Skip to main content
NCBI home page
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2026 Jul 2;35(13):e70455. doi: 10.1111/mec.70455

Population Genomics Highlight the Vulnerability of Coral‐Dwelling Gobies to Ecological Losses due to Climatic Disturbances

Catheline Y M Froehlich 1,2,3, Renae L Kirby 1, Samuel Greaves 4, Siobhan J Heatwole 1, O Selma Klanten 5, Martin Hing 1, Marian Y L Wong 1,✉
PMCID: PMC13324404  PMID: 42388061

ABSTRACT

Predicting species resilience to climate change generally focuses on species life history and ecology, overlooking genomic changes following disturbances. After consecutive cyclones and heatwaves, coral‐dwelling Gobiodon gobies have experienced extreme population and group size reductions, and slower recovery rates than their coral hosts. To assess genomic contributions to this vulnerability, we analysed multiple Gobiodon species for population structure (1) among three locations in western Oceania and (2) pre‐ and post‐disturbances at the central location that underwent the same cyclones and heatwaves that reduced population and group sizes. Species compared among locations all exhibited population structure. Post‐disturbances, two species (G. fuscoruber, G . rivulatus ) were extirpated, two species ( G. histrio, G. quinquestrigatus ) exhibited genetic bottlenecks, one species had no change to genetic diversity ( G. brochus ), and another species had consistently low diversity ( G. erythrospilus). Our results highlight that coral‐dwelling gobies face compounding ecological and genomic losses from climatic disturbances, warranting targeted conservation.

Keywords: compounding effects, cyclones, genetic bottlenecks, heatwaves, mutualisms, population structure

1. Introduction

Climate change is driving more frequent and intense disturbances, reshaping ecosystems at an alarming pace (AghaKouchak et al. 2018; Hughes, Anderson, et al. 2018; Seidl et al. 2017). Most ecosystems face one or more disturbances, such as drought, floods, fires, storms, heatwaves, ocean acidification, increasing salinity, species invasions, and pollution (Bergstrom et al. 2021). This can lead to a variety of changes, like disease outbreaks, habitat degradation, behavioural modifications, altered species interactions, population declines, extirpations, and even ecosystem collapse (Sergio et al. 2018; Thakur et al. 2022; Turner 2010). To understand to what extent organisms are facing climate‐related threats, it is crucial to study the impacts of climate change across multiple levels of biological organization (Attrill and Depledge 1997; Borges et al. 2021; Hourigan et al. 1988; Krishnan et al. 2013; Pinna et al. 2023). While ecological and physiological studies show that most species face mounting pressures due to climate change (Hill et al. 2021; Hughes, Kerry, et al. 2018; Johnstone et al. 2016; Pinsky et al. 2022), understanding the extent of gene flow among populations has also been central to predicting their resilience and refining management strategies (Baptista et al. 2025; Cooke et al. 2016; Pauls et al. 2013; Woodall et al. 2011). By investigating the link between population structure and the potential for genetic bottlenecks, we can provide greater insights into species persistence with continued disturbances and guide effective management strategies (Jangjoo et al. 2016; Munguía‐Vega et al. 2015; Ngeve et al. 2023). By combining insights from ecology, population connectivity, and genetic bottlenecks, we can better forecast species' resilience and refine conservation strategies that aim to mitigate the impacts of climate change (Teng et al. 2025).

Reef fishes face major ecological and genomic setbacks immediately following climatic disturbances with slow recovery, and coral‐dwelling fishes are particularly vulnerable (Bellwood et al. 2012; Froehlich et al. 2021; Gerlach et al. 2021; Richardson et al. 2018; Srinivasan et al. 2026). Our recent studies have shown that coral‐dwelling gobies (genus Gobiodon) declined considerably after consecutive cyclones and coral bleaching events, as they suffered extreme population losses and recovered slower than their coral hosts (Froehlich et al. 2021, 2023). Nine months after these consecutive disturbances, only a few adult and juvenile gobies were found. Although the community of coral hosts differed 3 years post‐disturbances, only a few goby species switched hosts and the majority of corals were left uninhabited (Froehlich et al. 2023). When forcibly displaced—as can occur during cyclones—gobies overwhelmingly risked predation to return to their home coral rather than accept a new host (Froehlich et al. 2022). Post‐disturbances, these gobies no longer lived in groups and were primarily solitary (Froehlich et al. 2024), despite the well‐documented benefits of group living for their survival and the need for pair‐living for reproduction (Wong and Buston 2013). Together, these findings paint a troubling picture of severe and multifaceted changes to goby ecology and population dynamics following extreme climatic events. Such negative changes raise crucial questions about whether gobies possess long‐term resilience to ongoing climate change.

Here, we combined two complementary approaches to determine the interplay between the genetic connectivity and diversity of Gobiodon populations and declines in their populations from climatic disturbances. We asked two questions: (1) from a spatial perspective, is there consistent gene flow among different areas within their ranges that could facilitate recovery? And (2) from a disturbance perspective, can genetic bottlenecks be detected only 3 years after consecutive disturbances that triggered ecological and population declines? Since our recent ecological studies showed that goby populations were very slow to recover from climatic disturbances (Froehlich et al. 2021, 2023), we hypothesized that population structure existed among geographic locations (limiting gene flow) and a loss of genetic diversity occurred after climatic disturbances (i.e., evidence of genetic bottlenecks). To address these questions, we sampled goby populations at three locations when in relatively healthy states: Papua New Guinea (Kimbe Bay), northern Great Barrier Reef (Lizard Island), and southern Great Barrier Reef (One Tree Island). We then revisited Lizard Island, the central location, after the disturbances that resulted in extreme population and ecological (i.e., host types and sociality) losses to determine whether there were also losses in genetic diversity and signals of genetic bottlenecks. Together, these approaches revealed how ecological changes and mass mortality following extreme disturbances were coupled with population structure to result in extirpations and genetic bottlenecks for several Gobiodon species. Our study underscores the high vulnerability of coral‐dwelling gobies to climatic disturbances and the need for additional protection.

2. Methods

2.1. Study Locations

Sampling was completed at three different locations in western Oceania (Figure 1). The northernmost location was Kimbe Bay, Papua New Guinea (PNG: −5.42896°, 150.09695°). We visited PNG from 25 Sep‐3 Nov 2018 (21 collection days) when the reef was relatively undisturbed. The central location was Lizard Island, northern Great Barrier Reef in Australia (LI; −14.687264°, 145.447039°). We visited LI from 30 Oct‐5 Nov 2013 (4 collection days) and 4 Feb‐26 Feb 2014 (22 collection days) when the reef was relatively undisturbed, and then from 15 Jan‐5 Mar 2020 (29 collection days) after the reef had been exposed to four consecutive years of extreme disturbances (category 4 cyclones Ita in 2014 and Nathan in 2015, and two mass bleaching events from extended marine heatwaves in 2016 and 2017). The southernmost location was One Tree Island, southern Great Barrier Reef in Australia (OTI; −23.506565°, 152.090954°). We visited OTI from 28 Jan‐19 Feb 2019 (16 collection days) when the reef was relatively undisturbed. The direct distance (point to point) between PNG and LI was ~1140 km, between LI and OTI was ~1200 km, and between PNG and OTI was ~2000 km. All collections were completed via SCUBA diving, which is dependent on safety restrictions with weather and repetitive diving. Each day consisted of 1–4 dives per day and 1.5–3 h per dive.

FIGURE 1.

FIGURE 1

Open in a new tab

Three sampling locations with timelines of sampling events (dashed lines) and climatic disturbances (downward arrows; black spiral symbol for cyclones and white coral symbol for marine heatwaves that caused mass bleaching) for all species studied in western Oceania.

2.2. Sample Collection

At each location, we searched for gobies within their Acropora coral hosts with the aid of an underwater flashlight. We selected goby species based on their abundance at all locations. When encountered, gobies were temporarily removed from the coral using an anaesthetic solution (1:1:1 clove oil: seawater: 70% ethanol) for anaesthesia and a hand wafting motion around the coral to direct them into hand nets. Gobies were identified to species level and a small fin clip was collected from their caudal fin, i.e., tail. Individuals were then returned to their coral hosts once they had recovered from the anaesthesia. Collected fin tissues were stored in 80% ethanol.

