Received: 26 May 2017
|
Revised: 30 August 2017
|
Accepted: 5 September 2017
DOI: 10.1111/mec.14352
ORIGINAL ARTICLE
Seascape genomics reveals fine-scale patterns of dispersal for a reef fish along the ecologically divergent coast of Northwestern Australia Joseph D. DiBattista1,2
| Michael J. Travers2,3
D. Evans5,6 | Stephen J. Newman3
| Glenn I. Moore2,4 | Richard
| Ming Feng2,7 | Samuel D. Moyle3 | Rebecca
J. Gorton8 | Thor Saunders9 | Oliver Berry2,7 1
Department of Environment and Agriculture, Curtin University, Perth, WA, Australia
2
Western Australian Marine Science Institution, Crawley, WA, Australia
3
Western Australia Fisheries and Marine Research Laboratories, Department of Primary Industries and Regional Development, Government of Western Australia, North Beach, WA, Australia 4
Department of Aquatic Zoology, Western Australian Museum, Welshpool, WA, Australia
5
Department of Biodiversity, Conservation and Attractions, Perth, WA, Australia
6
School of Biological Sciences and Oceans Institute, University of Western Australia, Crawley, WA, Australia
7
CSIRO National Collections and Marine Infrastructure, Level 4 – Indian Ocean Marine Research Centre, The University of Western Australia, Crawley, WA, Australia 8
CSIRO Oceans & Atmosphere, Hobart, TAS, Australia
9
Northern Territory Department of Primary Industry and Fisheries, Darwin, NT, Australia
Correspondence Joseph DiBattista, Department of Environment and Agriculture, Curtin University, Perth, WA, Australia. Email:
[email protected] Funding information Funding for Kimberley sampling was administered through the Western Australian Marine Science Institution (WAMSI) Kimberley Marine Research Program and the Department of Primary Industries and Regional Development (Fisheries; Government of Western Australia) Lalang-garram Camden Sound Marine Park Project. Pilbara and Gascoyne sample collections were funded by the Chevronoperated Wheatstone Project’s State Environmental Offsets Program and the Woodside-operated Pluto Project for the State Environmental Offsets Program administered by the Department of Biodiversity, Conservation and Attractions. The Wheatstone Project is a joint venture between Australian subsidiaries of Chevron, Kuwait Foreign Petroleum Exploration Company (KUFPEC), Woodside Petroleum Limited and Kyushu Electric Power Company, together with PE Wheatstone Pty Ltd (part owned by TEPCO).
Molecular Ecology. 2017;1–18.
Abstract Understanding the drivers of dispersal among populations is a central topic in marine ecology and fundamental for spatially explicit management of marine resources. The extensive coast of Northwestern Australia provides an emerging frontier for implementing new genomic tools to comparatively identify patterns of dispersal across diverse and extreme environmental conditions. Here, we focused on the stripey snapper (Lutjanus carponotatus), which is important to recreational, charterbased and customary fishers throughout the Indo-West Pacific. We collected 1,016 L. carponotatus samples at 51 locations in the coastal waters of Northwestern Australia ranging from the Northern Territory to Shark Bay and adopted a genotype-bysequencing approach to test whether realized connectivity (via larval dispersal) was related to extreme gradients in coastal hydrodynamics. Hydrodynamic simulations using CONNIE and a more detailed treatment in the Kimberley Bioregion provided null models for comparison. Based on 4,402 polymorphic single nucleotide polymorphism loci shared across all individuals, we demonstrated significant genetic subdivision between the Shark Bay Bioregion in the south and all locations within the remaining, more northern bioregions. More importantly, we identified a zone of admixture spanning a distance of 180 km at the border of the Kimberley and Canning bioregions, including the Buccaneer Archipelago and adjacent waters, which collectively experiences the largest tropical tidal range and some of the fastest tidal
wileyonlinelibrary.com/journal/mec
© 2017 John Wiley & Sons Ltd
|
1
2
|
DIBATTISTA
ET AL.
currents in the world. Further testing of the generality of this admixture zone in other shallow water species across broader geographic ranges will be critical for our understanding of the population dynamics and genetic structure of marine taxa in our tropical oceans. KEYWORDS
biodiversity, coral reef, DArTseq, environmental gradients, Kimberley, marine
1 | INTRODUCTION
between different environments (Nosil, Funk, & Ortiz-Barrientos, 2009; Rellstab, Gugerli, Eckert, Hancock, & Holderegger, 2015).
Coastal ecosystems contain some of the most productive, diverse
Access to remotely sensed environmental data such as sea surface
and valuable environments on the planet but are also exposed to
temperature and nutrient proxies (e.g., chlorophyll-a) from satellites
high anthropogenic stressors such as fishing, tourism, mineral and
also provides opportunities to identify the geographic and environ-
petrochemical industries, and coastal development (Moore et al.,
mental determinants of genetic structure (Balkenhol, Waits, & Dez-
2016). A robust understanding of the ecological drivers that under-
zani, 2009; Wang & Bradburd, 2014).
pin population regulation in marine systems is pivotal for the effec-
The globally unique Northwestern Australian (NWA) marine envi-
tive management of coastal aquatic resources. Larval dispersal is one
ronment and the Kimberley Bioregion in particular (Chin et al., 2008;
such driver with particular relevance to coastal management because
Halpern et al., 2015; Wilkinson, 2008) are likely to be subject to pro-
it has the potential to link the demographic fates of distant popula-
posed increases in industrial developments, fishing and tourist activi-
tions, as well as provide the genetic diversity necessary for evolu-
ties (Wood & Mills, 2008). In recognition of the potential for
tionary adaptation.
increased anthropogenic stressors in NWA, there has recently been
Many coastal marine species have restricted home ranges as
a growth in management activity, including the formation of major
adults (Cowen & Sponaugle, 2009), and so it is their dispersive
marine reserves (e.g., Department of Parks and Wildlife 2016). In
planktonic larval stages that promote demographic and genetic con-
some cases, management strategies have been implemented in the
nectivity between suitable habitats. Marine larvae exhibit diverse life
absence of basic knowledge of ecological processes operating within
histories, but typically they are small and have at least some capacity
these ecosystems (Moore et al., 2016). Most marine ecological
to determine their precise destination (Leis, 2015). The magnitude of
enquiry in NWA has focused on characterizing biodiversity, commu-
larval dispersal is determined by a combination of the physical envi-
nity assemblages and species distributions (Harvey et al., 2012;
ronment such as ocean currents, habitat distributions and coastal
Hutchins, 2001; McKinnon, Duggan, Holliday, & Brinkman, 2015;
topography, as well as larval attributes such as swimming speed, sen-
McLean et al., 2016; Moore & Morrison, 2009; Moore, Morrison,
sory capabilities and pelagic larval duration (PLD; Leis, 2015; Treml,
Hutchins, Allen, & Sampey, 2014; Poore & O’Hara, 2007; Travers,
Ford, Black, & Swearer, 2015). Some species will therefore operate
Newman, & Potter, 2006; Travers, Potter, Clarke, & Newman, 2012;
as closed demographic units on small spatial scales (within a few
Travers, Potter, Clarke, Newman, & Hutchins, 2010; Wilson, 2013).
kilometres), whereas others may remain connected over hundreds of
A limited number of studies have investigated connectivity among
kilometres (Almany et al., 2017; Berumen et al., 2012; Saenz-Agu-
the distinct but ecologically and economically important Australian
delo, Jones, Thorrold, & Planes, 2011). Coastal ecosystems can also
coastal ecosystems (Broderick et al., 2011; Horne, Momigliano, van
be topographically complex, which makes correct predictions of con-
Herwerden, & Newman, 2013; Horne, Momigliano, Welch, Newman,
nectivity among the network of populations elusive given the envi-
& van Herwerden, 2011, 2012; Johnson, Hebbert, & Moran, 1993;
ronmental variability among sites (Burgess et al., 2014). Even though
Johnson & Joll, 1993; Ovenden, Lloyd, Newman, Keenan, & Slater,
connectivity may dictate the resilience of entire ecosystems (Selkoe
2002; Taillebois et al., 2017; Veilleux, van Herwerden, Evans, Tra-
et al., 2016), it is often context dependent, variable among resident
vers, & Newman, 2011), and none have focused on inshore fishes
taxa (Drew & Barber, 2012) and difficult to directly measure (Krueck
with a comprehensive sample coverage across substantial environ-
et al., 2016). In the light of these challenges, we apply an indirect
mental gradients. This emerging frontier in NWA therefore provides
genomic approach (i.e., proxy) to measure connectivity (Riginos,
the ideal model system to increase our understanding of larval dis-
Crandall, Liggins, Bongaerts, & Treml, 2016).
persal and connectivity in a management context.
Single nucleotide polymorphism (SNP) discovery and genotyping
Herein, we evaluate genetic connectivity of the Spanish flag snap-
(Andrews, Good, Miller, Luikart, & Hohenlohe, 2016) via next-gen-
per or stripey snapper, Lutjanus carponotatus (Richardson, 1842),
eration sequencing (NGS) provide a means to quantify connectivity
across the entire NWA region using a genotyping-by-sequencing
within coastal ecosystems with high resolution (Riginos et al., 2016).
approach. Lutjanus carponotatus is abundant on tropical inshore reefs
The high genomic coverage provided by thousands of SNPs means
across the coast of Australia north of latitude ~27°S, but also occurs
that it is possible to detect signatures of differential natural selection
in coastal waters from India through to the Indo-West Pacific
DIBATTISTA
|
ET AL.
(Anderson & Allen, 2001; Newman & Williams, 1996). This species is of significant recreational fishing importance across NWA, comprising
3
2.2 | Sample collection
14% of the boat-based recreational catch in the North Coast Manage-
Tissue samples of L. carponotatus (Figure 1, Tables 1, and S1) were
ment Bioregion during 2013 and 2014 (Ryan et al., 2015). We also
collected from 51 coastal sites across NWA from the Anson Beagle
focus on L. carponotatus as a “model species” because its larval settle-
in the Northern Territory (hereafter referred to as NT) through the
ment behaviour and ecology are similar to other predatory species of
Kimberley, Canning, Pilbara, Ningaloo and Shark Bay bioregions of
commercial importance in NWA (e.g., Lethrinus spp., Fisher, Leis,
WA. In total, 1,016 samples were collected across 13° of latitude
Clark, & Wilson, 2005). A recent genetic investigation of L. carponota-
and 17° of longitude of the tropical Australian coastline (also see
tus on the Great Barrier Reef (GBR) in eastern Australia using mito-
Tables 1 and S1) and immediately preserved in 95% ethanol. Most
chondrial DNA markers found complete panmixia within and between
sampling was undertaken in 2014 and 2015; however, muscle tissue
inshore islands at a scale of 400 km (Evans, Van Herwerden, Russ, &
collected in 2002 and frozen at !80°C was obtained from four sites
Frisch, 2010). A companion study based on the same species and
(Cape Bossut, Cape Keraudren, Cape Preston, Locker Point).
molecular markers in Western Australia (WA), identified a comparable scenario of panmixia in this region at a scale of 800 km, although it was strongly differentiated from the GBR populations (Veilleux et al.,
2.3 | Reduced representation SNP genotyping
2011). We have significantly improved on and extended these studies
DNA was extracted from tissue samples using 96-well plates and the
by performing a genome-wide survey of L. carponotatus at 51 sites
salt extraction method described by Cawthorn, Steinman, and Wit-
along the extensive ~2,500 km coast of NWA to compare broadscale
thuhn (2011), followed by purification with Zymo ZR-96 DNA Clean
patterns of genomic divergence among bioregions that differ in reef
and Concentrator kits (Zymo Research, California, USA). Downstream
composition as well as oceanographic, tidal and current regimes. In
SNP genotyping was undertaken using a modified DArTseqTM proto-
our seascape genomic analysis, we include Euclidean distance, ocea-
col (Grewe et al., 2015), which is a proprietary method for reduced
nic-derived distance, biophysical modelling, environmental variables
representation genomic library preparation and NGS (Cruz, Kilian, &
and management boundaries to understand drivers of resistance to
Dierig, 2013; Kilian et al., 2012), where select loci at high coverage
gene flow over an extensive and globally significant coastline.
