Drone flying above a landscape
Photo by Becca Fenwick

Taking the pulse of California ecosystems

The California Heartbeat Initiative uses new technologies to track environmental change on a landscape scale

The climate crisis is changing California in ways unimaginable just a few decades ago. Droughts are growing longer and drier, wildfires are exploding to frightening sizes, and heat waves are setting temperature records on land and at sea. The magnitude of today’s disasters has scientists scrambling to measure the pace and scale of these shifts.

Biologist Todd Dawson believes that remote sensing is critical for monitoring the health of natural ecosystems going forward. “We can’t do this just with boots on the ground,” he says. “There’s not enough of us, and we can’t do it fast enough.”

Dawson and colleagues across the University of California are tackling this challenge with a pilot project called the California Heartbeat Initiative (CHI). Utilizing technologies such as drones and miniature climate stations in novel ways, they are tracking environmental change on a landscape scale.

A new toolkit

Map of water availability

Using drones, California Heartbeat Initiative researchers can track water stress in vegetation across an entire landscape. Data from multispectral cameras are reassembled into a map showing well-hydrated plants (green) versus water-stressed plants (orange) in California’s Channel Islands. Photo by Becca Fenwick.

Just as physicians gauge wellness by measuring temperature and blood pressure, CHI measures characteristics such as chlorophyll content, soil moisture, and microclimate to spot plants under stress—key indicators of waning ecosystem health or environmental change. But instead of assessing vital signs one plant at a time, drones generate data across hundreds of acres of forests and meadows per flight.

“We’re equipping ecologists with a new toolkit and demonstrating how well it can be used to detect change in natural environments,” says Becca Fenwick, CHI project coordinator and lead drone pilot.

Over time, regular CHI-style assessments can pinpoint forests at high fire risk, landscapes on the threshold of major shifts, and ecosystems in need of a human helping hand. “If we see signals of change, we understand that the way the ecosystem is knit together is fraying, and we can look for solutions to protect biodiversity or manage the land in a new way,” says Dawson, a professor in the Departments of Environmental Science, Policy, and Management (ESPM) and Integrative Biology.

Funding for CHI comes from the Gordon and Betty Moore Foundation, and collaborators leading the project include ESPM professor and Cooperative Extension specialist Maggi Kelly, Rausser College Dean David Ackerly, and Peggy Fiedler, MS ’80, PhD ’85, Forest Ecology and Botany, former executive director of the UC Natural Reserve System.

The CHI toolkit is being put through its paces across the UC Natural Reserve System (NRS), a network of 41 reserves in California used for research and teaching. NRS lands represent nearly every major ecosystem in the state, enabling CHI scientists to study a wide range of climate and vegetation types, from succulents in hot deserts to redwoods in coastal forests.

CHI’s versatile toolkit has proven particularly adept at tracking water availability across ecosystems. Water-starved plants are vulnerable to both disease and fire, making landscape water an important signal of ecosystem health in a drying region.

To spot thirsty plants, CHI researchers have mounted multispectral cameras, typically used by satellites, atop drones. They simultaneously capture images in six different light ranges, helping scientists quantify how green and active, or brown and water-stressed, one plant is versus another.

While each pixel in a satellite image covers roughly the area of a baseball diamond, a pixel from a drone camera covers just a few centimeters—a 30x boost in resolution. Drones also enable the scientists to obtain far more comprehensive surveys, since they can fly over places that are inaccessible on foot, like rugged ravines and stands of poison oak.

Mapping with drones

In 2018, the CHI team began using drones to map changes in water availability across seasons at nine NRS reserves featuring distinct ecosystems.

Having a bird’s-eye view of 25 tree canopies, rather than looking up at just one, has been a revelation for Dawson. “It’s like a new set of eyes,” he adds. “We can see one area changing while another remains the same, then go figure out why.”

Todd Dawson measuring a plant

Todd Dawson measures water pressure in leaves. The CHI team uses data from hand-sampling techniques in conjunction with data from drones and miniature weather stations. Photo by Kerri Johnson.

The team also installed miniature weather stations within each flight area. Called ClimaVues, they collect continuous measurements of standard meteorological data, as well as information relevant to plant health, such as leaf wetness and soil moisture. The researchers ground truthed this data with observations gleaned through traditional hand-sampling methods.

The comprehensive data from drones allows the CHI team to create digital surface models of their study areas. Resembling a blanket draped across the landscape, digital surface models enable researchers to deduce the volume and density of vegetation. “We can see changes as they happen—for example, some oak canopies are thinning while others are doing well,” Dawson says. When drone maps are melded with ClimaVue data, “we can explain the differences: microclimates, little pockets of resources or water, that we didn’t appreciate before.”

To date, the CHI team has gathered drone data over some 78 square miles, amassing more than 10 terabytes of data—about the amount of information generated annually by the Hubble Space Telescope. To organize it all, CHI has helped sponsor a geospatial data repository to make processed drone flight images available to anyone.