2.3. DNA Extraction and Sequencing

From Papua New Guinea, we sequenced 36 G. fuscoruber, 37  G. histrio , 43  G. quinquestrigatus , and 31  G. rivulatus . From Lizard Island pre‐disturbances, we sequenced 47  G. brochus , 43 G. erythrospilus, 38 G. fuscoruber, 50  G. histrio , 53  G. quinquestrigatus , and 47  G. rivulatus . From Lizard Island post‐disturbances, we sequenced 33  G. brochus , 33  G. erythrospilus , 40  G. histrio , and 41 G. quinquestrigatus. From One Tree Island, we sequenced 31 G. fuscoruber, 29  G. quinquestrigatus , and 29  G. rivulatus . We extracted DNA from the fin tissues at the University of Wollongong and Sydney Institute for Marine Science using the SV Genomic DNA Purification System (Promega) following the manufacturer's protocol. We concentrated the DNA at 60°C in a vacuum centrifuge. We sent the samples to the Australian Genome Research Facility (AGRF) for library preparation and sequencing. Samples were prepared following the Peterson et al. (2012) protocol for double digest restriction‐site associated digest (ddRAD) with restriction enzymes EcoRI and MseI, and size selected for a 60 bp region (350–410 bp fragments). The samples were sequenced on an Illumina NextSeq 500 (1 × 150 cycles mid‐output) at AGRF.

2.4. Data Analysis

Initial quality control and demultiplexing were performed by AGRF using STACKS v.2.3d software (Catchen et al. 2011; Catchen et al. 2013). Sequences that had exact barcodes were retained and were then trimmed to 150 bp. Genotyping and Single Nucleotide Polymorphism (SNP) filtering were performed using STACKS v2.54 (Rochette et al. 2019). Parameters used for denovo pipeline were as follows: six for the distance between stacks (−M), six for the distance between catalogue loci (−n), and three for the minimum stack depth of coverage (−m). These values were optimized following Rochette and Catchen (2017). Samples with less than 50% of loci called were excluded from analysis.

The data was divided into datasets for analysing (1) differences among locations in relatively healthy states (PNG 2018, LI 2014, OTI 2019) for each species and (2) differences among pre‐ (2014) and post‐disturbances (2020) (LI only) for each species. With the “populations” component of STACKS, loci were exported by calling more than 80% of samples (−r) that were present in all populations (−p; 3 for among locations, except 2 for G. histrio which was only found at two locations; 2 for pre/post‐disturbances), had less than 0.7 observed heterozygosity (−max‐obs‐het), and had a minimum of 2 minor allele count (−min‐mac). If loci were significantly out of Hardy–Weinberg Equilibrium (HWE) in two or more sampling locations, then they were removed. To properly calculate HWE, the Hardy–Weinberg (−hwe) option was used within populations and the p‐values were then adjusted with the “BY” correction of the “p.adjust” function in R v3.6.3 (RStudio V.1.4.1106) (R Core Team 2020; RStudio Team 2020) to account for False Discovery Rate. To avoid closely linked SNPs, the—write‐single‐snp option was used to export the final set of loci.

Population structure among (1) the different locations (PNG, LI, OTI) and (2) pre‐ and post‐disturbances (LI only) were quantified for analysis of molecular variance (AMOVA) using Arlequin v3.5.2.2 (Excoffier and Lischer 2010). For the AMOVA, 30,000 permutations were used and pairwise FST values were generated to assess hierarchical population structure among populations if the AMOVA was significant. The observed heterozygosity (Ho), expected heterozygozity (He), and inbreeding coefficient (FIS) were calculated as genetic diversity statistics using the “basic.stats” function in the R package hierfstat (Goudet 2005). To determine whether there was asymmetric genetic migration, we used the divMigrate function (Sundqvist et al. 2016) from the diveRsity package (Keenan et al. 2013) in R. We used the “Nm” statistic to estimate number of migrants and 1000 bootstrap samples when calculating significance.

To identify patterns of admixture in the populations with each analysis, STRUCTURE v.2.3.4 (Pritchard et al. 2000) was used. In each analysis, we included 100,000 iterations for burn‐in and 500,000 iterations of Markov‐chain Monte Carlo (MCMC). For each K value (1–9), we used 9–25 independent runs per K depending on the species to limit the number of minor modes created. By applying the △K Evanno method (Evanno et al. 2005), the most likely K value was determined through STRUCTURE HARVESTER (Earl and vonHoldt 2012). The Cluster Markov Packager Across Ks (CLUMPAK) online tool (Kopelman et al. 2015) was used to summarize the results. We produced the STRUCTURE graphs in R. The R package adegenet v1.7–16 (Jombart 2008; Jombart and Ahmed 2011) was used to conduct a Discriminant Analysis of Principal Components (DAPC). Loci were removed if there were more than 20% missing values, and individuals were removed if there were more than 25% missing values. For the remaining data, the missing values were inputted using the mean allele frequency for each locus calculated by the “tab” function. The “find.clusters” function was used to find the number of K values within each analysis regardless of a priori population groups sampled by using a combination of Bayesian Information Criterion (BIC) to identify the most likely K value, as well as a k‐means approach to identify genetic clusters. The “xvalDapc” function was used to cross‐validate results and identify the number of discriminant functions and principal components to retain for DAPC for each species per analysis.

3. Results

3.1. Population Structure Among Locations

After quality control and SNPs calling, 101 individuals from Gobiodon fuscoruber, 103  G. rivulatus , 99  G. quinquestrigatus , and 75  G. histrio were retained, with 5751, 6345, 6527, and 10,006 loci respectively passing filtering parameters for the analysis among locations. Each species had distinct population structure across the locations where it was observed. However, the degree of clustering, heterozygosity, and direction of migration asymmetry among locations varied by species.

Gobiodon fuscoruber exhibited the most distinct population structure among locations (AMOVA: FST = 0.14, p < 0.0001, see Supporting Information for all statistical outputs) compared to all other species (see results below). There was some shared ancestry between LI (central population) and PNG (northern population; pairwise FST = 0.12) as well as LI and OTI (southern population; pairwise FST = 0.11; Figure 2A,B). However, PNG and OTI were highly distinct from each other with little shared ancestry (pairwise FST = 0.22). Regardless of population designation, three clusters were identified (△K = 3, 9 STRUCTURE HARVESTER repetitions per K, no minor modes; Figure S1). The smallest cluster included all PNG individuals whereas OTI and LI each had individuals with ancestry dispersed among the three clusters (Figure 2B). The inbreeding coefficient (FIS) was high at LI, moderate at OTI, and low at PNG (Table 1). Heterozygosity values for both observed and expected were between 0.05 and 0.09 and observed (Ho) was only higher than expected (He) at PNG (Table 1). There was asymmetry in the migration rates from PNG to LI (100%) and PNG to OTI (61%; Figure 2C), yet there was no asymmetry in migration between LI and OTI.

FIGURE 2.

FIGURE 2

Open in a new tab

(A) Bayesian clustering algorithm STRUCTURE v2.3.4 with colour representing different clusters following Pritchard et al. 2000, (B) Principal Components Analysis of population differentiation, and (C) migration rates filtered for asymmetries for 3 sampled locations (purple = PNG, orange = LI, green = OTI) of each Gobiodon species. PNG = Papua New Guinea, LI = Lizard Island, OTI = One Tree Island. **p < 0.0001.

TABLE 1.

Genetic diversity measures for each population of Gobiodon species studied. To investigate inbreeding coefficients, we used the range suggested by Hartl 1999 in which 0–0.05 is low, 0.05–0.15 is moderate, 0.16–0.24 is high, and ≥ 0.25 is very high.

