© PLOSI o -
OPEN 3 ACCESS Freely available online
Connectivity between Migrating and Landlocked
Populations of a Diadromous Fish Species Investigated
Using Otolith Microchemistry
Ingrid Tu lp 1*, M arieke Keller1, Jacques Navez2'3, Hendrik V. W inter1, M artin de G raaf1, W illy Baeyens2'3
1 Institute for Marine Resources and Ecosystem Studies (IMARES), W ageningen University, IJmuiden, The Netherlands, 2 D epartm ent of Analytical and Environmental
Chemistry, Vrije Universiteit Brussel, Brussels, Belgium, 3 Section d e Pétrographie-Minéralogie-Géochimie, Koninklijk Museum voor Midden-Afrika, M usée Royal d e
l'Afrique Centrale, Tervuren, Belgium
Abstract
Smelt Osmerus eperlanus has tw o diffe re nt life history strategies In the Netherlands. The m igrating population Inhabits the
W adden Sea and spawns In freshwater areas. After the closure o f the A fsluitd ijk In 1932, part o f the smelt population
became landlocked. The fresh w ater smelt population has been In severe decline since 1990, and has strongly negatively
Impacted the numbers o f piscivorous w ater birds relying on sm elt as th e ir main prey. The lakes th a t were form ed after the
dike closure, Ijsselmeer and Markermeer have been assigned as Natura 2000 sites, based on th e ir Im portance fo r (among
others) piscivorous w ater birds. Because o f the declining fresh w ater sm elt population, the question arose w h ethe r this
population Is still supported by the diadrom ous population. O pportunities fo r exchange between fresh w ater and the sea
are however lim ited to discharge sluices. The relationship between the diadrom ous and landlocked sm elt population was
analysed by means o f o to lith m lcrochem lstry. Our Interpretation o f o to lith strontiu m (88Sr) patterns from sm elt specimens
collected In the fresh w ater area o f Lake Ijsselmeer and Markermeer, compared to those collected In the nearby marine
environm ent, Is tha t there Is currently no evidence fo r a substantial co n trib u tio n from the diadrom ous population to the
spawning stock o f the landlocked population.
C itation : Tulp I, Keller M, Navez J, Winter HV, d e Graaf M, e t al. (2013) Connectivity betw een Migrating and Landlocked Populations of a Diadromous Fish Species
Investigated Using Otolith Microchemistry. PLoS ONE 8(7): e69796. doi:10.1371/journal.pone.0069796
Editor: Sadie Jane Ryan, SUNY College o f Environmental Science and Forestry, United States of America
Received March 9, 2013; A ccepted June 17, 2013; Published July 29, 2013
C opyrigh t: © 2013 Tulp e t al. This is an open-access article distributed under the term s o f th e Creative Commons Attribution License, which permits unrestricted
use, distribution, and reproduction in any medium, provided th e original author and source are credited.
Funding: This research is part o f th e strategic research program KBIV "Sustainable spatial developm ent of ecosystems, landscapes, seas and regions" which is
funded by th e Dutch Ministry o f Economic Affairs, Agriculture and Innovation, and carried ou t by W ageningen University Research centre. Collection of som e of
th e material was m ade possible by th e National Program Sea and Coastal Research from The Netherlands Organisation for Scientific Research under contract nr
839.08.240. The research is also supported by th e Dutch research program ANT (Autonome N eerwaartse Trends) lead by Deltares (The Netherlands) and by the
Flemish Science Foundation (FWO) grant FWO G.0.096.08.N.10 via th e acquisition of th e LA-ICPMS instrument. The funders had no role in study design, data
collection and analysis, decision to publish, or preparation of th e manuscript.
C om peting Interests: The authors have declared th a t no com peting interests exist.
* E-mail: ingrid.tulp@wur.nl
Introduction
passage. D iadrom ous sm elt still inh ab it coastal w aters a n d the
W ad d e n Sea a n d can grow u p to 25 cm a n d eight years old, a n d
becom e m atu re a t 3 - 4 years [2]. T h e landlocked sm elt m orphs
becom e m atu re after only one y ear a n d stay sm aller (10-12 cm)
th a n a n ad ro m o u s sm elt m orphs [4]. Sm elt spaw n in early spring in
freshw ater, preferably in ru n n in g w ater in tributaries o f lakes or
along shallow shores o f lakes a n d rivers, on firm surfaces such as
sand, gravel a n d stones [3]. T h e m ajority o f landlocked smelt
spaw n only once a n d individuals bigger th a n 12 cm (and
p resum ably older th a n 2 years) m ake up a ro u n d 1 % o f the entire
landlocked po p u latio n (Drost & de W itte unpubl.).
T h e smelt p opulations o f Ijsselm eer a n d M a rk erm e er have b een
declining strongly since 1990 [5]. In w inter, the Lake Ijsselm eer
a rea is h om e to the largest po p u latio n o f w aterfow l in W estern
E urope [6], Based o n the n u m b e r o f (am ong o th er groups) fish
eating birds b o th lakes a re p ro tec te d as w etlands o f international
im p o rtan ce for w ater birds in the R a m sa r C onv en tio n a n d have
b e en assigned as N a tu ra 2000 areas. Sm elt fisheries used to be an
im p o rta n t activity locally a n d a re carried out du rin g the spaw ning
season, from m id -F eb ru ary till m id-M arch, w h en large num bers
aggregate n e ar h a rd substrate areas, such as dykes [7]. B ecause o f
S m elt Osmerus eperlanus occurs a lo n g th e coast o f E u ro p e
from N o rth W est R u ssia in the n o rth , in clu d in g the W h ite
S ea a n d Baltic Sea, to th e B ay o f B iscay in th e south [1],
A n a d ro m o u s sm elt resides in estu aries originally, b u t several
p o p u latio n s h a v e b e co m e lan d lo c k ed d u e to geological
processes o r h a b ita t a lte ra tio n s b y m a n [2]. M a n y fish species
a re k n o w n for th e ir large p h e n o ty p ic plasticity in life history
traits in re la tio n to e n v iro n m e n ta l c h aracteristics. S im ilarly
sm elt has show n th e ability to a d a p t to n ew circ u m stan c es in
m a n y locations. S m elt c a n show diverse life histo ry traits
a m o n g p o p u latio n s, ra n g in g from m ig ra to ry (diadrom ous) to
re sid en t in fresh w ater [3].
