Estuarine, Coastal and Shelf Science 149 (2014) 87e95

Contents lists available at ScienceDirect

Estuarine, Coastal and Shelf Science

journal homepage: www.elsevier.com/locate/ecss

Partial migration in introduced wild chinook salmon (Oncorhynchus

tshawytscha) of southern Chile
Miguel Araya a, *, Edwin J. Niklitschek b, Dave H. Secor c, Philip M. Piccoli d
a
University of Antofagasta, Doctorate Program in Applied Sciences, Coastal Marine Systems, Antofagasta, Chile
b
University of Los Lagos, i~mar Research Centre, Puerto Montt, Chile
c
University of Maryland Center for Environmental Science, Chesapeake Biological Laboratory, PO Box 38, Solomons, MD 20688, USA
d
University of Maryland, Department of Geology, College Park, MD 20742, USA

a r t i c l e i n f o a b s t r a c t

Article history: Partial migration, the incidence of opposing migration behaviors within the same population, has been a
Received 26 December 2013 key factor in the invasive ecology of Pacific salmon within South America. Here, we examined such life-
Accepted 20 July 2014 cycle variation in of an introduced chinook salmon population in the Ayse
n watershed, one of the largest
Available online 1 August 2014
fjord systems in NW Patagonia. The chinook salmon is the most successful invasive salmonid species in
Patagonia and has recently colonized numerous Patagonian watersheds of the Pacific and Atlantic
Keywords:
Oceans. Using analyses of fish scales and otolith strontium:calcium ratios, our results suggest the
introduced species
presence of two distinct ecotypes in the chinook population, an ocean type and a stream type, in a 3:2
life-history variation
otoliths microchemistry
ratio. The distribution of back-calculated length at the time of emigration from river to marine habitats
partial migration showed a mode of 14 cm for the ocean ecotype and 30 cm for the stream ecotype. River residence time
Sr:Ca Chile for the ocean ecotype ranged from 1 to 10 months, while that of the stream ecotype varied between 14
Ayse
n region and 20 months. Returning adults reproduced in riverine habitats between August and March, but
Ayse
n watershed reproduction by the stream ecotype was limited to the period between October and February. Our results
show that exotic chinook salmon populations established in NW Patagonia present a diversity of life-
history strategies, which seems to be as large as the ones exhibited by the species in its native distri-
bution range and in other invaded ecosystems. Chinook salmon have successfully invaded most major
rivers in Patagonia, placing priority on science and conservation related to their ecological impact.
© 2014 Elsevier Ltd. All rights reserved.

1. Introduction
At least 11 salmonid species have been introduced to rivers and
lakes in southern Chile since the beginning of the 20th century.
These introductions were mainly intended to sustain recreational
fishing (largely brown trout Salmo trutta and rainbow trout Onco-
rhynchus mykiss) or to develop extensive open cultivation or “sea
ranching” (several Pacific salmon species, including coho salmon
O. kisutch and chinook salmon Oncorhynchus tshawytscha) (Soto
et al., 2001, 2007). During the 1980s, salmon farming in sea cages
began to grow rapidly, making Chile the second-largest salmon
producer in the world after little more than 20 years, with 826,946 t
produced in 2012 (SERNAPESCA, 2013). The main species farmed in
* Corresponding author. Present address: Arturo Prat University, Faculty of
Renewable Natural Resources, Iquique, Chile.
E-mail addresses: maraya@unap.cl, miguelaray@gmail.com (M. Araya), edwin.
niklitschek@ulagos.cl (E.J. Niklitschek), secor@umces.edu (D.H. Secor), piccoli@
umd.edu (P.M. Piccoli).
Chile are Atlantic salmon Salmo salar, coho salmon, rainbow trout
and to a much lesser degree chinook salmon (SERNAPESCA, 2013).
Owing to deliberate fish introductions and fish escapes from
aquaculture facilities, the rivers, estuaries and coastal areas of
Argentinean and Chilean Patagonia now have considerable pop-
ulations of free-living salmonids. While some of these species
appear to have been established in local ecosystems for several
decades (Soto et al., 2006; Pascual and Ciancio, 2007), others (such
as chinook salmon) are undergoing expansion in their range
(Correa and Gross, 2008), and others, such as Atlantic and coho
salmon, seem to have not yet become established (Niklitschek et al.,
2013). Although a multitude of ecological impacts resulting from
these invasions can be predicted, they have seldom been docu-
mented and quantified (Buschmann et al., 2006; O’Ryan et al., 2010;
Niklitschek et al., 2013). These changes include various effects on
native wildlife, such as the introduction or transmission of diseases,
competition for habitat, trophic interactions, food-web perturba-
tions and changes in nutrient cycling and availability, including the
enhanced contribution of nutrients from the ocean to the rivers

