Dana, vol. 10, pp.

61-85, 1994

The Baltic salmon (Salmo salar L.):

its history, present situation and future
Lars Karlsson’ & Östen Karlstrårn
2
1. Swedish Salmon Research Institute, Forskarstigen, S-810 70 Alvkarleby, Sweden
2. Swedish National Board of Fisheries, Research Office Luleå,
Skeppsbrogatan 9, S-972 38 Lulel, Sweden

Abstract
Large-scale releases of hatchery-reared smolts in Baltic rivers were started in regulated Swedish rivers in
the 1950s. From 1988 to 1992 the countries around the Baltic released an average of 5.3 million smolts
per year in Baltic rivers. The salmon fishery in the Baltic has changed since the 1950s, i.e. the exploitation
rate has increased, and a major proportion of the catch is taken in the offshore fishery on mixed stocks
in the Baltic Main Basin. Because of the high fishing pressure, many wild srocks have declined to the cx
tent that smolt production is only about 20% of the estimated potential production.
This paper discusses the future management of Baltic salmon. A low total allowable catch (TAC) for
the offshore fishery combined with a late and variable opening date of the coastal fishery should give
wild stocks a chance to proliferate. But in 1992 and 1993 a pathological syndrome, called M74, killed
70-90% of the alevins of reared Swedish Baltic salmon stocks. Recent electrofishing surveys suggest that
wild stocks in the Gulf of Bothnia are also affected b the svndrome. If this is true, drastic reductions in
fishery exploitation are needed to avoid the depletion, and maybe even extinction, of wild stocks.
Keytuords: salmon, Baltic, M74, review, spawning run.

Introduction
This paper describes the present situation concerning the Baltic salmon stocks
(Salmo salar L.) and the biological and historical background that can explairi re
cent developments. Releases of reared salmon smolts to the Baltic have increased
gradually for about 40 years and the wild stocks have declined to a small proportion
of their original size. Only in recent years have a similar development been apparent
in other areas around the world. The Baltic experience may thus give some clues to
how a situation with increasing releases of reared salmon and decreasing wild
stocks ought to be handled. In addition, some current problems involved in man
aging Baltic salmon are also considered in more detail. Two focal points concern
problems with assessment of Baltic salmon stocks, because of the large proportion
of reared fish, and the ciosely related question of management and future changes
in fishery regulations. At present, however, these questions are partly overshadowed
by a syndrome which is particularly a threat to the remaining wild stocks. This syn
drome, called M74, is described in some detail in the last part of the paper, and suit
able means for limiting its impact on Baltic salmon stocks are discussed.
62 LARS KARLSSON & ÖSTEN KARLSTRÖM

Biology
Origin and history
The Baltic salmon is a geographically isolated group of Atlantic salmon populations
(Salmo salar L.). It is likely that North Atlantic salmon first migrated into the Baftic
area about ten thousand years ago as the ice cap over Scandinavia started to retreat
(Lundqvist 1965). Then a connection opened between the Baltic Ice Lake and the
Atlantic Ocean to the west, and the lake became a bay of the Atlantic Ocean, the
Yoldia Sea. Latei when the land started to elevate, the sea became cut off from the
Atlantic Ocean, and salmon in the area were isolated. The isolation from the
Atlantic lasted about two thousand years, which may explain why salmon do flot
migrate between the Baltic and the North Atlantic today. In Swedish tagging studies
of salmon smofts in the Baltic, only about 0.04% of the recaptured salmon had mi
grated outside the Danish Isles (Karlsson, unpubl.). Similarly, it was concluded that
the migration of salmon from Kattegat and Skagerrak into the Baltic is uncommon,
since less than 0.4% of the recaptured salmon that had been tagged on the Swedish
west coast were recovered in this area.
Berg (1948) did flot discriminate between Baltic and other Atlantic salmon popu
lations taxonomically, but later genetic studies have confirmed that Baltic salmon
form a distinct group. Ståhl (1987) compared salmon from many rivers in the Baltic
and other areas by gel electrophoresis of alleles of isozyrnes. He concluded that
salmon from Baltic populations formed a cluster that was genetically distinct from
populations outside the Baltic. He also confirmed that there were significant differ
ences between different Baltic stocks. In a couple of cases he also found significant
differences between salmon from different parts of a river. His work was partly re
peated by Jansson (1993) in the late 1980s and early 1990s, who confirmed the cx
istence of subpopulations in at least one of the few investigated Baltic rivers, the
River Kalix älv in the Gulf of Bothnia (Figure 1).

Life cycie of wild and reared salmon

Wild salmon spawn in September to November in rivers having rapids with coarse
gravel and stones, and the fertilized eggs are buried under a layer of gravel and
stones over the winter. Fecundity values from 1050 to 1500 eggs per kg female have
been reported for various Baltic river stocks (Carlin & Johansson 1971, Brännäs et
al. 1985), but fecundity is usually assumed to be in the range of 1100-1200 eggs per
kg female (Christensen & Larsson 1979; H. Börjeson, Swedish Salmon Res. Inst.,
pers. comm.). In rivers flowing into the Baltic Main Basin the eggs hatch in March
April, while in northern rivers ending in the Gulf of Bothnia hatching occurs in
April-May. The newly hatched alevins remain within the redd, where they obtain
nourishment from the yolk sac for about one month. Thereafter, in May or June,
the young salmon, now called a fry, emerges from the grave
1 where it seeks a well
protected area, often close to shore where it starts external feeding. These individ
uals soon establish territories which they defend. Territorial defense acts as an effi
cient spreading mechanism, with dominant individuals acquiring the best places
(Kalleberg 1958, Norman 1987). The salmon parr habitats are found in fairly fast-
 THE BALTIC SALMON 63

Kemij ok i

Figure i. Baltic drainage area.

Rivers and reaches of rivers
supporting salmon runs in the lijoki
past (thin line) and present
(thick line). Adapted from
Christensen & Larsson (1979).
Oulujoki

FINLAND
Ljungan

RUSSIA

SWED

flowing water having a bottom substratum of stones/boulders (Karlström 1977).

This territorial stage during which the parr ‘sit and wait’ for drifting live food
(Kalleberg 1958), lasts throughout the freshwater phase. In Main Basin rivers the
freshwater phase Iasts 1-2 years, while in Gulf of Bothnia rivers it usually lasts for
3 years, but occasionally 2 or 4 years (Alm 1934, Lindroth 1977, Karlström &
Byström 1994). At the end of this period the fish undergo physiological and be
havioural changes that result ifl a springtime smolt migration into the sea
64 LARS KARLSSON & ÖSTEN KARLSTRÖM

(Kalleberg 1958, Lundqvist 1983). The smolt run peaks when the water tempera
ture has increased to about 10°C which coincides with spring flow in the rivers, but
usually slightly after the peak flow (Karlström & Byström 1994). In a southern river
such as the Mörrumsån the smolt run occurs from late April to mid-June (Lindroth
1977). In northern rivers the emigration starts around mid-May and may last until
mid-July (Österdahl 1969, Karlström & Byström 1994).
Eggs for artificial rearing are collected by stripping salmon in the autumn. The
spawners are either caught in traps as they ascend the rivers or are kept as brood
stock in the hatchery. After hatching in the spring, the fish are kept in hatcheries for
two years in northern areas and for one or two years in southern areas. They are
then released at the smolt stage. Reared smolts are normally larger than wild ones.
For instaflce, the length of wild smolts in the River Mörrumsån varied from 10 to
16 cm (Lindroth 1977), and in River Torne the average length of wild smolts was
between 15 to 16 cm in the late 1980s (Karlström & Byström 1994). During the
same period Swedish reared smolts had a mean length of 18 cm at release (C. Eriks
son, Swedish Salmon Res. Inst., pers. comm.).
For almost all salmon stocks the central and southern parts of the Main Basin
constitute the main feeding area (Carlin 1969, Anon. 1992). Only River Neva
salmon from the Gulf of Finland do flot undergo long migration since they normally
stay in the Gulf (Kallio-Nyberg & Ikonen 1992). A minoi but variable, proportion
of the Gulf of Bothnia stocks stays within the Bothnian Sea, the southern part of
the Gulf of Bothnia (Salminen & Kuikka 1992). According to recapture data from
Finnish and Swedish tagging studies, the post-smolts migrate south along the
Swedish coast, generally swimming with the weak current (Larsson & Atheskar
1979, Ikonen & Auvinen 1985), which runs counter-clockwise in the Gulf of
Bothnia area. During their first period in the Baltic Sea smolts are presumed to feed
on airborne insects (Lindroth 1961) but in some areas they also prey to some extent
on small fish (Mitans 1970). After the post-smolts have attained a length of about
25 cm, they start eating fish, mainly clupeids (Thurow 1968). Sprat are the domi
nant food in the Main Basin, while herring are more common as food in the Gulf
of Bothnia (Thurow 1966, Christensen & Larsson 1979).
Feeding salmon remain in the sea for 1-4 years before returning to their native
rivers. They leave the southern Baltic in April-June (Thurow 1966), and Gulf of
Bothnia stocks migrate at least partly along the Finnish coast until becoming ciose
to their home river (Carlin 1969, Anon. 1992). The spawners seem to maintain a
high speed while swimming towards their rivers, at least in the Gulf of Bothnia.
Tagging studies of spawners reveal that migration speeds of about 40 km d are .