Species Population Inbreeding coefficient FIS Observed heterozygosity Ho Expected heterozygosity He
G. fuscoruber Papua New Guinea 0.048 0.06 0.05
Lizard Island 0.174 0.07 0.09
One Tree Island 0.106 0.06 0.07
G. histrio Papua New Guinea 0.125 0.08 0.09
Lizard Island 0.134 0.09 0.11
One Tree Island NA NA NA
G. quinquestrigatus Papua New Guinea 0.195 0.06 0.07
Lizard Island 0.156 0.05 0.06
One Tree Island 0.088 0.06 0.06
G. rivulatus Papua New Guinea 0.137 0.07 0.08
Lizard Island 0.128 0.06 0.07
One Tree island 0.061 0.04 0.04
Open in a new tab

We found Gobiodon histrio at only two of the three sampling locations (PNG and LI). Gobiodon histrio had a distinct population structure among the two locations (AMOVA: FST = 0.03, p < 0.0001), yet it had the least structure compared to other species. There were several individuals with shared ancestry among locations (Figure 2A,B). Regardless of population designation, three clusters were identified depending on the analysis used (△K = 3, 25 STRUCTURE HARVESTER repetitions per K, no minor modes; Figure S1). However, no cluster fully encapsulated the population from a single location (Figure 2B). The FIS was moderate for both locations and slightly higher at LI (Table 1). Heterozygosity values for both Ho and He were between 0.08 and 0.11, with Ho being lower than He at both locations (Table 1). There was no asymmetry in the migration rates among locations (Figure 2C).

Gobiodon quinquestrigatus had distinct population structure among the three locations (AMOVA: FST = 0.05, p < 0.0001), albeit with mixed ancestry present in all locations (pairwise FST values: 0.06 for PNG‐LI, 0.03 for LI‐OTI, 0.05 for PNG‐OTI; Figure 2A,B). Regardless of population designation, three clusters were identified (△K = 3, 25 STRUCTURE HARVESTER repetitions per K, 2 minor modes; Figures S1 and S2). However, no cluster fully encapsulated the population from a single location (Figure 2B). The FIS was high at PNG, high at LI, and moderate at OTI (Table 1). Heterozygosity values for both Ho and He were between 0.05 and 0.07, and Ho was lower or the same as He at all locations (Table 1). There was no asymmetry in the migration rates among locations (Figure 2C).

Gobiodon rivulatus had distinct population structure among the three locations (AMOVA: FST = 0.09, p < 0.0001). There was shared ancestry between LI and OTI (pairwise FST = 0.04) and some between LI and PNG (pairwise FST = 0.10), but less between PNG and OTI (pairwise FST = 0.13; Figure 2A,B). Regardless of population designation, three clusters were identified (△K = 3, 25 STRUCTURE HARVESTER repetitions per K, 2 minor modes; Figures S1 and S3). A single cluster encapsulated the PNG population and some individuals from LI and OTI, which were both dispersed among several clusters (Figure 2B). The FIS was moderate at all sites although it was two times higher at PNG and LI than OTI (Table 1). Heterozygosity values for both Ho and He were between 0.05 and 0.08 and Ho was lower or the same as He at all locations (Table 1). There is asymmetry in the migration rates from OTI to PNG (42%) and OTI to LI (100%, Figure 2C), but no asymmetry between PNG and LI.

3.2. Pre‐ Versus Post‐Disturbances at Lizard Island

Gobiodon fuscoruber and G. rivulatus were locally extirpated from Lizard Island following the climatic disturbances. For the pre‐ and post‐disturbances subsets from Lizard Island, we retained 62 G. brochus, 70  G. histrio , 59  G. quinquestrigatus , and 62  G. erythrospilus individuals with 9724, 9391, 7656, and 10,760 loci, respectively. Two out of the four species had distinct structure between pre‐ and post‐disturbances, and the degree of heterozygosity varied among species.

Gobiodon brochus had no distinct structure pre‐versus post‐disturbances (AMOVA: FST = 0.0007, p = 0.87). Regardless of population designation, four clusters were identified (△K = 4, 10 STRUCTURE HARVESTER repetitions per K, 0 minor modes; Figure S4). There was a loss of a single cluster made up of 2 individuals from pre‐ to post‐disturbances (Figure 3A,B). The FIS was moderate at both time points (Table 2). Heterozygosity values for both Ho and He were between 0.14 and 0.17, and Ho was lower than He at both time points (Table 2). There was no asymmetry in the migration rates among sampling periods (Figure 3C).

FIGURE 3.

FIGURE 3

Open in a new tab

(A) Bayesian clustering algorithm STRUCTURE v2.3.4 with colour representing different clusters Pritchard et al. 2000, (B) Principal Components Analysis showing population differentiation, and (C) asymmetric gene flow migration for pre‐ (PRE = orange) and post‐disturbances (POST = dark red) at Lizard Island of each Gobiodon species. **p < 0.0001, *p < 0.02.

TABLE 2.

Genetic diversity measures for each Gobiodon species studied pre‐ and post‐disturbances at Lizard Island, northern Great Barrier Reef.

Species Population Inbreeding coefficient Fis Observed heterozygosity Ho Expected heterozygosity He
G. brochus Pre‐disturbances 0.134 0.14 0.17
Post‐disturbances 0.127 0.14 0.17
G. erythrospilus Pre‐disturbances 0.130 0.13 0.16
Post‐disturbances 0.131 0.13 0.15
G. fuscoruber Pre‐disturbances 0.174 0.07 0.09
Post‐disturbances extirpated extirpated extirpated
G. histrio Pre‐disturbances 0.075 0.12 0.13
Post‐disturbances 0.108 0.08 0.10
G. quinquestrigatus Pre‐disturbances 0.183 0.08 0.10
Post‐disturbances 0.262 0.06 0.10
G. rivulatus Pre‐disturbances 0.128 0.06 0.07
Post‐disturbances extirpated extirpated extirpated
Open in a new tab

Gobiodon erythrospilus had barely significant population structure pre‐ and post‐disturbances (AMOVA: FST = 0.005, p = 0.018). However, only one cluster was primarily identified regardless of population designation (Figure S4). There were two clusters that disappeared from pre‐ to post‐disturbances, although those clusters were only found in two individuals pre‐disturbances. The rest of the individuals exhibited the same ancestry at both time periods (Figure 3A,B). The FIS was moderate at both time points (Table 2). Heterozygosity values for both Ho and He were between 0.13 and 0.16, and Ho was lower than He at both time points (Table 2). There was no asymmetry in the migration rates among sampling periods (Figure 3C).

Gobiodon histrio had distinct population structure pre‐ and post‐disturbances (AMOVA: FST = 0.08, p < 0.0001). The population structure homogenized post‐disturbances, with the loss of individuals with partial ancestry from another population (Figure 3A,B). Regardless of population designation, three clusters were identified (△K = 3, 10 STRUCTURE HARVESTER repetitions per K, 0 minor modes; Figure S4). The FIS was moderate at both time points but was 44% higher post‐disturbances than pre‐disturbances (Table 2). Heterozygosity values for both Ho and He were between 0.08 and 0.13, and Ho was lower than He at both time points (Table 2). Although there was asymmetric migration identified from post‐disturbances to pre‐disturbances, the context instead implies the opposite direction (Sundqvist et al. 2016) with a founder effect due to the disturbances (Figure 3C).

Gobiodon quinquestrigatus had distinct population structure pre‐ and post‐disturbances (AMOVA: FST = 0.08, p < 0.0001). This species homogenized post‐disturbances, with almost all individuals exhibiting full ancestry from only 2 populations (Figure 3A,B). Regardless of population designation, four clusters were identified (△K = 4, 10 STRUCTURE HARVESTER repetitions per K, 0 minor modes; Figure S4). Pre‐disturbances, individuals were found in each of the four clusters whereas one cluster was lost post‐disturbances. Most individuals post‐disturbances were from one of two clusters (dark blue or cyan) with a select few exhibiting a third cluster (red with dark blue) (Figure 3B). The FIS was high pre‐disturbances and 43% higher post‐disturbances sitting at a very high value (0.26), which was the highest in the whole study (Table 2). Heterozygosity values for both Ho and He were between 0.06 and 0.10 and Ho was lower than He at both time points (Table 2). Although there was asymmetric migration identified from post‐disturbances to pre‐disturbances, the context instead implies the opposite direction (Sundqvist et al. 2016) with a founder effect due to the disturbances (Figure 3C).

