figure 16 Sales strategies, when using the efficient frontiers model to assess food supply chains. Waste in the graph should be characterised as loss with the definition by Grolleaud, (2002), source: Broekmeulen and van Donselaar, (2019)...... 41 Figure 17 The left graph shows the scenario for fresh, iced cod in MAP to determine the most cost-effective on shelf availability (OSA). The right graph shows the remaining shelf-life to the consumers with the same model parameters as the left graph...... 42

xxv

1. Introduction

1.1 Cod fishery in Greenland and globally

Figure 2 World capture fisheries and aquaculture production (FAO, 2018).

The human population has increased from 1.65 billion in 1900 to more than 7.7 billion in 2020 and with the expectation of reaching 9.7 billion in 2050 (United Nations, 2019). This raises the question of how to feed the growing population and still provide nutritional food (Costello et al., 2019). In 2015, 17 % of the total protein intake originated from marine sources, and the average consumption per capita grew from 9.0 kg of fish in 1961 to 20.2 kg in 2015 (FAO, 2018). Production from the ocean can be divided into capture and aquaculture production, where capture production reached 80-90 million tonnes of whole fish in the late 1980s and has since stabilised at this level (Fig. 2). In the same period, the aquaculture production has increased dramatically and accounted for 50 % of the total marine production by 2016 (FAO, 2018).

Globally, the captured seafood production is concentrated on 25 species and genera, with a combined volume of >40 % of the total global production. Atlantic cod, as studied in Papers I, II and III, is a lean white fish, together with for example Alaska pollock (Gadus chalcogrammus), striped catfish (Pangasius hypophthalmus) and Nile tilapia (Oreochromis niloticus L.). Globally, Alaska pollock and Atlantic cod are the largest and 9th largest fishery (FAO, 2018), respectively, and lean white fish overall is the largest traded seafood category, with a trade value of USD 12 billion (Nikolik and Heinhuis, 2015). In the North Atlantic, the Atlantic cod represent a large proportion of the total wild catch fishery (Table 1).

Table 1 Wild catch fishery of selected species in the North Atlantic Ocean (Arctic Sea, Atlantic, Northeast and Atlantic, Northwest). Data modified from FishstatJ (FAO, 2020a).

Average production (tonnes) 2010 - 2017 Scientific name FAO name North Atlantica Greenlandb Gadus morhua L. Atlantic cod 1,223,599 33,953 Micromesistius poutassou Blue whiting 870,980 7,462 Melanogrammus aeglefinus L. Haddock 355,115 1,651 Pollachius virens L. Saithe 325,166 617 Pandalus borealis Northern prawn 281,884 97,382 Reinhardtius hippoglossoides Greenland halibut 108,791 36,319 Chionnoecetes opilio Queen crabc 99,172 1,976 Merluccius merluccius L. European hake 87,833 -d Trisopterus esmarkii Norway pout 59,120 -d Molva molva L. Ling 42,579 23 Brosme Brosme Tusk 22,798 252 Cyclopterus lumpus L. Lumpfish 15,255 8,964 Merluccius bilinearis Silver hake 15,083 -d Trisopterus luscus L. Pouting 10,305 -d Pollachius pollachius L. Pollock 8,973 1 a Including Greenland. b Greenland landings include swap quota with fishery in the Barents Sea. c Queen crab is also known as snow crab. d No catch registration.

The global fishery is a highly regulated sector with international agreements on fishing quotas

(TAC) based on stock assessments. In 2015, for 90 % of the global fishery, the fishing pressure (FPA) on the stocks were either at the maximum sustainable yield (FMSY) or above (FAO, 2018). The largest region for the Atlantic cod fishery is the Barents Sea and during the period 2012-2017, the fishery was sustainable, but in 2018 the FPA was greater than the FMSY by 9 % (ICES, 2019a). In West Greenland, the Atlantic cod was overfished by on average 116 % during 2012-2017 (ICES, 2019b).

Cod has been one of the most important commercial fisheries and has been the target for European fishing since man began fishing in the seas around Europe. The cod was the driver for a fleet of European fishermen travelling west already back in 985 in the pursuit for the large fishing banks near the coast of Newfoundland as described in the book by Mark Kurlansky “Cod: A Biography of the Fish That Changed the World” (Kurlansky, 1999).

“How did the Vikings survive in greenless Greenland and earthless Scotland? How did they have enough provisions to push on to Woodland and Vineland, where they dared not go inland to gather food, and yet they still had enough food to get back? What did these Norsemen eat on the five expeditions to America between 985 and 1011 that have been recorded in Icelandic sagas? They were able to travel to all these distant, barren shores because they had learned to preserve codfish by hanging it in the frosty winter air until it lost four-fifths of its weight and became a durable woodlike plank. They could break off pieces and chew them, eating it like hardtack”

(Kurlansky, 1999)

Fish processing has changed due to technical advances. Frozen seafood utilisation has increased dramatically since the 1960s and was in 2018 the second-most used storage method, after fresh or live fish. Cod has traditionally been utilised as either fresh or cured in the Northern European region and salted in the Southern European region (Oliveira et al., 2012). Freezing of cod has been utilised both for untreated cod and for salted or cured cod (FAO, 2018).

The global production of cod from 1950 to 2018 has fluctuated between 1 million to 4 million tonnes, with a peak in 1968 and a low during the 2000s. Between 2011-2016, the annual catch was above 1 million tonnes and with little fluctuation (FAO, 2020b). The cod fishery in Greenland has traditionally been divided between East and West Greenland, with the West Greenlandic fishery as the largest. From 1948 to 1973, the West Greenlandic fishery had an annual catch of more than 100,000 tonnes, with a maximum catch of >450,000 tonnes. After 1973, the fishery collapsed and the annual catch was below 100,000 tonnes with the exception of 1989 (Buch et al., 1994). In 1998, only 356 tonnes were landed in Greenland and mark the lowest point of the cod fishery in Greenland. Since 1998 the fishery has grown to above 10,000 tonnes/year since 2011 (Statistics Greenland, 2019).

The Atlantic cod in the North Atlantic Ocean all belong to the same species, Gadus morhua L., with subspecies or spawning population having possibly developed due to environmental pressure. One example of the environment driving such division is the Baltic Sea. These cod migrate from the North Sea and due to the lower salinity in the Baltic Sea, only cod that spawned egg with lower density survived in the Baltic Sea (Nissling and Westin, 1997). The cod in the Greenlandic waters are now recognised as belonging to four spawning population:

“Recruitment to the fisheries can involve four different stock components, with different spawning, larval drift, and migration patterns: (i) an offshore component spawning over the outer slope of various fishing banks off West Greenland; (ii) an offshore component from spawning areas located off Southeast and East Greenland; (iii) an Icelandic component, of which considerable numbers of larvae and pelagic 0-group stages are sometimes transported to East and West Greenland; (iv) a number of distinct local inshore populations, which spawn in separate fjord systems.” (Storr-Paulsen et al., 2004)

Based on knowledge of the subpopulations of Atlantic cod around Greenland, the Greenland Institute of Natural Resources in cooperation with International Council for Exploration of the Sea (ICES) advised for the first time for the 2013 TAC with two subpopulations, (i) an inshore population in West Greenland and (ii) an offshore population for West and East Greenland (Greenland Institute of Natural Resources, 2012). For the 2016 TAC, the population (i) was further split in separate West and East stock assessments (Greenland Institute of Natural Resources, 2015). A study to correlate the appearance of the cod and their DNA markers was performed together with experienced local fisherman. The conclusion of that study was that appearance, size and colour could not be used as a separator for these stocks (Hedeholm et al., 2016).

1.2 Aquaculture production of Atlantic cod The first FAO-registered aquaculture production of Atlantic cod was in 1987 with a production of 205 tonnes globally. Up until 2001, the annual production was below 1,000 tonnes. From 2006 to 2012, the production was higher than 10,000 tonnes and peaked at 22,728 tonnes in 2009 (Fig. 3) (FAO, 2020b).

)

n n

o 20

1 (

d o

r 10

e 5

a F 0 1990 1995 2000 2005 2010 2015 Year

Figure 3 Annual global production of farmed Atlantic cod in the period 1989 to 2015 (FAO, 2020b).

The first successful production of juvenile cod in 1985 also started extensive research on various vital parameters for a commercial aquaculture setup for Atlantic cod (Øiestad et al., 1985). Topics include hygienic egg and larvae production of cod (Hansen and Olafsen, 1989; McIntosh et al., 2008; Rosenlund and Halldórsson, 2007), small scale studies to identify if the cod was a suitable species for aquaculture (Audet et al., 1993; Quéméner et al., 2002), disease monitoring, with observation for Vibrio anguillarum infection at the cod fry stage (Buchmann et al., 1993), high level of mortality in 1-15 cm long-farmed cod due to Trichodina murmanica (Khan, 2004), identification of a granulomatous disease in mature cod caused by Francisella (Olsen et al., 2006), feed conversion rate (Lambert et al., 1994; Morais et al., 2001; Pérez-Casanova et al., 2009), freeze resistances of year 0 class fish (Fletcher et al., 1997; Gotceitas et al., 1999) feed formula (Gildberg et al., 1997; Olsen et al., 2007), the effect of light and temperature on grow rates (Hemre et al., 2002; Kolbeinshavn et al., 2012; Van Der Meeren and Jørstad, 2001), genetic composition of the brood stock (Dahle et al., 2006; Jørstad et al., 2006; Moen et al., 2008) and design of cages and net materials (Moe et al., 2009; Rillahan et al., 2011).

There was an increased production of farmed Atlantic cod up to 2009 (Fig. 3) and the export prices for whole gutted fish in Norway during the years 2002-2008 were in the range of 33-38 NOK/kg (3.23-3.72 €/kg with conversion rate from the 26 Feb 2020). In 2008, the global finance crisis broke out and in combination with increased costs of farming due to disease outbreaks and a significant reduction of export prices, partially due to increased stocks in the Barents Sea (ICES, 2019a) and around Iceland (ICES, 2019c), the price dropped to 25 NOK/kg (2.45 €/kg) in the years of 2009-2015 and lead to bankruptcy of important parts of the Norwegian aquaculture cod producers (Henriksen et al., 2018).

1.3 Capture-based aquaculture (CBA) Keeping fish, including cod, alive from catch to consumption has been practised for a long time. On an industrial scale, Norway has records of such production dating back to 1880 (Midling et al., 1996). In recent years one of the strategies to increase sales of cod has been to introduce CBA and by using feeding to increase the value of a quota by 100 % (Midling et al., 1996). The first attempt was based on the capture of juvenile cod, with a catch of 600,000 individuals in 1988. Only a small fraction survived to slaughter size, due to Vibrio salmonicida outbreak. Up to the outbreak, the fingerlings, i.e. the stage where juveniles have completed the transition from larval to fish, showed promising results, with low catch mortality, less than 1 %, and easy weaning to wet pellets (Jørgensen et al., 1989).

In Norway, the skrei cod has been the base for CBA in the 2000s and 2010s. Skrei cod is adult fish migrating from the Barents Sea to the coastal areas of Northern Norway for spawning around February to April. The cod was captured when they migrated during the winter months. The strategy for the Norwegian fishery was to keep the cod alive and in this way extending the season for fresh cod over the summer months (Hermansen and Eide, 2013). Outside Norway, Greenland is the only other country with an active CBA production of cod. In Greenland, the fishery was based on cod migrating to the coastal area of West Greenland for predation on capelin (Mallotus vilosus) and the fishing season was mainly during May-July and September-December (Storr-Paulsen et al., 2004, personal experience). The summer cod with high predation activity has shown lower WHC compared to cod from the colder months (Olsson et al., 2007), and the strategy in Greenland was to increase the WHC and texture quality by introducing a step where live cod was kept in net enclosures without feeding (Olsson et al., 2006).

Capture of wild cod by different types of gear, was observed to cause different mortality and gear damage on the individual cod. Fishing with trawl resulted in a 2.5 % mortality and in 16.5 to 28.3 %

6 had injuries or bruises (Digre et al., 2010). Compared to fishing with longline, the trawl fishing increased the occurrence of bruises and the fillet was poorly bled and had a significant lower water holding capacity (WHC) (Rotabakk et al., 2011). Damage could be reduced if the skipper of the fishing vessel has to change behaviour to decrease stress-induced mortality by slower rearing and smaller catch sizes (Anders et al., 2019; Digre et al., 2010). A modification of trawling procedure was studied, the idea was to introduce a “buffer towing” method. The method involves a step at 40 % of the maximum fishing- depth, buffer towing harmed the exsanguination and increase level of gear damage to the cod (Brinkhof et al., 2018). The most successful gear selection for fishing with the aim for CBA of cod, had been seine net (Danish or Scotttish seine) and since the 1990s several improvement to the construction and fishing procedure had resulted in mortality between 0 – 3 %, comparing to longline fishery with a mortality of 40 % (Dreyer and Nøstvold, 2008).

The CBA of cod in Greenland was based on poundnets (Fig. 6A), which is a passive fishing method (Paper II). There have not been any studies on the mortality rate during the capture of cod in Greenland, but according to personal observation and communication with local fishermen, the mortality rate due to capture has been low.

When captured, the cod need to recover from stress including gear and physical damage. By using seiners and trawl, the cod was brought up from depths of 130-200 meters to the surface in a relatively short time frame. This corresponded to a pressure loss from 131-201 bars to 1 bar and could lead to swim bladder rapture (Midling et al., 2012). The survival rate has been reported to vary between hauls from 49–93 % and was hypothesised to be related to the fishing depth (Digre et al., 2017). The damage to the cod would not be fatal if gas trapped in the swim bladder were to escape by rupture of the bladder. Midling et al., (2012) showed that the swim bladder would recover shortly and secure the welfare of the fish. If in contrast the swim bladder stays intact, the trapped gas will prevent the cod from resubmerging. One study showed that this occurs for approximately 40 % of a catch (Humborstad et al.,

2016). If not treated, as explained below, it would result in high mortality, 79 % CI95% 62–89 % (Humborstad et al., 2016). Depending on the fishing gear, the mortality of the portion of the cod that was able to submerge was 9–39 % and lowest for collapsible pots (fishing gear descripted in (Furevik and Løkkeborg, 1994)) and highest for longlined fish (Humborstad et al., 2016). Treatment would include a manual release of gas from the swim bladder by inducing a needle to puncture the bladder. The mortality due to this treatment has been shown to be low with the level of 6 % (Humborstad and Mangor-Jensen, 2013).

Fishing of Atlantic cod for CBA in Greenland is different from fishing in Norway. By mainly fishing close to the shore and in upper layers of the water (Fig. 6A), a lower catch-related mortality is obtained. However, some mortality still occurs due to catch handling. The behaviour of the capture cod, when released into net enclosures and well-boat (Fig. 6B, C & D), show that 50 % of the live fish search for the bottom of the cage. The behaviour is a response to the handling stress and by resting on the bottom, the cod would recover from the stress (Dreyer and Nøstvold, 2008). Further improvements of net design, i.e. increasing the bottom to volume ratio, and research studies for optimising the handling from the poundnet to the net enclosure would be necessary to increase the welfare and reduce mortality.

Stress and trawl handling induced damage to the fillet included increased blood content in the fillets. A recovery phase of >28 days was needed and a blood reduction of 9 % in the fillet was observed (Digre et al., 2017; Lindberg, 2019; Olsen et al., 2013). Generally, short-term storage of cod (<6 hours), did not show improved attributes of the fillets (Olsen et al., 2013) or a better colour quality (Erikson et al., 2019). One study did show a minor effect of short-term storage of cod, with recovery and reduced redness of the fillet after six hours of storage (Olsen et al., 2013).

The strategy for CBA production of cod depends on the distance to the market, the locally used fishing gear and the fishing season. In Norway, economic considerations suggest extending the selling season of fresh, live cod from the winter months to the spring and summer (Hermansen and Eide, 2013). The long-term storage of live fish requires feeding to avoid muscle, liver- and gonad-depletion and rapid onset of rigor mortis after slaughter (Ageeva et al., 2018a, 2018b, 2017). In Greenland, the current strategy involves two to four weeks of live storage of the fish and processing of fish in the pre-rigor mortis state for production of frozen fillets. Feed deprivation of more than 26 days for cod resulted in higher water content, gelatinous texture, a typical white colour and less fresh sea odour of fillets (Ageeva et al., 2018a). To be able to process the feed-deprived cod, the time from slaughter to the onset of rigor mortis is crucial. Starvation of cod for 23 days reduced the pre-rigor mortis time at 0-1 °C from 29 hours to 17 hours. Further starvation, of up to 79 days, did not decrease this period further (Ageeva et al., 2018b). Female cod were more prone to weight loss during starvation with the loss being most noticeably measured for total weight and liver weight. This could be related to the protein concentration in the fillets being higher in the male cod, 16.3 % ± 0.4, compared to female cod, 14.9% ± 0.4 (Ageeva et al., 2017).

CBA production of cod has recently increased in Norway and Greenland. In 2018, more than 8,000 tonnes of cod was placed in cages in Norway for feeding and later slaughtering (Fiskeribladet, 2018). The same year 8,000 tonnes of cod was the fishing goal for CBA cod in Greenland (Labansen, 2018)(Fig. 4).

50000

40000

30000

20000

10000

0 2013 2014 2015 2016 2017 2018 2019

Figure 4 Catchment of Atlantic cod in Greenland from 2013 to 2019; purple bars indicate the tonnes produce with the CBA concept (Statistics Greenland, 2019, personal communication with Royal Greenland).

The CBA production in Greenland is illustrated in Fig. 6 from catch (Fig. 6A) to the point of freezing (Fig. 6E). This process is described in details in Paper II and market by Royal Greenland A/S as Nutaaq® (the Greenlandic word for “new”) and is characterised by using more decentralised cages and shorter holding time compared to the process in Norway (Ageeva et al., 2018b). The Greenlandic fishery for cod is based on the use of poundnets for 73 % of the annual catch (ICES, 2019b), followed by longlines (14 %), gill nets (5 %) and hooks (8 %) (ICES, 2019b). In contrast, the fishery in the Barents Sea is dominated by trawling, 71 %, (ICES, 2019a) and around Iceland, trawling accounts for 51 % and longlines for 30 % (ICES, 2019c). Mobile net gages have been developed by Royal Greenland and were comfortable for the fisherman to take out with the small vessels (Fig. 5).

Figure 5 Fishermen in the preparation of gathering the cod inside the Royal Greenland developed mobile net cages. Photo: Jonas Steenholdt Sørensen

Figure 6 A) Poundnet fishery in the Greenlandic fjords, in the conventional method, cod would be gutted and rinsed in the small boat. B) In the CBA method, the cod is moved to small mobile net cages and starved for two to four weeks. C) The live starved fish are collected and transported by well-boat. D) At the process plant, the cod are released to a net enclosure and allowed 12 hours to rest. E) At the capacity of the process line, the cod is pumped into the plant and anesthetise by electricity, decapitated, rinsed and bled in circulating refrigerated water. The headed and gutted cod were machine filleted, hand-trimmed and individual quick frozen in a gyrofreezer.

2. Aim of the study

The overall aim of this PhD project was to evaluate the CBA production of cod in Maniitsoq, Greenland, as well as three different ways of distribution for fillets of Atlantic cod from this process to the primary consumer market in Europe. The three ways of distribution were (i) fresh and superchilled fillets, (ii) frozen fillets and (iii) frozen and thawed (refreshed) fillets. For each of these types of products shelf-life and relevant indices of spoilage and quality were evaluated as described in section 4.2 - 4.5. Each distribution route was covered by an individual research paper and based on these papers, a best- practice for distribution of Atlantic cod from Greenland was made.

Paper I Superchilling in combination with modified atmosphere packaging resulted in long shelf-life and limited microbial growth in Atlantic cod (Gadus morhua L.) from capture-based-aquaculture in Greenland.

“The objective of the present study was to determine shelf-life and indices of spoilage of iced and superchilled Atlantic cod from CBA in Greenland and thereby to evaluate the feasibility of non- frozen transportation to Europe. Firstly, sensory, chemical and microbial changes were studied in a storage trial with aerobically or MAP stored cod. The spoilage microbiota was studied by culture-dependent techniques and by 16S rRNA gene amplicon sequencing. Secondly, to point out SSO and evaluate indices of spoilage the spoilage potential and the spoilage activity of isolates from the spoilage microbiota were determined.” Paper I

Paper II Short-term capture-based aquaculture of Atlantic cod (Gadus morhua L.) in combination with optimised slaughter process gives better physicochemical properties during frozen storage

“The objective of the research was to evaluate a newly developed process involving capture-based aquaculture (CBA) of cod in comparison to the conventional process and give a recommendation for frozen shelf-life. Shelf-life would be based on a freeze durability study, analysing changes of colour, texture, physiochemical and sensory parameters for the product from the conventional process and the CBA process. Samples were taken every three months for one year and for the CBA cod it was investigated the effect of lowering the storage temperature would improve stability.” Paper II

Paper III Thawed and chilled Atlantic cod (Gadus morhua L.) from Greenland - Options for improved distribution

“The objective of the present study was to determine shelf-life and indices of spoilage for thawed Atlantic cod from CBA in Greenland. Firstly, the effect of two different bleeding methods on microbial contamination of cod fillets was evaluated. Secondly, sensory, chemical and microbial changes of frozen, thawed and chilled cod fillet pieces were studied in a storage trial with four treatments including

chilled storage at 0 °C and 3 °C in air or MAP (40% CO2 and 60% N2). Finally, and independent storage trial with cod in air was performed at ~1.5 °C to evaluate the results of the first storage trial.” Paper III

Figure 7 Graphical summary of papers (Next page)

Best practice Paper III Paper II Paper I 2 fillets, thawed at aquaculture frozen Capture-based Atlantic cod or conventional (c) aquaculture (CBA) Capture-based Atlantic cod aquaculture Capture-based Atlantic cod aquaculture Capture-based o C longer storage at -40 maximum 6months, if IQF, Stored at -20 TVC, spoilageSensory (IQF) Individual qiuckfrozen in airorMAP Iced orsuperchilled Shipment: P.p , IAandCFC o C for o C. pH TVBN Texture measurement profileSensory TVC, pH TVBN spoilageSensory sumers markets. land to thelarge con Shipment from Green Colour measurement Water holdingcapacity

Total volatile basic nitrogen (mg-N/100 g) 100 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 0 5 0 EU limit for limit EU 2 4 P.p 6 Gadidae 8 , IAandCFC 10 family 12 Storage period(days)Storage 14 16 18 20 22 24 26 28 - 30 - 32 34 • • • shelf-life:Sensory • • • • shelf-life:Sensory Shelf-life of12days 0 -2 Refreshed cod chilledat • • • • shelf-life:Sensory -40 >15 months, CBA at at -20 9 -12months, CBA 6 -9months, c chilled inMAP >32 days, super chilled inair >32 days, super 22 days, icedinMAP 15 days, icedinair chilled, inMAP >32 days, icedor 19 days, icedinair 13 days, 13 days, o C

o C

o C 1.4 2.9 o o C C inair inair - -

3. Shelf-life and indices of spoilage

3.1 Freshness and quality In term of seafood quality, proceedings of the Final Meeting of the Concerted Action “Evaluation of Fish Freshness” highlighted the many aspects of quality, depending on observer’s relationship to the seafood sector. (Olafsdóttir et al., 1997). Consumers are generally concern with sensory properties, price value, safety and convenience. Factors involves in determining the quality are summarised in Fig. 8, stated in 1997, but since then additional factors could be included, such as climate impact (Ziegler et al., 2013). Sensory properties and indices of spoilage are the factors studied in Papers I, II and III.

Figure 8 Quality factors contributing to the overall quality and emphasis of the studied indices of spoilage, yellow markers indicate quality factors studied in papers I, II and III. Modified from (Oehlenschläger and Sørensen, 1997).

3.2 Sensory quality Sensory evaluation of seafood is typically used to determine if products are fresh or spoiled and the chemical, physical and microbial changes can then be related to sensory shelf-life and can in some instrumental measurements then be used as indices of freshness or spoilage (Dalgaard, 2000; FAO, 1995). In the post-World War II years, sensory judgments were performed in all steps of the distribution chain, but the judgments were not performed in a systemic quantitative way. Shewan et al. (1953)

15 developed a standardised scoring system that could objectively quantify the sensory changes during storage post-mortem for fresh and cooked fish. The changes involved appearance, odour, flavour and texture. This Torry Scheme uses a base score of ten and during post-mortem changes the score may decrease to zero. The European Union has a different scheme for freshness evaluation of seafood. The EU system is based on a qualitative judgment of the skin, eye, gills odour and texture of flesh. Depending on the objective judgment of these factors, the fish is graded with Extra, A, B or not admitted for consumption (EC, 1996).

The EU grading scheme and the Torry scheme are mainly developed for white fish and other typical species landed in the northern hemisphere. Australia needed a system useful for the species of the southern hemisphere and in the late 1970s and early 1980s, this need resulted in development of the Quality Index Method (QIM). The QIM is different from the EU and Torry scheme by developing a single scheme for each species. The aim of the QIM is to measure the degree and rate of change post- mortem (Hyldig et al., 2012), inspired by the relative rate of spoilage model which can predict that seafood storage at 10 °C spoils four times faster than if storage at 0 °C (Ratkowsky et al., 1982). The first schemes developed for species from the southern hemisphere are blue grenadier (Macurinus novaezelandiae) (Statham and Bremmer, 1983), trevalla (Hyperoglyphe antarctica) (Statham and Bremmer, 1985) and sardine (Sardinops sagax) (Fitz-Gerald and Bremner, 1998). Species related to the northern hemisphere were later developed with schemes for cod, Atlantic herring (Clupea harengus L.) (Jespersen and Heldbo, 1991) and 11 other species are now published (https://www.qim- eurofish.com/). For Papers I and III a version of the QIM scheme for cod fillet was used to determine the sensory shelf-life (Archer, 2010).

Quantitative Descriptive Analysis (QDA) differs from the QIM and Torry scheme by using a scale with no number indicator. For each attribute assessed, the assessors mark the score on a 10 cm line. It is assumed that the assessors will use different part of the scale, which is the strength of the QDA. The information obtained from the QDA is not made in absolute scores, but the relative difference between products (Murray et al., 2001). Paper II uses QDA to assess the sensory quality between two processing methods and three different storage temperatures.

3.3 Indices of spoilage Determining the sensory shelf-life requires a tested and trained assessor panel. To better understand the causality of the sensory changes and the responsible spoilage reaction, various physical, chemical, biochemical and microbial measurements can be used and related to sensory spoilage and shelf-life. As single-compound quality index (SCQI) measurement of total volatile basic nitrogen (TVBN), trimethylamine (TMA) and concentrations for different groups of microorganisms have been extensively used for fresh fish and values for these SCQI corresponding to sensory spoilage have been suggested (Olafsdóttir et al., 1997; Dalgaard, 2000, ICMSF, 2011; Paper I, Paper III). Combining multiple measurements into one multi-compound quality index (MCQI) has also been suggested e.g. for ratios between concentrations of different biogenic amines (Mietz and Karmas, 1977) or concentrations of nucleotides resulting from degradation of adenosine triphosphate (ATP) (Hamm, 1977). The knowledge of the post-mortem changes was used by a Japanese research group to propose an indicator (k value) of freshness based on the ratios (Karube et al., 1984). Later multivariate statistical methods were used to identify MCQI e.g. cold-smoked salmon (Salmo salar L.) where MCQI based on biogenic amines, pH values (Jørgensen et al., 2000) and various physio-chemical and microbial changes (Leroi et al., 2001) were suggested.

3.3.1 Indices of spoilage for fresh and superchilled cod (Paper I) The sensory shelf-life of Atlantic cod from CBA in Greenland was 15 days when iced and stored in air and 22 days for iced MAP cod fillets. With superchilling at -1.7 °C, the shelf-life was extended to >32 days for cod in both air and MAP (Table2, Paper I). For easier comparison of shelf-lives between studies, the equivalent shelf-life at 0 °C was calculated by using the square-root model for relative rates of spoilage (RRS) of fresh seafood from temperate waters (FSSP, 2014). When stored in air, the shelf- lives of the CBA produced cod in Paper I were longer than for 51 other cod products which on average had a shelf-life of 12.9 days at 0 °C ± 3.9 SD (Table 2).

TVBN in the form of amines is widely used as a spoilage indicator. In the European Union legislation a maximum level of TVBN for cod is 35 mg-N/100 g muscle (EC, 2008), with the TVBN representing the sum of ammonia, dimethylamine (DMA), TMA and other nitrogenous compounds (Howgate, 2010a, 2010b). TVBN can be determined through a number of different methods, including

17 direct water vapour distillation (Antonacopulous, N., Vynke, 1989), distillation of an acidic extract (Etienne, 2005) or micro diffusion of an acidic extract (Conway and Byrne, 1933).

TVBN or TMA were used as SCQI for 46 products and proved to be a very reliable SCQI. On average the TVN index was 0.6 ±2.2 SD days faster to classify the fish as spoiled compared to the sensory shelf-life. The increased level of TVBN would result in an increased pH level, yet the pH level was not suitable as an index of spoilage, even though a suggestion of an index of 7.0 was made in Paper I. In the literature presented in Table 2, information of the pH levels from a total of 39 products was obtained at the point of sensory spoilage. The average pH value was 6.7 ± 0.3 SD; the drawback with using pH as an index is the different initial pH levels observed at time of slaughter depend on the species and energy reverse. The pH value drop post mortem varied and for wild cod a drop from 6.8 to 6.1-6.5 has been reported (FAO, 1995). The drop was more significant in July and could reach 5.9; the pH value drop was correlated with heavy predation (Love, 1979). Farmed cod with regular feeding experienced the same pH drop, regardless of seasonality and a pH level 6.0-6.3 at the point of spoilage was observed (Table 2, Hansen et al., 2016, 2007; Mørkøre et al., 2006; Sivertsvik, 2007).

The flesh and internal organs of a live fish has proven to be sterile, while the skin surface and digestive tracts contain a diverse microbiota. Post-mortem, microorganisms in the form of bacteria are the most important factor for spoilage of fish (Tarr, 1954). Dyer (1947) showed that not all microorganisms were capable of reducing trimethylamine N-oxide (TMAO) to TMA or forming other spoilage-associated compounds. The species capable of producing spoilage compounds (spoilage potential) and metabolites in quantities (spoilage activity) causing organoleptic rejection could be described as those responsible for the spoilage. These species are called specific spoilage organism (SSO) (Gram and Huss, 1996).

Using a SCQI for total viable count (TVC) of bacteria in seafood has been tried and the latest guide from Microorganism in Food 8, stated that “Spoilage is typically detected when specific spoilage bacteria are >107 CFU/g.” (ICMSF, 2011). In Table 2, that statement is true for 41 out of 71 products, including iced cod from Greenland (Paper I). The maximum growth rate, determined by fitting the logistic growth curve with delay to the TVC, did not differ for the cod from Greenland compared to the literature date (Fig. 11, Table 2). Variation in the enumeration of microorganism is challenged by different studies choosing different growth media. The variation is clear when considered a product as iced cod in MAP (Paper I, Fig. 9), which showed the CFU concentration was much higher when the growth media was Long and Hammer (LH) compared to Iron agar (IA) Lyngby (NMKL, 2006). Broekaert et

18 al., (2011) investigated four growth media; LH, IA, marine agar (MA) and plate count agar (PCA). After seven days on ice, the TVC for cod was 6.6 (LH), 6.5 (MA), 5.4 (IA) and 4.8 (PCA).

8 g

/ 6

g o

L 4

0 0 5 10 15 20 25 30 Storage period (days)

Figure 9 Concentrations of microorganisms in iced cod in MAP, enumerated by ■ Long and Hammer and ♦ Iron agar, modified from Paper I.

In Table 2, three common genera of microorganism associated with spoilage of cod are listed,

H2S-producing organisms like Shewanella spp., Photobacterium spp. and Pseudomonas spp. Critical levels depend on the spoilage potential and activity of each species. A large number of studies have been involved in determining the spoilage potential of a number of species in different seafood products (Laursen et al., 2005; Olafsdottir et al., 2005; Paludan-Müller et al., 1998). A few studies, including Paper I, have studied the spoilage activity and correlated the concentration of the microorganisms and the development of TVBN (Dalgaard, 1995; Xie et al., 2018; Paper I). The spoilage activity of Photobacterium spp., Shewanella spp. and Pseudomonas spp. was observed as being significant higher for Photobacterium spp. than the two other genera, and based on calculated TVBN formation in relation to a monoculture of the three genera(Fig. 8), a SCQI of 7.3 log CFU/g for Photobacterium spp., 8.1 – 8.8 log CFU/g for Shewanella spp. and 9.5 CFU/g for Pseudomonas spp. is proposed. The values for spoilage activity in Paper I were the same for Photobacterium spp. with the literature (Dalgaard, 1995). For Shewanella spp. the spoilage activity was determined to be a little higher in Paper I as compared to Dalgaard, (1995) (Fig. 10).

200 190

180

) g

170 0

0 160

1 /

N 150 -

g 140

m (

130 n

e 120 g

o 110

t i

n 100

i 90 s

a 80

e 70

l i

t 60

a l

o 50 v

EU limit for the Gadidae family

l 40

a t

o 30 T 20 10 0 0 1 2 3 4 5 6 7 8 9 10 Log CFU/g

Figure 10 The spoilage activity of ● Photobacterium spp., ■ Shewanella spp. and ▲ Pseudomonas spp., icons symbolised measurement and lines are the calculated formation of TVBN. The blue line is the calculated TVBN formation of Shewanella spp. with the spoilage activity obtained from (Dalgaard, 1995).

3 )

y 2

(

m 

 1

0 -4 -2 0 2 4 6 8 10 12 14 16 Storage temperature (oC)

Figure 11 Relation between storage temperature (°C) and the square root transformed maximum growth rates (1/days). Line shows regression of all data from literature (Table 2), ●) growth rate of product in air, ■) growth rate of product in MAP (Table 2), ▲) growth rate from paper I in air, ▼) growth rate from paper I in MAP.

Table 2 Selected studies of shelf-life and indices of shelf-life in fresh cod and haddock at different storage temperatures and gas compositions.