common (Carlin 1969). As the salmon approach their home river their speed
decreases and they seek the river mouth (Westerberg 1982). They remain at the
mouth for some time where they gradually enter a new migration phase adapted to
fresh water. Thereafter, they ascend the river in June-August. Wild salmon stay in
the river until spawning time, but in rivers having reared salmon stocks, spawners
may be caught for use as broodstock in hatcheries. Earlier, when the fishing
exploitation was lower, some spawners survived to spawn a second time, but now
adays repeat spawners are rare.
 THE BALTIC SALMON 65
Annual mean weight, kg
20
4 years in the sea

Spawaers, Gulf of Bothnia

15
3 years in the sea

ioC

i year in the sea

0 I I I
1930 1940 1950 1960 1970 1980 1990
Year
Figure 2. Annual mean weights of spawners in the Gulf of Bothnja. Values from 1930 to 1944 are based
on catch statistics for the Rivers Oulu and Torne. Values from 1953 to 1992 are based 00 Swedish tag
ging records. Because of the low number of 4 year old salmon jo the tagging records they were excluded
from the figure.

Growth at sea
The growth pattern of Baltic salmon is well documented (Järvi 1948, Lindroth
1965, Larsson 1983). A pronounced, and prolonged, change in mean weight
occurred in 1939, when the mean size at ali ages diminished considerably (Järvi
1948) (Figure 2). In recent years mean weights have started to increase again and
are now approaching weights recorded before 1940.
There are probably several reasons for the high growth rate observed between
1988 and 1992. The winters in recent years have been very mild, and the feeding
rate of A1+ salmon in particular, has evidently flot decreased in the manner ob
served earlier when the winters were colder (Anon. 1993a). Moreovei reared
smoits leaving the rivers have been larger than in earlier years, and smolt size is a
relevant predictor of future growth rate (Karlsson et al. 1991). The main part of the
offshore catch is taken by drift-nets which are highly size selective. In recent years
the drift-netting effort in the Main Basin has decreased considerably. It is probable
that a reduction in effort will give rapidly growing salmon a better chance of sur
viving, thus resulting in higher mean growth rates. Furthermore, the abundance of
sprat and herring has been very high in recent years. In 1992 the relation between
growth rate and the abundance of herring and sprat in the Baltic was studied (Anon.
1992). In the final multiple regression 55% of the variation in growth rate was
explained by various indicators of food availability. Since there is a considerable
covariation between smoit size, winter severity and food availability, it is, of course,
difficult to know how much any of these factors influence growth rate.

Development of naturally reproducing stocks

Prior to anthropogenic changes in the life cycle of Baltic salmon, it spawned in 60-
70 rivers (Figure 1). About 40 of these rivers are Swedish and most of them flow
into the Gulf of Bothnia (Christensen & Larsson 1979). Natural recruitment to the
66 LARS KARLSSON & ÖSTEN KARLSTROM

Baltic was then estimated to be 8-10 million smoits (Lindroth 1965). The decline of
Baltic salmon stocks started when German salmon stocks began disappearing in the
middle of the i 9th century. They were almost extinct by the end of the 1 9th century.
Damming and pollution have caused a serious decline in wild Scandinavian stocks
during the 2Oth century. The rate of this decline has been especially fast since the
late 1940s, when there was a rapid expansion of hydroelectric power production in
Sweden, and many of the country’s rivers were dammed. This caused suitable rearing
habitats for parr to disappear and decreased the accessibility of spawners to upper
parts of rivers where suitable rearing habitats could be found. Conditions in Esto
nia, Finland, Latvia, Lithuania and Russia deteriorated in a similar way with a con
siderable decline during the same period or a little later. Poland lost its famed River
Vistula stock in the 1960s, probably owirig to pollution but salmon were still pre
sent in a few other rivers such as the River Drawa (Bartel 1976). A weak stock may
1 occur in this and in a neighbouring river, but there is no current information on
stil
the status of these stocks (R. Bartel, ml. Fish. Inst. Gdansk, Poland, pers. comm.).
As a result of human intervention, natural salmon smolt production has decreased
from about 8 million smolts at the turn of the century to about 0.4 million at present
(Table 1). Of the original 44 naturally reproducing salmon stocks in the northern
rivers emptying into the Gulf of Bothnia only 12 remain in the wild. Some of the orig
inal stocks are also maintained in hatcheries. The smolt production in the northern
rivers amounts to only about 20-25% of the potential production (Table 1). Low
production can be attributed to the high exploitation by the fisheries from the 1950s
and onwards. The situation improved slightly in the late 1980s, but there are still
stocks close to extinction. For instance, River Lögde and Öre stocks produce only
a few hundred or thousand smolts per year. All but two of the rivers (i.e. Aby and
Kalix) have been stocked with eggs, fry, parr or smolts to supplement the natural
stocks. Some of these stocking efforts have been carried out at a rather large scale.
The production of stocks in many of the rivers emptying into the Main Basin is
also considerably below potential levels. Here the main reason for the low produc
tivity seems to be the pollution-caused reduction in the quality of the spawning
grounds and rearing habitats rather than overexploitation (Anon. 1993a). Most of
the salmon rivers in the Main Basin are situated in Latvia, where natural reproduc
tion occurs in about 10 rivers. Larger rivers there have been stocked with hatchery
reared smolts every year such that none of the Latvian salmon rivers have stocks of
entirely natural origin. To improve conditions for natural reproduction in some of
the larger rivers, some potential rearing areas have been restored.
Of the Main Basin rivers, the Mörrum seems to be the only one where the salmon
stock is producing close to its maximum potential capacity (Anon. 1993a).
However there is a serious problem in this river as well, where the disease furun
culosis has resulted in high mortality among spawners during the last three au
tumns.
To summarize, the situation for the Baltic natural reproducing salmon stocks is
serious. Almost ali of the stocks have suffered from various types of human activi
ties. Some of the stocks have disappeared, a few are close to extinction, and in 1992
only one was producing at fuli capacity.
 THE BALTIC SALMON 67
Table 1. Estimates of natural smolt production (thousands) by country and area in Baltic rivers having
natural stocks in 1992. Swedish names of rivers in the Gulf of Bothnia are followed by potential pro
duction figures within parentheses. Based on Anon. (1993a) and A. Johlander, Regional Fisheries Office
of the National Swedish Board of Fisheries, Jönköping, Sweden.

Gulf of Bothnia
Finland Sweden Russia Estouia Latvia Total
Tornio 23 Torne (500) 75 — — —

Simo 17 Råne(30) •
Kiiminki 0 Kalix (250) 75
Pyhäj •0 Pite (150)
Åby(6) •
Byske (60) 18
Sävarln (5) •
Rickleån (5) •0
Vindel (200) 20
Ore (5)
Lögde (5) •
Ljungan (45) 10
Sumof•2 Sumof 15
Sum of
potential (1261)

Sum Gulf of Bothnia

Finland 44 Sweden 213 257

Main Basin
Finland Sweden Russia Estonia Latvia Total
— Emån 5 — Plrnu • Irbe 10
Mörrumsån 90 Venta 15
Saka TO
Salaca 26
Vitrupe 5
Peterupe 5
Gauja 20
Daugava 5
Others 4
Sum Main Basin Sweden 95 Estonia • Latvia 100 195

Gulf of Finland
Finland Sweden Russia Estonia Latvia Total
Vantaa .0 — Neva • Kunda 4 —

Kymi •0 Loobu 7
Pirita 1
Keila 2
Vasalemma 1
Sum Gulf of Finland
Finland 0 Russia • Estonia 15 15

Total Baltic
Finland 44 Sweden 308 Russia • Estonia 15 Latvia 100 467

= low and uncertain production; •0 = original stock extinct. Present production based on stocked fish.
68 LARS KARLSSON & OSTEN KARLSTROM

No. of smoits
7000000

Other countries
6000000
Finland
Sweden

5000000

4000 000

3000000

uiIh11
2000000

1000000

0
1950 1960 1970 1980 1990
Year
Figure 3. Releases of reared salmon smoits to the Baltic, 1950-92. ‘?‘ indicates that in 1970 release of
smoits from ‘other countries’ was of uncertain magnitude.

