Asymmetric dispersal is a critical element of concordance between biophysical dispersal models and spatial genetic structure in Great Barrier Reef corals

Cynthia Riginos,Vimoksalehi Lukoschek,Karlo Hock,Madeleine J H Van Oppen,Ambrocio M Matias,Peter J Mumby,Eric Treml

doi:10.1111/ddi.12969

Abstract

AbstractAimWidespread coral bleaching, crown‐of‐thorns seastar outbreaks, and tropical storms all threaten foundational coral species of the Great Barrier Reef, with impacts differing over time and space. Yet, dispersal via larval propagules could aid reef recovery by supplying new settlers and enabling the spread of adaptive variation among regions. Documenting and predicting spatial connections arising from planktonic larval dispersal in marine species, however, remains a formidable challenge.LocationThe Great Barrier Reef, Australia.MethodsContemporary biophysical larval dispersal models were used to predict long‐distance multigenerational connections for two common and foundational coral species (Acropora tenuis and Acropora millepora). Spatially extensive genetic surveys allowed us to infer signatures of asymmetric dispersal for these species and evaluate concordance against expectations from biophysical models using coalescent genetic simulations, directions of inferred gene flow, and spatial eigenvector modelling.ResultsAt long distances, biophysical models predicted a preponderance of north–south connections and genetic results matched these expectations: coalescent genetic simulations rejected an alternative scenario of historical isolation; the strongest signals of inferred gene flow were from north–south; and asymmetric eigenvectors derived from north–south connections in the biophysical models were significantly better predictors of spatial genetic patterns than eigenvectors derived from symmetric null spatial models.Main conclusionsResults are consistent with biophysical dispersal models yielding approximate summaries of past multigenerational gene flow conditioned upon directionality of connections. For A. tenuis and A. millepora, northern and central reefs have been important sources to downstream southern reefs over the recent evolutionary past and should continue to provide southward gene flow. Endemic genetic diversity of southern reefs suggests substantial local recruitment and lack of long‐distance gene flow from south to north.

Highlights

Recurrent mass bleaching events on the Great Barrier Reef (GBR) have been increasing in severity and extent (Hughes et al, 2017) against a backdrop of multidecadal coral decline arising from tropical storms, crown‐of‐thorns seastar (COTS) predation, terrestrial run‐ off, and fishing pressure (De'ath, Fabricius, Sweatman, & Puotinen, 2012)
Directional spatial autocorrelation as assessed by asymmetric eigenvector mapping (AEM) explained a greater proportion of variance in allele frequencies for both species (Table 1), where the best models were based on along‐ shore north–south movements with inshore connections from the Swains to Keppel Island and to the Capricorn‐Bunker group
Similar re‐ sults were found for A. millepora (Figure S3) with AEM 1 describing the greatest amount of spatial variance in allele frequencies (5.9%, p = .003) followed by AEM 8 (5.7%, p = .007), and AEMs 1, 8, 5, and 3 individually significant below a p = .05 threshold out of 9 AEMs evaluated

Summary

| INTRODUCTION

Recurrent mass bleaching events on the Great Barrier Reef (GBR) have been increasing in severity and extent (Hughes et al, 2017) against a backdrop of multidecadal coral decline arising from tropical storms, crown‐of‐thorns seastar (COTS) predation, terrestrial run‐ off, and fishing pressure (De'ath, Fabricius, Sweatman, & Puotinen, 2012). Three studies turned to asymmetric eigenvector mapping (AEM: Blanchet, Legendre, Maranger, Monti, & Pepin, 2011) where asymmetric processes (such as biophysical migration probabilities) are statistically modelled as spatial autocorrelation structures; inclu‐ sion of AEMs substantially improved predictions of spatial genetic structure for American lobster (Benestan et al, 2016), California sea cucumbers (Xuereb et al, 2018), and Mediterranean striped red mul‐ let (Dalongeville et al, 2018) These first few studies that quan‐ titatively incorporate asymmetric biophysical predictions suggest that directions of larval dispersal are important elements of marine population connectivity. This study provides a framework for aligning spatially rich population genetic data against a priori predictions of asym‐ metric dispersal and represents the most comprehensive analysis of asymmetric gene flow along the full length of the GBR to date

| METHODS

Findings

| DISCUSSION