Values at sensory shelf-life Storage conditions Shelf-life (days) Log CFU

Temperature Gas composition RRS TVBN Drip loss a b c d Species (°C) (CO2/N2/O2) Sensory at 0 °C (>35 mg-N/100 g) TVC H2S P.p Ps pH (%) References Cod 0.4 Air 15 16.2 14 7.3 6.3 7.3 6.9 7.0 - Paper I 0.4 40/60/- 22 22.8 20 7.3 4.2 7.0 3.5 6.9 - -1.7 Air >32 22.0 >32 7.8 6.9 5.0 7.7 6.8 - -1.7 40/60/- >32 22.0 >32 3.9 2.6 3.5 2.3 6.7 - Cod 1.5 Air 9 11.9 8 7.5 6.7 - - 7.2 - (Wang et al., 2008) -0.9 Air 16 13.3 16 7.9 7.5 - - 7.0 - 1.5 50/45/5 14 18.5 12 7.4 5.7 - - 7.0 3.6 -0.9 50/45/5 >24 19.9 24 >7.4 >5.5 - - >6.84 5.2 Cod 0.5 Air 12 13.2 11 8.2 6.9 7.3 6.6 6.8 - Cod 1.9 Air 12 17.0 11 8.1 7.0 7.2 6.9 - (Olafsdottir et al., 2005) 2 Air 11.5 16.6 10.2 8.0 6.1 7.2 6.7 - (Olafsdottir et al., 2006b) 0.4 Air 13.5 14.6 >15 8.0 7.1 7.2 6.9 - 1.9 Air 10.5 14.9 11.5 8.0 6.6 7.4 6.9 - -0.9 Air 12.5 10.4 11 7.9 5.9 6.8 7.0 - -0.3 Air 14 13.2 12 8.0 7.4 6.8 6.9 - -1.3 Air >15 11.4 15 8.0 7.6 6.0 6.9 - -0.8 Air 15 12.7 15 8.1 7.5 6.7 7.0 - Haddock 0.9 Air 13.5 16.0 12 8.1 5.6 8.0 6.7 - 7.2 Air 6.5 19.2 4.5 8.4 6.0 8.1 6.7 - (Olafsdottir et al., 2006a) 15.1 Air 4.5 28.1 3 8.2 6.1 8.2 6.8 - 0.9 Air 11.5 13.7 12 7.5 5.8 7.5 6.7 - 3 Air 9 15.2 8 8.2 6.1 8.0 6.9 - Cod 16 Air 1.5 10.1 - 6.0 5.1 - - - - (Einarsson, 1992) 8 Air 3 9.7 - 5.3 4.4 - - - - 4 Air 4 7.8 - 5.0 4.8 - - - - 0 Air 10.5 10.5 - 5.7 4.6 - - - - -2 Air 21 13.4 - 6.7 5.7 - - - - -4 Air >45 16.2 - 4.5 1.9 - - - - 16 40/30/30 1.5 10.1 - 5.9 5.7 - - - - 8 40/30/30 2.5 8.1 - 4.2 5.2 - - - - 4 40/30/30 5 9.8 - 4.8 6.1 - - - - 0 40/30/30 12.5 12.5 - 5.5 5.5 - - - - -2 40/30/30 36 23.0 - 5.6 5.6 - - - - -4 40/30/30 >45 16.2 - 2.3 >1 - - - -

Table 2 continued

Temperature Gas composition RRS TVBN Drip loss a b c d Species (°C) (CO2/N2/O2) Sensory at 0 °C (>35 mg-N/100 g) TVC H2S P.p Ps pH (%) References 16 40/30/30 1.5 10.1 - 5.9 5.7 - - - - 8 40/30/30 2.5 8.1 - 4.2 5.2 - - - - 4 40/30/30 5 9.8 - 4.8 6.1 - - - - 0 40/30/30 12.5 12.5 - 5.5 5.5 - - - - -2 40/30/30 36 23.0 - 5.6 5.6 - - - - -4 40/30/30 >45 16.2 - 2.3 >1 - - - - Cod 0 Vacuum 9 9.0 9 7.6 7.3 - - 7.0 6.5 (Cann et al., 1993) 5 Vacuum 2.5 5.6 3 7.7 7.3 - - 6.9 2.9 10 Vacuum 2 8.0 2 7.3 5.9 - - 7.1 3.7 0 40/30/30 12 12.0 12 6.9 6.7 - - 6.9 7.3 5 40/30/30 5 11.3 4 7.7 6.5 - - 6.9 3.3 10 40/30/30 3 12.0 2 7.6 6.5 - - 7.0 3.9 Cod 0 Vacuum 13 13.0 ------(Huss, 1971) 4 Vacuum 7 13.7 ------Cod 0 Air 14 14.0 ------(Merritt, 1965) -1 Air 20 16.2 ------2 Air 26 16.6 ------3 Air 35 17.2 ------Cod 1.5 Air 14 18.5 ------(Olsen et al., 1993) -2 Air 19 12.2 ------Cod -1 Air 15 12.2 >16 7.1 5.6 - - - - (Eliasson et al., 2019) -0.73 Air 15 12.9 >16 6.9 5.1 - - - - -0.33 Air 12 11.2 16 7.0 4.8 - - - - 0.67 Air 12 13.7 13 7.3 6.3 - - - - Cod 0 Air 11 11.0 >13 6.5 4.5 5.5 6 4 to 5 (Lauzon et al., 2009) -2 Air 14 9.0 >15 7.0 6.0 4.9 6.4 3 to 4 -3.6 Air >15 - >15 4.4 2.2 ND 4 - 0 50/45/5 14 14.0 11 7.2 4.6 7.3 3.5 9 -2 50/45/5 21 13.4 23 6.1 4.7 5 3.9 6 -3.6 50/45/5 >28 13.7 >28 6.0 3.3 3.2 3.6 - -2 Air 12 7.7 14 7.2 6.2 5.7 5.7 1 to 2 -3.6 Air >15 - >15 5.9 5.2 2.4 5.8 - -2 50/45/5 13 8.3 25 5.6 3.4 2.4 3.7 >2 -3.6 50/45/5 >28 13.7 >28 5.0 3.1 1.8 4.3 -

Table 2 continued

Temperature Gas composition RRS TVBN Drip loss a b c d Species (°C) (CO2/N2/O2) Sensory at 0 °C (>35 mg-N/100 g) TVC H2S P.p Ps pH (%) References Cod 1 60/32/8 9 10.9 12e 5.5 4.9 - - - - (Woyewoda et al., 1984) 1 60/32/8 12 14.5 12e 5.8 4.3 - - - - 1 Air 9 10.9 9e 5.3 4.4 - - - - 1 Air 9 10.9 9e 5.2 4.9 - - - - Cod 4 Air <4 7.8 - 8.2 6.6 - 5.9 6.7 - (Kuuliala et al., 2018) 4 60/35/5 4 7.8 - 7.0 5.5 - 5.1 6.3 - 4 60/0/40 6 11.8 - 7.3 7.2 - 5.4 6.3 - 8 60/35/5 3 9.7 - 8.0 4.7 - 6.2 6.6 - 8 60/0/40 3 9.7 - 7.3 5.3 - 5.6 6.4 - Cod 1.9 Vacuum 7 9.9 - 5.7 5.0 - - 6.1 4 (Hansen et al., 2016) 1.9 Vacuumf 7 to 13 9.9 - 5.3 4.5 - - 6.0 4.2 1.9 60/40/- 7 to 13 9.9 - 5.7 4.0 - - 6.0 4.2 1.9 60/40f 10 to 16 14.2 - 5.5 4.1 - - 6.0 2.9 Cod 1.3 Vacuum 7 to 14 8.9 - - - - - 6.3 10.8 (Hansen et al., 2007) 1.3 Vacuumf 14 to 21 17.9 - - - - - 6.1 12.3 1.3 60/0/40 14 to 21 17.9 - - - - - 6.1 11.5 Cod 0 Air 9 9.0 - 8.0 - - - - - (Shewan, 1965) -1.5 Air 12 8.7 - 5.1 - - - - - 1 Air 13 15.7 - 8.9 ------1 Air 22 17.8 - 8.4 - - - - - Cod 0 100/0/0 >14 >14 - 6.3 - - - - 17.5 (Sivertsvik, 2007) 0 0/0/100 >14 >14 - 6.4 - - - - 7.5 0 0/100/0 10 to 14 14 - 7.2 - - - - 6.1 0 50/0/50 >14 >14 - 4.3 - - - - 7.1 0 50/50/0 >14 10 to 14 - 7.3 - - - - 9.2 0 0/50/50 >14 >14 - 6.3 - - - - 7.3 0 67/17/17 >14 >14 - 5.8 - - - - 8 0 17/17/67 >14 14 - 7.0 - - - - 7.6 0 17/67/17 >14 14 - 7.4 - - - - 9.5 0 33/33/33 >14 >14 - 6.4 - - - - 8.4 a Total variable count. b H2S-producing bacteria. c Photobacterium phosphoreum-like bacteria determined by conductance method (Dalgaard et al., 1996). d Pseudomonas spp. bacteria determined by spread plating on Pseudomonads agar (CM0559, Oxoid, Basingstoke, UK) with CFC selective supplement (SR0103, Oxoid, Basingstoke, UK). e Determined by TMA. f Packed together a CO2 emitter, containing citric acid. 23

3.3.2 Indices of spoilage for frozen cod (Paper II) The Torry scheme and QIM are both developed for fresh and chilled storage of seafood and the changes that happen to frozen seafood are characterised differently from those for chilled seafood. In frozen cod, a cold storage odour was identified from lipid oxidation of the cis-4-heptenal molecule (McGill et al., 1974). In the Recommendations for the Processing and Handling of Frozen Foods, the shelf-life of frozen foods are of different types; high-quality life (HQL) and practical storage life (PSL) (Bøgh-Sørensen, 2006) and is defined thus:

“The High-Quality Life is defined as the time elapsed between the freezing of an initially high-quality product and the moment when, by sensory assessment, a statistically significant difference (often P < 0.05) from the initial high quality (immediately after freezing) can be established. This is the Just Noticeable Difference (JND).”

“The practical storage life of a product is the period of frozen storage at a given temperature during which the product retains its characteristic properties and remains fully acceptable for consumption or the intended process.”

In Paper II it was logistically not possible to assess the sensory QDA immediately after freezing. The recommendation is to compare a sample with a sample store at a temperature below -60 °C. For the conventional method, 7 out of 13 attributes reach a significant difference from the control sample (CBA, -80 °) and the JND was at nine months. Using the CBA method storage at -20 °C, 5 out of 13 attributes reach a significant difference and the JND was at twelve months when excluding gaping, which may be due to poor thawing procedure. Lowering the temperature to -40 °C, only 2 out of 12 of the attributes reached a significant difference and the JND was at 12 months. The taste attributes were significantly changed after 15 months for the two products stored at -20 °C and might be an indication of end of the PSL. These determined shelf-lives were a bit longer for the HQL than other publish sensory HQLs and the PSL of 15 months at -20 °C was similar to observations from the Torry Research Station (Table 3).

Applying a SCQI based on SSP has not been proposed before. In Paper II, the suggestion of a PSL criteria of <60 % SSP or 0.2 mg SSP g-1 fish muscle, and a HQL criteria of >70 % SSP or 0.3 mg SSP g-1 fish muscle was made. There was no clear agreement between the studies of cod, hake and haddock for the reported HQL, compared to the reported HQL for lean fish by the Torry Research Station for temperature of -10 to -20 °C. Lowering the temperature to -30 °C, with the SCQI based on SSP, showed a

24 more promising application with an HQL of 14 months for both the Torry Research Station and multiple studies (Table 3).

Bøknæs et al., (2000) proposed criteria for SCQI based on WHC, the HQL criteria was when the muscle had a WHC above 70 % and PSL when the WHC drop below 60 %, the studied cod was storage at -28 °C. The HQLs determined by Burgaard and Jørgensen (2010) were longer than those found by Bøknæs et al. (2000). In Paper II, the cod were stored at -20 and -40 °C and are therefore not comparable to the HQL, at -28 °C, found by Bøknæs et al. (2000). While Burgaard & Jørgensen (2010) stored the cod at the same temperature as in Paper II and both the HQL and PSL was in similar range (Table 3).

Table 3 Frozen durability of cod and a few other seafood products.

Frozen storage shelf-life (months) Water holding capacity Salt soluble proteins Other criteria (WHC) (SSP) High High Quality Life Practical Storage Life Quality Life Practical (SSP >0.3mg proteins (SSP >0.2mg proteins Temperature High Practical (WHC >70 Storage Life g-1 fish muscle ) g-1 fish muscle ) Product (°C) Quality Life Storage Life %) (WHC >60 %) References Seafood, lean -12 - 4a - - - - (Bøgh-Sørensen, fish -18 - 9a - - - - 2006) -24 - >12a - - - - White fish -9.4 1a 4a - - - - (Torry Research (gutted) -20.6 4a 15a - - - - Station, 2001) -28.9 8a 48a - - - - Cod, -18.8 - 3 a - - - - IFF-IIR (TRS) haddock, etc. -25 - 6a - - - - -28.9 - 8a - - - - Cod -20 6 to 9b 12 to 15b 6c 7 10d >12d Paper II CBA cod -20 9 to 12b 12 to 15b 4c 10 10d >12d -40 >15b >15b >12c >12 >12d >12d -80 - - >12c >12 >12d >12d Cod -10 - - <1 <1 - - (Burgaard and -20 - - 4 10 - - Jørgensen, 2010) -30 - - >18 >18 - - -40 - - >18 >18 - - -50 - - >18 >18 - - -60 - - >18 >18 - - -70 - - >18 >18 - - -80 - - >18 >18 - - Cod -28 - - 7 9 - - (Bøknæs et al., 2001) -28 - - 7 9 - - -28 - - 7 >10 - - -28 - - 7 >10 - - -28 - - <7 7 - - Cod -20 - - 12 >12 - - (Bøknæs et al., 2002) -30 - - >12 >12 - - Cod -25 - - >1.5 - - - (Roiha et al., 2017) -25 - - >1.5 - - - -25 - - >1.5 - - -

Table 3 continued High Quality Life Practical High Quality Life Practical Storage Life Temperature High Practical (WHC >70 Storage Life (SSP >0.3mg proteins (SSP >0.2mg proteins Product (°C) Quality Life Storage Life %) (WHC >60 %) g-1 fish muscle ) g-1 fish muscle ) References (Sánchez-Alonso et - <1 Hake -10 - - - - al., 2012) (Godiksen et al., cod -30 >8.5e >8.5e - - - - 2003) -20 1.5e 3.5e - - - - -20 1.5e 6e - - - - cod -30 - - - - 2.7 >7 (Badii and Howell, -10 - - - - <1 3.3 2002a) -10 - - - - <0.5 1.5 -10 - - - - <0.5 1.5 -10 - - - - <0.5 <1 Hake -20 - - - - 7.0 10.5 (Del Mazo et al., -30 - - - - 9.0 >12 1999) Cod -20 - - - - 10.0 >14 -30 - - - - >14 >14 (Careche et al., 1998) Cod -14 - - - 6f - - (Schubring, 2005) -20 - - - 7f - - -28 - - - >13f - - -20 - - - 5f - - Cod -10 - - - 5f - - (Schubring, 2004) -20 - - - >10f - - -30 - - - >10f - - -20 - - - >10f - - (Siddaiah et al., Silver carp -18 - - - - >6 - 2001) cod -10 - - - - 0.3 0.5 (Badii and Howell, -30 - - - - 14 >14 2002b) haddock -10 - - - - 0.5 1 -30 - - - - 13 >14 cod -12 - - - - <10 10 (LeBlanc and LeBlanc, -15 - - - - <10 >10 1992) -22 - - - - >10 >10 -30 - - - - >10 >10 a Storage shelf-life define by reference. b Just notable difference for 12 attributes (gaping was excluded) to determine HQL: significant difference for metallic, sweet and bitter taste attributes for PSL. 27 c WHC (%) above 65. d SSP and water-soluble proteins were not separated, HQL limit for >70 % and PSL limit of >60 %. e 2+ -1 -1 2+ -1 -1 HQL Ca ATPase activity above 15µmol Pi min g muscle , PLS Ca ATPase activity above 5 µmol Pi min g muscle . f WHC measure as water loss, limit for PSL of 15 %.

3.3.3 Indices of spoilage for refreshed cod (Paper III) Table 4 includes 36 products of refreshed seafood with a defined shelf-life and storage in air. The equivalent shelf-life at 0 °C showed an average shelf-life of 12.3 days ± 4.0 SD. The shelf-life of refreshed seafood is in general not extended when compared to fresh seafood (Section 3.4 equivalent shelf-life at 0 °C was 12.9 days). The shelf-lives in Paper III were in general much longer with an equivalent shelf-life at 0°C of 21.9 days, using the RRS for fresh seafood from temperate water. Using the RRS for fresh seafood from tropical water (FSSP, 2014), the equivalent shelf-life at 0 °C was 17.9 days. The difference between the models is the temperaturedependent variable constant. Since the typical SSO, Photobacerterium spp. and Shewanella spp. for fish from temperate waters are psychrotolerant and not to same extent affected by temperatures from 0 to 15 °C (Dalgaard, 2003), the constant for the temperate water is higher in this temperature range compared to the RRS for tropical water. The tropical water RRS model might be more suitable for refreshed seafood, but only in cases where Photobacterium spp. and Shewanella spp. were inactivated by the frozen storage period. Magnússon and Martinsdóttir, (1995) observed an extended shelf-life of five days when the frozen storage at -25 °C was 52 weeks compared to 27 weeks or shorter.

TVBN or TMA were measured in 30 products and in contrast to fresh cod, the correlation between sensory spoilage and the SCQI with TVBN was poor in refreshed cod or redfish (Sebastes sp.). On average, TVBN spoilage was detected 3.8 days ± 4.3 SD after the sensory spoilage. Based on 48 products, the TVC at the point of sensory spoilage was higher in refreshed cod and redfish, TVC = 7.2 ± 1.0 SD, than in fresh cod. The poor performance of TVBN as an SCQI and increased TVC at the time of spoilage supports the suggestion of a spoilage course by Pseudomonas spp.. The TVBN formation is not associated with Pseudomonas spp. (Paper I), but generate an off-flavour characterised by being sweet and fruity (Castell and Greenough, 1956; Miller et al., 1973).

Previously, refreshed and chilled MAP cod often resulted in high drip losses (Table 4) in a level not acceptable for the producers (Paper III). The drip loss from CBA cod from Greenland in MAP was much lower, <3.6 %, a level acceptable for the producers. The reason behind the low drip loss has not been investigated, but the texture score for the QIM in paper I (data not shown) was also low for the superchilled cod. Further studies to identify the texture properties of the cod in relation to superchilling, freeze-thawing cycles and MAP storage would be interesting. Two main question are: is the cod from Greenland phenotypical different from cod from Iceland and Norway in relation to texture properties? Is the CBA method responsible for the texture properties?

Table 4 Selected refreshed seafood products and their shelf-life and indices of spoilage.

Storage conditions Frozen storage Shelf-life (days) Values at sensory shelf-life

Gas Drip Temperature composition Temperature Period loss a b c d e Species (°C) (CO2/N2/O2) Fresh/Thawed (°C) (Months) Sensory TVBN TVC H2S P.p Ps pH (%) Reference Cod 2.9 Air Thawed -20 4 13 14 8.1 N.D. - 8.3 7 2.4 Paper III 0.4 Air Thawed -20 4 19 22 8.7 N.D. - 8.7 6.9 2.5 2.9 40/60/- Thawed -20 4 >32 >32 7.7 N.D. - 5.9 6.8 3.6 0.4 40/60/- Thawed -20 4 >32 >32 5.4 N.D. - 3.5 7 3.4 1.4 Air Thawed -20 3 13 >18 7.2 N.D. - 7.5 6.7 - Cod 1.6 40/60/- Fresh - - 11 12F 7.6 7.5 - - 6.92 5.4 (Guldager et al., 1.6 40/40/20 Fresh - - 13 15F 7.3 7.4 - - 6.79 6.4 1998) 1.6 40/60/- Thawed -20 2 >20 >20 5.7 N.D. - - 6.71 14.5 Cod 2 60/40/- Thawedf -20 1.5 >17 >17 6.6 N.D. - - 6.7 - (Bøknæs, 2000) 2 60/40/- Thawedf -30 1.5 >17 >17 6.9 4 - - 6.7 - 2 60/40/- Thawedg -20 1.5 >17 >17 7.4 N.D. - - 6.7 - 2 60/40/- Thawedg -30 1.5 >17 >17 7 6 - - 6.7 - Cod 2.1 40/40/20 Thawed -29 3 14 14 - 8.1 - - - - (Bøknæs et al., 2.5 40/40/20 Thawed -31.5 6 14 14 - 7.1 - - - - 2002) 2.5 40/40/20 Thawed -24.6 6 21 >21 - 3.5 - - - - 2.1 40/40/20 Thawed -32.2 9 14 14 - 7 - - - - 2.1 40/40/20 Thawed -23.5 9 21 >21 - N.D. - - - - 2.5 40/40/20 Thawed -32.4 12 14 14 8.1 7.6 - - 6.9 11.8 2.5 40/40/20 Thawed -23.4 12 21 >21 8.5 N.D. - - 7 12.8 Cod 0 to 2 Air Thawed -25 1.5 14 14 8.3 - 7 - - 1.7 0 to 2 Air Thawed -25 1.5 14 14 8.4 - 7 - - 1.9 (Roiha et al., 2017) 0 to 2 Air Thawed -25 1.5 14 14 8.4 - 7.1 - - 1.3 Cod 2.9 Air Thawed -28 2 >6 >6 4.8 - 2.3 - 6.7 2.9 2.9 Air Thawed -28 2 >6 >6 4.2 - 1 - 6.8 3.3 (Roiha et al., 2018) Cod 0 Air Fresh - - 15 15 4 - - - - - 0 Air Thawed -25 2 15 - 4.5 - - - - - (Magnússon and 0 Air Fresh - - 10 17h 7.7 - - - - - Martinsdóttir, 1995) 0 Air Thawed -25 0.1 10 18h 8.3 - - - - - 0 Air Thawed -25 1 10 >20 7.9 - - - - - 0 Air Thawed -25 1.5 10 >20 7.5 - - - - - 0 Air Thawed -25 6 10 >20 7.3 - - - - - 0 Air Thawed -25 12 14 >20 7.9 - - - - -

Table 4 continued

Gas Drip Temperature composition Temperature Period loss a b c d e Species (°C) (CO2/N2/O2) Fresh/Thawed (°C) (Months) Sensory TVBN TVC H2S P.p Ps pH (%) Reference Sebastes 0 Air Fresh - - 9 18i 7.2 - 5.6 - - - sp. 0 Air Thawed -25 0.1 9 19i 5.6 - 4.3 - - - 0 Air Thawed -25 1.5 9 19i 5.9 - 3.8 - - - 0 Air Thawed -25 6 9 >20i 6.4 - 4.0 - - - Cod 0 Air Fresh - - 8 8 7.7 - - - 7.0 - (Vyncke, 1983) 0 Air Thawed -28 <1 10 10 7.7 - - - 7.0 - 0 Air Thawed -28 3 11 >13 7.7 - - - 6.9 - 0 Air Thawed -28 6 11 >13 7.5 - - - 7.0 - 0 Air Thawed -28 12 11 >13 7.8 - - - 6.9 - Cod 0 40/-/60 Thawed -23 10 19 - 7.0 - 6 6.4 6 - (Hansen et al., 1.7 40/-/60 Thawed -23 10 19 - 8.2 - 7.8 6.8 5.9 - 2015) 0 40/-/60 Thawed -23 10 19 - 7.5 - 7 7 7.2 - 1.7 40/-/60 Thawed -23 10 19 - 8.4 - 7.7 7.4 6 - 0 40/-/60 Thawed -23 10 19 - 8.0 - 7.5 6.9 6.3 - 0 Air Thawed -23 10 7 - 6.1 - 5.6 5 5.2 - Cod 0 Air Thawed -25 2 >12 >12 7.3 - 4.8 - 6.7 1 0 Air Thawed -25 2 >12 >12 7.8 - 5.0 - 6.7 1.2 (Martinsdottir and 4 Air Thawed -25 2 7 9 6.4 - 4.8 - 6.8 1 Magnusson, 2001) 4 Air Thawed -25 2 7 4 7.0 - 3.9 - 7.1 1.2 0 Air Thawed -25 6 12 ------0 Air Thawed -25 6 12 ------4 Air Thawed -25 6 6 ------4 Air Thawed -25 6 6 ------0 Air Thawed -25 12 10 >14 6.5 - 4.3 - 6.6 - 0 Air Thawed -25 12 10 >14 7.3 - 5.7 - 6.8 - 4 Air Thawed -25 12 7 >9 6.0 - 3.8 - 6.6 - 4 Air Thawed -25 12 7 7 7.0 - 3.8 - 6.7 - 0 Air Thawed -25 17 7 ------0 Air Thawed -25 17 7 ------4 Air Thawed -25 17 ------4 Air Thawed -25 17 ------Salmon 2 60/40/- Fresh - - 14 - - 7.3 4.5 - - - (Emborg et al., 2 60/40/- Thawed -20 1 >21 - - N.G. 4.7 - - - 2002) 2 60/40/- Fresh - - 21 - - 6.5 5.1 - - - 2 60/40/- Thawed -30 1 34 - - N.D. 7.0 - - -

Table 4 continued

Gas Drip Temperature composition Temperature Period loss a b c d e Species (°C) (CO2/N2/O2) Fresh/Thawed (°C) (Months) Sensory TVBN TVC H2S P.p Ps pH (%) Reference Whiting 4 Air Fresh - - >3 >3 4.14 - - - - 0 (Fagan et al., 2003) 4 Air Fresh - - >3 >3 5.54 - - - - 1 4 Air Thawed -30 0.1 >3 >3 4.04 - - - - 9 4 Air Thawed -30 0.1 >3 >3 5.24 - - - - 6 Mackerel 4 Air Fresh - - >3 >3 4.52 - - - - 0 4 Air Fresh - - >3 >3 5.34 - - - - 2 4 Air Thawed -30 0.1 >3 >3 4.26 - - - - 10 4 Air Thawed -30 0.1 >3 >3 5.14 - - - - 4 Salmon 4 Air Fresh - - >3 >3 5.08 - - - - 0 4 Air Fresh - - >3 >3 7.36 - - - - 1.7 4 Air Thawed -30 0.1 >3 >3 4.78 - - - - 3.3 4 Air Thawed -30 0.1 >3 >3 7.56 - - - - 3.2 Whiting 2 to 4 Air Thawed -30 0.1 >5 >5 4.81 - - - - 4.71 (Fagan et al., 2004) 2 to 4 40/30/30 Thawed -30 0.1 >5 >5 4.48 - - - - 9.4 2 to 4 100/-/- Thawed -30 0.1 >5 >5 4.34 - - - - 16.4 Mackerel 2 to 4 Air Thawed -30 0.1 >5 >5 4.88 - - - - 4.63 2 to 4 40/60/- Thawed -30 0.1 >5 >5 4.18 - - - - 5 2 to 4 100/-/- Thawed -30 0.1 >5 >5 3.99 - - - - 6.63 Salmon 2 to 4 Air Thawed -30 0.1 >7 >7 6.23 - - - - 2.45 2 to 4 40/60/- Thawed -30 0.1 >7 >7 5.04 - - - - 4.68 2 to 4 100/-/- Thawed -30 0.1 >7 >7 4.53 - - - - 5.84 a TVBN concentration above 35 mg-N/100 g as indicated by EU (2008). b Total variable count. c H2S-producing bacteria. d Photobacterium phosphoreum-like bacteria determined by conductance method (Dalgaard et al., 1996). e Pseudomonas spp. bacteria determined by spread plating on Pseudomonads agar (CM0559, Oxoid, Basingstoke, UK) with CFC selective supplement (SR0103, Oxoid, Ba-singstoke, UK). f Cod frozen after one-day storage on ice. g Cod frozen after eight-day storage on ice. h TVBN above 25 mg-N/100g as indicated by EU (2008). i TMA concentration above 20 mg-N/100 g. N.D. Not detected N.G. No growth

3.4 Best practice for CBA cod production, processing and distribution The CBA method for cod production has shown to improve the texture properties, reduce discolouration due to standardised bleeding procedure (Paper II) and the starvation period prior to slaughter (Olsson et al., 2006). Given the long distance from the fishing grounds in West Greenland to the primary market in Europe, the recommendation would be to freeze the cod to -40 °C. Using the low storage temperature would ensure a frozen HQL shelf-life of the highest quality for minimum one year (Paper II). If chosen to increase the storage temperature to -20 °C, a HQL shelf-life of six months is recommended to avoid poor WHC (Paper II). For convenience, consumers prefers to buy non-frozen seafood and in the case of distributing non-frozen seafood, Paper I showed that the margin for error is low, even for superchilled products. Based on good texture properties, the recommendation for distributing non-frozen cod would be to use refreshed cod (Paper III) with a shelf-life of maximum 12 days at 0 – 2 °C in air. MAP could extend the shelf-life, but only if the frozen temperature is-20 °C (Paper III). Longer shelf-life would increase the risk of growth by Listeria monocytogenes (Paper III) and potentially decrease the sensory taste and flavour qualities. Consumers from Iceland, Denmark and the Netherlands rated refreshed cod as the highest quality; higher than fresh, wild or farmed cod (Sveinsdóttir et al., 2010). Using the right marketing strategy, consumers would also consider buying refreshed cod (Altintzoglou et al., 2012).

4 Food waste and loss Food waste has been defined in different ways within scientific literature and regulations. To limit confusion the definition of Parfitt et al. (2010) is used within the present PhD thesis:

“(1) Wholesome edible material intended for human consumption, arising at any point in the FSC (Food supply chain, author) that is instead discarded, lost, degraded or consumed by pests. (2) as (1), but including edible material that is intentionally fed to animals or is a by-product of food processing diverted away from human food.”

Within the seafood area, examples of food waste include bycatch that is discarded in the sea as dead animals and species landed and utilised for animal feed. Furthermore, there might be an additional decrease in fillet yield, representing food waste, for seafood like the cod. When the fish is landed and aimed for human consumption, the edible parts of the fillet might be removed to maintain satisfactory standard as a result of damage to the fillet. Food loss is defined as food ready for human consumption

33 but which is spoiled or discarded in the food supply chain (FSC), including the home storage and lack of utilisation by the consumers (Grolleaud, 2002).

4.1 Food waste for cod fishing and processing In the seafood sector, bycatch is the first encounter in the FSC of food waste. Based on the whole global database of reported landing and discard information, during the ten years from 1992 to 2001, a weighted discard rate average of 8.0 % was found (Kelleher, 2005). The discard rate was highly dependent on geographical location, type of fishery and gear selection. The discard values range from 3.5 % in the Southeast Pacific to 37.7 % in the Western Central Atlantic. Shrimp fishing, and especially tropical fisheries, generate high discard rates and account for 27.3 % of the total global discard. Excluding data from the shrimp fishery, the discard rate for finfish was 3.5 % for midwater fishery, including Alaska pollock, 19.6 % for the demersal round fish fishery including Atlantic cod, 39.6 % for deep-water fisheries and 53.1 % for demersal flatfish fishery (Kelleher, 2005).

If bycatch cannot be prevented, the landed catch could be a source of new species for human consumption as exemplified by the low discard rate of 1.1 % in a demersal multispecies fishery. An explanation of the low discard rate in multispecies fishing is that the fishery includes the Chinese and Southeast Asian fisheries, where the utilisation of many different species which in other fisheries, such as the demersal round fish fishery, were discarded (Kelleher, 2005). Paper II and section 2.3 CBA describe the fishing gear used in West Greenland for Atlantic cod, where the pound nets in combination with the fishing ground and season provided a fishery with low bycatch. The typical bycatch includes lumpfish and Greenland cod (Gadus ogac), and these fish are released to the fjords with no mortality (Personal communication with fisherman).

The conventional fishing and processing method included transportation of slaughtered gutted cod packed in boxes and covered by ice. In the event of more fish being caught than could fit in available boxes, the cod might be transported as bulk. Cod and haddock transported and stored for 4-8 days were graded at the landing site, and two factors were the drivers for a lower grading. These factors were the fishing season (July being the worst) and storage in bulk (Savagaon and Power, 1976). The lower grading resulted in a decrease in production yield (fillet or whole) of 1.75 to 5.12 percent points (Savagaon and Power, 1976). The reduced production yield had been linked to a softer texture of the fish (Himelbloom et al., 1997).

Figure 12 Atlantic cod fished and produced with the conventional method (Fig. 5A), stored in boxes and covered by iced, source: Royal Greenland.

Short live storage (<12hours) of Atlantic cod and haddock has shown that bruises of the flesh could be reduced and texture quality increased (Olsen et al., 2013). The increased texture quality resulted in increased fillet yield (Himelbloom et al., 1997; Venugopal and Shahidi, 1998) and thereby reduced food waste. When handling live stored cod, gear damage to the fish could occur in the pumping unit and pipes. In 2017, the setup of the CBA process in Maniitsoq had issues with broken vertebrate and bruises due to sharp bends of the pipes from the well-boat to the processing plant (Fig. 5 D to E). The bruises increased waste in the same way as gear damage in trawling. After modification of the pipes, the waste related to broken vertebrate was significantly reduced (personal communication with plant manager).

5.2 Food losses for cod during distribution In tropical and developing countries the majority of food losses occur at the handling and processing stage of the FSC, while for industrial countries, the food losses occur mainly at the consumer stage (FAO, 2011). Love et al. (2015) estimated seafood losses in the American FSC during 2009 to 2013. Postharvest handling and storage resulted in a loss of 10,000 – 11,000 tonnes (Gustavsson et al., 2013; Love et al., 2015). Further processing and packaging of seafood resulted in a loss of 33,000 – 35,000 tonnes (Gustavsson et al., 2013; Love et al., 2015). In the distribution, retail and consumption stage of the FSC, 102,000 to 147,000 tonnes of fresh and frozen seafood were lost (Buzby et al., 2009; Love et al., 2015). The most substantial loss was estimated to happen in the consumption stage with 455,000 to 569,000 tonnes (Love et al., 2015; Muth et al., 2011).

In cases where the conventional method is split, such as the first process step being in Greenland and the second in China the cod is slaughtered as the conventionally processed cod (Paper II) and frozen as headed and gutted (H&G) in a vertical freezer in 20 kg bulk blocks. The H&G blocks are

35 transported to China and thawed, filleted and frozen once again The thawed-frozen cycles soften the texture (Fig. 12, Schubring, 1999) and the softer H&G cod result in lost fillet yield and thereby generate food loss at this stage of the FSC (Fig. 12, Himelbloom et al., 1997).

There is a social stigma against food loss in the retail sector. If food loss was successfully managed, it could help to reduce high food costs. Therefore, the Consumer Goods Forum, a network of >400 retailers and manufacturers from 70 countries with a total revenue of 2.5 trillion €, has set an aim to reduce food loss by 50 % in 2025 (The Consumer Goods Forum, 2015). To identify measurements to reduce food loss, qualitative studies haves been conducted. For the Norwegian FSCs, a case study identified a range of parameters leading to food losses. Grouping the parameters, the loss could be reduced by improved planning and handling between the wholesaler and retail sections of the FSC. The planning includes application of data to predict demand, like more accurate forecasts and based on data, the planning decisions could be optimised. The handling of the food products should minimise damage to the product, from storing products in wrong temperature zone to mechanical damage by operators, machines or customers. The chosen plan should be executed and ensure the right products is picked for delivery (Chabada et al., 2014).

The identified food losses were found in all parts of the FSCs and to reduce the losses, cooperation between actors was necessary (Göbel et al., 2015). For all perishable food bought in the UK, it is estimated that 2.0 million tonnes out of 4.2 million tonnes is lost due to “not being used in time”. It can be concluded that extending the shelf-life is one of the critical parameters to reduce food loss (Lee et al., 2015). Choosing the right storage condition for fresh, frozen or refreshed cod products could extend the shelf-life with >200 % across all products, taken from the shortest found shelf-lives to the longest shelf-lives (Paper I, II and III).

Spada et al. (2018) proposed a relationship between the number of non-sold products (returned goods) and shelf-life based on retail data from Italy with 826 food products, including 640 dairy and 186 non-dairy foods. There was no correlation for products with a shelf-life in the range of 0 – 30 days, due to large variation, consisting of 27.7 % of the dataset, while there was a correlation (R2 = 0.453) for products with a shelf-life of 31 - >700 days. The level of correlation indicates how well the model describes the food loss. The model for estimating the returned goods and thereby an indication of food loss is described in Eq. 1.

3.866 푅푒푡푢푟푛푒푑 푔표표푑푠 (%) = −0.009 + × 100 Eq. 1 푆ℎ푒푙푓 푙𝑖푓푒 (푑푎푦푠)

For frozen CBA cod, the HQL was extended from six to twelve months by lowering the temperature from -20 °C to -40 °C (Paper II), and the PSL could be extended from ten to twelve months. These shelf-life extensions and Eq. 1 suggest an 87 % reduction of food loss for premium cod when lowering the frozen storage temperature. The returned goods corresponded to 1.2 % and 0.2 % at the two storage temperatures. For the PSL the reduction was at a smaller magnitude at 0.2 % and a reduced food loss of 57.1 %.

D) Transport A) Slaugther, processing B) Sea transport. 8 C) Repackaging in E) Supermarket. F) Consumer +5 within Europe. 1 and packaging. 2 days. days. Europe. 1 day. +2 °C. °C. day, +2 °C.

Figure 13 Food Supply chain (FSC) of Atlantic cod from Greenland to Europe of fresh cod, the shelf-life determined in Paper I starts at stage A and the temperature in A-C is equal to those in Paper I.

