Boise State biologist and Associate Professor Trevor Caughlin won a fellowship through the National Science Foundation Established Program to Stimulate Competitive Research, also known as EPSCoR. Collaborating with University of New Mexico’s Jennifer Rudgers, distinguished professor of biology and director of Sevilleta Long-Term Ecological Research program, the team will investigate how weather patterns affect plant clusters in dryland ecosystems. The research will combine cutting-edge drone technology and field experiments to identify early warning signals of imperiled ecosystems.
What are drylands?
Though “drylands” might conjure images of barren and inhospitable wastelands, dryland ecosystems support vast networks of life across the globe. Accounting for 35% of the Earth’s landmass and housing 38% of the world’s human population, drylands include various ecosystems like the grassland, rangeland and steppe biomes that characterize much of the American West.
One of the largest threats to dryland ecosystems is desertification — when arid yet fertile lands degrade and become desert due to repeated challenges posed by climate, taxing agricultural practice and urbanization. A prime example of desertification is the 1930s Dust Bowl.
How can you tell when drylands are danger?
Bring in the drones.
Previous studies have relied on satellite imaging and fieldwork to track botanical landscape changes over time. However, satellites capture images in large pixels around 30 by 30 meters in size — an area bigger than the average American backyard. Such images lack the high resolution needed to identify specific plant species. And, though fieldwork can remedy the resolution issue, attempts to catalog plant ecosystems on foot can take months. Thankfully, modern technology offers a solution: drones.
“At Boise State, we’ve developed drone technology and modeling to identify up to hundreds of thousands of individual plants in a few hours with over 90% accuracy,” Caughlin said. “So we’re really excited to apply this technology to understand how changes in plant communities might provide early warning signals for major ecosystem degradation that’s really extremely difficult to reverse.”
By examining the presence, number and positioning of individual plant species, researchers can evaluate an ecosystem’s health by its ability to maintain healthy populations of native species. When there are enough clusters of healthy plants (such as native grasses, sagebrush and bitterbrush communities) to maintain soil structure, secure water supply and shade native seedlings, the environment is better able to withstand longer periods of drought.
“The Sevilleta LTER program has been collecting boots on the ground data since 1989 and we are thrilled to be able to add this new, drone-enabled window on how dryland ecosystems are structured,” Rudgers added.
Cultivating the next generation of scientists
“One thing I’m excited about at Boise State is that we are taking a lot of steps to train the next generation to use these technologies,” Caughlin said. “Our drone certification program is one of the only programs of its kind in the country. For states like Idaho, where managing our natural resources is so important, there has been an immediate demand for students with this training from a wide variety of sources.”
Though much of Caughlin and Rudgers’ efforts will take place at the University of New Mexico’s Sevilleta Long-Term Ecological Research Station and U.S. Fish and Wildlife Services Sevilleta National Wildlife Refuge, Caughlin aims to develop and manage similar long-term collaborative experiments at Dry Creek Watershed just north of Boise. He hopes this research will benefit land management organizations like the Bureau of Land Management and others that are working to conserve and restore lands, like the Boise foothills, so recently damaged by wildfires.
This ongoing, multisite approach will allow the researchers to track slow-moving ecosystem changes over time while training the next generation to use drone technology in the fight to save natural resources.
This publication was supported by the National Science Foundation EPSCoR Program under award number OIA-2531765.