http://dx.doi.org/10.1016/j.ecss.2014.07.011
0272-7714/© 2014 Elsevier Ltd. All rights reserved.
88 M. Araya et al. / Estuarine, Coastal and Shelf Science 149 (2014) 87e95
(Pizarro and Zolezzi, 2003; Correa and Gross, 2008; Arismendi and
Soto, 2012).
Researchers have speculated about which factors that may have
promoted the successful invasion of trout and salmon in southern
Chile (Soto et al., 2001) and in similar environments, such as New
Zealand (McDowall, 2003) and southern Argentina (Pascual and
Ciancio, 2007). Presumably, these salmonids have encountered
appropriate abiotic conditions, little or no predation pressure and
few competitors. However, this scenario does not explain the
apparent differences in invasion success of different salmonid
species, particularly of Atlantic and coho salmon, which have not
succeeded despite the millions of individuals escaped from salmon
farms during the last decades (Niklitschek et al., 2013).
The Pacific salmon species exhibiting the greatest success in
invading Patagonia is the chinook salmon, which has recently
colonized most Patagonian watersheds (Becker et al., 2007; Correa
and Gross, 2008) of the Pacific (39
- 53
S) and Atlantic Oceans
(48
e54
S). The rapid dispersal of this species has been attributed
to the similarities between the freshwater and marine habitats of
Patagonia and those of the North Pacific, facilitated by a low
ecological resistance speculated to characterize Patagonian eco-
systems (Correa and Gross, 2008). Riva-Rossi et al. (2012) have
proposed a third factor explaining this invasive success: the greater
genetic diversity and phenotypic plasticity of chinook salmon when
compared to other salmon species introduced into the region.
Multiple introductions of chinook salmon from different source
populations had resulted in higher genetic diversity and admixture,
while its higher phenotypic variability and capacity to disperse
geographically had permitted to this species a faster exploration
and colonization of suitable habitats.
Chinook salmon are considered semelparous and anadromous.
However, this is a species that exhibits partial migration (Chapman
et al., 2011), showing a considerable variability in its life-history
and life-cycle parameters, including the age of migration to the
sea, length of riverine and oceanic residence times, distribution and
patterns of ocean migration, and age and dates of the seasonal
migration to spawn in river headwaters (Healey, 1991). This di-
versity in life-history strategies may represent an adaptive mech-
anism to reduce the population consequences of environmental
variability (Secor et al., 2009). Therefore this diversity may be a
principal mean by which populations can persist and expand when
they experience or are introduced into new sets of environments
(Cury, 1994; Hendry et al., 2004; Secor, in press).
In the North Pacific, the life-history diversity of chinook salmon
is characterized by the presence of two principal ecotypes, stream
and ocean, which are defined by their residence times in freshwater
before migrating to the ocean. Individuals of the stream type
remain in freshwater for one to several years before migrating and
usually remain at sea for 2e4 years. Occasionally, males of this
ecotype mature precociously, without making an oceanic migra-
tion. Ocean-type individuals make an early migration to the sea
within their first year of life and remain at sea for 2e6 years
(Healey, 1991; Altukhov et al., 2000). These ecotypes have also been
detected in a handful of invaded areas that have been studied so far
in the southern hemisphere.
In New Zealand, most chinook salmon found in the Rakaia River
corresponded to the ocean ecotype (Unwin and Lucas, 1993). A
similar situation was found in three Pacific Ocean (western Pata-
gonia) watersheds, the Petrohue, Corcovado and Futalelfú river
basins, (Soto et al., 2007; Di Prinzio and Pascual, 2008). In contrast,
dominant stream ecotypes were found in another western Pata-
gonia watershed, the Pico River (Di Prinzio and Pascual, 2008), two
Tierra del Fuego basins, Lapataia and Ovando rivers (Ferna
ndez
et al., 2010), and one studied watershed which flows to the
Atlantic Ocean (eastern Patagonia): the Santa Cruz River basin.
In this study, we investigated how partial migration in chinook
salmon in Patagonian waters facilitated its invasion success. A
challenge in obtaining such information is the lack of regular
tagging/monitoring programs aimed to track the abundance and
dispersal of this species in Patagonian watersheds, which are now
inhabited by dozens of chinook salmon populations. This limitation
caused us to employ otolith microchemistry, which can yield in-
formation on the ontogenetic and lifetime movement behaviors of
salmon and other fishes (Elsdon et al., 2008).
For many applications, the relative proportion of strontium to
calcium (Sr:Ca) allows the estimation of residence times and the
timing of movement between freshwater and ocean for anadro-
mous species and other marine fishes that exhibit partial migration
(Secor and Piccoli, 1996; Gillanders, 2005; Miller et al., 2011).
The present study characterizes the life histories of chinook
salmon in the Ayse
n River watershed, one of the major watersheds
of northwestern Patagonia, by (i) identifying and estimating the
proportions of stream and ocean ecotypes; (ii) comparing the
growth patterns of these ecotypes; and, (iii) estimating their size-
and age-specific residency and movement patterns between ma-
rine and freshwater ecosystems.
2. Materials and methods
2.1. Sampling
Sampling was conducted between 2009 and 2010 in the Ayse
n
Fjord and in the known spawning grounds of chinook salmon in the
~
Nirehuao and Huemules Rivers (Fig. 1) during January, February
and March, which correspond to the main pre-reproductive and
reproductive period of this species in the Ayse
n River basin
(Niklitschek and Aedo, 2002). In the fjord, fish were captured using
gill nets (100  2 m; 10.1 cm stretch mesh). In spawning areas,
stranded carcasses and specimens donated by recreational fisher-
men were collected. Each sample was measured (fork length, ±
1 cm), and the sagittal otoliths and scales located above the lateral
line and near the anterior margin of the dorsal fin were extracted
and stored. In total, 63 individuals were captured in the fjord and
258 collected from spawning areas. The size distribution of the
samples ranged from 5 to 115 cm in the fjord, with a mode of 20 cm,
and from 60 to 115 cm in the spawning area, with a mode of 90 cm
(Fig 2).
2.2. Ecotype proportions
Life-history and life-cycle traits (ecotypes, age distribution,
growth, size distribution and age at river emigration) were deter-
mined by combining two types of information: (i) the circuli pat-
terns of the scales and (ii) the Sr:Ca ratio profiles of the otoliths.
After organic residues were removed, the three most readable
scales from each individual were photographed and used to
determine the age of the specimen by counting the annuli, which
were identified as abrupt decreases in the circuli spacing. The scales
were read twice by the same reader on different days, without
reference to the previous reading or to other information, such as
the specimen size or capture date and location. Thus, the readings
were based only on the scale-formation patterns. When two suc-
cessive readings were different, a third reading was performed with
reference to the length specimen data.
The final age estimate corresponded to the number of annuli
plus the time elapsed between a common hypothetical date of birth
and the date of capture for each individual. The hypothetical date of
birth was defined as July 15th of each year based on the median
spawning date reported for the species in the area (Niklitschek and
Aedo, 2002) and the estimated period of embryonic development,
 M. Araya et al. / Estuarine, Coastal and Shelf Science 149 (2014) 87e95 89

