Satellites offer new ways to study ecosystems—and maybe even save them

Jeannine Cavender-Bares began her scientific career with her hands in the dirt. Now, she has her eyes in the sky. As a teenager she helped her biologist dad dig deep into the leaf litter of Florida oak forests to catalog the diversity of slime molds, protists best known for uniting into unsightly blobs that creep across the landscape. But it was the oaks overhead that really fascinated her and became the focus of high school science fair entries and her graduate school research.

In forest plots across the southeastern United States, she planted acorns to study how local conditions affect photosynthesis and growth of different oak species. She froze the seedlings’ stems to study how they transport water to leaves and climbed into the forest canopy to measure gases emitted by mature trees.

But such studies could only provide a snapshot of one forest at a time. To get the big picture of forests around the world, Cavender-Bares has sought a higher vantage. Now a plant ecologist at the University of Minnesota, Twin Cities, Cavender-Bares has devised ways to translate light measured by spectrometers flown over forests into insights about their health and resilience. She and others have found this light, captured from an airplane or satellite, holds clues to intimate details such as photosynthesis levels, the genetic diversity of the trees, and even the microbial inhabitants of the soil they grow in.

Such remote sensing methods are not only revolutionizing how scientists such as Cavender-Bares study ecosystems, they’re also poised to become powerful new tools in the fight to protect them. Over the past year scientists have gathered to revise the most important international treaty aimed at conservation, the Convention on Biological Diversity (CBD). With the loss of plant and animal species accelerating, some researchers say conservation efforts should turn to remote sensing to monitor biodiversity in near–real time across wide swaths of the globe—and help policymakers prioritize the most critical areas.

Historically, researchers had to venture out to jungles, deserts, and mountaintops to document the flora and fauna. But species distributions and abundances are changing faster than ground-based surveys can track, because of climate change, human activities, and other factors. Remote sensing offers the possibility of faster, more standardized monitoring across the entire globe. “In the past decade, there’s been a revolution in the technology available to characterize ecosystems from space,” says David Schimel, a research scientist at NASA’s Jet Propulsion Laboratory.

Researchers are just beginning to figure out what remote sensing can and can’t do and how to incorporate it into global conservation efforts. But Schimel and others see great promise for supplementing ground-based measurements with a fuller picture of ecosystems’ health gleaned from aloft. “We want to transform the way biological research is done,” he says.

Cavender-Bares first learned that reflected or emitted light could help signal forest health during a graduate school visit to the Laboratory for the Use of ­Electromagnetic Radiation (LURE) at Paris-Sud University. In lab studies there almost 40 years ago, plant physiologist Jean-Marie Briantais and colleagues had shown they could measure photosynthetic activity by comparing how leaves fluoresce, or emit certain wavelengths of light, before and after being exposed to flashes of extremely bright white light. As the light displaces electrons inside chlorophyll, the intensity of red and infrared (IR) light emitted from the leaves varies, depending on how healthy the plant is. Later, Ismael Moya, a biophysicist at LURE, developed a sensor that required no extra light source, relying on sunlight instead. Moya went on to demonstrate that fluorescence could be detected from an airplane flying over crops, opening the way to surveying fields’ productivity remotely. “I just became fascinated with what we could learn from the interaction of plants with light and have been for 28 years,” Cavender-Bares says. In that time, researchers have mostly used methods based on reflected light, but fluorescence remains a valuable tool.

By now, airplanes, drones, and towers all provide spectroscopic data on vegetation. So does NASA and the U.S. Geological Survey’s series of Landsat satellites, the first of which were launched in the 1970s. Initially, the agencies expected the satellites’ cameras to primarily capture images in visible light, but an experimental spectral sensor on board proved the value of recording more of the electromagnetic spectrum, such as near-IR light, and by 2013 the satellites were monitoring 11 portions of the spectrum. With these “multispectral” data, researchers can monitor how “green” or productive a vegetated landscape is. Spectroscopically detected dips in chlorophyll can also signal a forest that is suffering because of drought or insect invasion—or has been cleared for development.