vs. the entire genome at low coverage are sequenced for individual samples. In our case, genomic DNA was digested with two restric-
2 | MATERIALS AND METHODS
tion enzymes (PstI-SphI and PstI-NspI) instead of one to generate
2.1 | Study region
step at 94°C for 1 min followed by 30 cycles of 94°C for 20 s, 58°C
more SNP loci. PCR conditions consisted of an initial denaturation for 30 s and 72°C for 45 s, with a final extension step at 72°C for
The remote coast of NWA between ~13°S and 27°S is largely
7 min. After PCR, equimolar amplification products from each sam-
unpopulated and geologically ancient, with low productivity, and is
ple were pooled and applied to a cBot bridge PCR system followed
characterized by limited accessibility and extreme marine conditions
by sequencing on an Illumina Hiseq2500. The sequencing (single
(Molony, Newman, Joll, Lenanton, & Wise, 2011; Wilson, 2013). This
read) was run for 77 cycles.
coastline contains nearly 3,000 islands, two World Heritage sites
Read assembly, quality control (QC) and SNP calling were under-
(Ningaloo Reef and Shark Bay), four National Heritage areas of out-
taken using DArT PLD’s software DArTsoft14, a program that pro-
standing natural and indigenous significance, several marine reserves,
duces scoring consistency derived from technical sample replicates
and hosts a diverse assemblage of fishes and corals (McLean et al.,
(i.e., samples processed twice, from DNA library preparation to SNP
2016; Moore et al., 2014; Richards, Garcia, Wallace, Rosser, & Muir,
calling). Testing for Mendelian distribution of alleles in these popula-
2015; Richards & O’Leary, 2015; Travers et al., 2006, 2010, 2012).
tions facilitated selection of technical parameters discriminating true
Environmental conditions vary significantly across this coast but the
allelic variants from paralogous sequences. A total of 17,007 SNP
most extreme are the tides that range from ~12 m in the Kimberley
loci were identified during this process.
Bioregion, the highest tropical tidal range in the world, down to 1 m at Shark Bay, with maximum monthly water temperatures ranging from >32°C in the Kimberley to 24°C in Shark Bay (Lowe, Pivan, Falter, Symonds, & Gruber, 2016; Wilson, 2013). Existing biogeographic data, along with information on environ-
2.4 | Quality filtering Following SNP genotyping, additional QC steps were performed to the 17,007 loci identified prior to genetic analyses: (i) rare alleles (fre-
mental attributes including ocean currents, geology and marine sedi-
quency <0.05) and highly variable loci (heterozygosity >0.75) were
ments has formed the foundation of bioregional classifications such
removed, (ii) loci with coverage less than 209 and greater than 2009
as the provinces and ecoregions of Spalding et al. (2007), as well as
were removed and (iii) individuals with more than 1% missing data
the provincial and mesoscale bioregions of the Integrated Marine
were removed as suggested by thresholds within the R package dartR
and Coastal Regionalisation of Australia (IMCRA). Indeed, the NWA
(https://github.com/green-striped-gecko/dartR). Following these fil-
coast spans seven IMCRA mesoscale marine bioregions (Table 1;
tering steps, we were left with 5,094 loci. We tested for departures
sensu Commonwealth of Australia, 2006).
from Hardy–Weinberg equilibrium (HWE) and linkage disequilibrium
4
|
DIBATTISTA
T A B L E 1 Site, IMCRA classification, sample size (N) and molecular metrics (Na = number of alleles; Ho = observed heterozygosity; He = expected heterozygosity; FIS = inbreeding coefficient) for Lutjanus carponotatus based on 4,402 SNP loci. IMCRA mesoscale bioregions are derived from the Commonwealth of Australia (2006) Site
IMCRA bioregion
N
Na
Ho
He
FIS
TABLE 1
ET AL.
(Continued)
Site
IMCRA bioregion
N
Na
Ho
He
FIS
Depuch Island
Pilbara (Nearshore)
27
1.940
0.218
0.228
0.036
West Moore
Pilbara (Nearshore)
25
1.937
0.218
0.230
0.044
Gidley Island
Pilbara (Nearshore)
27
1.937
0.199
0.223
0.084
Rosemary Island
Pilbara (Nearshore)
27
1.937
0.206
0.225
0.068
Cape Preston
Pilbara (Nearshore)
30
1.950
0.255
0.239
!0.043
Passage Island
Pilbara (Nearshore)
26
1.930
0.197
0.222
0.091
Pilbara (Offshore)
26
1.935
0.198
0.221
0.082
Bass Reef
Anson Beagle
13
1.833
0.208
0.219
0.034
Point Blaze
Anson Beagle
10
1.784
0.205
0.218
0.040
Sail City
Anson Beagle
2
1.385
0.195
0.158
!0.246
Long Reef
Kimberley
28
1.939
0.224
0.229
0.023
Pascal Island
Kimberley
15
1.868
0.217
0.224
0.024
Jamieson Reef
Kimberley
12
1.829
0.205
0.218
0.041
Cape Voltaire
Kimberley
49
1.970
0.214
0.229
0.058
Montebello North
Cleghorn Island
Kimberley
9
1.763
0.243
0.225
!0.071
Montebello South
Pilbara (Offshore)
24
1.922
0.210
0.226
0.053
Bigge Island
Kimberley
43
1.968
0.222
0.231
0.037
1.930
0.199
0.222
0.086
Kimberley
30
1.938
0.200
0.222
0.081
Pilbara (Nearshore)
25
Blue Holes
Thevenard Islands
Adieu Point
Kimberley
22
1.912
0.211
0.224
0.046
Paroo Shoal
Pilbara (Nearshore)
26
1.931
0.213
0.226
0.049
Wailgwin Island West
Kimberley
19
1.892
0.207
0.220
0.045
Locker Point
Pilbara (Nearshore)
24
1.926
0.256
0.239
!0.050
Bay of Rest
Pilbara (Nearshore)
28
1.940
0.203
0.222
0.070
Roberts Island
Pilbara (Nearshore)
27
1.938
0.201
0.224
0.090
Hall Point
Kimberley
22
1.921
0.207
0.224
0.059
Ngalanguru Island
Kimberley
17
1.877
0.225
0.227
0.004
Montgomery Reef
Kimberley
4
1.580
0.216
0.198
!0.100
Dugong Bay
Kimberley
21
1.905
0.207
0.223
0.056
Bathurst Island
Kimberley
3
1.495
0.221
0.183
Shark Bay
25
1.901
0.193
0.218
0.088
Kimberley
8
1.845
0.348
0.262
!0.203
Bernier Island
Irvine Island
!0.256
Dorre Island
Shark Bay
27
1.902
0.194
0.218
0.089
Tantabiddi
Ningaloo
24
1.917
0.200
0.221
0.075
Milyering
Ningaloo
25
1.927
0.206
0.223
0.057
Fraser Island
Kimberley
7
1.711
0.207
0.210
Longitude Island
Kimberley
5
1.638
0.204
0.202
Asshlyn Island
Kimberley
3
1.517
0.215
0.190
Pope Island
Kimberley
4
1.579
0.209
0.195
Gregory Island
Kimberley
2
1.421
0.227
0.173
Mermaid Island
Kimberley
5
1.633
0.237
0.210
Jorrol
Kimberley
14
1.863
0.209
0.221
0.035
Hal’s Pool
Kimberley
14
1.850
0.267
0.239
Tallon Island
Kimberley
28
1.938
0.240
0.234
!0.086
Jackson Island
Kimberley
7
1.717
0.224
0.215
Bowlun
Kimberley
3
1.503
0.224
0.187
Shenton Bluff
Kimberley
41
1.962
0.208
0.226
Ngamakoon
Canning
23
1.923
0.229
0.229
0.003
version 2.0.5.1 (Lischer & Excoffier, 2012). Downstream genetic anal-
Emeriau Point
Canning
30
1.946
0.241
0.236
yses were performed with all samples from all sites (full data set; 51
Cape Bossut
Canning
30
1.945
0.267
0.241
!0.009
sites) unless noted, in which case these were performed with only
Cape Keraudren
Canning
30
1.949
0.271
0.245
!0.071 !0.073
those populations with N > 20 individuals collected (reduced data set;
!0.006 !0.031 !0.144 !0.085 !0.315 !0.123
!0.013 !0.047 !0.197
0.068
(Continues)
IMCRA, Integrated Marine and Coastal Regionalisation of Australia; SNP, single nucleotide polymorphism.
(LD) expectations within each sample site and chose to exclude loci out of HWE in more than 10 populations and loci that exhibited LD in greater than five populations. Testing for HWE was implemented within the R packages dartR, SNPassoc (Gonz!alez et al., 2007) and pegas (Gonz!alez et al., 2007; Paradis, 2010). Testing for LD made use of the R packages doParallel (Calaway, Weston, & Tenenbaum, 2014), Adegenet (Jombart, 2008) and dartR. After these filtering steps, 4,402 loci remained. We filtered SNP loci using a number of different values for HWE, LD and QC ("15% of threshold), but it did not impact the overall outcome (data not shown). The resulting genind formatted file was converted to other program specific input files using PGDSPIDER
30 sites) to mitigate the effects of low sample size, and in some cases, all three NT sites were pooled to facilitate.
DIBATTISTA
|
ET AL.
(a)
5
Indonesian
Indone sia
So u th
Java
Curre n
t
C
er
ter Win
m
m
Su
Anson Beagle
mn
utu
a llow Ho
Ningaloo B Current
Pilbara
Le e
Shark Bay
uw
Gascoyne
t i n C urre n
(b) B
113.10 E
Rosemary Island
Montebello North Montebello South Thevenard Islands
Tantabiddi Milyering
23.20 S
A nt rre y Cu
Canning
WESTERN AUSTRALIA
118.90 E
(c)
Depuche Island West Moore Island Gidley Island Cape Preston 14.40 S Passage Island Paroo Shoal Locker Point Bay of Rest
Roberts Island
NORTHERN TERRITORY
Kimberley
23.20 S
123.20 S
Jamieson Reef Pascal Island Cape Voltaire
14.40 S
Long Reef
Bigge Island Adieu Point Blue Holes Wailgwin Island Hall Point Montgomery Reef
Buccaneer Archipelago
Cleghorn Island
Ngalanguru Island
Sunday Strait
Ngamakoon
Bernier Island
Emeriau Point
King Sound
16.80 S
Dorre Island
Dugong Bay
Dampier Peninsula
26.10 S 113.10 E
Marine Reserve Boundary 26.10 S 0 75 150km 118.90 E
123.20 S
16.80 S Marine Reserve Boundary
123.20 S
F I G U R E 1 Map of sampling sites (yellow dots) for Lutjanus carponotatus across (a) the entire sampling range in Northwestern Australia, (b) central western Australia (Pilbara to Gascoyne coasts) and (c) the Kimberley coast. The Holloway Current is the dominant current affecting coastal waters of the Kimberley, Canning and Pilbara bioregions, and the Leeuwin Current significantly impacts the Ningaloo and Shark Bay bioregions (adapted from Sprintall, Wijffels, Chereskin, & Bray, 2002; Domingues, Maltrud, Wijffels, Church, & Tomczak, 2007; D’Adamo, Fandry, Buchan, & Domingues, 2009; Schiller, 2011). Illustration of L. carponotatus © R.Swainston/www.anima.net.au
2.5 | Descriptive statistics, genetic subdivision and outlier detection
otherwise noted, given that they had a marginal effect on the out-
FST, FIS and genetic diversity metrics (percentage of polymorphic loci,
2.6 | Model-based clustering analysis
average number of alleles, observed and expected heterozygosity)
come of these tests.
