How Interconnected Is Life in the Ocean?

When it comes to telling manta rays apart, Asia Armstrong is an expert. The University of Queensland PhD student is studying populations of Mobula alfredi, the reef manta, in the Great Barrier Reef Marine Park off the northeast coast of Australia, and has spent countless hours poring over photos of the fish—snapped by citizen scientists as well as by Armstrong and her colleagues over several decades—with the aim of identifying individuals. “Manta rays have a unique spot pattern on their ventral surface—smudges, dots, stripes,” she explains. In a database of around 1,300 individual animals, “I’d probably recognize half of them now.” 

For the past few years, researchers have been working under the assumption that local manta populations are split between two main regions, a northern one and a southern one, separated by hundreds of kilometers. But a video Armstrong received last June from a dive site off the coast of northern Queensland in between the two supposed ranges threw that assumption into doubt. The video showed two manta rays that Armstrong immediately recognized as members of a population inhabiting the southern region. Indeed, the rays had last been seen some 1,150 kilometers south of the dive site—a distance almost double that of the longest recorded movement for a reef manta. Photographs submitted a few weeks later confirmed the find: one of the two mantas had been spotted again swimming around the same site. 

“Until this point, everything we had [on the mantas] from northern Queensland didn’t match anything from southern Queensland . . . and we didn’t have anything in between,” says Queensland marine scientist Christine Dudgeon, who coauthored the study documenting the finding in July. Although it’s unclear whether the northern and southern populations overlap, the new data extend the southern population’s range by hundreds of kilometers to the north, she says. This information could help researchers devise better plans for worldwide conservation of manta rays, whose numbers are decreasing in part due to the many threats they face from humans, including harvesting for use in traditional Chinese medicine. 

TRACKING GIANTS: With the help of citizen scientists, researchers such as University of Queensland PhD student Chris Lawson have been photographing reef mantas for an online database that notes when and where the massive rays have been spotted.

ASIA ARMSTRONG

The findings are a surprise to local manta researchers, but they hammer home the importance of considering marine organisms’ movements through the world’s oceans when trying to protect them. Research suggests that lots of animal species, many of them commercially, culturally, and ecologically important, could regularly traverse much larger areas than previously realized, whether as swimming adults, or, more often, at other life stages such as larvae or juveniles swept along by or propelling themselves within ocean currents. Until now, most species have been managed and conserved locally, often in marine protected areas (MPAs)—from strict “no-take” zones to areas with more-nuanced rules—overseen by national or regional governments. But the constant traffic of individuals from one place to another means that animals are often moving between areas with different levels of conservation protection and administrative oversight. Unless such geographical links are taken into account, an organism’s protection in one area could easily by undermined by its vulnerability in another. 

Recognizing this, scientists and policymakers are taking the concept of marine connectivity—a term biologists use broadly to refer to the exchange of individuals, genetic sequences, or food and other material between regions or populations in the ocean—into account as they wrestle with how best to protect marine ecosystems from overfishing, climate change, and other anthropogenic pressures. Quantifying marine connectivity is consequently becoming a central focus of marine conservation research, with biologists developing new methods to assess species’ movements and building management guidelines from the results. 

“It is the next issue that policymakers are being faced with,” says Anna Metaxas, a biological oceanographer at Dalhousie University in Halifax, Canada. “The science is telling us, ‘Whoa, you have to consider [connectivity]. You can’t ignore it if you want your MPAs to be viable.’”

The Language of the Sea A relatively young concept in marine biology, marine connectivity is loosely defined as the exchange of individual organisms, food, or other material between habitats or populations in the ocean. In addition to this umbrella term, researchers also use several non-mutually-exclusive subtypes of marine connectivity to describe species’ patterns. These terms have equivalents in research on connectivity in terrestrial habitats, too.
	Structural connectivity Movement due to the physical characteristics of the environment. Ocean currents and seafloor topography help determine structural connectivity.
	Demographic connectivity The effect of individuals’ movement on the size, growth, and other characteristics of a population.
	Functional connectivity Movement due to the ecological traits of an organism. Dispersal ability plus habitat and food preferences influence functional connectivity.
	Genetic connectivity The effect of gene flow on evolutionary processes in a population.