The meteorological data is served up by a new, open-source data portal called Dendra. Developed by CHI data manager Collin Bode, the portal now collates information from climate stations across the NRS, and it will soon be a one-stop resource for meteorological data from entities such as California State Parks, California Department of Fish and Wildlife, and The Nature Conservancy.

With Dendra, users can easily query the conditions at a reserve across any time period. Dawson contrasts that to the two years he and his colleagues spent calling airports and government offices to locate weather records for a study on fog and redwoods. “If people want to analyze climate trends across the NRS, Dendra is going to be their go-to tool,” Dawson says.

Woman flying a drone.

Becca Fenwick releases a research drone at the Landels Hill Big Creek Reserve in 2020 to capture images of the impacts of the Dolan Fire on the landscape. Photo by Wendy Baxter.

Fire followers

In August 2020, a lightning storm sparked more than 650 wildfires across California. Nine reserves, many of which the CHI team had already mapped, were among the record 2.7 million acres burned.

Amid the devastation, the researchers saw a rare opportunity to observe how landscapes rebound from fire. With additional Moore Foundation funding, they flew drones over the burned reserves a few weeks after the fire, and again six months later. They rounded out the data set with ground surveys of vegetation and arthropods, and collections of soil samples for environmental DNA analysis, to identify returning species. As the scientists are particularly interested in how fire intensity affects ecosystem recovery, CHI postdoctoral fellow Yinan He also used remote sensing data from satellites to generate burn severity maps.

CHI data from before and after the fires is revealing how burn intensity affects vegetation survival and recolonization by plants and animals. Such knowledge about the ecological impacts of fire will only become more important as wildfires increase nationwide.

Stopping a gully in its tracks

In California’s Channel Islands, postdoctoral fellow Kerri Johnson has been applying the CHI toolkit to examine how environmental disturbance can destabilize the very integrity of landscapes.

Johnson’s laboratory is the NRS’ Santa Cruz Island Reserve on California’s largest Channel Island. For 150 years, introduced livestock consumed a considerable amount of the native vegetation. This caused much of the island’s soil to be lost, and some areas even developed gullies—gashes in the land cut by water. Some of these gullies have stabilized since the grazers were removed in the 1990s, but others continue to erode.

“I wanted to figure out how to give these landscapes their best fighting chance."
Kerri Johnson

“Gullies are thought of as a canary in the coal mine, an indication of significant change,” Johnson says. As a gully expands and steepens, water and soil sheet off ever faster, leaving less for plants. This destructive cycle can eventually produce badlands, landscapes unable to absorb water or support vegetation.

“I wanted to figure out how to give these landscapes their best fighting chance to once again support vegetation even during droughts and sequester carbon by building up soil again,” Johnson says.

Using the CHI tools, she’s been able to pinpoint climate as the main factor controlling gully recovery on the island. First, Johnson used lidar data to map the extent of gullies along a section of the island’s mountainous spine. Next, she installed six ClimaVue stations in a line across the ridge, and mapped the area with drones to gauge plant hydration.

Meteorological data revealed that the wind blew fog directly onto the north-facing slope, bathing its plants in extra moisture during driest summer months. The drone images confirmed that these plants were relatively well hydrated year-round. Yet plants on the south-facing slope experienced both a fog shadow and received extra sun, and appeared parched in summer.

Three computer-generated map images

One drone flight can supply different types of information about a single site. Here, the research site on Santa Cruz Island can be seen as an elevation map, a visual spectrum image, and a multispectral water status image.

Johnson’s analyses showed that on the wetter slope, thriving shrubs had stabilized gullies, whereas in the drier area, vegetation has been unable to overcome the history of disturbance enough to stanch gully growth. “We found that specific differences in climate across the ridge are enough that, without intervention, one side will stabilize, and the other side will continue to unravel toward a bedrock badland,” Johnson says.

These findings will help the National Park Service prioritize sites for future restoration, says Channel Islands National Park botanist Cameron Williams, PhD ’17 Integrative Biology. “We’ll be able to put our effort and money into places where we’re pretty sure it’s going to work,” he says.

Funding for CHI ends this fall, but its achievements will continue to benefit environmental research. In addition to helping establish the Dendra data portal, CHI has inspired others to adopt remote sensing tools in wildland management. For example, California State Parks plans to use drones to monitor natural and archeological areas in places such as wildfire-damaged Big Basin State Park. On a broader scale, CHI’s techniques will help California and the world meet the goal of conserving at least 30 percent of lands and waters by 2030 (see "The New Conservation").

“Visualization through remote sensing—and calibration by people on the ground like we’re doing through CHI—will teach us where we need to intervene to rescue our biodiversity and protect our ecosystems in the face of climate change,” Dawson says.

Two sides of a hill, one dry and a one with vegetation

Ocean fog supplied plants on one side of the ridge with abundant moisture, while drier conditions on the other side resulted in the formation of gullies. Photos by Kerri Johnson.