Sm elt was a b u n d a n t in the form er Z uiderzee estuary (The
N etherlands) before its closure w ith a large d a n i (Afsluitdijk) in
1932 (Fig. 1). A fter the Afsluitdijk was built, the Z uiderzee estuary
was closed off from the sea resulting in the freshw ater Lake
Ijsselm eer, in w hich a landlocked po p u latio n o f sm elt developed.
M igrating opportunities b e cam e severely red u ced to instants w hen
d a n i sluices are o p e n ed a n d w ater c u rre n t is low e n ough to allow
PLOS ONE I www.plosone.org
1
July 2013 I Volume 8 | Issue 7 | e69796
Connectivity in Migrating and Landlocked Smelt
North S ea
W adden
S ea
IJsselm ee?
Afsluitdijk
M arkerm ee
0 ®
IJssel
2 5 .5
km
Figure 1. M ap o f th e study area showing th e locations w here samples w ere taken.
d o i:1 0 .1 3 7 1 /jo u rn a l.p o n e .0 0 6 9 7 9 6 .g 0 0 1
A m eth o d com m only used to distinguish freshw ater from
m arin e origin in fish is co m p arin g the S trontium ( Sr) profiles
along sagittal otoliths [8,9,10,11,12]. T h e utility o f this m eth o d
stems from the fact th a t otoliths a re largely m etabolically inert,
deposit a th in layer calcium c a rb o n a te continuously w ith age a n d
reflect a m b ie n t chem istry [13]. T h e origin o f smelt collected at
different locations in Lake Ijsselm eer a n d M a rk erm e er d u rin g the
peak spaw ning p erio d was investigated, b y co m p arin g the
S tro n tiu m (88Sr) signature o f their sagittal otoliths w ith those o f
specim ens collected in the D u tc h W ad d e n Sea (Fig. 1). If the
d iadrom ous p o p u latio n still contributes to the landlocked p o p u ­
lation, we expected to find individuals w ith a m arin e signal in their
88Sr otolith profiles in the spaw ning period. T h e 88Sr c o n ce n tra ­
tions in the edge o f the otoliths (the p a rt m ost recently deposited) o f
W ad d e n Sea individuals was used as a m arin e reference. T hese
concentrations w ere c o m p a red to 88Sr concentrations in specim ens
collected in b o th lakes at spaw ning tim e.
low biom ass, smelt fishery was closed d u rin g five years in the last
d ecade (unpubl. data).
T h e life histo ry o f a n a d ro m o u s fish d e p en d s o n b o th the
q u a lity o f th e ir h a b ita t a n d th e co n n ectiv ity b e tw ee n h a b ita ts
A fter th e c re a tio n o f b o th lakes tw o possible dev elo p m en ts m ay
h ave o c cu rre d : 1) sm elt b e c a m e se p a ra te d in a lan d lo c k ed a n d
a d ia d ro m o u s p o p u la tio n , w ith th e lan d lo c k ed p o p u la tio n
closing its life cycle w ith in th e lakes a n d the d iad ro m o u s
p o p u la tio n using spaw ning sites elsew here in a rea s a d jo in in g
th e W a d d e n Sea; 2) th e tw o p o p u latio n s a re m ix ed a n d the
d iad ro m o u s p o p u la tio n consists o f th e individuals w a sh e d o u t
th ro u g h th e sluices; re tu rn in g from th e W a d d e n Sea to L ake
Ijsse lm e er is lim ite d to events w h e n sluices in th e dyke are
o p e n ed .
G iven the decline in the landlocked smelt a n d the potential for
recovery o f the p opulation, the question w e aim to answ er is: do
an ad ro m o u s sm elt still c ontribute to the total spaw ning biom ass o f
sm elt in L ake Ijsselm eer a n d M arkerm eer? C ould the landlocked
sm elt stock be resupplied w ith W ad d e n Sea smelt? Insight in the
im p o rtan ce o f the m arin e stock for the landlocked stock w ould
con trib u te to the u n d e rstan d in g o f the po p u latio n dynam ics o f
smelt, a n d consequently o f those b ird species th a t rely o n smelt as a
food source.
PLOS ONE I www.plosone.org
Methods
Study Area
T h e estuary Z uiderzee was closed off from the D u tc h W ad d e n
Sea sea in 1932 by a d a n i resulting in the form ation o f the
2
July 2013 I Volume 8 | Issue 7 | e69796
Connectivity in Migrating and Landlocked Smelt
freshw ater lake Ijsselm eer (Fig. 1). T h is was p a rt o f a larger plan
w ith a threefold aim : pro tectin g central N eth erlan d s from the
effects o f the N o rth Sea, increasing the D u tc h food supply w ith
new agricultural la n d a n d im proving w ater m an ag em en t by
creating a lake from the form er u ncontrolled salt w ater inlet. T h is
lake is now adays the largest freshw ater b o d y in the N etherlands. It
is a shallow lake w ith a m ean d e p th o f 4 m a n d w ith u p to 8 m
deep channels. T h e lake was furth er divided b y three la n d
reclam atio n projects, red u cin g its surface w ith 40% to ab o u t
1900 km~ [5]. T h e lake M a rk erm e er w ith a surface a rea o f
700 km~, was c rea te d after the construction o f a d a n i in 1975, w ith
the objective o f reclaim ing L ake M ark erm eer, a p lan th a t was later
a b an d o n ed .