Measuring connectivity among marine animals

Clumps of edible marine mussels (Mytilus galloprovincialis) are a familiar sight along the rockier parts of Portugal’s Atlantic coast. Ecologically and socioeconomically important, Mytilus species help shape the structure and food webs of tidal ecosystems around the world. For the last few years, Henrique Queiroga and his colleagues have been studying the connectivity of wild populations of these mussels—specifically, the exchange of larvae between populations distributed along the coast. “In most marine species, dispersal takes place during the larval phase,” says Queiroga, a marine ecologist at Universidade de Aveiro in Portugal. “Most of the marine species that we know and most of the marine species that we eat have a larval phase—tuna, cod, herring, clams, mussels, crabs, shrimps.” 

Making direct measurements of dispersal in the ocean is a hugely intensive effort.

 —Simon Thorrold, Woods Hole Oceanographic Institution

Because, completely unlike manta rays, these larvae are too small to see with the naked eye, Queiroga’s team recently took advantage of an indirect method of assessing connectivity for a study of mussel populations around Lisbon. Right after it’s spawned, a mussel larva begins growing a calcium carbonate shell. For the next few weeks, the animal is ferried about by ocean currents until it settles on a rock or other surface along the shoreline to begin metamorphosing into a juvenile and later maturing into an adult. The chemical composition of each section of shell, Queiroga explains, depends on the seawater in which it develops, meaning that the base of a mussel’s shell provides a permanent “elemental fingerprint” of where that animal started life.

To create a database of these fingerprints, the team harvested thousands of larvae spawned from mussels in the lab, and deposited them in batches of about 20,000 into larval homes—small PVC tubes with mesh covering either end. They then distributed these homes along more than 120 kilometers of Portuguese coastline, waited a few days for the larvae to begin making their shells, and then hauled them back into the lab for chemical analysis. The result was an atlas of chemical signatures of different spawning regions.

By comparing the shell bases of wild mussels collected along the shoreline to this chemical atlas, the researchers discovered that mussel larvae move quite a bit. Mussels in an MPA south of Lisbon seemed to have contributed offspring not only to their own population during the study period, but also to populations to the north, including another MPA more than 100 kilometers up the coast. The findings underline the practical importance of considering connectivity. Often “you cannot just protect” one population, says Queiroga. “Because if this one is supplied by larvae that come from the other population and the other population is not protected, then your management plans are worthless.”

Such geochemical analyses can offer insight into many taxa beyond mollusks. For example, researchers often use calcium carbonate structures known as otoliths, present in the inner ear of many vertebrate species, to determine the origins of individual fish. One recent study that analyzed the otoliths of more than 100 Atlantic herring (Clupea harengus) demonstrated that measuring the relative concentrations of 17 chemical elements could pinpoint the specific bay or estuary where that fish had been spawned—serving as a “chemical ‘birth certificate’ of their natal origin,” the authors write. 

But the approach is just one of many now being used to assess marine connectivity. Telemetry—using acoustic or satellite-based markers, for example—can offer more-detailed information about the peregrinations of organisms large enough to capture and tag. Work by Jay Rooker’s group at Texas A&M University, for example, has used multiple types of tags to map the movements of sharks and commercially important fish across jurisdictional boundaries in the Gulf of Mexico, while Dudgeon and her colleagues have recently deployed satellite tags on their manta rays to track them in and around the Great Barrier Reef. DNA-based methods, meanwhile, can generate data not just about animals’ journeys between populations, but the resulting exchange of genomic material as well. 

Most of the genetics techniques used until now to study marine connectivity have focused on evaluating genetic diversity as a way to estimate gene flow between seemingly isolated populations. But an increasingly popular method is parentage analysis, which uses genetic markers such as single-nucleotide polymorphisms or repetitive, fast-mutating sequences of DNA called micro-satellites to identify an offspring’s parents. “What’s amazing about the tool is that I can literally match [a larva] with the mother or father that produced it,” says Mark Carr, a marine ecologist at the University of California, Santa Cruz. “As long as the adult doesn’t move, we know where that larva came from.”

Connectivity may not always be necessary to include. But it should always be considered.

