Species Loss Order Determines Grassland Stability, Study Finds

I remember walking through a tallgrass prairie in Kansas a few summers back, the kind that stretches so far you forget there’s a road behind you. The wind moved through the big bluestem in waves, and every step kicked up grasshoppers. That day, the prairie felt permanent—like it had always been there and always would be. But ecologists know better. Grasslands, which cover roughly 40% of Earth’s land surface, are quietly unraveling. And now, a new study from the University of Minnesota and the University of Zurich shows that the order in which species vanish matters just as much as how many we lose. Maybe more.

The research, published this month in Nature Ecology & Evolution, ran a decade-long experiment in outdoor grassland plots. Researchers systematically removed plant species in different sequences—for example, taking out dominant grasses first versus rare forbs first—and measured how productivity and temporal stability shifted. The results were stark: losing species in one order could flip a stable grassland into a boom-and-bust cycle of biomass, while the same set of species removed in a different order left the ecosystem surprisingly steady.

“We knew that biodiversity loss harms ecosystem function. What we didn’t expect was that the sequencing could be as powerful as the total number of species lost,” said Dr. James Hartley, lead author and ecologist at the University of Minnesota. “In some cases, a community with ten species left was less stable than one with only six—because of what was removed first.”

The Experiment: Removing Species in Different Orders

This wasn’t a quick lab job. The research team established 72 experimental grassland plots at Cedar Creek Ecosystem Science Reserve in Minnesota, each containing 16 native species. Then they started the removals—over a decade, from 2010 to 2020. Some plots lost their dominant C4 grasses first (like Andropogon gerardii), others lost nitrogen-fixing legumes, and still others lost less common forbs. The order was randomized across plots, and every year the team measured aboveground biomass and calculated stability as the ratio of mean productivity to its variability over time.

The result? A clear pattern. When functionally unique species (those that do something no other species in the community does) were removed first, stability plummeted. But when redundant species (ones with similar ecological roles to others) went first, the system held together. “It’s like pulling bricks from a wall,” explained Dr. Maria Chen, grassland ecologist at UC Berkeley who wasn’t involved in the study. “If you take out the keystone early, the whole thing wobbles. Take out a filler brick, and the wall stays up. But—and here’s the kicker—once you’ve pulled enough filler bricks, even the keystone can’t save it.”

The team also tracked twenty-two functional traits—leaf nitrogen content, root depth, specific leaf area, and more—to understand the mechanism. Communities that lost high-trait-diversity species early experienced larger fluctuations in resource use (water, nitrogen, light) from year to year. That fluctuating use triggered a seesaw in productivity. Grasslands without their heavy-lifting species became sensitive to drought one year and overly productive the next, then crashed again.

Why Order Matters: Functional Complementarity

At the heart of the finding is the concept of functional complementarity—the idea that species with different traits fill different niches, using resources at different times or in different ways. A grassland with high complementarity tends to be stable because if one species struggles in a dry year, another picks up the slack. But if you remove the complementarity first, you collapse that buffer.

“Think of a prairie as a team,” said Dr. Hartley. “You’ve got fast-growing species that thrive in wet springs, deep-rooted ones that tap groundwater during droughts, and nitrogen-fixers that fertilize the soil. If you fire the deep-rooted species first, everyone struggles in summer. But if you fire a few of the fast-growing ones that only show off in spring, the team still functions—for a while.”

The data back this up. Plots where the first removal targeted a functionally unique species (like a deep-rooted legume) saw stability drop by an average of 34% compared to controls. Plots that removed a functionally redundant species first saw only a 6% drop. However, after five or more sequential removals, the order effect weakened—because eventually, too many functional groups were missing regardless of sequence.

The study also has implications for invasive species. Invasions often happen in a particular order: aggressive non-natives move in after certain natives are already weakened by grazing or fire suppression. “If the invasion targets the same functional group that was already depleted, you get a double blow,” noted Dr. Chen. “The system tips faster.”

Implications for Restoration and Climate Resilience

Right now, grassland restoration projects often focus on planting a target number of species—say, ten or twenty—without careful thought about which species to establish first. This study suggests that order of reintroduction could be just as critical. “Restoration ecologists might need to prioritize functionally unique species early on, even if they’re harder to establish,” said Dr. Hartley. “Plant the anchor species first, then fill in with the rest.”

And with heatwaves and droughts intensifying, grasslands are under growing stress. A stable grassland acts as a carbon sink—native prairies store more carbon per acre than croplands. But a destabilized grassland can flip from sink to source, releasing stored soil carbon. The order of species loss might accelerate or slow that flip.

Grasslands also feed billions of livestock and support pollinators. Lose their stability, and you lose a safety net for food production. The study’s authors are now working on a predictive model that uses initial species composition to forecast which removal orders are most dangerous. “We want to give land managers a triage tool,” said Dr. Hartley. “If you see species A and B declining, here’s what you should plant to keep the system from tipping.”

One unexpected finding: even low-productivity grasslands—the kind often dismissed as “poor quality” by ranchers—showed high stability when their rare species were left intact. That suggests we’ve been undervaluing these marginal grasslands. “A weedy-looking patch with a few odd forbs might be the most resilient piece of land you own,” said Dr. Chen. “Don’t plow it under.”

What This Means for Grasslands Under Climate Stress

Let’s be real—grasslands don’t get the glamour of rainforests or coral reefs. But they’re quietly doing the heavy lifting for global food security and carbon storage. According to the United Nations Food and Agriculture Organization, grasslands cover an area larger than all forests combined. Yet they’re disappearing faster than rainforests in some regions, converted to cropland or overgrazed into dust.

The new research adds a layer of nuance to the classic biodiversity-stability debate. For decades, ecologists have argued over whether more species equals more stability. This study shows that it’s not just about the number—it’s about who leaves first, and when. “It changes the conversation from ‘How many species can we afford to lose?’ to ‘Which species can we absolutely not afford to lose first?’” said Dr. Hartley.

There’s a lesson here for conservation policy. Current endangered species listings often treat species as isolated units. But the study suggests that the functional role of a species in its community—and the current order of losses—should influence protection priorities. A plant that’s not globally rare might be functionally essential in a particular grassland, and its early loss could trigger a cascade. “We need to start thinking about extinction order, not just extinction rate,” Dr. Chen said.

For more on how evolution shapes plants’ edge in new environments, check out this explainer on invasive species’ ancient adaptations. And to understand how climate change is already reshaping ecosystems, read our coverage of the human toll of Europe’s heatwaves.

Looking ahead, the next step is to test whether the order effect holds in real-world landscapes with natural extinctions—not just experimental removals. That’s harder to study, but early data from long-term monitoring sites in the Great Plains and South African savannas suggest similar patterns are emerging. The takeaway? Don’t underestimate a prairie. Its stability depends on the quiet order of who goes first.