Researchers examine a visualization made by the Moorea Island Digital Ecosystem Avatar at a physical-modeling workshop in Zurich. PHOTO: Neil Davies

Measuring and Modeling

Two Berkeley projects assess the present and predict the future of Earth’s ecosystem

Harnessed to a 100-foot-high metal tower surrounded by conifers in Colorado’s Rocky Mountains, a research technician checks sensors that measure how fast the trees and the atmosphere are exchanging carbon dioxide. A world away on the French Polynesian island of Moorea, researchers scrutinize the threat that the same greenhouse gas poses to the tiny organisms that make coral. 

Two centuries of burning carbon-rich fuels have pushed global greenhouse gas levels higher than they’ve been in more than half a million years. The warnings are familiar: Global warming and increased ocean acidity will likely bring rising seas, dying corals, struggling wildlife, and disrupted coastal communities.  

But how soon will disruptions come, and how problematic will they be? The urgent need for answers requires refined models of global climate change and its impacts. And the models need data—a great deal of data, drawn from ecosystems ranging from alpine treetops to delta wetlands to tropical seas.

“If we’re going to live sustainably on the planet—on this island in space—we need a much deeper grasp of how ecosystems function and our likely impact on them,” says Neil Davies, executive director of Berkeley’s Richard B. Gump South Pacific Research Station, on Moorea.

Davies leads one of two ambitious and vastly different research projects in which Berkeley scientists are applying advanced technologies and intense analyses to measure an array of biological and physical processes in unprecedented detail, to refine models of likely local and global change.

Building a digital avatar

The younger project is the Island Digital Ecosystem Avatar, or IDEA. It was launched in 2015 to generate, as Davies puts it, “digital models of island ecosystems—islands in silico.” It’s a collaborative effort by UC Berkeley; the Moorea-based French Insular Research Center and Environment Observatory; the Swiss Federal Institute of Technology; Oxford University; and UC Santa Barbara, which leads the National Science Foundation’s Moorea Coral Reef Long-Term Ecological Research site.

Scientists from these and 18 other institutions are focusing intensely on the Moorea ecosystem, from its reefs to its green-cloaked peaks. Their conception of the ecosystem includes not only the organisms and their physical environment but also social-ecological systems—the local population’s resource use and economy.

The pioneering project is intended to develop a model of island ecosystem dynamics in order to inform still-larger models of the impacts of global climate change. Davies expects that the effort to understand the true ecology of a single populated island will provide a template for modeling ecosystem change, both on other islands and, ultimately, on Earth, our island in space.

Screenshot of the Moorea Avatar

A screenshot of the Moorea Earth Avatar interface. Users can manipulate the avatar to see visualizations of data sets and time periods. 

To convey IDEA’s overarching goal, Davies cites fruit flies and nematodes—the model organisms of biomedical research—as an example. “These creatures are not as complex as humans, but they’ve yielded enormous insights into human biology,” he says. “We’re developing Moorea as a ‘model ecosystem,’ and we expect the digital avatar—a computational data science infrastructure—to provide insights that can be applied to other, much larger-scale ecosystems.”

Developing the digital ecosystem model involves drawing on dynamic “time series” data ranging from island organisms’ genomes to 3-D satellite images of Moorea’s entire terrestrial and marine environment. “IDEA is not just a video game,” Davies stresses. “We want to put Moorea on a supercomputer, but it must also correspond to how the island works in the real world.” It makes good sense to concentrate on a limited physical area because it’s tractable, he says. Although exuberantly diverse in habitat and species, the island is just 50 square miles—about the size of San Francisco—with a human population of only 17,000. For that reason, it’s already one of the best-studied islands in the world.