Development of hatchery-reared stocks

In Sweden and other countries the artificial propagation of salmon began as early
as the 1860s with stocking of alevins. From about 1930 onwards, fry were retained
in ponds for autumn releases. Beginning in the late 1940s, many of the Swedish
salmon rivers were dammed. In order to replace the production dispiaced by the
hydrodams, methods for rearing salmon up to the smolt stage were developed, and
the first smolt releases were made around 1950 (Figure 3). Once these were success
ful, fishery authorities requested, and water courts decided, that power companies
had to compensate for reductions in natural smolt recruitment by artificial smolt
production. Tt was also decided from the beginning that the different river stocks
should be kept separate. For a given river, this was accomplished by only releasing
smolt of the stock native to that river and by using spawners returning to the river
as broodstock in the hatchery. At present, about 2 million smolts are introduced
annually in Sweden. Production levels are maintained by annually catching and
stripping fish from the spawning run.
Beginning in the early 1980s, Finland increased its compensatory smolt produc
tion considerably and nowadays Finnish releases are of about the same magnitude
as Swedish releases. The Finnish rearing programme is, however, mainly based on
brood stocks kept in hatcheries. Other countries around the Baltic supply about 1
million smoits annually. On average over the period 1988-92, 5.33 million reared
 THE BALTIC SALMON 69
smolts have been released annually in Baltic rivers. About 0.52 million smolts were
released in the Gulf of Finland although, as these fish do flot leave the Gulf, it is
considered a separate management unit. Of the remaining 4.81 million hatchery
reared smoits, about 75% were released in rivers flowing into the Gulf of Bothnia.
Because reared smoits account for about 85-90% of the total smolt production in
the Baltic region, the Gulf of Bothnia continues to dominate as a smolt-producing
area everi though the status of the wild stocks is presently very wealc.

Fishery and regulatory measures

Deuelopment offisheries
Saimon fishing in the Baltic was originally corifined to the capture of ascending
spawners in the rivers. This way of exploiting salmon stocks dominated until the
end of the l9th century. Between 1915 and the mid-1940s, the total catch was rela
1 (Figure 4A). Thereafter it increased
tively constant, at about 1000 tonnes.year
rapidly to about 3000 tonnes as a consequerice of an increase in the intensity of off
shore drift gul-net fishing for feeding salmon. Offshore fishing primarily takes place
in the Main Basin since this is the main feeding areas for almost ali salmon stocks.
Between 1945 and 1990, the offshore fishery steadily increased its share of the total
Catch, tonnes
5000

4000
A

3000-

2000

Sea
Coast Catoh, %
River 100-

Figure 4. A: catch of salmon in the

Baltic (5-year means) by area between
1915 and 1992; B: catch of salmon in
the Baltic, in per cent, by area. River
and coastal catches not separated
before 1930. oil r$

,) /,t/,zs , , ,) ,)
5
E
70 LARS KARLSSON & OSTEN KARLSTROM

catch to about 80% (Figure 4B) although it has decreased since 1990. At present
about 75% of the offshore catch is taken with drift-nets, and the remainder with
long lines. There is a closed season in the summei from 15 June -15 September.
Salmon are exploited in the offshore fishery from the autumn at the age of A1+.
The coastal fishery normally catch spawning migrators; thus the size of catches
in the coastal fishery depends on both the effectiveness and the number of coastal
fishing gears as x.vell as the riumber of spawning migrators leaving the feeding areas.
Coastal fisheries employ a number of different gears with trap-nets and pound-nets
being used the most.
River catches have decreased almost coritinuously since 1945 owifig mainly to
the decreased aburidance of salmon in the rivers. The fewer opportunities to catch
salmon in the rivers used for hydroelectric power production have also contributed
to this decline. As a result, even after the slight increase in 1990-92, river catches
stil! make up only about 3% of the total catch. At preserit, most of the river catch
is taken by Finland and Sweden, primarily in the Gulf of Bothnia. Most of these fish
are caught with gill-nets and rod and line by recreational fishermen. In Sweden, a con
siderable portion of the catch in dammed rivers is also used for breeding purposes.
For many years Denmark was the !eading sa!mon fishirig nation owing to the high
Danish catch in offshore fisheries. However, since 1986 Sweden and Finland have
dominated. This deve!opment coincided with the slight increase in the coasta! and
river fisheries and the corresponding decrease in offshore fishery (Figure 4A & B).

International and national regulation of the salmon fishery

‘The Baltic Salmon Fisheries Convention of 1962’ took effect in 1966 after ratifica
tion by Denmark, Federal Repub!ic of Germany, Finland and Sweden. In 1976 the
articies of the 1962 Convention were adopted by the newly established ‘Inter
national Baltic Sea Fishery Commission’ (IBFSC). The convention area inciudes the
entire Baltic Sea to the coasts. The Commission has accepted a number of regula
ting measures of the salmon fishery within the Convention area. Sa!rnon-producing
nations have also taken a number of steps to gradually strengthen their national
regulations. However, it can be conciuded that the regulations have been insufficient
as the wild stocks, particularly in the Gulf of Bothnia, decreased more or less con
tinuously up until the end of the 1980s. Around that time there was a gradual
strengthening in the effectiveness of the measures undertaken. One of the most
importarit chafiges was the division of the disputed White Zone east of Got!and
between Sweden and former Sovjet Union in 1988. This stopped large-scale off
shore fishing activities outside national fishing zones. Furthermore, commercial
prices of Baltic salmon have decreased considerably over the past decade or 50
owing to competition from farmed sa!mofl. At the end of the 1970s fishermen
received 40-50 SEK/kg, whereas by the end of the 1980s the price had dropped to
15-25 SEK/kg. This made the salmon fishery !ess attractive to fishermen and
decreased incefitives to fish salmon cornrnercial!y.
A considerable shift in attitude of the IBFSC became evident at their meeting in
1989 when, for the first time, they decided to base future decisions on ‘safe biolo
gical limits defined to safeguard wild Baltic salmon stocks’ (Anon. 1990) instead of
 THE BALTIC SALMON 71
only ‘safe biological limits for Baltic salmon’ (Anon. 1989). Tt seems unlikely that
many of the delegates at IBFSC understood the full implications of this change.
Since some wild stocks were on the verge of extinction, the only biologically mean
ingful action would have been to largely decrease exploitation of these stocks. This
would require a ban mi both the offshore fishery in the Main Basin and most of the
coastal fishery in the Gulf of Bothnia. Although changes in regulatory measures of
this magnitude have not been implemented, in 1991 a Total Aliowable Catch (TAC)
for Baltic salmon was established for the first time. TACs were also established in
1992 (expressed in weight) and 1993 (expressed in numbers). There has been con
siderable resistance to the agreements, afid it has, for several reasofis, been difficult
for Sweden and Finland to keep within the agreed catch quota.

