Hotter and Drier with a Chance of Extinction: Forecasting Biodiversity with Big Data
With $2.5 Million from the National Science Foundation, a national team of interdisciplinary collaborators aims to harness the power of big data and CyVerse to predict the fate of global biodiversity in light of environmental change.
Species are on the move. Those that thrive in moist, cooler environments are quietly disappearing, while those that tolerate hotter, drier conditions are moving in to take their places.
These are just some of the changes among plants, animals, and ecosystems that Brian Enquist, professor of ecology and evolutionary biology at the University of Arizona, has witnessed – over time, through data, and with his eyes.
Enquist frequently conducts fieldwork in the tropics and in alpine mountain environments, where, he said, biodiversity differences in response to global environmental changes are now clearly apparent, and accelerating.
Enquist and then UArizona doctoral candidate Lindsey Sloat work in one of their research areas in Luquillo, Puerto Rico, where they study changes in forest dynamics. Photo by Benjamin Blonder.
"When we talk about climate change, we tend to think about changes in temperature and rainfall," Enquist said. While scientists have created increasingly more accurate models to predict how global temperatures, forest fire patterns, or sea levels will change under different greenhouse gas emissions scenarios, little has been done to predict similar changes in the biosphere, he noted. "What if we can develop similar forecasting ability to predict biospheric changes and what these changes will mean for biodiversity, ecosystems, and human well-being, including for agriculture and emerging diseases and pests?"
Enquist teamed up with Nirav Merchant, CyVerse co-principal investigator and director of UArizona's Data Science Institute, to lead an interdisciplinary collaboration of the nation's scientists aiming to harness the power of big data and cyberinfrastructure to predict global biodiversity changes under different climate outcome scenarios.
The project was funded this year at $2.5 million under the National Science Foundation's Harnessing the Data Revolution program, with just over $966 thousand awarded to UArizona. The grant stemmed from work done by the Bridging Biodiversity and Conservation Science group, a new interdisciplinary initiative at UArizona.
"We want to envision what's potentially going to happen to our world's biodiversity, in particular Earth's forests and how they function to impact water quality and weather, and how much carbon dioxide, a greenhouse gas, they can absorb," Enquist explained. "We want to create a framework like the National Weather Service, but for biodiversity and biosphere functioning."
The year 2020 marks the beginning of the United Nations agreement for the Convention on Biological Diversity (CBD). The CBD has been agreed by 195 nations to require "...maintenance and recovery of viable populations of species for the conservation of biodiversity."
However, what isn't clear, Enquist said, is where on the planet limited resources and time should be spent to help protect the majority of Earth's biodiversity from extinction. With climate change, species are now moving at different rates and in different directions, resulting in altered combinations of species in any one location over time. The interaction between climate change and complex landscapes makes knowing where to conserve species more complicated.
All the same story
Drier climates, depleted snow-packs, disappearing species.
Kosñipata Valley in Manu National Park, Peru as seen from the Wayquecha biological field station. The analyses completed by Enquist’s team may be able to help parks such as Manu strategize and prioritize conservation methods. Photo by Brian Enquist/University of Arizona.
"Depending on where you are in the world, the changes are different," Enquist said, "but it's all underlying the same story. We can see it from the ground: a shift in the species present." Changing atmospheric conditions, changing climate conditions, and changes to the biosphere – the interactive network of Earth's living things – are all linked, he said.
The team's work focuses on changes in plant diversity, but what happens first to plants, happens to animals as well, Enquist explained. The connection makes plants reliable first indicators to watch in anticipating ecosystem changes.
To meet global deadlines for conservation, the team is rapidly assessing current and potential future distributions of species under different climate and land use scenarios. "We are modeling both present and future extinction risk under two greenhouse gas concentration pathways: the best-case scenario as outlined by the Paris Climate Agreement, and the worst-case, business-as-usual scenario."
The researchers also are comparing the outcomes of different conservation scenarios for land-based species, from the roughly 14.7% of organisms that are currently protected up to a hopeful 50%. The possibility of reducing extinction risk through conservation is discussed in a recent Ecography publication.