 —Anna Metaxas, Dalhousie University

Carr, along with a large group of colleagues, graduate students, and volunteers, recently carried out a parentage analysis for kelp rockfish (Sebastes atrovirens), an abundant species in temperate ecosystems. The team sampled 6,000 fish—clipping a bit of fin tissue from individuals caught with a hook and line, or nicking the skin of freely swimming fish with a customized pole spear while scuba diving—along 25 kilometers of California’s coastline that included multiple MPAs. Analyzing nearly 100 different genetic markers in the samples, the researchers identified eight parent-offspring pairs. At least two of those offspring had moved out of the MPA where their parents lived into areas where fishing is allowed. Four other offspring had also dispersed out of their parents’ MPA, but subsequently found their way into another.

Unfortunately, such projects are expensive and time-consuming, notes Simon Thorrold, an ocean ecologist at the Woods Hole Oceanographic Institution whose group used parentage analysis a couple of years ago to reveal high levels of connectivity among clownfish (Amphiprion percula) and butterflyfish (Chaetodon vagabundus) populations around Papua New Guinea. “[Making] direct measurements of dispersal in the ocean is a hugely intensive effort,” Thorrold says. Because researchers can’t study every species in the ocean using these methods, he adds, “we’re always going to be extrapolating from the few species we know considerably more about.”

Theoretical approaches to measuring connectivity

To aid in making such inferences, researchers turn to computer models. Simulations based on a species’ distribution and life history, combined with physical data such as ocean current measurements, can help labs predict the movements of marine organisms. Metaxas and colleagues, for example, recently modeled larval dispersal to assess connectivity between populations of deep-water corals, ecologically important but difficult-to-study organisms, in canyons in the ocean floor off the coast of Nova Scotia. Using physical data on ocean currents as well as the distribution of corals—which the team visually assessed using a remotely operated vehicle—the researchers estimated the movement of larvae from one population to another and found evidence for frequent exchange between Canadian and US waters.

When possible, researchers complement physical data on ocean currents with biological data on larval behavior. Larvae of many species can swim up and down in the water column, for example, or toward or away from stimuli such as light, noise, and certain environmental chemicals. Nandini Ramesh, a postdoc in atmospheric scientist William Boos’s lab at the University of California, Berkeley, and colleagues recently developed one such biophysical model to estimate the larval movement of more than 700 commercially important fish species around the globe. Results generated by the model suggested that the world’s fisheries are highly interconnected, with many fish spawned in one country’s waters going on to be caught in another’s. Accounting for the species’ economic value to the fishing industry, the team estimated that this international connectivity helps generate more than $10 billion a year.

Such biophysical models can predict, on a large scale, both where individual animals might end up and how they get there. But the findings are difficult to validate, says Thorrold. “While [simulations] can produce quite compelling visuals, there really has been no way of determining how accurate those biophysical models are,” he says. “It’s going to come down to combining [them] with other methods.”

Back in Portugal, Queiroga and his colleagues have been working on just such a combined approach for their mussel populations. Last year, the researchers published a biophysical model of larval dispersal based on ocean currents along the coast and compared it to observations from their geochemical analyses. The results of the two approaches, the team found, were highly correlated. “If the different methods converge, [and their] estimates are similar to each other,” says Queiroga, “then we begin to be more confident about the description of the process.”

Measuring Movement MESA SCHUMACHER Researchers study the movement of marine organisms using both direct and indirect techniques. Some methods are only appropriate for some species, and most groups try to use multiple methods to validate their results.

	TRACKING The most direct way of measuring marine movement is to record observations of ocean-dwelling organisms in situ. Those data can be collected through photographs, videos, or visual surveys or by using  tags that store or remotely transmit location information to researchers. Tracking is particularly effective for large, migratory animals, but is practically infeasible for smaller species and larvae.
	GENETICS Tissue samples from marine organisms allow researchers to quantify genetic diversity within populations and gene flow between them. One particularly resource-intensive and time-consuming approach known as parentage analysis even allows researchers to identify which adults spawned a particular larva. For species that move relatively little in adulthood, such as many species of reef fish, researchers can then extrapolate to figure out where that larva was spawned.
	BIOGEOCHEMISTRY Calcium carbonate-containing structures such as fish otoliths and mollusk shells contain chemical signatures of the composition of seawater in which they formed. By comparing the otoliths or shells of wild organisms to a database of chemical signatures, researchers can establish where an organism developed.
	SIMULATIONS Researchers turn to computer simulations when fieldwork is impractical, or when they want to compare experimental data with a particular theory. This is an especially common approach for estimating the travel patterns of larvae or planktonic organisms, which can be moved great distances by ocean currents. More-complex models incorporate data on larval behavior to try to boost predictive power.