Current pro blems in assessrnent

The situatiofi in the Baltic is complicated by the fact that reared salmon nowadays
constitute 80 to 90% of the efitire recruitmeflt. Reared salmofi and wild salmon dif
fer ifl the level of exploitation that they can withstafld. If brood stocks for rearing
are kept in a hatchery as in Finland, reared stocks may even be able to withstand
exploitation rates approaching 100%. Since 1980, the ICES group ‘Baltic Salmon
and Trout Assessment Working Group’ has carried out assessments and suggested
catch options based on conditions for wild stocks. Due to a lack of appropriate data
many assumptions were made and tagging data or catch data for reared stocks were
used for caiculations of levels of fishery exploitation. In 1989, the IBSFC endorsed
the principle of safeguarding wild stocks, and it became even more important to
base calculations on quantitative data from important wild stocks. Since fishing is
the major factor leading to the depletiofi of wild stocks in the Gulf of Bothnia, these
stocks have been the basis for assessments. There are, howevei sorne major diffi
culties in carrying out such evaluations.
Firstly, since wild salmon make up only about 10-20% of the entire recruitrnent,
and both reared and wild salmon are caught mainly ifl a mixed-stock fishery, reli
able methods are needed to distinguish between the two componefits. For several
years, identification has been carried out mainly by scale-reading analysis. Wild fish
and reared fish differ in the structure of the ‘freshwater zone’ of the scale, with wild
fish having more pronounced winter zones (Antere & Ikonen 1983). The rate of
misclassification of reared fish as wild fish may be more than 10% of the examined
reared fish (Anon. 1993a). It is therefore not possible to get any reliable estimates
of the proportion of wild fish when the true proportion is on the order of 0-15%
(Anon. 1993a). The method is better suited to areas where wild fish constitute more
than 30% of the catch as is usually only the case in parts of the Gulf of Bothnia.
The tagging of wild fish has been carried out on a rather small scale. Such data can
be useful in charting feeding areas and migration routes, but are flot sufficient for
any type of cofltinuous quantitative assessments. At preseflt it seems unlikely that
new methods will be found with which wild and reared fish can be distinguished on
a scale sufficient to allow a separate assessment of wild stocks although it would be
straightforward to cut the adipose fin on all reared fish. In a Swedish hatchery the
cost of adipose fin cutting is about 0.18 SEK per parr. Thus, to cut the fins of all
72 LARS KARLSSON & OSTEN KARLSTROM

5.25 million reared smolts would cost 0.94 million SEK. This cost may seem high,
but the cost of producing one reared smolt is 15-20 SEK, and fin-cutting would only
add about 1% to the production cost. Until an easy and cheap way of distinguishing
between wild and reared fish is found, assessments will have to be made on the
entire stock, inciuding both wild and reared fish. An alternative is to base models
on tagging data from selected reared stocks, which resembie wild stocks in terms of
life histories and exploitation patterns.
Most wild salmon stocks in the Gulf of Bothnia are in a poor state and occur in
large rivers having mean water flows at the river mouth in the rarige of 30-400
3 s These factors contribute to a lack of reliable estimates of spawner numbers,
m .

age-specific survival rates and size of smolt production. Present figures are mainly
based on catch statistics, data from electrofishing surveys and smolt traps (Karlström
1989). Estimates made in the 1960s or earliei when stocks were greater have often
been used for assessment purposes. For instance, it has often been assumed that for
wild salmon stocks in the Gulf of Bothnia a survival from egg to smolt is about 2%
(Lindroth 1965, Anon. 1988). If an average female spawner weighs 5 kg and can
tributes 1100 eggskg’, each female would produce 110 smolts. To maintain the
stock, only one male and one female of these 110 smolts would have to return to
the river. The needed escapement would thus be 2/110 smoits 0.018 or 1.8% of
the smoits. Another figure proposed ifl the 1960s is the so called ‘artificial smolt
unit’ (ASU), which was used to calculate the value of reared smoits compared to wild
ones. In setting the value of the ASU, it was assumed that one reared smolt is, on
average, worth half a wild smolt owing to lower survival during the post-smolt pe
riod. In population models of Baltic salmon, reared smolts are normally assumed to
have a mortality of, on average, 75-85% during ther first year at sea (Larsson 1983).
Only fish surviving the post-smolt period are available to the fishery. According to the
ASU unit, wild smolts would then have corresponding mortalities of only 50-70%.
These values of relative survival of wild and reared smoits also originate from the
1960s when comparative taggings of wild and reared smoits were made in Sweden
(Osterdahl 1969) and Finland (Toivonen 1977). In later investigations it was found
that the survival of reared smolts was higher than assumed by the ASU (Larsson et
al. 1979). These findings support the fact that reared smolts have increased in size
since the 1960s. Thus, since survival is positively correlated with length, an increase
in the survival of reared smoits compared with wild anes may be expected.
Unfortunately, there is no Baltic river with a wild salmon stock where the entire
smolt run can be trapped and also the ascending spawners monitored to provide
suitable data. The Swedish River Ume älv (Figure 1) where ascending fish can be
monitored has both reared and wild salmon stocks. The wild stock breeds in the
tributary Vindelälven. The two populations can be distinguished because ali reared
smoits released in the river since 1970 have had their adipose fin cut. Spawners of
both stock components are caught in a trap situated at the lowest dam in Stor
norrfors, about 20 km from the river mouth. Wild spawners are released above the
dam, allowing them to migrate to their natural spawning areas in the Vindelälven.
Catch statistics from the trap, the remainder of the river fishery and the coastal fish
ery close to the river mouth were gathered in 1973-84. Together with data on the
 THE BALTIC SALMON 73
Coastal catch in reared smolt equivalents
250000

200000
.

150000
Figure 5. Stock-recruitment curve
for the wild river Vindelälven 100000 • •
salmon stock from 1973 to 1984. .
Dots show values in individual 50000
years, and the line is a Ricker curve
fitted to the data. 0
0 0.5 1 1.5 2 2.5 3 3.5
Female spawners, tonnes

catch of spawners in the trap, this information was used to formulate a stock re
cruitment relationship (Figure 5, adapted from Andersson 1988). Ali wild smoits
were assumed to be three years old and the entire catch was divided into age ciasses
based on weights in order to determine their smolt year-class affiliation. The num
ber of reared smoits released each year is known, whereas the number of wild smoits
is flot known. By comparing the catch of adipose fin-cut fish with the catch of wild
fish (still having an adipose fin) it is possible to express the recruitment in terms of
the catch derived from reared smolts. A Ricker curve was then fitted to the individ
ual points from different years (Ricker 1954, Hilborn & Walters 1992):
(Catch in reared smolt equivalents)
(Female spawriers kg) x e4541 Female spawners kg)135748);
—

2
r = 0.208.
The potential smolt productiori capacity of the river Vindeldlven has been estirnated
to be about 200000 smolts (Table 1). If a wild smolt is equal to two reared smoits,
a catch of 115 000 smolt equivalents is equal to a production of 57500 wild smoits
which would be equal to 57500/200 000 = 0.288 or 28.8% of full production. This
was rather close to the estimated average smolt production level of salmon rivers in
the Gulf of Bothnia according to Table 1. Thus the river Vindelälven may be repre
sentative for other rivers in the Gulf of Bothnia. According to the Ricker curve,
1500 kg fernales produce a catch of 115266 reared smolt equivalents. The average
weight of wild female spawners in River Ume älv was 5.02 kg (Andersson 1988),
thus each of the 299 females gave a recruitment comparabie to that of 115266/
299 = 386 reared smolts. If a wild smolt was equal to two reared smolts owing to
differences in survival, 386 reared smolts equal 193 wild smolts. The discrepancy
between this value and the ‘average’ value mentioned above of 110 smolts perS kg
female may have arisen any time during the life cycle. There may be an egg-to-smolt
survival of more than 2% in wild salmon, or the survival of wild smolts is even more
than twice as high as that of reared smolts. A third factor is the possibility of dif
ferential mixed-stock exploitation rates for wild and reared components of the
stock. This example iliustrates that it may be of little use only to concentrate on
fishery exploitation. Instead, it may be more worthwhile to try to cover the entire
salmon life cycle in the assessment of wild Baltic salmon. This would also contribute
to a more long-term view of the management of salmon.
74 LARS KARLSSON & ÖSTEN KARLSTRÖM

Although TAC has been the international way to regulate the salmon fishery in
199 1-93, several concerns have been raised regarding use of a TAC as the main in
strument for regulating the fishery on Baltic salmon. Most importantly, salmon
have rapidly growing individuals, where each year class is exploited by man for
about two years. The IBFSC normally meets in September of year N to decide on
the TAG for the following year, N+1. On that occasion there must be a scientifically
based estimate of the number of smoits in the year class leaving the river in year N
and the survival of these post-smoits. To date, no method for providing this type of
data has been developed (Anon. 1992). Tt is therefore necessary that the TAG be low
enough to allow the required number of wild salmon to survive the fishery and start
their spawning migration, even when post-smolt survival is low.
A reduction in the mixed-stock fisheries in the Main Basin aimed at protecting
wild salmon would result in large numbers of reared spawners returning to the Gulf
of Bothnia. If these are flot caught in coastal fisheries, they will enter rivers in large
numbers but in many rivers there are limited possibilities to fish thus leadmg to an
under-utilization of a valuable resource. Furthermore reared salmon may stray to
neighbouring rivers having wild salmon stocks. Since reared and wild salmon may
tolerate different exploitation levels, it is important to try to find selective mearis
for protecting wild stocks. The coastal fishery in the Gulf of Bothnia provide such
means as wild spawners start migrating earlier than reared fish. Scale-sampling
resuits indicate that wild salmon are, on average, caught a few days earlier than
reared salmon in the coastal fishery on spawners of mixed origin (Anon. 1992).
Swedish data also show that catches outside rivers carrying wild salmon stocks in
the northern Gulf of Bothnia were on average about 10 days earlier than catches
outside rivers carrying reared stocks (Anon. 1993a). Similarly for the River Ume
älv, the mean coastal catch date was earlier for wild salmon than for reared salmon,
i.e. 2 days for grilse, 3 days for 2+ fish and 5 days for 3+ salmon (Figure 6). For