Cavender-Bares’s occasional collaborator, remote sensing expert Michael Schaepman, aims to glean more insights from spectroscopic data. At a test area on Lägern mountain near Zürich, Schaepman—who is now president of the University of Zurich (UZH)—and colleagues have studied for decades how the beech, ash, and spruce trees there change their biochemistry throughout the seasons. In 2009, the researchers began to crisscross their study site in an airplane equipped with advanced sensors that capture 100 bands of reflected light—enough to identify the unique spectrum resulting from each tree’s combination of chlorophyll, pigments, water content, and other biochemical and physical features. By cross-referencing the data with ground-based measurements of plant physiology, the researchers were able to develop computer models that turn the spectroscopic surveys into useful information about the health and diversity of the forest.

Other remote sensing methods can fill out the picture. Lidar, which typically bounces pulses from an airborne laser off the terrain below, can determine topography and the height of vegetation. (Recently, a lidar sensor was added to the International Space Station.) Based on tree heights obtained through lidar and light reflected from the plants back to a plane or satellite, Schaepman’s team can calculate a forest’s biomass. The researchers have also developed computer models that predict forest productivity based on changes in biomass and in the biochemical composition of the plants. And they have confirmed that satellite-borne sensors measuring long IR wavelengths can track the effects of drought. Plants normally cool their leaves by transpiring water; when they run short, their leaf temperature rises.

Schaepman hopes the lessons they’ve learned—and software they’ve developed—at Lägern can be scaled up to survey the health of larger landscapes from afar, something they are testing across 154 square kilometers of subtropical forest in China.

At the same time, he and others are also trying to zoom in—by developing remote sensing tools that capture finer detail, such as the diversity of species and communities, and even the genetics and molecular characteristics of individual trees. “My dream,” says Bernhard Schmid, a UZH plant ecologist, “is that without even disturbing the system, we can measure every tree.” Such fine-scale data could help fill out a picture of ecosystem health. Greater genetic diversity, for example, makes ecosystems more resilient by enabling some individuals to cope better with disease outbreaks, drought, or other environmental changes.

Again the clues lie in the spectrum of light reflected from foliage. Over the past decade, experiments have shown that varying concentrations of pigment, wax, and other chemicals, as well as leaf shape and other features, can subtly alter the color or brightness of the light. All of those traits have a genetic basis, and in a study published in 2020 in Ecology and Evolution, Schaepman’s team demonstrated that spectral signatures collected from a plane matched up well with genetic data collected on the ground at Lägern.

Cavender-Bares, Schaepman, and colleagues have also shown that sunlight reflected off leaves can reveal evolutionary relationships among diverse plants. Comparing spectral signatures can show which trees in a given forest are most closely related, they reported in 2020 in New Phytologist. The finding suggests satellite-borne sensors may one day be able to map genetic variation in a forest.

Satellite data can even uncover microbial activity and nutrient cycling in the soil—processes that also contribute to an ecosystem’s overall health and productivity. At the Cedar Creek Ecosystem Science Reserve’s Long-Term Ecological Research site in Minnesota, Cavender-Bares’s team dug into the soil to measure respiration, microbial biomass and enzyme activity, and fungal and bacterial composition. The chemical makeup of decaying plants and even chemicals released by living plants into soil can affect the microbes present there, so she and her colleagues used computer modeling to look for correlations between what they found on the ground and lidar and spectral measurements of Cedar Creek’s foliage. They discovered they could infer microbial activity based on remote sensing data, results they reported on 8 October in Ecological Monographs.

Other researchers, meanwhile, are trying to use satellite-derived information about vegetation to gain insights into the animals living there. Walter Jetz, a biodiversity scientist at Yale University, and colleagues have been developing computer models that incorporate data from the research literature about the conditions that various animal species require to thrive, such as plants used for food or nests. They feed remote sensing information about an area’s climate, vegetation, and topography into the models to predict where a given animal species should be found and at what densities.

The result is the Map of Life, a freely available online repository of maps of species’ ranges that draws from 600 million observations worldwide of 44,351 species of vertebrates, plants, and insects. Jetz began to build this resource 10 years ago, aiming to show how remote sensing data can help compensate for the lack of enough conservationists and ecologists to physically count every species and population throughout its range. By constantly updating the observations, Jetz’s team can predict population declines or detect new threats to at-risk species.