Lake Ijsselm eer gets n u trie n t in p u t from the R iver IJssel, a
trib u ta ry o f the R iver R hine. P rim ary p ro d u c tio n a n d fish
p ro d u c tio n in Lake M a rk erm e er is low er th a n in L ake Ijsselm eer
[14,15]. B oth lakes have m ultiple functions, fisheries, recreation,
drinking w ater supply, tran sp o rt a n d as a rest a n d forage a re a for
birds [16], N u trien t reductions in the early 1980s lead to a d ro p in
p h o sp h ate levels. In Lake Ijsselm eer reten tio n tim e o f w ater is
a ro u n d 6 m onths, w hereas in Lake M a rk erm e er it is a ro u n d 12
m onths [5].
M ig ra to ry possibilities for sm elt to pass th e different d am s are
lim ited. In th e A fsluitdijk, tw o sluice g ro u p s (one in th e W est
a n d one in th e East) c o n tro l th e outflow o f fresh w a te r a n d th ere
a re tw o sh ip p in g locks. T h ese a re th e o nly d irec t c o n n ectio n s
b e tw ee n lake Ijsse lm e er w ith th e IJssel a n d R h in e riv er basin.
T h e w esterly sluice c o m p le x has th re e g ro u p s o f discharge
sluices w ith five tubes in e a c h gro u p . T h e e asterly sluice
c o m p le x has tw o g ro u p s w ith five tu b es each . E ac h tu b e is 12 m
w ide. In ste a d o f th e fo rm er o p e n in g o f 35 km , now adays only 25
sluices m ea su rin g 12 m form the c o n n ec tio n b e tw ee n the
W a d d e n S ea to Ijsse lm e er, a re d u c tio n o f 99% . T h e discharge
is g o v e rn ed b y ra in fall, so in sp rin g disch arg e is g enerally high
a n d c h an ces for in la n d m ig ra tio n , especially for w eak sw im m ers,
a re lim ite d to p e rio d s a t th e e n d o f disch arg e p e rio d s w h e n the
w a te r level outside a n d inside th e dike a re a t a sim ilar level.
M ig ra tio n possibilities h ave n o t c h a n g e d greatly since the
closure in 1932. T h e n e are st possibility for sm elt to e n te r a n
u n o b stru c te d fresh w ater a re a a n d spaw n is th e E m s-D o lla rd
estuary, w h ic h is ca 150 km E ast from th e A fsluitdijk. M ig ra tio n
possibilities b e tw ee n L ake Ijsse lm e er to L ake M a rk e rm e e r a n d
reverse a re lim ite d to tw o com plexes, a t th e N o rth side th e re are
tw o ship p in g locks a n d a disch arg e sluice (two tu b es o f 20 m
wide) a n d a t th e S o u th side th e re is a sh ip p in g lock a n d a
disch arg e sluice (six tu b es o f 16 m wide).
M e a n m onthly b ack g ro u n d 88Sr values v ary betw een Ijsselm eer
a n d M ark erm eer, w ith low er levels du rin g sum m er th a n w inter
a n d a generally low er level in M a rk erm e er th a n in Ijsselm eer
(Fig. 2, d a ta sta n d ard m on ito rin g R ijksw aterstaat, c o ncentration
m easured in surface w ater). Sim ilar d a ta for the W ad d e n Sea are
not available.
550
500
"Sj
g
§
g
400
350
— MARKERMEER
IJSSELMEER
300
J
F
M
A
M
J
J
A
S
O
N
D
Figure 2. Seasonal variatio n o f Strontium concentrations in
M arkerm eer and Ijsselm eer in 2 0 0 9 (data Rijksw aterstaat).
do i:1 0.13 7 1 /jo u rn a l.p o n e .0 0 6 9 7 9 6 .g 0 0 2
m ens o f 11 a n d 16 cm), a n d 9 smelt (length 13 -2 3 cm) from the
W ad d e n Sea (2 hauls/locations) w ere collected. T h e perm ission to
collect fish in the areas u n d e r study was issued by the M inistry o f
E conom ic Affairs in the N etherlands, d e p artm e n t im plem entation
fisheries regulation (nr 60898). In T h e N eth erlan d s there is no
need for special approval to catch o r kill fish b y a n ethics
com m ittee (the experim ents o n anim als act, BW B R 0003081).
Sm elt is a fragile fish a n d m ost o f th em w ere already d e a d w hen
taken o n b o a rd . F or the few fish still alive, suffering was
am elio rated by im m ediate freezing.
T h e total length o f all specim ens w ere m easured ( ± 1 nun) a n d
otoliths w ere collected in the laboratory. O n e otolith from each
specim en was cleaned a n d e m b ed d ed in resin. T h e otoliths w ere
placed on a first layer o f nearly cu red resin, after w hich the otoliths
w ere fixed evenly w ith a second layer o f resin. E m b e d d ed otoliths
w ere sectioned th ro u g h the nucleus. T h e thickness o f the slides
ra n g ed b etw een 0.5 a n d 0.6 nini. T h ree resin strips, c o ntaining 10
sectioned otoliths each, w ere glued to a glass slide w ith clear resin.
N o coverslip was used. T h e slides w ere stored in covered
c ontainers until analysed b y laser ablation (LA) inductively
coupled m ass spectrom etry (ICPM S).
Laser Ablation-ICP-MS Analyses
A n IC P -M S (T herm o Scientific X-serie2) coupled to a laser (ESI
N ew W ave LTP-193 Fast E xcim er ArF) was used to d eterm ine the
co n cen tratio n o f 88Sr along sagittal otoliths. T h e N IS T 612
reference m aterial was used for the o ptim ization o f following
instrum ental param eters: the m ovem ent rate, the dwell tim e, the
flow o f A r, the flow o f H e, the o u tp u t a n d the re petition rate. T h e
choices w ere m ad e according to the intensities a n d stabilities
m easured. Logically, w ith a longer dwell tim e, the stability
increases. T h e configuration set allow ed a resolution o f 3.7 pm ,
an ablation d iam eter o f 25 p m (35 p m a t the p re -a b latio n point)
a n d a scan speed o f 25 pins ).