Coast River

[
29 August

Wild spawners

Reared spav ners 14 August

30 JuIy

15 JuIy
Figure 6. Mean of annual
median dates of catch, by age
class of wild and reared
30 June
(adipose fin cut) spawners on
the coast and iii a trap in the
River Ume älv between 1974
and 1982. 15 June
Grilse 2+ 3+ Grilse 2+ 3+
 THE BALTIC SALMON 75
river catches in the trap, the mean time lag was larger between wild and reared
salmon. Since wild salmon are caught earlier than reared salmon and large salmon
are caught earlier thari smaller ones, a ciosed spring fishery would selectively save
the most valuable part of the spawning migrators. Regulations of this type are flow
in force in Sweden for the trap-net fishery in the Gulf of Bothflia. Finlafld used a
similar maflagement system ifl 1986-91 and has plans to re-establish it in 1994. In
both cases there is a fixed opening date of the fishery, afid the aflflual variability in
rufi-timing has flot beefi considered.
Tagging data for reared fish from the Swedish rivers Lule älv, Angermanälven
and Indalsiilven have been used to investigate the long-term variation in run-timing.
Coastal tag recoveries in the Gulf of Bothflia in 1957-92, excluding grilse recoveries,
were included in the analysis in cases where the catch date was known. An annual
mean catch date, weighted by ifidividual fish weights, was calculated for each river
stock. Finally, a single annual arithmetic mean catch date was calculated from the
mean catch dates for the three individual stocks. This estimate was based on annual
catches varying in the rafige of 52-719 salmofi. Figure 7 shows a considerable varia
tion in mean catch date where maximum and minimum values differed by 27 days.
Within each year most of the catch takes place within a short period. The mean an
nuals.d. of catch date was 16.0, 16.6 and 18.2 days in the three river stocks Lule
älv, Angermanälven and Ifldalsälvefl, respectively. If we assume that the catch ifl
each year is normally distributed, then 68.2% of the efitire catch is taken within
mean ± s.d. or 32, 33 and 36 days, respectively. Considerifig the annual variation
in run-timing and the small s.d., it ma be expected that large variations in fishery
exploitations would occur with a fixed opefliflg date in the fishery.
Mean catch date Water temperature ifl April, °C
30 July

15 July.
5

4
Cateh date

3
30 Jufle

2
Water temperature

l5Jufle. I I I
1957 1962 1967 1972

Figure 7. Meart annual coastal catch date of three Swedish Gulf of Bothnia stocks from 1957 to 1992.
The hars show mean April seawater temperature along the Trelleborg-Sassnitz ferry route from 1974 to
1992.
76 LARS KARLSSON & OSTEN KARLSTROM

To study the possible connection between run-timing and the size of the escape
ment without spring ciosures of the fisheries, data were used on catches of spawners
in the rivers. The three rivers above have fixed traps to catch spawners ciose to the
lowest dam in the river. Normally the traps are operated throughout the season;
thus the catch leve! of this fishery can be used as an index of escapement. First, the
annual deviation from the mean leve! of female spawners caught in each river in
1974-92 was caiculated. Then a single annual catch level was calculated as an arith
metic mean from the three series of deviations from the mean for the individual
stocks. The correlation coefficient between the average annual coastal catch date
and the index of escapement was —0.826 although cause and effect have not been
demonstrated. Irrespective of this, early salmon runs produced high escapement and
vice versa. If the opening date of the coastal fishery is fixed, the annual variation in
escapement should increase further.
One way of avoiding this might be to change the opening date in response to the
expected run-tirning, to ensure that the spawning run is exploited in a more stable
manner from year to year. To investigate this possibility, the relationship between
seawater temperature in the Main Basin and run-timing was studied (Table 2). The
correlation increased gradually from January to April and then decreased in May.
Data for the April temperature (Figure 7) and a scatterplot (Figure 8) produce the
linear regression:
(Annual mean catch date) = 212 — 4.6 >< (water temperature in April);
r 0.769, p <0.001.
=
2
Table 2. Correlation between rnonthlv mean seawater temperatures measured from the ferry on the
Trelleborg-Sassnitz route, in the south of the Baltic, and mean day of coastal catch of three salmon stocks
in the Gulf of Bothnia, 1974-1992.

Temperature rneasurements in
January February March April May February- March- April-
April May May

Correlation coefficient —0.672 —0.772 —0.826 —0.877 —0.711 —0.851 —0.867 —0.855

Annual mean catch date

30 July

•
15 July .
•

Figure 8. Scatter plot of 30 June

annual April seawater
temperarure and mean
catch date. The line shows
the linear regression. 15
0 2 4 6 8
Water temperature in April, °C
 THE BALTIC SALMON 77
If water temperature in April is used as an indicator to annually change opening
dates of the fisheries in response to expected run-timing, the variation in escapement
may be decreased. The simplest way to use the data would be to change the opeflifig
date by approximately 14 days, depending on whether the water temperature in
April was below or above 4.0°C.
If the opening date is sufficiently late, a regulatiori of the type proposed may very
well be the main measure for the coastal salmon fishery in the Gulf of Bothnia. A
TAG is flot suitable for the coastal fishery as the fishery will be intensive at the start
of the seasori when wild salmon migrate and it may be stopped when only reared
grilse remain to be caught.

Reproductive disturbance in Baltic salmon

The syndrome M74
In 1974, the Swedish Salmon Research Institute detected increased mortality among
salmon alevins in some Swedish hatcheries. The first visible symptoms occurred
during the resorption of the yolk sac which was followed by rapidly accelerating
mortality. The disease was given the name ‘M74’ by Jonas Sahlin, manager of
Bergeforsen Hatchery on the river Indalsälven. The letter ‘M’ suggests that the cause
is some factor in the environment (‘miljö’ in Swedish), while ‘74’ refers to the year
when the problem was detected.
M74 has only been found in Baltic salmon females that have returned to the
rivers to spawn. For these females mortality among their offspring is close to 100%.
Tt has flot been observed among salmon on the Swedish west coast which feed in
the Atlantic. The survival of alevins produced by farmed females kept in the Baltic
is normal and the syndrome has not been confirmed from fish produced in the
Finnish compensatory programme, which is based on keeping broodstocks of
salmon in hatcheries.
The cause of the syndrome is still unknown although microorganisms are prob
ably flot involved (Johansson et al. 1993). Similar symptoms were described from
Lake Michigan (Johnson & Pecor 1969) where high mortality among coho salmon
alevins occurred in connection with the transition from endogenous to exogenous
feed. This mortality was related to increased concentrations of DDT in the egg.
Investigations to date in the Baltic have not detected increased levels of any envi
ronmental poisons (PCDD/F, sDDT, sPCB, PCN and HCB) in M74-alevins com
pared with healthy ones (Johansson et al. 1993). During 1989-91 biological and
chemical investigations were carried out on spawners whose progeny (alevins) were
affected by M74 (Norrgren et al. 1993). Biochemical investigations showed that the
detoxification system EROD-activity (cytochrome P450-dependent etoxyresorufin
0-deethylase) in the livers of sick alevins was elevated compared with levels in
healthy individuals. This suggests that the problems are caused by environmental
poisons. It was recently found that females, whose offspring show signs of M74,
exhibit a number of behavioural abnormalities such as sluggishness. New data also
indicate that there is a close connection between disturbed balance in females and
the occurrence of M74 (H. Börjeson, Swedish Salmon Res. Inst., pers. comm.).
78 LARS KARLSSON & ÖSTEN KARLSTROM

Table 3. M74-related mortality in Swedish stocks of Baltic salmon in 1992

and 1993. Ali data are from hatcheries where spawners of wild or reared
origin were caught for stripping. The mortality occurs from hatching to the
Start of feeding. )All data supplied by I-I. Börjeson, Swedish Salmon
Research Institute, Älvkarleby, Sweden.)