4. Discussion

We quantified population structure and signals of genetic bottlenecks that resulted from the severe declines in Gobiodon populations and changes in their social structures and host‐use patterns following repeated cyclones and heatwave‐induced bleaching events (Froehlich et al. 2021, 2023, 2024). While all species showed some degree of population structure across the study locations, they differed in genetic diversity and directional gene flow. For example, G. fuscoruber exhibited migration patterns from north to south and the opposite was found for G. rivulatus , while G. histrio and G. quinquestrigatus had no clear pattern. Following disturbances at the central location, G. fuscoruber and G. rivulatus were extirpated and we detected genetic bottlenecks in G. quinquestrigatus and G. histrio . Our study highlights that population structure and genetic bottlenecks increase the vulnerability of these fishes to population collapse during climatic disturbances, alongside other ecological losses observed for these fishes. Given these results, we recommend expanding conservation efforts to include more Gobiodon species, which are facing multiple climate‐driven threats.

Population structure was evident for each Gobiodon species across all locations in relatively healthy states, suggesting these populations may be especially vulnerable to climatic disturbances. Most species exhibited low to moderate levels of genetic isolation, as indicated by FST values (0.03–0.13), while G. fuscoruber showed moderate to high FST values (0.11–0.22) (Hall 2022). The distances between our study locations (1000–2000 km) are below typical thresholds for genetic isolation at sea (Cooke et al. 2016). For example, populations of another coral reef fish, Pomancentrus amboinensis, were found to be genetically similar in Papua New Guinea and Lizard Island (Jones et al. 2010). However, G. fuscoruber's minimum FST of 0.15 (Frankham et al. 2002) indicates that its populations are instead genetically isolated between the same locations. Such populations may be either predominantly recruitment‐limited or post‐recruitment regulated (Jones et al. 2009). High self‐recruitment rates are observed in many reef fishes with short larval durations (D'Aloia et al. 2018; Jones et al. 1999, 2005; Saenz‐Agudelo et al. 2012; Selwyn et al. 2016; Swearer et al. 2002), like the coral‐dwelling goby Paragobiodon xanthosoma (Rueger, Buston, et al. 2021), which suggests a lack of recruitment‐limitation. As Gobiodon species have very similar biology, sociality, and ecology to P. xanthosoma (Rueger, Branconi, et al. 2021), and exhibit a similarly short pelagic larval duration (19.6 days for Gobiodon; Pereira et al. 2015), this suggests that they are not recruitment limited. Instead, they are likely regulated by post‐recruitment limitations under normal conditions, which reduces the risks of extirpation but increases recovery time (Jones et al. 2009). Nonetheless, two species (G. fuscoruber and G. rivulatus ) experienced local extirpation in the disturbed location. Following these severe climatic events, which dramatically reduced adult populations (Froehlich et al. 2021), Gobiodon species may have become recruitment‐limited as well.

Inbreeding levels further highlight species‐specific vulnerabilities. The statistical significance of inbreeding coefficients in our study should be further validated for some Gobiodon species, as they were below 10,000 SNPs for full validation for some species (two had above 10,000 SNPs and all others had between 5751 and 9724 SNPs). Nonetheless, inbreeding coefficients ranged from low to high (Hartl 1999) at different locations in relatively heathly states depending on the species (e.g., G. fuscoruber, G. quinquestrigatus ), or moderate throughout ( G. rivulatus ). At the central location (Lizard Island), inbreeding coefficients was moderate for G. histrio and G. rivulatus and high for G. fuscoruber and G. quinquestrigatus . Such inbreeding patterns are unsurprising since Gobiodon gobies, like other habitat specialists, tend to settle near natal reefs with siblings (Buston et al. 2009; Jones et al. 2005). Limited genetic diversity may occur from such settlement behaviour, thus increasing their susceptibility to extreme disturbances (Jones et al. 2009). Accordingly, low to high population structure coupled with moderate to high inbreeding patterns for all species limit gene flow and genetic diversity, which likely contribute to Gobiodon's slow recovery rates from impacts of climatic disturbances (Froehlich et al. 2021).

Ocean currents play crucial roles in larval dispersal and the recovery potential of marine species (Gouvêa et al. 2023; Snead et al. 2023; Wilson et al. 2016), but they do not fully explain the patterns of directional genetic migration in Gobiodon species. Based on prevailing ocean currents in western Oceania (Ridgway and Hill 2009), the central populations in our study (Lizard Island) would be expected to act as a source for the northern populations (Papua New Guinea). Additionally, the presence of competing currents between the southern (One Tree Island) and central populations (East Australian and Hirri Currents; Ridgway and Hill 2009) further complicates the predictions of source‐sink dynamics between these locations. Of the four species studied among locations, only two species (G. fuscoruber and G. rivulatus ) exhibited directional genetic migration, and both were extirpated from the central location following climatic disturbances. Interestingly, their migration patterns were in opposite directions: G. fuscoruber showed directional genetic migration from the northern population to the central and southern populations, whereas G. rivulatus exhibited the opposite. It is possible that we did not sample intermediate populations found between the study locations, which may help explain the contrasting patterns among species. Ocean currents themselves can also be impacted by climatic disturbances via directional and speed changes (Hasselmann 1991; Wilson et al. 2016). Such changes to ocean circulations may alter expected directions of gene flow and larval dispersal (Wilson et al. 2016). Larval dispersal can also be impacted directly by increasing water temperatures, whereby larval dispersal is shortened, resulting in larvae settling closer to natal reefs (O'Connor et al. 2007). Gobiodon populations may thus be recovering slower due to climate‐driven shifts in the direction and distance of larval dispersal from other populations. However, ocean currents alone cannot explain the overall pattern of gene flow in all Gobiodon species, especially given that the central populations would be expected to have been a source population. Although the dispersal and genetic structure of many marine fishes is often driven by ocean currents (Baltazar‐Soares et al. 2014; Munguia‐Vega et al. 2017; Snead et al. 2023), not all tropical species exhibit such patterns (Shulman and Bermingham 1995). Since two Gobiodon species exhibited contrasting patterns despite their overlapping ranges, their genetic migration is likely influenced by other ecological factors, like their sociality (Froehlich et al. 2024; Rueger, Branconi, et al. 2021) and the distribution of their preferred hosts (Froehlich et al. 2023). Overall, our findings indicate that ocean currents do not always influence the population structure of all species within a single genus, which has flow‐on implications for their resilience to climatic disturbances.

When investigating whether climatic disturbances at the central location affected genetic diversity, not all species responded in the same manner. There were few changes to the population structure of Gobiodon brochus and G. erythrospilus following the disturbances. The latter species had little genetic diversity at both time points, which aligns with its already vulnerable status on the IUCN red list (Larson 2020). Other taxa that underwent mass mortality following storms and heatwaves had little changes to their genetic diversity, such as the submerged aquatic plant Vallisneria americana and the red gorgonian Paramuricea clavata (Ngeve et al. 2023; Pilczynska et al. 2016). For the alpine butterfly Parnassius smintheus , high connectivity pre‐disturbances resulted in little change to their genetic diversity after mass mortality (Jangjoo et al. 2016). However, not all species were resilient. Two species in our study, G. fuscoruber and G. rivulatus , were extirpated. The other two species, G. histrio and G. quinquestrigatus , showed a clear loss of genetic diversity, increased inbreeding, and marked directional genetic migration post‐disturbances. In both species, post‐disturbance populations were more genetically uniform (i.e., less diverse), and at least one ancestral cluster that was initially present pre‐disturbances disappeared entirely. There were characteristics similar to a founder effect (Sundqvist et al. 2016), whereby a few individuals pre‐disturbances gave rise to the post‐disturbance populations. This was evident from the reduction in the number of ancestral clusters, the increase in assignment of individuals to a single cluster, and the directional genetic migration observed. Broader patterns in marine ecosystems showed similar results in which population structure leads to lower genetic diversity post‐disturbances and the potential for genetic bottlenecks (Noreen et al. 2009; Selkoe et al. 2016). Similarly, seaweed populations underwent up to 80–100% loss in genetic diversity following marine heatwaves (Gurgel et al. 2020). Taken together, our results suggest that genetic bottlenecks pose a real threat to several Gobiodon species. Only 4 of 28 Gobiodon species are listed as vulnerable on the IUCN red list, and all others are of least concern or data deficient. Conservation strategies aimed at protecting coral reef ecosystems should include coral‐dwelling fishes and incorporate species‐specific management plans at a population level. As climatic disturbances continue to increase in frequency and severity, genetic bottlenecks may become even more common, placing many more reef‐associated species at risk.