~
Fig. 1. Sampling sites. A: Nirehuao River; B: Huemules River;
n Fjord.
: Ayse

which was computed according to Billard and Jensen (1996),

assuming an average water temperature of 5
C.
Individuals were classified into stream and ocean ecotypes
(Healey, 1991) based on the formation pattern of the circuli around
the scale focus, following the method for chinook salmon suggested
by Koo and Isarankura (1967) and applied to non-native pop-
ulations in the southern hemisphere by Unwin and Lucas (1993) in
New Zealand and by Ciancio et al. (2005) in Argentina. This method

0.4
Relative frequency

0.3

0.2

0.1

0.0
5 15 25 35 45 55 65 75 85 95 105 115
Fork length (cm)

Fjord River

Fig. 2. Relative frequency distribution of the fork lengths of chinook salmon captured Fig. 3. Back-scattered electron photomicrograph of a transverse section of a chi-
in the fjord and rivers. nookesalmon otolith showing the microprobe measurement path (white dots).
90 M. Araya et al. / Estuarine, Coastal and Shelf Science 149 (2014) 87e95

is based on the fact that the circuli formed during freshwater age at river emigration was determined from Li, removing t from
residence are more closely spaced than those formed during ocean the Gompertz model for each ecotype.
residence. Therefore, it is relatively easy to separate life-history
periods and to determine the timing of river emigration. If the 3. Results
first annulus falls within the river period, the specimen is classified
as stream type. If the first annulus lies within the marine stage, the 3.1. Ecotype proportions
individual is classified as ocean type.
To confirm the pattern of the scales, Sr:Ca profiles were obtained Scale analysis indicated that of 199 individuals analyzed, 128
from transverse sections of the otoliths (sagittae). Otolith were (64%) corresponded to the ocean ecotype and 71 (36%) to the
cleaned ultrasonically and then sectioned and prepared using stream ecotype. There was no evidence of a significant difference in
techniques described by Secor and Piccoli (1996, 2007). After the these proportions between sexes (chi-squared test; c2 ¼ 0.62;
otolith sections were carbon-coated in a high-vacuum evaporator, d.f. ¼ 1; p > 0.4 or between the sub-watersheds of the Huemul and
Sr and Ca concentrations were measured via wavelength-dispersive ~
Nirehuao Rivers (chi-squared test; c2 ¼ 0.48; d.f. ¼ 1; p > 0.4)
spectrometry (WDS) using a JEOL JXA-8900 electron probe micro- (Table 1).
analyzer. The operating conditions included an accelerating voltage
of 25 kV and a current of 20 nA (Secor and Piccoli, 1996). A transect
3.2. Growth patterns
was taken from the otolith core to the dorsal edge, with a nominal
spot size of 5 mm and 10e50 mm of spacing between points (Fig. 3).
The fitted Gompertz models (Fig. 4) showed significant differ-
Raw x-ray intensities were converted to concentrations using a ZAF
ences in the three growth parameters between the two ecotypes
algorithm. The Sr and Ca concentrations were converted to molar
(Table 2). The asymptotic length was greater for the ocean ecotype
concentrations and expressed as a ratio (Sr:Ca). The time of river
(105.3 cm), which exhibited greater sizes at all ages than the stream
emigration was identified as the inflection point of the Sr:Ca profile
ecotype, whose asymptotic length was estimated in only 93 cm.
(Miller et al., 2011), where the Sr:Ca value exceeded its initial mean
Inflection points corresponded to 1.1 and 1.5 years, for ocean and
by 2 standard deviations. Sr:Ca profiles were analyzed in a random
stream ecotypes, respectively (Table 3). The absolute growth rate
sub-sample of spawners from all sampled habitats (fjord and both
was higher in the stream ecotype up to 3 years of age, but was
rivers). A total of 34 of these fish corresponded to ocean-type in-
slightly higher in the ocean ecotype after this age (Fig. 5).
dividuals (59%) and 23 to stream-type individuals (41%).

3.3. Partial migration characteristics

2.3. Growth patterns
The otoliths showed a low initial Sr:Ca ratio
The von Bertalanffy and Gompertz growth models were fitted (mean ¼ 0.37 mmol mol1; s.d. ¼ 0.23) followed by a strong and
using age (from scale readings) and fork-length data (Ricker, 1979). rapid increase to much higher values (mean ¼ 1.38 mmol mol1;
Gompertz was found to be the most informative model (lowest s.d. ¼ 0.12). These average concentrations were significantly
Akaike information criterion value), and was expressed as. different (t-test; d.f. ¼ 108; p < 0.0001), and were interpreted as
gðtt0Þ indicative of residence periods in freshwater (low Sr:Ca) and salt-
Lt ¼ L∞ ee water (high Sr:Ca) habitats (Fig. 6). The two ecotypes also differed
The absolute growth rate (G) at age t was calculated as. significantly (t-test; d.f. ¼ 55; p < 0.0001) with respect to the otolith
radius corresponding to river emigration time (i.e., the otolith
Gt ¼ gLt ðlnL∞  lnLt Þ radius at which the Sr:Ca ratio increased). This increase occurred at
an average distance of 708.9 mm (s.d. ¼ 144.74) from the otolith
where Lt is the length at age t, L∞ is the asymptotic length, g is the center in the ocean ecotype and 858.2 mm (s.d. ¼ 68.34) in the
instantaneous growth rate when t ¼ t0, and t0 corresponds to the stream ecotype (Fig. 6). None of the profiles of spawning in-
inflection point of the curve. The model was fitted using a non- dividuals or carcasses showed a decrease in the Sr:Ca ratio, which
linear Marquardt curve-fitting algorithm assuming additive er- would have indicated an extended return to freshwater habitats.
rors. Likelihood-ratio tests (Kimura, 1980) were used to compare The distribution of back-calculated fork length at river emigra-
the growth parameters of the ocean and stream types. tion showed a mode of 14 cm for the ocean ecotype and 30 cm for
the stream ecotype. For the ocean ecotype, the range was between
12 and 28 cm, with an average size of 18.2 cm (s.d. ¼ 5.07); for the
2.4. Life-history characterization
stream ecotype, the range was between 16 and 38 cm, with an
average of 29.9 cm (s.d. ¼ 3.71). The two ecotypes differed signif-
The fork length at the time of river emigration was back-
icantly in their mean back-calculated size at river emigrations (t-
calculated from the otolith Sr:Ca profile using the formula pro-
test; d.f. ¼ 55; p < 0.0001) (Fig. 7).
posed by Campana (1990):
River residence time of the ocean ecotype ranged from 1 to 10
ðRi  Rc Þ months with a mode of 4 months, while river residence of the
Li ¼ Lc þ ðLc  L0 Þ
ðRc  R0 Þ
Table 1
where Li is the fork length (cm) at the time of river emigration, Ri is Proportions of ecotype between sexes and rivers.
the transverse otolith radius (mm) at the time of river emigration, Rc
Ecotype
is the transverse otolith radius (mm) at the time of capture, Lc is the
fork length (cm) at the time of capture, R0 and L0 are the transverse Ocean Stream
otolith radius (mm) and fork length (cm), respectively, corre- Sex Male 66.0% 60.0%
sponding to the initial otolith-body length ratio. For the latter two Female 34.0% 40.0%
parameters, values for chinook salmon in the northern hemisphere River Huemul 48.5% 43.1%
~
Nirehuao 51.5% 56.9%
were adopted: R0 ¼ 191 mm and L0 ¼ 3.1 cm (Zabel et al., 2010). The
 M. Araya et al. / Estuarine, Coastal and Shelf Science 149 (2014) 87e95 91