Using connectivity estimates to build networks of marine reserves

As researchers continue to demonstrate the practical importance of marine connectivity, the concept is slowly being incorporated into conservation planning. Metaxas, who acts as an advisor to the Canadian government’s Department of Fisheries and Oceans, notes that policymakers around the world increasingly use ecological data alongside social and economic information to guide the design of MPAs. But marine connectivity is “still the new kid on the block.”

Metaxas and Dalhousie University graduate student Arieanna Balbar recently reviewed nearly 750 MPAs in regions of the globe with advanced conservation initiatives. They found that just 11 percent included considerations of marine connectivity, and most of those that did were in Australia and California, which have the largest and second-largest networks of MPAs, respectively. New research guidelines take a while to trickle down into management plans, Metaxas says, and there’s still a lack of data for many of the marine areas that policymakers want to protect. “You have a disconnect between what managers can do and what scientists write in the literature that they should do.” 

To help conservation planners consider marine connectivity even when data are sparse, Carr and others have promoted the idea of “rules of thumb” to guide the size and spacing of individual MPAs within a larger network of protected areas. According to these rules, developed in California after the state passed the 1999 Marine Life Protection Act, each MPA should have an area of at least 50 square kilometers—large enough to contain all of the adults in most local reef or coastal fish populations. Furthermore, MPAs should be spaced closely enough that larvae from one MPA can disperse to an adjacent one—ideally, a distance of no more than 100 kilometers. 

California’s MPA network is considered a model example, says Carr, and other managers, such as those in Oregon, have based their  own marine protection plans on a similar framework. Yet some researchers argue that rules-of-thumb approaches risk oversimplifying connectivity by ignoring factors specific to individual species or regions, and overlooking empirical data that may be on hand. 

Protected Area California has one of the largest marine management programs in the world, with 119 marine protected areas (MPAs) and five state marine recreational management areas covering a total of 852 square miles. Marine connectivity is an important factor to consider during MPA network design, as the movement of organisms outside protected areas can render them vulnerable. In particular, connectivity influences how conservation planners decide the minimum area of an MPA, to make sure that most adults in local populations will be contained within it, and the spacing between different MPAs, as populations in different regions of coastline may depend on one another for long-term survival.

Metaxas and undergrad student Jenny Smith recently developed a decision tree that tries to incorporate more of that local specificity by walking conservation planners through a series of connectivity-relevant questions about a target species or habitat. For example, according to the tree, a species with larvae that disperse over short distances would likely require that MPAs be large enough to capture that dispersal and ensure each population is adequately protected locally, but the spacing of the MPAs may not matter if dispersal outside the region is rare. The approach helps planners identify the aspects of marine connectivity that are most relevant and weigh them against other factors guiding MPA design—from ecological considerations such as a region’s existing biodiversity to socioeconomic factors such as the income and livelihoods of local fishermen, says Metaxas. Connectivity “may not always be necessary to include,” she explains. “But it should always be considered.” 

Efforts to design effective networks of MPAs are made particularly tough by the fact that populations’ interconnectedness is likely to evolve for most species and ecosystems in the coming years as anthropogenic pressures on the oceans intensify. With climate change, for example, altered weather patterns could drastically reduce the effectiveness of existing MPA networks by disrupting dispersal between areas, recent computational studies have found. “There are a number of things that are going to change,” explains Ramesh. “Ocean circulation is going to change, but also the fish themselves are going to be responding to warmer waters and moving into habitats that they find most suitable. . . . This whole [connectivity] network is going to be shifting over time as the Earth warms.” 

Catherine Offord is an associate editor at The Scientist. Email her at cofford@the-scientist.com.