Our study is one of the rare cases in which ecological and genomic axes have been combine to study impacts of the same extreme climatic disturbances (Froehlich et al. 2021, 2023, 2024). Other coral reef fish studies have primarily measured the ecological impacts of climate change (e.g., Fabricius et al. 2008; Richardson et al. 2018; Srinivasan et al. 2026), with only a few genomic studies (e.g., Gerlach et al. 2021). Here, by linking genomic, life history, and ecological outcomes of the same extreme and consecutive disturbances on Gobiodon species, a clearer picture emerges as to why some species recover more slowly than others. For example, the two species that were extirpated post‐disturbances (G. fuscoruber and G. rivulatus ) had high FST values among populations. Pre‐disturbances, both species lived primarily in groups rather than pairs, unlike other goby species in this study (Froehlich et al. 2024; Hing et al. 2019). Post‐disturbances, the absence of established social groups for G. fuscoruber and G. rivulatus likely reduced successful juvenile settlement, with these impacts compounded by high inbreeding rates and population structure. All Gobiodon species—regardless of whether they typically lived in pairs or groups before the disturbances—were mostly juveniles living alone and only a few adult pairs post‐disturbances (Froehlich et al. 2024), likely reducing their reproductive success. Additionally, a species' growth potential also shapes resilience. For example, G. rivulatus is one of the smallest species in the genus (Hing et al. 2018, 2019), making it more vulnerable to predation, especially since smaller fish are more frequently targeted post‐disturbances (Brandl et al. 2018; Goatley and Bellwood 2016). Since Gobiodon species fully depend on live coral, how well their preferred coral hosts recover can directly affect their own recovery. Gobiodon species that were extirpated post‐disturbances (G. fuscoruber and G. rivulatus ) primarily inhabited corals that became rare post‐disturbances ( Acropora millepora and A. gemmifera , respectively) (Froehlich et al. 2021, 2023). However, Gobiodon species that persisted either were able to change the coral host they inhabited or benefited from their preferred host recovering quickly (Froehlich et al. 2023). Yet even Gobiodon species that survived were slower to recover than their coral hosts and several exhibited genetic bottlenecks, which further reduces their ability to respond to future disturbances. Several studies have shown that coral reef fishes need considerable recovery time from cyclones and heatwaves (e.g., Fabricius et al. 2008; Gerlach et al. 2021; Halford et al. 2004; Richardson et al. 2018; Srinivasan et al. 2026), and we show that coral‐dwelling fishes are a highly vulnerable taxon. The compounding ecological and genomic pressures on gobies highlight the urgent need for conservation efforts that specifically target coral‐dwelling fishes as they play important roles for coral health and resilience to bleaching.

Author Contributions

C.Y.M.F., O.S.K., M.W. designed the study; C.Y.M.F., M.H., R.L.K., S.H., O.S.K., M.W. collected the data; C.Y.M.F., R.L.K. analysed the data; S.G. provided and supervised analytical methods; C.Y.M.F. wrote the draft; all authors contributed to revisions.

Funding

Field and laboratory work were funded by the Hermon Slade Foundation to MW and M. Dowton, Sea World Research and Rescue Foundation to MW and T. Rueger, a target grant from the Faculty of Science, Medicine and Health at the University of Wollongong (UOW) to MW, funding initiatives from the Centre for Sustainable Ecosystem Solutions to MW, CF and SH, the Zoltan Florian Marine Biology Fellowship as part of the Lizard Island Doctoral Fellowship program of the Australian Museum to CF, and an Australian Coral Reef Society Fellowship to CF. For student support, CF was supported under a University Postgraduate Award, RK and SH by an Australian Government Research Training Program Scholarship, both administered through UOW. SG is supported by the National Science Foundation Graduate Research Fellowship Program under Grant No. DGE‐2035702.

Ethics Statement

The study was approved under animal ethics protocols AE1404 and AE1725 at the University of Wollongong. This research was conducted under research permits issued by the Great Barrier Reef Marine Park Authority (G13/36197.1, G15/37533.1, and G18/41020.1) and Papua New Guinea Research Visa Permit AA654347.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1: (a) DeltaK graphs from HARVESTER determining the best K‐value, (b) Discriminant Analysis Principal Components and (c) eigenvalues for each axis for species compared among locations (purple = Papua New Guinea, orange = Lizard Island, green = One Tree Island); asterisks mean significant differences observed among groups.

Figure S2: Two minor modes were created for Gobiodon quinquestrigatus with a K‐value of three.

Figure S3: Two minor modes were created for Gobiodon rivulatus with a K‐value of three.

Figure S4: (a) DeltaK graphs from HARVESTER determining the best K‐value, (b) Discriminant Analysis Principal Components and (c) eigenvalues for each axis for species compared pre‐disturbances (orange) and post‐disturbances (dark red) at Lizard Island; ** means p < 0.001; * means p < 0.02.

Table S1: Amova results showing the population differentiation among three locations for Gobiodon species. Sample sizes included as denoted by n. Number of variable loci that passed filtering parameters denoted by no. loci. PNG = Papua New Guinea, LI = Lizard Island, OTI = One Tree Island.

Table S2: Pairwise FST values and corresponding p‐values between locations for Gobiodon fuscoruber calculated through Arlequin software. Below the grey line are the pairwise FST values between the given locations, while above the grey line are the corresponding p‐values for the FST statistic. PNG = Papua New Guinea, LI = Lizard Island, OTI = One Tree Island.

Table S3: Pairwise FST values and corresponding p‐values between sampled locations for Gobiodon quinquestrigatus calculated through Arlequin software. Below the grey line are the pairwise FST values between the given populations, while above the grey line are the corresponding p‐values for the FST statistic.

Table S4: Pairwise FST values and corresponding p‐values between locations for Gobiodon rivulatus calculated through Arlequin software. Below the grey line are the pairwise FST values between the given locations, while above the grey line are the corresponding p‐values for the FST statistic.

Table S5: Amova results of population differentiation pre‐ and post‐disturbances at Lizard Island for Gobiodon species. Number of variable loci that passed filtering parameters denoted by no. loci.

MEC-35-e70455-s001.pdf (472.5KB, pdf)

Acknowledgements

We pay respect to the traditional owners past and present of the Kilu and Tamare communities of Kimbe Bay in Papua New Guinea, the Dingaal Aboriginal Traditional Owners of Lizard Island in Australia, and the Gurang, Gooreng Gooreng, Bailai and Bunda Traditional Owners of One Tree Island in Australia. We greatly appreciate our field helpers: Kylie Brown, Karen Hing, Al Alder, Theresa Rueger, Rituraj Sharma, Jemma Okapie Smith, Nelson and Jerry Sikatura. We are thankful to Mark Dowton and Tracey Gibson for guidance regarding laboratory practices. We are grateful to the facilities and management of field stations at the Mahonia Na Dari Research and Conservation Centre of Kimbe Bay, Lizard Island Research Station, and One Tree Island Research Station, with particular appreciation to Ainsley and Paul Carin, Anne Hoggett, Lyle Vail, Somei Jonda, Peter and Jane Miller for their assistance with our field trips. We would like to thank Courtney Hildebrandt for her Gobiodon artwork and staff at the Sydney Institute of Marine Science for help while carrying out laboratory work in the facility. Open access publishing facilitated by University of Wollongong, as part of the Wiley ‐ University of Wollongong agreement via the Council of Australasian University Librarians.

Data Availability Statement

The data is available on NCBI project: PRJNA1309028.