“We’re developing Moorea as a ‘model ecosystem,’ and we expect the digital avatar to provide insights that can be applied to other, much larger-scale ecosystems.”
— Neil Davies

The IDEA project builds on another ambitious undertaking, called the Moorea Biocode Project, which Davies launched 10 years ago as a first step in building the Moorea model ecosystem. With over $5 million from the Gordon and Betty Moore Foundation, the Berkeley-led collaboration identified and cataloged the genomes of—or at least a snippet of DNA from—every land and sea species visible to the naked eye on Moorea, totaling 6,000 species to date. Biocode involved many College of Natural Resources faculty, graduate students, and postdoctoral researchers, including environmental science, policy, and management (ESPM) professors George Roderick and Matteo Garbelotto, who coordinated the terrestrial invertebrate and fungal campaigns respectively. “The IDEA project takes a systems approach,” Davies says, “and we cannot fully understand systems without a reasonably complete parts list. For an ecosystem, that means all of its constituent species.” IDEA investigators are now using Biocode’s unprecedented database of marker genes to build a high-throughput “genomic observatory” in which to study how all the organisms interact—for example, through pollination networks and food webs. The effort is, as Davies puts it, “dynamically mapping the island’s interactome.”

Corals—the canaries in the coal mine

Agriculture and fishing, as well as water and energy consumption, drive a good deal of the ecosystem dynamics on Moorea. In turn, human activities are influenced by those dynamics—for example, logging and increased sediment runoff damage coral reefs and deplete fish populations, and these changes loop back to potentially stress the local economy.

“We all live and act primarily at the local level, and our avatar models must include human communities as critical participants in the system of interactions that constitute a place,” Davies says.

Essentially, the IDEA project is an exhaustive assessment of connectivity, mapping the biological and social networks that characterize complex living systems. At the hub of the island’s ecological and economic connections are the coral reefs and lagoons that protect Moorea from battering surf and provide nurseries for a spectacular diversity of fish and other marine life vital to the island’s economy and the ocean’s food web—the rainforests of the sea.

Neil Davies
Photo: Maheata White

On many tropical islands, sediment and other pollutants pouring in from construction and agriculture degrade the ability of the tiny coral polyps to construct their calcified skeletons. Bleaching induced by warming waters can cause widespread coral die-offs. Another devastating potential: Increased atmospheric carbon dioxide (CO2) levels are leading to an increase in ocean acidity that might doom many calcifying marine organisms. Lab experiments are already demonstrating that the acidity can dissolve their skeletons, says Davies.

For now, French Polynesia maintains an almost ideal pH and a healthy coral habitat. “That was the case when Captain Cook arrived in Tahiti, and Polynesia is still the sweet spot for coral,” he observes. But models show that in 25 or 30 years Polynesia’s coral habitat will be only marginally good, and by 2100 its corals are really going to be struggling. “It will be difficult to reverse,” he says.

For the most part, researchers don’t know if coral species are uniformly vulnerable, or if genetic differences might make some species more resilient. The Biocode project promises to distinguish between closely related corals, allowing lab experiments to determine which genotypes may be the hardiest.

Many species of flora and fauna will find themselves maladapted to new climate-induced conditions, and the changes may occur too fast for genetic variation to catch up. “We might need to have ‘assisted migration,’ ” Davies says. “Like it or not, we’re going to be designing ecosystems. We’d better get good at it very quickly.”

Looking to invasive species

Changes in land use have ushered in aggressively invasive species, such as the purple-leafed miconia tree, which now dominates parts of Moorea’s hillside terrain and alters erosion patterns and water availability. These changes, along with higher temperatures, can boost populations of mosquitoes that carry the dengue and Zika viruses. Rising temperatures allow invasive species such as mosquitoes to expand their ranges and alter interactions among other species, says ESPM’s Roderick, an expert in insect invasions and chair of his department. After Biocode, Roderick joined the IDEA project, investigating how breakthroughs in molecular biology and a new generation of genetic technologies could be safely and ethically used for insect control, including the reduction of human exposure to disease without disrupting the surrounding ecosystem.

