Morning Overview

Ravens are far craftier hunters than anyone realized — scientists tracking them in Yellowstone found the birds memorize exactly where wolves make their kills

In Yellowstone’s Lamar Valley, wolves tear into an elk carcass in the predawn cold. Within hours, ravens descend. For decades, biologists assumed the birds got there by shadowing the pack, riding the hunt like feathered scouts. A study published in Science in May 2025 tells a different story: the ravens already knew where to go. They were not following the wolves. They were remembering the landscape.

A GPS study that upended a long-held assumption

A research team led by behavioral ecologist Matthias Loretto, working across the University of Washington and the Max Planck Institute of Animal Behavior, fitted satellite transmitters to 69 ravens and 20 wolves inside Yellowstone National Park. Over roughly two and a half years, the tags recorded each animal’s position at regular intervals, producing one of the largest GPS movement datasets ever assembled for wild corvids.

The results were defined by what they did not show. Across the entire monitoring period, the team documented only a single extended event in which a raven traveled alongside a moving wolf pack. One. Yet ravens turned up at a substantial share of all observed wolf kills within seven days of the predation event. If the birds were not trailing the predators in real time, something else was guiding them.

That something, according to the peer-reviewed analysis, is spatial memory. Ravens preferentially returned to landscape patches where wolves had historically made the most kills, behavior the researchers describe as reliance on a “carcass abundance map.” The birds were not reacting to a hunt in progress. They were forecasting where the next carcass was likely to appear, then flying to that spot on their own schedule.

Some of those flights were staggering. Individual ravens covered more than 150 kilometers to reach a predicted kill zone. The Max Planck Institute noted in its summary of the work that ravens can fly six hours nonstop to reach a site. These are not short, opportunistic hops between nearby meadows. They are deliberate, energy-costly navigations to a remembered destination.

Why ground observers saw a different picture

The U.S. National Park Service has long noted that ravens are attracted to wolf kills and may follow wolves while they hunt. That description is based on decades of ranger and biologist sightings across Yellowstone’s Northern Range, and it is not wrong. People do see ravens near active wolf packs. The limitation is that visual observation from the ground captures only the behavior that happens to be visible.

The GPS data adds a layer that no pair of binoculars could. A raven quietly departing a roost at dawn, flying 100 kilometers to a ridgeline where wolves killed elk three times last winter, and landing near a fresh carcass hours later leaves no trace for a ground observer. The majority strategy, quiet revisitation of known kill zones, is invisible without satellite tracking. The minority behavior, the occasional raven that does tag along with a hunting pack, is the one that generations of field biologists happened to witness.

Both accounts can coexist. The NPS profile accurately reflects what observers have seen on the ground, while the GPS evidence reveals an additional, previously hidden layer of behavior and suggests that real-time following is the exception rather than the rule.

What the study cannot yet answer

The interpretation rests on GPS fixes taken at intervals rather than continuous video. Because transmitters logged positions periodically, short bursts of following behavior between fixes could theoretically go undetected. A raven might briefly track a moving pack for a few minutes between location points without leaving a clear signature in the data. The researchers acknowledge this limitation, but the possibility of brief, unrecorded trailing episodes has not been fully ruled out.

A second open question involves how ravens build and update their mental maps. The paper describes the pattern of revisitation but does not specify how quickly a young raven, one that has never encountered a particular kill zone, learns to fold it into a foraging routine. Whether the map is socially transmitted through communal roosts, individually learned through trial and error, or some combination of both remains unresolved. Earlier research on corvid information sharing at roost sites suggests communal exchange is plausible, but this study does not test that mechanism directly.

There is also the question of flexibility. Yellowstone’s wolf populations, elk herds, and winter severity all fluctuate year to year. The study period offers a detailed snapshot, but it does not yet reveal how quickly ravens adjust their routes when wolf territories shift or when a historically productive valley goes quiet. If the birds revise their maps rapidly, that points to a highly dynamic cognitive system. If they persist in visiting declining sites, it suggests a trade-off between memory and current information, one that could carry real survival costs in a harsh, seasonal landscape.

What ravens’ strategy reveals about Yellowstone’s food web

Ravens are not passive recipients of wolf generosity. The GPS tracks recast them as active participants in Yellowstone’s carrion economy: animals that invest significant time and energy to forecast where food is likely to appear, then navigate long distances to exploit those predictions. That strategy helps explain how a large-bodied, warm-blooded bird sustains itself through Yellowstone winters, when temperatures can plunge well below zero and carcasses are both vital and unpredictable.

It also raises practical questions for wildlife managers. If ravens reliably converge on high-kill-density zones, their presence could serve as an indirect signal of wolf activity, useful for monitoring pack territories without additional collaring. Conversely, heavy raven scavenging at kill sites may reduce the caloric return wolves get from each kill, a dynamic that could influence pack size, kill rates, and prey population models. Those interactions are not addressed in the current study, but they sit squarely in the path of future research.

For now, the clearest takeaway is a reminder that familiar species often turn out to be making far more complex decisions than casual observation suggests. It took satellite tags on 69 birds and 20 wolves to reveal that Yellowstone’s ravens have been running a quiet, sophisticated forecasting operation all along, one that was hiding in plain sight above the valley floor.

More from Morning Overview

*This article was researched with the help of AI, with human editors creating the final content.