120 Table 3
Estimated growth parameters of the Gompertz models for the ocean and stream
100 ecotypes. C.I.: 95% confidence interval.
Fork length (cm)

Parameter Ocean (n ¼ 129) Stream (n ¼ 72)

80
Value C.I. Value C.I.
60 L∞ (cm) 105.28 100.93e109.64 93.03 87.83e98.22
G (year1) 0.83 0.73e1.18 1.32 0.91e1.73
40 T0 (year) 1.06 0.95e1.18 1.52 1.34e1.69

20

0 and the Futalelfú basins (Soto et al., 2007; Di Prinzio and Pascual,
0 1 2 3
Age (year)
Ocean-type Stream-type
Fig. 4. Relationship between age and fork length for the sampled chinook salmon,
showing the Gompertz models fitted separately to the two ecotypes.
stream ecotype ranged between 14 and 20 months, with a mode of
18 months (Fig. 8). Based on the Sr:Ca profiles, river emigration
occurred between August and March (Fig. 9), but the majority (78%)
of the stream ecotype emigrated between November and January.
The distribution of emigration dates differed significantly between
the two ecotypes (KolmogoroveSmirnov test; ks ¼ 0.6667;
p < 0.01).
4. Discussion
Our results indicate the presence of at least two main ecotypes
of invasive chinook salmon in the Ayse
n River basin. The emer-
gence of partial migration, regardless of differences related to the
invasion dynamics, has occurred elsewhere for introduced chinook
salmon including New Zealand (Flain and Glova, 1988; Unwin and
Lucas, 1993) and Argentina (Ciancio et al., 2005). The migration
patterns we found emulate closely those reported for chinook
salmon in their native range. Here chinook salmon juvenile show
two basic patterns of river residence (Gilbert, 1912), where ocean
type individuals commonly emigrate from rivers during the first
three months of life, while stream type individuals migrate to the
sea after spending a year or more in the river (Healey, 1991).
We do not know whether the predominance of the ocean
ecotype in the Ayse
n River watershed results from local condi-
tioning factors, or responds to a latitudinal or other broad-scale
regional differences. In the northern hemisphere, a transitional
area between the two ecotypes has been identified north of British
Columbia and southeast of Alaska (56
N). The stream ecotype
predominates northward, while the ocean ecotype predominates
4 5 2008). In contrast, the stream ecotype has been found to be the
dominant form in the few eastern Patagonia and Tierra del Fuego
basins studied until now (Ciancio et al., 2005; Ferna
ndez et al.,
2010). Although it is too soon to conclude there is a predomi-
nance of the ocean ecotype in Pacific watersheds, this expectation
arises from hypotheses that relate ecotype composition to relative
differences in individual growth conditions and to distances be-
tween spawning areas and river mouths (Taylor, 1990; Liss et al.,
2006). Beyond local variability expected from density-dependent
triggering factors (Chapman et al., 2011), stream ecotypes tend to
predominate in systems that exhibit spawning areas farther from
river mouths and provide relatively poor growth conditions,
perhaps due to low river temperatures and limited availability of
nutrients. Ocean ecotypes tend be more frequent in systems where
nursery areas are located nearer river mouths and/or where more
favorable growth conditions are found by early juveniles (Liss et al.,
2006). Overall, spawning areas are located closer to the river mouth
in western Patagonia than in eastern Patagonia. While most
spawning areas in the western side are <100 km from the river
mouth, spawning areas in eastern Patagonia can be located hun-
dreds of kilometers from the ocean, as reported in the Santa Cruz
basin (Ciancio et al., 2005).
The Sr:Ca profiles of the chinook salmon otoliths exhibited
patterns and concentrations similar to those reported in other
studies of the same species (Miller and Kent, 2009; Miller et al.,
2011). These patterns reflected transition from fresh to seawater
consistent with expectations and previous laboratory (Kalish, 1990;
Fowler et al., 1995; Zimmerman and Nielsen, 2003; Volk et al.,
2010) studies. Otolith Sr:Ca results were also highly consistent
with results from our analysis of circuli patterns in scales, indi-
cating a delayed migration in ocean-type fish. It must be considered
that all our Sr:Ca profiles corresponded to returning (surviving)
adults. Therefore, they do not represent, necessarily, the migratory
patterns experienced by juveniles.
50

southward (Healey, 1983; Taylor, 1990). A latitudinal increase in the

Absolute growth rate (cm yr -1)

prevalence of the stream ecotype has also been reported in New 40

Zealand (Taylor, 1990).
Within Patagonia, the predominance of the ocean ecotype varies 30
between Pacific and Atlantic rivers. The ocean ecotype has been
observed in three out of four watersheds invaded by chinook 20
salmon and studied in western Patagonia: the Petrohue, Corcovado
10
Table 2
Likelihood-ratio tests comparing the estimated Gompertz parameters for stream
type (S) and ocean type (O) chinook salmon. d.f.: degrees of freedom. 0
0 1 2 3 4 5
Null hypothesis d.f. p-value
Age (years)
S (L∞ , g, t0) ¼ O (L∞, g, t0) 3 <0.0001
S (L∞) ¼ O (L∞) 1 0.0011
Ocean type Stream type
S (g) ¼ O (g) 1 0.0055
S (t0) ¼ O (t0) 1 <0.0001
Fig. 5. Absolute growth rate by age for the two chinook salmon ecotypes.
92 M. Araya et al. / Estuarine, Coastal and Shelf Science 149 (2014) 87e95

Fig. 6. Sr:Ca profiles along the dorsal axis in chinook salmon otoliths representing ocean-type and stream-type individuals.