were estimated using Genodive version 2.0 (Meirmans & Van Tien-
To explore genetic structure across sampling sites, a model-based
deren, 2004); the significance of pairwise FST values was tested by
clustering analysis was performed with STRUCTURE version 2.3.4
10,000 permutations via bootstrapping. To identify SNPs that may
(Pritchard, Stephens, & Donnelly, 2000) with a priori information on
be under divergent selection, we performed outlier scans between
the geographic origin of each sample. The analyses were run on the
all pairs of sites using Outflank version 0.1 (Whitlock & Lotterhos,
CSIRO Accelerator Cluster “Bragg” under the admixture model with
2015). The approach implemented in Outflank is based on an
correlated allele frequencies, a burn-in of 200,000 MCMC iterations,
improved method for deriving the null distribution of population dif-
followed by 500,000 iterations for each run (Falush, Stephens, &
ferentiation for neutral loci and results in fewer false positive than
Pritchard, 2003). The number of K (putative populations) ranged
other outlier tests (Lotterhos & Whitlock, 2015). We ran Outflank
from 1 to 8, and 20 replicate analyses were run for each value of K.
with 5% left and right trim for the null distribution of FST, minimum
The number of clusters was inferred by comparing the ln Pr (X|K)
heterozygosity for loci of 0.1% and a 5% false discovery rate (q-
among different values of K; the value that was highest or reached a
value). Sixty-six SNPs under putative directional selection were iden-
plateau was selected as the most parsimonious number of popula-
tified. These loci were included in all downstream analyses, unless
tions in our sample. The ad hoc statistic DK (Evanno, Regnaut, &
6
|
Goudet, 2005) was also considered. The outcome was no different when no a priori information on geographic origin of each sample
DIBATTISTA
ET AL.
2.9 | Determinants of genetic differentiation
was included, and so we only present results with a priori informa-
We used an information-theoretic approach (Anderson, 2008) to
tion. To objectively assess whether geographic origin correlates with
determine the geographic and oceanographic variables that best
the ancestry profiles from STRUCTURE, we estimated a correlation
explain the observed genetic subdivision (FST) in L. carponotatus
coefficient (R2) statistic using the program ObStruct version 1.0
throughout NWA. This method ranks alternative models per their
(Gayevskiy, Klaere, Knight, & Goddard, 2014) and a default of
relative empirical support (Correa & Hendry, 2012). Sample sites
10,000 permutations.
with N < 20 individuals were excluded from the analysis given uncertainty in FST estimates when based on low sample sizes (Will-
2.7 | Discriminant analysis of principle components
ing, Dreyer, & Van Oosterhout, 2012). Geographic variables were
We employed discriminant analysis of principle components (DAPC)
based on their AIC score, evidence ratio and posterior probability.
implemented in the R package Adegenet to further describe
Geographic factors included the Euclidean distance between sites
genetic groups present within our data. Initially, the k-means algo-
(geo) and the presence of three putative barriers to larval dispersal
rithm was employed to evaluate all potential clusters (K) in the
based on the regional designations described below. We also calcu-
data. For this analysis, we retained all principle components and
lated hydrodynamic distance between sites (hyd; see details below).
then evaluated the Bayesian information content for all values of
Although the northern and western coasts of Australia have been
K. A linear discriminant analysis was then conducted based on 338
classified and reclassified according to a number of marine biogeo-
retained principle components (N individuals divided by 3) identi-
graphic boundaries (Fox & Beckley, 2005; Spalding et al., 2007;
fied as optimal based on the optim.a.score command, and 50 dis-
Thackway, 1998), we follow the marine ecoregions of the world
criminant functions retained (N ! 1 populations) to describe the
(MEOW) of Spalding et al. (2007), as it utilizes the most recent
clusters evident in the data. For this analysis, we did not restrict
quantitative data on marine fish distributions in this area. The ecore-
the number of clusters to the number identified in the find.clusters
gional units we specifically test (and thus the barriers between them)
analysis. All analyses were run on all loci as well as only on the
are the Bonaparte Coast, Exmouth to Broome and Ningaloo Reef to
outlier loci.
Shark Bay. The NT ecoregion was not included in this analysis due
included in the model selection process, and each model was ranked
to a limited sample size. It should be noted that the marine ecoregional units differ from
2.8 | Spatial autocorrelation and isolation by distance
the management bioregions used for the purposes of fisheries man-
GenAlEx version 6.502 (Peakall & Smouse, 2006) was used to quan-
bioregion boundaries, we also specifically explored the following
tify spatial autocorrelation based on all sites with N > 6. Calculations
divisions: NT (all NT sites), Kimberley (Long Reef to Cape Bossut),
were made separately for the three distinct “clusters” identified in
Pilbara (Cape Keraudren to Paroo Shoal) and Gascoyne (Locker Point
model-based clustering (see above). Within the Kimberley cluster,
to Dorre Island), in addition to the MEOW ecoregional units as out-
distances were 0–256 km (N = 266); within the Pilbara cluster, dis-
lined above. Indeed, these management bioregions most closely align
tances were 0–426 km (N = 391); and within the “transition zone”
with the bioregional boundaries developed by IMCRA (Common-
between the Kimberley and Pilbara clusters, distances were
wealth of Australia, 2006) rather than MEOW (Spalding et al., 2007).
agement. To examine potential connectivity across management
0–148 km (N = 193). We conducted multiple distance-class spatial
The common feature of the various bioregional classifications
autocorrelations rather than conventional correlograms to accommo-
and other quantitative fish assemblage studies is the pronounced
date uneven sample sizes and distances typical of reef topography
faunal break in the Cape Leveque region at the northern tip of the
(Peakall, Ruibal, & Lindenmayer, 2003). Geographic distances
Dampier Peninsula (Figure 1; Spalding et al., 2007; Thackway, 1998;
between sites were calculated based on the shortest across-water
Travers et al., 2010; Wilson, 2013). Barriers were modelled as a fac-
distance with a minimum low tide water depth of 1 m. These esti-
tor with values of 0–3 (or 0–2), where sites compared on the same
mates were calculated with the Marmap R package (Pante & Simon-
side of the barrier received a score of 0 and sites compared on dif-
Bouhet, 2013) and based on the GEBCO 2014 30-s bathymetry
ferent sides of a barrier received a score of 1–3 (or 1–2) depending
available from the British Oceanographic Data Centre.
on the number of barriers in between.
We applied Mantel tests to evaluate the relationship between
Overall, 15 models were fitted using both linear models (lm) and
linearized FST (FST/(1 ! FST) and distance (isolation by distance, IBD).
linear mixed effects models (lmer). Linear mixed effects models
This analysis was based on 9,999 permutations of the data calcu-
included site ID as a random effect to compensate for the fact that
lated with the vegan R package (Oksanen et al., 2007). Mantel tests
pairwise FST values are not independent among sites. For each
were applied to the entire data set, as well as the Kimberley and Pil-
model, the sample size-corrected Akaike information criterion (AICc)
bara bioregions separately. A generalized linear model (GLM) imple-
was calculated. Models were then ranked based on increasing AICc
mented in R was used to test for regional isolation-by-distance
and further interpretation based on model probabilities (w) and evi-
effects in the Pilbara and Kimberley bioregions.
dence ratios (Anderson, 2008).
DIBATTISTA
|
ET AL.
7
surface current velocities (0–5 m) extracted from the model output
2.10 | Broadscale hydrodynamic model and connectivity calculations
and used for particle tracking. A total of 100 particles were seeded at
Null connectivity of L. carponotatus was modelled using CONNIE,
January), at 3-day intervals with a PLD of 40 days. This particle
the engine behind the CONNIE3 web interface (www.csiro.au/con
release period was chosen to capture all possible spawning periods
sampling sites during the austral spring–summer period (September–
nie). This tool combines currents from oceanographic models, parti-
for L. carponotatus (Kritzer, 2004; R. Evans et al., unpublished data). A
cle tracking techniques and simple behavioural models to estimate
fourth-order Runge-Kutta sub-time-stepping scheme was used to
connectivity statistics (Condie, Hepburn, & Mansbridge, 2012). Esti-
update the particle locations every hour. Using the random walk
mates of the oceanographic conditions along the WA coast were
effect of 1 m2/s, particles were tracked for 35 days based on the
based on an output of the 10 km resolution Bluelink Ocean ReANa-
!re ! & Leis, 2010). The potential PLD of this species (33–38 days; Que
lyis (BRAN2.1) model (Schiller et al., 2008). Although finer resolution
settlement of the particles was set to be within 500 m of a particular
models do exist, none include the entire geographic span of the 51
site. Connectivity among sampling sites was estimated as the average
L. carponotatus sampling sites. Particles were released from each of
number of particles released from site i that were tracked to be in site
the 51 sites, and the parameterization of the particles was selected
j, this ranged from 0 to 7.49 per release period, based on 48 simula-
to resemble the behaviour of L. carponotatus larvae; particles were
tion replicates. The oceanographic connectivity matrix was visualized
released in the final quarter (1 October to 31 December) of each
using the package qgraph (Epskamp, Cramer, Waldorp, Schmittmann,
year from 2003 to 2008. Over the course of each day, 100 particles
& Borsboom, 2012).
were released randomly within a 0.2 degree square around the lati-
For the purposes of comparison to genetic data, the output matrix
tude and longitude of each sampling site. Individual particles were
was converted into a distance matrix by dividing by the number of
then tracked as they were dispersed by the horizontal component of
released particles and subtracting from 1 and averaging the genetic
modelled ocean currents over a total dispersal time of 40 days. For
connectivity between pairwise sites. To increase the confidence in
the first 5 days, particles remained at a depth of 5 m. From day 5
the FST results for correlation purposes, we used only sites with N > 7
onwards, each particle underwent diel vertical migration, spending
samples (see Willing et al., 2012), resulting in 18 sites for the Mantel
the day at 15 m and the night at 5 m, where the length of day
test comparing pairwise FST to larval particle tracking results and the
depends on the time of the year.
oceanographic distance in the Kimberley Bioregion. Mantel tests with
The number of particles in a simulation released from a source
10,000 repetitions were run in the R package vegan.
site that arrive at a destination site here represents the potential connectivity between sites (Hock, Wolff, Condie, Anthony, & Mumby, 2014). The particle trajectories were used to calculate the potential connectivity between each pair of sites (i, j) and this was combined to create a combined connectivity matrix for the region.
3 | RESULTS 3.1 | Genetic diversity
The particles were only allowed to “settle” (and contribute to the
A summary of the principal statistics (number of individuals per
connectivity matrix) 33–40 days from their release. If a particle
site, percentage of polymorphic loci, average number of alleles,
moved within 0.1 degree of a potential settlement site, that particle
observed and expected heterozygosity and FIS) obtained for 1,016
was assumed to have settled and would not move onwards to any
individual samples from 51 locations in NWA is presented in
other potential settlement sites. The connectivity value between any
Tables 1 and S1. Observed heterozygosity was significantly higher
two sites (i, j) is the percentage of particles released from site i that
in the southwestern bioregions (combined Canning, Pilbara and
“settled” at site j. Self-recruitment (i, i) was also estimated.