In o rd e r to b e tte r answ er o u r questions, a q u antitative analytical
m eth o d was developed instead o f w orking w ith 88S r / 44C a ratios.
T o do this, calibration standards N IS T 612 a n d 610 (lines o f
600 pm ) w ere chosen. Finally, to correct the drift a n d differences
in efficiency, the Cia was used as the in tern al standard. G iven the
size o f sm elt otoliths (ca 1 nun), the analysis is fast e n ough a n d the
extent o f standards was estim ated every 10 sam ples. T o calculate
concentrations, norm ally the standards enclosing the sam ples w ere
chosen. G iven the stability o f the m easures (r~ = 0.99), all the
c oncentrations w ere calculated based o n a line w ith all standards
m easured d u rin g one day. In p relim in ary tests, the resin itself was
Sample Collection
In M a rch 2010 smelt w ere sam pled w ith a 2,5 m b e a m traw l
(rigged w ith fine-m eshed cod-end; 0.5 cm) in Lake Ijsse lm e er/
M ark erm eer. All know n spaw ning aggregations w ere sam pled
(Fig. 1). A dult fish w ere collected from the W ad d e n Sea w ith a 3 m
b e am traw l (rigged w ith a fine-m eshed cod-end; 20 nini) betw een
A pril a n d O c to b er 2009 to serve as a m arin e reference. A fter
c ap tu re fish w ere kept frozen a n d taken to the lab o rato ry for later
dissection. All fish w ere m easured to the n earest n u n . In total 110
smelt, 56 from Ijsselm eer (collected in 4 hauls/locations) a n d 54
from M a rk erm e er (4 hau ls/lo catio n s, length 5 -7 cm , tw o speci-
PLOS ONE I www.plosone.org
450
Ü
3
July 2013 I Volume 8 | Issue 7 | e69796
Connectivity in Migrating and Landlocked Smelt
Wadden Sea
5000
4000
-
3000
-
2000
1000
5000
E
Q.
Q.
4000
CC
3000
-
00
co.
2000
-
1000
-
cn
H r
8
5000
-
4000
-
3000
2000
1000 H
500
1000
1500
500
1000
1500
500
1000
1500
distance to core(iunn)
Figure 3. Strontium concentration profiles (in ppm ) fo r sm elt caught in th e W adden Sea, going from th e nucleus (left) to the o uter
edg e (rig ht). T h e c o lo u rs in d ic a te th e d iffe re n t h a u ls, all p rofiles in o n e c o lo u r o rig in a te fro m th e s a m e haul.
d o i:1 0 .1 3 7 1 /jo u rn a l.p o n e .0 0 6 9 7 9 6 .g 0 0 3
m easured a n d it show ed a very special signature, com pletely
different from the otoliths. Such signatures w ere not observed in
the otolith m easurem ents. In addition, to avoid the risk o f
c o n tam in atio n b y the resin o r b y dust, the line o f m easurem ent
was situated in the m iddle o f the sam ple a n d avoided the extrem e
edges. M o reo v er a pre-ab latio n was carried out. T ests indicated
th a t a th ird ab lation did n o t differ from the second one. H en ce, all
results refer to the second ablation m easu rem en t th a t was always
p reced ed by a p re-ab latio n step. All the m easures are in T R A
m ode (time resolved analyses). T h e laser was led th ro u g h the core
o f the otolith.
PLOS ONE I www.plosone.org
Data Analyses
O f each profile cross-cutting the otolith from the rostrum ,
th ro u g h the nucleus, to the post-rostrum , the p a rt from the nucleus
to the ro stru m was used for analyses. 88Sr at the cores show the
conditions the fish e n co u n tere d lived w h en they b eg an to grow.
O nly the o u ter 50 m easurem ents (roughly representing 185 pm ) o f
W ad d e n Sea individuals w ere used to define the threshold for the
m arin e signal, to ascertain th a t the 88S r co n cen tratio n truly
rep resen ted the m arin e e nvironm ent a n d to avoid potential
ontogenetic effects in 88Sr deposition. Sim ilarly the last 50
m easurem ents (outer edge) o f the otoliths o f fish caught in the
4
July 2013 I Volume 8 | Issue 7 | e69796
Connectivity in Migrating and Landlocked Smelt
Ijsselm eer
1
I
2
3
4
5
6
8
7
2000
1500
*
1000
-
2000
-
9
10
11
12
13
14
15
17
18
19
20
21
22
23
1500
1000
16
mb*
24
2000
^
E
Q.
3
iw/'
1000
*ev*wv
wAt
25
26
27
28
29
30
31
32
40
2000
^
1500
cf )
1000 -
/ M
00
CO
2000
1500
1000
33
34
35
36
37
38
39
41
42
43
44
45
46
47
50
51
52
53
54
55
-
M jjjjIiJ J ™
'I ' p | P “ y
2000
1500 1000
Hàdi.
uaX1
49
2000
1500
1000
48
56
J ju ,
-
8CM ^8 80 C8DSO C8M 8 8© 8CO8O 8CM 8' t 80 8CO8O 8CM 8 80 8CD8O C8M 8-Í 80 80 8O 8CM8Tf 80 80 8O 8CM 8Tf 80 80 8O 8CM8 80 80 8O
distance to core(nm)
Figure 4. Strontium concentration profiles (in ppm ) fo r sm elt caught in Lake Ijsselm eer, going from th e nucleus (left) to th e o uter
edg e (rig ht). T h e c o lo u rs in d ic a te th e d iffe re n t h a u ls, all p rofiles in o n e c o lo u r o rig in a te fro m th e s a m e haul.
d o i:1 0 .1 3 7 1 /jo u rn a l.p o n e .0 0 6 9 7 9 6 .g 0 0 4
values below 2000 p p m a t the edge. T h e overall p a tte rn o f this
individual is how ever different from Ijsselm eer individuals, th at
never show v a lu e s> 2 0 0 0 p p m . T h e individual profiles o f all
specim ens collected in fresh w ater show ed 88S r levels ra nging
betw een 500 a n d 2000 pp m . Profiles o f specim ens in Lake
Ijsselm eer w ere m ore variable th a n those from Lake M arkerm eer.