River Reared/ Wild origin, M74-mortality, %

R/W 1992 1993

Torne älv W 80-90 >90

Lule älv R 60 76
Skellefte älv R 45 55
Byske älv W 75 95
Ume älv/Vindel/ilven RIW 75/50 96
Lögdeälv W 60 >90
AngermanSlven R 50 85
Indalsälven R 55 > 82
Ljungan R!W 60 97
Ljusnan R 42 87
Dalälven R 80 83
Mörrumsån RJW 55 95

The incidence of M74 has varied from year to yeai and for many years it did flot
give rise to any major problems in Swedish hatcheries. However, in 1992 the frequency
of M74 iflcreased dramatically, and the figures in 1993 were even higher when losses
varied from 60 to 95% in most of the hatcheries (Table 3). Since losses of such mag
nitude cannot be compensated ifl normal hatchery routines, the release of smolts pro
duced in Swedish hatcheries will be lower ifl 1994 and 1995. Although Finnish sal-
mon rearing with broodstocks in hatcheries have flot been affected, mortality among
River Neva stock spawners ifl the Gulf of Finland and the Gulf of Bothnia were high
for the first time ifl 1993 (E. Ikonen, Finnish Game & Fish. Res. Inst., Helsinki, pers.
comm). This is an important point because River Neva fish do flot leave their area of
release. Thus it can be concluded that the agents causiflg M74 occur ifl both the Gulf
of Finland and the southern part of the Gulf of Bothnia, the Bothnian Sea.
In Sweden, plans are flow being considered for keeping broodstocks in hatcheries,
as ifl the Finnish compensatory programme. The Swedish compeflsatory programme
has worked rather well for many years 50 it is understandable that many persons con
sider it unreasonable to give up one of the key elements in the programme. A further
reason for proceeding cautiously is, of course, that no one knows whether the cur
rent outbreak of M74 is temporary or permanent. The hesitance of key people is
further increased by the fact that for some hatcheries it would cost several million
SEK to make the changes required for keeping a brood stock in a suitable manner.
Though M74 affects reared stocks seriously, the influence of M74 on wild stocks
is of greater concern. Ifl cases where offspring from wild spawners were brought
into a hatchery, their mortality rates were found to be similar to those of reared
spawner offspring (Table 3). These findings suggest that wild stocks are also affect
ed by the syndrome. Howevei electrofishing surveys of salmon parr in rivers in the
Gulf of Bothnia in 1992 gave somewhat conflicting results: densities of 0+ parr in
the River Lögde älv were only 0.1 parr •100 m compared with 3.0 parr .100 m ,
2
 THE BALTIC SALMON 79
in 1990-91 (Anon. 1993a). Densities in the River Torne dlv were also low; but den
sities in the Byske dlv afid Kalix dlv were rather similar to the levels measured in
1990-9 1. Because the locatioris of spawniflg sites chafige ifl a random manner and
0+ part tend to remain close to their place of birth, the fiumber of such parr is a
rather uficertaifi indicator of year-class strength. There was therefore flO strofig ev
idence that wild stocks in ali rivers had beefi seriously affected by M74 ifl 1992.
Ifistead, it was decided to wait for data from surveys in 1993 which are considered
more reliable, because they sample 1+ part which disperse more. Once it became
apparerit that reared salmofi suffered even higher mortality ifl 1993 than ifl 1992,
the piafined surveys ifl autumfi 1993 took Ofi evefi more importance. They will not
only give clear indicatiofis of what happened to the year ciass hatching in 1992 but
will also provide the first information about the fate Ofi the 1993 year class.
In 1993 the northern part of Sweden received an unusually large amount of snow,
and meltifig started late. Later ifl the summer, rains were uflusually heavy and alt
larger rivers reached flood stage. Coflsequently, almost ali electrofishing had to be
postponed ufitil water levels had decreased and serious eiectrofishing work in rivers
emptying into the Gulf of Bothflia did flot begin untd late August, or two weeks
later than normally. The surveys were made in a stafidardized manner (Karlström
1994), afid several sites were monitored from 1976 oflwards (Tabte 4). The low parr
Table 4. Densities of saimon parr (no 100 m ) measured by elecirofishing surveys in the River Lainio liv
2
(a tributary to the Torne liv), the middle part of river Torne liv and in the lower part of the River Byske liv
in 1976 to 1993. Catch figures represent the total catch in the River Torne system and the River Byskc lIv.

Age Sampling Salmon catch,

Year 0÷ 1+ 2÷/older sites, N tonnes

River Lainio liv 1976-89 0.15 0.16 0.14 1

4-7year 3.7
1986 3.3
1987 2.8
1988 2.7
1989 7.7
1990 0.70 0.72 0.48 14 13.9
1991 3.3 1.1 0.68 18 15.9
1992 0.05 4.70 1.1 12 22.6
1993 0.26 0.29 2.0 12

River Torne liv (middle) 1976-89 0.37 0.25 0.21 2-4year

1990 0.72 1.1 0.17 4
1991 5.7 1.4 0.72 4
1992 0.21 4.3 0.70 4
1993 0.29 0.09 2.4 4

River Ilysice liv 1980-88 1.1 1.0 0.94 17 in total 0.3-0.9

1986 0.373
1987 0.416
1988 0.307
1989 3.9 1.0 1.3 4 0.760
1990 3.6 0.35 0.73 6 2.968
1991 10.6 3.0 2.0 4 2.421
1992 3.4 9.3 2.6 6 1.089
1993 0.84 0.85 3.3 4
80 LARS KARLSSON & ÖSTEN KARLSTRÖM

No. of salmon parr.100 m

2 Salmon catch, kg
5 3000
A

• 2500
4

3 [ 0+ parr
parr
• 2000

I2+ and older parr 1500

—ç— Salmen catch ifl preceding year

2
1000

500

0 0

1+ parr/2+ parr density quotient

1.00

Figure 9. A: annual mean number of salmon parr 0.75

) in electrofishing surveys in the entire
(no .100 m
2
River Byske älv 1980-93. Salmon catch in the river
is shown from 1978 to 1992; B: 1+ parr/2÷ parr 0.50
density quotients in the River Byske älv in 1989-93.
The parr densities were corrected for differing 0.25
spawner biomass by dividing by the salmon catch
in the parent year. 0
1989 1990 1991 1992 1993
Year

densities in the 1980s can be explained by the small spawning stocks during that
period. The number of ascending spawners has been increasing since 1990 as mdi
cated by the increased salmon catch in the rivers. This has resulted in a considerable
increase in parr densities (Table 4, Figure 9A). In 1993 the overall parr densities
were low; the 0+ and 1+ groups were smaller than had been expected based on the
size of the spawning stocks in 1991 and 1992. Tt is necessary to know the escape
ment in each year to correctly analyse these data. If catch is assumed to be directly
proportional to escapement, then catch can be used to correct for differences in
spawner biomass between year classes. Parr densities were divided by the salmon
catch in the parent year, whereupon the quotient between the numbers of 1+ and
2+ parr was calculated (Figure 9B). The quotient remained fairly constant from
1989 to 1992, averaging 0.72, but it decreased to 0.31 in 1993, or 43% of the
average level iii 1989-92. This strongly suggests that the stock suffered from M74
in 1992.
 THE BALTIC SALMON 81

