Saving Oaks: Remote Sensing Could Detect Invasive Fungus Earlier

Cleaner air, carbon sequestration, habitat for wildlife, erosion prevention—the list of environmental benefits an oak tree boasts may surprise some. All told, these trees have an estimated economic value of $22.3 billion per year in the United States.

For a portion of these trees, however, a pressing threat looms. Oak wilt, a disease caused by an invasive species of fungus, kills hundreds of thousands of red oak trees in the upper Midwest every year, according to the US Forest Service. The Minnesota Department of Natural Resources finds the disease is widespread in the southern half of Minnesota and inching north.

Jeannine Cavender-Bares, PhD, Distinguished McKnight University Professor of ecology, evolution, and behavior in the University of Minnesota’s College of Biological Sciences, aims to develop a faster and more accurate way to detect oak wilt early on in its infection of a tree. The improved monitoring method stands to better inform government agencies and individual land managers’ decisions as to where to target the time consuming and costly efforts needed to stanch the spread of the disease.

In addition to working with scientists in the Department of Forest Resources in the College of Food, Agricultural, and Natural Resource Sciences, Cavender-Bares is also collaborating with researchers from the Minnesota DNR, the US Forest Service, NASA, the University of Wisconsin-Madison, and the University of Nebraska. The project is one of many to receive funding from a recent $8.75 million investment in UMN invasive species research by the state Environment and Natural Resources Trust Fund. These funds were awarded to two invasive species research centers—the Minnesota Invasive Terrestrial Pests and Plants Center (MITPPC) and the Minnesota Aquatic Invasive Species Research Center (MAISRC)—to support individual research projects.

Oak wilt is caused by an invasive species of fungus called Bretziella fagacearum. It can spread above ground over long distances when beetles (drawn to the fruity odor of fungal growths in diseased oaks) pick up fungal spores. The fungus can also spread underground when root systems from two nearby trees graft together, allowing the disease to transfer from one tree to another. After infection, a red oak tree cannot be saved. It will wilt, lose its leaves, and die within a few weeks.

When infected trees are spotted early, killing the individual tree may be enough to prevent the disease from spreading. In more advanced cases of oak wilt, land managers must use machinery to cut slits into the ground to sever root connections around an infected tree and others within a buffer area. The infected trees have to be removed. In some cases, they may also spray a chemical fungicide on nearby trees for added short-term protection.

“The treatment is much easier if we catch it early,” Cavender-Bares said. “Whereas if it starts spreading, it becomes an expensive operation. There’s a lot of effort going into finding these pockets of oak wilt to eliminate them.”

Reflections of Tree Health

Cavender-Bares, an oak specialist, has been studying oak wilt for several years and is building on decades of work conducted by US Forest Service pathologist and collaborator Jennifer Juzwik, PhD. While previous efforts to track the spread of the disease have relied on regular photographs taken from above (generally by human-flown aircraft or drones) her team is the first to apply hyperspectral imaging to detect the disease.

This form of spectroscopy involves looking at the various wavelengths of electromagnetic radiation (which includes both visible light and wavelengths we can’t see) that reflect off of a given material. The method is used widely in science, including on Mars, where it helps scientists find out what the soil is made of and track weather changes. For oak wilt, the reflected wavelengths can reveal key characteristics about a tree’s health, from carbohydrate changes to water loss to how well it performs photosynthesis to sustain itself.

“It’s more accurate and provides more information to use hyperspectral imaging,” Cavender-Bares said. “We can tell you why a tree is sick, not just if it doesn’t look right from above.”

Using spectroscopy, Cavender-Bares can first separate red oaks (which are much more susceptible to oak wilt) from white oaks. Then, looking at the red oaks only, her team can differentiate between the healthy and unhealthy trees. Under experimental conditions, the process can even separate the trees suffering from other ailments (such as drought) from those with oak wilt.

The method isn’t 100 percent accurate; for complete certainty, researchers would need to take a sample of the individual tree in question and run tests on its DNA. It should, however, give researchers a high-probability way to determine which trees need attention.

Gerard Sapés, PhD, lead post-doctoral researcher on the study and a member of Cavender-Bares’s lab team, is studying the physiological processes that take place in infected oaks, as well as how to detect the infected trees across the landscape using spectral imaging of the tree canopy. He hopes the method will provide a much-needed way to detect the disease earlier than current methods allow.

“I hope we can use features in the light reflected by tree canopies to infer the physiological status of the trees, which in turn could allow forest managers to detect infected trees before they show visual symptoms of oak wilt infection,” Sapés said. “Such a tool would highly enhance the chances of controlling the disease before it spreads further.”

From a Higher Vantage Point

Right now, the researchers gather hyperspectral data in several ways. They can use instruments held in the hand or clamped to leaves to study individual trees. Monitoring larger areas involves taking to the air, gathering spectral data from piloted aircraft or unpiloted drones. Over the summer, for example, the team flew a drone with a custom-built sensor attached to collect data at the wavelengths on the electromagnetic spectrum that they believed could be the most important in detecting oak wilt.

For all its potential benefits, spectral imaging does currently face some drawbacks. It can be difficult to obtain the data and then expensive and time-consuming to process it to the point where it could help guide land decision-making. Cavender-Bares is looking at ways to address these challenges and tune the technology specifically for use with oak wilt. Part of the solution, she said, may be in adapting it for use with imagery that’s already being taken from much, much higher above the ground.

“Ultimately we’re hoping it works with satellite data coming online through NASA,” she said. “We’d like to have this be available to the Minnesota DNR and to the Forest Service. Meanwhile, we are also hoping the drone-based sensor we’re developing will be useful to individual land managers for finding the disease.”