The very fast transition from freshwater to seawater suggested
by Sr:Ca ratios was somewhat unexpected. The Aysen fjord is an
extensive, highly stratified and productive estuarine system that
could sustain chinook salmon juveniles for several months.
Moreover, immature fish up to 45 cm were captured by us in the
fjord. Thus, although a rapid outmigration to the sea cannot be
discarded, it is possible that the Sr:Ca ratio was an unsuitable tracer
to detect residence at this intermediate salinity waters (Walther
and Limburg, 2012). Sr is expected to mix conservatively across

0.4

0.6
Relative frequency

0.3 0.5
Relative frequency

0.4
0.2
0.3

0.2
0.1
0.1

0 0
10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 2 4 6 8 10 12 14 16 18 20 22
Fork length at river emigration (cm) Age at river emigration (months)

Ocean type Stream type Ocean type Stream type

Fig. 7. Back-calculated fork lengths at river emigration for ocean-type and stream-type Fig. 8. Frequency distribution of the age at river emigration for ocean type and stream
individuals. type individuals.
 M. Araya et al. / Estuarine, Coastal and Shelf Science 149 (2014) 87e95
0.5
Relative frequency
0.4
0.3
0.2
0.1
0 in spawning and migratory runs, and a large variability of habitat
Aug Sep Oct Nov Dec Jan Feb Mar
Month
Ocean type Stream type
Fig. 9. Month at river emigration for ocean type and stream type individuals.
estuarine gradients, and to exhibit a good correlation with salinity
(Walther and Limburg, 2012), which was, in fact, observed in a
complementary work conducted by us in the study area (data not
shown here). However, when Sr and Ca mixing dynamics are
combined as a ratio in the otolith, the ratio tends to show an
asymptotic relationship at salinities above 8e10 (Kraus and Secor,
2004; Miller et al., 2010). This asymptotic response associated to
a rapid transition of juveniles from fresh to brackish-water layers
once in the estuary may explain the absence of intermediary signals
in the otoliths.
Our limited ability to estimate residence time of chinook salmon
smolts in the Aysen Fjord and adjacent areas was unfortunate. The
Patagonian Estuarine System is a highly productive area
(Niklitschek et al., 2014), where early chinook salmon immigrants
(ocean-type fish) may reach much higher growth rates than their
stream-type counterparts in freshwater nursery areas. Ocean-type
fish exhibited, in fact, higher size at all ages, particularly, at ages 1
and 2. Nonetheless, we were unable to learn whether this enhanced
juvenile growth was related to estuarine or fully marine waters,
which would allow much improved understanding on the under-
lying factors explaining invasion success and the extent of the in-
vasion's impacts upon the native fauna.
A second difficulty we faced was to detect immigration from the
ocean to the river in the otoliths, as Sr:Ca profiles remained rela-
tively constant near the otolith edge in all sampled adults. This
unexpected pattern may have resulted from insufficient sampling
resolution near the otolith border and/or from curtailed somatic
and otolith growth after the fish re-entered the river. Reduced so-
matic and otolith growth in freshwater would occur if fish entrance
into the freshwater system occurred shortly before spawning, a
likely situation given the relatively short distances existing be-
tween the two studied spawning areas and the sea, and the absence
of lakes along both rivers. As an alternative explanation, this
apparent stability of the Sr:Ca ratio could be an artifact of physio-
logical changes associated with reproduction. The levels of free Ca
may be reduced due to an increase in the Ca bound by plasma
proteins, thus increasing the Sr:Ca ratio in the endolymph and in
the otolith, obscuring the influence of diminished ambient Sr levels
in the river (Kalish, 1991).
The successful invasion of chinook salmon in eastern and
western Patagonia raises various questions that are relevant both to
understanding biological invasions and to resource and biodiversity
management in Patagonia. The presence of abiotic conditions
similar to those of original habitats and the absence or low abun-
dance of predators and competitors may have facilitated the
establishment of different salmonid species in Patagonia (Soto
et al., 2001, 2006). However, the species-specific factors that have
93
enabled the chinook salmon to be a more successful invader than
other introduced or escaped Pacific salmons in the area remains
unknown. One such factor appears to be the greater plasticity and
life-history diversity of chinook salmon (Riva-Rossi et al., 2012),
which leads to a grater spread of the risk of mortality in fluctuating
environments (den Boer, 1968; Kerr et al., 2010). Our results sup-
port the idea that chinook salmon populations established in
Patagonia maintain a considerable portion of the life cycle plasticity
observed in their natural range. Such plasticity exceeds by far the
existence of two ecotypes, as it includes a large temporal variability
Apr May Jun Jul use patterns within ecotypes (Healey, 1991).
Of particular interest to our results, is chinook salmon's capa-
bility to advance or delay emigration of juveniles from freshwater.
This is a particular dimension of partial migration behaviors, which
has not been fully investigated in terms of its costs and benefits
(Chapman et al., 2012). Early emigrants (ocean ecotype) may escape
sooner from low growth/low survival conditions in the extremely
oligotrophic and highly variable freshwater areas of western Pata-
gonia. These early emigrants would benefit from a prompter use of
the much more productive and stable Patagonian Estuarine System,
where higher growth rates and larger size at age would be expected
(Jessop et al., 2008; Kerr and Secor, 2009). This expectation is
actually supported by our data that show ocean-type fish were
larger than stream-type ones at all ages. There is, however, a trade-
off between the bioenergetic advantages of an early migration and
the increased risk of predation expected for smaller juveniles
arriving into the estuarine system (Beamish and Mahnken, 2001;
Chapman et al., 2012). From a rather different point of view, the
earlier spawning schedule and the ability to emigrate sooner from
freshwater, would be a key factor reducing overlap and inter-
specific competition with juveniles of brown and, mainly, rainbow
trout, already established in Patagonian watersheds several decades
before the ongoing chinook salmon invasion (Arismendi et al., in
press). O’Neal and Stanford (2011) documented an important posi-
tive feedback between invasive ecology and partial migration in
introduced brown trout. Successful colonization by principally
resident forms resulted in loss of forage bass and emergence of sea-
run brown trout. Increased ocean ecotypes of chinook salmon in
Chilean systems may have resulted from similar feedbacks.
Important ecological consequences of partial migration include
niche partitioning, compensating for a poor start, capitalizing on a
good start, reproductive success among ecotypes, and colonization
success. Partial migration is common in fishes, a latent attribute
that can become more fully expressed in novel environments
(Secor, in press). To better understand and anticipate the impacts of
invasive chinook salmon upon Patagonian ecosystems and partic-
ular native populations, it is necessary to increase our knowledge of
partial migration in the diverse watersheds of Patagonia (Pascual
and Ciancio, 2007).
Despite the progress made in the present study, important as-
pects of the biology and life cycle of chinook salmon in the Ayse
n
River basin remain unknown. These aspects include fertility, mor-
tality, reproductive migration seasonality (run timing), population-
abundance, population (genetic) structure, habitat-use patterns
and trophic ecology in estuarine, coastal and open-ocean habitats.
Additional information on the biology of this and other chinook
salmon populations in Patagonia is urgently needed because the
invasion process across this vast region remains as a large-scale
ongoing process, whose effects are still poorly understood. Still,
existing knowledge gaps cannot justify the remarkable lack of
public action regarding this invasion. It is imperative to design,
evaluate and implement immediate remedial management mea-
sures for the control or removal of this species, particularly in
protected and/or vulnerable areas of Patagonia.
94 M. Araya et al. / Estuarine, Coastal and Shelf Science 149 (2014) 87e95
Acknowledgements
This work was supported by a grant to M. Araya of CONICYT.
Additional funds were provided by Chilean Fisheries Research Fund
(FIP) grant N
2008-30, University of Antofagasta, Arturo Prat
University and the Regional Government of Tarapaca. Laboratories
and specialized equipment was contributed by universities Austral
of Chile, Los Lagos (Chile) and Maryland (USA). The authors thank
Chris Conroy for help with the preparation of otoliths for Sr:Ca
