As Earth warms, satellite images send strong indications of Arctic greening. But a greening signal belies greater complexity. Climate change brings a ground war to the tundra as plants compete for dominance. In many tundra regions, taller shrubs are invading areas once bare or with short cover like lichen. Understanding Arctic vegetation change is vital to understanding carbon storage and feedback mechanisms to help improve climate change models. But revealing granular details is challenging in an environment that is remote, difficult to access, and sparsely populated. For decades, remote sensing has provided Arctic eyes in the sky but has drawbacks. Ecologists are challenged with trying to extrapolate fine-scale patterns from coarse-grained satellite observations. Emerging research suggests drones could help bridge mismatches of scale.
Much of the evidence for Arctic greening comes from data from satellites orbiting Earth since the 1970s. Satellite data provide coarse-scale resolution. Pixel sizes can correspond to areas as large as 64 square kilometers, explains vegetation ecologist Isla Myers-Smith at the University of Edinburgh.
In contrast, for nearly 2 decades at Qikiqtaruk in Canada’s Yukon Territory, Myers-Smith’s research team has quantified Arctic vegetation change in square-meter plots about the size of a coffee table. Year after year, her team dropped 100 pins in each plot, recording every plant, leaf, and stem the pins contact. It’s painstaking work. Assessing each square takes hours.
But the tundra is vast. Only tiny Arctic fragments can be examined in such detail. The overall greening signal provided by satellites—the normalized difference vegetation index (NDVI)—is unmistakable, but are plants growing bigger? Are different plants encroaching? Are changes homogeneous? These questions intrigue tundra ecologists wanting to know what’s happening on the ground.
It’s difficult to scale from square-meter plots to what satellites see over large spatial extents. “You end up with that gap in between,” said Andrew Cunliffe, research fellow at the University of Exeter in the United Kingdom. He led a recent study addressing this gap, published in Environmental Research Letters. Coauthored with Myers-Smith and three others, the study represents a broader effort to bridge scale gaps using drones.
Sharpening the Fuzzy Lens
Satellites tell us about the Arctic, “but through a fuzzy lens,” said study coauthor Jeff Kerby at Aarhus University in Denmark. Satellite data dating back to the 1970s and 1980s can be helpful, but “the pixels are maybe the size of Manhattan,” he said. “From this satellite data, we have evidence of change, just not the information to understand the change.”
The tundra is a fluctuating and complex ecosystem, with climate variables affecting accurate interpretation of satellite data. Arctic snow cover can occur any time of year and obscure what’s happening with the plants below. Often cloudy, the Arctic is also dark for half the year. When present, the angle of Arctic sunlight can create huge shadows. “Shadows are great if you’re taking landscape photos for fun but bad if you’re trying to understand plants with a computer,” said Kerby. A green plant in shadow doesn’t look green.
Enter the drone. Even when fitted with fairly simple, off-the-shelf digital cameras, drones can create a clear picture of what’s happening on the ground. Photos of the same thing from different angles, stitched together, can produce 3D models. Kerby and Myers-Smith have formed the High-Latitude Drone Ecology Network, creating a standardized protocol for tundra vegetation monitoring.
Initially skeptical about the utility of drones, Northern Arizona University’s Scott Goetz, who was not involved in the recent study, is now convinced of their value. “Scale is one of the key issues with remote sensing,” said Goetz, science lead of NASA’s Arctic Boreal Vulnerability Experiment (ABoVE) and deputy principal investigator for science on NASA’s Global Ecosystem Dynamics Investigation. A complete picture isn’t possible with field measurements alone, but linking satellite remote sensing with field data has been a long and challenging path, he explained, also noting that remote sensing resolution is improving. “It’s not that NDVI doesn’t work or that we can’t monitor [plant growth] in a systematic way. It’s more an issue of the component of the system that you want to measure,” he explained.
NDVI data, found Cunliffe and collaborators, performed poorly when it came to indicating plant biomass because this broad-scale indicator of greenness doesn’t discriminate between tiny green organisms like moss or lichen and larger forms like shrubs.
Alemu Gonsamo, a remote sensing vegetation and climate change scientist at McMaster University in Canada who was not involved in the current study, said that if drone-derived structural measures are properly integrated with lidar and greenness measures, “they provide an unprecedented opportunity to monitor changes both in tundra greenness and canopy structure such as canopy height and aboveground biomass.”
When it comes to the utility of drones in this context, people are just getting started, said Northern Arizona University’s Logan Berner, a collaborator on NASA’s ABoVE project. Of Cunliffe’s study, Berner, who leads a study assessing Landsat NDVI trends across the Arctic tundra biome since the 1980s, said, “There is tremendous potential for the sort of work that they have done to improve our understanding for what these changes in tundra greenness mean, why they’re happening, and how the Arctic might change in the future.”
—Lesley Evans Ogden (@ljevanso), Science Writer