Implications of M74 for the future of wild stocks

Since most wild stocks suffer from a considerable lack of spawners, they will prob
ably react linearly to changes in mortality at the alevin stage. M74 in reared Swedish
stocks increase mortality by about 70%. If wild stocks are affected by M74 in the
same manner and this mortality continues in the future, wild stocks would decrease
rapidly. Within two generations, or approximately 12 years, smolt production
would decrease from the present 15-20% of the potential level to less than 5%.
Long before this point, however, several of the weakest stocks would disappear.
Furthermore, the smolt year ciasses in 1993 and 1994 would be the last large ones,
and, consequently, the last buffer against disappearance of several minor river
stocks. It will therefore be important to erisure that high proportions of these year
classes return as spawners.
The selective pressure on a stock where almost ali offspring from one female die
while offspring from other females remain unaffected must be very strong.
Therefore the chance of developing eventual resistance to the disease should be high
with a surpius of spawners. From this point of view a large number of spawners
would also be the best method of counteracting M74.
The fishery regulations needed to counteract M74 will have to be extensive and
must decrease exploitation so that the number of spawners increases by at least 3.3
times the present levet to counteract a 70% mortality rate. This would bring the
stock up to the present levet for 1-3 years before the weak smolt year classes 1995-
97 come back to the river to spawn two-three years later. The exploitation of these
year classes will have to decrease so that 10.9 (3.3 X 3.3) times as many spawners
escape the fisheries. At present, the level of fishery exploitation may be on the order
of 90-94% of all fish surviving the post-smolt stage (Anon. 1988, Anori. 1992). A
3.3-fold increase in the number of returning spawners would require a decrease in
exploitation to 60-70% of salmon surviving the post-smolt stage. The coastal ex
ploitation rate close to the river rnouth is often about 50%, and the river exploita
tion rate may be about 30%. If alt fishing on the discrete stock is prohibited the ex
ploitation rate would probably decrease to about 75%. This type of measure is
therefore irisufficient, implying that a reduction in the mixed-stock fishery is also
necessary. The only reliable way to achieve a 10.9-fold increase in escapement
would be to stop ali fishery exploitation on wild stocks.
Even the banning of al! fishery exploitation may not be enough to save ali wild
stocks if M74 becomes prevalent. As a precaution for such an eventuality, other
measures ought to be undertaken. If M74 occurs in ali parts of the Baltic an inter
national programme will be needed to establish a gene bank to protect both reared
and wild stocks. Fish from endangered stocks would be collected and brought into
hatcheries where they would be kept isolated from other stocks. These fish may later
be used as broodstock to produce fish to be released in nature. As a complement,
mik samples would be collected from a number of males from each stock. These
samples are kept deep frozen. A gene bank has been established in Norway where
both of the methods mentioned above currently are used to preserve 161 of the
country’s salmon stocks (Lillehammer & Gausen 1992).
82 LARS KARLSSON & ÖSTEN KARLSTRÖM

Conciusions
Many wild salmon stocks in the Baltic have disappeared, and others are approach
ing extinction. Since 1989, the level of the fishery exploitation has decreased and as
a result some wild stocks have undergone a minor increase. In other parts of the dis
tribution area of Atlantic salmon the exploitation pattern has changed considerably
more in the last ten years. In the middle of the 1980s, a significant fishing exploita
tion on mixed stocks still occurred in several areas. Most of these fisheries have,
howevei disappeared in the last few years. In 1989 the salmon drift-net fishery in
the Norwegian Sea was forbidden. This fishery exploited both Norwegian, Swedish,
Finnish and Russian salmon stocks (Anon. 1993b). In 1991 the salmon quota at the
Faroes was bought by a private organization, the North Atlantic Salmon Fund, and
now only a research fishery is allowed there. Last summer (1993) the sarne organi
zation purchased the quota at West Greenland. In Canada, the government has re
cently bought out the Newfoundland coastal salmon fishery, which exploited both
Canadian and USA salmon.
Of course fisheries on mixed salmon stocks still exist in several areas, but it is only
in the Baltic that the major part of the fisheries is based completely on mixed stocks.
Part of the explanation for this is that a large number of countries must act jointly
to implement major changes in Baltic fishery politics. Profits made by the commercial
salmon fishery have decreased considerably in the last few years, so it should be
easier to make unpopular decisions. Perhaps the threat from M74 can act as a cata
lyst to bring the countries together in a cooperative effort to take some large steps
forward in the management of Baltic salmon. Let us hope that M74 is only a tern
porary phenomenon, but that changes in the management system will survive longer.

Addendum
In the year that passed since this paper was written, in late 1993, a number of sig
nificant developments have occurred with regard to Baltic salmon. This addendum
brings the situation up-to-date, reflecting the situation in late 1994.
M74 continued to cause high mortality in reared salmon alevins in 1994. In
Swedish hatcheries, mortality averaged 70% (range 50-90% in different stocks).
Results of electrofishings made in a few northern rivers up until 1993 were shown
in Table 4, and results from surveys in 1994 are added in Table 5.
Densities of 0+ parr in the river sections were higher in 1994 than in 1992-93.
This was generally also true in other river sections and other rivers (Karlström
1995), the exceptions being the rivers in the southern part of the county of
Västerbotten, i.e. the Vindelälven, Öre älv and Lögde älv. In these rivers densities
were as low in 1994 as they had been during preceding years (Carlsson 1995).
Small numbers of 1+ parr in 1994 in all rivers suggest that the 1993 year class
was poor as indicated by 0+ densities in 1993. This resembles the situation in 1993,
when 1+ parr densities were also very low in all rivers examined.
The main part of the group 2+/older in 1994 was four-summer olds and had orig
inated from the strong year class 1991 (Karlström 1995). The normal smolt-age in
 THE BALTIC SALMON 83
Table 5.
Densities of salmon parr (no. 100 m 2> in the River Lainio älv, in the middle
part of the River Torne and jo the lower part of the River Byske in 1994.

River Age Sampling

sites

0+ 1+ 2+/older (N)

Lainioälv 1.2 0.14 1.3 14

Torne älv (middle) 1.3 0.55 1.8 5
Byske älv (lower) 4.2 0.86 4.2 6

the rivers is three years, but evidently the proportion of parr delaying their smolti
fication beyond three years of age was sufficient to increase parr densities in 1994
substantially.
The salmon catch (and probably spawning stock) in 1993 was at about the same
level as the catch in 1992 in ali rivers, except the river Vindelälven, where the num
ber of spawners was higher (Carlsson 1995, Karlström 1995).
Table 1 gave estimates of the natural smolt production in 1992. Estimates and
predictions of the smolt production in the years 1993-95 ifldicated that large fluc
tuations occurred during this period (Anon. 1994). In the Gulf of Bothnia, smoit
production was estimated to be 290 000 in 1993, 440 000 in 1994 and 105 000 in
1995. These figures were 24, 37 and 9% of the estimated potential production
(1200000 smolts). The high smolt production in 1994 was a result of the very
strong 1991 year class, and the low production for 1995 can be ascribed to the poor
year class in 1992, Similarly the smolt production in 1996-97 will be low. Thus the
situation boks bleak for salmon stocks in coming years. These additional data sup
port the earlier statement, given in this paper, about the serious effects of the M74-
svndrome on wild stocks.
As a result of the impact of M74 on wild stocks, the biobogicaily based advice
given to the Baltic Sea Fishery Commission in autumn 1994 was that alI fishery ex
ploitatiofi of wild salmon be prohibited. IBSFC responded by lowerifig the total
TAC of Baltic salmon from 600 000 individuals in 1994 to 500 000 in 1995. In ad
dition, IBSFC changed their management policy, i.e. their main goal of assigning
‘safe biological limits defined to safeguard wild Baltic salmon stocks’ was dowu
graded in priority and became only one of three equally importafit goals.
Sweden unilaterally imposed a number of restrictions on the salmon fishery.
Fisheries in river mouths and within rivers having wild salmon stoclcs were closed
ifl 1994, and will be closed in 1995. The exception will be angling in rivers for a
short period. A unilateral cbosure of the Swedish offshore fishery for salmon was
also imposed for the winter 1994-95.
84 LARS KARLSSON & ÖSTEN KARLSTRÖM

References
Alm, G., 1934: Salmon in the Baltic precincts. —Rapp. p.-v. Réun. Cons. int. Explor. Mer 91: 17-18.
Andersson, T, 1988: Utvecklingen av Ostersjöns laxbestånd under senare decennier. En utvrdering av
laxkompensationsverksamheten. PM, May 1988. Fiskeriverkets urredningskontor i Härnösand,
-.

pp. 1-91.
Anon., 1988: Report of the Baltic Salmon and Trout Assessment Working Group. ICES, C.M.
—

1988/Assess: 19.
Anon., 1989: Report of the Baltic Salmon and Trout Assessment Working Group. ICES, C.M. 1989/
Assess: 16.
Anon., 1990: Report of the Baltic Salmon and Trout Assessment Working Group. ICES, C.M. 1990/
Assess: 21.
Anon., 1992: Report of the Baltic Salmon and Trout Assessment Working Group. ICES, C.M. 1992/
Assess: 10.
Anon., 1993a: Report of the Baltic Salmon and Trout Assessment Working Group. ICES, C.M. 1993/
Assess: 14.
Anon., 1993b: Report of the North Atlantic Salmon Working Group. ICES, C.M. 1993/Assess: 10.
Anon., 1994: Report of the Baltic Salmon and Trout Assessment Working Group. ICES C.M.
—

