The way trees sway may communicate water stress

The Western U.S. is experiencing periods of increasingly intense drought, and higher temperatures are exacerbating the issue by drawing even more moisture from plants and soils into the atmosphere. We see the results in the ever-expanding wildfire footprint, with many of the region’s largest recorded wildfires occurring over the last decade. One DRI scientist is innovating a new way to monitor water stress in trees – one that may help monitor wildfire risk across the region.  

Joe Ammatelli is an assistant research scientist developing a method that analyzes water stress in trees using existing traffic camera networks. He began assessing the feasibility of the process during his Masters thesis work and is refining it now with the help of a partnership with University of Connecticut researchers Bob Fahey and Sidnee Everhart. The group is teaming up to understand how changes in the frequency of a tree’s movement, or sway, reflects water availability. At the heart of their research lies the question about how the amount of water in a tree impacts the tree’s movement and stability in the wind. If this is measured and quantified, then video processing algorithms can use traffic cameras to remotely assess water stress, informing risk assessments for wildfires and possibly powerline-disruptions.  

“Water stress impacts how much water is in the tree, and in turn, how the tree moves in the wind,” Ammatelli says. “So, the cool thing about tree sway is, if we can measure how it moves in the wind, we can learn something about how much water it might have. We may also be able to use the sway frequency to tell if it’s more likely to blow over.”  

Ammatelli and his collaborators are using the Harvard Forest Climate Interactions with Forest Fragmentation (CLIFF) site in Massachusetts to further their work. The forest is the perfect experimental site because it is set up with a system that can control the amount of water available along with a natural stand that serves as a control site. This allows researchers to examine various climate impacts to the forest. 

Ammatelli is deploying cameras while his collaborators set up contact sensors on the trees that can measure slight changes in the tree’s tilt. As the tree moves, the trunk’s angle moves back and forth, creating something known as  “periodic motion.”  Sap flow sensors offer a more direct way to measure water stress, while dendrometers measure the expansion of the tree as it grows and expands from water loss and gain. Precipitation gauges and soil moisture data also offer information on how much water is available for the trees. One prior study found that trees generally sway at a lower rate when they’re drier. 

“If you’re measuring the sway frequency over time and you see it decrease, you can maybe say, oh, this tree is experiencing water stress,” Ammatelli said. “That’s kind of my long-term vision.” 

During his master’s work, Ammatelli used a video camera to film swaying trees and developed an algorithm to pull out the sway frequency from the video for any tree in the camera’s field of view. He hopes this can provide a scalable and inexpensive way to monitor water stress, and therefore wildfire danger, in the Western U.S.  

Although other research has examined the relationship between a tree’s sway and its water content, Ammatelli is one of the pioneering scientists using video processing to measure tree sway. He has already produced two algorithms that work well for tall, slender trees with a distinct trunk by measuring changes in brightness levels across video pixels as the tree moves. Unfortunately, he’s found that trees with a lot of limbs often don’t lend themselves to the process as well because their movement is more distributed throughout their limbs.   

“The underlying idea is that you model a tree as a cantilever beam –basically a rigid beam fixed on one end,” he says. “Because it’s fixed at one end and has a vibration, we can measure that vibration to obtain information about its mass and stiffness.” 

When trees fall into powerlines, they often spark wildfires and disrupt power supply, so Ammatelli is hoping that his method can also pinpoint trees likely to fall so that utility companies can address the issue proactively. Existing traffic camera networks like the ones available from Caltrans and Nevada Department of Transportation already provide widespread coverage.  

“Anywhere there’s a video camera that is also observing trees, you could, in theory, monitor water stress there,” he says. “That’s a huge network that already exists, and we can tap into for free.”  

The method also has potential for determining how urban trees experience heat stress, which could provide a way to better strategize their care and planting to maximize their shade benefits in rapidly warming cities like Reno and Las Vegas.  

Ammatelli and his collaborators will remove the cameras and sensors at the end of the summer, when they’ll be able to use the data collected to measure tree sway over the full summer season. They expect to publish the results of their research sometime next year. They are also aiming to produce tangible benefits by commercializing the method, and have received support from the Knowledge Fund — an initiative by the Governor’s Office of Economic Development, which provides support for potential commercialization pathways.

———- 

*This research is possible with support from the Governor’s Office of Economic Development’s Nevada Knowledge Fund.  

Read more: Measuring tree sway frequency with videos for ecohydrologic applications: Assessing the efficacy of Eulerian processing algorithms 

Can’t get enough DRI science in your life? Support DRI with a donation, attend one of our events, or sign up for our monthly newsletter.