In Lake Ijsselm eer m ore th a n h a lf o f the individuals show ed lowest
values at the core a n d a n increase in 88Sr c o n ce n tra tio n going
from the core to the edge, b u t never as high as the level identified
as m arin e (2260 ppm ) (Fig. 4). Individuals from the 3ld haul
(brown, 40 -5 4 ) h a d low er values at the core (500-1000) th a n m ost
specim ens from o th er hauls. Levels varied betw een 5 0 0 2000 p p m . T h e two largest anim als (Fig. 4, individuals 20 a n d
21) h a d a distinct p a tte rn : after constant flat profiles b o th show ed a
strong dip in 88Sr concentrations to v a lu e s< 1 0 0 0 p p m . Profiles
from Lake M a rk erm e er specim en’s otoliths w ere generally flat,
w ith no clear tre n d going from the core to the edge a n d no obvious
differences b etw een individuals originating from different hauls
(Fig. 5). 88Sr concentrations varied b etw een 5 0 0 -1 2 0 0 ppm .
M e a n 88Sr concentrations at the otolith edge w ere significantly
higher in the W ad d e n Sea th a n in b o th lakes, b u t also differed
betw een the two lakes, w ith lowest values found in Lake
M a rk erm e er (T able 1, Fig. 6). A t the core 88Srconcentrations
did not differ significantly b etw een a n y o f the areas (T able 1,
Fig. 6). A lthough m ea n core values w ere generally low er at lake
M a rk erm e er the difference was n o t statistically significant (Fig. 6).
lakes w ere chosen to represent the conditions a t tim e a n d shortly
before tim e o f capture.
S pecim ens o riginated from different hauls, b u t a varying
n u m b e r o f individuals w ere collected from each haul. T herefore,
the overall differences b etw een the concentrations o f 88Sr
co n cen tratio n in the section n e a r the edge o f the otoliths from
specim ens collected in the W ad d e n Sea a n d b o th lakes w ere tested
w ith a m ixed effect m odel w ith individual nested w ithin haul as
ra n d o m effect a n d location as fixed effect [17]. T h e ra n d o m
intercep t allow for correlations betw een observations from the
sam e site a n d individual. Because o f heterogeneity in the d ata, Sr
concentrations w ere log-transform ed p rio r to analyses.
All d a ta analyses w ere carried out in R 2.15.1.
Results
88S r profiles o f individuals caught in the W ad d e n Sea show ed
strong variations w ith 88Sr levels at the o u ter edge generally
varying b etw een 2000 a n d 6000 pp m , b u t not in the core (Fig. 3).
T h e core values w ere all below 2000 ppm . T h e level at the o u ter
edge, the p a rt o f the otolith th a t was m ost recently deposited, was
used as a reference for the m arin e environm ent. E ight out o f the
nine individuals show ed m ean values g re ater th a n 2000 in the final
50 m easurem ents (m ean o f the final 50 m easurem ents o f the 9:
2260 p p m (range 1 6 0 0-3100 ppm ). Individual 5 from the
W ad d e n Sea has a different profile from the o th er eight, show ing
PLOS ONE I www.plosone.org
5
July 2013 | Volume 8 | Issue 7 | e69796
Connectivity in Migrating and Landlocked Smelt
Markermeer
1 1 1 1 1
1
i i i i i
2
i i i i i
3
l l l l l
4
i i i i i
5
i i i i i
6
i i i i i
7
i i i i i
8
l l l l l
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
39
40
41
42
43
44
45
53
54
2000
1500
1000
V
2000
1500
1000
2000
^
§;
1500
1000
E 1
CO
CO
CO
2000
1500
1000
-irtffK»,
M L éééJU
'I' 111 H P 1* 1
37
38
2000
1500
jyl
«L.
1000
47
48
i i i i i
lí
I I I I I
46
49
m J
50
52
51
2000
1500
1000
/ S
- w \
i i i i i
88888
CM
^
0
00
O
88888 88888
CM
' í
C D
C O
O
CM
' í
0
CO
O
i i i i i
i i i i i
88888
CM
' í
C D
iNfS
N fw W iI
ik A v l
C O
O
i i i i i
88888
CM
T f
0
C O
O
i i i i i
88888 88888
C V J ^ C O C O O
CM
^
C O
00
O
...................
88888
CM
’t
O
C O
O
* r‘
'■
88888
CM
^
0
00 O
distance to core(|j.m)
Figure 5. Strontium concentration profiles (in ppm ) for sm elt caught in Lake M arkerm eer, going from the nucleus (left) to the o uter
edg e (rig ht). T h e c o lo u rs in d ic a te th e d iffe re n t h a u ls, all p rofiles in o n e c o lo u r o rig in a te fro m th e s a m e haul.
d o i:1 0 .1 3 7 1 /jo u rn a l.p o n e .0 0 6 9 7 9 6 .g 0 0 5
on the m ark et o r auction anym ore (unpubl.data). A scenario in
w hich the larger diadrom ous sm elt e n te r Lake Ijsselm eer a n d
possibly even M a rk erm e er for a very short p e rio d a n d in low
num bers is possibly. O riad ro m o u s sm elt m ay quickly pass Lake
Ijsselm eer a n d h e a d to historically know n spaw ning sites, such as
R iver IJssel [19]. In such scenario the recruits still c ontribute to the
po p u latio n in L ake Ijsselm eer. G iven the fact th a t older a n d larger
individuals in m ost fish species (and also show n in smelt) contribute
disproportionally to the recru itm en t [20,21], a con trib u tio n from
the d iadrom ous to the landlocked m o rp h s is possible, even if only a
few larger sized individuals spaw n in Lake Ijsselm eer or
M a rk erm e er.Since landlocked smelt populations are generally less
fecund th a n estuarine (m igrant) sm elt populations, a disp ro p o r­
tional con trib u tio n from estuarine populations could even be
expected [22].