Shark Bay) than the northeastern bioregions (NT and Kimberley; t
2.11 | Fine-scale biophysical dispersal model: Effects in the Kimberley Bioregion
correlated with the latitude of each site (r = .159), suggesting a
The biophysical dispersal model was based on the regional ocean
the northeastern bioregions they were mostly negative (t test:
modelling system hydrodynamic model with 2 km resolution (M.
t = !3.17 and p = .0026). FIS values were similarly weakly corre-
test: t = !4.19 and p < .001). Observed heterozygosity was weakly
Feng, unpublished data) to construct a pairwise matrix of larval con-
bioregional effect rather than a direct effect of latitude. FIS values were mostly positive in the southwestern bioregions, whereas in
lated with the latitude of each site (r = .137). We also identified 66
nectivity among sampling sites. We used only a subset of sites based
outlier loci using Outflank; analyses based on the outlier loci were,
on: (i) the quality (i.e., resolution) of available data and (ii) our interest
in most cases, no different from the same analyses based on all
in larval dispersal within the oceanographically complex Kimberley
4,402 SNP loci.
Bioregion. The hydrodynamic model, which included fine-scale tidal data, was nested within the Ocean Forecasting for Australia Model 3 (OFAM3) simulation (Zhang et al., 2016) and forced by three hourly
3.2 | Genetic subdivision
meteorological measures derived from Kobayashi et al. (2015). The
Patterns of pairwise genetic differentiation revealed small but signifi-
simulation was carried out over the 2011 time period, with hourly sea
cant genetic differences among most sampling sites (i.e., 424 of 496
8
|
DIBATTISTA
ET AL.
tests were significant in Figure 2 and Table S2), which suggests
Kimberley bioregional boundary exhibited ancestry profiles that tran-
restrictions in gene flow between adjacent sites and geographically
sitioned between those typical of other Pilbara and Kimberley sites
distant ones (e.g., NT and Shark Bay). The historical samples col-
(i.e., from Dugong Bay to Emeriau Point). In all such results, q values
lected from sites in the Pilbara in 2002 exhibited higher levels of
of 0.5 to 0.7 were attributed to an uninformative common cluster in
genetic differentiation than those collected in 2014 and 2015 (Fig-
all individuals.
ure 2 and Table S2). Pairwise differentiation was more often signifi-
When applied to the outlier loci data set (66 loci), STRUCTURE
cantly greater than zero in the Kimberley (92% pairwise comparisons
analysis revealed the same pattern of three clear genetic clusters
significant) than in the Pilbara (63% pairwise comparisons significant;
plus a transition zone between the Canning–Kimberley bioregional
Figure 2 and Table S2).
boundary, but with greater clarity since the uninformative cluster across all individuals was no longer evident (Figure 3b). For the outlier loci, the optimum K was 2 (DK = 416.427), which was further
3.3 | Model-based clustering analysis
supported by the ObStruct analysis (r2 = .93, p < .0001); DK was
Bayesian clustering analysis suggested K = 3 populations as the most
markedly higher at K = 3 (DK = 0.773) and K = 4 (DK = 2.096).
parsimonious partitioning of individuals based on the DK metric (Evanno et al., 2005; also see Fig. S1, DK = 1.233). Inspection of the
3.4 | Discriminant analysis of principle components
STRUCTURE ancestry profiles revealed that at K = 3, there was a clear distinction between the sites in Shark Bay and the remaining
The k-means clustering algorithm was optimized at K = 2 in both the
sites to the north (ObStruct analysis: r2 = .96, p < .0001), and at
neutral and outlier data sets. Linear discriminant analysis of principal
K = 4 (DK = 0.179), a further distinction between the Kimberley and
components revealed that for the neutral data set these groups cor-
Pilbara sites was evident (ObStruct analysis: r2 = .98, p < .0001; also
responded to the Shark Bay Bioregion vs. all remaining sites to the
see Figure 3a). In addition, several sites spanning the Canning–
north (Figure 4a). The DAPC analysis of the outlier data set revealed
Bernier Island
Bay of Rest
Passage Island
* * * * * *
* * * *
* * * *
* * * * * * * * * * * * * * * * *
* * * * * * * * * * * * *
* * * * * * * * * * * * * * * *
Paroo Shoal
* * * * * *
Thevenard Island
* * * * * * * * * * * * * * * * * * * * * * * * * *
* * * * * * * * *
Montebello South
* *
* * * * * * *
Passage Island
* * * *
West Moore
*
Ngamakoon
* * *
* * * * * * * * * * * *
* * * * * * * * * * * * * * *
* * * * * * * * * * * * * * * *
Rosemary Island
NE
* * * * * * * * * * *
* * * * * * * * *
Tallon Island
Long Reef
Northern Territory
Blue Holes
Long Reef
Adieu Point
Cape Voltaire
* * * * * * * * * * * * * * Bigge Island
Bigge Island
Cape Voltaire
Blue Holes
Buccaneer Arch
Adieu Point
* * * * * *
Hall Point
Hall Point
Dugong Bay
Dugong Bay
Shenton Bluff
Tallon Island Buccaneer Arch
* * * * * * * * * * *
Emeriau Point
Ngamakoon Shenton Bluff
* * * * * * * * * * * * *
! Cape Bossut
Emeriau Point
* * * * * * * * * * * * * *
! Cape Keraudren
! Cape Bossut
Depuch Island
! Cape Keraudren
Gidley Island
West Moore Depuch Island
! Cape Preston
Gidley Island
* * * * * * * * * * * * * * * * * * *
Montebello North
! Cape Preston Rosemary Island
*
* *
* * * * * * * * * * * * * * * * * * * * *
* * * * * * * * * * * * * *
* * * * * * * * * * * * * * * *
Tantabiddi
*
Montebello North
* * * *
Bay of Rest
Montebello South
* * * * * * * * * * * * * * * * * * * * * * * * *
! Locker Point
Paroo Shoal Thevenard Island
Roberts Island
! Locker Point
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
FST
0.015 0.010 0.005 0.000
Dorre Island
Roberts Island
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Milyering
Milyering Tantabiddi
Bernier Island
SW
SW
F I G U R E 2 Heatmap of pairwise FST values for Lutjanus carponotatus populations with 20 or more individuals in Northwestern Australia based on 4,402 SNP loci. *Indicates significant difference after Narum correction (p < .0074). The four historical sample sites (i.e., 2002) are indicated by small, red arrows. SNP, single nucleotide polymorphism
DIBATTISTA
|
ET AL.
9
(a) K = 2 (–2846366.74 +/– 2.78)
K = 3 (–2844170.61 +/– 4108.04)
Milyering
Bernier Island
Dorre Island
Bernier Island
Dorre Island
Tantabiddi Tantabiddi
Milyering
Bay of Rest
Roberts Island
Bay of Rest
Paroo Shoal
Locker Point
Paroo Shoal
Locker Point
Roberts Island
Passage Island
Thevenard Island Thevenard Island
Montebello South Montebello South
Passage Island
Cape Preston
Montebello North Montebello North
Rosemary Island Rosemary Island
Cape Preston
West Moore
Gidley Island Gidley Island
Depuch Island Depuch Island
West Moore
Cape Bossut
Cape Keraudren
Cape Bossut
Cape Keraudren
Ngamakoon
Emeriau Point
Tallon Island
Shenton Bluff
Blue Holes
Adieu Point Hall Point Dugong Bay
Bigge Island
Long Reef
(b) K = 2 (–44677.63 +/– 0.10)
K = 3 (–38695.01 +/– 13.63)
NE
Ngamakoon
Emeriau Point
Tallon Island
Shenton Bluff
Blue Holes
Adieu Point Hall Point Dugong Bay
Bigge Island
Cape Voltaire
Long Reef
K = 4 (–38388.15 +/– 22.79)
N. Territory
F I G U R E 3 Results of Bayesian clustering for Lutjanus carponotatus populations with 20 or more individuals in Northwestern Australia based on (a) 4,402 SNP loci and (b) 66 outlier loci. Results from K = 2, K = 3 and K = 4 (with mean LnP[K] " SD in parentheses) are presented; K = 2 was the most likely number of clusters in both cases (see Fig. S1). Prior locations were input as priors in these runs. Individuals are represented by vertical bars, each divided according to their estimated probability of ancestry from each of the genetic clusters (represented by blue and orange for K = 2). Sites are ordered northeast to southwest and from left to right. SNP, single nucleotide polymorphism
Cape Voltaire
N. Territory
K = 4 (–2841791.59 +/– 3980.55)
SW
a similar pattern (Figure 4b); however, Shark Bay sites were less dif-
apart geographically are more likely to be separated by multiple bar-
ferentiated. Sites spanning the Canning–Kimberley bioregional
riers and less connected via prevailing currents. Further, this analysis
boundary appeared intermediate between clusters composed of
reinforces the IBD effect, with the slope of IBD changing within dif-
remaining Pilbara and Kimberley sites in both the analysis of neutral
ferent bioregions as outlined below. It should also be noted that
and outlier loci.
hydrodynamic distance was an order of magnitude higher on average
3.5 | Determinants of genetic differentiation
51 sites (0.003 " 0.010 SD), a potential indicator of strong self-
within all 51 sites (0.053 " 0.063 SD) relative to between all
recruitment in this system (Table S4).
For analyses based on all 4,402 SNP loci, four models outperformed all the others, which included the variables geo, hyd, the presence of all three barriers, as well as an interaction between the terms
3.6 | Spatial autocorrelation and IBD
(Table S3). We repeated this analysis with outlier loci only and got
Spatial autocorrelation analysis revealed modest (r ~ .0025), but sig-
the same result. These outcomes did not change when the MEOWs
nificant local-scale genetic structure that dropped away from its ini-
or fisheries management divisions for outlining barriers to dispersal
tial plateau (Figure 5) to cross the x-axis at approximately 300 km
were considered across all sites. Thus, modelling the effects of barri-
in all regions, except in the transition zone (i.e., Dugong Bay to
ers to dispersal, hydrodynamic distance (i.e., connectivity) and geo-
Emeriau Point), where the axis was crossed at 80 km. This crossing
graphic distance on genetic differentiation in this species revealed
point indicates the distance where the random effects of genetic
strong effects for all three factors, but in most cases, these effects
drift, not gene flow, are the primary determinants of genetic com-
could not be distinguished from each other. Indeed, sites that are far
position. A Mantel test revealed that when considering all data,
10
|
DIBATTISTA
(a)
1 2 3 4 5 6 7 8 9
10 11 12 13 14 15 16 17 18
19 20 21 22 23 24 25 26 27
46 47 48 49 50 51 52
37 38 39 40 41 42 43 44 45
28 29 30 31 32 33 34 35 36
(b) 1 2 3 4 5 6 7 8 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 37 38 39 40 41 42 43 44 45
ET AL.
DIBATTISTA
|
ET AL.