E) Transport D) Thawing F) A) Slaugther and B) Packaging C) Sea within G) Consumer. and Supermarket. processing. and freezing. transport. Europe. 1 day, +5 °C. repackaging. +2 °C. +2 °C.

Figure 14 Food Supply chain (FSC) of Atlantic cod from Greenland to Europe of frozen and thawed cod, the shelf-life determined in Paper III starts at stage D.

Common for fresh and refreshed cod studied in Papers I and III, are that the shelf-lives are in the range of 0-30 days (Table 2 and 4) and the model of Spada et al. (2018) for frozen products is thus not applicable. However, four other models are available to correlate shelf-life of perishable products with food loss, such as those in Papers I and III (Broekmeulen and van Donselaar, 2019; Buisman et al., 2019; Eriksson et al., 2016). For these models, one of the input parameters is the remaining shelf-life at the retail stage of the FSC (Fig. 13 for fresh cod, Fig. 14 for refreshed cod). The observed shelf-lives in Table 2 are all from stage A in the FSC (Fig. 13) for fresh cod and the shelf-lives for refreshed cod in Table 4 are from stage D (Fig. 14). The remaining or available shelf-life (ASL) at the retail stage can be calculated if the previous stages of the FSC has a known time and temperature history. Transportation from wholesaler to supermarket is assumed to take one day and the storage temperature during transportation to the supermarket and at the supermarket is assumed to be within the legal EU limit for fresh fish, 2 °C (EC, 2004).

To calculate the ASL for each product studied in Paper I, the RRS model (Eq. 2), is used. Here, T is the storage temperature (°C) and Tref is the reference temperature (°C) at which the shelf-life is determined. In practice, the calculations were performed with the Food Spoilage and Safety Predictor software (FSSP, 2014).

푇+10 √푅푅푆 = Eq. 2 푇푟푒푓+10

The ASL for the four different storage conditions in Paper I are calculated to be two days, seven days and nine days (the same shelf-life for the two products stored at superchilled condition), rounded to nearest whole day (Table 5). In the same way, the ASL was calculated for cod products in Paper III, the difference being that the starting point for shelf-life was at the repackaging stage in Europe (Fig. 9 D) and it was assumed that the frozen storage was within shelf-life range determined in Paper II for frozen cod. The ASL for the five different storage conditions in Paper III were 11, 13, 14, 30 and 30 days (Table 5).

The model of Buisman et al. (2019) estimates food loss at retail stages of FSCs. The parameters used are order-up-to level of 16 and an average demand of five products a day, with an applied Poisson distribution to randomise the actual demand per day. Other parameters included in the model are the shelf-life, profit margin of the product, order size from the wholesaler and consumer behaviour (Buisman et al., 2019). With ASL as the model input, the predicted food loss decreased rapidly while extending the shelf-life (Table 5). With an ASL of nine days or more, the model estimated no food loss and limited occurrences of shortages of products in the supermarket.

Eriksson et al. (2016) published a model to estimate relative food loss in relation to demand, ASL and minimum order size. The correlation between the ASL and relative food loss was based on a dataset of 984 products from Swedish supermarkets, including 92 cheese products, 258 dairy products, 333 deli products and 331 meat products. The multilinear regression has an R2 of 0.666. By shifting from fresh cod (Paper I) to refreshed cod (Paper III), with their respective shelf-lives, the predicted food loss was reduced by 48-53 % using fresh cod in air as a reference and by 16-23 % if fresh cod in MAP was used a reference product (Table 5).

The model of Broekmeulen and von Donselaar (2019) for food loss is based on 17,000 different perishable foods within three types of products; fruits and vegetables, fresh meat and convenience products. The supermarket data were obtained in Europe and covered small, medium and large stores. Within this model two different concepts are used to access the potential of reducing food loss in the

38 supermarket. The first and most simplistic was the fresh case cover (FCC) concept (Eq. 3), which was developed to be applied in the retail sector for decision makers without the mathematical background of more elaborate models (Broekmeulen and van Donselaar, 2019). The inputs for calculating the FCC are product case pack size, average daily sales and the ASL. The FCC is calculated by Eq. 3, where Q is the case pack size, m is the remaining shelf-life (ASL) and µ is the average daily demand.

푄 퐹퐶퐶 = Eq. 3 푚×휇

If the demand is assumed to be deterministic and constant, i.e. no randomised demand from day to day, food loss will only occur if the FCC > 1. In reality, the demand is stochastic and with a stochastic demand the food loss will still occur if FCC < 1. The other concept is the efficient frontier (EF), which estimates the food loss depending on the same inputs as for the FCC and a chosen level of on- shelf availability (OSA).

Table 5 Fresh case cover estimates for cod with shelf-lives from Paper I and III.

(Buisman et al., (Eriksson et (Broekmeulen and van Donselaar, 2019)a al., 2016) 2019)c

Product, storage Efficient conditions Available Relative Fresh frontiers, and shelf-life Waste Shortage food lossb case Corresponding food loss Source temperature (days) (%) (%) (%) cover food loss (%) (%) Paper Air, 0.1 °C 2 25.13 0.52 Ref 3.00 200.7 I MAP, 0.1 °C 7 0.05 0.85 -38 Ref 0.86 22.0 13.3 Air, -1.7 °C 9 0.00 0.85 -44 -9 0.67 17.1 5.7 MAP, -1.7 °C 9 0.00 0.85 -44 -9 0.67 17.1 5.7 Paper Air, 2.9 °C 14 0.00 0.85 -53 -23 0.43 11.5 0 III Air, 1.4 °C 11 0.00 0.85 -48 -16 0.55 14.3 2.7 Air, 0.4 °C 13 0.00 0.85 -51 -21 0.46 12.3 0 MAP, 2.9 °C 30 0.00 0.85 -65 -43 0.20 5.3 0 MAP, 0.4 °C 30 0.00 0.85 -65 -43 0.20 5.3 0 a Modified unpublished model of the published model in Buisman et al. (2019). b The model returns a relative food loss and not exact quantitative date, food loss reduction is calculated with fresh MAP iced cod. c Assumption of an average daily demand of 2 (CU/day) case pack size (CU) of 12.

From the FCC results in Table 5, it is clear, due to a FCC >> 1, that shipment of fresh iced cod in air from Greenland to Europe is not feasible as this results in high levels of predicted food loss. Correlating the FCC values with the average food loss of the 17,093 items, that is the basis of the model, showed that even for the refreshed cod in MAP with an FCC of 0.2, a food loss of 5.3 % is estimated to

39 happen on average. That is a substantial reduction from the fresh iced cod in MAP with an FCC of 0.86, correlating to a 22.0 % food loss. The R2 of the correlation between the FCC and food loss was equal to 0.42, meaning that 42 % of the variance is explained. When applying the FCC to quantify food loss of a specific single product as in Table 5, the actual food loss could vary a lot from the estimated food loss since the FCC is correlated to the average food loss.

To minimise the uncertainty of the FCC concept, with averaging the food loss from many different products, the EF is a more advanced indicator with product-specific inputs. In Fig. 15, the food loss to OSA relation is plotted, based on the ASL in Table 5 and a setting for the model with an average daily demand at the supermarket of two items and a case pack size of 12, the smallest possible order for the supermarket. The curve for fresh iced cod in air (Paper I) is not shown in Fig. 15, as the curve is above 100 % food loss.

Changing the EF curve by changing the FSC and thereby increasing the ASL leads to three different sales strategies (Fig. 16). 1) sell more, by keeping the food loss constant and increasing the OSA, 2) reduce food loss, by keeping the OSA constant and 3) sell more and reduce food loss, by changing both the OSA and reducing the food loss. For fresh iced cod in MAP (Fig. 15 ●) and an OSA value of 95 %, the predicted food loss would be 13 %. If reducing food loss is the main goal, then changing to refreshed cod in air will reduce food loss to 0-2.7 % without changing the OSA (Fig. 15 ♦ and▼).

d 20

% 15

0 80 85 90 95 100 % OSA

Figure 15 Efficient frontiers for the food supply chains studied in Paper I and Paper III, ● fresh iced cod in MAP, ■ Superchilled fresh cod in air or MAP, ▼ refreshed cod in air at 1.4 °C, ♦ refreshed cod in air at 0.4 °C, ▲ refreshed cod in air at 2.9 °C and ● refreshed cod in MAP.

Figure 16 Sales strategies, when using the efficient frontiers model to assess food supply chains. Waste in the graph should be characterised as loss with the definition by Grolleaud, (2002), source: Broekmeulen and van Donselaar, (2019).

The EF model can be correlated to the profits of different FSCs. With more retail data it is possible to estimate the most cost-effective OSA for each FSC. In addition to the initial three inputs, four cost parameters must be included; “% margin of sales price”, “% out of stock substitution”, “% lost sales cost of sales price” and “% ordering cost of sales price”. With these additional data, the OSA with lowest cost per demand can be predicted. Fig. 17 is calculated based on retail data supplied by Broekmeulen and van Donselaar, (2019) and is not actual obtained data for Atlantic cod. With these cost parameters, the fresh iced cod in MAP has a most cost-effective OSA of 97 %. To choose the right OSA for the products in Papers I and III, more reliable data directly linked to Atlantic cod should be investigated.

Choosing a sales strategy for the retail stage of the FSC with limited time for the consumer to utilise the fish at home might be the wrong strategy. The EF model can estimate the remaining shelf-life to the consumers correlated with the OSA (Fig. 17). The drawback of the estimated remaining shelf-life is the model’s assumption that the spoilage rate is not affected by the temperature changes between the storage temperature at the supermarket and the fridge in the consumer’s home. With a highly perishable and temperature-sensitive product such as raw fish, the assumption in Figure 17 is a temperature at the consumers’ household of 2 °C and the same as in the supermarket and no significant transportation time and temperature abuse from the supermarket to the consumers. The consumers’ refrigerators were between 3.9 °C and 5 °C as factory settings, but with regular behaviour of door opening and the additional room environment, the temperature was often higher (Rodriguez-Martinez et al., 2019). A change from 2 °C to 4 °C results a 36 % increase on the RRS (FSSP, 2014).

0.30 6

0.25 5

0.20 4

life life (days) - 0.15 3

0.10 2

Cost per Costdemand € per 0.05 1 Remaining shelf Remaining 0.00 0 80 85 90 95 100 80 85 90 95 100 % OSA % OSA

Figure 17 The left graph shows the scenario for fresh, iced cod in MAP to determine the most cost-effective on shelf availability (OSA). The right graph shows the remaining shelf-life to the consumers with the same model parameters as the left graph.

The EF model has a number of an assumptions, including a fixed one-day delivery and unpacking period from the wholesaler to the shelf at the supermarket, a one-day fixed reviewing and ordering time for restocking, no change of demand during the week and first-in first-out (FIFO) buying habit for the consumers (Broekmeulen and van Donselaar, 2019). Additionally, the model does not consider availability at the wholesaler either; for products in Paper I the supply from Greenland to Europe depends on sail routes and in the year 2017-2019 only a weekly supply was available. Refreshed products (Paper III) has the advantage of more flexible ordering, as the demand and ordering can easier be timed with the forecast. When the retailer can order in smaller quantities and the flexibility of spreading orders during the week, it has been identified that this would reduce food loss (Göbel et al., 2015).

6 Sustainable Development Goals In September 2015, the United Nations agreed on 17 goals to continue the global development from the eight 2015 goals adopted in 2000 (United Nations, 2020). The new goals were called Sustainable Development Goals (SDG) and covered large parts of food supply chains, from the environment to the labour force, polices and international cooperation (United Nations, 2020). The impact of this PhD project and the changed fishing and processing methods in Maniitsoq was evaluated by identifying the SDG that was most influenced both positively and negatively.

6.1 Positively impacted SDG SDG number 12 with the title “responsible consumption and production” includes eight defined targets. The two targets that are most improved by the PhD project are “12.2 By 2030, achieve the sustainable management and efficient use of natural resources” and “12.3 By 2030, halve per capita global food waste at the retail and consumer levels and reduce food losses along production and supply chains, including post-harvest losses”. The indicator for the two targets are “12.2.1 Material footprint, material footprint per capita, and material footprint per GDP (Gross Domestic Product, author)”, “12.2.2 Domestic material consumption, domestic material consumption per capita, and domestic material consumption per GDP” and “12.3.1 Global food loss index”, source of the description of targets and indicators are from the United Nations (United Nations, 2020).

12.2 is improved by CBA production of cod, as described in Paper II. The improved texture quality, in combination with frozen storage stability of the proteins, increases product yield (Venugopal and Shahidi, 1998). The material footprint for each product available to the market in Europe is significantly reduced compared to the previous processing with frozen H&G cod shipped from Greenland over Europe to China, where it was thawed, processed, repacked and frozen before being shipped back to Europe (Tybjerg, 2018). Greenhouse gas (GHG) emissions for Atlantic cod, filleted in

Norway and transported frozen to Paris was at 2.51 kg CO2e/kg edible product when reaching the wholesaler. The GHG emission increased to 3.78 kg CO2e/kg edible product, for the same cod, with filleting in China instead of in Norway (Ziegler et al., 2013). Due to the longer transport from Greenland to Europe the difference in GHG emission is expected to be slightly smaller than the 34% reported by Ziegler et al. (2013).

12.3 can be improved by the knowledge of the shelf-life from Papers I and III. With the optimal FSC, the food loss can be reduced by 80 % if the cod is sold as a refreshed iced product in air in comparison with fresh iced cod in MAP as discussed in section 6.2.

SDG number 10 with the title “Reduced inequality within and among countries” has seven targets. The target 10.1 “10.1 By 2030, progressively achieve and sustain income growth of the bottom 40 per cent of the population at a rate higher than the national average” has indirectly been improved by changing the fishing and processing methods in Maniitsoq, Greenland. Maniitsoq is a town with 2,501 – 2,670 inhabitants during the years 2013-2020 (Statistics Greenland, 2020a) and a large part of the inhabitants work at the local fish factory or in the industry servicing the factory. From 2013, the year before the first trial fishery with CBA cod in Maniitsoq, to 2017, the average income increased from DKK 190,531 to DKK 227,132 annually corresponding to a 19.2 % increase (Statistics Greenland, 2020b). The 19.2 % increase is higher than the average in any of the four municipalities in Greenland. The increased average income in the municipalities were: Kujalleq 13.1 %, Sermersoq 10.8 %, Qaasuitsup 13.4 % and Qeqqata (including Maniitsoq) 15.3 % (Statistics Greenland, 2020c). Sermersoq Municipality is the richest in Greenland and the increased income in Maniitsoq was also higher in absolute value (36,601 DKK) than in Sermersoq (28,535 DKK), indicating that inequality between Maniitsoq and Sermersoq decreased in the years of CBA production.

SDG number 14, “Conserve and sustainably use the oceans, seas and marine resources for sustainable development” has seven targets and the one improved mostly by the CBA fishing and procession of cod is “14.B Provide access for small-scale artisanal fishers to marine resources and

44 market”, source of the description of targets and indicators are from the United Nations (United Nations, 2020). The foundation of the CBA strategy is a close corporation with local small-scale artisanal fishers, these fishers operate the pound nets with small vessels, typical with one or two for each household (Fig. 5A). The number of local artisanal fishers, that supply Royal Greenland has increased from 47 in 2016 to 188 in 2019.

6.2 Negatively impacted SDG SDG number 14 was also negatively impacted directly and indirectly by the CBA of cod. The two targets that are negatively impacted are

“14.4 By 2020, effectively regulate harvesting and end overfishing, illegal, unreported and unregulated fishing and destructive fishing practices and implement science- based management plans, in order to restore fish stocks in the shortest time feasible, at least to levels that can produce maximum sustainable yield as determined by their biological characteristics” and

“14.6 By 2020, prohibit certain forms of fisheries subsidies which contribute to overcapacity and overfishing, eliminate subsidies that contribute to illegal, unreported and unregulated fishing and refrain from introducing new such subsidies, recognizing that appropriate and effective special and differential treatment for developing and least developed countries should be an integral part of the World Trade Organization fisheries subsidies negotiation”, source of the description of targets and indicators are from the United Nations (United Nations, 2020).

14.6 was directly linked to Royal Greenland and the CBA concept, by the new fishery, has increased profitability in the fishing for Atlantic cod. The cod in West Greenland is overfished in the years involving in the CBA production. Royal Greenland has subsidized local fishers by financial loans for fishing gear, enabling a larger part of the population to take part in the fishery. No part of the fishing was illegal or unregulated as other parts of the goal aimed at preventing.

The CBA production of cod indirectly impacts 14.4. The fishery for cod as illustrated in Fig. 6 and was regulated by the government of Greenland. During the years 2013 to 2019, the annual catch has been within the regulation of the government (ICES, 2019b). In the same years, the biological assessment advised to lower the annual catchment to keep the stock cod at a sustainable level. The FMSY

45 for the inshore cod from West Greenland was introduced in 2018, and the biological stock could sustain a catch between 6,806-8,858 tonnes annually. The politically chosen TAC was significantly higher than the FMSY and the FPA for 2018 and 2019, corresponding to an overfishery every year of 252 – 355 % (ICES,

2019b; Statistics Greenland, 2019). The FMSY for West Greenland has dropped in the years since 2017, for restoring the fishing stock of Atlantic cod in West Greenland, it is important to reduce the TAC for the coming years. From a sales perspective, consumers choose seafood with the Marine Stewardship Council (MSC) label (Thrane et al., 2009) and to apply to label the cod from Greenland with the MSC label, the core principle should be followed and the first of these principles are; the fishery should be sustainable and within the FMSY (MSC, 2020).

SDG number 8,“Promote sustained, inclusive and sustainable economic growth, full and productive employment and decent work for all” and the target “8.5 By 2030, achieve full and productive employment and decent work for all women and men, including for young people and persons with disabilities, and equal pay for work of equal value” is improved, but work is still needed with the implementation of the CBA fishing and processing in Maniitsoq. Unemployment in Maniitsoq has dropped from an annual maximum in 2013 of 186 to 135 in 2018 (Statistics Greenland, 2020d).

6. Conclusions

The present PhD project had enlighten on shelf-lives for fresh, frozen and refreshed CBA cod from Greenland. A sensory shelf-life of 15 days was found for iced cod in air and could be extended when stored in MAP. More efficient was to lower the temperature to -1.7 °C and when combining the superchilled condition with MAP, the microbial growth was limited. The frozen shelf-life was most depended on storage temperature, the HQL was six to nine month for the conventional and CBA methods at -20 °C. Lowering the temperature to -40 °C extended the HQL to more than 15 months. The change of WHC and SSP was only depending on storage temperature and not the fishing and processing methods. Cod that had been frozen for five months, was shown to have a 19 days sensory shelf-life after thawing and stored on ice. The extension for the fresh cod, was due to the inhibition of Photobacterium spp. and Shewanella spp. in the freezing stage.

The laboratory research obtained shelf-lives was applied to simulate the food loss of FSC of fresh or refreshed cod. A change from fresh iced cod in MAP to refreshed cod in air, had a potential of

46 reducing the food loss by 80 %. The reduction is an important step toward the UN SDG number 12, with an overall aim of reducing the world food loss by 50 %.

7. Perspectives

The CBA production and related processing of cod, have shown to generate some high quality texture properties of the Atlantic cod. Stored as fresh and superchilled, the sensory texture score was stable through the storage trial of 32 days. The frozen CBA cod had a higher texture hardness measurement compared to the conventional method and in a storage trial, in MAP, the drip loss is low. It would be interesting to investigate if the texture properties originate from the CBA production method, the genetic of the cod or something else. Especially the properties to maintain the protein structure of the superchilled or the refreshed MAP cod is unique and without explanation with the current knowledge.

One factor that could influence the texture, but also other aspects of sensory evaluation, is the season for the fishery. It is well known that the summer period results in softer texture of the fillet and with a lower WHC compared to cod from the fall or early spring. It would be interesting and vital to know if the cod from the summer months have a shorter frozen shelf-life and thereby a higher likelihood of a market complain of poor texture quality.

The SSO for fresh iced cod from Greenland was identified to be P. carnosum and the associated spoilage potential and activity was found. For refreshed cod, no SSO was identified and a deeper understanding of the microbiota with amplicon sequencing of the 16S rRNA gene would be necessary if further work of spoilage potential and activity should be conducted. Controlled sequencing with the gyrB gene might identify more species, cod and salmon fillet had a higher species richness when performing the amplicon sequencing with gyrB instead of 16S rRNA, while the opposite was true for pork sausages (Poirier et., 2018)

“Oh choosing a fresh cod: “The head should be large; tail small; shoulders thick; liver creamy white; and the skin clear and silvery with a bronze like sheen.” – British admiralty, Manual of Naval Cookery, 1921 (Kurlansky, 1999)

Marine Stewardship Council (MSC) is one of the main consumer driven labels within the seafood industry. The first core principle of MSC is that the fishery should be sustainable and within the FMSY (MSC, 2020; Thrane et al., 2009). Currently, the TAC of Atlantic cod has been 3 times higher than the

FMSY in West Greenland during the years 2016-2019. To get the same production mass from fewer individuals and thereby to maintain the profitably of the processing plant, a perspective would be to start feeding on the captured cod. One Norwegian producer started farmed cod production in January 2020 with 123,000 individuals of juvenile cod of an average size of 140 grams (Fiskeribladet, 2020).

References

Ageeva, T.N., Jobling, M., Olsen, R.L., Esaiassen, M., 2017. Gender-specific responses of mature Atlantic cod (Gadus morhua L.) to feed deprivation. Fish. Res. 188, 95–99. https://doi.org/10.1016/j.fishres.2016.12.010

Ageeva, T.N., Olsen, R.L., Joensen, S., Esaiassen, M., 2018a. Quality aspects of fillet, loin and tail products made from live-stored feed-deprived Atlantic cod (Gadus morhua L.) at different times post mortem. Lwt 97, 656–661. https://doi.org/10.1016/j.lwt.2018.06.031

Ageeva, T.N., Olsen, R.L., Joensen, S., Esaiassen, M., 2018b. Effects of long-term feed deprivation on the development of rigor mortis and aspects of muscle quality in live-stored mature Atlantic cod (Gadus morhua L.). J. Aquat. Food Prod. Technol. 27, 477–485. https://doi.org/10.1080/10498850.2018.1448919

Altintzoglou, T., Nøstvold, B.H., Carlehög, M., Heide, M., Østli, J., Egeness, F.A., 2012. The influence of labelling on consumers’ evaluations of fresh and thawed cod fillets in England. Br. Food J. 114, 1558–1570. https://doi.org/10.1108/00070701211273027

Anders, N., Howarth, K., Totland, B., Handegard, N.O., Tenningen, M., Breen, M., 2019. Effects on individual level behaviour in mackerel (Scomber scombrus) of sub-lethal capture related stressors: Crowding and hypoxia. PLoS One 14, 1–20. https://doi.org/10.1371/journal.pone.0213709

Antonacopulous, N., Vynke, W., 1989. Determination of volatile basic nitrogen in fish: a third collaborative study by the West European Fish Technologists’ Association (WEFTA). Z Leb. Unters Forsch 309–316.

Archer, M., 2010. Sensory assessment scoresheets for fish and shellfish https://www.seafish.org/media/Publications/sensory_assessment_scoresheets_14_5_10.pdf (2010), Accessed 26th Sep 2019

Audet, C., Besner, M., Munro, J., Dutil, J.-D., 1993. Seasonal and diel variations of various blood parameters in Atlantic cod (Gadus morhua) and American plaice (Hippoglossoides platessoides). Can. J. Zool. 71, 611–618. https://doi.org/10.1139/z93-083

Badii, F., Howell, N.K., 2002a. Effect of antioxidants, citrate, and cryoprotectants on protein denaturation and texture of frozen cod (Gadus morhua). J. Agric. Food Chem. 50, 2053–2061.

https://doi.org/10.1021/jf010824f

Badii, F., Howell, N.K., 2002b. A comparison of biochemical changes in cod (Gadus morhua) and haddock (Melanogrammus aeglefinus) fillets during frozen storage. J. Sci. Food Agric. 82, 87–97. https://doi.org/10.1002/jsfa.998

Bøgh-Sørensen, L., 2006. Recommendations for the Processing and Handling of Frozen Foods., 4th ed. International Institute of Refrigeration, France.

Bøknæs, N., Guldager, H.S., Østerberg, C., Nielsen, J., 2001. Production of high quality frozen cod (Gadus morhua) fillets and portions on a freezer trawler. J. Aquat. Food Prod. Technol. 10, 33–47.

Bøknæs, N., Jensen, K.N., Guldager, H.S., Østerberg, C., Nielsen, J., Dalgaard, P., 2002. Thawed chilled Barents Sea cod fillets in modified atmosphere packaging-application of multivariate data analysis to select key parameters in good manufacturing practice. LWT - Food Sci. Technol. 35, 436–443. https://doi.org/8

Bøknæs, N., Østerberg, C., Nielsen, J., Dalgaard, P., 2000. Influence of freshness and frozen storage temperature on quality of thawed cod fillets stored in modified atmosphere packaging. LWT - Food Sci. Technol. 33, 244–248. https://doi.org/10.1006/fstl.2000.0634

Brinkhof, J., Larsen, R.B., Herrmann, B., Olsen, S.H., 2018. Assessing the impact of buffer towing on the quality of Northeast Atlantic cod (Gadus morhua) caught with a bottom trawl. Fish. Res. 206, 209– 219. https://doi.org/10.1016/j.fishres.2018.05.021

Broekaert, K., Heyndrickx, M., Herman, L., Devlieghere, F., Vlaemynck, G., 2011. Seafood quality analysis: Molecular identification of dominant microbiota after ice storage on several general growth media. Food Microbiol. 28, 1162–1169. https://doi.org/10.1016/j.fm.2011.03.009

Broekmeulen, R.A.C.M., van Donselaar, K.H., 2019. Quantifying the potential to improve on food waste, freshness and sales for perishables in supermarkets. Int. J. Prod. Econ. 209, 265–273. https://doi.org/10.1016/j.ijpe.2017.10.003

Buch, E., Horsted, S.A., Hovgård, H., 1994. Fluctuations in the occurrence of cod in Greenland waters and their possible causes. ICES Mar. Sci. Symp. 198, 158–174.

Buchmann, K., Larsen, J.L.L., Dalsgaard, I., 1993. Diseases and injuries associated with mortality of hatchery-reared Baltic cod (Gadus morhua L) larvae. Acta Vet. Scand. 34, 385–390.

Buisman, M.E., Haijema, R., Bloemhof-Ruwaard, J.M., 2019. Discounting and dynamic shelf life to reduce fresh food waste at retailers. Int. J. Prod. Econ. 209, 274–284. https://doi.org/10.1016/j.ijpe.2017.07.016

Burgaard, M.G., Jørgensen, B.M., 2010. Effect of temperature on quality-related changes in cod (Gadus morhua) during short- and long-term frozen storage. J. Aquat. Food Prod. Technol. 19, 249–263. https://doi.org/10.1080/10498850.2010.510944

Buzby, J.C., Wells, H.F., Axtman, B., Mickey, J., 2009. Supermarket loss estimates for fresh fruit, vegetables, meat, poultry, and seafood and their use in the ERS loss-adjusted food availability data. Econ. Inf. Bull. 44, 43–68.

Cann, D.C., Smith, G.L., Houston, N.C., 1993. Further studies on marine fish stored under modified atmosphere packaging. Torry Research Station, Aberdeen.

Careche, M., Del Mazo, M.L., Torrejón, P., Tejada, M., 1998. Importance of frozen storage temperature in the type of aggregation of myofibrillar proteins in cod (Gadus morhua) fillets. J. Agric. Food Chem. 46, 1539–1546. https://doi.org/10.1021/jf970841y

Castell, C.H., Greenough, M.F., 1956. The action of Pseudornonas on fish muscle: 1. Organisms responsible for odours produced during incipient spoilage of chilled fish muscle. J. Fish. Res. Board Canada 14, 617–625. https://doi.org/https://doi.org/10.1139/f57-020

Chabada, L., Damgaard, C.M., Dreyer, H.C., Hvolby, H.-H., Dukovska-Popovska, I., 2014. Logistical causes of food waste: A case study of a Norwegian distribution chain of chilled food products, in: Grabot, B., Vallespir, B., Gomes, S., Bouras, A., Kiritsis, D., Turner, A.J., Goedicke, M., Murayama, Y. (Eds.), Advances in Production Management Systems Innovative and Knowledge-Based Production Management in a Global-Local World. Springer, pp. 273–280.

Conway, E.J., Byrne, A., 1933. An absorption apparatus for the micro-determination of certain volatile substances: The micro-determination of ammonia. Biochem. J. 27, 419-429.

Costello, C., Cao, L., Gelcich, S., Cisneros, M.A., Free, C.M., Froehlich, H.E., et al., 2019. The Future of Food from the Sea.

Dahle, G., Jørstad, K.E., Rusaas, H.E., Otterå, H., 2006. Genetic characteristics of broodstock collected from four Norwegian coastal cod (Gadus morhua) populations. ICES J. Mar. Sci. 63, 209–215.

https://doi.org/10.1016/j.icesjms.2005.10.015

Dalgaard, P., 2003. Spoilage of seafood, in: Caballero, B., Trugo, L., Finglas, P. (Eds.), Encyclopedia of Food Sciences and Nutrition. Elsevier Science Ltd, London, pp. 2462–2471.

Dalgaard, P., 2000. Fresh and lightly preserved seafood, in: Man, C.M.., Jones, A.. (Eds.), Shelf-Life Evaluation of Foods. Aspen publishing inc., Gaithersburg, pp. 110–139.

Dalgaard, P., 1995. Qualitative and quantitative characterization of spoilage bacteria from packed fish. Int. J. Food Microbiol. 26, 319–333. https://doi.org/https://doi.org/10.1016/0168-1605(94)00137- U

Dalgaard, P., Mejlholm, O., Huss, H.H., 1996. Conductance method for quantitative determination of Photobacterium phosphoreum in fish products. J. Appl. Bacteriol. 81, 57–64. https://doi.org/10.1111/j.1365-2672.1996.tb03282.x

Del Mazo, M.L., Torrejón, P., Careche, M., Tejada, M., 1999. Characteristics of the salt-soluble fraction of hake (Merluccius merluccius) fillets stored at -20 and -30 °C. J. Agric. Food Chem. 47, 1372–1377. https://doi.org/10.1021/jf9807420

Digre, H., Hansen, U.J., Erikson, U., 2010. Effect of trawling with traditional and “T90” trawl codends on fish size and on different quality parameters of cod Gadus morhua and haddock Melanogrammus aeglefinus. Fish. Sci. 76, 549–559. https://doi.org/10.1007/s12562-010-0254-2

Digre, H., Rosten, C., Erikson, U., Mathiassen, J.R., Aursand, I.G., 2017. The on-board live storage of Atlantic cod (Gadus morhua) and haddock (Melanogrammus aeglefinus) caught by trawl: Fish behaviour, stress and fillet quality. Fish. Res. 189, 42–54. https://doi.org/10.1016/j.fishres.2017.01.004

Dreyer, B.M., Nøstvold, B.H., 2008. Capture-based aquaculture of cod, in: Lovatelli, A. Holthus, P.F. (Eds.), Capture-based aquaculture. Global overview. FAO Fisheries Technical Paper. No. 508. Rome, FAO. pp. 183–198.

Dyer, F.E., 1947. Microorganisms from Atlantic cod. J. Fish. Res. Board Canada 7a, 128–136. https://doi.org/https://doi.org/10.1139/f47-014

EC, 2008. Commission Regulation (EC) No 1022/2008 of 17 October 2008 amending Regulation (EC) No 2074/2005 as regards the total volatile basic nitrogen (TVB-N) limits. Eur. Comm. 18–20.

EC, 2004. Commission Regulation (EC) No 853/2004 of the European Parlamient and of the Council of 29 April 2004 laying down specific hygiene rules for on the hygiene of foodstuffs. Eur. Comm. L 139, 55.

EC, 1996. Council Regulation (EC) No 2406/96 of 26 November 1996 laying down common marketing standards for certain fishery products. Off. J. Eur. Communities L344, 1–15.

Einarsson, H., 1992. Predicting the shelf life of cod (Gadus morhua) fillets stored in air and modified atmosphere at temperatures between -4 °C and +16 °C, in: Huss, H.H., Jakobsen, M., Liston, J. (Eds.), Quality Assurance in the Fish Industry. Elsevier B.V, Amsterdam, pp. 479–500.

Eliasson, S., Arason, S., Margeirsson, B., Bergsson, A.B., Palsson, O.P., 2019. The effects of superchilling on shelf-life and quality indicators of whole Atlantic cod and fillets. LWT - Food Sci. Technol. 100, 426–434. https://doi.org/10.1016/j.lwt.2018.10.062

Emborg, J., Laursen, B.G., Rathjen, T., Dalgaard, P., 2002. Microbial spoilage and formation of biogenic amines in fresh and thawed modified atmosphere-packed salmon (Salmo salar) at 2°C. J. Appl. Microbiol. 92, 790–799. https://doi.org/10.1046/j.1365-2672.2002.01588.x

Erikson, U., Tveit, G.M., Bondø, M., Digre, H., 2019. On-board live storage of Atlantic cod (Gadus morhua): Effects of capture stress, recovery, delayed processing, and frozen storage on fillet color characteristics. J. Aquat. Food Prod. Technol. 28, 1076–1091. https://doi.org/10.1080/10498850.2019.1684406

Eriksson, M., Strid, I., Hansson, P.A., 2016. Food waste reduction in supermarkets - Net costs and benefits of reduced storage temperature. Resour. Conserv. Recycl. 107, 73–81. https://doi.org/10.1016/j.resconrec.2015.11.022

Etienne, M., 2005. Volatile amines as criteria for chemical quality assessment. Seaf. Plus 1–22.

Fagan, J.D., Gormley, T.R., Mhuircheartaigh, M.U., 2003. Effect of freeze-chilling, in comparison with fresh, chilling and freezing, on some quality parameters of raw whiting, mackerel and salmon portions. LWT - Food Sci. Technol. 36, 647–655. https://doi.org/10.1016/S0023-6438(03)00084-7

Fagan, J.D., Gormley, T.R., Uí Mhuircheartaigh, M.M., 2004. Effect of modified atmosphere packaging with freeze-chilling on some quality parameters of raw whiting, mackerel and salmon portions. Innov. Food Sci. Emerg. Technol. 5, 205–214. https://doi.org/10.1016/j.ifset.2004.01.001

FAO, 2020a. FishStatJ - Software for fishery and aquaculture statistical time series.

FAO, 2020b. FAO Fisheries & Aquaculture - Species Fact Sheets - Gadus morhua (Linnaeus, 1758) [WWW Document]. URL http://www.fao.org/fishery/species/2218/en (accessed 1.26.20).

FAO, 2018. The State of World Fisheries and Aquaculture 2018 - Meeting the sustainable development goals. Rome.

FAO, 2011. Global food losses and food waste – Extent, causes and prevention. Roma

FAO, 1995. Quality and quality changes in fresh fish, Fisheries. ed. Food and agriculture organization of the United Nations, Rome.

Fiskeribladet, 2020. Norcod har satt sin første torskeyngel i sjøen - Tekfisk [WWW Document]. URL https://fiskeribladet.no/tekfisk/nyheter/?artikkel=71122 (accessed 3.24.20).

Fiskeribladet, 2018. I fire år har de forsket på levendelagring av torsk. Dette har de funnet ut. - Tekfisk [WWW Document]. URL https://fiskeribladet.no/tekfisk/nyheter/?artikkel=63092 (accessed 10.4.19).