analysis.
References
Altukhov, Y.P., Salmenkova, E.A., Omelchenko, V.T., 2000. Salmonid Fishes. Popu-
lation Biology, Genetics and Management. Blackwell Science Ltd, Oxford, p. 354.
Arismendi, I., Soto, D., 2012. Are salmon-derived nutrients being incorporated in
food webs of invaded streams? Evidence from southern Chile. Knowl. Manag.
Aquat. Ecosyst. 405, 01.
Arismendi, I., Penaluna, B.E., Dunham, J.B., de Leaniz, C.G., Soto, D., Fleming, I.A.,
Gomez-Uchida, D., Gajardo, G., Vargas, P.V., Leo
n-Mun ~ oz, J., 2014. Differential
invasion success of salmonids in southern Chile: patterns and hypotheses. in
press Rev. Fish. Biol. Fish.. http://dx.doi.org/10.1007/s11160-014-9351-0. Online.
Beamish, R.J., Mahnken, C., 2001. A critical size and period hypothesis to explain
natural regulation of salmon abundance and the linkage to climate and climate
change. Prog. Oceanogr. 49, 423e437.
Becker, L.A., Pascual, M.A., Basso, N.G., 2007. Colonization of the Southern Patagonia
ocean by exotic chinook salmon. Conserv. Biol. 21, 1347e1352. http://dx.doi.org/
10.1111/j.1523-1739.2007.00761.x.
Billard, R., Jensen, J.O.T., 1996. Gamete removal, fertilization and incubation. In:
Pennell, W., Barton, B.A. (Eds.), Developments in Aquaculture and Fisheries
Science, Principles of Salmonid Culture, vol. 29. Elsevier, Amsterdam,
pp. 291e363.
Buschmann, A.H., Riquelme, V.A., Herna
ndez-Gonz
alez, M.C., Varela, D.A.,
Jime
nez, J.E., Henríquez, L.A., Vergara, P.A., Guin
~ ez, R., Filún, L., 2006. A review of
the impacts of salmonid farming on marine coastal ecosystems in the southeast
Pacific. ICES J. Mar. Sci. 63, 1338e1345.
Campana, S.E., 1990. How reliable are growth back-calculation based on otoliths?
Can. J. Fish. Aquat. Sci. 47, 2219e2227.
Chapman, B.B., Bro € nmark, C., Nilsson, J.-A., Hansson, L.-A., 2011. The ecology and
evolution of partial migration. Oikos 120, 1764e1775.
Chapman, B.B., Hulthen, K., Brodersen, J., Nilsson, P.A., Skov, C., Hansson, L.-A.,
Bro€ nmark, C., 2012. Partial migration in fishes II: causes and consequences.
J. Fish. Biol. 81, 456e478.
Ciancio, J.E., Pascual, M.A., Lancelotti, J., Rossi, C.M.R., Botto, F., 2005. Natural colo-
nization and establishment of a Chinook salmon, Oncorhynchus tshawytscha,
population in the Santa Cruz River, an Atlantic basin of Patagonia. Environ. Biol.
Fishes 74, 219e227.
Correa, C., Gross, M.R., 2008. Chinook salmon invade southern South America. Biol.
Invasions 10, 615e639.
Cury, P., 1994. Obstinate nature: an ecology of individuals. Thoughts on reproduc-
tive behavior and biodiversity. Can. J. Fish. Aquat. Sci. 51, 1664e1673.
Den Boer, P.J., 1968. Spreading of risk and stabilization of animal numbers. Acta
Biotheorical 18, 165e194.
Di Prinzio, C.Y., Pascual, M.A., 2008. The establishment of exotic Chinook salmon
(Oncorhynchus tshawytscha) in Pacific rivers of Chubut, Patagonia, Argentina.
Ann. De. Limnol. e Int. J. Limnol. 44, 25e32.
Elsdon, T.S., Wells, B.K., Campana, S.E., Gillanders, B.M., Jones, C.M., Limburg, K.E.,
Secor, D.E., Thorrold, S.R., Walther, B.D., 2008. Otolith chemistry to describe
movements and life-history measurements of fishes: hypotheses, assumptions,
limitations, and inferences using five methods. Oceanogr. Mar. Biol. Annu. Rev.
46, 297e330.
Ferna
ndez, D.A., Ciancio, J., Ceballos, S.G., Riva-Rossi, C., Pascual, M.A., 2010. Chinook
salmon (Oncorhynchus tshawytscha, Walbaum 1792) in the Beagle Channel,
Tierra del Fuego: the onset of an invasion. Biol. Invasions 12, 2991e2997.
Flain, M., Glova, G.J., 1988. A test of the reliability of otolith and scale readings of
chinook salmon (Oncorhynchus tshawytscha). N. Z. J. Mar. Freshw. Res. 22,
497e500.
Fowler, A.J., Campana, S.E., Jones, C.M., Thorrold, S.R., 1995. Experimental assess-
ment of the effect of temperature and salinity on elemental composition of
otoliths using laser ablation ICPMS. Can. J. Fish. Aquat. Sci. 52, 1431e1441.
Gilbert, C.H., 1912. Age at maturity of the Pacific coast salmon of the Genus
Oncorhynchus. Bull. United States Bureau Fish. 32, 1e22.
Gillanders, B.M., 2005. Using elemental chemistry of fish otoliths to determine
connectivity between estuarine and coastal habitats. Estuarine. Coast. Shelf Sci.
64, 47e57.
Healey, M.C., 1983. Coastwide distribution and ocean migration patterns of stream-
and ocean-type chinook salmon, Oncorhynchus tshawytscha. Can. Field-Natu-
ralist 97, 427e433.
Healey, M.C., 1991. Life history of Chinook salmon (Oncorhynchus tshawytscha). In:
Groot, C., Margolis, L. (Eds.), Pacific salmon Life Histories. University of British
Columbia Press, Vancouver, British Columbia, Canada, pp. 311e393.
Hendry, A.P., Castric, V., Kinnison, M.T., Quinn, T.P., 2004. The evolution of phil-
opatry and dispersal: homing versus straying in Salmonids. In: Hendry, A.P.,
Stearns, S.C. (Eds.), Evolution Illuminated, Salmon and Their Relatives. Oxford
Press, NY, pp. 52e90.
Jessop, B.M., Cairns, D.K., Thibault, I., Tzeng, W.N., 2008. Life history of American eel
Anguilla rostrata: new insights from otolith microchemistry. Aquat. Biol. 1,
205e216.
Kalish, J.M., 1990. Use of otolith microchemistry to distinguish the progeny of sym-
patric anadromous and non-anadromous salmonids. Fish. Bull. U.S 88, 657e666.
Kalish, J.M., 1991. Determinants of otolith chemistry: seasonal variation in the
composition of blood plasma, endolymph and otolith bearded rock cod Pseu-
dophycis barbatus. Mar. Ecol. Prog. Ser. 74, 137e159.
Kerr, L.A., Secor, D.H., 2009. Bioenergetic trajectories underlying partial migration in
Patuxent River (Chesapeake Bay) white perch (Morone americana). Can. J. Fish.
Aquat. Sci. 66, 602e612.
Kerr, L.A., Cadrin, S.X., Secor, D.H., 2010. The role of spatial dynamics in the stability,
resilience, and productivity of an estuarine fish population. Ecol. Appl. 20,
497e507.
Kimura, D.K., 1980. Likelihood methods for the von Bertalanffy growth curve. Fish.
Bull. U.S. 77, 765e776.
Koo, T.S.Y., Isarankura, A., 1967. Objective studies of scales of Columbia River Chi-
nook salmon, Oncorhynchus tshawytscha (Walbaum). Fish. Bull. U.S. 66,
165e180.
Kraus, R.T., Secor, D.H., 2004. Incorporation of strontium into otoliths of an estua-
rine fish. J. Exp. Mar. Biol. Ecol. 302 (1), 85e106.
Liss, W.J., Stanford, J.A., Lichatowich, J.A., Williams, R.N., Coutant, C.C., Mundy, P.R.,
Whitney, R.R., 2006. A foundation for restoration. In: Williams, R.N. (Ed.), Re-
turn to the River: Restoring Salmon to the Columbia River. Elsevier Academic,
Amsterdam, The Netherlands, pp. 51e98.
McDowall, R.M., 2003. Impacts of introduced salmonids on native galaxiids in New
Zealand upland streams: a new look at an old problem. Trans. Am. Fish. Soc.
132, 229e238.
Miller, J.A., Kent, A.J.R., 2009. The determination of maternal run time in Chinook
salmon (Oncorhynchus tshawytscha) based on Sr/Ca and 87Sr/86Sr within
otolith cores. Fish. Res. 95, 373e378.
Miller, J.A., Gray, A., Merz, J., 2010. Quantifying the contribution of juvenile
migratory phenotypes in a population of Chinook salmon (Oncorhynchus
tshawytscha). Mar. Ecol. Prog. Ser. 408, 227e240.
Miller, J.A., Butler, V.L., Simenstad, C.A., Backus, D.H., Kent, A.J.R., 2011. Life history
variation in Columbia River Chinook salmon (Oncorhynchus tshawytscha): a
comparison using modern and ~500 yr-old archaeological otoliths. Can. J. Fish.
Aquat. Sci. 68, 603e617.
Niklitschek, E., Aedo, E., 2002. Estudio del ciclo reproductivo de las principales
especies objetivo de la pesca deportiva en XI Regio
n. [Study of the reproductive
cycle of principal species for recreational fishing in Region XI]. Universidad
Austral de Chile, Coyhaique, Chile, p. 68. Reporte Final FIP.
Niklitschek, E.J., Soto, D., Lafon, A., Molinet, C., Toledo, P., 2013. Southward expan-
sion of the Chilean salmon industry in the Patagonian Fjords: main environ-
mental challenges. Rev. Aquac. 5, 172e195.
Niklitschek, E.J., Secor, D.H., Toledo, P., Valenzuela, X., Cubillos, L.A., Zuleta, A., 2014.
Nursery systems for Patagonian grenadier off Western Patagonia: large inner
sea or narrow continental shelf? Ices J. Mar. Sci. 71, 374e390.
O'Neal, S.L., Stanford, J.A., 2011. Partial migration in a robust brown trout population
of a Patagonian river. Trans. Am. Fish. Soc. 140, 623e635.
O’Ryan, R., Niklitschek, M., Niklitschek, E., Ulloa, A., Gligo, N., 2010. Trade liber-
alization, rural poverty and the environment: a case study of the Forest and
Salmon sectors in Chile. In: Cook, J., Cylke, O., Larson, D., Nash y, J., Stedman-
Edwards, P. (Eds.), En Vulnerable Places, Vulnerable People: Trade Liberaliza-
tion, Rural Poverty y the Environment. Elgar, Northampton, MA.
Pascual, M.A., Ciancio, J.E., 2007. Introduced anadromous salmonids in Patagonia: risks,
uses, and a conservation paradox. In: Bert, T.M. (Ed.), Ecological and Genetic Im-
plications of Aquaculture Activities. Springer, New York City, New York, USA.
Pizarro, R., Zolezzi, C., 2003. Impactos ambientales del escape de salmo
nidos.
[Environmental impacts of escaped salmonids]. In: An
alisis de políticas Púb-