References

  1. AghaKouchak, A. , Huning L. S., Chiang F., et al. 2018. “How do Natural Hazards Cascade to Cause Disasters?” Nature 561: 458–460. [DOI] [PubMed] [Google Scholar]
  2. Attrill, M. J. , and Depledge M. H.. 1997. “Community and Population Indicators of Ecosystem Health: Targeting Links Between Levels of Biological Organisation.” Aquatic Toxicology 38: 183–197. [Google Scholar]
  3. Baltazar‐Soares, M. , Biastoch A., Harrod C., et al. 2014. “Recruitment Collapse and Population Structure of the European Eel Shaped by Local Ocean Current Dynamics.” Current Biology 24: 104–108. [DOI] [PubMed] [Google Scholar]
  4. Baptista, M. S. P. , Keuroghlian A., Tambosi L. R., et al. 2025. “Agriculturally Developed Areas Reduce Genetic Connectivity for a Keystone Neotropical Ungulate.” Biotropica 57: e13418. [Google Scholar]
  5. Bellwood, D. R. , Baird A. H., Depczynski M., et al. 2012. “Coral Recovery May Not Herald the Return of Fishes on Damaged Coral Reefs.” Oecologia 170: 567–573. [DOI] [PubMed] [Google Scholar]
  6. Bergstrom, D. M. , Wienecke B. C., Hoff J., et al. 2021. “Combating Ecosystem Collapse From the Tropics to the Antarctic.” Global Change Biology 27, no. 9: 1692–1703. [DOI] [PubMed] [Google Scholar]
  7. Borges, F. L. G. , Oliveira M. D. R., De Almeida T. C., Majer J. D., and Garcia L. C.. 2021. “Terrestrial Invertebrates as Bioindicators in Restoration Ecology: A Global Bibliometric Survey.” Ecological Indicators 125: 107458. [Google Scholar]
  8. Brandl, S. J. , Goatley C. H. R., Bellwood D. R., and Tornabene L.. 2018. “The Hidden Half: Ecology and Evolution of Cryptobenthic Fishes on Coral Reefs.” Biological Reviews 93: 1846–1873. [DOI] [PubMed] [Google Scholar]
  9. Buston, P. M. , Fauvelot C., Wong M. Y. L., and Planes S.. 2009. “Genetic Relatedness in Groups of the Humbug Damselfish Dascyllus aruanus : Small, Similar‐Sized Individuals May Be Close Kin.” Molecular Ecology 18: 4707–4715. [DOI] [PubMed] [Google Scholar]
  10. Catchen, J. , Hohenlohe P. A., Bassham S., Amores A., and Cresko W. A.. 2013. “Stacks: An Analysis Tool Set for Population Genomics.” Molecular Ecology 22: 3124–3140. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Catchen, J. M. , Amores A., Hohenlohe P., Cresko W., and Postlethwait J. H.. 2011. “Stacks: Building and Genotyping Loci De Novo From Short‐Read Sequences.” G3: Genes, Genomes, Genetics 1: 171–182. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Cooke, G. M. , Schlub T. E., Sherwin W. B., and Ord T. J.. 2016. “Understanding the Spatial Scale of Genetic Connectivity at Sea: Unique Insights From a Land Fish and a Meta‐Analysis.” PLoS One 11: e0150991. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. D'Aloia, C. C. , Xuereb A., Fortin M.‐J., Bogdanowicz S. M., and Buston P. M.. 2018. “Limited Dispersal Explains the Spatial Distribution of Siblings in a Reef Fish Population.” Marine Ecology Progress Series 607: 143–154. [Google Scholar]
  14. Earl, D. A. , and vonHoldt B. M.. 2012. “STRUCTURE HARVESTER: A Website and Program for Visualizing STRUCTURE Output and Implementing the Evanno Method.” Conservation Genetics Resources 4: 359–361. [Google Scholar]
  15. Evanno, G. , Regnaut S., and Goudet J.. 2005. “Detecting the Number of Clusters of Individuals Using the Software Structure: A Simulation Study.” Molecular Ecology 14: 2611–2620. [DOI] [PubMed] [Google Scholar]
  16. Excoffier, L. , and Lischer H. E. L.. 2010. “Arlequin Suite Ver 3.5: A New Series of Programs to Perform Population Genetics Analyses Under Linux and Windows.” Molecular Ecology Resources 10: 564–567. [DOI] [PubMed] [Google Scholar]
  17. Fabricius, K. E. , De'ath G., Puotinen M. L., Done T., Cooper T. F., and Burgess S. C.. 2008. “Disturbance Gradients on Inshore and Offshore Coral Reefs Caused by a Severe Tropical Cyclone.” Limnology and Oceanography 53: 690–704. [Google Scholar]
  18. Frankham, R. , Ballou J. D., Briscoe D. A., and McInnes K. H.. 2002. Introduction to Conservation Genetics. Cambridge University Press. [Google Scholar]
  19. Froehlich, C. Y. M. , Heatwole S. J., Klanten O. S., and Wong M. Y. L.. 2022. “Habitat Health, Size and Saturation Do Not Alter Movement Decisions in a Social Coral Reef Fish.” Animal Behaviour 191: 125–133. [Google Scholar]
  20. Froehlich, C. Y. M. , Heatwole S. J., Selma Klanten O., et al. 2024. “Multi‐Level Framework to Assess Social Variation in Response to Ecological and Social Factors: Modeled With Coral Gobies.” Oikos 2024: e10669. [Google Scholar]
  21. Froehlich, C. Y. M. , Klanten O. S., Hing M. L., Dowton M., and Wong M. Y. L.. 2021. “Uneven Declines Between Corals and Cryptobenthic Fish Symbionts From Multiple Disturbances.” Scientific Reports 11: 16420. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Froehlich, C. Y. M. , Klanten O. S., Hing M. L., Dowton M., and Wong M. Y. L.. 2023. “Delayed Recovery and Host Specialization May Spell Disaster for Coral‐Fish Mutualism.” Ecology and Evolution 13: e10209. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Gerlach, G. , Kraemer P., Weist P., Eickelmann L., and Kingsford M. J.. 2021. “Impact of Cyclones on Hard Coral and Metapopulation Structure, Connectivity and Genetic Diversity of Coral Reef Fish.” Coral Reefs 40: 999–1011. [Google Scholar]
  24. Goatley, C. H. R. , and Bellwood D. R.. 2016. “Body Size and Mortality Rates in Coral Reef Fishes: A Three‐Phase Relationship.” Proceedings of the Royal Society B: Biological Sciences 283: 20161858. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Goudet, J. 2005. “Hierfstat, a Package for r to Compute and Test Hierarchical F‐Statistics.” Molecular Ecology Notes 5: 184–186. [Google Scholar]
  26. Gouvêa, L. P. , Fragkopoulou E., Cavanaugh K., et al. 2023. “Oceanographic Connectivity Explains the Intra‐Specific Diversity of Mangrove Forests at Global Scales.” Proceedings of the National Academy of Sciences 120: e2209637120. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Gurgel, C. F. D. , Camacho O., Minne A. J. P., Wernberg T., and Coleman M. A.. 2020. “Marine Heatwave Drives Cryptic Loss of Genetic Diversity in Underwater Forests.” Current Biology 30: 1199–1206.e2. [DOI] [PubMed] [Google Scholar]
  28. Halford, A. , Cheal A. J., Ryan D., et al. 2004. “Resilience to Large‐Scale Disturbance in Coral and Fish Assemblages on the Great Barrier Reef.” Ecology 85: 7. [Google Scholar]
  29. Hall, S. J. G. 2022. “Genetic Differentiation Among Livestock Breeds—Values for Fst.” Animals 12: 1115. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Hartl, D. L. 1999. A Primer of Population Genetics. 3rd ed. Sinauer Associates. [Google Scholar]
  31. Hasselmann, K. 1991. “Ocean Circulation and Climate Change.” Tellus Series A: Dynamic Meteorology and Oceanography 43: 82–103. [Google Scholar]
  32. Hill, G. M. , Kawahara A. Y., Daniels J. C., Bateman C. C., and Scheffers B. R.. 2021. “Climate Change Effects on Animal Ecology: Butterflies and Moths as a Case Study.” Biological Reviews 96: 2113–2126. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Hing, M. L. , Klanten O. S., Dowton M., Brown K. R., and Wong M. Y. L.. 2018. “Repeated Cyclone Events Reveal Potential Causes of Sociality in Coral‐Dwelling Gobiodon Fishes.” PLoS One 13: e0202407. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Hing, M. L. , Klanten O. S., Wong M. Y. L., and Dowton M.. 2019. “Drivers of Sociality in Gobiodon Fishes: An Assessment of Phylogeny, Ecology and Life‐History.” Molecular Phylogenetics and Evolution 137: 263–273. [DOI] [PubMed] [Google Scholar]
  35. Hourigan, T. F. , Timothy C. T., and Reese E. S.. 1988. “Coral Reef Fishes as Indicators of Environmental Stress in Coral Reefs.” In Marine Organisms as Indicators, edited by Soule D. F. and Kleppel G. S., 107–135. Springer. [Google Scholar]
  36. Hughes, T. P. , Anderson K. D., Connolly S. R., et al. 2018. “Spatial and Temporal Patterns of Mass Bleaching of Corals in the Anthropocene.” Science 359: 80–83. [DOI] [PubMed] [Google Scholar]
  37. Hughes, T. P. , Kerry J. T., Baird A. H., et al. 2018. “Global Warming Transforms Coral Reef Assemblages.” Nature 556: 492–496. [DOI] [PubMed] [Google Scholar]
  38. Jangjoo, M. , Matter S. F., Roland J., and Keyghobadi N.. 2016. “Connectivity Rescues Genetic Diversity After a Demographic Bottleneck in a Butterfly Population Network.” Proceedings of the National Academy of Sciences 113: 10914–10919. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Johnstone, J. F. , Allen C. D., Franklin J. F., et al. 2016. “Changing Disturbance Regimes, Ecological Memory, and Forest Resilience.” Frontiers in Ecology and the Environment 14: 369–378. [Google Scholar]
  40. Jombart, T. 2008. “Adegenet: A R Package for the Multivariate Analysis of Genetic Markers.” Bioinformatics 24: 1403–1405. [DOI] [PubMed] [Google Scholar]
  41. Jombart, T. , and Ahmed I.. 2011. “Adegenet 1.3‐1: New Tools for the Analysis of Genome‐Wide SNP Data.” Bioinformatics 27: 3070–3071. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Jones, D. B. , Jerry D. R., McCormick M. I., and Bay L. K.. 2010. “The Population Genetic Structure of a Common Tropical Damselfish on the Great Barrier Reef and Eastern Papua New Guinea.” Coral Reefs 29: 455–467. [Google Scholar]
  43. Jones, G. P. , Almany G. R., Russ G. R., et al. 2009. “Larval Retention and Connectivity Among Populations of Corals and Reef Fishes: History, Advances and Challenges.” Coral Reefs 28: 307–325. [Google Scholar]
  44. Jones, G. P. , Milicich M. J., Emslie M. J., and Lunow C.. 1999. “Self‐Recruitment in a Coral Reef Fish Population.” Nature 402: 802–804. [Google Scholar]
  45. Jones, G. P. , Planes S., and Thorrold S. R.. 2005. “Coral Reef Fish Larvae Settle Close to Home.” Current Biology 15: 1314–1318. [DOI] [PubMed] [Google Scholar]