Fitz-Gerald, C.H., Bremner, A., 1998. The oxidative stability of chilled and frozen pilchards used as feed for captive southern bluefin tuna. J. Aquat. Food Prod. Technol. 7, 27–44. https://doi.org/10.1300/J030v07n01

Fletcher, G.L., Wroblewski, J.S., Hickey, M.M., Blanchard, B., Kao, M.H., Goddard, S. V., 1997. Freezing resistance of caged atlantic cod (Gadus morhua) during a Newfoundland winter. Can. J. Fish. Aquat. Sci. 54, 94–98. https://doi.org/10.1139/f96-162

FSSP, 2014. Food Spoilage and Safety Predictor. National Food Institute (DTU Food). http://fssp.food.dtu.dk.

Furevik, D.M., Løkkeborg, S., 1994. Fishing trials in Norway for torsk (Brosme brosme) and cod (Gadus morhua) using baited commercial pots. Fish. Res. 19, 219–229. https://doi.org/10.1016/0165- 7836(94)90040-X

Gildberg, A., Mikkelsen, H., Sandaker, E., Ringø, E., 1997. Probiotic effect of lactic acid bacteria in the feed on growth and survival of fry of Atlantic cod (Gadus morhua). Hydrobiologia 352, 279–285. https://doi.org/10.1007/978-94-011-5234-1_27

Göbel, C., Langen, N., Blumenthal, A., Teitscheid, P., Ritter, G., 2015. Cutting food waste through cooperation along the food supply chain. Sustain. 7, 1429–1445. https://doi.org/10.3390/su7021429

Godiksen, H., Hyldig, G., Jessen, F., 2003. Sarcoplasmic reticulum Ca2+-ATPase and cytochrome oxidase as indicators of frozen storage in cod (Gadus morhua). J. Food Sci. 68, 2579–2585. https://doi.org/10.1111/j.1365-2621.2003.tb07064.x

Gotceitas, V., Methven, D.A., Fraser, S., Brown, J.A., 1999. Effects of body size and food ration on over- winter survival and growth of age-0 Atlantic cod, Gadus morhua. Environ. Biol. Fishes 54, 413–420. https://doi.org/10.1023/A:1007587101669

Gram, L., Huss, H. H., 1996. Microbiological spoilage of fish and fish products. Int. J. Food Microbiol. 33, 121–37.

Greenland Institute of natural Resources, 2015. Den biologiske rådgivning for 2016 fra ICES og NAFO.

Greenland Institute of natural Resources, 2012. Den biologiske rådgivning for fiskebestande for 2013 fra ICES.

Grolleaud, M., 2002. Post-harvest losses: discovering the full story, in: Overview of the Phenomenon of Losses during the Post-Harvest System. FAO, Rome.

Guldager, H.S., Bøknæs, N., Østerberg, C., Nielsen, J., Dalgaard, P., 1998. Thawed cod fillets spoil less rapidly than unfrozen fillets when stored under modified atmosphere at 2°C. J. Food Prot. 61, 1129–1136. https://doi.org/10.4315/0362-028X-61.9.1129

Gustavsson, J., Cederberg, C., Sonesson, U., Emanuelsson, A., 2013. The methodology of the FAO study : “ Global Food Losses and Food Waste - extent , causes and prevention ” - FAO , 2011, SIK report No. 857.

Hamm, R., 1977. Postmortem breakdown of ATP and glycogen in ground muscle: A review. Meat Sci. 15– 39.

Hansen, A.Å., Moen, B., Rødbotten, M., Berget, I., Pettersen, M.K., 2016. Effect of vacuum or modified

atmosphere packaging (MAP) in combination with a CO2 emitter on quality parameters of cod loins (Gadus morhua). Food Packag. Shelf Life 9, 29–37. https://doi.org/10.1016/j.fpsl.2016.05.005

Hansen, A.Å., Mørkøre, T., Rudi, K., Olsen, E., Eie, T., 2007. Quality changes during refrigerated storage

of MA-packaged pre-rigor fillets of farmed atlantic cod (Gadus morhua L.) using traditional MAP,

CO2 emitter, and vacuum. J. Food Sci. 72, 423–430. https://doi.org/10.1111/j.1750- 3841.2007.00561.x

Hansen, A.Å., Rødbotten, M., Lea, P., Rotabakk, B.T., Birkeland, S., Pettersen, M.K., 2015. Effect of transport packaging and repackaging into modified atmosphere on shelf life and quality of thawed Atlantic cod loins. Packag. Technol. Sci. 28, 925–938. https://doi.org/10.1002/pts.2139

Hansen, G.H., Olafsen, J.A., 1989. Bacterial colonization of cod (Gadus morhua L.) and halibut (Hippoglossus hippoglossus) eggs in marine aquaculture. Appl. Environ. Microbiol. 55, 1435–46.

Hedeholm, R.B., Jacobsen, R.B., Nielsen, E.E., 2016. Learning from “apparent consensus” in TAC disputes: Exploring knowledge overlaps in LEK and genetic categorization of Atlantic cod. Mar. Policy 69, 114–120. https://doi.org/10.1016/j.marpol.2016.04.020

Hemre, G.I., Taranger, G.L., Hansen, T., 2002. Gonadal development influences nutrient utilisation in cod (Gadus morhua). Aquaculture 214, 201–209. https://doi.org/10.1016/S0044-8486(02)00359-9

Henriksen, E., Heide, M., Hansen, Ø.J., Mortensen, A., 2018. Kunnskaps- og erfaringsgrunnlag for torskeoppdrett.

Hermansen, Ø., Eide, A., 2013. Bioeconomics of capture-based aquaculture of cod (Gadus morhua). Aquac. Econ. Manag. 17, 31–50. https://doi.org/10.1080/13657305.2013.747225

Himelbloom, B.H., Crapo, C.A., Brown, E.K., Doyle, J.P., 1997. Factors affecting quality of rock sole (Lepidopsetta bilineata) fillets. J. Aquat. Food Prod. Technol. 3, 45–56. https://doi.org/10.1081/E- EEE2-120046011

Howgate, P.F., 2010a. A critical review of total volatile bases and trimethylamine as indices of freshness of fish. part 1. determination. J. Environ. Agric. Food Chem. 9, 29–57.

Howgate, P.F., 2010b. A critical review of total volatile bases and trimethylamine as indices of freshness of fish. part 2. formation of the bases, and application in quality assurance. Electron. J. Environ. Agric. Food Chem. 9, 58–88.

Humborstad, O.B., Breen, M., Davis, M.W., Løkkeborg, S., Mangor-Jensen, A., Midling, K.T., Olsen, R.E., 2016. Survival and recovery of longline- and pot-caught cod (Gadus morhua) for use in capture- based aquaculture (CBA). Fish. Res. 174, 103–108. https://doi.org/10.1016/j.fishres.2015.09.001

Humborstad, O.B., Mangor-Jensen, A., 2013. Buoyancy adjustment after swimbladder puncture in cod Gadus morhua: An experimental study on the effect of rapid decompression in capture-based aquaculture. Mar. Biol. Res. 9, 383–393. https://doi.org/10.1080/17451000.2012.742546

Huss, H. H., 1971. Prepacked fresh fish, in: Kreuzer, R. (Ed.) Fish inspection and quality control. Fishing News Ltd, FAO, London.

Hyldig, G., Bremner, A., Martinsdóttir, E., Schelvis, R., 2012. Quality Index Mehods, in: Nollet, L.M.L. (Ed.), Handbook of Meat, Poultry and Seafood Quality. Wiley-Blackwell, Oxford, pp. 437–458.

ICES, 2019a. Cod (Gadus morhua) in subareas 1 and 2 (Northeast Arctic). Rep. ICES Advis. Comm. 27, 1– 8. https://doi.org/10.17895/ices.advice.4710

ICES, 2019b. Cod (Gadus morhua) in NAFO Subarea 1, inshore (West Greenland cod). Rep. ICES Advis. Comm. 21, 1–8.

ICES, 2019c. Cod (Gadus morhua) in Division 5.a (Iceland grounds). Rep. ICES Advis. Comm. 27, 1–8. https://doi.org/10.17895/ices.pub.3102

ICMSF, 2011. Microorganisms in Foods 8, International Commission on Microbiological Specifications for Foods (ICMSF). Springer, New York. https://doi.org/10.1007/978-1-4419-9374-8_19

Jespersen, C.M., Heldbo, J., 1991. Sensorisk kvalitetsindeks til torsk og sild, in: Kvalitet Og Fiskeforædling. Report to the Nordic Council of Ministres, Copenhagen.

Jørgensen, L.V., Dalgaard, P., Huss, H.H., 2000. Multiple compound quality index for cold-smoked salmon (Salmo salar) developed by multivariate regression of biogenic amines and pH. J. Agric. Food Chem. 48, 2448–2453. https://doi.org/10.1021/jf9909407

Jørgensen, T., Midling, K., Espelid, S., Nilsen, R., Stensvåg, K., 1989. Vibrio salmonicida, a pathogen in salmonids, also causes mortality in net-pen captured cod (Gadus morhua). Bull. Eur. Assoc. Fish Pathol. 9, 42–44.

Jørstad, K.E., Karlsen, Svåsand, T., Otterå, H., 2006. Comparison of growth rate among different protein genotypes in Atlantic cod, Gadus morhua, under farmed conditions. ICES J. Mar. Sci. 63, 235–245. https://doi.org/10.1016/j.icesjms.2005.11.005

Karube, I., Matsuoka, H., Suzuki, S., Watanabe, E., Toyama, K., 1984. Determination of Fish Freshness with an Enzyme Sensor System. J. Agric. Food Chem. 32, 314–319.

https://doi.org/10.1021/jf00122a034

Kelleher, K., 2005. Discards in the World’s Marine Fisheries. An Update. FAO Fish. Tech. Pap. 131.

Khan, R.A., 2004. Disease outbreaks and mass mortality in cultured Atlantic cod, Gadus morhua L., associated with Trichodina murmanica (Ciliophora). J. Fish Dis. 27, 181–184. https://doi.org/10.1111/j.1365-2761.2004.00525.x

Kolbeinshavn, A.G., Vestergaard, P., Patursson, Ø., Gislason, H., 2012. Rapid growth of farmed cod in sea cages in the Faroe Islands. Aquaculture 358–359, 224–233. https://doi.org/10.1016/j.aquaculture.2012.06.036

Kurlansky, M., 1999. Cod - A biography of the fish that changed the world. Vintage, London.

Kuuliala, L., Al Hage, Y., Ioannidis, A.G., Sader, M., Kerckhof, F.M., Vanderroost, M., et al, 2018. Microbiological, chemical and sensory spoilage analysis of raw Atlantic cod (Gadus morhua) stored under modified atmospheres. Food Microbiol. 70, 232–244. https://doi.org/10.1016/j.fm.2017.10.011

Labansen, V., 2018. Indhandlede mængder af levende torsk 2017 /2018 [WWW Document]. URL https://gembaseafood.dk/wp-content/uploads/2018/09/13-Vivi-Labasen-Royal-Greenland.pdf (accessed 10.4.19).

Lambert, Y., Dutil, J.D., Munro, J., 1994. Effects of intermediate and low salinity conditions on growth rate and food conversion of Atlantic cod (Gadus morhua). Can. J. Fish. Aquat. Sci. 51, 1569–1576. https://doi.org/10.1139/f94-155

Laursen, B.G., Bay, L., Cleenwerck, I., Vancanneyt, M., Swings, J., Dalgaard, P., Leisner, J.J., 2005. Carnobacterium divergens and Carnobacterium maltaromaticum as spoilers or protective cultures in meat and seafood: Phenotypic and genotypic characterization. Syst. Appl. Microbiol. 28, 151– 164. https://doi.org/10.1016/j.syapm.2004.12.001

Lauzon, H.L., Magnússon, H., Sveinsdóttir, K., Gudjónsdóttir, M., Martinsdóttir, E., 2009. Effect of brining, modified atmosphere packaging, and superchilling on the shelf life of cod (Gadus morhua) loins. J. Food Sci. 74. https://doi.org/10.1111/j.1750-3841.2009.01200.x

LeBlanc, E.L., LeBlanc, R.J., 1992. Determination of hydrophobicity and reactive groups in proteins of cod (Gadus morhua) muscle during frozen storage. Food Chem. 43, 3–11.

https://doi.org/10.1016/0308-8146(92)90234-S

Lee, P., Osborn, S., Whitehead, P., 2015. Reducing food waste by extending product life [WWW Document]. URL http://www.wrap.org.uk/sites/files/wrap/Product Life Report Final_0.pdf (accessed 2.17.20).

Leroi, F., Joffraud, J.J., Chevalier, F., Cardinal, M., 2001. Research of quality indices for cold-smoked salmon using a stepwise multiple regression of microbiological counts and physico-chemical parameters. J. Appl. Microbiol. 90, 578–587. https://doi.org/10.1046/j.1365-2672.2001.01283.x

Lindberg, S., 2019. Restitution of trawl-caught fish and its effect on blood concentration in fillets, in: WEFTA 2019, Thorshavn.

Love, D.C., Fry, J.P., Milli, M.C., Neff, R.A., 2015. Wasted seafood in the United States: Quantifying loss from production to consumption and moving toward solutions. Glob. Environ. Chang. 35, 116–124. https://doi.org/10.1016/j.gloenvcha.2015.08.013

Love, R.M., 1979. The post‐mortem pH of cod and haddock muscle and its seasonal variation. J. Sci. Food Agric. 30, 433–438. https://doi.org/10.1002/jsfa.2740300414

Magnússon, H., Martinsdóttir, E., 1995. Storage quality of fresh and frozen‐thawed fish in ice. J. Food Sci. 60, 273–278. https://doi.org/10.1111/j.1365-2621.1995.tb05654.x

Martinsdottir, E., Magnusson, H., 2001. Keeping quality of sea-frozen thawed cod fillets on ice. J. Food Sci. 66, 1402–1408. https://doi.org/10.1111/j.1365-2621.2001.tb15222.x

McGill, A.S., Hardy, R., Burt, J.R., Gunstone, F.D., 1974. Hept‐cis‐4‐enal and its contribution to the off‐ flavour in cold stored cod. J. Sci. Food Agric. 25, 1477–1489. https://doi.org/10.1002/jsfa.2740251208

McIntosh, D., Ji, B., Forward, B.S., Puvanendran, V., Boyce, D., Ritchie, R., 2008. Culture-independent characterization of the bacterial populations associated with cod (Gadus morhua L.) and live feed at an experimental hatchery facility using denaturing gradient gel electrophoresis. Aquaculture 275, 42–50. https://doi.org/10.1016/j.aquaculture.2007.12.021

Merritt, J.H., 1965. Superchilling on board trawlers. Bull. Int. Inst. Refrig. Aneex 1965 (5), 183-190.

Midling, K., Beltestad, A., Isaksen, B., 1996. Live fish technology - Historical convenience to modern multispecies strategy in Norway, in: Bremner, A., Davis, C., Austin, B. (Eds.), Making the Most of the

Catch... Proceedings from an International Post-Harvest Seafood Symposium. The national seafood centre, Brisbane, Queensland, Australia, pp. 63–67.

Midling, K.Ø., Koren, C., Humborstad, O.B., Sæther, B.S., 2012. Swimbladder healing in Atlantic cod (Gadus morhua), after decompression and rupture in capture-based aquaculture. Mar. Biol. Res. 8, 373–379. https://doi.org/10.1080/17451000.2011.638640

Mietz, J.L., Karmas, E., 1977. Chemical quality index of canned tuna as determined by high‐pressure liquid chromatography. J. Food Sci. 42, 155–158. https://doi.org/10.1111/j.1365- 2621.1977.tb01240.x

Miller, A., Scanlan, R.A., Lee, J.S., Libbey, L.M., Morgan, M.E., 1973. Identification of the volatile compounds produced in sterile fish muscle (Sebastes melanops) by Pseudomonas fragi. Appl. Microbiol. 25, 257–261. https://doi.org/10.1128/aem.25.2.257-261.1973

Moe, H., Gaarder, R.H., Olsen, A., Hopperstad, O.S., 2009. Resistance of aquaculture net cage materials to biting by Atlantic Cod (Gadus morhua). Aquac. Eng. 40, 126–134. https://doi.org/10.1016/j.aquaeng.2009.02.001

Moen, T., Hayes, B., Nilsen, F., Delghandi, M., Fjalestad, K.T., Fevolden, S.E., Berg, P.R., Lien, S., 2008. Identification and characterisation of novel SNP markers in Atlantic cod: Evidence for directional selection. BMC Genet. 9, 1–9. https://doi.org/10.1186/1471-2156-9-18

Morais, S., Bell, J.G., Robertson, D.A., Roy, W.J., Morris, P.C., 2001. Protein/lipid ratios in extruded diets for Atlantic cod (Gadus morhua L.): Effects on growth, feed utilisation, muscle composition and liver histology. Aquaculture 203, 101–119. https://doi.org/10.1016/S0044-8486(01)00618-4

Mørkøre, T., Hansen, S.J., Rørvik, K., 2006. Relevance of storage temperature for contraction and gaping of pre rigor filleted farmed cod (Gadus morhua L.), in: Luten, J.B., Jacobsen, C., Bekaert, K., Sæbø, A., Oehlenschläger, J. (Eds.), Seafood Research from Fish to Dish. Wageningen Academic, Wageningen.

MSC, 2020. The MSC Fisheries Standard | Marine Stewardship Council [WWW Document]. URL https://www.msc.org/standards-and-certification/fisheries-standard (accessed 3.24.20).

Murray, J.M., Delahunty, C.M., Baxter, I.A., 2001. Descriptive sensory analysis: Past, present and future. Food Res. Int. 34, 461–471. https://doi.org/10.1016/S0963-9969(01)00070-9

Muth, M.K., Karns, S.A., Nielsen, S.J., Buzby, J.C., Wells, H.F., 2011. Consumer-Level food loss estimates and their use in the ERS loss-adjusted food availability data. Tech. Bull. 1927, 123.

Nikolik, G., Heinhuis, L., 2015. Rabobank World Seafood Trade Map 2015, Rabobank Industry Note.

Nissling, A., Westin, L., 1997. Salinity requirements for successful spawning of Baltic and Belt Sea cod and the potential for cod stock interactions in the Baltic Sea. Mar. Ecol. Prog. Ser. 152, 261–271. https://doi.org/10.3354/meps152261

NMKL, 2006. NÆRINGSMIDLER No . 184 Aerobic count and specific spoilage organisms in fish and fish products. Nord. Comm. food Anal. 2–7.

Oehlenschläger, J., Sørensen, N.-K., 1997. Criteria of seafish freshness and quality aspects, in: Olafsdottir, G., Luten, J., Dalgaard, P., Careche, M., Verrez-Bagnis, V., Martinsdóttir, E., Heia, K. (Eds.), Methods to Determine the Freshness of Fish in Research and Industry. International institute of refrigeration, Paris, pp. 30–35.

Olafsdóttir, G., Luten, J., Dalgaard, P., Careche, M., Verrez-Bagnis, V., Martinsdóttir, E., Heia, K., 1997. Methods to Determine the Freshness of Fish in Research and Industry. International institute of refrigeration, Paris.

Olafsdottir, G., Jonsdottir, R., Lauzon, H.L., Luten, J., Kristbergsson, K., 2005. Characterization of volatile compounds in chilled cod (Gadus morhua) fillets by gas chromatography and detection of quality indicators by an electronic nose. J. Agric. Food Chem. 53, 10140–10147. https://doi.org/10.1021/jf0517804

Olafsdottir, G., Lauzon, H.L., Martinsdottir, E., Kristbergsson, K., 2006a. Influence of storage temperature on microbial spoilage characteristics of haddock fillets (Melanogrammus aeglefinus) evaluated by multivariate quality prediction. Int. J. Food Microbiol. 111, 112–125. https://doi.org/10.1016/j.ijfoodmicro.2006.04.045

Olafsdottir, G., Lauzon, H.L., Martinsdóttir, E., Oehlenschläger, J., Kristbergsson, K., 2006b. Evaluation of shelf life of superchilled cod (Gadus morhua) fillets and the influence of temperature fluctuations during storage on microbial and chemical quality indicators. J. Food Sci. 71.

Oliveira, H., Pedro, S., Nunes, M.L., Costa, R., Vaz-Pires, P., 2012. Processing of salted cod (Gadus spp.): A Review. Compr. Rev. Food Sci. Food Saf. 11, 546–564. https://doi.org/10.1111/j.1541-

4337.2012.00202.x

Olsen, A.B., Mikalsen, J., Rode, M., Alfjorden, A., Hoel, E., Straum-Lie, K., Haldorsen, R., Colquhoun, D.J., 2006. A novel systemic granulomatous inflammatory disease in farmed Atlantic cod, Gadus morhua L., associated with a bacterium belonging to the genus Francisella. J. Fish Dis. 29, 307–311. https://doi.org/10.1111/j.1365-2761.2006.00714.x

Olsen, K.B., Whittle, K., Strachen, N., Veenstra, F.A., Storbeck, F., van Leeuwen, P., 1993. Integrated quality assurance of chilled food fish at sea. Kgs. Lyngby.

Olsen, R.E., Hansen, A.C., Rosenlund, G., Hemre, G.I., Mayhew, T.M., Knudsen, D.L., Tufan Eroldoǧan, O., Myklebust, R., Karlsen, Ø., 2007. Total replacement of fish meal with plant proteins in diets for Atlantic cod (Gadus morhua L.) II - Health aspects. Aquaculture 272, 612–624. https://doi.org/10.1016/j.aquaculture.2007.05.010

Olsen, S.H., Tobiassen, T., Akse, L., Evensen, T.H., Midling, K.T., 2013. Capture induced stress and live storage of Atlantic cod (Gadus morhua) caught by trawl: Consequences for the flesh quality. Fish. Res. 147, 446–453. https://doi.org/10.1016/j.fishres.2013.03.009

Olsson, G.B., Gundersen, B., Esaiassen, M., 2006. Pre-slaughter starvation of farmed Atlantic cod fed vegatable proteins: Effects on quality parameters, in: Luten, J.B., Jacobsen, C., Bekaert, K., Sæbø, A., Oehlenschläger, J. (Eds.), Seafood Research from Fish to Dish. Wageningen Academic, Wageningen, pp. 139–148.

Olsson, G.B., Seppola, M.A., Olsen, R.L., 2007. Water-holding capacity of wild and farmed cod (Gadus morhua) and haddock (Melanogrammus aeglefinus) muscle during ice storage. LWT - Food Sci. Technol. 40, 793–799. https://doi.org/10.1016/j.lwt.2006.04.004

Paludan-Müller, C., Dalgaard, P., Huss, H.H., Gram, L., 1998. Evaluation of the role of Carnobacterium piscicola in spoilage of vacuum- and modified-atmosphere-packed cold-smoked salmon stored at 5°C. Int. J. Food Microbiol. 39, 155–166. https://doi.org/10.1016/S0168-1605(97)00133-5

Parfitt, J., Barthel, M., MacNaughton, S., 2010. Food waste within food supply chains: Quantification and potential for change to 2050. Philos. Trans. R. Soc. B Biol. Sci. 365, 3065–3081. https://doi.org/10.1098/rstb.2010.0126

Pérez-Casanova, J.C., Lall, S.P., Gamperl, A.K., 2009. Effect of feed composition and temperature on food

consumption, growth and gastric evacuation of juvenile Atlantic cod (Gadus morhua L.) and haddock (Melanogrammus aeglefinus L.). Aquaculture 294, 228–235. https://doi.org/10.1016/j.aquaculture.2009.06.005

Poirier, S., Rué, O., Peguilhan, R., Coeuret, G., Zagorec, M., Champomier-Vergès, M.C., Loux, V., Chaillou, S., 2018. Deciphering intra-species bacterial diversity of meat and seafood spoilage microbiota using gyrB amplicon sequencing : A comparative analysis with 16S rDNA V3-V4 amplicon sequencing. PLoS One 13(9), e0204629. https://doi.org/10.1371/journal.pone.0204629

Quéméner, L., Suquet, M., Mero, D., Gaignon, J.L., 2002. Selection method of new candidates for finfish aquaculture: The case of the French Atlantic, the Channel and the North Sea coasts. Aquat. Living Resour. 15, 293–302. https://doi.org/10.1016/S0990-7440(02)01187-7

Ratkowsky, D.A., Olley, J., McMeekin, T.A., Ball, A., 1982. Relationship between temperature and growth rate of bacterial cultures. J. Bacteriol. 149, 1–5.

Rillahan, C., Chambers, M.D., Howell, W.H., Watson, W.H., 2011. The behavior of cod (Gadus morhua) in an offshore aquaculture net pen. Aquaculture 310, 361–368. https://doi.org/10.1016/j.aquaculture.2010.10.038

Rodriguez-Martinez, V., Velazquez, G., Massa-Barrera, S., Welti-Chanes, J., Fagotti, F., Torres, J.A., 2019. Estimation of safety and quality losses of foods stored in residential refrigerators. Food Eng. Rev. 11, 184–199. https://doi.org/10.1007/s12393-019-09192-1

Roiha, I.S., Jónsson, Á., Backi, C.J., Lunestad, B.T., Karlsdóttir, M.G., 2017. A comparative study of quality and safety of Atlantic cod (Gadus morhua) fillets during cold storage, as affected by different thawing methods of pre-rigor frozen headed and gutted fish. J. Sci. Food Agric. 98, 400–409. https://doi.org/10.1002/jsfa.8649

Roiha, I.S., Tveit, G.M., Backi, C.J., Jónsson, Á., Karlsdóttir, M., Lunestad, B.T., 2018. Effects of controlled thawing media temperatures on quality and safety of pre-rigor frozen Atlantic cod (Gadus morhua). LWT - Food Sci. Technol. 90, 138–144. https://doi.org/10.1016/j.lwt.2017.12.030

Rosenlund, G., Halldórsson, Ó., 2007. Cod juvenile production: Research and commercial developments. Aquaculture 268, 188–194. https://doi.org/10.1016/j.aquaculture.2007.04.040

Rotabakk, B.T., Skipnes, D., Akse, L., Birkeland, S., 2011. Quality assessment of Atlantic cod (Gadus

morhua) caught by longlining and trawling at the same time and location. Fish. Res. 112, 44–51. https://doi.org/10.1016/j.fishres.2011.08.009

Royal Greenland A/S, 2018. Årsrapport Royal Greenland A/S 1. januar 2018 - 31. december 2018. Nuuk.

Royal Greenland A/S, 2013. Årsrapport Royal Greenland A/S 2012/2013. Nuuk.

Sánchez-Alonso, I., Martinez, I., Sánchez-Valencia, J., Careche, M., 2012. Estimation of freezing storage time and quality changes in hake (Merluccius merluccius, L.) by low field NMR. Food Chem. 135, 1626–1634. https://doi.org/10.1016/j.foodchem.2012.06.038

Savagaon, K.A., Power, H.E., 1976. Effect of box stowage on quality of fish. fish. technol. 13, 4–8.

Schubring, R., 2005. Changes in texture, water holding capacity, colour and thermal stability of frozen cod (Gadus morhua) fillets: Effect of frozen storage temperature. Dtsch. Leb. 101, 484–493.

Schubring, R., 2004. Instrumental colour, texture, water holding and DSC measurements on frozen cod fillets (Gadus morhua) during long term storage at different temperatures. Dtsch. Leb. 100, 247– 254.

Schubring, R., 1999. Influence of twice-freezing on quality parameters of Pollack (Pollachius virens) fillets. Dtsch. Leb. 95, 161.

Shewan, J.M., 1965. Bacteriology of fish stored in chilled sea water, in: Fish Handling and Perservation. Paris, pp. 95–109.

Shewan, J.M., Macintosh, R.G., Tucker, C.G., Ehrenberg, A.S.C., 1953. The development of a numerical scoring system for the sensory assessment of the spoilage of wet white fish stored in ice. J. Sci. Food Agric. 4, 283–298. https://doi.org/10.1002/jsfa.2740040607

Siddaiah, D., Sagar Reddy, G.V., Raju, C. V., Chandrasekhar, T.C., 2001. Changes in lipids, proteins and kamaboko forming ability of silver carp (Hypophthalmichthys molitrix) mince during frozen storage. Food Res. Int. 34, 47–53. https://doi.org/10.1016/S0963-9969(00)00127-7

Sivertsvik, M., 2007. The optimized modified atmosphere for packaging of pre-rigor filleted farmed cod (Gadus morhua) is 63 ml/100 ml oxygen and 37 ml/100 ml carbon dioxide. LWT - Food Sci. Technol. 40, 430–438. https://doi.org/10.1016/j.lwt.2005.12.010

Spada, A., Conte, A., Del Nobile, M.A., 2018. The influence of shelf life on food waste: A model-based

approach by empirical market evidence. J. Clean. Prod. 172, 3410–3414. https://doi.org/10.1016/j.jclepro.2017.11.071

Statham, J.A., Bremmer, H.A., 1985. Acceptability of trevalla (Hyperoglyphe-porosa richardson) after storage in carbon-dioxide. Food Technol. Aust. 37, 212–215.

Statham, J.A., Bremmer, H.A., 1983. Effect of potassium sorbate on spoilage of blue grenadier (Macruronus novaezelandiae) as assessed by microbiology and sensory profiles. J. Food Prot. 46, 1084–1091. https://doi.org/10.4315/0362-028x-46.12.1084

Statistics Greenland, 2020a. Befolkningen i lokaliteterne 1. januar efter fødested, lokalitet og tid [WWW Document]. Befolk. i Lokal. pr 1. Jan. 1977-2020 [BEDST4]. URL http://bank.stat.gl/pxweb/da/Greenland/Greenland__BE__BE01__BE0120/BEXST4.PX/table/table ViewLayout1/?rxid=BEXST418-02-2020 12:31:43 (accessed 2.18.20).

Statistics Greenland, 2020b. Gennemsnitlige personindkomster efter lokalitet, tid og indkomsttype [WWW Document]. Gennemsnitlige Pers. fordelt på Lokal. [INDP3]. URL http://bank.stat.gl/pxweb/da/Greenland/Greenland__IN__IN30/INXP3.px/table/tableViewLayout1 /?rxid=INXP318-02-2020 12:38:16

Statistics Greenland, 2020c. Gennemsnitlige personindkomster efter tid, kommune og indkomsttype [WWW Document]. Indk. OG FORBRUG. URL http://www.stat.gl/dialog/topmain.asp?lang=da&subject=Indkomster&sc=IN (accessed 2.18.20).

Statistics Greenland, 2020d. Antal ledige i gennemsnit pr. måned blandt fastboende 18-65-årige efter distrikt, tid og kvartal [WWW Document]. URL http://bank.stat.gl/pxweb/da/Greenland/Greenland__AR__AR40/ARXLED3.px/table/tableViewLay out1/?rxid=INXP318-02-2020 12:38:16 (accessed 2.18.20).

Statistics Greenland, 2019. Total landings of fish and shellfish by time, municipality, species, unit and month [FIE001] [WWW Document]. FIX001. URL http://bank.stat.gl/pxweb/da/Greenland/Greenland__FI__FI10/FIX001.px/chart/chartViewLine/?rx id=FIX00109-08-2019 07:30:24 (accessed 8.9.19).

Storr-Paulsen, M., Wieland, K., Hovgård, H., Rätz, H.J., 2004. Stock structure of Atlantic cod (Gadus morhua) in West Greenland waters: Implications of transport and migration. ICES J. Mar. Sci. 61, 972–982. https://doi.org/10.1016/j.icesjms.2004.07.021

Sveinsdóttir, K., Martinsdóttir, E., Hyldig, G., Sigurgísladóttir, S., 2010. Sensory characteristics of different cod products. J. Sens. Stud. 25, 294–314. https://doi.org/10.1111/j.1745-459X.2009.00259.x

Tarr, H.L.A., 1954. Microbiological deterioration of fish post mortem, its detection and control. Bacteriol. Rev. 18, 1–15.

The Consumer Goods Forum, 2015. Consumer goods industry commits to food waste reduction [WWW Document]. URL https://www.theconsumergoodsforum.com/press_releases/consumer-goods- industry-commits-to-food-waste-reduction/ (accessed 2.17.20).

Thrane, M., Ziegler, F., Sonesson, U., 2009. Eco-labelling of wild-caught seafood products. J. Clean. Prod. 17, 416–423. https://doi.org/10.1016/j.jclepro.2008.08.007

Torry Research Station, 2001. Cold storage of frozen fish [WWW Document]. Torry Advis. note No. 28. URL http://www.fao.org/3/x5907e/x5907e00.htm (accessed 1.26.20).

Tybjerg, J., 2018. Denne torsk har været den halve vej rundt om jorden og tilbage. Sådan blev den billigst for alle (bortset fra kloden) [WWW Document]. Zetland. URL https://www.zetland.dk/historie/s8YxXnJR-a85E2rzd-7752a (accessed 2.18.20).

United Nations, 2020. SDGs: Sustainable Development Knowledge Platform [WWW Document]. URL https://sustainabledevelopment.un.org/sdgs (accessed 3.22.20).

United Nations, 2019. World population prospects 2019: Highlights, Department of Economic and Social Affairs.

Van Der Meeren, T., Jørstad, K.E., 2001. Growth and survival of Arcto-Norwegian and Norwegian coastal cod larvae (Gadus morhua L.) reared together in mesocosms under different light regimes. Aquac. Res. 32, 549–563. https://doi.org/10.1046/j.1365-2109.2001.00578.x

Venugopal, V., Shahidi, F., 1998. Traditional methods to process underutilized fish species for human consumption. Food Rev. Int. 14, 35–97. https://doi.org/10.1080/87559129809541149

Vyncke, W., 1983. Shelf life of thawed cod fillets kept in ice. Z. Lebensm. Unters. Forsch. 177, 19–21. https://doi.org/10.1007/BF01042489

Wang, T., Sveinsdóttir, K., Magnússon, H., Martinsdóttir, E., 2008. Combined application of modified atmosphere packaging and superchilled storage to extend the shelf life of fresh cod (Gadus morhua) loins. J. Food Sci. 73. https://doi.org/10.1111/j.1750-3841.2007.00590.x

Woyewoda, A.D., Bligh, E.G., Shaw, S.J., 1984. Controlled and modified atmosphere storage of cod fillets. Can. Inst. Food Sci. Technol. J. 17, 24–27. https://doi.org/10.1016/s0315-5463(84)72311-x

Xie, J., Zhang, Z., Yang, S.P., Cheng, Y., Qian, Y.F., 2018. Study on the spoilage potential of Pseudomonas fluorescens on salmon stored at different temperatures. J. Food Sci. Technol. 55, 217–225. https://doi.org/10.1007/s13197-017-2916-x

Ziegler, F., Winther, U., Hognes, E.S., Emanuelsson, A., Sund, V., Ellingsen, H., 2013. The Carbon Footprint of Norwegian Seafood Products on the Global Seafood Market. J. Ind. Ecol. 17, 103–116. https://doi.org/10.1111/j.1530-9290.2012.00485.x

Øiestad, V., Kvenseth, P.G., Folkvord, A., 1985. Mass production of Atlantic cod juveniles Gadus morhua in a Norwegian saltwater pond. Transsactions Am. Fish. Soc. 114, 590–595. https://doi.org/https://doi.org/10.1577/1548-8659(1985)114<590:MPOACJ>2.0.CO;2

Paper I

Jonas Steenholdt Sørensen, Niels Bøknæs, Ole Mejlholm & Paw Dalgaard

Superchilling in combination with modified atmosphere packaging resulted in long shelf-life and limited microbial growth in Atlantic cod (Gadus morhua L.) from capture-based-aquaculture in Greenland

Food Microbiology, Volume 88, June 2020,

103405, https://doi.org/10.1016/j.fm.2019.103405.