T h e fact th a t smelt older th a n two years a re rarely observed in
Lake Ijsselm eer can be explained if they pass the sluices a n d stay
in the W ad d e n Sea w ithout re tu rn in g to Lake Ijsselm eer. A nnually
(with a p eak in O ctober) large n u m b ers 0 gro u p sm elt are passing
the sluices to the W ad d e n Sea (195 tonnes, 65.000.000 individuals,
Discussion
C onnectivity and Dispersal between Freshwater and
Marine Environment
T h e 88Sr concentrations in otoliths o f individuals collected in
the freshw ater lakes w ere notas high as reco rd ed in the otolith
edges o f specim ens caught in the W ad d e n Sea. T herefore, based
o n o u r study we argue th a t th ere is no evidence o f a substantial
con trib u tio n from d iadrom ous individuals to the landlocked smelt
p opulation. H ow ever, the possibility for a con trib u tio n can n o t be
com pletely excluded. O bservations n e a r the discharge sluices,have
show n th a t W ad d e n Sea sm elt occasionally enters Lake Ijsselm eer
(K ruitw agen unpubl. data). S paw ning generally takes place w ithin
a few weeks [18], D u e to ship tim e lim itations the specim ens w ere
collected in a short p erio d in M a rc h only, b u t covered all know n
spaw ning locations in L ake IJsselm eer/M ark erm eer. T h e sam ­
pling g ear used is p ro b ab ly suboptim al for catching larg er size
classes o f smelt if they h a d occurred. L arge sm elt w ere caught in
fyke nets o r gillnets in Lake IJsse lm e er/M a rk erm ee r in the late
20th century. C urrently, these size classes are never e n co u n tere d
PLOS ONE I www.plosone.org
6
July 2013 I Volume 8 | Issue 7 | e69796
Connectivity in Migrating and Landlocked Smelt
Table 1. Estimates, estim ated degrees o f freedom and significance fo r parameters testing area effects fo r log88Sr concentrations at
the core and the edge o f the oto liths in mixed models w ith individual fish nested w ith in sample as random effects.
response v a ria b le
88Sr at edge
88Sr at core
pa ram e te r
estim ate
s tan dard error
df
t-v alu e
P value
intercept
3.1207
0.03942
6142
79.164
0.0000
area: Markermeer
-0.1715
0.05524
9
-3 .1 0 5
0.0126
area: W adden Sea
0.1866
0.05688
9
3.281
0.0095
intercept
3.0323
0.04071
6172
74.477
0.0000
area: Markermeer
-0.0831
0.05672
9
-1 .4 6 5
0.1767
area: W adden Sea
0.0171
0.05995
9
0.285
0.7815
doi:10.1371 /journal.pone.0069796.t001
w hich roughly equals one th ird o f the total Ijsselm eer p opulation
(Drost & de W itte unpubl.).
88S r levels for the W ad d e n Sea fish w ere highly variable. T h e
W ad d en Sea is a very dynam ic habitat w ith strong variations in
salinity (24—32 P S U [23]) caused b y the tide a n d the outflow of
freshw ater through the sluices in the Afsluitdijk. A lthough 88S r values
from the W ad d en Sea are not available, they are likely to v ary greatly
a n d given the strong positive correlation betw een Sr a n d salinity, to
follow similar seasonal a n d daily fluctuations as salinity (h ttp ://
w w w o ld .n io z .n l/n io z _ n l/c c b a 2 4 6 4 b a 7 9 8 5 d le b l9 0 6 b 9 5 1 b lc 7 f6 .
php). T h is h igh variab ility in salinity is likely c ausing th e strong
fluctuations in
Sr profiles from W a d d e n Sea individuals a n d a
sm all o verlap in 88S r values b etw ee n W a d d e n Sea a n d Lake
Ijsselm eer m ea su re d in the o u te r 50 o f the otoliths. B ecause
W a d d e n Sea sam ples consisted o f specim ens th a t m ust h ave b e en
2 - 4 years old [18,22], we e x p ected to find dips in th e S r profile
(other th a n the core a re a th a t was p ro b a b ly dep o sited in th eir
early life in fresh w ater), p o ten tially in d ic a tin g a spaw ning event in
freshw ater. T h e absence o f such dips could m e a n th a t e ith e r these
individuals did n o t go to a freshw ater a re a for spaw ning, did not
spaw n yet o r d id n o t grow (and feed) w hile b e in g in th e ir spaw ning
areas. It is know n th a t fem ales leave the spaw ning areas earlier
th a n m ales, b u t it is n o t know n w h e th er tim e spent in rivers differs
b e tw ee n th e sexes [24], In all spaw ning does n o t take m o re th a n
1 -4 weeks [21]. T h e very high c o n ce n tra tio n s o f > 3 0 0 0 p p m
(Fig. 3) m ay in d icate the influence o f N o rth S ea w a ter inflow o f
h ig h er salinity.
b elong to a very local p opulation. Sr profiles w ere very sim ilar
w ithin a n d betw een hauls. Sim ilar profiles w ithin individuals o f
one haul b u t different from a n o th e r haul can b e expected if fish
stay to g eth er in one school for m ost o f th eir life a n d e n co u n ter
sim ilar circum stances as they m ove aro u n d . A lternatively the
sim ilar p attern s are caused by the
Sr signature o f the specific
region, w hich w ould indicate th a t sm elt do not m ove a ro u n d a lot,
a finding w hich shows agreem ent w ith a closely related species the
rainbow sm elt Osmerus mordax [25,26,27]
otolith core
4000
a
E
2000
a.
