Gravlax – a buried salmon

by Guillemette Barthouil

One of our great sources of inspiration are the food cultures of East Asia. Our exploration of umami taste, for example, has made us rediscover the wildness of our own region’s fermentations. The bridges between these cultures are not only contemporary, but can also be traced down through history.

While looking into these foodways, an unexpected similarity arose between gravlax and sushi. These two preparations are nowadays eaten raw or lightly cured. Through looking at their etymology we understand that both were once fermented fish. ‘Gravlax’ means ‘buried salmon’ or ‘grave salmon’. It is part of the wider family of the Scandinavian fermented fishes which includes Swedish surlax (‘sour salmon’) and Norwegian rakfisk (‘soaked fish’) [Falk and Torp, 1906]. Harold McGee explains that these techniques were used in remote places where huge quantities of fish were caught in a short period of time and where (and when) salt was a rare good [McGee, 2004]. The solution was to bury the clean and lightly salted fish in a ‘grave’ dug into the earth, add some carbohydrates (bark, whey or malted barley) and some antioxidants (pine needles or berries) [Levin and Al., 1964]. This traditional method creates the conditions for the lacto-fermentation process that preserves the fish. Enzymes and bacteria from the fish flesh would break down protein and fat to produce a buttery texture with a cheesy, ammoniated smell. An ‘acquired’ taste as one would say, though not so pleasant to most of us nowadays.

Sushi literally means ‘preserved fish’, revealing its fermented roots [Mouritsen, 2009]. As with gravlax, sushi in its original form, called nare-zushi, had an added carbohydrate – rice in this context – to favour the lactic acid bacteria. During this process enzymes would also break down proteins into amino acids, developing an umami taste coming mainly from glutamic acid and aspartic acid [Hariono, 2005]. Even though some regional sushi in Japan is still fermented, like funa-zushi from the shores of Lake Biwa north-east of Kyōtō, this method is no longer used for most sushi. Haya-zushi, quick sushi, appeared at the end of the 17th century during the Edō period and since then sushi became a quickly-prepared and quickly-eaten food that urban people could eat standing on a street corner [Barber, 2011]. Even though the technique has changed dramatically, some patterns of taste remain: the sourness of the lactic acid fermentation is reproduced by adding vinegar to the rice, and the umami notes of the fermented fish are replaced by soy sauce.

In the Nordic countries, there are few if any modern adaptations of this traditional buried salmon. And while there are fermented fish products, like the pungent, sulphuric and ammoniated Swedish surströmming [Skara and Al, 2015; Valeri, 2010], it isn’t buried but rather sealed in a tin.

The challenge of this experiment was to reproduce a gravlax, in its old fermented version, that is delicious to us now – or let’s say, for this first trial, at least palatable.

So we got a whole salmon, scaled it, filleted it and cut it into 2cm-wide slabs.

Barley kōji proved a great source of not only carbohydrates but also enzymes and, of course, flavour. During fermentation, kōji, Aspergillus oryzae grown on grains and/or legumes, serves as a source of a variety of enzymes which catalyse the degradation of solid raw materials to soluble products that provide fermentable substrates for yeast and bacteria in the subsequent fermentation stages [Mheen, 1972]. This catalysis speeds up the fermentation process and also allows specific flavours to develop – flavours that seem to be more pleasant than those made by the historical autolytic method [Kaoru and Al., 2006]. The main taste pattern we have found in most of our kōji-based fermentations are nuttiness, pineapple, tropical fruit and undergrowth.

Barley kōji

Our first 2 recipes made on 25 may 2013:

Scale, gut and fillet a whole salmon. Cut fillets into 2cm-wide slabs.
Layer in a food-grade plastic container 2kg of salmon, 400g of pearled barley Kōji (15%),  200g of salt (7.5%), and:

RECIPE 1: + 1 handful of spruce shoots
RECIPE 2: + 1 handful of cranberries and 1 handful of blueberries

Press them so they are covered by their own brine.
Leave them to ferment slightly cooler than room temperature (15-20˚C).

We started the first experiment with low salt content (from 5 to 7%) as the original recipes seemed to have. We checked on them after a month. They were very challenging taste- and texture-wise. We decided to take a bit more freedom from the original gravlax and increase the salt content for a product more adapted to nowadays’ taste buds.

After a few trials, 15% of salt gave better results.