11
F I G U R E 4 Scatterplot of DAPC performed on all Lutjanus carponotatus samples based on (a) 4,402 SNP loci and (b) 66 outlier loci. Populations are coloured in north to south order with 95% inertia ellipses. Dots represent individual genotypes, and axes show the first two discriminant functions. IMCRA bioregions: site 1–3, Anson Beagle Bioregion; site 4–31, Kimberley Bioregion; site 32–35, Canning Bioregion; site 36–48, Pilbara Bioregion; site 49–50, Ningaloo Bioregion; site 51–52, Shark Bay Bioregion. IMCRA, Integrated Marine and Coastal Regionalisation of Australia; SNP, single nucleotide polymorphism; DAPC, discriminant analysis of principle components
distance was significantly correlated with genetic differentiation between sites (r = .25, p < .001; Figure 6). Distance was not a sig-
3.7 | Hydrodynamic models
nificant correlate with genetic differentiation when considering sites
As outlined above, hydrodynamic distance in the CONNIE model
in the Kimberley only (r = .08, p = .23), but was significantly corre-
correlated with genetic differentiation across all sampling sites (Shark
lated for the sites in the Pilbara only (r = .50, p = .01). Supporting
Bay to Kimberley) but this could not be distinguished from other fac-
these results, a GLM showed that pairwise FST was on average sig-
tors (geographic distance and barriers to dispersal; Table S3). In con-
nificantly higher in the Kimberley than the Pilbara after removing
trast, the custom fine-scale modelling analysis did not predict the
the effect of distance (p < .0001), but also provided a significant
observed genetic structure in the Kimberley Bioregion. Indeed, a
interaction term, indicating that the relationship between distance
Mantel test with linearized FST found no correlation between particle
and FST differed significantly between these regions (p < .001).
tracking outputs from fine-scale hydrodynamic models (r = .1044,
F I G U R E 5 Spatial autocorrelation as a function of cumulative geographic distance (in kilometres) for Lutjanus carponotatus populations with more than six individuals in Northwestern Australia based on 4,402 SNP loci in all of the Kimberley Bioregion, the Kimberley transition zone only, the Northern Kimberley only, or the Pilbara Bioregion. SNP, single nucleotide polymorphism
F I G U R E 6 Isolation by distance for all Lutjanus carponotatus samples illustrating the relationship between geographic distance and linearized FST. The dashed line indicates the best linear fit
12
|
DIBATTISTA
ET AL.
p = .1846) or Log oceanographic distance (r = .0635, p = .2634).
This genetic break coincides with a well-recognized biogeographic
This indicates that larval behaviour and not particle tracking (i.e., pas-
boundary and oceanographic features at North West Cape near Nin-
sive dispersal) may be strongly influencing connectivity at fine-scales
galoo Reef (Commonwealth of Australia, 2006; Hutchins, 2001;
(e.g., Dixson, Abrego, & Hay, 2014).
Spalding et al., 2007; Woo, Pattiaratchi, & Schroeder, 2006). Wilson (2013) suggested that the effect of the poleward flowing Leeuwin Current across this region, the most dominant boundary current of
4 | DISCUSSION
the western coast of Australia (Figure 1; Godfrey & Ridgway, 1985), may not be effective in preventing exchange for species with plank-
This study is one of the first to investigate coastal connectivity
totrophic larvae (such as L. carponotatus). Some studies have sug-
within the remote Kimberley Bioregion of WA and includes exten-
gested that this barrier is gradual rather than abrupt (Johnson et al.,
sive sampling across ~2,500 km covering all of the major marine
1993; Thomas et al., 2014, 2017; Whitaker, 2006), and potentially
bioregions in NWA. We provide evidence of restricted connectivity
results from a mesoscale eddy that advects larvae offshore (Woo
between geographically distant sites and, in some cases, neighbour-
et al., 2006).
ing sites within bioregions separated by a few kilometres. These
A second apparent boundary was observed between the Kim-
results support earlier observations of restricted connectivity in
berley and Canning bioregions (Figures 2–4). This pattern was most
fishes and corals between the Pilbara, Ningaloo Reef and Shark Bay
evident in the STRUCTURE analysis of outlier SNPs as a region of
bioregions (Johnson et al., 1993; Thomas, Kennington, Evans, Ken-
progressive admixture between two apparently homogenous genetic
drick, & Stat, 2017; Underwood, 2009; Whitaker, 2006). In addition,
clusters representing the Kimberley and NT, and the combined Pil-
we identified a localized region of admixture within the Buccaneer
bara and Canning bioregions (Figure 3). The result was also
Archipelago and adjacent waters between the otherwise genetically
reflected in the DAPC analysis of both neutral and outlier SNPs
distinct Kimberley and Canning bioregions. This region of admixture
(Figure 4), but again it was not as clear for the neutral data set.
corresponds to a well-defined biogeographic boundary based on
These results, supported by a distinctive pattern of low spatial
shifts in faunal composition for several taxa, including fishes (Hutch-
autocorrelation in this region (Figure 5), indicate that it likely repre-
ins, 2001; Travers et al., 2010), mangroves and molluscs (Wilson,
sents restricted dispersal over ~180 km near the tip of the Dampier
2013), and experiences the most extreme tidal fluctuations of tropi-
Peninsula and Buccaneer Archipelago (Figure 1). This is consistent
cal waters globally.
with a recognized biogeographic break in marine communities (Wilson, 2013) that parallels a change in the underlying coastal geology
4.1 | Genetic diversity and broadscale subdivision across NWA
(e.g., from Proterozoic sandstone and basic intrusive rock in the
Levels of genetic diversity were similar throughout the sampling
reefs in the north to soft substrate communities in the south). This
range of L. carponotatus (Tables 1 and S1), which may be due to the
break also corresponds with an abrupt change of coastal geomor-
! re ! & Leis, 2010) relatively long PLD of this species (33–38 days; Que
phology from the ria coast of the Kimberley with its numerous
and large genetic neighbourhoods that we observed (i.e., positive
islands and embayments to the Canning coast comprised of coastal
spatial autocorrelation up to 450 km; Figure 5). In contrast, two
dunes and tidal flats (Brocx & Semeniuk, 2010). As discussed below,
dominant patterns of genetic subdivision were evident from the SNP
a change in tidal regimes across this region is also a likely driver of
genotyping of L. carponotatus. The first was an IBD effect, where on
this dispersal pattern.
north to Mesozoic sandstone with diverse sedimentary structures in the south) and associated dominant benthic habitat (e.g., from coral
average sampling sites were genetically most similar to their closest
This study is the first example of an investigation of a coastal fish
neighbours and least similar to distant sites (Figure 6). By implication,
in this region using SNPs. Studies on several other marine taxa from
dispersal is limited on the scale of this investigation (~2,500 km) and
this region including a brooding reef fish Pomacentrus milleri (Berry,
proceeds in a stepping-stone manner. This pattern is likely also true
Travers, Evans, Moore, & Hernawan, 2017) and a mangrove Avicennia
for other species in the region from the Lutjanidae or Lethrinidae
marina (R. M. Binks, unpublished data) show similar broadscale
families with similar life histories, although, like L. carponotatus, we
genetic structure with the use of SNPs. The genetic discontinuities
would expect this correlation to be weak due to the abundance and
between broadscale bioregions evident in the coastal L. carponotatus
fecundity of these species in addition to their larval dispersal capabil-
samples contrast to the widespread connectivity observed in deep-
ities (Berry, England, Marriott, Burridge, & Newman, 2012; Kritzer,
water and epipelagic species with similar PLDs. Studies using nuclear
2004; Marriott et al., 2010; Newman, Cappo, & Williams, 2000; Tra-
microsatellite markers and mitochondrial DNA found evidence for
vers et al., 2006).
weak genetic structure between sites along the edge of the continen-
In addition to the IBD effect, several pronounced genetic discon-
tal shelf in this region (Bentley et al., 2014; Gaither et al., 2011;
tinuities were evident among samples of L. carponotatus from across
Johnson et al., 1993; Kennington et al., 2017), suggesting that the
NWA. One of the obvious genetic breaks was between the Shark
processes operating offshore and across continental shelves and
Bay Bioregion and all locations of the northwest continental shelf
islands may differ significantly from those processes influencing dis-
(Ningaloo Reef, Pilbara, Canning and Kimberley) including the NT.
persal in coastal species. Alternatively, the genetic structure detected
DIBATTISTA
|
ET AL.
13
among coastal sites in the current study, relative to the offshore and
L. carponotatus, in most cases, linear distance provided a comparable
deepwater studies, may be a function of the higher resolution mark-
explanation
ers that we used. Veilleux et al. (2011) found no genetic structure in
(Table S3). This reflects the fact that the hydrodynamic distance and
L. carponotatus at similar coastal sampling sites spread across the
number of barriers almost exactly tracked linear distance (i.e.,
Kimberley to the Pilbara bioregions, albeit with only one mitochon-
for
the
observed
patterns
of
genetic
structure
collinearity; r = !.25, p < .001), due in part to the large spatial scale
drial marker. Although connectivity estimated using the different
of the study. That is, large distances between sampling sites (up to
markers is not directly comparable between the overlapping studies
hundreds of kilometres) over a gradual latitudinal gradient lend itself
given their different rates of evolution and mode of inheritance, as
to other forms of change relative to that gradient.
well as a focus on fine-scale vs. broadscale information, comparisons
Fine-scale genetic partitioning in the Kimberley Bioregion was not
with the offshore studies highlight the need to incorporate NGS
predicted by the custom particle tracking model used in this study nor
approaches into stock assessments.
the oceanographic distance. This highlights the complex biological and physical characteristics of larval dispersal in the Kimberley. First, the 2-
4.2 | Fine-scale connectivity across NWA
km scale of this model likely does not account for the fine-scale eddy
The broadscale genetic discontinuities between bioregions were
whole it is considered well mixed; this disconnect may affect the sig-
overlaid by subtle genetic differentiation within each bioregion.
nal-to-noise ratio in our custom model. Second, late-stage larval L. car-
Genetic patterns also differed between the bioregions, indicating
ponotatus are capable of swimming speeds up to 33 cm/s and may be
that L. carponotatus dispersal behaviour may be variable across its
able to limit their dispersal by swimming to larval retention zones near
range. On average, genetic differentiation between sites was higher
! re ! & Leis, 2010) or via vertical migration by interacting with reefs (Que
in the Kimberley than the Pilbara bioregions (Figure 2 and Table S2),
currents at different depths (Stephenson, Power, Laffan, & Suthers,
implying that dispersal is more restricted in the Kimberley. A moder-
2015). Indeed, Wolanski and Kingsford (2014) demonstrated that coral
ate IBD effect was evident among Pilbara samples, yet not in the
reef fish larvae with directional swimming abilities using olfactory and
Kimberley (Figure 6). This also suggests greater restriction to gene
auditory cues appear to self-recruit an order of magnitude more fre-
flow in the Kimberley than the Pilbara, and the more idiosyncratic
quently than passive particles in their predictive model. The genetic
patterning likely reflects the more powerful tidal regime (e.g., tidal
results support the larval behaviour hypothesis given that greater
pumping at the mouth of King Sound) and complex coastal topogra-
retention or less dispersal of L. carponotatus in the Kimberley Biore-
phy (e.g., numerous islands, rocky islets and bays) present in the
gion was observed vs. elsewhere on the coast of NWA.
effects of the large tides in the Kimberley Bioregion, although on the
Kimberley in contrast to the more linear current flow along the Canning and Pilbara coasts (Feng, Colberg, Slawinski, Berry, & Babcock, 2016). Larval L. carponotatus on the GBR are capable of actively
4.4 | Caveats and future directions
influencing their dispersal and settlement through swimming and
Management of L. carponotatus in NWA is currently based in part on
!re ! & Leis, 2010). However, the maximum sensory capabilities (Que
recognizing three overarching management units corresponding to:
larval swimming speed recorded for this species is ~33 cm/s, which
(i) the NT, (ii) combined Kimberley, Canning and Pilbara Bioregions
is considerably less than the maximum tidal velocity in the vicinity of
and (iii) the Gascoyne Bioregion, which includes both Shark Bay and
the transition zone identified here (100 cm/s; Lowe, Leon, Symonds,
Ningaloo Reef in fisheries management arrangements. Our results
Falter, & Gruber, 2015; Wolanski & Spagnol, 2003). Although spawn-
indicate that the management boundaries of these stocks require re-
!re ! & Leis, 2010), L. caring probably occurs during neap tides (Que
evaluation or alternatively that the barriers to connectivity need to
! re ! & Leis, ponotatus have a relatively long PLD (33–38 days; Que
be considered within management arrangements. Moreover, this
2010), which would expose them to the full spectrum of tidally dri-
highlights the need for continuous improvement of adaptive man-
ven water movement in this region. Tidal currents and their interac-
agement processes as new techniques arise. In this case, NGS and
tions with complex coastal topography may thus play an active role
SNP genotyping has improved our ability to determine restrictions to
in larval retention by limiting the opportunities for net larval trans-
gene flow over ecological timescales.