ce
1000
w
CO
Wadden Sea
Ijsselmeer
Markermeer
Exchange between Freshwater Areas
T h e two individuals w ith a dip in th eir profile late r in life (Fig. 4,
20 a n d 21) suggest th a t they m oved tow ards w ater w ith a low er
88Sr profile th a n Lake Ijsselm eer. T h e 88Sr level is sim ilar to th at
found in sm elt in Lake M ark erm eer. A n o th e r possibility is th at
they sw ani up the river IJssel, a n im p o rta n t spaw ning location in
form er days w h en the lakes w ere not closed off from the sea yet
[19]. T h e re are records o f sm elt (ca 20 cm) sw arm ing a t the R iver
IJssel m o u th a n d m igrating upstream for several tens o f kilom etres
d atin g back to 1840s [19], A dedicated search for smelt d u rin g the
spaw ning p erio d in the R iver IJssel m o u th , a n d analyses o f their
otoliths 88S r profiles could clarify if this scenario cu rren tly still
happens. T hese two individuals a n d the Lake Ijsselm eer individ­
uals 38 a n d 4 0 -5 5 th a t show core values closer to the range o f
M a rk erm e er individuals, could be a n in dication o f exchange
betw een Lake Ijsselm eer a n d Lake M ark erm eer. T h e only group
th a t has ra th e r low 88Sr values (around 1000 ppm ) are the ones
caught in the east o f Lake Ijsselm eer (3ld sam ple in Fig. 4); they
c an n o t be distinguished from M a rk erm e er individuals. T h e flat
88Sr profiles found in M a rk erm e er p ro b ab ly indicates these fish
PLOS ONE I www.plosone.org
otolith edge
ra
cu
E
o
IO
4000
3000
E
2000
Q.
CL
CC
Ul
1000
Wadden Sea
Ijsselmeer
Markermeer
Figure 6. M ean Sr (in ppm ) concentrations in otoliths o f the
inner 50 m easurem ents (core) and th e o u te r 50 m easurem ents
o f sm elt (edge) caught in th e W adden Sea, Lake Ijsselm eer and
Lake M arkerm eer.
do i:1 0.13 7 1 /jo u rn a l.p o n e .0 0 6 9 7 9 6 .g 0 0 6
7
July 2013 I Volume 8 | Issue 7 | e69796
Connectivity in Migrating and Landlocked Smelt
Im plications for M anagem ent o f Landlocked Population
O u r results show th a t u n d e r the c u rre n t m an a g em e n t th ere is
no indication o f a strong con trib u tio n o f the W ad d e n Sea sm elt to
the p opulations o f b o th lakes. C u rre n tly the possibilities for smelt
to re -e n ter L ake Ijsselm eer th ro u g h the sluices at either
K o rn w erd e rza n d or D en O e v er are lim ited to periods w hen
sluices are opened. Passage o f the discharge sluices tow ards the
W ad d e n Sea is only possible shortly after op en in g a n d tow ards
Lake Ijsselm eer shortly before closure. D iadrom ous sm elt tends to
aggregate in front o f the sluices in M a rch , w here in 2009 ca
50 tonnes o f a d u lt fish w ere caught in the com m ercial smelt
fisheries (Drost & de W itte unpubl.). G iven the lim ited en tran ce
possibilities, m an a g em e n t to en h an ce the con trib u tio n o f d iad ro ­
m ous smelt to the Lake Ijsselm eer pop u latio n should con cen trate
o n im proving the o p p o rtu n ity to en ter by adjusting the discharge
regim e o r increasing the n u m b e r o f e n tra n ce points in the
spaw ning period.
Acknowledgments
We want to thank the crew of the Tjeûkemar (property of the Netherlands
Institute of Ecology, N IO O , Wageningen, The Netherlands) that helped in
collecting the fish samples. H. van der Veer and H. Witte from the Royal
Netherlands Institute for Sea Research kindly provided Wadden Sea
specimens. N. van der Meeren (IMARES) prepared the otoliths.
Author Contributions
Conceived and designed the experiments: IT M K HW WB MDG.
Performed the experiments: JN M K. Analyzed the data: IT JN .
Contributed reagents/materials/analysis tools: M K JN WB. Wrote the
paper: IT.
References
1. Lassalle G , C ro u ze t P, R o ch a rd E (2009) M odelling th e cu rren t distribution o f
E u ro p ean diadrom ous fishes: an ap p ro ach integrating regional anthropogenic
pressures. Freshw ater Biology 54: 587—606.
2. N ellbring S (1989) T h e ecology o f smelts (G enus O sm erus): a literatu re review.
N ordic J Freshw Res 65: 116-145.
3. K ottelat M , F reyhof J (2007) H an d b o o k o f E u ro p ean freshw ater fishes. C om oi,
Switzerland: Publications K ottelat.
4. W illem sen J (1977) P opulation dynam ics o f percids in Lake IJssel a n d some
smaller lakes in th e N etherlands. J Fish R es B d C a n 34: 1710-1719.
5. D e Leeuw J J , D ekker W , Buijse A D (2008) A im ing a t a m oving target, a slow
h a n d fails! 75 yeasr o f fisheries m anagem ent in lake Ijsselm eer, th e N etherlands
J o u rn a l o f Sea R esearch 60: 21-31.
6. v an R oo m en M , H o rn m an M , Flink S, Langendoen T , van W in d en E, e t al.
(2012) Flyw ay-trends for w aterbird species im p o rtan t in Lakes Ijsselm eer an d
M arkerm eer. S O V O N . 2012-22 2012-22.
7. M ous I J (2004) Interactions betw een fisheries a n d birds in Ijsselm eer, T h e
N etherlands: W ageningen University, T h e N etherlands.