The following February we decided to make some new trials. Coming back to the Nordic tradition, we experimented with different batches: adding 2% of dried blackcurrant to one, 2% of juniper wood to another and a different kōji to the last (made with sunflower seeds and an heritage barley variety called Nøgen Byg, or ‘naked barley’).

trials 3-6, left to right

Recipes 1 February 2014:

Scale, gut and fillet a whole salmon. Cut fillets into 2cm-wide slabs.
Layer in a food-grade plastic container 2kg of salmon, 615g of pearled barley Kōji (20%),  460g of salt (15%), and:

RECIPE 3: nothing (basic rakfisk)
RECIPE 4: + 60g (2%) dried blackcurrant
RECIPE 5: + 60g (2%) juniper wood
RECIPE 6. instead of pearled barley koji, Nøgen Byg and sunflower seed kōji

We layered all these ingredients, fit an identical food-grade plastic container overtop and pressed them overnight at room temperature to extract the brine. It is important that the fermenting substrates are immersed in their own brine both for ideal fermentation conditions and also to reduce rancidification of the fat from contact with oxygen in the air.

Once the brine covers the salmon we can remove the weight in the container on top, and add a little water to keep it down.
Leave to ferment in a cold room (ours was 8˚C) for at least 4 months.

We also did an underwater version of Recipe 3 to see how the extra pressure might alter the fermentation. We kept it in a hard-sealed vacuum bag 7m under 2˚C water off of the boat deck. But then a big storm came and broke the string. I tried to dive down and find it on the harbour floor, but it was too deep.

The four batches gave pretty different results, but all had amazingly a mellower and nicer smell than we expected.

Tasting notes – 3 June 2014

3. Salmon, Koji, Salt:Beautiful and soft texture. Significant layer of fat on the top. Quite sour and really umami, the salmon is not pungent at all and has developed a mushroomy/morel taste. Salmon-like and clean aftertaste.

4. Salmon, Koji, Salt, Blackcurrant:Tougher texture, it looks like the surface has been over-dried by the salt. The smell is stronger and more like Swedish surströmming. Taste-wise it also has developed a similar character to surströmming but softer and more balanced with a light salmon aftertaste.

5. Salmon, Koji, Salt, Juniper wood:Soft and melty texture. The wood has brought unpleasant notes such as bitterness, mouldy, cleaning product and astringent. We suppose that this may be from juniper and will try the same recipe replacing it with inner birch bark.

6. Salmon, Heritage koji, salt:Nice texture similar to 1). The smell is incredibly fruity/ tropical fruit. The taste is quite sour but very complex. Very nice aftertaste, clean mouth.

Trials 3 and 6 turned out best. An umami, salty, sour and sweet fish with nutty and fruity notes and mellow salmon taste. A cleaner and softer version of Swedish surströmming. And as in the ancient Japanese nare-zushi tradition (particularly with nama-zushi, or ‘raw sushi’, which had a shorter fermentation time of only month [Mouritsen, 2009]), the grain (here kōjied barley) can be eaten. A taste of another time. « I don’t know anymore if I like it or not » as Josh said, but great to understand the limits of our contemporary concept of deliciousness.

In the middle of June last year I left the lab. A couple weeks after, Josh and Roberto threw together a snack with the gravlax to try to figure it out a bit more, find a context for it and wrap up this stage of the research into something appetising.

Recipe 27.6.14: Rye-birch cracker, gravlax, viili, pineapple weed


400g rye flour
50g inner birch bark flour
15g salt
150g butter
Combine ingredients into dough, roll out thinly, punch out small discs (~3cm diameter), bake at 150˚C for 12 min.


We used trial six (with nøgen byg and sunflower seed koji). Make a brunoise of gravlax. Mix in minced pineappleweed heads (Matricaria discoidea) to taste.


Plate gravlax mixture into ring mould on cracker to form a little bowl that reaches the cracker in the middle. Fill the hole with mature viili. Top with hand-pulled buds of pineapple weed.

Recipe 6


Barber, K. (2011), Hishio, taste of Japan in humble microbes, Oxford Symposium on Food and Cookery. Oxford, England.

Falk and Torp: “Etymologisk ordbok over det norske og det danske sprog”, 1906

Hariono, I and al. (2005), Use of koji and protease in fish sauce fermentation, Singapore J Pri Ind 32: 19-29.

Kaoru, I and al. (2006), Comparison of characteristics of fermented salmon fish sauce using wheat gluten Koji with those using soy sauce Koji, Food Sci. Technol. Res., 12(3), 206-212.

Levin, MG and Potapov, LP. (1964), The people of siberia, The university of chicago press, USA, p 595

McGee, H. (2004), Food § Cooking: an encyclopedia of kitchen science, history and culture, Hodder and Stoughton, UK, p235

Mheen, T I (1972), Korean fermented foods. Selected paper from the UNESCO Work Study on Waste Recovery by Microorganisms, University of Malaya, Kuala Lumpur, Malaysia.

Mouritsen Ole G. (2009), Sushi, food for the eye, the body and the soul, Springer, New York

Skara and Al, (2015), Fermented and ripened fish products in the northern European countries, Journal of Ethnic Foods, 2 (1), 18-24

Valeri, R. (2010), Surstromming, Sweden’s famous fermented herrings, Oxford Symposium on Food and Cookery. Oxford, England.