Superchilling in combination with modified atmosphere packaging resulted in long shelf-life and limited microbial growth in Atlantic cod (Gadus morhua L.) from capture-based-aquaculture in Greenland

Jonas Steenholdt Sørensen1,2*, Niels Bøknæs2, Ole Mejlholm2, Paw Dalgaard1

1National Food Institute (DTU Food), Technical University of Denmark, Kgs. Lyngby, Denmark

2Royal Greenland A/S, Svenstrup J, Denmark

* Corresponding author: Food Microbiology and Hygiene, National Food Institute, Technical University of Denmark, Kemitorvet, Building 202, 2800, Kgs. Lyngby, Denmark. E-mail: [email protected]

Abstract

Sensory, chemical and microbial changes for Atlantic cod (Gadus morhua L.) filets from capture-based- aquaculture in Greenland were studied. The objective was to determine shelf-life and indices of spoilage for iced or superchilled fillets when stored in air, or modified atmosphere packed (MAP; 40% CO2 and 60%

N2). MAP iced storage extended the sensory shelf-life from 15 days to 21 days compared to storage in air. With superchilling at -1.7 ○C sensory shelf-life was above 32 days, and no formation of total volatile nitrogen (TVN) was observed irrespective of storage in air or MAP. pH of ≥7.0, TVN (≥35 mg-N/100g) and trimethylamine (≥20 mg-N TMA/100g) were promising indices of spoilage. Aerobic viable counts were less valuable indices of spoilage as the dominating microbiota of cod in air (Pseudomonas spp., Photobacterium spp., Shewanella spp., Acinetobacter spp.) changed to Photobacterium spp. in MAP cod. Spoilage activity determined as the yield factor for TVN formation was 6-200 folds higher for Photobacterium spp. compared to Shewanella spp. and Pseudomonas spp. Photobacterium carnosum was responsible for TVN formation in iced cod irrespective of storage in air or MAP, and it was identified at the specific spoilage organism that limited iced product shelf-life.

Keywords: Storage trial, indices of spoilage, total volatile nitrogen (TVN), 16S rRNA gene amplicon sequencing, specific spoilage organisms (SSO), spoilage activity.

1. Introduction

Fisheries make-up more than 90 % of the total export value from Greenland and Atlantic cod (Gadus morhua L.) has been an important species (Statistics Greenland, 2018). After a collapse of stocks for Atlantic cod in the 1990s (Buch et al., 1994; Storr-Paulsen and Wieland, 2006) this fishery is now regaining volume, primarily with in-shore catches by smaller fishing vessels (Statistics Greenland, 2019). Traditionally whitefish, including cod, from Greenland were shipped frozen to the primary export market in Europe. The frozen distribution was appropriate due to variable fish landings and typical transit time by ship of approximated two weeks.

Capture-based aquaculture (CBA) was newly introduced as a fishery and processing technic for cod in Greenland, with full scale production starting in 2015. The cod was caught by pound nets near the costal line and kept alive up to about two weeks in net enclosure next to the fishing ground and without additional feeding. Live cod, where then transported by well-boat to a processing plant where the fish was kept in large nets for two-four days. Cod were electrically stunned, slaughtered by machine decapitation, viscera removed, bleed and fileted within two hours from the time fishes are pumped from nets to the processing line. The combination of CBA and fast processing gives new options for distribution, which include shipment of non-frozen cod with reduced energy consumption and carbon footprint compared to frozen shipment. One option was superchilling at about -2 °C with the potential benefit that cod do not need to be covered with iced and therefore more fish can be transported per volume of ship hull. This reduce the carbon footprint for transport (Claussen et al., 2011; Hoang et al., 2016) and provide a potential to meet market demands for non-frozen fish (Altintzoglou et al., 2012). However, for non- frozen cod from Greenland to be transported and marketed in Europe a shelf-life above 19 days is needed with about 10 days for superchilled transport by ship, seven days for ground distribution including display in supermarkets at +2 °C and one to two day of consumer’s storage at +5 °C (James and James, 2014; Koutsoumanis and Gougouli, 2015).

Shelf-life and quality changes of Atlantic cod from Europe and the North-East Atlantic have been extensively studied. During chilled storage, quality changes were characterised by a loss of freshness due to autolytic reactions followed by sensory spoilage resulting from microbial activity including the reduction of trimethylamine-oxide (TMAO) to trimethylamine (TMA) by specific spoilage organisms (SSO) that grow to high concentrations. For iced cod the sensory shelf-life of aerobically stored and MAP fillets were typical of 10-14 days and 14-26 days, respectively, with indices of spoilage including total volatile nitrogen (TVN), TMA and concentrations of SSO (Dalgaard, 2000; DeWitt and Oliveira, 2016; Sivertsvik et

74 al., 2002). For aerobic storage in ice, H2S-producing Shewanella and Photobacterium phosphoreum have been identified as the SSO responsible for spoilage and TMA formation. P. phosphoreum with high resistance to CO2 and pronounced TMA formation was the SSO in chilled MAP cod. Growth of H2S- producing Shewanella was markedly reduced by CO2, and therefore they were not responsible for spoilage and TMA formation in iced MAP cod (Dalgaard, 2006; Hovda et al., 2007; Kuuliala et al., 2018; Olafsdottir et al., 2005; Reynisson et al., 2009). The microbiota of live Atlantic cod differs for fish from the Baltic sea, the Norths sea and the North-East Atlantic ocean (Wilson et al., 2008). In Greenland, the low seawater temperature of less than 1 °C to 5 °C (Buch, 2002) may select for a microbiota that differs from those of cod caught in warmer waters, but we have found no previous studies of this or of the potential effect on product spoilage and shelf-life.

Compared to iced storage, superchilling of Atlantic cod has been little studied. However, shelf-life of from 12 to >42 days has been observed for storage at -0.9 to -2.2 °C (Duun and Rustad, 2007; Eliasson et al., 2019; Lauzon et al., 2009; Olafsdottir et al., 2006; Wang et al., 2008). The combination of superchilling and MAP with 50% CO2, 5% O2 and 45% N2 resulted in shelf-life of 21 to >24 days for Atlantic cod from Iceland and similar results have been found for other fish species including wolfish and salmon from Norway (Lauzon et al., 2009; Rosnes et al., 2006; Sivertsvik et al., 2003; Wang et al., 2008). Thus, superchilling may provide sufficient shelf-life of Atlantic cod for non-frozen transport from Greenland to Europe.

The objective of the present study was to determine shelf-life and indices of spoilage of iced and superchilled Atlantic cod from CBA in Greenland and thereby to evaluate the feasibility of non- frozen transportation to Europe. Firstly, sensory, chemical and microbial changes were studied in a storage trial with aerobically or MAP stored cod. The spoilage microbiota was studied by culture- dependent techniques and by 16S rRNA gene amplicon sequencing. Secondly, to point out SSO and evaluate indices of spoilage the spoilage potential and the spoilage activity of isolates from the spoilage microbiota were determined.

2. Materials and methods

2.1 Storage trial with fresh Atlantic cod from capture-based aquaculture.

2.1.1 Fish raw material, packaging and storage conditions.

Atlantic cod (Gadus morhua L.) were captured inshore by pound net at NAFO fishing ground 1C during September 2018, kept alive in net enclosures next to the fishing ground, transported by well- boat to a fish factory in Maniitsoq, Greenland and kept alive in net enclosures until the time of processing. The cod was slaughtered by decapitation and rinsed before machine filleting. One hundred fourteen filets were taken directly from the production line and cut by hand into 342 pieces, each with a weight of approximated 100 grams. Between the cuttings of each filet, cutting boards and knives were rinsed with 96% ethanol to avoid cross-contamination of microorganism between fillets. Each piece of cod was packed individually in plastic bags of 70 µm thick polyethylene ﬁlm with high permeability of >6 g m-2 d-1 for water

3 -2 -1 -1 3 -2 -1 -1 vapour, >3,000 cm m d atm for O2 and >14,000 cm m d atm for CO2 (H902, Topiplast A/S, Greve, Denmark) and transported by aeroplane to DTU Food, Denmark in polystyrene boxes where the fish was cooled with gel ice packs (Sorbatek, Sorbafreeze, Glenrothes, UK). At DTU Food, the pieces of cod were randomly divided into four treatments, with 90 pieces for each treatment except treatment (i) with 63 pieces. In addition, nine pieces were used for sensory, chemical and microbial evaluation of fresh cod prior to initiation of storage (See 2.1.2-2.1.5). For the four treatments, each cod piece was placed in plastic trays (71-51A hvid/PS, Færch Plast, Holstebro, Denmark).

A storage trial with four treatments were carried out including (i) aerobic storage in ice; (ii) aerobic superchilled storage in slurry ice; (iii) MAP (40% CO2 and 60% N2) storage in ice and (iv) superchilled MAP (40% CO2 and 60% N2) storage in slurry ice. Trays with cod were packed using bags of a

3 -2 -1 -1 3 -2 -1 - 117 ± 6 µm laminate film with low gas permeability of 0.45 cm m d atm for O2 and 1.8 cm m d atm

1 for CO2 (NEN 40 HOB/LLPDE 75, Amcore, Horsens, Denmark). The gas to product ratio was about fifteen to one for both aerobic and MAP treatments with the large ratio selected to ensure relatively stable gas composition during storage. Bags for aerobic storage were sealed, without altering the composition of the air inside the bags, by using a Multivac C500 packaging machine (Multivac A/S, Vejle, Denmark). Bags for MAP storage were prepared by removal of air (20 mbar), followed by injection a 40% CO2 and 60% N2 gas mixture (AGA, Copenhagen, Denmark) to atmospheric pressure and finalised by sealing of the bags (Multivac C500, Multivac A/S, Vejle, Denmark). All samples (n = 342) were stored in a chilled room. Iced samples, both aerobic (n = 63) and MAP (n = 90), were entirely covered with flake ice which was regularly refilled during storage, as the ice melted. For superchilling, bags (n = 180) were submerged in slurry ice, produced by mixing sodium chloride, ice flakes and water to obtain a target temperature of -2.0 °C. The temperature was recorded every 30 min. during transport of the cod from Greenland to Denmark and during storage of all treatment by using a minimum of two temperature loggers for each treatment (TinyTag Plus, Gemini Data Loggers Ltd., Chichester, UK). After processing in Maniitsoq, triplicate samples

76 for aerobic viable counts (See 2.1.4) were taken, and at DTU Food, before dividing the cod pieces into the four treatment, fresh samples were analysed using sensory, chemical and microbial (enumeration and amplicon sequencing) methods (See 2.1.2-2.1.5). For each treatment, sampling was performed with intervals of two to four days during a total storage period of 21 days in ice to 32 days for superchilling. At each sampling time, three randomly picked bags, from each treatment, were analysed for microbiological and chemical changes. Five other randomly picked bags, from each treatment, were chosen for sensory evaluation.

2.1.2 Sensory changes during iced and superchilled storage

Sensory evaluation was performed by using the Quality Index Method (QIM) for thawed Atlantic cod filets, including scores for “Texture”, “Odour”, “Colour”, “Bloodstains” and “Parasites” (Archer, 2010). Compared to the original scheme, the odour attribute was expanded from the original score of 0, 1 and 2 to include “Acetic, ammonia” with a score value of 3. At each day of analyses, five pieces of cod from each treatment were each given a random three digits code and placed on cooling plates, to avoid changing sensory scores during the session. Samples were presented under artificial daylight (6500K, L 36W 965 Lumilux De Luxe, Osram, Germany) to a tested and trained panel consisting of four to seven assessors per session. At each sensory session during the storage trial, six mock samples were included randomly to prevent assessors from guessing the evolution of QI scores. The mock samples were prepared by thawing cod from the same fishing ground and by storing these samples at the same conditions as the real samples, but with different storage times. The results of the mock samples were not included in the presented data. End of shelf-life was determined after completion of the storage trial and based on the evolution of scores for the four treatments as well as variability of scores for the mock samples.

2.1.3 Chemical changes

Chemical changes as potential indices of spoilage were determined throughout the storage trial: Trimethylamine-oxide (TMAO), trimethylamine (TMA) and total volatile nitrogen (TVN) was determined in duplicate for each bag by a modified Conway and Byrne method (Conway and Byrne, 1933). pH was recorded for each sample as part of the Conway and Byrne protocol by using a pH meter (HQ411D Benchtop Meter, HACH Company, Loveland, USA). Lactic and acetic acids were determined,

77 with duplicate extract of each of the three cod pieces from each treatment, by HPLC with external standards for identification and quantification (Dalgaard and Jørgensen, 2000). Headspace gas composition was determined on each bag for microbiological and chemical analysis by using a gas analyser to measure CO2 and O2 concentrations (Checkmate3, MOCON Dansensor®, Ringsted, Denmark).

2.1.4 Culture-depended microbiology

The microbiota was quantified in triplicate, i.e. three separate bags, for each treatment and for each sampling time by diluting 20.0 grams of cod without skin tenfold in chilled physiological saline with 0.1% peptone (PSP) (NMKL, 2006) followed by homogenisation for 60 seconds in a Stomacher 400 (Seward Medical, London, UK). Further 10-fold dilutions with PSP were performed as required. Aerobic viable counts (AVC) was determined by spread plating on chilled Long and Hammer (LH) agar with 1% NaCl (7 d; 15°C) (NMKL , 2006). Pseudomonas spp. was determined by spread plating on Pseudomonads agar (CM0559, Oxoid, Basingstoke, UK) with CFC selective supplement (SR0103, Oxoid, Basingstoke, UK) and incubation for 48 h at 25⁰C. H2S-producing Shewanella spp. was determine as black colonies by pour plating in Iron Agar Lyngby (CM0964, Oxoid, Basingstoke, UK) with L-cysteine hydrochloride and incubation for three days at 25 ⁰C (NMKL, 2006). Photobacterium phosphoreum was enumerated by using a conductance method with incubation at 15 °C (Dalgaard et al., 1996). Lactic Acid Bacteria (LAB) were quantified by pour plating in nitrite actidione polymyxin (NAP) agar and counted after incubation for four days at 25 ⁰C (Davidson and Cronin, 1973).

To identify the dominating microbiota for each treatment, all countable colonies on LH plates with the highest dilution factor (3 plates for each treatment) were divided into groups based on colony characteristics (size, profile, elevation, boundary, colour). 13 colonies for iced cod in air, nine colonies for iced cod in MAP, 28 colonies for superchilled cod in air and 16 colonies for superchilled cod in MAP were divided into groups and for each group of cololonies, their proportion of the concentration of countable colonies was calculated. To identify the groups of colonies present for each treatment, ten colonies were isolated from LH plates (highest dilutions) at the time of sensory spoilage or at the end of the storage period, with the exception of iced cod in MAP, where only eight isolates were sequenced. For identification of isolates these were pure-cultured using the GMB medium (Dalgaard et al., 1994) and LH plates. DNA was extracted using the DNeasy Blood & Tissue Kit (Qiagen, Germany), the 16S rRNA gene was targeted by PCR reactions with specific primers (Forward (27F): AGAGTTTGATCMTGGCTCAG, Reverse

(1492R): ACCTTGTTACGACTT) and the PCR products were purified by using the MinElute PCR Purification Kit (Qiagen, Germany). The purified PCR products were sent to Eurofins Genomic for sequencing (Mix2Seq, Eurofins Genomics). Sequences were trimmed with the CLC workbench, by removing part of the sequences with low quality score (limit of 0.05) and the trimmed sequences were not allowed to have more than two ambiguous nucleotides (CLC workbench 8.1, Qiagen, Aarhus, Denmark). The trimmed sequences were assembled to reads with a minimum of 50 aligned base pairs. For identification, the 16S rRNA gene reads of the isolates were compared to the NCBI 16S ribosomal RNA sequence Database using their website service tool (https://blast.ncbi.nlm.nih.gov/Blast.cgi).

Isolates identified as Photobacterium spp. by 16S rRNA gene sequencing, were further analysed by partial sequencing of their gyrB gene (Forward (22fVf): GAAGTTATCATGACGGTACTTC, Reverse (1240rVf): AGCGTACGAATGTGAGAACC) (Ast and Dunlap, 2004) and speciated by Maximum Likelihood Phylogeny, constructed based on the Neighbour-joining methods and by using a general time-reversible nucleotide substitution model. Bootstrap values were calculated based on 1,000 replicates (CLC workbench 8.1, Qiagen, Aarhus, Denmark).

A subsection (n=19) of the isolates (3x10+8) were chosen to represent the different identified species and analysed for spoilage potential and spoilage activity (Table 1, See 2.2).

2.1.5 Culture-independent microbiology

DNA amplicon sequencing was performed to analyse the microbiota at the start of the storage period, and at the point of sensory spoilage or at the end of the storage period for the different treatments. DNA amplicon sequencing was performed in triplicate resulting in 15 samples with three samples before storage and three samples taken after storage for each of the four treatments. One gram

79 of filet meat was sampled and stored at -80 ⁰C until the end of the experiment. Samples were shipped covered in dry ice to Eurofins Genomics for DNA extraction and amplicon Illumina MiSeq sequencing of the V1-V3 region in the 16S rRNA gene (Forward: AGAGTTTGATCATGGCTCAG, Reverse: GTATTACCGCGGCTGCTG) (Leser et al., 2002; Weisburg et al., 1991). Eurofins performed quality control check, and the primer sequences were removed from the sequences. Quantitative Insights Into Microbial Ecology 2 (QIIME2) (Bolyen et al., 2019) using the DADA2 pipeline (Callahan et al., 2016) and following the standard operating procedure (https://github.com/LangilleLab/microbiome_helper/wiki/Amplicon-SOP- v2-(qiime2-2018.6)) were used to assign Amplicon Sequence Variants (ASV) from reads. The DADA2 approach was selected for identification of real biological variants, and the term ASV was used to separate these from Operational Taxonomic Units (OTUs) (Callahan et al., 2016). To minimise sequencing carry- over contamination between MiSeq runs, ASV with an abundance of less than 0.1% of the total observations were filtered out, and the sampling depth of the analysis was based on the number of reads in the sample with fewest reads. To assign taxonomy for the ASVs, a classifier was generated based on the SILVA 132 SSU Ref NR 99 database (Quast et al., 2013). The V1-V3 region based on the amplicon sequencing primers were extracted from the database to minimise false-positive and used as the classifier. Merged reads were deposited at the NIH NCBI Sequence Read Archive with the accession number PRJNA565897.

The total number of ASVs were used for each treatment to represent the species richness and Faith phylogenetic diversity (Faith, 1992) was used as a measure of phylogenetic differences within a treatment. Phylogenetic beta-diversities were calculated using the unweighted UniFrac matrix and used for pairwise comparison of microbial composition between treatments with values ranging from 0.0 for complete similarity to 1.0 for complete dissimilarity (Lozupone and Knight, 2005).

2.2 Spoilage potential and spoilage activity of the dominating microbiota

To identify the bacteria responsible for spoilage of cod in different treatments, the qualitative spoilage potential and the quantitative spoilage activity was determined (Dalgaard, 1995). Spoilage potential was determined as the ability of isolates to produce off-odours when growing in cod muscle blocks (MB) at 0 °C. Spoilage activity was quantified as the yield factor for, respectively, TVN

(YTVN/CFU, mg-N/CFU) and TMA (YTMA/CFU, mg-N/CFU) formation of isolates growing in cod MB at 0 °C. Spoilage potential and spoilage activity were determined in duplicate for 19 different isolates (Table 1).

Each isolate was pre-cultured in GMB medium at 15 °C (Dalgaard et al., 1994) and thawed cod MB were then inoculated with 4 log CFU/g. This high inoculum was chosen to ensure that the background microbiota was not contributing to off-odour or amine formation which was further evaluated for non- inoculated MB. MB was, respectively, stored in the same atmosphere as used for the treatments in the storage trial where the different isolates were isolated from (Table 1). After storage at 0°C during ten days in air or 14 days in MAP, the MB were placed in plastic trays at room temperature for 15 minutes before off-odour evaluation by five assessors. The assessors provided a score based on off-odour attributes; (i) no off-odour, (ii) weak off-odour or (iii) strong off-odour. An isolate was determined to have a spoilage potential if the average score was above 2.0.

The same 19 isolates and the same inoculation and storage conditions were used to determine both spoilage potential and spoilage activity. For each isolate, the cell concentration and the concentrations of TVN and TMA were determined after inoculation of MB and at the end of the storage period. Cell concentrations were determined using LH (See 2.1.4), and the volatile amines were quantified by using Conway titration (See 2.1.3). The yield factors (mg-N/CFU) were calculated using the equation presented by Dalgaard, (1995), with the modification that yield factors for both TVN formation and TMA formation were determined (Eq. 1).

푚푔−푁 푚푔−푁 1 (푇푉푁퐹𝑖푛푎푙 ( ⁄100 푔) − 푇푉푁 퐼푛𝑖푡𝑖푎푙( ⁄100 푔))× 푌 = 100 (1) 푇푉푁/퐶퐹푈 10log (퐶퐹푈/푔)퐹𝑖푛푎푙−10log (퐶퐹푈/푔)퐼푛𝑖푡𝑖푎푙

The yield factor for a group of isolates was calculated as the average of the log-transformed yield factor values for the isolates within the group. These average yield factors for each bacterial group were used to determine calculated concentrations of TVN (mg-N/100g) during the storage period for the four treatments analysed in the storage trial (See 2.1). Calculated concentrations of TVN (mg-N/100g)

log CFU/g were determined from measured concentrations (10 ) of Pseudomonas spp., H2S-producing Shewanella and Photobacterium spp. (See 2.1.4) as shown in Eq. (2). The initial TVN concentration of 14 mg-N TVN/100g in cod was added to the amount of TVN formed by the three groups of bacteria (Eq. 2).

푙표푔퐶퐹푈 (푃푠푒푢푑표푚표푛푎푠) 푙표푔퐶퐹푈 (푆ℎ푒푤푎푛푒푙푙푎) 푔 푔 퐶푎푙푐푢푙푎푡푒푑푇푉푁 (푚푔 − 푁/100푔) = 퐼푛𝑖푡𝑖푎푙 푇푉푁 + (10 × 푌푇푉푁 (푃푠푒푢푑표푚표푛푎푠) + 10 × 푌푇푉푁 (푆ℎ푒푤푎푛푒푙푙푎) +

푙표푔퐶퐹푈 (푃ℎ표푡표푏푎푐푡푒푟𝑖푢푚) 푔 10 × 푌푇푉푁/푃ℎ표푡표푏푎푐푡푒푟𝑖푢푚) × 100 (2)

2.2 Statistical analyses

Statistical analyses for the difference between the initial values of pH or lactic acid and the values at sensory spoilage or at the end of the storage trial were performed using with a two-tailed homoscedastic distribution t-Test (Microsoft Excel 2016, Microsoft Corp., Redmond, WA, USA). Maximum specific growth

-1 rate (μmax, h ) of P. carnosum were determined by fitting the results of the conductance methods to the log-transformed 3-parameter logistic model (Dalgaard et al., 1997b).

3. Results

3.1 Storage trial

3.1.1 Storage conditions

During transport for 38 hours from Maniitsoq in Greenland to the laboratory at DTU Food in Denmark, the pieces of cod were at 1.5 ± 1.1oC. The average temperatures for the following iced or superchilled storage were, respectively, 0.4 ± 0.06°C and -1.7 ± 0.08°C (Table 2). During storage, the headspace gas composition surrounding the cod changed with the major changes occurring after the point of sensory spoilage, i.e. CO2 and O2 concentrations changed in the iced and aerobically stored cod from the point of spoilage on day 15 to the end of the storage period on day 21 (Table 2).

3.1.2 Sensory changes

Scores of odour attributes resulted in a shelf-life for iced cod of 15 days when stored in air and of 22 days in MAP. The end of shelf-life was set to a score value of 1.0. Superchilled cod did not reached sensory spoilage within the storage period of 32 days, irrespective of storage in air or MAP (Fig. 1, Table 3). With the applied QIM scheme, the total QI scores did not show a clear development during the storage period (Results not shown).

) r

u 2

(

I 1 Q

0 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 Storage period (days)

Fig. 1. QI odour scores during storage of iced cod in air (●), iced cod in MAP (∆), superchilled cod in air (■) and superchilled cod in MAP (◊). Symbols and error bars indicate Avg ± SD.

3.1.3 Chemical changes

The cod meat pH of 6.7 ± 0.4, measured 1.6 days after processing in Greenland, increased during storage to 7.3 ± 0.4 and 7.4 ± 0.6 for iced cod in air and MAP, respectively (Table 4). Using pH of 7.0 as an index of spoilage, it took 15 days for iced cod in air and 24 days for cod in MAP to reach this value (Table 2). For superchilled cod, pH did not reach 7.0 (Table 4). A similar development was observed for the concentration of volatile amines where TVN reached the EU limit for Gadidae of 35 mg-N TVN/100 g fish muscle (EC, 2008) after 14 and 20 days, respectively, for iced cod in air or MAP (Fig. 2; Table 3). Superchilled cod showed no significant increase in TVN concentrations (p > 0.05; linear regression (slope), Fig. 2). The initial TMAO concentration of 73 ± 16 mg-N TMAO/100 g fish muscle was reduced when TMA was produced during iced storage. TMA concentrations increased with the same rate as TVN, from zero to 59 mg-N TMA/100 g and to 63 mg-N TMA/100 g for iced cod in air or MAP, respectively. The concentration of TMA in superchilled MAP samples remained below ten mg-N TMA/100 g during storage. The initial lactic acid concentration of 3,209 ± 373 ppm was reduced to 1,805 ± 316 ppm during iced aerobic storage whereas significant changes were not observed (p > 0.05, t-test) for other storage conditions (Table 5). Other chemical changes included an increased level of isobutyric acid (tentatively identified by HPLC retention time of 1.54 relative to lactic acid) for iced cod in air (Results not shown).

100 95 90 85

) 80

0 75

0 1

/ 70 N

- 65

g m

( 60

e 55 g

o 50

t i

n 45

e l

i 40 EU limit for Gadidae family

t a

l 35

o v

l a

t 25 o T 20 15 10 5 0 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 Storage period (days)

Fig. 2. Formation of total volatile nitrogen (TVN) in cod pieces stored iced in air (●), iced in MAP (∆), superchilled in air (■) and superchilled in MAP (◊). Symbols and error bars indicate Avg ± SD. Dashed line represent the critical EU limit of 35 mg-N TVN/100g (EC, 2008).

3.1.4 Culture-based microbial changes

At the time of filleting in Greenland, cod had AVC of 2.7 ± 0.6 log CFU/g. For storage in ice, AVC reached 7.0 log CFU/g after 13 days of aerobic storage and after 17 days in MAP (Fig. 3, Table 3). Superchilled cod in air showed a slower growth rate for AVC and reached 7.0 log CFU/g after 23 days, and for superchilled MAP cod, no clear microbial growth was observed with AVC reaching 3.9 ± 1.8 log CFU/g after 32 days of storage (Fig. 3a, Table 3). Based on concentrations of bacteria as determined by using selective media, Photobacterium spp. dominated the microbiota of iced cod in air during the period of from six to 14 days of storage (Fig. 3b, c and d). From 14 days of storage the microbiota were also dominated by Pseudomonas spp. and H2S-producing Shewanella as they reached concentrations of 7.7 log CFU/g and similar to those of Photobacterium spp. (Fig. 3c, b, d). For superchilled cod in air,

Pseudomonas spp. dominated the microbiota with H2S-producing Shewanella and Photobacterium spp. being at slightly lower concentrations (Fig. 3). However the cod was not sensory spoiled (Fig. 1, Table 3). Photobacterium spp. dominated the spoilage microbiota of iced MAP cod and at the time of sensory spoilage (22 d), their concentration was 4 log CFU/g higher than those of both Pseudomonas spp. and H2S- producing Shewanella (Fig. 3b, c, d). This was reflected by identification of the isolated bacteria where 100 % were Photobacterium spp. (Table 6). Finally, for superchilled MAP cod growth of the studied groups of microorganisms were limited, and no particular group of microorganisms seemed to dominate the microbiota (Table 6). For LAB limited growth was observed and exclusively for iced cod in air where they reached 4.7 log CFU/g (Results not shown).

10 10

9 A 9 B

8 8

7 7

6 6 5 5

4 4

3 3

2 2 g

/ 1 1 U

F 0 0

C 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34

g 10

o 10

L C D 9 9

8 8

7 7

6 6

5 5

4 4

3 3

2 2

1 1

0 0 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 Storage period (days)

Fig. 3. Microbial growth during storage of cod fillets: Aerobic viable counts (A), Photobacterium spp. (B),

Pseudomonas spp. (C) and H2S-producing Shewanella (D). Cod were stored iced in air (●), iced in MAP (∆), superchilled in air (■) and superchilled in MAP (◊). Symbols and error bars indicate Avg ± SD.

3.1.5 Identification of isolates

Thirty-eight isolates from the dominating microbiota on L&H agar plates were identified by 16S rRNA gene sequencing (Table 6). For cod in air, Pseudomonas spp. dominated the microbiota with 92 % and 67 % in iced and superchilled cod, respectively. These percentages were calculated from identified isolates and from the proportion of their colonies in the microbiota (See 2.1.4). For MAP storage, Photobacterium spp. dominated the microbiota with 100 % and 85 % of the isolates from iced cod (n = 8) and superchilled cod (n = 7), respectively. Of the 15 isolates of Photobacterium spp. from the two MAP treatments (Table 6), 11 had their gyrB gene successfully sequenced and used for identification. Ten of these isolates were identified as P. carnosum, and one isolate (M0.6) may belong to a not yet described Photobacterium species (Fig. 4).

Fig. 4. Maximum likelihood phylogeny tree based on 1045 bp of the gyrB gene for 11 Photobacterium isolates from the present study (M0.1, M0.3, M0.5, M0.6, M0.7, M0.8, M2.1, M2.4, M2.5, M2.6 and M2.8). Sequences for Photobacterium species were obtained from literature and NCBI accession numbers were shown in braches. Bootstrap values were calculated for 1,000 replicas. Branch length represent the number of nucleotide substitutions per site, the number of substitutions was indicated with the scale bar.

3.1.6 Analyses of microbiota by 16S rRNA gene amplicon sequencing

Of the 15 analysed samples, 13 samples passed the performed quality control. One of the triplicate fresh cod samples (before storage) and one superchilled MAP sample did not pass the quality control. The 13 samples resulted in a total of 923,177 individual reads. After rarefication, the sampling depth was chosen to 9,300 reads per sample, at this depth, all rarefication curves had levelled off. Species richness was lowest for iced MAP cod with an average of 34 ASVs compared to 98 ASVs for fresh samples (Fig. 5). Within treatments, the fresh sample had the highest phylogenetic differences with Faith phylogenetic diversity of 23.6, followed by superchilled MAP cod with limited microbial growth (5.8), iced cod in air (2.3), superchilled cod in air (2.0) and iced cod in MAP (1.2). The phylogenetic differences were

88 also shown by the composition of genera for the treatments of cod (Table 6) and this corresponded to number of ASVs (Fig. 5). Between treatments, the unweighted UniFrac distance matrix showed the microbial composition of fresh samples to differ from samples at the time of sensory spoilage or at the end of experiments (0.92 - 0.95) although the value for superchilled MAP cod was slightly lower (0.83; Table 6). Compared to iced cod in air with shelf-life of 15 days, MAP iced cod with shelf-life of 22 days changed the microbial composition (UniFrac distance of 0.60) more than superchilled cod in air with shelf- life > 32 days (UniFrac distance of 0.43) (Table 7).

Fresh cod

Iced cod in air

Iced cod in MAP

Superchilled cod in air

Superchilled cod in MAP

20 30 40 50 60 70 80 90 100 110 Numbers of ASV

Fig. 5. Numbers of Amplicon Sequence Variant (ASV) in fresh and stored cod pieces. Samples included fresh cod (•) and iced cod in air (●), iced MAP cod (∆), superchilled cod in air (■) and superchilled MAP cod (◊). Symbols and error bars indicate Avg ± SD.

3.2 Spoilage potential and activity

Common for isolates of Pseudomonas, Shewanella and Photobacterium, was that the majority of the isolates had spoilage potential (Table. 8). The Photobacterium carnosum isolates had six time’s higher yield-factor for TVN and TMA formation than the Shewanella baltica and Shewanella putrefaciens isolates and 200 times higher yield-factor for TVN formation than Pseudomonas spp. (Table 8). For non-inoculated MB, cell concentrations remained below 6.1 log CFU/g, and less than 5.0 mg-N TVN/100g was formed indicating that the TVN concentration used for yield factor determination was formed by the studied isolates (Results not shown).

For iced cod in air, the TVN-concentrations calculated from enumerated bacteria and yield factors were close to TVN-concentrations observed in the storage trial both concerning the final concentration and changes during the storage time (Fig. 6A). For iced MAP cod, the observed and calculated TVN-concentrations were in close agreement until the end of sensory shelf-life (22 d). After 28 days of storage the calculated TVN-concentration was lower than the observed TVN-concentration and this may reflect an underestimation of the concentration of Photobacterium spp. (Fig. 2B, Fig. 6). For both treatments, Photobacterium spp. were responsible for at least 97 % of the calculated TVN formation (Table 9). For iced cod in air, Photobacterium spp. made up 65 % of the spoilage microbiota based on enumeration by selective media (Fig. 3) and 24.1 % based on amplicon sequencing, but they contributed 97 % of the produced TVN whereas Pseudomonas spp. contribute less than one percent of the formed TVN concentrations, which was significantly less than their large proportion of the microbiota (Table 9).

100 100 )

g B

90 A 90

0 1

/ 80 80

- g

70 70

(

e 60 60

r t

i 50 50

e l

i 40 40

a l

o 30 30

l a

t 20 20

o T 10 10

0 0 0 2 4 6 8 10 12 14 16 18 20 22 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 Storage period (days)

Fig. 6. Development of observed total volatile nitrogen (TVN) during storage (●) and calculated TVN formation based on microbial count (Fig. 3b, c, d) and yield factors for TVN-formation (Table 8) (◊). Treatments included iced cod in air (A) and iced cod in MAP (B). Symbols and error bars indicate Avg ± SD.

4. Discussion

For Atlantic cod from Iceland, Wang et al. (2008) and Lauzon et al. (2009) found shelf-life of 24-26 days for superchilled MAP products at -0.9 or -2.0 °C an d these shelf-lives included 3-5 days of iced storage before packaging and superchilling, these results were similar to the shelf-life of superchilled MAP cod from Greenland (Table 3). However, they found sensory shelf-lives of 14-17 days for aerobic superchilled cod and these were markedly shorter than the > 32 days observed in the present study. To evaluate distribution of cod at changing temperatures, we calculated 32 days at -1.7 °C as equivalent to 22 days at 0 °C according to the relative rate of spoilage (RRS) model suggested by Dalgaard and Huss (1997) for fresh fish from temperate waters. By using this RRS model, as included in the FSSP software (FSSP, 2014), a distribution chain with ten days superchilled transportation from

Greenland to Europe at -1.7 °C followed by seven days chilled distribution at +2 °C e.g. through supermarkets and one day consumer storage at +5 °C corresponds to 19.2 days at 0 °C. The remaining shelf-life of > 2.8 days at 0 °C suggests it would be possible to ship superchilled cod from Greenland to Europe with a final chilled distribution on land. This distribution, however, requires careful temperature management. The remaining shelf-life of > 2.8 days would be reduced to zero if the ten days of superchilling was carried out at -0.1 °C instead of -1.7 °C or the seven days of chilled distribution was at +3.6 °C instead of +2.0 °C.

For storage of Atlantic cod in ice Olafsdottir et al. (2006) and Wang et al. (2008) found 27- 49 mg-N TVN/100g at sensory spoilage after 9-13 days. MAP extended shelf life to 14->24 days with >36- 46 mg-N TVN/100g at sensory spoilage. This TVN formation and shelf-lives corresponded reasonably to iced cod in the present study (Table 3). However, these Icelandic studies found 37-59 mg-N TVN/100g at sensory spoilage after 12-17 days for superchilled cod in air. The markedly slower TVN-development (Fig. 2) and longer shelf-life for superchilled cod in the present study (Table 3) could, at least partly, be explained by a slightly lower storage temperature of -1.7 °C compared to -0.9 °C for Wang et al. (2008) and -1.3 °C for Olafsdottir et al. (2006).