  46. Keenan, K. , McGinnity P., Cross T. F., Crozier W. W., and Prodöhl P. A.. 2013. “diveRsity: An R Package for the Estimation and Exploration of Population Genetics Parameters and Their Associated Errors.” Methods in Ecology and Evolution 4: 782–788. [Google Scholar]
  47. Kopelman, N. M. , Mayzel J., Jakobsson M., Rosenberg N. A., and Mayrose I.. 2015. “Clumpak: A Program for Identifying Clustering Modes and Packaging Population Structure Inferences Across K.” Molecular Ecology Resources 15: 1179–1191. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Krishnan, P. , George G., Immanuel T., Bitopan‐Malakar M., and Anand A.. 2013. “Studies on the Recovery of Bleached Corals in Andaman: Fishes as Indicators of Reef Health.” In Ecology and Conservation of Tropical Marine Faunal Communities, edited by Venkataraman K., Sivaperuman C., and Raghunathan C., 395–407. Springer. [Google Scholar]
  49. Larson, H. K. 2020. “ Gobiodon erythrospilus .” IUCN Red List of Threatened Species 2020: e.T68330883A68333739. [Google Scholar]
  50. Munguia‐Vega, A. , Marinone S. G., Paz‐Garcia D. A., et al. 2017. “Anisotropic Larval Connectivity and Metapopulation Structure Driven by Directional Oceanic Currents in a Marine Fish Targeted by Small‐Scale Fisheries.” Marine Biology 165: 16. [Google Scholar]
  51. Munguía‐Vega, A. , Sáenz‐Arroyo A., Greenley A. P., et al. 2015. “Marine Reserves Help Preserve Genetic Diversity After Impacts Derived From Climate Variability: Lessons From the Pink Abalone in Baja California.” Global Ecology and Conservation 4: 264–276. [Google Scholar]
  52. Ngeve, M. N. , Engelhardt K. A. M., Gray M., and Neel M. C.. 2023. “Calm After the Storm? Similar Patterns of Genetic Variation in a Riverine Foundation Species Before and After Severe Disturbance.” Ecology and Evolution 13: e10670. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Noreen, A. M. E. , Harrison P. L., and Van Oppen M. J. H.. 2009. “Genetic Diversity and Connectivity in a Brooding Reef Coral at the Limit of Its Distribution.” Proceedings of the Royal Society B: Biological Sciences 276: 3927–3935. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. O'Connor, M. I. , Bruno J. F., Gaines S. D., et al. 2007. “Temperature Control of Larval Dispersal and the Implications for Marine Ecology, Evolution, and Conservation.” Proceedings of the National Academy of Sciences 104: 1266–1271. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Pauls, S. U. , Nowak C., Bálint M., and Pfenninger M.. 2013. “The Impact of Global Climate Change on Genetic Diversity Within Populations and Species.” Molecular Ecology 22: 925–946. [DOI] [PubMed] [Google Scholar]
  56. Pereira, P. H. C. , Munday P. L., and Jones G. P.. 2015. “Competitive Mechanisms Change With Ontogeny in Coral‐Dwelling Gobies.” Ecology 96: 3090–3101. [DOI] [PubMed] [Google Scholar]
  57. Peterson, B. K. , Weber J. N., Kay E. H., Fisher H. S., and Hoekstra H. E.. 2012. “Double Digest RADseq: An Inexpensive Method for De Novo SNP Discovery and Genotyping in Model and Non‐Model Species.” PLoS One 7: e37135. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Pilczynska, J. , Cocito S., Boavida J., Serrão E., and Queiroga H.. 2016. “Genetic Diversity and Local Connectivity in the Mediterranean Red Gorgonian Coral After Mass Mortality Events.” PLoS One 11: e0150590. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Pinna, M. , Zangaro F., Saccomanno B., et al. 2023. “An Overview of Ecological Indicators of Fish to Evaluate the Anthropogenic Pressures in Aquatic Ecosystems: From Traditional to Innovative DNA‐Based Approaches.” Water 15: 949. [Google Scholar]
  60. Pinsky, M. L. , Comte L., and Sax D. F.. 2022. “Unifying Climate Change Biology Across Realms and Taxa.” Trends in Ecology & Evolution 37: 672–682. [DOI] [PubMed] [Google Scholar]
  61. Pritchard, J. K. , Stephens M., and Donnelly P.. 2000. “Inference of Population Structure Using Multilocus Genotype Data.” Genetics 155: 945–959. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. R Core Team . 2020. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing. https://www.R‐project.org. [Google Scholar]
  63. Richardson, L. E. , Graham N. A. J., Pratchett M. S., Eurich J. G., and Hoey A. S.. 2018. “Mass Coral Bleaching Causes Biotic Homogenization of Reef Fish Assemblages.” Global Change Biology 24: 3117–3129. [DOI] [PubMed] [Google Scholar]
  64. Ridgway, K. , and Hill K.. 2009. “The East Australian Current (EAC).” In A Marine Climate Change Impacts and Adaptation Report Card for Australia 2009, 1–16. NCCARF. [Google Scholar]
  65. Rochette, N. C. , and Catchen J. M.. 2017. “Deriving Genotypes From RAD‐Seq Short‐Read Data Using Stacks.” Nature Protocols 12: 2640–2659. [DOI] [PubMed] [Google Scholar]
  66. Rochette, N. C. , Rivera‐Colón A. G., and Catchen J. M.. 2019. “Stacks 2: Analytical Methods for Paired‐End Sequencing Improve RADseq‐Based Population Genomics.” Molecular Ecology 28: 4737–4754. [DOI] [PubMed] [Google Scholar]
  67. RStudio Team . 2020. RStudio: Integrated Development Environment for R. RStudio PBC. http://www.rstudio.com/. [Google Scholar]
  68. Rueger, T. , Branconi R., Froehlich C. Y. M., Heatwole S. J., Wong M. Y. L., and Buston P. M.. 2021. “The Next Frontier in Understanding the Evolution of Coral Reef Fish Societies.” Frontiers in Marine Science 8: 665780. [Google Scholar]
  69. Rueger, T. , Buston P. M., Bogdanowicz S. M., and Wong M. Y.. 2021. “Genetic Relatedness in Social Groups of the Emerald Coral Goby Paragobiodon xanthosoma Creates Potential for Weak Kin Selection.” Molecular Ecology 30: 1311–1321. [DOI] [PubMed] [Google Scholar]
  70. Saenz‐Agudelo, P. , Jones G. P., Thorrold S. R., and Planes S.. 2012. “Patterns and Persistence of Larval Retention and Connectivity in a Marine Fish Metapopulation.” Molecular Ecology 21: 4695–4705. [DOI] [PubMed] [Google Scholar]
  71. Seidl, R. , Thom D., Kautz M., et al. 2017. “Forest Disturbances Under Climate Change.” Nature Climate Change 7: 395–402. [DOI] [PMC free article] [PubMed] [Google Scholar]
  72. Selkoe, K. A. , Gaggiotti O. E., Treml E. A., Wren J. L. K., Donovan M. K., and Toonen R. J.. 2016. “The DNA of Coral Reef Biodiversity: Predicting and Protecting Genetic Diversity of Reef Assemblages.” Proceedings of the Royal Society B: Biological Sciences 283: 20160354. [DOI] [PMC free article] [PubMed] [Google Scholar]
  73. Selwyn, J. D. , Hogan J. D., Downey‐Wall A. M., Gurski L. M., Portnoy D. S., and Heath D. D.. 2016. “Kin‐Aggregations Explain Chaotic Genetic Patchiness, a Commonly Observed Genetic Pattern, in a Marine Fish.” PLoS One 11: e0153381. [DOI] [PMC free article] [PubMed] [Google Scholar]
  74. Sergio, F. , Blas J., and Hiraldo F.. 2018. “Animal Responses to Natural Disturbance and Climate Extremes: A Review.” Global and Planetary Change 161: 28–40. [Google Scholar]
  75. Shulman, M. J. , and Bermingham E.. 1995. “Early Life Histories, Ocean Currents, and the Population Genetics of Caribbean Reef Fishes.” Evolution 49: 897–910. [DOI] [PubMed] [Google Scholar]
  76. Snead, A. A. , Tatarenkov A., Avise J. C., et al. 2023. “Out to Sea: Ocean Currents and Patterns of Asymmetric Gene Flow in an Intertidal Fish Species.” Frontiers in Genetics 14: 1206543. [DOI] [PMC free article] [PubMed] [Google Scholar]
  77. Srinivasan, M. , Galbraith G. F., Ceccarelli D. M., Cresswell B. J., Strähl S. J., and Williamson D. H.. 2026. “Long‐Term Effects of a Severe Tropical Cyclone on Coral Reef Habitat and Fish Assemblages at the Whitsunday Islands, Central Great Barrier Reef.” PLoS One 21: e0329995. [DOI] [PMC free article] [PubMed] [Google Scholar]
  78. Sundqvist, L. , Keenan K., Zackrisson M., Prodöhl P., and Kleinhans D.. 2016. “Directional Genetic Differentiation and Relative Migration.” Ecology and Evolution 6: 3461–3475. [DOI] [PMC free article] [PubMed] [Google Scholar]
  79. Swearer, S. E. , Shima J. S., Hellberg M. E., et al. 2002. “Evidence of Self‐Recruitment in Demersal Marine Populations.” Bulletin of Marine Science 70: 251–271. [Google Scholar]
  80. Teng, Y. , Li W., Wang X., et al. 2025. “Integrating Population Genomics and Environmental Data to Predict Adaptation to Climate Change in Post‐Bottleneck Tibetan Macaques.” Science Advances 11: eadw0562. [DOI] [PMC free article] [PubMed] [Google Scholar]
  81. Thakur, M. P. , Risch A. C., and van der Putten W. H.. 2022. “Biotic Responses to Climate Extremes in Terrestrial Ecosystems.” iScience 25: 104559. [DOI] [PMC free article] [PubMed] [Google Scholar]
  82. Turner, M. G. 2010. “Disturbance and Landscape Dynamics in a Changing World.” Ecology 91: 2833–2849. [DOI] [PubMed] [Google Scholar]
  83. Wilson, L. J. , Fulton C. J., Hogg A. M., Joyce K. E., Radford B. T. M., and Fraser C. I.. 2016. “Climate‐Driven Changes to Ocean Circulation and Their Inferred Impacts on Marine Dispersal Patterns.” Global Ecology and Biogeography 25: 923–939. [Google Scholar]
  84. Wong, M. Y. L. , and Buston P. M.. 2013. “Social Systems in Habitat‐Specialist Reef Fishes: Key Concepts in Evolutionary Ecology.” Bioscience 63: 453–463. [Google Scholar]
  85. Woodall, L. C. , Koldewey H. J., and Shaw P. W.. 2011. “Historical and Contemporary Population Genetic Connectivity of the European Short‐Snouted Seahorse Hippocampus Hippocampus and Implications for Management.” Journal of Fish Biology 78: 1738–1756. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Figure S1: (a) DeltaK graphs from HARVESTER determining the best K‐value, (b) Discriminant Analysis Principal Components and (c) eigenvalues for each axis for species compared among locations (purple = Papua New Guinea, orange = Lizard Island, green = One Tree Island); asterisks mean significant differences observed among groups.