8. B o rch e rd in g J, P ickhardt G , W in ter H V , Becker J S (2008) M igratory history o f
N o rth Sea H o u tin g Coregonus oxyrinchus (L.) caught in Lake Ijsselm eer (The
N etherlands) inferred from scale transects o f 88S r:44C a ratios. A quatic Sciences
69: 4 7 -56.
9. G effen AJ, N ash R D M , D ickey-Collas M (2011) C haracterizatio n o f h erring
populations w est o f th e British Isles: a n investigation o f m ixing based o n otolith
m icrochem istry. IC E S Jo u rn a l o f M arine Science: 1447—1458.
10. M oráis P , D ias E , B a b a lu k J , A n tunes G (2011) T h e m igration pattern s o f the
E u ro p ean flounder Platichthys flesus (Linnaeus, 1758) (Pleuronectidae, Pisces) a t
th e southern lim it o f its distribution range: Ecological im plications a n d fishery
m anagem ent. J o u rn a l o f Sea R esearch 65: 235—246.
11. A rai T , C hino N (2012) D iverse m igration strategy betw een freshw ater an d
seaw ater habitats in th e freshw ater eel genus A nguilla. J o u rn a l o f Fish Biology
81: 4 4 2 ^ 5 5 .
12. W alther BD , L im burg K E (2012) T h e use o f otolith chem istry to characterize
diadrom ous m igrations. J o u rn a l o f Fish Biology 81: 796—825.
13. Secor D H , H endersonarzapalo A , Piccoli P M (1995) C a n otolith m icro chem istry
ch a rt patterns o f m igration a n d h a b ita t utilization in anadrom ous fishes. J o u rn a l
o f E xperim ental M arine Biology a n d Ecology 192: 15—33.
14. M ous I J , V an D ensen W L T , M achiels M A M (2004) V ertical distribution
patterns o f zooplanktivorous fish in a shallow, eu trophic lake, m ediated by w ater
transparency. Ecology o f Freshw ater Fish 13: 61—69.
15. v a n D en sen W L T , G azem ier W G , D ekker W , O u d e la a r H G J (1990)
M anagem ent o f th e fish stock in lake IJssel, th e N etherlands. In: S teinm etz B,
PLOS ONE I www.plosone.org
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
H ughes R H , editors. Proceedings o f a sym posium organized b y th e E u ro p ean
In la n d Fisheries Advisory Com m ission. G oeteborg, Sweden: Pudoc. pp. 313—
327.
Buijse A D , v an E erd en M R , D ekker W , v an D ensen W L T (1993) Elem ents o f a
trophic m odel for Ijsselm eer (the N etherlands), a shallow eutro p h ic lake. In:
C hristensen V , Pauly D , editors. Proceedings o f th e T ro p h ic m odels o f aq u atic
ecosystems, IC L A R M C onference Proc. pp. 90—94.
Z u u r A F, Ieno E N , W alker N J, Saveliev AA, Sm ith G M (2009) M ixed effects
m odels a n d extensions in ecology w ith R . N ew York: Springer.
D o h erty D , M c C h a rty T K (2004) T h e ecology a n d co n serv atio n o f E u ro p e an
S m elt (Osmerus eperlanus) fro m W a terfo rd estu ary , in so u th ea stern Irelan d .
Biology & en v iro n m en t, P roceedings o f th e R o y al Irish A cad em y 104B: 125—
130.
v an d en E n d e W P (1847) V erslagen v an de V ereeniging to t B evordering der
Inlandsche Ichthyologie. Eerste deel. A n. Nijhoff, A rnhem , th e N etherlands.
Berkeley S, H ix o n M , Larson R , Love M (2004) Fisheries sustainability via
p ro tectio n o f age structure a n d spatial distribution o f fish populations. Fisheries
29: 2 3-32.
H u tch in so n P , Mills D H (1987) C haracteristics o f spaw ning-run sm elt Osmerus
eperlanus (L.), fro m a Scottish river, w ith recom m endations for their conservation
a n d m anagem ent. A quaculture a n d fisheries m an ag em en t 18: 249—258.
B elyania T J (1969) Synopsis o f biological d ata o n smelt, Osmerus eperlanus L.
R om e: F ood a n d A griculture O rg an isatio n o f th e U n ited N ation.
v an A ken H M (2008) V ariability o f th e salinity in th e w estern W ad d en Sea o n
tid al to centennial tim e scales. J o u rn a l o f Sea R esearch 59: 121—132.
Q uigley D , Igoe F, O ’C o n n o r W (2004) T h e E u ro p ean Sm elt Osmerus eperlanus L.
in Ireland: G en eral biology, ecology, distribution a n d status w ith conservation
recom m endations Biology & E nvironm ent: Proceedings o f th e R o y al Irish
A cadem y 104: 5 7-66.
B radbury IR , C am p an a SE, B entzen P (2008) O to lith elem ental com position
a n d ad u lt tagging reveal spaw ning site fidelity a n d estuarine dependency in
rain b o w smelt. M arin e Ecology-Progress Series 368: 255—268.
B rad b u ry IR , G oulson M W , C a m p a n a SE, Baggs E, B entzen P (2009)
Postglacial recolonization a n d th e loss o f an ad ro m y in rain b o w sm elt from
coastal N ew foundland. In: H a ro A , Sm ith K L , R ulifson R A , M offitt C M ,
K lau d a R J e t al., editors. Challenges for diadrom ous fishes in a dynam ic global
environm ent, pp. 79—96.
B radbury IR , DiBacco C , T h o rro ld SR , Snelgrove P V R , C a m p a n a SE (2011)
Resolving n atal tags using otolith geochem istry in a n estuarine fish, rainbow
sm elt Osmerus mordax. M arin e Ecology-Progress Series 433: 195-U576.
July 2013 I Volume 8 | Issue 7 | e69796