port to periods of slack water in the ebb/flood tidal cycle and during
The distinctiveness of the Shark Bay samples from all other
neap tides known as selective tidal-stream transport (Forward &
bioregions indicates that the Gascoyne management boundary is not
Tankersley, 2001), when the swimming speed of larvae exceeds the
supported because sites north of Shark Bay have greater affinities to
speed of tidal currents.
sites in the Pilbara Bioregion. In addition, support for separate management of L. carponotatus from the NT is equivocal. NT samples
4.3 | Seascape genomics
were significantly, albeit weakly, genetically differentiated from all
The integration of geographic and oceanographic variables to explain
STRUCTURE and DAPC analyses. However, a large sampling gap
genetic signals of differentiation, often referred to as seascape geno-
exists between the Kimberley and NT sites despite available habitat,
mics, is a growing field (Selkoe et al., 2016). Although we explored
and so it is unclear whether the genetic differentiation of the NT
hydrodynamic distance across the entire sampling range of
samples reflects a genuine discontinuity, or a continuation of the
other samples (Figure 2), and appeared weakly divergent in both
14
|
DIBATTISTA
ET AL.
isolation-by-distance effect observed elsewhere in the range of
dispersal to limit any latitudinal effects on genetic diversity across
L. carponotatus. Although unavoidable in this study due to the
the sampling range. Through multiple lines of evidence however, and
remoteness of the region, this gap in the data highlights the neces-
despite an apparent stepping-stone process to dispersal, genetic con-
sity to sample evenly across a species’ range (Meirmans, 2015).
nectivity was not homogeneous across the sampling range. Indeed,
The transition zone identified around the Dampier Peninsula that
the bioregion with the most extreme hydrodynamic environment in
separates the Kimberley from the Pilbara/Canning populations may
our tropical oceans, the Kimberley, displayed less connectivity among
be uniquely influenced by the extreme tidal flushing and seasonal
sampling sites than did the Pilbara/Canning bioregion, which exhibits
9
3
!1
;
lower tidal flows and a less complex coastal topography. Thus, the
Wolanski & Spagnol, 2003) at the mouth of King Sound. Fine-scale
spatial scale of sampling, hydrodynamic effects within the sampling
parentage analysis (Harrison et al., 2012; Pusack, Christie, Johnson,
range, genetic approach used and the biology of the study species
Stallings, & Hixon, 2014) coupled with larval dispersal models that
are all important considerations when inferring patterns of dispersal
include both tides and currents, explored over multiple lunar phases
under a management framework.
reductions in salinity (i.e., mean runoff 5.75 9 10 m L year
and spawning seasons, may be needed to understand the sourcesink dynamics of larvae and successful recruitment in this zone. This investigation would be beneficial given its applicability to other harvested fish species occurring within the transition zone.
ACKNOWLEDGEMENTS The development and implementation of this project benefited
Samples included in this analysis were collected over a span of
greatly from our collaborating scientists on the WAMSI 1.1.3 project:
15 years. Although older samples exhibited their closest affinity to
Kathryn McMahon, Jim Underwood, Zoe Richards and James Gil-
newer samples from the same bioregion, in most cases they also
mour. For cultural advice and permissions, we thank the Kimberley
exhibited a distinct genetic composition. Based on post hoc analyses,
Land Council and the Wunambal Gaambera, Dambimangari, Mayala,
we were able to exclude at least one mechanism of DNA degrada-
Bardi-Jawi and Karajarri people. For field assistance, we thank Fran-
tion, which could have accounted for this pattern (deamination or C/
cis Woolagoodja, Danial Oades, Kevin George, Damon Pyke, Kevin
T transitions; Fig. S2), by comparing SNP type (i.e., transition vs.
Ejai, Azden Howard, Kevin Dougal, Tesha Stumpagee, Phillip
transversion) between samples. This result may instead represent a
McCarthy, Peter Hunter, Zac Ejai, Paul Davey, Trevor Sampi, Chris
real shift in allele frequencies over time, indicating the potential for
Sampi, Sandy Isaac, Alec Isaac, Craig Skepper, Grace Davis, Gabby
changing patterns of connectivity among sites, and in this case, an
Mitsopoulos, Fiona Webster, Todd Bond, Matt Birt, Tyrone Ridgway
increase in connectivity.
and Nicole Ryan. We also thank The Western Australian Marine
The coastal-associated L. carponotatus is harvested by commer-
Science Institute, Kelly Waples, Kim Friedman and Stuart Field. For
cial, recreational, charter and indigenous fishers at various locations
logistics and advice, we thank Cygnet Bay Pearls, James Brown, Erin
throughout its range. Currently, spatial partitioning based on genetic
McGinty, Ali McCarthy, Scott Whitlam and Duncan Smith. For
separation in the more northern bioregions largely conform to exist-
molecular guidance and genotyping service, we thank DaRT and
ing management boundaries. However, the genetic transition zone
Andrzej Killian. For assistance implementing the information-theore-
that spans a distance of ~180 km in the Buccaneer Archipelago and
tic approach, we thank Pablo Saenz-Agudelo. For computing assis-
adjacent waters within the Kimberley Bioregion is unique. The iden-
tance, we thank Ian LeCoultre, Tim Ho, CSIRO High Performance
tification of this transition zone was unique to the NGS and SNP
Computing Centre, Philippe Moncuquet, Annette McGrath, the
genotyping approach; this pattern was absent in a study of Kimber-
CSIRO Bioinformatics Core and Bernd Gruber for R scripting lessons.
ley to Pilbara/Canning populations of L. carponotatus using mitochondrial markers (Veilleux et al., 2011). That said, it is still unclear whether these transition zones isolate populations on either side of
DATA ACCESSIBILITY
them, and influence the evolution of life history traits such as repro-
The genetic data are permanently and publically available from the
duction, spawning dynamics and growth.
CSIRO Data Access Portal (DAP) at https://data.csiro.au/dap/landing page?pid=csiro:20195. Individual SNP genotypes for 1,016 individuals are available in the DRYAD data depository https://doi.org/10.
5 | CONCLUSION The use of NGS and SNP genotyping in this population connectivity study has highlighted the advantages of adopting higher resolution
5061/dryad.4hk28.
AUTHOR CONTRIBUTIONS
genetic markers in a seascape genomics context to re-evaluate the
O.B., M.T., G.M. and R.E. designed the study. M.T., G.M., R.E., S.N.,
potential for restrictions to gene flow in species that once were con-
S.M. and T.S. collected the samples. J.D. and O.B. executed the
sidered panmictic. Our study also highlights the benefits of collecting
genetic analyses. M.F. and R.G. executed the oceanographic mod-
samples across an extensive and broad range of habitats and hydro-
elling analyses. J.D., O.B., M.T., G.M. and R.E. interpreted the data
dynamic regimes within a species range. Our model species, L. car-
and led the writing of the manuscript. All authors contributed to the
ponotatus, has a PLD of 33–38 days, which likely enables sufficient
final draft of the manuscript.
DIBATTISTA
|
ET AL.
ORCID Joseph D. DiBattista Michael J. Travers Stephen J. Newman
http://orcid.org/0000-0002-5696-7574 http://orcid.org/0000-0002-3072-1699 http://orcid.org/0000-0002-5324-5568
REFERENCES Almany, G. R., Planes, S., Thorrold, S. R., Berumen, M. L., Bode, M., Saenz-Agudelo, P., . . . Nanninga, G. B. (2017). Larval fish dispersal in a coral-reef seascape. Nature Ecology & Evolution, 1, s41559–017. Anderson, D. R. (2008). Information theory and entropy. New York, NY: Springer. Anderson, W. D., & Allen, G. R. (2001). Lutjanidae. In K. E. Carpenter & V. H. Niem (Eds.), FAO species identification guide for fisheries purposes. The living marine resources of the Western Central Pacific, vol. 5, part 3 (pp. 2840–2918). Rome: FAO. Andrews, K. R., Good, J. M., Miller, M. R., Luikart, G., & Hohenlohe, P. A. (2016). Harnessing the power of RADseq for ecological and evolutionary genomics. Nature Reviews Genetics, 17, 81–92. Balkenhol, N., Waits, L. P., & Dezzani, R. J. (2009). Statistical approaches in landscape genetics: An evaluation of methods for linking landscape and genetic data. Ecography, 32, 818–830. Bentley, B. P., Harvey, E. S., Newman, S. J., Welch, D. J., Smith, A. K., & Kennington, W. J. (2014). Local genetic patchiness but no regional differences between Indo-West Pacific populations of the dogtooth tuna Gymnosarda unicolor. Marine Ecology Progress Series, 506, 267– 277. Berry, O., England, P., Marriott, R. J., Burridge, C. P., & Newman, S. J. (2012). Understanding age-specific dispersal in fishes through hydrodynamic modelling, genetic simulations and microsatellite DNA analysis. Molecular Ecology, 21, 2145–2159. Berry, O., Travers, M., Evans, R., Moore, G., & Hernawan, U. (2017). Genomic connectivity in a tropical reef fish from the Kimberley, Pilbara and Gascoyne bioregions of Western Australia. Report 1.1.3.4a submitted to the Western Australian Marine Science Institution. 26 pp. Berumen, M. L., Almany, G. R., Planes, S., Jones, G. P., Saenz-Agudelo, P., & Thorrold, S. R. (2012). Persistence of self-recruitment and patterns of larval connectivity in a marine protected area network. Ecology and Evolution, 2, 444–452. Brocx, M., & Semeniuk, V. (2010). Coastal geoheritage: A hierarchical approach to classifying coastal types as a basis for identifying geodiversity and sites of significance in Western Australia. Journal of the Royal Society of Western Australia, 93, 81–113. Broderick, D., Ovenden, J. R., Buckworth, R. C., Newman, S. J., Lester, R. J. G., & Welch, D. J. (2011). Genetic population structure of grey mackerel Scomberomorus semifasciatus in northern Australia. Journal of Fish Biology, 79, 633–661. Burgess, S. C., Nickols, K. J., Griesemer, C. D., Barnett, L. A., Dedrick, A. G., Satterthwaite, E. V., . . . Botsford, L. W. (2014). Beyond connectivity: How empirical methods can quantify population persistence to improve marine protected-area design. Ecological Applications, 24, 257–270. Calaway, R., Weston, S., & Tenenbaum, D. (2014). doParallel: Foreach parallel adaptor for the ‘parallel’ package. Retrieved from https://cra n.r-project.org/package=doParallel Cawthorn, D.-M., Steinman, H. A., & Witthuhn, R. C. (2011). Comparative study of different methods for the extraction of DNA from fish species commercially available in South Africa. Food Control, 22, 231– 244. Chin, A., Sweatman, H., Forbes, S., Perks, H., Walker, R., Jones, G., . . . Edgar, G. (2008). 2008 status of the coral reefs in Australia and
15