The team will integrate diverse datasets, including biomass and plant trait data from the research community, detailed plot data from the National Ecological Observatory Network (NEON), and data from NASA's Global Ecosystem Dynamics Investigation (GEDI), a project that uses high resolution lasers to map Earth's forests and topography from the International Space Station.
"We need robust and reliable data sources in order to make these different forecasts for our biosphere," he said.
And, they need a reliable way to integrate those data – which is where Merchant and CyVerse, the NSF's premiere data management platform headquartered at UArizona, come in.
Enquist in the field in the tropical forests of the Kosñipata Valley, Manu National Park, Peru. Photo by Brad Boyle/University of Arizona.
"There is a substantial amount of data that needs to be synthesized in order to make these large-scale biodiversity predictions," said Merchant. "CyVerse provides an efficient pipeline whereby researchers worldwide can share new ecological findings to update forecasts, and continually improve our understanding of our biosphere's responses to global change."
The project's challenge, Enquist said, is to develop models that forecast biosphere changes under different levels of human land use, including deforestation, urbanization, and agriculture, as well as under different potential future climate scenarios: the worst-case, the best-case, and everything in between.
"If we don't change anything, don't get off fossil fuels, then carbon dioxide in the atmosphere is going to continue to increase very quickly. That's the worst-case scenario," Enquist explained. "The best-case scenario is if society actively worked to reduce greenhouse gas emissions and remove carbon dioxide from the atmosphere."
Building a best-case scenario
It could be said that an alternative best-case scenario would be if we used our knowledge and technology to understand which species are most at risk, and how to save as many as we can.
That's why Enquist's and Merchant's team is partnering with organizations such as Conservation International: to apply the information they gather from biodiversity models with knowledge of the world's protected areas in order to mitigate the effects of environmental change and protect the most endangered species and ecosystems.
"Conservation International is very excited to be participating in this project," said Patrick Roehrdanz, a managing scientist with the organization. "It's important from my perspective to know what the potential shifts and impacts of climate change are in making those decisions, because they often aren't considered in developing conservation strategies."
The project builds upon work Enquist and the Botanical Information and Ecology Network team has done previously in collaboration with Conservation International. The goal of the earlier project funded by the Global Environment Facility, Roehrdanz explained, was to use biological information to create models of where species are now, how their populations might shift, and how we might plan for conservation action to minimize impacts on biological biodiversity."
"With this new grant, we aim to take the modeling methods and data pipelines that we developed through the previous collaboration and put them on a rocket ship," he said, "being able to rapidly assimilate multiple data sources to produce updated maps and priorities for conservation."
An exciting aspect of the new project, Roehrdanz added, is the much shorter timeframe of the forecasts. While earlier initiatives created biodiversity models able to forecast to 2070, the researchers' new goal is to predict changes in the 2030s and earlier.
The team will work with several U.S. universities and people in different areas of
Brad Boyle, then a research technician in Enquist’s lab, in the tropical forests in the Kosñipata Valley of Manu National Park, Peru. The team monitors forests along an elevation gradient from the Amazon to the top of the Andes Mountains as part of the Andes Biodiversity and Ecosystem Research Group. Photo courtesy of Brian Enquist/University of Arizona.
governments, including environmental ministers, national parks directors, and reserver managers. Their collaborations extend beyond the U.S., and include partners in South America, Africa, and Asia.
Most of these connections are relationships that Enquist and collaborators have fostered through work on previous grants. In order to learn what information is most useful, he said, he's learned it's important to listen to the people who are making a difference on the ground.
"During these intensive two years, we'll have a mutual conversation to understand what information they need to protect parks, to help governmental policy decisions, and to anticipate future changes in specific locations," Enquist said. "We'll be working with them to present that information in a way that is understandable to visitors and government associations."
Enquist, Merchant, and their collaborators aim to create the foundations of a national effort to improve our forecasts for the changing fate of global biodiversity.
Partner institutions include Arizona State University, the University of Maryland, the University of Connecticut, Lawrence Berkeley National Lab, Conservation International, the National Center for Ecological Analysis and Synthesis, NASA, and remote sensing partners at Planet Labs.