There is even progress on eyeing animals directly from the sky. Already researchers have used satellite images to count penguins in Antarctica and follow whales and wildebeests through their migrations. And since 2018, the International Cooperation for Animal Research Using Space project has used a giant antenna attached to the International Space Station to track individual birds, mammals, and any other organisms outfitted with GPS loggers. Tags have shrunk to 4 grams and are getting smaller, meaning ever-tinier birds and even insects such as dragonflies can be followed from space.

Remote sensing does have limitations. Satellite measures of biodiversity are mostly indirect, and they are limited by factors such as the spatial resolution of the satellite’s sensors, how often a given area is imaged, and how many wavelengths are sensed. For biologists used to working with smaller data sets, moreover, interpreting satellite data can be a challenge. “Although there is more spectral data than ever before from airborne, satellite, and ground sensors, the data are not always easy to use,” says Danielle Wood, an applied earth scientist at the Massachusetts Institute of Technology.

New satellites on the horizon, such as Landsat Next, NASA’s Surface Biology and Geology mission, and the European Space Agency’s (ESA’s) Copernicus Hyperspectral Imaging Mission for the Environment, will have much greater capabilities than earlier Landsat satellites, capturing hundreds of bands of light. And even current technologies can be bolstered by combining different types of remote sensing, says Frank Muller-Karger, a biological oceanographer at the University of South Florida. Already, some researchers are purchasing very high-resolution imagery from private satellite companies and overlaying fine scale spectral information and lidar and radar data from other sources to reveal the forest health in ever greater detail. Both lidar and fine-scale spectral measurements from space “are really game-changing technologies, especially in combination,” Jetz says.

He and other researchers stress that remote monitoring still works best if validated and enhanced with data collected on the ground, for example by community scientists, fieldwork, or monitoring efforts such as the National Ecological Observatory Network (NEON), which operates 81 sites across the United States. A team at the University of Wisconsin (UW), Madison, recently attempted to do this by combining NEON field data on birds and vegetation with remote sensing data on land cover and ecosystem health. The study, published in February in Remote Sensing of Environment, showed birds were most abundant where vegetation was diverse, a connection that would not have emerged from just one data set, says UW landscape ecologist and co-author Volker Radeloff.

Such remote sensing data can guide conservation efforts as well as research. In the past, governments and conservation organizations have relied primarily on expert opinion or on field studies covering only a small area to decide how much suitable habitat remains for a particular endangered species—and how much is needed. “The power of remote sensing is that it is possible to extrapolate field measurements to cover entire continents,” Radeloff says.

One recent project combined remote sensing and fieldwork to map suitable habitat for the endangered Siberian white crane, Leucogeranus leucogeranus. The birds breed in wetlands, where the cranberries, tubers, insects, fish, and small mammals they depend on are plentiful. Field researchers working in a small 3-square-kilometer area that included some of the bird’s breeding grounds had determined the cranes tend to avoid breeding where shrubs are invading the wetlands. Because the broader region is mostly inaccessible, even by foot, conservationists could only guess how much good habitat was available. “But with the satellite images, we were able to expand the area of investigation to a region of around 16,000 square kilometers,” says Claudia Röösli, a UZH remote sensing expert. With these images, conservationists can monitor the effects of floods or dry summers to assess management needs for the crane.

Many other conservation uses of remote sensing are in the proposal or pilot study phase. But wider adoption may be coming soon, if the technology is incorporated into the CBD, the landmark treaty signed in 1992 by 196 countries hoping to stem the loss of the world’s flora and fauna. “We are at a key moment,” says Nathalie Pettorelli, a conservation biologist at the Zoological Society of London. “I am hoping satellite data will gain momentum with the CBD, and I am hoping that trust in this source of information will increase.”

The previous CBD goals, set in 2010, included halving the rate of loss of forests and other natural habitats, slowing the spread of invasive species, and protecting coral reefs by 2020. None has been met. And the lack of a standardized way to monitor progress “was part of the reason we failed,” says Andrew Gonzalez, a conservation biologist at McGill University. That’s where some scientists think satellites and other remote sensing methods can help.