The initial pH of 6.7 for Atlantic cod from CBA in Greenland (Table 4) was similar to previously reported values of 6.3-6.7 for wild-caught cod (Debevere and Boskou, 1996; Rustad, 1992) and in more than 1,900 cod from an area ranging from West Greenland, Spitzbergen and Aberdeen, only 44 individuals had a pH below 6.25 (Love, 1979). However, Atlantic cod from aquaculture typically have much lower pH of 6.10-6.13 (Duun and Rustad, 2007; Hansen et al., 2007; Sivertsvik, 2007). Common for all studies, the pH increased during storage and at the time of sensory spoilage Wang et al. (2008) and Olafsdottir et al., (2006) found pH of 6.7-7.2. These findings support the use of >7.0 as an index for spoilage of Atlantic cod from CBA or wild-caught (Table 3).

For fresh and spoiled MAP cod (20 % CO2 and 80 % N2) Chaillou et al. (2015) found a 79 % reduction of OTUs from 225 to 48 and Kuuliala et al. (2018) reported OTUs being reduced at the end of storage for salted cod. This corresponds well with the 65 % reduction of ASVs between fresh and spoiled MAP cod in the present study (Fig. 5). 4). A reduction of OTUs during storage for fresh and lightly preserved food have been observed for different seafood and meat products (Chaillou et al. 2015) and this is in agreement with the hypothesis of SSOs being selected during storage (Dalgaard, 2000).

The observed shift in the dominating spoilage microbiota from Pseudomonas spp., Photobacterium spp., Shewanella spp. and Acinetobacter spp. under aerobic storage to Photobacterium spp. in MAP cod (Table 6) also seemed in agreement with several previous studies of cod. P. phosphoreum made up less than 1 % of the spoilage microbiota in iced aerobic Atlantic cod from Denmark whereas they made up more than 90 % of the spoilage microbiota in MAP cod from Denmark and Iceland (Dalgaard et al. 1997c). By cultivation, Reynisson et al. (2009) found 29 % Pseudomonas spp. and 6 % P. phosphoreum in the spoilage microbiota for cod in air and these values changed to < 1 % Pseudomonas spp. and 21-99 % P. phosphoreum for MAP storage. However, by using 16S rRNA clone analysis and t-RFLP Reynisson et al. (2009) found higher percentages (84-100 %) of P. phosphoreum for both aerobic and MAP storage. Similarly, based on 16S rRNA gene amplicon sequencing data, Photobacterium spp. made up 81.2-92.5 % of the microbiota in chilled vacuum-packed cod and 96.3-97.5 % in chilled MAP cod (Hansen et al., 2016).

> 7.0 log CFU/g has been suggested as the microbial concentration where spoilage of fresh fish generally are detected (ICMSF, 2011) and more recently Eliasson et al. (2019) found this to be appropriate for iced and superchilled Atlantic cod in air. However, this microbial index underestimated sensory shelf-life in the present study (Table 3) as also observed by Olafsdottir et al. (2006) and Wang et al. (2008) where AVC of 7.4-8.1 log CFU/g were found at the time of sensory spoilage for both aerobic and MAP stored cod. With the concentration of TVN being a reasonable index of sensory spoilage (Table 3) and with the spoilage microbiota being selected depending on storage conditions (Table 6) it must be expected that microbial concentrations at the time of spoilage will depend on product storage. This is due to the marked difference in TVN formation by different groups of bacteria. That Photobacterium spp. in the present study produced 6-10 time more TVN and TMA per cell than H2S-producing Shewanella (Table 8) is in close agreement with previous studies of their yield-factors (Dalgaard, 1995). The quantitatively very low TVN-formation per cell by Pseudomonas spp. (Table 8) also corresponded to available data for this organism (Koutsoumanis and Nychas, 2000; Xie et al., 2018). With 35 mg-N TVN/100g as index of spoilage, the calculated yield-factors for TVN suggest minimal spoilage levels of 7.2 log CFU/g for

Photobacterium spp., 7.8 log CFU/g for H2S-producing Shewanella spp. and 9.5 log CFU/g for Pseudomonas spp. These microbial indices of spoilage were calculated as the concentration required for each group of bacteria to produce 35 mg-N TVN/100g in cod.

Based on their pronounced TVN-formation Photobacterium spp. were identified as the SSO responsible for spoilage of both aerobically stored and MAP Atlantic cod from CBA in Greenland (Table 8). Photobacterium spp. did not completely dominate the spoilage microbiota of the aerobically stored cod

93 but concentrations of other groups of bacteria, including H2S-producing Shewanella, could not form the concentration of TVN that correlated with sensory spoilage. A similar situation was previously reported for vacuum-packed and MAP cod fillets where both P. phosphoreum and S. putrefaciens were present and where the pronounced TMA-formation by P. phosphoreum made it the SSO (Dalgaard, 1995). Those highly TMA producing cells were bioluminescent and non-bioluminescent P. phosphoreum with some variability in their characteristics (Dalgaard, 1995; Dalgaard et al., 1997a). The non-bioluminescent P. carnosum isolates (n = 10) and the potentially new Photobacterium species (n = 1) from the present study (Fig. 4) also belongs to the Photobacterium phosphoreum clade (Labella et al., 2018; Le Doujet et al., 2019). P. carnosum is a recently described species that have been isolated from MAP poultry meat (Hilgarth et al., 2018) and it is an important part of the gut microbiota for Atlantic cod (Le Doujet et al., 2019).

P. carnosum was enumerated in cod by using a conductance-based method developed for bioluminescent and non-bioluminescent P. phosphoreum (Fig. 2; Dalgaard et al., 1996), and found P. carnosum to have the same yield factor for TMA-formation as the previously studied bioluminescent and non-bioluminescent P. phosphoreum isolates (Table 8; Dalgaard, 1995). However, shelf-life of cod from CBA in Greenland was longer than reported for Atlantic cod from other regions were bioluminescent, and non-bioluminescent P. phosphoreum was the SSO. The longer shelf-life could be due to low initial concentrations of P. carnosum or that they grow slower in cod than bioluminescent, and non- bioluminescent P. phosphoreum. The initial concentration of 1.0 log CFU/g for P. carnosum in CBA cod from Greenland was similar to initial concentrations of P. phosphoreum in Icelandic iced cod in air (Olafsdottir et al., 2006) and Danish iced MAP cod (Dalgaard et al., 1997b). However, the shelf-life of, respectively, 15 days and 22 days for CBA cod from Greenland (Fig. 2b; Table 3) was longer than the 11-

○ 14 days and 16-17 days for cod from Iceland and Denmark. At 0.4 C and with 0 % or 35 % CO2 in MAP the maximum specific growth rate of P. carnosum in cod from Greenland was, respectively, 0.079 h-1 and 0.040 h-1 (Fig 2b). For these conditions the growth rate of P. phosphoreum was 0.092 h-1 and 0.066 h-1 as predicted by the growth model of Dalgaard et al. (1997b) for MAP cod. Thus, P. carnosum may grow slower than P. phosphoreum previously identified as SSO in cod. Bioluminescent and non-bioluminescent P. phosphoreum previously determined as part of the spoilage microbiota for cod and other fresh fishes included P. aquimaris, P. iliopiscarium, P. kishitanii, P. phosphoreum, and P. piscicola (Ast and Dunlap, 2004; Dalgaard, 1995; Dalgaard et al., 1997a; Figge et al., 2014; Poirier et al., 2018). The present study suggests P. carnosum and probably other Photobacterium species are part of this group of spoilage bacteria. Further studies are needed to evaluate their occurrence and importance for spoilage of various foods.

5. Conclusions

For cod fillets from CBA in Greenland a lowering of the storage temperature from +0.4 °C to -1.7 °C had a more pronounced effect on shelf-life extension compared to changing the storage atmosphere from air to MAP with 35 % CO2 at equilibrium. Superchilling increased the sensory shelf-life to more than 32 days. The combination of superchilling and MAP inhibited microbial growth and TVN formation during the studied 32 days of storage. Based on the determination of spoilage activity, P. carnosum was pointed out as the specific spoilage organism that limited shelf-life of iced cod irrespective of storage in air or MAP conditions. Further studies are suggested to evaluate the long shelf-life of cod fillets from CBA in Greenland and of the occurrence and importance for P. carnosum in food spoilage.

Acknowledgments

This research was funded by Innovation Fund Denmark (grant no. 5189-00175B). We thank employees at the fish factory in Maniitsoq, Greenland for their assistance and the laboratory technicians; Mia Laursen, Rannvá Høgnadóttir Houmann and Margrethe Carlsen for skilful assistance at DTU Food. Senior research scientist Grethe Hyldig and food technician Rie Sørensen provided input on sensory evaluation and we thank them and the sensory panel for their contribution.

References

Archer, M., 2010. Sensory assessment scoresheets for fish and shellfish. https://www.seafish.org/media/Publications/sensory_assessment_scoresheets_14_5_10.pdf (accessed 26 September 2019).

Ast, J.C., Dunlap, P. V., 2004. Phylogenetic analysis of the lux operon distinguishes two evolutionarily distinct clades of Photobacterium leiognathi. Arch. Microbiol. 181, 352–361. https://doi.org/10.1007/s00203-004-0663-7.

Bolyen, E., Dillon, M., Bokulich, N., Abnet, C., Al-Ghalith, G., Alexander, H., et al, 2018. QIIME 2:

Reproducible, interactive, scalable, and extensible microbiome data science. PeerJ Prepr. https://doi.org/10.7287/peerj.preprints.27295.

Bolyen, E., Rideout, J.R., Dillon, M.R., Bokulich, N.A., Abnet, C.C., Al-Ghalith, et al., 2019. Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nat. Biotechnol. 37, 852–857. https://doi.org/10.1038/s41587-019-0209-9.

Buch, E., 2002. Present Oceanographic Conditions in Greenland Waters, Danish Meteorological Institute, Copenhagen.

Buch, E., Horsted, S.A., Hovgård, H., 1994. Fluctuations in the occurrence of cod in Greenland waters and their possible causes. ICES Mar. Sci. Symp. 198, 158–174.

Callahan, B.J., McMurdie, P.J., Rosen, M.J., Han, A.W., Johnson, A.J.A., Holmes, S.P., 2016. DADA2: High- resolution sample inference from Illumina amplicon data. Nat. Methods 13, 581–583. https://doi.org/10.1038/nmeth.3869.

Chaillou, S., Chaulot-Talmon, A., Caekebeke, H., Cardinal, M., Christieans, S., Denis, C., et al, 2015. Origin and ecological selection of core and food-specific bacterial communities associated with meat and seafood spoilage. ISME J. 9, 1105–1118. https://doi.org/10.1038/ismej.2014.202.

Claussen, I.C., Indergård, E., Grinde, M., 2011. Comparative life cycle Assessment (LCA) of production and transport of chilled versus superchilled haddock (Melanogrammus aeglefinus) fillets from Norway to France. Procedia Food Sci. 1, 1091–1098. https://doi.org/10.1016/j.profoo.2011.09.163.

Conway, E.J., Byrne, A., 1933. An absorption apparatus for the micro-determination of certain volatile substances: The micro-determination of ammonia. Biochem. J. 27, 419-429.

Dalgaard, P., Ross, T., Kamperman, L., Neumeyer, K., McMeekin, T.A., 1994. Estimation of bacterial growth rates from turbidimetric and viable count data. Int. J. Food Microbiol. 23, 391–404. https://doi.org/10.1016/0168-1605(94)90165-1.

Dalgaard, P., Mejlholm, O., Huss, H.H., 1996. Conductance method for quantitative determination of Photobacterium phosphoreum in fish products. J. Appl. Bacteriol. 81, 57–64.

https://doi.org/10.1111/j.1365-2672.1996.tb03282.x.

Dalgaard, P., Huss, H.H., 1997. Mathematical modeling used for evaluation and prediction of microbial fish ppoilage, in: Shahidi, F., Jones, Y., Kiits, D. (Eds.), Seafood Safety, Processing and Biotechnology. Technomic pub. Co. Inc., Lancaster, pp. 73–89.

Dalgaard, P., Manfio, G.P., Goodfellow, M., 1997a. Classification of photobacteria associated with spoilage of fish products by numerical taxonomy and pyrolysis mass spectrometry. Zentralblatt fur Bakteriol. 285, 157–168. https://doi.org/10.1016/S0934-8840(97)80024-2.

Dalgaard, P., Mejlholm, O., Huss, H.H., 1997b. Application of an iterative approach for development of a microbial model predicting the shelf-life of packed fish. Int. J. Food Microbiol. 38, 169–179. https://doi.org/10.1016/S0168-1605(97)00101-3.

Dalgaard, P., Mejlholm, O., Christiansen, T.J., Huss, H.H., 1997c. Importance of Photobacterium phosphoreum in relation to spoilage of modified atmosphere-packed fish products. Lett. Appl. Microbiol. 24, 373–378. https://doi.org/10.1046/j.1472-765X.1997.00152.x.

Dalgaard, P., 2000. Fresh and lightly preserved seafood. in: Man, C.M.D., Jones, A.A. (Eds.), Shelf-Life Evaluation of Foods. Aspen Publishing Inc., Gaithersburg, pp. 110–139.

Dalgaard, P., Jørgensen, L. V, 2000. Cooked and brined shrimps packed in a modified atmosphere have a shelf-life of >7 months at 0 ○C, but spoil in 4-6 days at 25 ○C. Int. J. Food Sci. Technol. 35, 431–442. https://doi.org/10.1046/j.1365-2621.2000.00402.x.

Dalgaard, P., 2006. Microbiology of marine muscle foods, in: Hui, Y. (Ed.), Handbook of Food Science: Technology and Engineering. CRC Press, Boca Raton, pp. 1–20.

Davidson, C.M., Cronin, F., 1973. Medium for the selective enumeration of lactic acid bacteria from foods. Appl. Microbiol. 26, 439–40.

Debevere, J., Boskou, G., 1996. Effect of modified atmosphere packaging on the TVB/TMA-producing microflora of cod fillets. Int. J. Food Microbiol. 31, 221–229. https://doi.org/10.1016/0168- 1605(96)01001-X.

DeWitt, C., Oliveira, A., 2016. Modified atmosphere systems and shelf life extension of fish and fishery products. Foods 5, 48. https://doi.org/10.3390/foods5030048.

Duun, A.S., Rustad, T., 2007. Quality changes during superchilled storage of cod (Gadus morhua) fillets.

Food Chem. 105, 1067–1075. https://doi.org/10.1016/j.foodchem.2007.05.020.

EC, 2008. Commission Regulation (EC) No 1022/2008 of 17 October 2008 amending Regulation (EC) No 2074/2005 as regards the total volatile basic nitrogen (TVB-N) limits. Eur. Comm. 18–20.

Faith, D.P., 1992. Conservation evaluation and phylogenetic diversity. Biol. Conserv. 61, 1–10. https://doi.org/10.1016/0006-3207(92)91201-3.

Figge, M.J., Cleenwerck, I., van Uijen, A., De Vos, P., Huys, G., Robertson, L., 2014. Photobacterium piscicola sp. nov., isolated from marine fish and spoiled packed cod. Syst. Appl. Microbiol. 37, 329– 335. https://doi.org/10.1016/j.syapm.2014.05.003.

FSSP, 2014. Food Spoilage and Safety Predictor. http://fssp.food.dtu.dk/ (accessed 26 September 2019)

Hansen, A.Å., Mørkøre, T., Rudi, K., Olsen, E., Eie, T., 2007. Quality changes during refrigerated storage of MA-packaged pre-rigor fillets of farmed atlantic cod (Gadus morhua L.) using traditional MAP,

CO2 emitter, and vacuum. J. Food Sci. 72, 423–430. https://doi.org/10.1111/j.1750- 3841.2007.00561.x.

Hansen, A.Å., Moen, B., Rødbotten, M., Berget, I., Pettersen, M.K., 2016. Effect of vacuum or modified atmosphere packaging (MAP) in combination with a CO2 emitter on quality parameters of cod loins (Gadus morhua). Food Packag. Shelf Life 9, 29–37. https://doi.org/10.1016/j.fpsl.2016.05.005.

Hilgarth, M., Fuertes, S., Ehrmann, M., Vogel, R.F., 2018. Photobacterium carnosum sp. nov., isolated from spoiled modified atmosphere packaged poultry meat. Syst. Appl. Microbiol. 41, 44–50. https://doi.org/10.1016/j.syapm.2017.11.002.

Hoang, H.M., Brown, T., Indergard, E., Leducq, D., Alvarez, G., 2016. Life cycle assessment of salmon cold chains: Comparison between chilling and superchilling technologies. J. Clean. Prod. 126, 363–372. https://doi.org/10.1016/j.jclepro.2016.03.049.

Hovda, M.B., Sivertsvik, M., Lunestad, B.T., Rosnes, J.T., 2007. Microflora assessments using PCR- denaturing gradient gel electrophoresis of ozone-treated and modified-atmosphere-packaged farmed cod fillets. J. Food Prot. 70, 2460–2465. https://doi.org/10.4315/0362-028X-70.11.2460.

ICMSF, 2011. Fish and seafood products, in: Swanson, K.M.J. (Eds.), International Commission on Microbiological Specifications for Foods (ICMSF) Microorganisms in Foods 8. Springer, Boston, pp. 107-135. https://doi.org/10.1007/978-1-4419-9374-8_19.

James, S.J., James, C., 2014. Chilling and freezing, in: Motarjemi, Y., Levievel, H. (Eds.), Food Safety Management: A Practical Guide for the Food Industry, Academic Press, London, pp. 481-510. https://doi.org/10.1016/B978-0-12-381504-0.00020-2.

Koutsoumanis, K., Nychas, G.J.E., 2000. Application of a systematic experimental procedure to develop a microbial model for rapid fish shelf life predictions. Int. J. Food Microbiol. 60, 171–184. https://doi.org/10.1016/S0168-1605(00)00309-3.

Koutsoumanis, K.P., Gougouli, M., 2015. Use of time temperature Integrators in food safety management. Trends Food Sci. Technol. 43, 236–244. https://doi.org/10.1016/j.tifs.2015.02.008.

Labella, A.M., Castro, M.D., Manchado, M., Lucena, T., Arahal, D.R., Borrego, J.J., 2018. Photobacterium malacitanum sp. nov., and Photobacterium andalusiense sp. nov., two new bacteria isolated from diseased farmed fish in Southern Spain. Syst. Appl. Microbiol. 41, 444–451. https://doi.org/10.1016/j.syapm.2018.04.005.

Le Doujet, T., De Santi, C., Klemetsen, T., Hjerde, E., Willassen, N.P., Haugen, P., 2019. Closely-related Photobacterium strains comprise the majority of bacteria in the gut of migrating Atlantic cod (Gadus morhua). Microbiome 7. https://doi.org/10.1186/s40168-019-0681-y.

Leser, T.D., Amenuvor, J.Z., Jensen, T.K., Lindecrona, R.H., Boye, M., Møller, K., 2002. Culture- independent analysis of gut bacteria: The pig intestinal tract revisited. Appl. Environ. Microbiol. 68, 673–690. https://doi.org/10.1128/AEM.68.2.673.

Lozupone, C., Knight, R., 2005. UniFrac: A new phylogenetic method for comparing microbial communities. Appl. Environ. Microbiol. 71, 8228–8235. https://doi.org/10.1128/AEM.71.12.8228- 8235.2005.

Love, R.M., 1979. The post‐mortem pH of cod and haddock muscle and its seasonal variation. J. Sci. Food Agric. 30, 433–438. https://doi.org/10.1002/jsfa.2740300414.

NMKL, 2006. Aerobic count and specific spoilage organisms in fish and fish products. Nordisk Metodikkomité for Næringsmidler (NMK; Nordic Committee on Food Analysis). No. 184.

Olafsdottir, G., Lauzon, H.L., Martinsdóttir, E., Oehlenschläger, J., Kristbergsson, K., 2006. Evaluation of shelf life of superchilled cod(Gadus morhua) fillets and the influence of temperature fluctuations during storage on microbial and chemical quality indicators. J. Food Sci. 71. https://doi.org/10.1111/j.1365-2621.2006.tb08928.x.

Poirier, S., Rue, O., Peguilhan, R., Coeurent, G., Zagorec, M., Champomier-Vergès, M.-C., et al, 2018. Deciphering intra-species bacterial diversity of meat and seafood spoilage microbiota using gyrB amplicon sequencing : A comparative analysis with 16S rDNA V3-V4 amplicon sequencing. PLoS One 1–26. https://doi.org/10.6084/m9.figshare.7083209.

Quast, C., Pruesse, E., Yilmaz, P., Gerken, J., Schweer, T., Yarza, P., et al, 2013. The SILVA ribosomal RNA gene database project: Improved data processing and web-based tools. Nucleic Acids Res. 41, 590– 596. https://doi.org/10.1093/nar/gks1219.

Reynisson, E., Lauzon, H.L., Magnsson, H., Jánsdáttir, R., Lafsdáttir, G., Marteinsson, V., Hreggvisson, G.L., 2009. Bacterial composition and succession during storage of North-Atlantic cod (Gadus morhua) at superchilled temperatures. BMC Microbiol. 9, 1–12. https://doi.org/10.1186/1471- 2180-9-250.

Rosnes, J.T., Kleiberg, G., Sivertsvik, M., Lunestad, B.T., Lorentzen, G., 2006. Effect of modified atmosphere packaging and superchilled storage on the shelf-life of farmed ready-to-cook spotted wolf-fish (Anarhichas minor). Packag. Technol. Sci. 19, 325–333. https://doi.org/10.1002/pts.735.

100

Rustad, T., 1992. Muscle chemistry and the quality of wild and farmed cod, in: Huss, H.H., Jakobsen, M., Liston, J. (Eds.), Quality Assurance in the Fish Industry: Proceedings of an International Conferenc, Copenhagen, Denmark, 26-30 August 1991. Elsevier Science publishers, Amsterdam, pp. 19–27.

Sivertsvik, M., Jeksrud, W.K., Rosnes, J.T., 2002. A review of modified atmosphere packaging of fish and fishery products - Significance of microbial growth, activities and safety. Int. J. Food Sci. Technol. 37, 107–127. https://doi.org/10.1046/j.1365-2621.2002.00548.x.

Sivertsvik, M., Rosnes, J.T., Kleiberg, G.H., 2003. Effect of modified atmosphere packaging and superchilled storage on the microbial and sensory quality of Atlantic salmon (Salmo salar) fillets. J. Food Sci. 68, 1467–1472. https://doi.org/10.1111/j.1365-2621.2003.tb09668.x.

Statistics Greenland, 2019. Total landings of fish and shellfish by time, municipality, species, unit and month. http://bank.stat.gl/pxweb/en/Greenland/Greenland__FI__FI10/FIX001.px/ (accessed 8 September 2019).

Statistics Greenland, 2018. Greenland in Figures 2018. http://www.stat.gl/publ/da/GF/2018/pdf/Greenland%20in%20Figures%202018.pdf (accessed 26 September 2019).

Storr-Paulsen, M., Wieland, K., 2006. Den Atlantiske torsk i de Grønlandske farvande 2005. http://www.natur.gl/fileadmin/user_files/Dokumenter/Tekniske_rapporter/Teknisk_Rapport_63_t orsk.pdf (accessed 26 September 2019)

Weisburg, W.G., Barns, S.M., Pelletier, D.A., Lane, D.J., 1991. 16S Ribosomal DNA Amplification for Phylogenetic Study. J. Bacteriol. 173, 697–703. 10.1128/jb.173.2.697-703.1991.

Wilson, B., Danilowicz, B.S., Meijer, W.G., 2008. The diversity of bacterial communities associated with Atlantic cod Gadus morhua. Microb. Ecol. 55, 425–434. https://doi.org/10.1007/s00248-007-9288-

101

102

103

104

Paper II

Jonas Steenholdt Sørensen, Niels Bøknæs, Ole Mejlholm, Karsten Heia, Paw Dalgaard and Flemming Jessen

Short-term capture-based aquaculture of Atlantic cod (Gadus morhua L.) generates good

physicochemical properties and high sensory quality during frozen storage.

Manuscript prepared to be submitted to Innovative Food Science and Emerging Technologies

105

Short-term capture-based aquaculture of Atlantic cod (Gadus morhua L.) generates good physicochemical properties and high sensory quality during frozen storage.

Jonas Steenholdt Sørensen 1, 2, Niels Bøknæs2, Ole Mejlholm2, Karsten Heia3, Paw Dalgaard1 and Flemming

Jessen1

1National Food Institute (DTU Food), Technical University of Denmark, Kgs. Lyngby, Denmark

2Royal Greenland Seafood A/S, Svenstrup J, Denmark

3Nofima, Tromsø, Norway

* Corresponding author: Food Microbiology and Hygiene, National Food Institute, Technical University of

Denmark, Kemi torvet, Building 202, 2800, Kgs. Lyngby, Denmark. E-mail: [email protected]

106

Abstract

The annual global catch of Atlantic cod (Gadus morhua L.) is above 1,000,000 tons, and a large proportion is preserved and sold as frozen products. In Greenland, a new innovative fishing method centred on capture-based aquaculture has been developed. To compare the new fishing and processing method with the conventional method, the texture, colour, water holding capacity, salt soluble proteins and sensory properties of fillets were evaluated during 12 months of frozen storage. High quality life (HQL) and practical storage life (PSL) were determined for each fisting and processing method based on water holding capacity and the salt soluble protein fraction. Capture-based aquaculture fishing and related processing increased frozen storage durability regarding texture attributes, blood content and sensory properties compared to the conventional method. Lowering the storage temperature to -40 °C increase the HQL to >12 months compared to a HQL of 4-6 months at -20 °C.

Keywords: Durability study, salt soluble protein, water holding capacity, texture profile analysis (TPA), Greenland, shelf-life

107

1. Introduction

Atlantic cod (Gadus morhua L.) is one of the most important commercially fished species for more than a century. The annual global catch of wild cod has been above 1,000,000 tons, except for the years 2000-2010 (FAO, 2020). In the period 2006 to 2012, attempts were made to farm cod and aquaculture production reached 22,728 tons in 2009 but since collapsed (FAO, 2020; Henriksen, Heide, Hansen, & Mortensen, 2018). Capture-based aquaculture (CBA) is a new fishing method for Atlantic cod fishery that keeps the fish alive after capture and during transport to holding nets adjacent to a process plant. The live fish are then available for slaughter in a non-stressed state and quantities matching the production capacity can be taken from the holding nets. In addition, CBA combined with holding live fish in net enclosures give the possibility for rapid processing and packaging, both for frozen or wholesale distribution of fresh cod (Sønvisen & Standal, 2019). In Norway, adult cod is captured during the winter and kept alive for processing in the late spring and summer. The main driver is to reduce seasonality of the landings to stabilise prices throughout the year (Hermansen & Eide, 2013). Furthermore, rested and rapidly processed cod from CBA has a long shelf-life and limited microbial growth when superchilled at - 1.7 °C (Sørensen, Bøknæs, Mejlholm, & Dalgaard, 2020) and after frozen storage the thawed CBA cod has a low drip loss and long chilled shelf-life when stored in modified atmosphere packaging (Sørensen, Ørnfeld-Jensen, Bøknæs, Mejlholm, Jessen, & Dalgaard, 2020b). With CBA, cod is not typically fed after the catch, and some degree of starvation takes place. A starvation period of no longer than 54 days has been recommended (Ageeva, Jobling, Olsen, & Esaiassen, 2017). For salmon and carp, with an empty gut, the metabolic and physical activity is reduced (Jobling, 1981), which leads to reduced stress during handing and pumping and thereby reduced mortality (the Farm Animal Welfare Committee (FAWC), 2014; Waagbø, Jørgensen, Timmerhaus, Breck, & Olsvik, 2017). Within the first 54 days of starvation, no significant sensory or physicochemical changes occur in the fillet (Ageeva, Olsen, Joensen, & Esaiassen, 2018).

Greenland and Newfoundland have a long history of cod fishing with large quantities of fish and long distances to the primary markets in Europe. Traditionally salting and later freezing were used to preserve cod during the long transportation from the Northwest Atlantic to Europe (Cole, 1990). Freezing and frozen storage of cod below -10 °C inhibit the microbial spoilage (Ratkowsky, Olley, McMeekin, & Ball, 1982), while the autolytic and chemical reactions may occur at slow rates for Atlantic cod (Rehbein, 1988). Protein denaturation and dehydration are the main spoilage reactions that determine the frozen storage shelf-life when the temperature is below -10 °C (Torry Research Station, 2001). Protein changes during

108 frozen storage depend on the frozen storage temperature (Burgaard & Jørgensen, 2010) and temperature variability during frozen storage (LeBlanc, LeBlanc, & Blum, 1988). Temperatures below -28 °C can prolong the frozen shelf-life to 4 years before white fish become inedible compared to 15 months when stored at -21 °C (Torry Research Station, 2001). Furthermore, storage at -5 to -15 °C resulted in markedly reduced shelf-life due to enzymatic formaldehyde formation from trimethylamine oxide (TMAO) (LeBlanc & LeBlanc, 1988). However, we have found no previous studies of storage durability of CBA produced cod nor how frozen storage temperature influences the shelf-life.

The cod fishery in West Greenland occurs primary inshore, with small vessels setting pound nets during the period from May to November when the cod follow their primary feedstock of capelin close to the shore (Statistics Greenland, 2019). On-board freezing facilities are not available on these small day- boats, and they bring their catch to a local processing and freezing plant. Different challenges are related to the processing and distribution of high-quality cod products from this fishery. Firstly, the time from catch to bleeding and slaughter may be longer than optimal due to limited capacity on the small vessels. Secondly, in high season, quantities of iced cod delivered to processing plants may be larger than what their filleting and freezing capacity can handle leading to storage of whole gutted cod in ice for several days prior to processing. Consequently, frozen whole cod from this fishery can have soft texture after thawing resulting in food and profit losses (Personal communication with Royal Greenland).

The objective of the present study was to evaluate the effect on frozen storage shelf-life at -20° C of cod fillets from a newly developed process involving CBA of cod and to compare with the conventional processing in Greenland. A durability study of frozen storage with cod fillets was performed to determine shelf-life based on sensory changes and changes of colour, texture and other physicochemical characteristics and to compare the two types of cod processing. Additionally, the effect of frozen storage at -20, -40 and -80 °C on shelf-life and quality attributes for cod fillets from CBA was investigated.

2. Materials and methods

2.1 Materials

Atlantic cod (Gadus morhua L.) were conventionally caught with pond nets close to the shoreline of West Greenland. This fishing with pond nets is a passive fishing method, as the fish are trapped in net enclosures. From the net enclosure, the cod were processed either by a conventional method (2.1.1) or by a new short term capture-based aquaculture (CBA) processing method (2.1.2).

109

2.1.1 Conventional processing method

Fish caught in poundnets were brought to the surface by pulling the net into the fishing vessel and the cod was trapped in a confined space. From there, the fish were then hand-picked, throat cut, eviscerated, and bleed in the air at the bottom of the small open surface vessels. At the end of this processing, the cod were packed in fish boxes and covered with ice (Fig. 1A). From the fjords, the fish were then transported to a nearby processing plant. At the processing plant, the two first days after landing of the fish, no further processing could be performed as the rigor mortis process had to end prior to machine filleting. The post-rigor mortis cod were processed by machine decapitation (Baader 417, Lübeck, Germany) and filleted (Baader 192, Lübeck, Germany). The fillets were hand-trimmed and individually quick frozen (IQF) as skin-on fillet using a gyro freezer (Carnitech, Støvring, Denmark) with a 10 % (w/w) protective glaze layer of water.

2.1.2 Short term capture-based aquaculture (CBA) processing method

After being caught by poundnets (Fig. 1A), the fish was gently and alive lead into a transportable net enclosure without leaving the water at any point and then kept alive for two to three weeks to reduce their stomach content (Fig. 1B). The starved cod was pumped up to a well-boat, and all cod smaller than 42 cm were released directly to the fjord (42 cm was the minimum cod landing size in Greenland in 2018) (Fig. 1C). The well-boat transported the cod to the processing plant, where the fish was pump into a large ocean net enclosure adjacent to the processing plant (Fig. 1D). At this step, the cod was kept for a minimum of 12 hours to recover from transportation and handling stress. The rested cod was pumped into the processing plant at a rate corresponding to the workforce and production capacity of a given day and then stunned immediately after exposure to the air. The stunning was performed by electricity (20 V, 400 Hz for 5 seconds) and the cod was slaughtered by decapitation (Baader 417, Lübeck, Germany). The viscera were removed by hand and the headed and gutted cod was bleed for 30 min. in circulating water at 3 - 5 °C. The fish was then placed 30 - 60 minutes in ice water to reduce the muscle tension creative by ATP activation of the action potential (Hibberd & Trentham, 1986), but not long enough for the rigor- mortis stiffness to begin. The cod was machine filleted (Baader 192, Lubeck; Germany) and hand-trimmed followed by freezing as IQF fillets in the same way as described for the conventional method (See 2.1.1).

110

Fig. 1. Illustration of the capture-based aquaculture fishery method. The Atlantic cod was captured inshore by poundnets (A) similar to the conventional fishery. The cod was transferred without leaving the water to net enclosures and hold for two to three weeks without feeding (B). The cod, now without gastrointestinal content, was pumped into a well-boat (C) and transported alive to a processing plant, where the cod kept into large net enclosures until processed (D).

2.1.3 Transport to DTU Food

In early June 2018, fish caught by the conventional (2.1.1) and CBA (2.1.2) processing methods, were transported by ship from Greenland to Denmark. The cod arrived in Denmark one month after slaughter and processing. During transport from Greenland to DTU Food, the temperature of the cod fillets was recorded regularly by loggers (TinyTag Plus, Gemini Data Loggers Ltd., Chichester, UK) and the average storage temperature was -21.2 °C ± 0.1 (SD), and it never exceeded -16.2 °C.

111

2.2 Experimental plan and analyses

2.2.1 Experimental plan

From arrival at DTU Food, the conventionally produced fillets were stored at -20 °C (i). Fillets from CBA were randomly split into three groups, that were stored at -20 °C (ii), -40 °C (iii) and -80 °C (iv), respectively. After storage for three, six, nine and 12 months, eight cod fillets from each treatment (i-iv) were analysed. The fillets were thawed overnight in a 2 °C cooling room, and the thawed fillets were cut into loin (Fig. 2A) and tail parts (Fig. 2B). The loin parts were studied by image analyses (2.2.2), texture (2.2.3), water holding capacity (2.2.4), and salt soluble proteins (2.2.4) in that order. The tail parts were studied by image analyses (2.2.2).

Using four different fillets from each treatment (i, ii, iii and iv) sensory profiling (2.2.5) was conducted by Royal Greenland Seafood A/S, after three, six, nine, 12, and 15 months of frozen storage.

Fig 2. Photo of cod fillet showing the position of the sampling area for hyperspectral imaging (A and B), texture analysing (A), water holding capacity (A), salt soluble proteins (A).

2.2.2 Hyperspectral imaging (HSI)

HSI was performed with a VideometerLab 2 (VideometerLab 2, version 2.13.54, Videometer A/S, Denmark) installed with 18 channels from 375 nm to 1050 nm. Before the experiment, a series of HSI of cod from different production dates were taken to optimise the exposure time of each channel and saved

112 as the systematic light setting for the colour analyse. The system was calibrated radiometrically using a diffuse white, dark and geometrical disc before each sampling point. The samples (loin and tail, Fig. 2A and 2B) were placed in a plastic petri dish, and the HSI picture was taken, using the software provided by Videometer. The HSI spectra were transformed to daylight50 CIELab colour space and processed with the software accommodating the Videometer (VideometerLab 2, version 2.13.54, Videometer A/S, Denmark) to obtain the average L, a* and b* values for each fillet using MATLAB (R2018b, The Mathworks Inc., MA, USA).

2.2.2.1 Blood content by HSI analysis

The quantification of blood in the fillets was analysed using an indirect method for the Videometer HSI from three and six months of storage. The 18 channel HSI was converted from a reflectance spectrum into an absorbance spectrum by using Eq. 1.