Papua New Guinea. In C. Wilkinson (Ed.), Status of coral reefs of the world: 2008 (pp. 159–176). Townsville, Qld: Global Coral Reef Monitoring Network. Commonwealth of Australia (2006). A guide to the integrated marine and coastal regionalisation of Australia version 4.0. Canberra: Department of the Environment and Heritage. Condie, S. A., Hepburn, M., & Mansbridge, J. (2012). Modelling and visualisation of connectivity on the Great Barrier Reef. Proceedings of the 12th International Coral Reef Symposium, Cairns, Australia, 9–13 July 2012. Correa, C., & Hendry, A. P. (2012). Invasive salmonids and lake order interact in the decline of puye grande Galaxias platei in western Patagonia lakes. Ecological Applications, 22, 828–842. Cowen, R. K., & Sponaugle, S. (2009). Larval dispersal and marine population connectivity. Marine Science, 1, 443–466. Cruz, V. M., Kilian, A., & Dierig, D. A. (2013). Development of DArT marker platforms and genetic diversity assessment of the U.S. collection of the new oilseed crop Lesquerella and related species. PLoS ONE, 8, e64062. D’Adamo, N. D., Fandry, C., Buchan, S., & Domingues, C. (2009). Northern sources of the Leeuwin Current and the ‘Holloway Current’ on the North West Shelf. Journal of the Royal Society of Western Australia, 92, 53–66. Department of Parks and Wildlife (2016). North Kimberley Marine Park Joint management plan Uunguu, Balanggarra, Miriuwung Gajerrong, and Wilinggin management areas. Perth, WA: Management Plan 89, Department of Parks and Wildlife. Dixson, D. L., Abrego, D., & Hay, M. E. (2014). Chemically mediated behavior of recruiting corals and fishes: A tipping point that may limit reef recovery. Science, 345, 892–897. Domingues, C. M., Maltrud, M. E., Wijffels, S. E., Church, J. A., & Tomczak, M. (2007). Simulated Lagrangian pathways between the Leeuwin Current System and the upper-ocean circulation of the southeast Indian Ocean. Deep-Sea Research Part II—Topical Studies in Oceanography, 54, 797–817. Drew, J. A., & Barber, P. H. (2012). Comparative phylogeography in Fijian coral reef fishes: A multi-taxa approach towards marine reserve design. PLoS ONE, 7, e47710. Epskamp, S., Cramer, A. O., Waldorp, L. J., Schmittmann, V. D., & Borsboom, D. (2012). qgraph: Network visualizations of relationships in psychometric data. Journal of Statistical Software, 48, 1–18. Evanno, G., Regnaut, S., & Goudet, J. (2005). Detecting the number of clusters of individuals using the software STRUCTURE: A simulation study. Molecular Ecology, 14, 2611–2620. Evans, R. D., Van Herwerden, L., Russ, G. R., & Frisch, A. J. (2010). Strong genetic but not spatial subdivision of two reef fish species targeted by fishers on the Great Barrier Reef. Fisheries Research, 102, 16–25. Falush, D., Stephens, M., & Pritchard, J. K. (2003). Inference of population structure using multilocus genotype data: Linked loci and correlated allele frequencies. Genetics, 164, 1567–1587. Feng, M., Colberg, F., Slawinski, D., Berry, O., & Babcock, R. (2016). Ocean circulation drives heterogeneous recruitments and connectivity among coral populations on the North West Shelf of Australia. Journal of Marine Systems, 164, 1–12. Fisher, R., Leis, J. M., Clark, D. L., & Wilson, S. K. (2005). Critical swimming speeds of late-stage coral reef fish larvae: Variation within species, among species and between locations. Marine Biology, 147, 1201–1212. Forward, R. B., & Tankersley, R. A. (2001). Selective tidal-stream transport of marine animals. Oceanography and Marine Biology: An Annual Review, 39, 305–353. Fox, N. J., & Beckley, L. E. (2005). Priority areas for conservation of Western Australian coastal fishes: A comparison of hotspot, biogeographical and complementarity approaches. Biological Conservation, 125, 399–410.
16
|
Gaither, M. R., Jones, S. A., Kelley, C., Newman, S. J., Sorenson, L., & Bowen, B. W. (2011). High connectivity in the deepwater snapper Pristipomoides filamentosus (Lutjanidae) across the Indo-Pacific with isolation of the Hawaiian Archipelago. PLoS ONE, 6, e28913. Gayevskiy, V., Klaere, S., Knight, S., & Goddard, M. R. (2014). ObStruct: A method to objectively analyse factors driving population structure using Bayesian ancestry profiles. PLoS ONE, 9, e85196. Godfrey, J. S., & Ridgway, K. R. (1985). The large-scale environment of the poleward-flowing Leeuwin Current, Western Australia: Longshore steric height gradients, wind stresses and geostrophic flow. Journal of Physical Oceanography, 15, 481–495. !, X., Guino ! , E., Mercader, J. M., Estivill, Gonz! alez, J. R., Armengol, L., Sole X., & Moreno, V. (2007). SNPassoc: An R package to perform whole genome association studies. Bioinformatics, 23, 654–655. Grewe, P. M., Feutry, P., Hill, P. L., Gunasekera, R. M., Schaefer, K. M., Itano, D. G., . . . Davies, C. R. (2015). Evidence of discrete yellowfin tuna (Thunnus albacares) populations demands rethink of management for this globally important resource. Scientific Reports, 5, 16916. Halpern, B. S., Frazier, M., Potapenko, J., Casey, K. S., Koenig, K., Longo, C., . . . Walbridge, S. (2015). Spatial and temporal changes in cumulative human impacts on the world’s ocean. Nature Communications, 6, 7615. Harrison, H. B., Williamson, D. H., Evans, R. D., Almany, G. R., Thorrold, S. R., Russ, G. R., . . . Berumen, M. L. (2012). Larval export from marine reserves and the recruitment benefit for fish and fisheries. Current Biology, 22, 1023–1028. Harvey, E. S., Newman, S. J., McLean, D. L., Cappo, M., Meeuwig, J. J., & Skepper, C. L. (2012). Comparison of the relative efficiencies of stereo-BRUVs and traps for sampling tropical continental shelf demersal fishes. Fisheries Research, 125–126, 108–120. Hock, K., Wolff, N. H., Condie, S. A., Anthony, K. R. N., & Mumby, P. J. (2014). Connectivity networks reveal the risks of crown-of-thorns starfish outbreaks on the Great Barrier Reef. Journal of Applied Ecology, 51, 1188–1196. Horne, J. B., Momigliano, P., van Herwerden, L., & Newman, S. J. (2013). Murky waters: Searching for structure in genetically depauperate blue threadfin populations of Western Australia. Fisheries Research, 146, 1–6. Horne, J. B., Momigliano, P., Welch, D. J., Newman, S. J., & van Herwerden, L. (2011). Limited ecological population connectivity suggests low demands on self-recruitment in a tropical inshore marine fish (Eleutheronema tetradactylum: Polynemidae). Molecular Ecology, 20, 2291–2306. Horne, J. B., Momigliano, P., Welch, D. J., Newman, S. J., & van Herwerden, L. (2012). Searching for common threads in threadfins: Phylogeography of Australian polynemids in space and time. Marine Ecology Progress Series, 449, 263–276. Hutchins, J. B. (2001). Biodiversity of shallow reef fish assemblages in Western Australia using a rapid censusing technique. Records of the Western Australian Museum, 20, 247–270. Johnson, M. S., Hebbert, D. R., & Moran, M. J. (1993). Genetic analysis of populations of north-western Australian fish species. Marine and Freshwater Research, 44, 673–685. Johnson, M. S., & Joll, L. M. (1993). Genetic subdivision of the pearl oyster Pinctata maxima (Jameson, 1901) (Mollusca: Pteriidae) in northern Australia. Marine and Freshwater Research, 44, 519–526. Jombart, T. (2008). adegenet: A R package for the multivariate analysis of genetic markers. Bioinformatics, 24, 1403–1405. Kennington, W. J., Keron, P. W., Harvey, E. S., Wakefield, C. B., Williams, A. J., Halafihi, T., & Newman, S. J. (2017). High intra-ocean, but limited inter-ocean genetic connectivity in populations of the deepwater oblique-banded snapper Pristipomoides zonatus (Pisces: Lutjanidae). Fisheries Research, 193, 242–249. Kilian, A., Wenzl, P., Huttner, E., Carling, J., Xia, L., Blois, H., . . . Uszynski, G. (2012). Diversity arrays technology: A generic genome profiling technology on open platforms. In F. Pompanon & A. Bonin (Eds.),
DIBATTISTA
ET AL.
Data production and analysis in population genomics: Methods and protocols (pp. 67–89). Totowa, NJ: Humana Press. Kobayashi, S., Ota, Y., Harada, Y., Ebita, A., Moriya, M., Onoda, H., . . . Takahashi, K. (2015). The JRA-55 reanalysis: General specifications and basic characteristics. Journal of the Meteorological Society of Japan, 93, 5–48. Kritzer, J. P. (2004). Sex-specific growth and mortality, spawning season, and female maturation of the stripey bass (Lutjanus carponotatus) on the Great Barrier Reef. Fisheries Bulletin, 102, 94–107. Krueck, N. C., Ahmadia, G. N., Green, A., Jones, G. P., Possingham, H. P., Riginos, C., . . . Mumby, P. J. (2016). Incorporating larval dispersal into MPA design for both conservation and fisheries. Ecological Applications, 27, 925–941. Leis, J. M. (2015). Is dispersal of larval reef fishes passive? In C. Mora (Ed.), Ecology of fishes on coral reefs (pp 223). Cambridge: Cambridge University Press. Lischer, H. E. L., & Excoffier, L. (2012). PGDSpider: An automated data conversion tool for connecting population genetics and genomics programs. Bioinformatics, 28, 298–299. Lotterhos, K. E., & Whitlock, M. C. (2015). The relative power of genome scans to detect local adaptation depends on sampling design and statistical method. Molecular Ecology, 24, 1031–1046. Lowe, R. J., Leon, A. S., Symonds, G., Falter, J. L., & Gruber, R. (2015). The intertidal hydraulics of tide-dominated reef platforms. Journal of Geophysical Research: Oceans, 120, 4845–4868. Lowe, R. J., Pivan, X., Falter, J., Symonds, G., & Gruber, R. (2016). Rising sea levels will reduce extreme temperature variations in tide-dominated reef habitats. Science Advances, 2, e1600825. Marriott, R. J., Jarvis, N. D. C., Adams, D. J., Gallash, A. E., Norriss, J., & Newman, S. J. (2010). Maturation and sexual ontogeny in the spangled emperor Lethrinus nebulosus. Journal of Fish Biology, 76, 1396–1414. McKinnon, A., Duggan, S., Holliday, D., & Brinkman, R. (2015). Plankton community structure and connectivity in the Kimberley-Browse region of NW Australia. Estuarine and Coastal Shelf Science, 153, 156–167. McLean, D. L., Langlois, T. J., Newman, S. J., Holmes, T. H., Birt, M. J., Bornt, K. R., . . . Fisher, R. (2016). Distribution, abundance, diversity and habitat associations of fishes across a bioregion experiencing rapid coastal development. Estuarine and Coastal Shelf Science, 178, 36–47. Meirmans, P. G. (2015). Seven common mistakes in population genetics and how to avoid them. Molecular Ecology, 24, 3223–3231. Meirmans, P. G., & Van Tienderen, P. H. (2004). GENOTYPE and GENODIVE: Two programs for the analysis of genetic diversity of asexual organisms. Molecular Ecology Notes, 4, 792–794. Molony, B. W., Newman, S. J., Joll, L., Lenanton, R. C. J., & Wise, B. (2011). Are Western Australian waters the least productive waters for finfish across two oceans? A review with a focus on finfish resources in the Kimberley region and North Coast Bioregion. Journal of the Royal Society of Western Australia, 94, 323–332. Moore, G. I., & Morrison, S. M. (2009). Fishes of three North West Shelf atolls off Western Australia: Mermaid (Rowley Shoals), Scott and Seringapatam Reefs. Records of the Western Australian Museum, Supplement, 77, 221–255. Moore, G. I., Morrison, S. M., Hutchins, J. B., Allen, G. R., & Sampey, A. (2014). Kimberley marine biota. Historical data: Fishes. Records of the Western Australian Museum Supplement, 84, 161–206. Moore, C. H., Radford, B. T., Possingham, H. P., Heyward, A. J., Stewart, R. R., Watts, M. E., . . . Bryce, C. W. (2016). Improving spatial prioritisation for remote marine regions: Optimising biodiversity conservation and sustainable development trade-offs. Scientific Reports, 6, 32029. Newman, S. J., Cappo, M., & Williams, D. M. c. B. (2000). Age, growth and mortality of the stripey, Lutjanus carponotatus (Richardson) and the brown-stripe snapper, L. vitta (Quoy and Gaimard) from the central Great Barrier Reef, Australia. Fisheries Research, 48, 263–275.