licas. Publicaciones Terram. Serie APP, N 22. http://www.terram.cl.
Ricker, W.E., 1979. Growth Rates and Models. In: Fish Physiology VIII Series. Aca-
demic Press, San Diego, CA, pp. 677e743 (Chapter 11).
Riva-Rossi, C.M., Pascual, M.A., Aedo Marchant, E., Basso, N., Ciancio, J.E., Mezga, B.,
Fern
andez, D.A., Ernst-Elizalde, B., 2012. The invasion of Patagonia by Chinook
salmon (Oncorhynchus tshawytscha): inferences from mitochondrial DNA pat-
terns. Genetica 140, 439e453.
Secor, D.H., Piccoli, P.M., 1996. Age- and sex-dependent migrations of the Hudson
River striped bass population determined from otolith microanalysis. Estuaries
19, 778e793.
Secor, D.H., Piccoli, P.M., 2007. Oceanic migration rates of Upper Chesapeake Bay
striped bass (Morone saxatilis), determined by otolith microchemical anal-
ysis2007. Fish. Bull. 105, 62e73.
Secor, D.H., Kerr, L.A., Cadrin, S.X., 2009. Connectivity effects on productivity, sta-
bility, and persistence in a herring metapopulation model. ICES J. Mar. Sci. 66,
1726e1732.
Secor, D.H., 2014. Migration Ecology of Marine Fishes. The Johns Hopkins University
Press (in press).
SERNAPESCA, 2013. Servicio Nacional de Pesca, Anuario Estadístico. [National
Fishing Service Statistical Yearbook]. Ministerio de Economía, Fomento y
Reconstruccio
n, Valparaíso.
 M. Araya et al. / Estuarine, Coastal and Shelf Science 149 (2014) 87e95
Soto, D., Jara, F., Moreno, C., 2001. Escaped salmon in the inner seas, southern Chile:
facing ecological and social conflicts. Ecol. Appl. 11, 1750e1762.
Soto, D., Arismendi, I., Gonza
lez, J., Sanzana, J., Jara, F., Jara, C., Guzm