The question is exactly what should be measured and how. One potential solution follows the lead of climate scientists, who at the beginning of the 21st century established the “essential climate variables,” a set of standardized measures of carbon dioxide, temperature, and other physical characteristics that countries now use to monitor their progress toward stemming climate change. But Earth’s life is more complex than climate, says Henrique Pereira, a conservation biologist at the German Centre for Integrative Biodiversity Research, making it more challenging to document how it’s changing.

Pereira, Schaepman, and about 1500 other scientists, working under the auspices of an international organization called the Group on Earth Observations Biodiversity Observation Network (GEO BON), have spent the past 8 years developing a set of essential biodiversity variables (EBVs) for assessing and tracking changes in biodiversity across the planet. “To forge international cooperation, we need apples to apples comparisons of progress made in different countries,” says Becky Chaplin-Kramer, an ecologist at Stanford University.

The 20 EBVs fall into six spatial scales, from DNA to species to ecosystems (see graphic, above). Remote sensing can’t monitor all of them, but it shows promise for at least the two most overarching scales, ecosystem structure and ecosystem function, Röösli says. From 2017 through 2020 she coordinated the ESA-funded GlobDiversity project, which tested the potential of satellites to provide three EBVs: the chlorophyll content of the canopy, the degree of habitat fragmentation, and the timing of the vegetation’s yearly growth cycle, known as phenology.

Röösli and her colleagues used field data from 10 reserves—including arctic tundra, wetlands, and a temperate forest—to develop and test algorithms for estimating these EBVs based on high-resolution satellite images. The results show how EBVs derived from satellite data could be used as an ecological early warning system, the researchers say. In Germany, for example, changes in the spectral signatures of spruce forests coincided with a wave of dying trees infected by an invasive spruce bark beetle.

GEO BON recommends countries or regions set up biodiversity observation networks, analogous to existing national climate observatories, that will allow each country to measure biodiversity by whatever means they have available—not just remote sensing but also ground-based methods such as camera traps and citizens counting animals. Computer models will then convert the data into EBV values and biodiversity “indicators,” which can be compared globally.

One example of such a national observatory is already being piloted in Finland. That country has embraced a mix of remote sensing and ground measurements to assess biodiversity using EBVs. Given Finland’s low population density, “remote sensing is the most promising, and perhaps only, way for comprehensive wall-to-wall ecosystem monitoring,” says Petteri Vihervaara, an ecologist at the Finnish Environment Institute.

Over the next 3 years, a Finnish Ecosystem Observatory (FEO) project will test the feasibility of determining the boundaries of wetlands, forests, and other ecosystems in Finland, in particular Lapland, using a combination of satellites, airborne lidar, aerial photographs, field surveys, and machine learning algorithms that pull all these data together to come up with an EBV called “ecosystem distribution.” Knowing where these ecosystems are is a first step toward preserving them.

In addition, FEO is using satellites to monitor water levels over time in EU-protected boreal wetlands called appa mires. This variation, an EBV called ecosystem phenology, depends on rainfall. But ditches dug nearby also affect water flow in a mire, and the remote sensing could provide an early warning of any damaging effects.

CBD negotiators are now considering proposals to incorporate indicators based on at least two EBVs into the treaty: “species distribution” to track invasive species and “ecosystem distribution” (each one’s extent and fragmentation) to monitor the richness of habitats. But whether the world is ready to embrace remote sensing in general and the EBV strategy, in particular, is unclear. Many ecologists back this approach, but it may be a harder sell for the political appointees and government representatives who need to sign off on the new CBD, because of its novelty and a lack of remote sensing resources and expertise in many lower and middle-income countries. Jetz worries that without such monitoring, countries won’t be able to track their progress toward better protection of biodiversity. But there’s not much time left to convince delegates before they meet in May 2022 to finalize the treaty.

Cavender-Bares and some other scientists aren’t convinced that reducing complex ecosystems to a small set of numbers is the best way to protect them. But regardless of whether the revised treaty embraces EBVs, Cavender-Bares is convinced remote sensing will play a critical role in tracking and combating the massive global losses now underway. “We have to be measuring biodiversity from space,” she says.