1994/Assess. 15.
Antere, I. & E. Ikonen, 1983: A merhod of distinguishing wild salmon from those originating from fish
farms on the basis of scale structure. ICES, C.M. I 983/M: 26.
—

Bartel, R., 1976: The Drawa river salmon in the light of some recent tagging experimenrs. ICES, C.M.—

1976/M: 6.
Berg, L.S., 1948: Fresh water fishes of the Sovjet Union and adjacent countries. Part 1. Moscow-Lenin
—

grad. 446 pp. (Translated from Russian by Israel Program for Scientific Translations, Jerusalem 1962.)
Brännäs, B., 1<. Brännäs & L-O. Eriksson, 1985: Egg characteristics and hatchery survival in a Baltic
salmon, Salmo salar L., population. Rep. Inst. Freshwat. Res. Drottningholm. 62: 5-11.
—

Carlin, B., 1969: Salmon tagging experiments. Swedish Salmon Res. Inst. Rep. 1969)3): 8-13.
—

Carlin, B. & N. Johansson, 1971: Undersökningar över rommSngden hos lax i Ljusnan 1950 och 1970
—en jämförelse. Laxforskningsinstitutet Medd. 8: 1-6.
—

Carlsson, U., 1995: Rekrytering av laxungar i några ålvar i Våsterhotten. Redovisning av elfiskeresultat
1980-1994. PM 1995-01-11. Projekt Västerbottenslax. 1-8.
—

Christensen, 0. & P-O. Larsson, 1979: Review of Baltic salmon research, ICES. Coop. Res. Rep. No
—

89, 124 pp.

Hilborn, R. & C.J. Walters, 1992: Quantitative fisheries stock assessment: choice, dynamics and uncer
tainty. Chapman & Hall, London. 570 pp.
—

lkonen, B & H. Auvinen, 1985: Migration of salmon postsmolts (Salmo salar L.) in the Baltic Sea. —

ICES, CM. l985/M: 5.

Jansson, H., 1993: Allozyme evidence for local populations of Atlantic salmon (Salmo salar L.) in the
Kalix River, northern Sweden. Working Document 1993.— ICES Study Group on Stock Identificarion
Protocols for Finfish and Shellfish Stocks.
Johansson, N., P. Jonsson, 0. Svanberg, A. Södergren &J. Thulin, 1993: Reproduktionsstörningar hos
Ostersjöfisk. Forskningsutredning. Naturvårdsverket Rapport 4222: 1-34
-.

Johnson, H.E. & C. Pecor, 1969: Coho salmon mortality and DDT in Lake Michigan. Trans. N. Arn.
—

Wildl. Nat. Res. Conf. 34: 159-166.

Järvi, TH., 1948: On the periodicity of salmon reproduction in the Northern Baltic area and its causes.
—Rapp. P.-v. Réun.Cons. int. Explor. Mer 119: 1-131.
Kalleberg, H., 1958: Observations in a stream tank of territoriality and competition in juvenile salmon
and trout. Rep. Inst. Freshwat. Res., Drottningholm 39: 55-98.
—

Kallio-Nyberg, I. & B. Ikonen, 1992: Migration pattern of two salmon stocks in the Baltic Sea. ICES, —

CM. 1985/M: 5.
Karlsson, L., T Eriksson & C. Eriksson, 1990: Production trends in reared Swedish stocks of Baltic
salmon. DFO Symp. ‘Biological Interactions of Enhanced and Wild salmonids’. June 17-20, 1991,
—

Nanaimo, BC, Canada, pp. 1-36.

 THE BALTIC SALMON 85
Karlström, 0., 1977: Habitat selection and population densities of salmon (Salmo salar L) and trout
(Salmo trutta L) parr in Swedish rivers with some references to human activities. PILD. thesis. Acta
—

Universitatis tipsaliensis 404. 12 pp.

Karlström, 0., 1989: Situationen för de narurliga laxålvarnas laxbestånd. Fiskeristyrelsen, Utred
—

ningskontoret Luleå, Sverige, Medd. 5/1989: 1-14.

Karlströni, 0., 1994. Salmon (Salmo salar L.) reproduction in rivers in northern Sweden. Working
Document 1994. ICES Baltic Salmon and Trout Assessment Working Group in 1994.
—

Karlström, 0., 3995. Naturlaxproduktion i vattendrag i norra Sverige. Elfisken 1976-94. Rapport
19950120. Fiskeriverkets utredningskontor i Luleå. 1-17.
—

Karlström, 0. & 0. Bvström, 1994: Estimates of the smolt run in River Torne älv 1987-1993. ICES —

CM. 1994/M:19.
Larsson, P-O., 1983: Growth of Baltic salmon in the sea. Some characteristics of the Baltic salmon
(Salmo salar L.) population. Ph.D. thesis, Univ. Stockholm. 80 pp.
—

Larsson, P-O. & S. Athesl,kar, 1979: Laxsmoltens vandring från Lulellven. Fiskeritidskrift för Fin
—

land, 23(1): 8-9.

Larsson, 0-0., H.0. Larsson & C. Eriksson, 1979: Oversiktlig bedömning av resultaten av mlrknings
försök på olika stammar av lax (Salmo salar L.). Laxforskningsinstitutet, Medd. 5: 1-28.
—

Lillehammei; L. & D. Gausen, 1992: Genbank Prosjektet. Nasjonal satsning for å verne våre villaks
stammer. Laks & Miljö 1: 42-44.
—

Lindroth, A., 1961: Sea food of Baltic salmon smolt. ICES, CM. 1961/8.
—

Lindroth, A., 1965: The Baltic salmon stock. —Mitt. int. Verein. theor. angew. Limnol., 1.3: 163-192.
Lindroth, A., 1977: The smolt migration in the river Mörrumsån (Sweden) 1963-1 966.— ICES, CM.
1977/M: 8.
Lundquist, 1.. 1965: The Quaternary of Sweden. b K. Rankama (cd): The Quarernary, Vol. i., pp.
—

139-198.
Lundquist, H., 1983: Precocious sexual maturation and smolting in Baltic salmon (Salnio salar L.);
Photoperiodic synchronization and adaptive significance of annual biological cycles. HiD. thesis,
—

Univ. Umeå. 75 pp.

Mitans, A.P., 1970: Feeding of Baltic salmon smolts in the river and sea. Vopr. Ikthiol. 10)1): 60.
—

Norman, L., 1987: Akvarieobservationer av revirhävdandet hos laxungar (Salmo salar L.) av vild och
odlad härstamning. Laxforskningsinstitutet. Medd. 2: 1-7.
—

Norrgren, L., T Andersson, PA. Bergquist & 1. Björk!und, 1993: Chemical, physiological and morpho
logical studies of feral Baltic salmon (Salnio salar) suffering from abnormal fry mortality. Envir.
—

Toxicol. Chem. 12: 2065-2075.

Rickei WE., 1954. Stock and recruitment. —J. Fish. Res. Bd Can. Ii: 559-623.
Salminen, ivl. & 5. Kuikka, 1992: The divergent feeding migration of Baltic salmon (Salmo salar L.); The
significance of smolt size. NJF Seminar ‘Post-smolt biology of salmonids in ranching systems’. 10-
—

12Nov. 1992., Umeå, Sweden, pp. 1-19.

Ståbl, G., 1987: Genetic population structure of Atlantic salmon. — Fn N. Ryman & F. Urter (cds):
Population genetics and fishery management, pp. 121-140. Univ. Washington Press, Seattle.
Thurou4 1-i, 1966: Beitråge zur Biologie und Bestandkunde die Atlantischen Lachses (Salino salar L.) in
der Ostsee. Ber. dt. wiss. Komm. Meeresforsch. 18(3/4): 223-379.
—

Thurow, E, 1968: On food, behaviour and population mechanism of Baltic salmon. Swedish Salmon
—

Res. Inst. Rep. 1968(4): 1-16.

Toiuonen, J., 1977: Difference in recaptures of wild and hatchery-reared salmon smolts. ICES, C.M.
—

1977/M: 7.
Westerberg, [-I., 1982: Ultrasonic tracking of Atlantic salmon (Salmo salar L.) 1. Movements in coastal
regions. Rep. Inst. Freshwat. Res., Drortningholm 60: 82-101.
—

Osterdahl, L., 1969: The smolt run of a small Swedish river. Laxforskningsinstitutet. Medd. 8: 1-11.
—