an, E., Lara, A.,
2006. Southern Chile, trout and salmon country: invasion patterns and threats
for native species. Rev. Chil. Hist. Nat. 79, 97e117.
Soto, D., Arismendi, I., Di Prinzio, C., Jara, F., 2007. Establishment of chinook salmon
(Oncorhynchus tshawytscha) in Pacific basins of southern South America and its
potential ecosystem implications. Rev. Chil. Hist. Nat. 80, 81e98.
Taylor, E.B., 1990. Environmental correlates of life-history variation in juvenile
chinook salmon, Oncorhynchus tshawytscha (Walbaum). J. Fish. Biol. 37, 1e17.
Unwin, M.J., Lucas, D.H., 1993. Scale characteristics of wild and hatchery chinook
salmon (Oncorhynchus tshawytscha) in the Rakaia river, New Zealand, and their
use in stock identification. Can. J. Fish. Aquat. Sci. 50, 2475e2484.
95
Volk, E.C., Bottom, D.L., Jones, K.K., Simenstad, C.A., 2010. Reconstructing juvenile
Chinook Salmon life history in the Salmon river estuary, Oregon, using otolith
microchemistry and microstructure. Trans. Am. Fish. Soc. 139 (2), 535e549.
Walther, B.D., Limburg, K.E., 2012. The use of otolith chemistry to characterize
diadromous migrations. J. Fish. Biol. 81, 796e825.
Zabel, R., Haught, K., Chittaro, P., 2010. Variability in fish size/otolith radius re-
lationships among populations of Chinook salmon. Environ. Biol. Fishes 89,
267e278.
Zimmerman, C.E., Nielsen, R.L., 2003. Effect of analytical conditions on the mea-
surement of strontium to calcium (Sr/Ca) ratios in otoliths of anadromous
salmonids using wavelength dispersive electron microprobe analysis. Fish. Bull.
101, 712e718.