Integrating larval connectivity with local demography reveals regional dynamics of a marine metapopulation

Abstract

Many ocean species exist within what are called marine metapopulations: networks of otherwise isolated local populations connected by the exchange of larval offspring. In order to manage these species as effectively as possible (e.g., by designing and implementing effective networks of marine protected areas), we must know how many offspring are produced within each local population (i.e., local demography), and where those offspring disperse (i.e., larval connectivity). Although there is much interest in estimating connectivity in the relatively simple sense of identifying the locations of spawning parents and their settling offspring, true measures of demographic connectivity that account for among-site variation in offspring production have been lacking. We combined detailed studies of local reproductive output and larval dispersal of a coral reef fish to quantify demographic connectivity within a regional metapopulation that included four widely spaced islands in the Bahamas. We present a new method for estimating demographic connectivity when the levels of dispersal among populations are inferred by the collection of genetically "tagged" offspring. We estimated that 13.3% of recruits returned to natal islands, on average (95% CI = 1.1-50.3%), that local retention was high on one of the islands (41%, 95% CI = 6.0-97.0%), and that larval connectivity was appreciable, even between islands 129 km apart (mean = 1.6%, 95% CI = 0.20-8.8%). Our results emphasize the importance of properly integrating measurements of production with measurements of connectivity. Had we not accounted for among-site variation in offspring production, our estimates of connectivity would have been inaccurate by a factor as much as 6.5. At a generational timescale, lifetime offspring production varied substantially (a fivefold difference among islands) and the importance of each island to long-term metapopulation growth was dictated by both larval production and connectivity. At the scale of our study (local populations inhabiting 5-ha reefs), the regional metapopulation could not grow without external input. However, an exploratory analysis simulating a network of four marine protected areas suggested that reserves of >65 ha each would ensure persistence of this network. Thus, integrating studies of larval connectivity and local demography hold promise for both managing and conserving marine metapopulations effectively.

Accounting for ocean connectivity and hydroclimate in fish recruitment fluctuations within transboundary metapopulations

“…, Johnson et al. ). For large fish stocks, connectivity information derived from genetic, tagging and otolith is often used to spatially delineate subpopulations rather than to model temporal dynamics (e.g., Holmes et al.…”

“…Earlier studies have explored how estimates of larval dispersion could be integrated into fisheries or demographic models. They include theoretical perspectives (e.g., Botsford et al 2009b, Castorani et al 2015, but also studies for small regions and highly sedentary species (e.g., Rochette et al 2013, Johnson et al 2018. For large fish stocks, connectivity information derived from genetic, tagging and otolith is often used to spatially delineate subpopulations rather than to model temporal dynamics (e.g., Holmes et al 2014).…”

Connectivity, Dispersal, and Recruitment: Connecting Benthic Communities and the Coastal Ocean

“…This type of information is key to informed marine spatial planning and adaptive management, particularly of marine protected areas (e.g., . Recently, Johnson et al (2018) used information on production, transport (based on parentage), and survival to actually make the necessary calculations (see Box 1) to estimate population persistence over different spatial scales for a coral reef fish. Similarly, Garavelli et al (2018) combined circulation model predictions with a population model to predict that some Caribbean nations could support self-persistent populations of spiny lobster, despite the basin-wide scale of dispersal in that species.…”

A spatially explicit and mechanistic model for exploring coral reef dynamics

“…Reducing this uncertainty requires better estimates of recruitment rates, which can be achieved with tile experiments (e.g., Ritson-Williams et al 2016) or visual assessment of new recruits along transects (e.g., Gilmour et al, 2013; Holbrook et al, 2018). The proportion of locally versus regionally recruited larvae can be estimated with population genetics (e.g., Almany et al, 2017; Johnson et al, 2018) or by modelling larvae plumes (e.g., Golbuu et al, 2012; Wolanski and Kingsford, 2014).…”