Figure S2: Two minor modes were created for Gobiodon quinquestrigatus with a K‐value of three.

Figure S3: Two minor modes were created for Gobiodon rivulatus with a K‐value of three.

Figure S4: (a) DeltaK graphs from HARVESTER determining the best K‐value, (b) Discriminant Analysis Principal Components and (c) eigenvalues for each axis for species compared pre‐disturbances (orange) and post‐disturbances (dark red) at Lizard Island; ** means p < 0.001; * means p < 0.02.

Table S1: Amova results showing the population differentiation among three locations for Gobiodon species. Sample sizes included as denoted by n. Number of variable loci that passed filtering parameters denoted by no. loci. PNG = Papua New Guinea, LI = Lizard Island, OTI = One Tree Island.

Table S2: Pairwise FST values and corresponding p‐values between locations for Gobiodon fuscoruber calculated through Arlequin software. Below the grey line are the pairwise FST values between the given locations, while above the grey line are the corresponding p‐values for the FST statistic. PNG = Papua New Guinea, LI = Lizard Island, OTI = One Tree Island.

Table S3: Pairwise FST values and corresponding p‐values between sampled locations for Gobiodon quinquestrigatus calculated through Arlequin software. Below the grey line are the pairwise FST values between the given populations, while above the grey line are the corresponding p‐values for the FST statistic.

Table S4: Pairwise FST values and corresponding p‐values between locations for Gobiodon rivulatus calculated through Arlequin software. Below the grey line are the pairwise FST values between the given locations, while above the grey line are the corresponding p‐values for the FST statistic.

Table S5: Amova results of population differentiation pre‐ and post‐disturbances at Lizard Island for Gobiodon species. Number of variable loci that passed filtering parameters denoted by no. loci.

MEC-35-e70455-s001.pdf (472.5KB, pdf)

Data Availability Statement

The data is available on NCBI project: PRJNA1309028.

Articles from Molecular Ecology are provided here courtesy of Wiley

RESOURCES