1 퐴푏푠표푟푏푎푛푐푒 = 푙표푔 ( ⁄푅푒푓푙푒푐푡푎푛푐푒) Eq. 1

The obtained absorbance spectra were analysed using Nofima’s software (IDL 8.71., Harris Geospatial Solution Inc., Colorado, USA) for quantitative estimation of blood in the surface of cod fillets based on a constrained spectral un-mixing approach (Skjelvareid, Heia, Olsen, & Stormo, 2017). In addition to HSI obtained from the Videometer, four whole fillets (stored for nine months) for each treatment were transported from DTU Food to Nofima (Tromsø, Norway). A high-resolution HSI were taken using a VNIR- 1024 imaging spectrograph (Norsk Elektro Optikk, Skedsmokorset, Norway), operated in a push broom imaging mode. The imaging was performed in an interactance setup, that means that the fillets are illuminated with two focussed light lines. The HSI measured between the two light lines as the fillets were moving on a conveyor belt (speed 400 mm/s). In this way the light had to travel some distance inside the fillet before being recorded, giving an image of both the surface and the inside of the fillet. The spatial and spectral resolution was 0.28 x 0.56 mm and 2.7 nm in the range of 410-990nm. Using a Teflon reference the HSI data can be transformed into absorbance and run through the constrained spectral unmixing analysis and the blood abundancies were recalculated into chemical blood concentration. (Skjelvareid et al., 2017).

2.2.3 Texture profile analysis

113

The loin part previously used for HSI was cut into two 4x4x2cm muscle pieces (MP), and the texture profile analysis (TPA) was determined in duplicates for each fillet. The texture of the MP was measured using the TA.XTplus Texture Analyser (Stable Micro System, Surry, England) equipped with a 30 kg load cell. A P/75 compression platen compressed the MC to 50 % of the total MC height, with a five mm/s speed. Texture curves were recorded with a resolution of 200 points/s and analysed by the instrumental software (Texture Exponent, Version 6.1.15, Stable Micro System, Surry, England). From the time-force graphs, the TPA parameters, hardness and springiness were calculated (Bourne, 2002).

2.2.4 Water holding capacity and salt soluble proteins

The loin used for TPA were blended for 5 seconds using a kitchen table blender, and the blended loin was used to determine dry matter (DM), water holding capacity (WHC) and salt soluble proteins (SSP). DM was measured as the percentage of remaining mass after 24 hours drying at 105 °C in relation to the two grams wet weight of loin and performed in duplicates. WHC was measured in quadrupled, by weighted two gram of the blended fillet into a cylindrical tube with a filter bottom that allowed the liquid to pass through. The cylindrical tube was placed on top of 40 gram of glass marbles in a centrifuge tube as illustrated in Eide, Børresen, & Strøm (1982) and centrifuged as described in Burgaard & Jørgensen (2010). The WHC was expressed as the water retention from the original water content and was calculated by Eq. 2. Where ∆r is the mass difference of the sample before and after centrifugation.

100−퐷푀−Δr 푊퐻퐶 % = × 100 Eq. 2 100−퐷푀

SSP was quantified by determining the fraction of proteins dissolved in a salt solution to the total protein in the muscle (Dyer, French, & Snow, 1950). The proteins were dissolved in a salt solution, 1.75 g of blended loin muscle was mixed for 3 minutes (Ultraturrex, 8,500 rpm) with a salt solution prepared in a measurement cylinder. In a 50 mL cylinder, 30 mL of 0.86M NaCl and 0.6M NaHCO3, cooled with ice, was topped to the 35 mL mark with flake ice. The mix was centrifuged at 10,000 g for 20 minutes at 4 °C (Sigma 4-16KS, Sigma Laborzentrifugen GmbH, Osterode am Harz, Germany). The supernatant of the sample was filtered through a metal tea filter and stored at -20 °C until further analysis. The protein, both the total protein content of the fillet and the protein content in the salt solution, was measured as the nitrogen value determined by the Dumas combustion procedure (Etheridge, Pesti, & Foster, 1998). The parameters of the combustion profiles were optimised for the nitrogen value of the samples. For total protein in the muscle, 0.2 g sample was combusted with 300

114 mL/min O2 for 120 seconds and the salt solution, 2.5 g sample was combusted with 150 mL/min O2 for 90 seconds and recorded with Rapid MAXn (Elementar, V.1.0.14, elementar Analysensteme GmbH, Langenselbold, Germany). The conversion rate between the nitrogen values and proteins was set to a factor of 6.25.

2.2.5 Sensory profiling analysis

Quantitative descriptive analysis (QDA) as described by Stone, Sidel, Oliver, Woolsey, & Singleton, (1974) was used to evaluate the sensory profile of cod fillets from differents treatments and frozen storage times. Eight to thirteen tested and trained assessors from Royal Greenland Seafood A/S (RG), with extensive knowledge of seafood, were the base for the sensory panel. Fourteen attributes were previously chosen at RG to describe the sensory profile of cod. These included, appearance (gaping, whiteness, frayed, firmness) and odour (fishy, ocean/seaweed) for raw cod . The cooked cod was assessed using odour (warm milky), texture (juiciness, toughness, flakiness) and flavour (sweet, metallic, fishiness, bitter) attributes. For each session, one raw and one cooked fillet of each treatment (i, ii, iii and iv) was evaluated and scored on an unstructured scale with the size of 100 mm.

The cod fillets were shipped from DTU Food to RG in Styrofoam boxes cooled with dry ice. At RG, the fillets were placed in a stainless steel container (1/1 gastronorm) and thawed overnight in a 2 °C refrigerator. The cooked fillets were heated with the skin side up in a preheated oven at 80 °C with 60% air circulation and steam for 8 minutes (SCC 61, Rational, Heerbrugg, Germany). The raw and cooked fillets were presented to the panel on stainless steel plates, and each product was given a random 3-digit number.

2.3 Data analysis

To determine the blood content of fillets from the 18 channels Videometer HSI, these were transformed into a seven-channel HSI including three channels with information of blood concentration by an IDL widget application (IDL, Harris Geospatial Solution Inc., Colorado, USA). The region of interest (ROI) was picked by a MATLAB script (MATLAB R2018b, The MathWorks Inc., Massachusetts, USA) and RStudio was used for statistical analysis (RStudio Version 1.2.5001, Massachusetts, USA) for both HSI obtain from the Videometer and Nofima’s platform. Linear models (lm) in combination with ANOVA were used to check differences between frozen storage time and treatment, post hoc analysis in the form of Tukey to analyse which treatment was significantly different. The analysis for the development of WHC,

115

SSP, DM, texture and colour during storage were performed by using GraphPad Prism 8.3.0 (GraphPad Software, California, USA), with simple linear regression and focus on if the slope was significantly different from zero. In case of a slope significantly different from zero, the mean of the values was analysed with two-way ANOVA. Principle component analysis (PCA) of the sensory data was performed with PanelCheck v1.4.2 (PanelCheck, 2012) and the raw PCA data was exported to Microsoft Excel 2016 to produce the graphs (Microsoft Corp., Redmond, WA, USA).

3. Results

3.1 Sensory profiling

The overall variation in the sensory quality of the cod was shown in Fig. 3A as the scores of the first two principal components (PCs) from a principal component analysis (PCA) on the data from the sensory profiles. The main variation along with the PC 1 (70.6 % of the explained variance) showed that cod from the conventional processing after 12 and 15 months of storage was positioned in the right side of the scores plot separated from all the other treatments and sampling times (Fig. 3A). The scores plot in Fig. 3A also shows that in the PC 2 (7.0 % of explained variance) direction the CBA cod stored for 12 and 15 months at -20 °C were positioned apart from (in the bottom of the plot) the CBA cod stored at -40 ○C and -80 ○C for 12 and 15 months. From the corresponding loading plot (Fig. 3B), it can be seen that the sensory attributes causing the variation in the PC 1 positive direction were mainly the texture attribute toughness and the flavour attribute metallic. In contrast, the negative direction was primarily associated with the texture attributes flakiness and juicy and the flavour attribute sweet. In the PC 2 direction, the attributes responsible for most of the variation between samples were the appearance attribute firmness of the raw cod (negative direction) and the sweet flavour of the cooked cod (positive direction) (Fig. 3B).

116

117

Fig. 3. Principal component analysis of the sensory profile scores for the two fishing methods, A (scores), conventional storage at -20 °C (i) and capture-based aquaculture (CBA) at -20 °C (ii), -40 °C (iii) and -80 °C (iv) for a storage period of 15 months (m). Circles highlight the samples storage after 12 and 15 months for each treatment. Loadings (B) based on the 14 words of the sensory profile, for the raw fillet appearance (A) (gaping, whiteness, frayed, firmness), odour (O) (fishy, ocean/seaweed) and for the cooked fillet (O) (warm milky), texture (T) (juiciness, toughness, flakiness) and flavour (F) (sweet, metallic, fishiness, bitter).

Frozen shelf-life determined by sensory changes was defined in two ways, high-quality life (HQL) and practical storage life (PSL), with HQL as the time to reached a significantly different from the control sample and PSL as the time when the product was not acceptable (Bøgh-Sørensen, 2006). The just noticeable difference between the control sample (CBA cod at -80 °C) and the conventionally processed cod, the CBA cod at -20 °C and the CBA cod at -40 °C gave a HQL of six to nine months for the conventionally processed cod and the CBA cod stored at -20 °C. Storage at -40 °C extended the HQL for CBA cod to >15 months (Table 1).

After 15 months of storage, the scores for flakiness and juicy were significantly lower in the conventionally processed cod compared to CBA cod (p < 0.05) (Fig. 4A and 4B). Contrary, toughness scores, as well as metallic taste scores, were significantly higher for conventional cod compared to CBA cod after 12 and 15 months (p < 0.01) of storage (Fig. 4C and 4D). The sweetness score decreased both in the conventional cod (p < 0.01) and in the CBA cod (p < 0.05) during storage at -20 °C (Fig. 4E). The firmness score in the conventional cod was lower than in the CBA cod (Fig. 4F) during most of the storage time, although this was exclusively shown to be significant (p < 0.05) after 3, 12 and 15 months of storage.

118

100 100 A B 80 80

60 60

40 40

20 20

0 0 0 3 6 9 12 15 0 3 6 9 12 15

100 50 )

0 C D

0 1

- 80 40

(

r 60 30

c s

40 20

r o

s 20 10

n e S 0 0 0 3 6 9 12 15 0 3 6 9 12 15

100 100 E F 80 80

60 60

40 40

20 20

0 0 0 3 6 9 12 15 0 3 6 9 12 15 Storage time (months)

Fig. 4. Sensory scores for three texture parameters of the cooked cod, flakiness (A), juicy (B), toughness (C), two taste parameters of the cooked cod, metallic (D), sweet (E) and the appearance parameter, firmness (F), for the raw cod. Conventional cod storage at -20 °C was indicated by ● and capture-based aquaculture storage at -20 °C (∆), -40 °C (■) and -80 °C (◊). Symbols and error bars indicate Avg. ± SEM.

119

3.2 Blood content

Frozen storage for three and six months did not affect the blood content in the cod (Conventional and CBA) (p = 0.79) and it was assumed that the blood content was independent of storage period, henceforth the HSI for nine and 12 months was not analysed for blood. For the CBA cod, the storage temperature did not affect the blood content (p = 0.93). All the obtained HSI spectra taken from the conventionally processed cod after three and six months of storage were pooled (n = 16), and so were the spectra from the CBA processed cod, independent of storage temperature (n = 48) and the two methods were compared. The comparison of the two processing methods showed a significantly lower blood content in the CBA cod compared to the conventionally produced cod, both for the loin and the tail regions of the fillet (p < 0.001). For both the conventionally and the CBA produced cod, the tail part contained significantly more blood than the loin part (p < 0.05).

HSI of whole fillets, acquired with high-resolution VNIR, showed an even more clear difference in blood content between cod from the two processing methods (p < 0.0001, n = 16 with n = 4 for conventional and n = 12 for CBA). The average blood content of conventionally processed cod was 0.09 mg haemoglobin/g muscle and 0.07 mg haemoglobin/g muscle for CBA cod. To illustrate the difference between the conventional and the CBA fishing method, an analysed HSI spectrum of a fillet from each method is presented in Fig. 5.

120

A B

Fig. 5. Blood concentration images of one filet from the conventional fishing method (A) and from the capture-based aquaculture fishing method (B). The images are representative of the 16 fillets studied.

3.3 Physiochemical changes during storage

The WHC changed during storage as an effect of storage temperature, after nine months of storage, both the conventionally produced and the CBA cod stored at -20 °C had significantly lower WHC than CBA cod stored at -40 °C and -80 °C (p < 0.05) (Fig. 6A) and remained lower after 12 months of storage (p < 0.0001). No significant effect was observed between the two production methods, and no changes were found for the CBA cod when stored at the lower temperatures of -40 °C and -80 °C (Fig. 6A). Similar to WHC, the SSP was exclusively affected by storage temperature and not by the processing method. The SSP for CBA produced cod stored at -40 °C and -80 °C did not change during storage and were after 12 months significantly (p < 0.05) higher than in cod stored at -20 °C (Fig. 6B).

121

100 100

) A B

)

(

% (

y 90

n i

c 80

o a

80 r

n i

b 60 d

70 u

e a

t 60 a

S 40 W 50 0 3 6 9 12 0 3 6 9 12 Storage period (months)

Fig. 6. Water holding capacity (A) and salt soluble proteins (B) during 12 months of frozen storage for conventionally processed cod (●) stored at -20 °C and capture-based aquaculture cod stored at -20 °C (∆), -40 °C (■) and -80 °C (◊). Symbols and error bars indicate Avg. ± SEM.

3.4 Texture changes during storage

Comparing the two processes at -20 °C, showed that CBA had an increased hardness of the loin and was significantly more firm after three and twelve months (Table 2). The conventional process resulted in a texture of significant higher springiness throughout the storage period. Comparing the hardness of the loin stored at different temperatures did not show any significant difference for CBA cod at the same frozen storage time. However, the CBA cod stored at -20 °C showed a significant increase in hardness during the storage period (Table 2). The springiness remained stable when stored at -20 °C, whereas, at lower temperatures, the springiness significantly decreased during the frozen storage time. The resilience of the cod did not change during storage, and there was no significant difference between the treatments (data not shown).

122

3.5 Colour changes during storage

The colour of the loin was influenced by the production method. The L and a* value of CBA cod remained stable during the twelve months of storage, while the a* value of the conventionally processed cod showed a significant increase (Table 2). Storage of cod at -40 °C or -80 °C resulted in stable b* values during the entire storage period, whereas at -20 °C the b* value increased significantly during the storage.

4. Discussion

The two production methods included some fundamentally different processing steps of the frozen cod. In the CBA production, the cod was slaughtered in a rested state after a minimum of stress in the antemortem phase, and the fillets were frozen pre-rigor mortis. In the conventional production method, the cod was more stressed before slaughter, and the fillets were cut and frozen post-rigor mortis.

In the present study, the conventional method resulted in a softer fillet, with lower hardness scores, throughout the entire frozen storage period (Table 2). The harder texture found in the CBA cod

123 could be caused by several factor, including starvation, time of filleting and stress. Olsson, Gundersen, & Esaiassen (2006) showed that prolonging the starvation period from two to three weeks before slaughtering increased the hardness (shear force) of cod fillets after storage on ice for two weeks. The CBA cod from Greenland starved between two to four weeks in comparison to the conventional method, including no starvation of the cod.

In the conventional method, the cod goes through the rigor mortis process before filleting. With the muscle structure physically attached to the vertebrate frame of the fish during rigor mortis it results in more gaping compared to pre-rigor mortis filleted cod (Kristoffersen, Tobiassen, Steinsund, & Olsen, 2006). The increased gaping observed by Kristoffersen et al. (2006) indicates that the connective tissue was broken and this maybe a contributing factor to the softer fillet texture of the conventional method. Mørkøre, Hansen, & Rørvik, (2006) showed that pre-rigor mortis filleting results in better sensory texture. The texture differences between stressed and rested cod had previously been reported, and it was found that the stressed cod (corresponding to the conventional cod in the present study, Table 2) had a softer texture (Hultmann, Phu, Tobiassen, Aas-Hansen, & Rustad, 2012) compared to rested cod. However, others have shown no significant textural different between the rested and stressed cod (Ulf Erikson, Digre, & Misimi, 2011).

The differences in discolouration seen in the present study could be explained by the lower blood content found in fillets from CBA cod compare to the traditionally processed cod. It has been hypothesised that such a low blood content could be due to the combination of an optimal stunning practice, live storage and proper bleeding of the cod, three factors that have been addressed in the CBA production chain (Fig. 1). The stunning by electricity in the CBA method (Fig. 1D) has been investigated previously and no adverse effect on sensory, pH and texture quality when compared to anaesthesia was found, which must be assumed to be the least stress-induced stunning method (Erikson et al., 2012). The increased redness and yellowness of fillets from the conventional fishing method (Table 2) might be related to induced stress antemortem. Studies have documented stress-induced discolouration, such as increased yellowness and redness in cod fillets (Hultmann et al., 2012; Jørpeland, Imsland, Stien, Bleie, & Roth, 2015). A resting period after the stress has a positive influence on discolouration, and it has been shown that six hours live storage of cod captured by trawl was sufficient to avoid the stress-induced discolouration (Olsen, Tobiassen, Akse, Evensen, & Midling, 2013). In this context, the importance of bleeding was underlined, although the type of bleeding method was unimportant (Olsen et al., 2014).

124

The storage temperature and duration of frozen storage has an impact on the enzyme activity in the muscle and on the ability of the proteins to dissolve in a NaCl solution. The Ca2+ATPase activity was lost more rapid at -20 °C compared to -30 °C (Godiksen, Hyldig, & Jessen, 2003) and it is well known that the SSP fraction during higher frozen storage temperatures (e.g. -10 °C, -15 °C) decreases fast. Lowering of the temperature to -30 °C reduces this loss in protein solubility dramatically (Badii & Howell, 2002a; Careche, Del Mazo, Torrejón, & Tejada, 1998; Del Mazo, Torrejón, Careche, & Tejada, 1999; LeBlanc & LeBlanc, 1992). There is no common agreement between sensory shelf-life and the SSP faction and to compare studies, we suggest a high quality life (HQL) and a practical storage life (PSL) criteria of <70 % SSP or 0.3 mg SSP g-1 fish muscle and <60 % SSP or 0.2 mg SSP g-1 fish muscle, respectively. Applying this criteria, gave a PSL at -30 °C or below of more than 12 months in the present study (Table 1) and the period was supported by various studies with cod (Badii & Howell, 2002b, 2002a; LeBlanc & LeBlanc, 1992), hake (Del Mazo et al., 1999) and haddock (Badii & Howell, 2002a). Rising the storage temperature to -20 °C decreased the PSL to 10 months (Fig. 6B) which is similar to results from previous studies on hake and cod (Del Mazo et al., 1999; LeBlanc & LeBlanc, 1992). A further increase of the temperature to -10 °C resulted in a PSL around one month (Badii & Howell, 2002b, 2002a).

The negative impact on the proteins during frozen storage, also affects the WHC of the muscle proteins. The freezing storage temperature and time have previously been shown to influence the WHC. Cod from the Barents Sea showed a significantly lower WHC after 12 months of storage at -20 °C compared to -30 ° C (Bøknæs et al., 2002). Bøknæs, Guldager, Østerberg, & Nielsen, (2001) suggested a criterion for HQL of 70 % WHC and a PLS of 60 % WHC. In the present study, a HQL of 65 % WHC was chosen as the initial WHC was approximately 70 % (Fig. 6A) and this gave a HQL of 4-6 months and a PSL of 8-10 months when stored at -20 °C. Lowering the temperature to -40 °C increased both the HQL and PSL to >12 months. A similar trend was seen by Burgaard & Jørgensen, (2010) with a HQL/PSL of 4/10 months at -20 °C that increased to >18 months for both HQL and PSL at a storage temperature of -30 °C. The change of temperature from -20 °C to -30 °C as the critical range was supported by stabilisation of water loss when lowering the storage temperature to -30 °C as shown by (Schubring, 2005).

The frozen HQL shelf-life of cod as determined by sensory changes was six to nine months for cod at -20 °C whereas the measured values for WHC and SSP, used as indices of spoilage, indicate a HQL of five to six months (Table 1). We suggest the sensory changes for the metallic, sweet and bitter taste attributes defined the PLS and resulted in a PLS exclusively depending on the frozen storage temperature. Storage at -20 °C gave a PLS of 12 – 15 months and >15 months by lowering the temperature (Table 1).

125

Further studies are needed on the correlation between sensory changes and levels of WHC and SSP in frozen cod to give a better criterion for HQL and PSL.

To accommodate the market demand for more convenience products, refreshed cod fillet (frozen and thawed) can supply the market with ready-to-use fish product. The CBA method would be recommended as filleting in the pre-rigor mortis phase gives a higher sensory score for the refreshed cod after five days chilled storage regardless of the frozen storage period (Martinsdottir & Magnusson, 2001). As to the use of lower freezing storage temperatures in order to preserve the protein structure and thereby the textural quality this could have an adverse effect on the refreshed product as it has been shown that keeping cod at -30 °C, also protect the microbiota of the fish (Bøknæs et al. 2002; Emborg, Laursen, Rathjen, & Dalgaard, 2002). Sørensen et al. (2020b) found a 15 days safe shelf-life of chilled MAP CBA cod, after frozen storage at -20 °C for five months. Accurate recommendation to the industry of an optimal frozen storage period and temperature for refreshed cod, require further studies with different combinations of freezing periods, temperatures and sensory scores at chilled condition after thawing.

5. Conclusion

The CBA method results in a product with better durability in the sensory and colour quality. Initially and throughout the storage period, the CBA method gave a more firm fillet, and the better bleeding procedure resulted in a whiter fillet with a lower blood content. The conventional method gave a lower texture score at the end of the frozen storage, and they were rated lower for flakiness, juicy for the cooked fillet, and higher for a metallic taste. The frozen storage temperature had an impact on the frozen shelf-life. Storage at -20 °C gave a HQL of six to nine months, determined by sensory evaluation and supported by WHC and SSP. Lowering the storage temperature to -40 °C, extended the HQL to >15 months.

Acknowledgement

Innovation Fund Denmark (grant no. 5189-00175B) funded this research. We thank employees at the process plant in Maniitsoq, Greenland for their assistance, the local fisherman to supply us with the material and the sensory panellist at Royal Greenland for their time to evaluated cod samples, with a special thanks to Jan Zoutenbier for all the practical work associated with the sensory evaluation.

126

References

Ageeva, T. N., Jobling, M., Olsen, R. L., & Esaiassen, M. (2017). Gender-specific responses of mature Atlantic cod (Gadus morhua L.) to feed deprivation. Fisheries Research, 188, 95–99. https://doi.org/10.1016/j.fishres.2016.12.010

Ageeva, T. N., Olsen, R. L., Joensen, S., & Esaiassen, M. (2018). Quality aspects of fillet, loin and tail products made from live-stored feed-deprived Atlantic cod (Gadus morhua L.) at different times post mortem. Lwt, 97(9291), 656–661. https://doi.org/10.1016/j.lwt.2018.06.031

Badii, F., & Howell, N. K. (2002a). A comparison of biochemical changes in cod (Gadus morhua) and haddock (Melanogrammus aeglefinus) fillets during frozen storage. Journal of the Science of Food and Agriculture, 82(1), 87–97. https://doi.org/10.1002/jsfa.998

Badii, F., & Howell, N. K. (2002b). Effect of antioxidants, citrate, and cryoprotectants on protein denaturation and texture of frozen cod (Gadus morhua). Journal of Agricultural and Food Chemistry, 50(7), 2053–2061. https://doi.org/10.1021/jf010824f

Bøgh-Sørensen, L. (2006). Recommendations for the Processing and Handling of Frozen Foods. (4th ed.). France: International Institute of Refrigeration (IIR).

Bøknæs, N., Guldager, H. S., Østerberg, C., & Nielsen, J. (2001). Production of high quality frozen cod ( Gadus morhua ) fillets and portions on a freezer trawler. Journal of Aquatic Food Product Technology, 10, 33–47.

Bøknæs, N., Jensen, K. N., Guldager, H. S., Østerberg, C., Nielsen, J., & Dalgaard, P. (2002). Thawed chilled Barents Sea cod fillets in modified atmosphere packaging-application of multivariate data analysis to select key parameters in good manufacturing practice. LWT - Food Science and Technology, 35(5), 436–443. https://doi.org/8

Bourne, M. (2002). Food texture and viscosity: concept and measurement. Retrieved from https://books.google.com/books?hl=en&lr=&id=S2HNnvSOuf8C&oi=fnd&pg=PP2&dq=Bourne,+M. C.,+2002.+Principles+of+objective+texture+measurement.+Food+texture+viscosity:+concept+meas urement.+2nd+Ed.&ots=u_NNokXOsl&sig=H8zG7IXJ43J-ZIAm5AGZak3iw-E

Burgaard, M. G., & Jørgensen, B. M. (2010). Effect of temperature on quality-related changes in cod (Gadus morhua) during short- and long-term frozen storage. Journal of Aquatic Food Product

127

Technology, 19(3–4), 249–263. https://doi.org/10.1080/10498850.2010.510944

Careche, M., Del Mazo, M. L., Torrejón, P., & Tejada, M. (1998). Importance of Frozen Storage Temperature in the Type of Aggregation of Myofibrillar Proteins in Cod (Gadus morhua) Fillets. Journal of Agricultural and Food Chemistry, 46(4), 1539–1546. https://doi.org/10.1021/jf970841y

Cole, S. C. (1990). Cod, God, Country and Family. Maritime Studies, 3(1), 1–29.

Del Mazo, M. L., Torrejón, P., Careche, M., & Tejada, M. (1999). Characteristics of the salt-soluble fraction of hake (Merluccius merluccius) fillets stored at -20 and -30 °C. Journal of Agricultural and Food Chemistry, 47(4), 1372–1377. https://doi.org/10.1021/jf9807420

Dyer, W. J., French, H. V., & Snow, J. M. (1950). Proteins in Fish Muscle.: I. Extraction of Protein Fractions in Fresh Fish. Journal of the Fisheries Research Board of Canada, 7d(10), 585–593. https://doi.org/10.1139/f47-052

Eide, O., Børresen, T., & Strøm, T. (1982). Minced Fish Production From Capelin (Mallotus villosus). A New Method for Gutting, Skinning and Removal of Fat from Small Fatty Fish Species. Journal of Food Science, 47(2), 347–349. https://doi.org/10.1111/j.1365-2621.1982.tb10078.x

Emborg, J., Laursen, B. G., Rathjen, T., & Dalgaard, P. (2002). Microbial spoilage and formation of biogenic amines in fresh and thawed modified atmosphere-packed salmon (Salmo salar) at 2°C. Journal of Applied Microbiology, 92(4), 790–799. https://doi.org/10.1046/j.1365- 2672.2002.01588.x

Erikson, U., Lambooij, B., Digre, H., Reimert, H. G. M., Bondø, M., & van der Vis, H. (2012). Conditions for instant electrical stunning of farmed Atlantic cod after de-watering, maintenance of unconsciousness, effects of stress, and fillet quality - A comparison with AQUI-STM. Aquaculture, 324–325, 135–144. https://doi.org/10.1016/j.aquaculture.2011.10.011

Erikson, Ulf, Digre, H., & Misimi, E. (2011). Effects of Perimortem Stress on Farmed Atlantic Cod Product Quality: A Baseline Study. Journal of Food Science, 76(4). https://doi.org/10.1111/j.1750- 3841.2011.02141.x

Etheridge, R. D., Pesti, G. M., & Foster, E. H. (1998). A comparison of nitrogen values obtained utilizing the Kjeldahl nitrogen and Dumas combustion methodologies (Leco CNS 2000) on samples typical of an animal nutrition analytical laboratory. Animal Feed Science and Technology, 73(1–2), 21–28.

128

https://doi.org/10.1016/S0377-8401(98)00136-9

FAO. (2020). FAO Fisheries & Aquaculture - Species Fact Sheets - Gadus morhua (Linnaeus, 1758). Retrieved January 26, 2020, from http://www.fao.org/fishery/species/2218/en

Godiksen, H., Hyldig, G., & Jessen, F. (2003). Sarcoplasmic Reticulum Ca2+-ATPase and Cytochrome Oxidase as Indicators of Frozen Storage in Cod (Gadus morhua). Journal of Food Science, 68(8), 2579–2585. https://doi.org/10.1111/j.1365-2621.2003.tb07064.x

Henriksen, E., Heide, M., Hansen, Ø. J., & Mortensen, A. (2018). Kunnskaps- og erfaringsgrunnlag for torskeoppdrett.

Hermansen, Ø., & Eide, A. (2013). Bioeconomics of Capture-Based Aquaculture of Cod (Gadus Morhua). Aquaculture Economics and Management, 17(1), 31–50. https://doi.org/10.1080/13657305.2013.747225

Hibberd, M. G., & Trentham, D. R. (1986). Relationships between chemical and mechanical events during muscular contraction. The Annual Review of Biophysics, 15, 119–161. https://doi.org/10.1146/annurev.bb.15.060186.001003

Hultmann, L., Phu, T. M., Tobiassen, T., Aas-Hansen, O., & Rustad, T. (2012). Effects of pre-slaughter stress on proteolytic enzyme activities and muscle quality of farmed Atlantic cod (Gadus morhua). Food Chemistry, 134(3), 1399–1408. https://doi.org/10.1016/j.foodchem.2012.03.038

Jobling, M. (1981). The influences of feeding on the metabolic rate of fishes: a short review: NOTE ADDED IN PROOF. Journal of Fish Biology, 18(5), 615–615. https://doi.org/10.1111/j.1095- 8649.1981.tb03802.x

Jørpeland, G., Imsland, A., Stien, L. H., Bleie, H., & Roth, B. (2015). Effects of filleting method, stress, storage and season on the quality of farmed Atlantic cod ( Gadus morhua L.). Aquaculture Research, 46(7), 1597–1607. https://doi.org/10.1111/are.12312

Kristoffersen, S., Tobiassen, T., Steinsund, V., & Olsen, R. L. (2006). Slaughter stress, postmortem muscle pH and rigor development in farmed Atlantic cod (Gadus morhua L.). International Journal of Food Science and Technology, 41(7), 861–864. https://doi.org/10.1111/j.1365-2621.2005.01149.x

LeBlanc, E. L., & LeBlanc, R. J. (1988). Effect of frozen storage temperature on free and bound formaldehyde content of god (Gadus morhua) fillets. 12, 95–113. https://doi.org/10.1111/j.1745-

129

4549.1988.tb00070.x

LeBlanc, E. L., & LeBlanc, R. J. (1992). Determination of hydrophobicity and reactive groups in proteins of cod (Gadus morhua) muscle during frozen storage. Food Chemistry, 43(1), 3–11. https://doi.org/10.1016/0308-8146(92)90234-S

LeBlanc, E. L., LeBlanc, R. J., & Blum, I. E. (1988). Prediction of Quality in Frozen Cod (Gadus morhua) Fillets. Journal of Food Science, 53(2), 328–340. https://doi.org/10.1111/j.1365- 2621.1988.tb07700.x

Martinsdottir, E., & Magnusson, H. (2001). Keeping Quality of Sea-Frozen Thawed Cod Fillets on Ice. Journal of Food Science, 66(9), 1402–1408. https://doi.org/10.1111/j.1365-2621.2001.tb15222.x

Mørkøre, T., Hansen, S. J., & Rørvik, K. (2006). Relevance of storage temperature for contraction and gaping of pre rigor filleted farmed cod (Gadus morhua L.). In J. B. Luten, C. Jacobsen, K. Bekaert, A. Sæbø, & J. Oehlenschläger (Eds.), Seafood research from fish to dish. Wageningen: Wageningen Academic.

Olsen, S. H., Joensen, S., Tobiassen, T., Heia, K., Akse, L., & Nilsen, H. (2014). Quality consequences of bleeding fish after capture. Fisheries Research, 153, 103–107. https://doi.org/10.1016/j.fishres.2014.01.011

Olsen, S. H., Tobiassen, T., Akse, L., Evensen, T. H., & Midling, K. T. (2013). Capture induced stress and live storage of Atlantic cod (Gadus morhua) caught by trawl: Consequences for the flesh quality. Fisheries Research, 147, 446–453. https://doi.org/10.1016/j.fishres.2013.03.009

Olsson, G. B., Gundersen, B., & Esaiassen, M. (2006). Pre-slaughter starvation of farmed Atlantic cod fed vegatable proteins: Effects on quality parameters. In J. B. Luten, C. Jacobsen, K. Bekaert, A. Sæbø, & J. Oehlenschläger (Eds.), Seafood research from fish to dish (pp. 139–148). Wageningen: Wageningen Academic.

PanelCheck. (2012). PanelCheck. Retrieved December 11, 2019, from http://www.panelcheck.com/

Ratkowsky, D. A., Olley, J., McMeekin, T. A., & Ball, A. (1982). Relationship between temperature and growth rate of bacterial cultures. Journal of Bacteriology, 149(1), 1–5. https://doi.org/10.1128/jb.149.1.1-5.1982

Rehbein, H. (1988). Relevance of trimethylamine oxide demethylase activity and haemoglobin content

130

to formaldehyde production and texture deterioration in frozen stored minced fish muscle. Journal of the Science of Food and Agriculture, 43(3), 261–276. https://doi.org/10.1002/jsfa.2740430308

Schubring, R. (2005). Changes in texture, water holding capacity, colour and thermal stability of frozen cod (Gadus morhua) fillets: Effect of frozen storage temperature. Deutsche Lebensmittel- Rundschau, 101(11), 484–493.

Skjelvareid, M. H., Heia, K., Olsen, S. H., & Stormo, S. K. (2017). Detection of blood in fish muscle by constrained spectral unmixing of hyperspectral images. Journal of Food Engineering, 212, 252–261. https://doi.org/10.1016/j.jfoodeng.2017.05.029

Sønvisen, S. A., & Standal, D. (2019). Catch-based aquaculture in Norway - Institutional challenges in the development of a new marine industry. Marine Policy, 104(April 2018), 118–124. https://doi.org/10.1016/j.marpol.2019.02.039

Sørensen, J. S. J. S., Bøknæs, N., Mejlholm, O., & Dalgaard, P. (2020). Superchilling in combination with modified atmosphere packaging resulted in long shelf-life and limited microbial growth in Atlantic cod ( Gadus morhua L .) from capture-based-aquaculture in Greenland. Journal Of Food Microbiology, 88(December 2019), 103405. https://doi.org/10.1016/j.fm.2019.103405

Statistics Greenland. (2019). Total landings of fish and shellfish by time, municipality, species, unit and month [FIE001]. Retrieved August 9, 2019, from FIX001 website: http://bank.stat.gl/pxweb/da/Greenland/Greenland__FI__FI10/FIX001.px/chart/chartViewLine/?rx id=FIX00109-08-2019 07:30:24

Stone, H., Sidel, J., Oliver, S., Woolsey, A., & Singleton, R. (1974). Sensory evaluation by quantitative descriptive analysis. Food Technology, 28(11), 24–33. the Farm Animal Welfare Committee (FAWC). (2014). Opinion on the welfare of farmed fish at the time of killing welf. London.