DIBATTISTA
ET AL.
Newman, S. J., & Williams, D. M. c. B. (1996). Variation in reef associated assemblages of the Lutjanidae and Lethrinidae at different distances offshore in the central Great Barrier Reef. Environmental Biology of Fishes, 46, 123–138. Nosil, P., Funk, D. J., & Ortiz-Barrientos, D. (2009). Divergent selection and heterogeneous genomic divergence. Molecular Ecology, 18, 375–402. Oksanen, J., Kindt, R., Legendre, P., O’Hara, B., Stevens, M. H. H., Oksanen, M. J., & Suggests, M. A. S. S. (2007). The vegan package. Community Ecology Package, 10, 631–637. Ovenden, J. R., Lloyd, J., Newman, S. J., Keenan, C. P., & Slater, L. S. (2002). Spatial genetic subdivision between northern Australian and southeast Asian populations of Pristipomoides multidens: A tropical marine reef fish species. Fisheries Research, 59, 57–69. Pante, E., & Simon-Bouhet, B. (2013). marmap: A package for importing, plotting and analyzing bathymetric and topographic data in R. PLoS ONE, 8, e73051. Paradis, E. (2010). pegas: An R package for population genetics with an integrated–modular approach. Bioinformatics, 26, 419–420. Peakall, R., Ruibal, M., & Lindenmayer, D. B. (2003). Spatial autocorrelation analysis offers new insights into gene flow in the Australian bush rat, Rattus fuscipes. Evolution, 57, 1182–1195. Peakall, R. O. D., & Smouse, P. E. (2006). GENALEX 6: Genetic analysis in Excel. Population genetic software for teaching and research. Molecular Ecology Notes, 6, 288–295. Poore, G. C. B., & O’Hara, T. D. (2007). Marine biogeography and biodiversity of Australia. In S. D. Connell & B. M. Gillanders (Eds.), Marine ecology (pp. 177–198). Melbourne, Vic.: Oxford University Press. Pritchard, J. K., Stephens, M., & Donnelly, P. (2000). Inference of population structure using multilocus genotype data. Genetics, 155, 945– 959. Pusack, T. J., Christie, M. R., Johnson, D. W., Stallings, C. D., & Hixon, M. A. (2014). Spatial and temporal patterns of larval dispersal in a coralreef fish metapopulation: Evidence of variable reproductive success. Molecular Ecology, 23, 3396–3408. !re !, G., & Leis, J. M. (2010). Settlement behaviour of larvae of the Que Stripey Snapper, Lutjanus carponotatus (Teleostei: Lutjanidae). Environmental Biology of Fishes, 88, 227–238. Rellstab, C., Gugerli, F., Eckert, A. J., Hancock, A. M., & Holderegger, R. (2015). A practical guide to environmental association analysis in landscape genomics. Molecular Ecology, 24, 4348–4370. Richards, Z. T., Garcia, R. A., Wallace, C. C., Rosser, N. L., & Muir, P. R. (2015). A diverse assemblage of reef corals thriving in a dynamic intertidal reef setting (Bonaparte Archipelago, Kimberley, Australia). PLoS ONE, 10, e0117791. Richards, Z. T., & O’Leary, M. J. (2015). The coralline algal cascades of Tallon Island (Jalan) fringing reef, NW Australia. Coral Reefs, 34, 595. Riginos, C., Crandall, E. D., Liggins, L., Bongaerts, P., & Treml, E. A. (2016). Navigating the currents of seascape genomics: How spatial analyses can augment population genomic studies. Current Zoology, 62, 1–21. Ryan, K. L., Hall, N. G., Lai, E. K., Smallwood, C. B., Taylor, S. M., & Wise, B. S. (2015). State-wide survey of boat-based recreational fishing in Western Australia 2013/14. Fisheries Research Report No. 268. North Beach, WA: Fisheries Research Division, Department of Fisheries, 168 pp. Saenz-Agudelo, P., Jones, G. P., Thorrold, S. R., & Planes, S. (2011). Connectivity dominates larval replenishment in a coastal reef fish metapopulation. Proceedings of the Royal Society of London B: Biological Sciences, 278, 2954–2961. Schiller, A. (2011). Ocean circulation on the North Australian Shelf. Continental Shelf Research, 31, 1087–1095. Schiller, A., Oke, P. R., Brassington, G. B., Entel, M., Fiedler, R., Griffin, D. A., & Mansbridge, J. (2008). Eddy-resolving ocean circulation in the Asian-Australian region inferred from an ocean reanalysis effort. Progress in Oceanography, 76, 334–365.
|
17
Selkoe, K., D’Aloia, C., Crandall, E., Iacchei, M., Liggins, L., Puritz, J. B., . . . Toonen, R. J. (2016). A decade of seascape genetics: Contributions to basic and applied marine connectivity. Marine Ecology Progress Series, 554, 1–19. ~a, Z. A., FinSpalding, M. D., Fox, H. E., Allen, G. R., Davidson, N., Ferdan layson, M., . . . Lourie, S. A. (2007). Marine ecoregions of the world: A bioregionalization of coastal and shelf areas. BioScience, 57, 573–583. Sprintall, J., Wijffels, S., Chereskin, T., & Bray, N. (2002). The JADE and WOCE I10/IR6 throughflow sections in the southeast Indian Ocean. Part 2: Velocity and transports. Deep-Sea Research Part II—Topical Studies in Oceanography, 49, 1363–1389. Stephenson, R. L., Power, M. J., Laffan, S. W., & Suthers, I. M. (2015). Tests of larval retention in a tidally energetic environment reveal the complexity of the spatial structure in herring populations. Fisheries Oceanography, 24, 553–570. Taillebois, L., Barton, D. P., Crook, D. A., Saunders, T., Taylor, J., Hearnden, M., . . . Ovenden, J. R. (2017). Strong population structure deduced from genetics, otolith chemistry and parasite abundances explains vulnerability to localized fishery collapse in a large Sciaenid fish, Protonibea diacanthus. Evolutionary Applications. https://doi.org/ 10.1111/eva.12499 Thackway, R. (Ed.). (1998). Interim marine and coastal regionalisation for Australia: An ecosystem-based classification for marine and coastal environments. Canberra, Australia: Environment Australia, Department of the Environment. Thomas, L., Kendrick, G., Stat, M., Travaille, K. L., Shedrawi, G., & Kennington, W. J. (2014). Population genetic structure of the Pocillopora damicornis morphospecies along Ningaloo Reef, Western Australia. Marine Ecology Progress Series, 513, 111–119. Thomas, L., Kennington, W. J., Evans, R. D., Kendrick, G. A., & Stat, M. (2017). Restricted gene flow and local adaptation highlight the vulnerability of high latitude reefs to rapid environmental change. Global Change Biology, 23, 2197–2205. Travers, M. J., Newman, S. J., & Potter, I. C. (2006). Influence of latitude, water depth, day v. night and wet v. dry periods on the species composition of reef fish communities in tropical Western Australia. Journal of Fish Biology, 69, 987–1017. Travers, M. J., Potter, I. C., Clarke, K. R., & Newman, S. J. (2012). Relationships between latitude and environmental conditions and the species richness, abundance and composition of tropical fish assemblages over soft substrata. Marine Ecology Progress Series, 446, 221– 241. Travers, M. J., Potter, I. C., Clarke, K. R., Newman, S. J., & Hutchins, J. B. (2010). The inshore fish faunas over soft substrates and reefs on the tropical west coast of Australia differ and change with latitude and bioregion. Journal of Biogeography, 37, 148–169. Treml, E. A., Ford, J. R., Black, K. P., & Swearer, S. E. (2015). Identifying the key biophysical drivers, connectivity outcomes, and metapopulation consequences of larval dispersal in the sea. Movement Ecology, 3, 1. Underwood, J. N. (2009). Genetic diversity and divergence among coastal and offshore reefs in a hard coral depend on geographic discontinuity and oceanic currents. Evolutionary Applications, 2, 222–233. Veilleux, H. D., van Herwerden, L., Evans, R. D., Travers, M. J., & Newman, S. J. (2011). Strong genetic subdivision generates high genetic variability among eastern and western Australian populations of Lutjanus carponotatus (Richardson). Fisheries Research, 108, 74–80. Wang, I. J., & Bradburd, G. S. (2014). Isolation by environment. Molecular Ecology, 23, 5649–5662. Whitaker, K. (2006). Genetic evidence for mixed modes of reproduction in the coral Pocillopora damicornis and its effect on population structure. Marine Ecology Progress Series, 306, 115–124. Whitlock, M. C., & Lotterhos, K. E. (2015). Reliable detection of loci responsible for local adaptation: Inference of a null model through trimming the distribution of FST. The American Naturalist, 186, S24–S36.
18
|
Wilkinson, C. (2008). Status of coral reefs of the World: 2008. Townsville, Qld: Global Coral Reef Monitoring Network and Reef and Rainforest Research Centre, 296 pp. Willing, E. M., Dreyer, C., & Van Oosterhout, C. (2012). Estimates of genetic differentiation measured by F ST do not necessarily require large sample sizes when using many SNP markers. PLoS ONE, 7, e42649. Wilson, B. R. (2013). The biogeography of the Australian North West Shelf: Environmental change and life’s response. Burlington, MA: Elsevier. Wolanski, E., & Kingsford, M. J. (2014). Oceanographic and behavioural assumptions in models of the fate of coral and coral reef fish larvae. Journal of the Royal Society Interface, 11, 20140209. Wolanski, E., & Spagnol, S. (2003). Dynamics of the turbidity maximum in King Sound, tropical Western Australia. Estuarine Coastal and Shelf Science, 56, 877–890. Woo, M., Pattiaratchi, C., & Schroeder, W. (2006). Dynamics of the Ningaloo current off point cloates, Western Australia. Marine and Freshwater Research, 57, 291. Wood, M., & Mills, D. (2008). A turning of the tide: Science for decisions in the Kimberley-Browse marine region. Perth, WA: Western Australian Marine Science Institute. Retrieved from www.wamsi.org.au/publi cations-scientific-publications/turning-tide-science-decisions-kimberle y-browse-marine-region
View publication stats
DIBATTISTA
ET AL.
Zhang, X., Oke, P. R., Feng, M., Chamberlain, M., Church, J. A., Monselesan, D., . . . Fiedler, R. (2016). A near-global eddy-resolving OGCM for climate studies. Geoscientific Model Development Discussions. https:// doi.org/10.5194/gmd-2016-17
SUPPORTING INFORMATION Additional Supporting Information may be found online in the supporting information tab for this article.
How to cite this article: DiBattista JD, Travers MJ, Moore GI, et al. Seascape genomics reveals fine-scale patterns of dispersal for a reef fish along the ecologically divergent coast of Northwestern Australia. Mol Ecol. 2017;00:1–18. https://doi.org/10.1111/mec.14352