Image of a mosquito and headshot of George Roderick

Changes in land use can boost populations of mosquitoes that carry the dengue and Zika viruses, says George Roderick. Photos from Left: iStock; Mel Roderick

Researchers found that a 2013 Zika outbreak infected two-thirds of the people in French Polynesia, including Moorea. And it’s from this region that Zika traveled to Brazil and the rest of the Americas, most likely via infected people. Such are the consequences of globalization, says Roderick.

Recent hints of a startling cause-and-effect relationship between at least five very different species on Moorea, from mosquitoes to humans, drive home the potential of the IDEA strategy to uncover utterly unsuspected connections. Scientists have noted that coral reefs in French Polynesia bounce back from the impact of cyclones or bleaching events, while those in the Caribbean and some other regions are more vulnerable after these assaults, often decimated by hardy strands of algae that smother the coral polyps.

A key factor in the ecological resilience of Moorea’s coral reefs seems to be the fact that coastal development and fishing pressure have not yet undermined the island’s herbivorous fish populations. For example, flashy blue and green parrot fish feed voraciously on algae. UC Santa Barbara marine biologists Sally Holbrook and Russell Schmitt, both members of the IDEA consortium, have found that shallow, near-shore lagoons appear to protect parrot fish larvae. This assures a thriving adult population of the algae-eaters, allowing corals to regain their foothold even after severe weather.

Because IDEA scientists investigate so many different levels of interactions, Holbrook and Schmitt learned that the alarming 2013 spike in Zika infections on Moorea may have coincided with a coral die-off in some—but not all—island lagoons. Other investigators noted that the areas of concern were favored fishing sites.

Avatar researchers began making the connections: They suspect that when mosquito-borne viruses sickened family members, the illness prompted people to fish more to pay for mosquito repellents and medical services. This in turn may have reduced the parrot fish populations in these fishing lagoons. As a result, algae growth spiked and corals died.

So buzzing mosquitoes might threaten coral below the surface of the ocean. Who knew? Along the way, people suffer, lagoons are overfished, and algae snuffs out the island’s biological and economic engines.

For now, this chain of events remains a working hypothesis, but it shows how integrating interdisciplinary data across an entire system can lead to new discoveries. The implications extend from health and economics to ecology and ecosystem stability.

 “We have everything to gain by recognizing the drivers of ecosystem dynamics both at the local level and across the planet,” Roderick says. “Many of the players influence each other in ways we haven’t even considered.

“We recognize the overall threats of increased greenhouse gases and globalization, but we need more data at every level to refine our models and predict change and resilience more precisely. That will inform our ability to develop effective responses. This is not a challenge we can ignore.”

Measuring the cycles of climate change

Margaret Torn next to an eddy covariance tower

Photo: Roy Kaltschmidt/Lawrence Berkeley National Lab

Clearly, to reach their goals, IDEA’s scientists need input of all kinds, and they’re committed to sharing data and insights with researchers working at different scales. Not surprisingly, climate science is truly a big-data enterprise. It takes intense analysis and increasing collaboration to incorporate the mother lode of data into complex, sophisticated models.

Another research effort, AmeriFlux Management Project, gathers massive amounts of environmental data of the kind that could greatly inform the IDEA model. AmeriFlux aims to refine understanding of how land ecosystems function and how climates and ecosystems interact. The 20-year-old network, managed by the Lawrence Berkeley National Laboratory (Berkeley Lab), integrates data from field stations that monitor carbon dioxide, water, methane, and energy flux in nearly every type of terrestrial habitat—from tundra and savanna to boreal woodlands and tropical forests. It offers data about what controls photosynthesis and plant growth, soil carbon cycling, and the transpiration of water through plants from soil to the atmosphere. Experts in biology, atmospheric science, and biogeochemistry study this interplay at more than 100 mainland research stations from the Arctic to South America.