Torry Research Station. (2001). Cold storage of frozen fish. Retrieved January 26, 2020, from Torry advisory note No. 28 website: http://www.fao.org/3/x5907e/x5907e00.htm

Waagbø, R., Jørgensen, S. M., Timmerhaus, G., Breck, O., & Olsvik, P. A. (2017). Short-term starvation at low temperature prior to harvest does not impact the health and acute stress response of adult Atlantic salmon. PeerJ, 2017(4), 1–22. https://doi.org/10.7717/peerj.3273

131

132

133

134

Paper III

Jonas Steenholdt Sørensen, Niels Bøknæs, Ole Mejlholm, Flemming Jessen & Paw Dalgaard

Thawed and chilled Atlantic cod (Gadus morhua L.) from Greenland - Options for improved distribution

Manuscript submitted to LWT – Food Science and Technology

135

Thawed and chilled Atlantic cod (Gadus morhua L.) from Greenland - Options for improved distribution

Jonas Steenholdt Sørensen 1,2, Oliver Ørnfeld-Jensen1, Niels Bøknæs2, Ole Mejlholm2, Flemming Jessen1 and Paw Dalgaard1

1National Food Institute (DTU Food), Technical University of Denmark, Kgs. Lyngby, Denmark

2Royal Greenland Seafood A/S, Svenstrup J, Denmark

* Corresponding author: Food Microbiology and Hygiene, National Food Institute, Technical University of

Denmark, Kemitorvet, Building 202, 2800, Kgs. Lyngby, Denmark. E-mail: [email protected]

136

Abstract

Frozen Atlantic cod can have a long shelf-life, but some markets demand convenience products and thawed and chilled (refreshed) fish may fulfil this demand. Sensory, chemical and microbiological changes for refreshed cod from Greenland were studied to determine shelf-life and potential indices of spoilage. Aerobic sensory shelf-life was 13 days at 2.9 °C and 19 days at 0.4 °C, with modified atmosphere packaging

(MAP: 40% CO2 and 60% N2) extending shelf-life to >32 days. Low drip loss during chilled storage of 2.3- 2.5% for refreshed cod in air and 3.4-3.6% in MAP suggested the studied fish material was suitable for a combination of frozen and chilled distribution. Pseudomonas spp. and Psychrobacter spp. dominated the spoilage microbiota of chilled cod in air, while Carnobacterium maltaromaticum and Rahnella aquatilis dominated the microbiota of chilled MAP cod. A specific spoilage organism, that limited sensory shelf-life and caused the observed chemical product changes, including the formation of total volatile basic nitrogen (TVBN), was not identified.

Keywords: Sensory shelf-life, Modified atmosphere packaging (MAP), Microbial changes, H2S-producing bacteria, Pseudomonas.

137

1. Introduction

Food waste and losses must be reduced within all food industries to meet the UN Sustainable Development Goals (FAO, IFAD, UNICEF, WFP, & WHO, 2018). 27 % of all landed fish has been estimated to be wasted or lost (FAO, 2018). Microbial spoilage was a considerable cause of seafood losses and this may be reduced by improved hygiene, food preservation, packaging and management of conditions in distribution (Dalgaard, 2000; Ghaly, Dave, Budge, & Brooks, 2010; Svanevik, Roiha, Levsen, & Lunestad, 2015).

Frozen cod with shelf-life of 8-12 months at -24 to -30 °C (Bøgh-Sørensen, 2006) allowed catching of cod at high season and spreading the sales and distribution globally throughout the year (Hermansen & Dreyer, 2010; Kearney, 2010). Capture-based aquaculture (CBA), including live storage prior to processing and filleting of fish in pre-rigor mortis state, has been shown to improve the colour of the cod fillet, by decreasing discolouration and may improve other sensory attribures (Martinsdottir & Magnusson, 2001; Olsen, Tobiassen, Akse, Evensen, & Midling, 2013). Furthermore, live fish in net enclosures can be kept close to a processing facility and time from slaughter to freezing can be as short as two hours which was beneficial for the sensory quality of the fish (Martinsdottir & Magnusson, 2001). After distribution of frozen cod it can be marketed frozen or alternative as thawed and chilled (refreshed) products for catering or in consumer sizes packaging.

Compared to chilled fresh cod, the shelf-life of refreshed and aerobically stored cod has been extended marginally by less than 3-4 days. In contrast, shelf-life of refreshed chilled cod fillets in modified atmosphere packaging (MAP) has been extended by more than one to two weeks compared to unfrozen products (Guldager, Bøknæs, Østerberg, Nielsen, & Dalgaard, 1998). For refreshed MAP cod, markedly reduced production of both trimethylamine (TMA) and total volatile basic nitrogen (TVBN) was observed, and this was due to inactivation of the spoilage bacterium Photobacterium spp. by freezing and frozen storage (Bøknæs, Østerberg, Nielsen & Dalgaard, 2000; Bøknæs et al., 2002; Guldager et al., 1998). Although, refreshed MAP cod had relatively long chilled shelf-life this product was challenged by high drip loss (Bøknæs et al., 2000; 2002; Guldager et al., 1998).

The objective of the present study was to determine shelf-life and indices of spoilage for thawed Atlantic cod from CBA in Greenland. Firstly, the effect of two different bleeding methods on microbial contamination of cod fillets was evaluated. Secondly, sensory, chemical and microbial changes of frozen, thawed and chilled cod fillet pieces were studied in a storage trial with four treatments including

138 chilled storage at 0 °C and 3 °C in air or MAP (40% CO2 and 60% N2). Finally, and independent storage trial with cod in air was performed at ~1.5 °C to evaluate the results of the first storage trial.

2. Materials and methods

2.1 Effect of bleeding methods on microbial quality of cod fillets

The effect of two different bleeding methods on the microbial quality of cod fillets was evaluated. For method (I) stunned cod was double cut at the dorsal aorta and washed for three minutes in circulating refrigerated water (CRW). Then, fish were transferred to a larger tank with CRW at 4-8 °C where they were bled for 30 minutes. With method (II) the stunned cod was decapitated and eviscerated manually followed by washing and bleeding as for method (I). Evaluation of the concentration of microorganisms in nine cod fillet for both bleeding methods was performed during a two hour full-scale production in Maniitsoq, Greenland. When the fish was filleted knives and workbench were cleaned with ethanol between each fish. Samples were kept in individual plastic containers and transported in styrofoam boxes, cooled with ice from Maniitsoq to a laboratory in Nuuk, Greenland for enumeration of bacteria as described in section 2.4.1.

2.2 Storage trial with thawed Atlantic cod from capture-based aquaculture (Batch A).

2.2.1 Raw material, packaging and storage conditions.

Atlantic cod (Gadus morhua L.) were captured inshore by pound net and transported alive to a fish factory in Maniitsoq, Greenland. The fish was handled by method II (See 2.1) and machine filleted. Fillets were individually quick frozen (IQF). The studied fish raw material was produced on the 27th of November of 2017 and transferred to DTU Food, Kgs. Lyngby, Denmark, in March 2018. Storage and transport were at -20 °C. One-hundred fillets were thawed overnight at +2 °C. The thawed fillets were cut by hand into 300 pieces of approximated 100 g each. In between the cutting of each fillet, the cutting board and knives were rinsed with 96 % ethanol to avoid cross-contamination of microorganism between fillets.

A storage trial was performed with four treatments including (i) aerobic storage in ice; (ii)

aerobic chilled storage at 3 °C; (iii) MAP (40 % CO2 and 60% N2) storage in ice and (iv) chilled MAP storage

139 at 3 °C. Pieces of cod were packed as previously described (Sørensen et al., 2020). The iced samples, both aerobic and MAP were entirely covered with flake ice, which was regularly refilled during storage, as the ice melted. The temperature was recorded every 30 minutes (TinyTaq Plus, Gemini Data Loggers Ltd., Chichester, UK).

After thawing of the cod, before dividing the pieces of fillet meat into the four treatments, samples to determine the initial sensory, chemical and microbiological quality attributes were analysed using methods described in the sections 2.2.2-2.2.4. During storage and for each treatment, sampling was performed with intervals of three to five days with a total storage period of up to 26 days for aerobic storage and of up to 32 days for MAP storage. At each sampling time, three randomly picked bags, from each treatment, were analysed for microbiological and chemical changes. Five other randomly selected bags, from each treatment, were chosen for sensory evaluation.

2.2.2 Sensory changes of refreshed and chilled cod

Sensory evaluation of batch A cod was performed by using the Quality Index Method for thawed Atlantic cod fillets as previously described (Sørensen et al., 2020).

2.2.3 Chemical changes

Chemical changes as potential indices of spoilage were determined throughout the storage trial: Concentrations of trimethylamine-oxide (TMAO), trimethylamine (TMA) and total volatile basic nitrogen (TVBN) were determined in duplicate for each bag by a modified Conway and Byrne method

(Conway & Byrne, 1933). pH was recorded in 25 g fish mixed with 75 mL H20 for each sample as part of the first step in the Conway and Byrne protocol, lactic acid was quantified by HPLC and the headspace gas composition was determined on each bag for microbiological and chemical analysis by using a gas analyser as previously described (Sørensen et al. 2020). Drip loss was measured by gravity draining of liquid in each bag for one minute and calculated as the percentage loss of the total weight (Guldager et al., 1998).

2.2.4 Microbiological changes

140

The microbiota was quantified by diluting 20.0 grams of cod flesh without skin tenfold in chilled physiological saline with 0.1 % peptone (PSP) (NMKL, 2006) followed by homogenisation for 60 seconds in a Stomacher 400 (Seward Medical, London, UK). Total viable counts (TVC) was determined by spread plating on chilled Long and Hammer (LH) ager (NMKL, 2006), Pseudomonas spp. was determined by spread plating on Pseudomonads agar (CM0559, Oxoid, Basingstoke, UK) with CFC selective supplement (SR0103, Oxoid, Basingstoke, UK), H2S-producing bacteria were determined as black colonies by pour plating in Iron Agar (IA) Lyngby (CM0964, Oxoid, Basingstoke, UK) with L-cysteine hydrochloride, Photobacterium spp. was enumerated by using a conductance method and Lactic acid bacteria (LAB) were quantified by pour plating in nitrite actidione polymyxin (NAP) agar using methods and incubation as described by Sørensen et al. (2020).

To identify the dominating microbiota for treatments of the storage trial, all countable colonies on LH plates were divided into groups based on colony characteristics (size, profile, elevation, boundary, colour) and for each group of colonies, their proportion of the concentration of countable colonies was calculated. To identify the groups of colonies present for each treatment, a total of 30 colonies with five to nine colonies for each treatment were isolated from LH plates (highest dilutions) at the time of sensory spoilage or at the end of the storage period. To identify isolates, these were pure- cultured and their 16S rRNA gene was sequenced as as previously described (Sørensen et al.,2020).

2.3 Additional storage trial with refreshed cod in air (Batch B)

An independent storage trial was performed with cod pieces which were produced, packed and analysed as described above in section 2.2 with the following modifications. The cod was processed on 27th July 2017 and transferred to DTU Food in October 2017. The additional storage trial included a single treatment (v) with aerobic storage at 1.5 °C to fill the gap between treatment i and ii. Sensory evaluation was performed in triplicate with a minimum of five assessors to evaluate off-odours by using a simple scale with three grades (Dalgaard, 2000). An average score of 2.5 or above was used as the point of spoilage. Cod pieces were stored and evaluated during 18 days and at start and end of storage pH was measured (See 2.2.3). Ten colonies were isolated at the end of the storage trial and their 16S rRNA gene was sequenced to characterise the dominating microbiota as described in section 2.2.4.

2.4 Statistical analyses

141

To evaluate differences between the microbial concentrations resulting from the two studied bleeding methods, differences in product pH and in lactic acid concentrations a two-tailed homoscedastic distribution t-Test was preformed using Excel 2016 (Microsoft Corp., Redmond, WA, USA).

A most-probable-number technique was used to determine low concentrations of H2S-producing bacteria in IA (Jarvis, Wilrich, & Wilrich, 2010). To evaluate drip loss an one-way ANOVA was performed using GraphPad Prism 8.3.0 (GraphPad Software, San Diego, CA, USA).

3. Results

3.1 Effect of bleeding methods on concentrations of microorganisms in cod fillets

TVC and concentrations for Pseudomonas spp. in cod fillets did not differ (p > 0.05) for the two studied bleeding methods. However, the bleeding method II with decapitation resulted in significantly lower concentrations in IA (p < 0.0001) and of H2S-producing bacteria (p < 0.0001) in the cod fillets (Table 1). Irrespective of the studied bleeding methods, the microbiota in cod fillets, was dominated (> 96.8%) by psychrotolerant microorganisms unable to grow in IA after pour plating but with the ability to grow on the surface of chilled LH-agar plates at 15 °C (Table 1).

142

3.2 Storage trial with cod from batch A

3.2.1 Storage conditions

After freezing in Greenland, the cod fillets were stored for 4.5 months at -20 °C and after thawing and packaging at DTU Food the fish was stored at 2.9 ± 0.4 °C (Chilled) and at 0.4 ± 0.1 °C (Iced).

The equilibrium CO2 concentration for headspace gas in MAP decreased from 36% to 29% during storage in ice but remained constant at 2.9 °C (Table 2).

3.2.2 Sensory changes

Refreshed cod in air at 2.9 °C had a sensory shelf-life of 13 days, determined by total QI scores, with a cut-off level of 5 (Fig. 1). Refreshed MAP cod had shelf-life above 32 days at both 2.9 °C and 0.4 °C as total QI scores did not increase during storage (Fig. 1).

143

9 4.5

8 4.0

7 3.5

6 3.0

o s

e 5 2.5 r

s c

c S

I 4 2.0 r

e Q

3 1.5

2 1.0

1 0.5

0 0.0 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 Storage period (days)

Fig. 1. Total Quality Index (QI) scores during storage of refreshed cod at difference storage conditions:

(■) Chilled cod in air batch A, (■) iced cod in air batch A, (●) chilled cod in MAP batch A, (●) iced cod in

MAP batch A, (♦) chilled cod in air batch B. Symbols and error bars indicate Avg. ± SD. Dashed black line indicate limit for sensory spoilage

3.2.3 Chemical changes

Average drip loss for the four treatment ranged from 2.3% to 3.6% and did not change significantly during storage (p > 0.4, linear regression). There was a small but significant difference in drip loss between the treatments (p < 0.01) with the highest drip for MAP cod fillets (Table 2). An increase in pH of 0.3 units from the initial value was evaluated as an index of spoilage and this resulted in shelf-life for refreshed cod in air of 14 days at 2.9 °C and 22 days at 0.4 °C (Table 3). The EU critical limit of 35 mg- N TVBN/100 g (EC, 2008) was suitable as an index of spoilage for refreshed cod in air from batch A but TVBN concentrations did not increase for refreshed MAP cod (Fig. 2; Table 3; Table 4). The formation of TVBN could not be explained by a formation of TMA. In fact, TMA concentrations remained below 7.5 ± 2.9 mg-N/100 g of cod flesh for all treatments and the initial TMAO concentration of 74 ± 11 mg/100 g remained close to this value during storage (Results not shown). The initial lactic acid concentration for

144 refreshed cod was 2177 ± 89 ppm. For cod in air, the lactic acid concentrations decreased towards the end of the storage period (Table 4). Dry matter was on average for all treatment 19.6 ± 1.0 %.

100 95

) 90

0 85

0 1

/ 80 N

- 75 g

m 70

(

n 65 e

g 60

o r

t 55

i n

c i

s 45 a

b 40 EU limit for Gadidae family

e l

i 35

t a

l 30

o v

l a

t 20 o

T 15 10 5 0 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 Storage period (days)

145

Fig. 2. Formation of total volatile basic nitrogen (TVBN) during storage of refreshed cod at different storage conditions: (■) Chilled cod in air batch A, (■) iced cod in air batch A, (●) chilled cod in MAP batch

A, (●) iced cod in MAP batch A, (♦) chilled cod in air batch B. Symbols and error bars indicate Avg. ± SD.

The dashed line represent the critical EU limit of 35 mg-N TVBN/100g (EC, 2008).

3.2.3 Microbiological changes

The time for TVC to reach 7.0 log CFU/g underestimated sensory shelf-life and was not suitable as an index of spoilage (Table 3). The initial concentration of Pseudomonas spp. in cod after thawing was 1.4 log CFU/g. For refreshed cod in air Pseudomonas spp. grew to 9.0 log CFU/g after 13 days at 2.9 °C and after 19 days at 0.4 °C. When stored in MAP, growth of Pseudomonas spp. was slower and reached 5.9 log CFU/g at 2.9 ⁰C and 3.5 log CFU/g at 0.4 °C after 32 days of storage (Fig. 3). For refreshed MAP cod concentrations of Pseudomonas spp. were approximately two log CFU/g lower than concentrations of TVC (Fig. 3). H2S-producing bacteria, determined in IA as black colonies, was detected after 11 days of storage and never reached more than 3.0 log CFU/g in any of the treatments (data not shown). Photobacterium spp., determined by a Malthus conductance method, was not detected nor showed any growth during the storage for any of the four treatments (data not shown).

10 10 A B 9 9

8 8

7 7

g /

6 6

F C

5 5 g

o 4 4 L

3 3

2 2

1 1

0 0 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34

Storage period (days)

146

Fig. 3. Microbial changes determined by total viable counts (A) and on selective media (Pseudomonas

CFC Agar) for Pseudomonas spp. (B) at different storage conditions: (■) Chilled cod in air batch A, (■) iced cod in air batch A, (●) chilled cod in MAP batch A, (●) iced cod in MAP batch A, (♦) chilled cod in air batch B. Symbols and error bars indicate Avg. ± SD.

3.2.4 Identification of isolates from the dominating microbiota

Pseudomonas spp. dominated the microbiota with other identified species being Rahnella aquatilis, Carnobacterium maltaromaticum and Serratia conticola for refreshed cod in air, when stored in MAP, C. maltaromaticum and R. aquatilis dominated the microbiota (Table 5).

3.3 Additional storage trial with refreshed cod in air (Batch B)

The cod in batch B was stored in air at 1.4 ± 1.0 °C (Table 2) and had a sensory shelf-life of 13 days based on average odour scores exceeding 2.5. pH was lower than observed with cod from batch A (Table 4). TVBN and lactic acid concentrations did not change significantly during the storage period and TVBN concentrations remained below the EU critical limit (Table 3; Fig. 2). With 7.0 log CFU/g for TVC as a potential index of spoilage, the corresponding shelf-life for refreshed batch B cod in air was ten days at 1.4 °C and concentrations of TVC or Pseudomonas spp. never reached 9.0 log CFU/g (Fig. 3). Thus, the studied potential indices of spoilage (pH, TVBN, TVC and Pseudomonas spp.) did not corresponded to the observed sensory shelf-life of 13 days. Concentrations of TVC on LH and of Pseudomonas spp. on CFC agar were similar during the storage trial (Fig. 3) and the isolated microbiota was dominated by Pseudomonas

147 spp. and Psycrobacter spp. Photobacterium spp. was not detected and no growth of H2S producing bacteria was observed (data not shown).

4. Discussion

The observed drip losses (Table 2) was markedly lower than previously observed with cod from other regions and production methods. Frozen-at-sea cod from the Norwegian Sea had a drip loss in the range of 1.7 – 3.3 % for refreshed fillets in air (Roiha, Jónsson, Backi, Lunestad, & Karlsdóttir 2017; Roiha et al. 2018). Whiting had drip losses of, respectively, 6.0 – 9.0 % and 9.4 – 16.4 % for refreshed fish in air and MAP (Fagan, Gormley & Uí Mhuircheartaigh, 2003; Fagan, Gormley & Uí Mhuircheartaigh, 2004). Cod from the Baltic Sea and frozen post-rigor mortis had drip loss of 13 – 19 % for refreshed MAP fillets, which was much higher than the 4.6 – 5.4 % observed for fresh MAP fillets (Bøknæs et al., 2002; Guldager et al., 1998). Bøknæs et al. (2002) found drip losses of 11.4 – 12.8 % for frozen-at-sea refreshed MAP cod from the Barents Sea. The pronounced difference in drip loss for refreshed MAP cod from CBA in Greenland (3.4 - 3.6 %) and frozen-at-sea refreshed MAP cod from the Barents Sea (11.4 - 12.8 %) was not due to differences in product pH of 6.8 - 7.0. Low drip loss for refreshed cod from CBA in Greenland suggest this fish is suitable for a combination of frozen and chilled distribution. Further studies are relevant to determine if the difference in drip loss was due to production method, region and sub-group of Atlantic cod.

Frozen storage for less than one months extended shelf-life of chilled refreshed cod marginally whereas sensory shelf-life was extended 3-4 days in ice following frozen storage periods up to twelve months (Magnússon & Martinsdóttir, 1995; Vyncke, 1983). After frozen storage during one to 12 months at –20 °C to –28 °C, several studies with cod from Belgium, Iceland and Norway found sensory shelf-life of 7 – 15 days in ice for refreshed cod in air (Hansen et al., 2015; Martinsdottir & Magnusson, 2001; Roiha et al., 2017; Vyncke, 1983). Fresh cod from CBA in Greenland had sensory shelf-life of 15 days in air (Sørensen et al., 2020). Thus, the sensory shelf-life extension from 15 days to 19 days for iced refreshed cod in air (Table 3) was similar to cod from other regions and production methods. In contrast, sensory shelf-life of refreshed MAP cod from CBA in Greenland of > 32 days at both 0.4 °C and 2.9 °C (Table 3) was markedly longer than previously observed with cod from other regions. Baltic Sea cod had sensory shelf-life extended from 11 - 12 days at 1.6 °C for fresh MAP fish to more than 20 days at 1.6 °C for refreshed MAP cod, previously frozen at -21 °C during eight weeks (Guldager et al., 1998). For Barents Sea

148

MAP (13 % CO2, 83 % O2) refreshed cod, sensory shelf-life was 19 days at 0 °C after frozen storage at -23

°C for ten months (Hansen et al., 2015). Also for Barents Sea MAP (34 % CO2 and 66 % N2) refreshed cod Bøknæs et al. (2002) found sensory shelf-life of 21 days at 2.1 – 2.5 °C after frozen storage at -20 °C during six to twelve months.

Refreshed cod will typically be cooked before consumption and pathogenic microorganisms will then be inactivated. Raw refreshed cod can be used for ready-to-eat dishes like ceviche where the occurrence of L. monocytogenes can be a challenge (Fuchs & Sirvas, 1991). To avoid more than 100-fold growth of L. monocytogenes we suggest limiting the safe shelf-life of refreshed cod to 15 days at 2°C in air and 20 days at 2°C in MAP. These suggestions were based on product characteristics for refreshed cod (Table 2; Table 4) and predictions by the L. monocytogenes growth model of Mejlholm & Dalgaard (2009), as included in the Food Spoilage and Safety Predictor software (http://fssp.food.dtu.dk).

Sørensen et al. (2020) pointed out Photobacterium carnosum as the specific spoilage organism (SSO), that limited the sensory shelf-life of fresh cod from CBA in Greenland. The absence of Photobacterium spp. in the refreshed cod (See 3.2.3 and Table 5) showed P. carnosum to be inactivated by freezing and frozen storage as previously observed for other species from the P. phosphoreum clade (Bøknæs et al., 2000, 2002; Dalgaard et al., 2006; Emborg et al., 2002; Guldager et al., 1998). The absence Photobacterium spp. explained the limited TMA formation for refreshed cod. However, Bøknæs et al. (2002) found Photobacterium spp. to survive 12 months frozen storage of Barents Sea cod at -30 °C, and this resulted in pronounced TMA formation in the refreshed MAP fish. If kept at -30 °C, a similar survival of Photobacterium spp. with associated TMA formation and the shelf-life limitation must be expected for refreshed cod from Greenland.

Growth of H2S-producing bacteria to more than 7.0 log CFU/g has been observed for aerobically stored refreshed cod from Iceland and Norway (Magnússon & Martinsdóttir, 1995; Martinsdottir & Magnusson, 2001; Roiha et al., 2017, 2018). With cod from CBA in Greenland, no or very limited growth of H2S-producing bacteria was observed and this was probably explained by the production method (Table

1). Furthermore, H2S-producing Shewanella was to some extend inactivated by freezing and frozen storage. Based on decimal reduction times and frozen storage of 4.5 months, a reduction of 3.2 log would be expected (Emborg et al., 2002).

149

The observed long sensory shelf-life (Table 3) seems related to the absence or very low concentrations of Photobacterium spp. and H2S-producing bacteria in the studied cod. Avoiding contamination of the thawed cod therefore becomes important to maintain the long shelf-life. Process contamination can markedly reduce the sensory product shelf-life, as shown with thawed shrimp that was contaminated prior to chilled distribution (Mejlholm et al., 2008). Contamination with H2S-producing bacteria may be more problematic to avoid than contamination with Photobacterium spp. as H2S- producing bacteria was shown to be present in a fish processing environment after sanitation, while Photobacterium spp. were more likely to originate from the fish being processed (Møretrø, Moen, Heir, Hansen, & Langsrud, 2016).

In the present and some previous studies with refreshed chilled cod in air, TVC reached 8 - 9 log CFU/g during storage and the spoilage microbiota was dominated by Pseudomonas spp. unable to produce TMA (Fig. 3; Table 3; Magnússon & Martinsdóttir, 1995; Roiha et al., 2017). However, a dominating microbiota, including Psychrobacter spp., as observed for batch B, was also previously reported (Hansen et al., 2015). The measured concentrations of Pseudomonas spp. in refreshed cod could not alone explain the observed formation of TVBN based on their spoilage activity and yield factor for

TVBN formation (log (YTVBN/CFU)) of -10.2 (Table 5; Fig. 2; Sørensen et al., 2020). The remaining formation of TVBN must have been formed by other members of the dominating microbiota including C. maltaromaticum, R. aquatilis, and S. conticola (Fig. 2; Table 5). C. maltaromaticum has a high resistance to freezing and frozen storage and it was previously determined as a dominating part of the spoilage microbiota in refreshed seafood including MAP cod, garfish and salmon (Dalgaard et al., 2006; Emborg et al., 2002; Guldager et al., 1998). R. aquatilis was previously found as part of the spoilage microbiota for chilled cold-smoked salmon (Paludan-Müller, Dalgaard, Huss, & Gram, 1998). Indices of spoilage or an SSO responsible for spoilage and TVBN formation was not identified for refreshed cod from CBA in Greenland neither for storage in air or MAP (Table 3). For refreshed cod in air 35 mg-N TVBN/100 g corresponded to the determined sensory shelf-life for batch A. However, this was not the case for batch B (Fig. 2; Table 3). The EU critical limit of 35 mg-N/100 g (EC, 2008) therefore could not be confirmed as spoilage index for chilled refreshed cod in air although Roiha et al., (2017) found this index of spoilage appropriate. The absent or very limited TMA-formation for refreshed cod in air (See 3.2.3) previously kept 4.5 months at -20 °C corresponded to previous studies with cod from other regions. Magnússon & Martinsdóttir (1995) found <8.0 mg-N TMA/100g in aerobically stored refreshed cod from Iceland when previously kept from 5 to 52 weeks at -25 °C. Martinsdottir & Magnusson, (2001) confirmed this effect of frozen storage time on TMA formation and showed markedly less TMA development in chilled refreshed

150 cod when frozen in the pre-rigor mortis state compared to freezing post-rigor mortis. For refreshed MAP cod previously stored at close to -20 °C, the observed absence of TVBN and TMA formation (Fig. 2) has previously been observed with cod for other regions as well as for refreshed whiting, mackerel and salmon (Bøknæs et al., 2000, 2002; Fagan et al., 2004)

5. Conclusions

The long sensory shelf-life and the low drip loss for refreshed cod fillets from CBA in Greenland makes this fish raw material particularly suitable for a combination of frozen and chilled distribution. Sensory shelf-life of refreshed MAP cod was above 32 days at 2.9 °C, however since the product is frozen within most of the distribution chain, this long chilled shelf-life after thawing is not needed. A safe shelf-life of no more than 15-20 days at 2°C is recommended, to prevent more than 100- fold potential growth of Listeria monocytogenes in these products.

Acknowledgements

Innovation Fund Denmark (grant no. 5189- 00175B) funded this research. We thank laboratory technician Mia Laursen for help with the storage trials and senior research scientist Grethe Hyldig and food technician Rie Sørensen for input on sensory evaluation. We thank the sensory panel for their contribution.

ReferencesBøgh-Sørensen, L. (2006). Recommendations for the Processing and Handling of Frozen Foods. (4th ed.). France: IFF-IIR.

Bøknæs, N., Østerberg, C., Nielsen, J., & Dalgaard, P. (2000). Influence of freshness and frozen storage temperature on quality of thawed cod fillets stored in modified atmosphere packaging. LWT - Food Science and Technology, 33(3), 244–248. https://doi.org/10.1006/fstl.2000.0634

151

Connell, J. J., & Howgate, P. F. (1986). Fish and fish products. In S. M. Herschdoerfer (Ed.), Quality Control in the Food Industry -- Volume 2 (2nd ed., pp. 347–405). London: Academic Press, Inc.

Conway, E. J., & Byrne, A. (1933). An absorption apparatus for the micro-determination of certain volatile substances: The micro-determination of ammonia. Biochemical Journal, 27(2), 419-429.

Dalgaard, P. (1995). Qualitative and quantitative characterization of spoilage bacteria from packed fish. International Journal of Food Microbiology, 26(94), 319–333. https://doi.org/https://doi.org/10.1016/0168-1605(94)00137-U

Dalgaard, P. (2000). Fresh and lightly preserved seafood. In C. M. . Man & A. . Jones (Eds.), Shelf-life evaluation of foods (pp. 110–139). Gaithersburg: Aspen publishing inc.

Dalgaard, P., Madsen, H. L., Samieian, N., & Emborg, J. (2006). Biogenic amine formation and microbial spoilage in chilled garfish (Belone belone belone) - Effect of modified atmosphere packaging and previous frozen storage. Journal of Applied Microbiology, 101(1), 80–95. https://doi.org/10.1111/j.1365-2672.2006.02905.x

EC. (2008). Commission Regulation (EC) No 1022/2008 of 17 October 2008 amending Regulation (EC) No 2074/2005 as regards the total volatile basic nitrogen (TVB-N) limits. European Commission, (1022), 18–20. Retrieved from https://www.fsai.ie/uploadedFiles/Reg1022_2008.pdf

Fagan, J. D., Gormley, T. R., & Uí Mhuircheartaigh, M. M. (2004). Effect of modified atmosphere packaging with freeze-chilling on some quality parameters of raw whiting, mackerel and salmon portions. Innovative Food Science and Emerging Technologies, 5(2), 205–214. https://doi.org/10.1016/j.ifset.2004.01.001

FAO, IFAD, UNICEF, WFP, & WHO. (2018). The state of food security and nutrition in the world 2018 - Building climate resilience for food security and nutrition. Retrieved from www.fao.org/publications

Gates, K. W. (2015). Seafood Processing: Technology, Quality and Safety. Journal of Aquatic Food Product Technology, 24(1), 91–97. https://doi.org/10.1080/10498850.2014.954475

152

Ghaly, A. E., Dave, D., Budge, S., & Brooks, M. S. (2010). Fish spoilage mechanisms and preservation techniques: Review. American Journal of Applied Sciences, 7(7), 859–877. https://doi.org/10.3844/ajassp.2010.859.877

Guldager, H. S., Bøknæs, N., Østerberg, C., Nielsen, J., & Dalgaard, P. (1998). Thawed cod fillets spoil less rapidly than unfrozen fillets when stored under modified atmosphere at 2°C. Journal of Food Protection, 61(9), 1129–1136. https://doi.org/10.4315/0362-028X-61.9.1129

Hansen, A. Å., Rødbotten, M., Lea, P., Rotabakk, B. T., Birkeland, S., & Pettersen, M. K. (2015). Effect of transport packaging and repackaging into modified atmosphere on shelf life and quality of thawed atlantic cod loins. Packaging Technology and Science, 28(11), 925–938. https://doi.org/10.1002/pts.2139

Hermansen, Ø., & Dreyer, B. (2010). Challenging spatial and seasonal distribution of fish landings-The experiences from rural community quotas in Norway. Marine Policy, 34(3), 567–574. https://doi.org/10.1016/j.marpol.2009.11.003

Jarvis, B., Wilrich, C., & Wilrich, P. T. (2010). Reconsideration of the derivation of Most Probable Numbers, their standard deviations, confidence bounds and rarity values. Journal of Applied Microbiology, 109(5), 1660–1667. https://doi.org/10.1111/j.1365-2672.2010.04792.x

Kearney, J. (2010). Food consumption trends and drivers. Philosophical Transactions of the Royal Society B: Biological Sciences, 365(1554), 2793–2807. https://doi.org/10.1098/rstb.2010.0149

Lorenzo, J. M., Cachaldora, A., Fonseca, S., Gómez, M., Franco, I., & Carballo, J. (2010). Production of biogenic amines “in vitro” in relation to the growth phase by Enterobacteriaceae species isolated from traditional sausages. Meat Science, 86(3), 684–691. https://doi.org/10.1016/j.meatsci.2010.06.005

Magnússon, H., & Martinsdóttir, E. (1995). Storage Quality of Fresh and Frozen‐thawed Fish in Ice. Journal of Food Science, 60(2), 273–278. https://doi.org/10.1111/j.1365-2621.1995.tb05654.x

Martinsdottir, E., & Magnusson, H. (2001). Keeping Quality of Sea-Frozen Thawed Cod Fillets on Ice. Journal of Food Science, 66(9), 1402–1408. https://doi.org/10.1111/j.1365-2621.2001.tb15222.x

Mejlholm, O., & Dalgaard, P. (2009). Development and validation of an extensive growth and growth boundary model for Listeria monocytogenes in lightly preserved and ready-to-eat shrimp. Journal

153

of Food Protection, 72(10), 2132–2143. https://doi.org/10.4315/0362-028X-72.10.2132

Mejlholm, O., Kjeldgaard, J., Modberg, A., Vest, M. B., Bøknæs, N., Koort, J., … Dalgaard, P. (2008). Microbial changes and growth of Listeria monocytogenes during chilled storage of brined shrimp (Pandalus borealis). International Journal of Food Microbiology, 124(3), 250–259. https://doi.org/10.1016/j.ijfoodmicro.2008.03.022

Møretrø, T., Moen, B., Heir, E., Hansen, A., & Langsrud, S. (2016). Contamination of salmon fillets and processing plants with spoilage bacteria. International Journal of Food Microbiology, 237, 98–108. https://doi.org/10.1016/j.ijfoodmicro.2016.08.016

NMKL. (2006). NÆRINGSMIDLER No . 184 Aerobic count and specific spoilage organisms in fish and fish products. Nordic Committee on Food Analysis, (184), 2–7.

Roiha, I. S., Jónsson, Á., Backi, C. J., Lunestad, B. T., & Karlsdóttir, M. G. (2017). A comparative study of quality and safety of Atlantic cod (Gadus morhua) fillets during cold storage, as affected by different thawing methods of pre-rigor frozen headed and gutted fish. Journal of the Science of Food and Agriculture, 98(1), 400–409. https://doi.org/10.1002/jsfa.8649

Roiha, I. S., Tveit, G. M., Backi, C. J., Jónsson, Á., Karlsdóttir, M., & Lunestad, B. T. (2018). Effects of controlled thawing media temperatures on quality and safety of pre-rigor frozen Atlantic cod (Gadus morhua). LWT - Food Science and Technology, 90(April 2018), 138–144. https://doi.org/10.1016/j.lwt.2017.12.030

Sivertsvik, M., Jeksrud, W. K., & Rosnes, J. T. (2002). A review of modified atmosphere packaging of fish and fishery products - Significance of microbial growth, activities and safety. International Journal of Food Science and Technology, 37(2), 107–127. https://doi.org/10.1046/j.1365- 2621.2002.00548.x

154

Svanevik, C. S., Roiha, I. S., Levsen, A., & Lunestad, B. T. (2015). Microbiological assessment along the fish production chain of the Norwegian pelagic fisheries sector - Results from a spot sampling programme. Food Microbiology, 51, 144–153. https://doi.org/10.1016/j.fm.2015.05.016

Vyncke, W. (1983). Shelf life of thawed cod fillets kept in ice. Zeitschrift Für Lebensmittel-Untersuchung Und -Forschung, 177(1), 19–21. https://doi.org/10.1007/BF01042489

155

156

National Food Institute Technical University of Denmark

Kemitorvet, Building DK-2800 Kgs. Lyngby Tlf. 35 88 70 00 https://www.food.dtu.dk/