“Terrestrial ecosystems are the very largest sources and sinks of atmospheric carbon dioxide and water,” says Margaret Torn, an ecologist, biogeochemist, and leader of AmeriFlux management in Berkeley Lab’s Climate and Ecosystem Sciences Division—as well as an adjunct professor in CNR’s Energy and Resources Group. “If we want to understand climate and climate feedback within ecosystems, we need to measure the exchange of CO2, methane, and water between the earth’s surface and the atmosphere.”

The levels of gases in any ecosystem vary throughout the year. Carbon dioxide concentrations in the air decrease during the growing season—when photosynthesis outpaces respiration—and increase in the dormant period, when decomposition and respiration release CO2. Levels never exceeded 300 parts per million over the last 800,000 years until the industrial age. We’re not likely to see that level again in many lifetimes.

“When I started graduate school in 1977, some of my atmospheric carbon dioxide measurements were 325 parts per million,” says Dennis Baldocchi, an ESPM professor and one of the founders of the AmeriFlux research network. “When I came to Berkeley in 1999, concentrations in the same season of the year were up to 370 parts per million. Then they reached 400. We’ve already waited too long. With increasing temperatures, ecosystems will change over time. Our biggest problem is that it’s happening very fast, and we have seven billion people who depend on this biosphere.”

“We’ve already waited too long... Our biggest problem is that it’s happening very fast, and we have seven billion people who depend on this biosphere.”
— Dennis Baldocchi

In order to measure the levels of carbon dioxide, water, and energy in any given ecosystem—what Baldocchi calls the “breathing of the biosphere”—AmeriFlux researchers use a technique known as eddy covariance, which combines precise, continuous measurements with sophisticated statistical computation. Turbulent eddies are found within all moving airflows, with both vertical and horizontal components, Baldocchi explains. Eddy covariance measures the velocity of air drafts as they move up and down and captures the instantaneous concentration of trace gases in the drafts. Direct second-by-second measurements are made without interfering with the airflow. Sensors are mounted on towers of varying heights, about 10 feet for grassland, 300 feet for conifer forests.

An accelerating rate of atmospheric change

Rising temperatures caused by increased atmospheric carbon dioxide will likely stress many plants, which in turn can affect atmospheric processes. And, in at least one key habitat, the changes can unleash a vicious cycle. In Arctic latitudes, when the heat of summer thaws the upper few feet of frozen ground, soil microbes decompose and convert dead plant material into greenhouse gases. Specifically, the summer’s microbial metabolism releases huge volumes of carbon dioxide and another potent greenhouse gas: methane.

Dennis Baldocchi in the field and eddy covariance equipment

Left: Dennis Baldocchi wades through a restored wetland at a research site in the Sacramento-San Joaquin River Delta. PHOTO: Joseph Verfaillie. Right: Equipment on the top of an eddy covariance tower in the wetlands of North Carolina’s Alligator River. PHOTO: Guofang Miao.

If  global warming raises summer temperatures a few degrees, the increased heat will push the thaw deeper, allowing microbes access to much more organic material. This could greatly boost their growth and might release much more carbon and methane throughout the world’s northern latitudes.

“Right now, we can’t put precise numbers on this at the scale of the Arctic or the ecosystem,” Baldocchi says. “But we’re running experiments where we compare the CO2 emissions in permafrost soil heated to different temperatures.”

AmeriFlux aims to strengthen communication within the far-flung community of scientists working at different sites. “A plant physiologist assigns a positive number to the flow of CO2 from the atmosphere into a photosynthesizing plant,” says Baldocchi. “But an atmospheric scientist would record the flow of CO2 as a negative number because it’s leaving the atmosphere. We need to use the same language.” Says Torn, “If we get people talking and working together, we can get a look at a bigger, richer picture.”


Banner photo at top right: Researchers examine a visualization made by the Moorea Island Digital Ecosystem Avatar at a physical-modeling workshop in Zurich. Photo by Neil Davies.