When Coexistence Stops Working: A Threshold Test for Predator Management
A new urban-coyote framework asks a broader wildlife-management question: What behavior should trigger education, deterrence, investigation or targeted removal?
“Coexistence” is an appealing goal for wildlife management.
It is also an incomplete instruction.
People and predators increasingly share working lands, suburbs, parks and the edges of expanding cities. Most encounters do not justify removing an animal. Many conflicts can be reduced by securing food, protecting livestock, changing human behavior, using deterrents and giving wildlife room to retreat.
But coexistence cannot mean repeating the same response regardless of what the animal is doing.
A coyote passing through a neighborhood is not the same management problem as one repeatedly approaching people. Defensive behavior near a den is not the same as stalking. A wolf living near cattle is not the same as an animal repeatedly attacking livestock.
The species may be the same. The behavior, context and risk may not be.
A recent peer-reviewed Perspective in Frontiers in Ecology and Evolution argues that urban coyote policy often fails to make these distinctions operational. Its authors propose a threshold-based system in which predefined behavioral markers—not political pressure or improvised judgment—determine when management should escalate.
Their proposal concerns urban coyotes, not every predator in every landscape. But the question it raises is much broader:
What evidence should cause a wildlife agency to change its response?
Coexistence is active management
Properly understood, coexistence does not mean leaving wildlife problems unattended.
The Frontiers authors describe coexistence as a dynamic condition in which wildlife populations persist alongside tolerable risk and social legitimacy. That definition permits a range of responses, including public education, attractant control, access restrictions, deterrence and, under serious circumstances, targeted removal.
The management failure they identify is not coexistence itself. It is the transformation of coexistence into institutional passivity.
That can happen when an agency adopts education and hazing as standing policies but never defines the conditions under which those tools have failed. Intervention remains discretionary. Decisions are delayed while incidents accumulate. Officials are forced to negotiate standards during a crisis rather than apply standards established beforehand.
A better system would begin with tolerance—but would not end there automatically.
Behavior matters more than presence
The proposed framework separates ordinary presence from escalating behavior.
At its baseline level, coyotes behaving normally require monitoring and public education. An animal moving through habitat, hunting natural prey and avoiding people has not crossed a meaningful intervention threshold merely by being seen.
Early habituation calls for prevention and mitigation: removing food sources, enforcing feeding restrictions and, where appropriate, using active deterrence.
The paper makes a narrower claim about hazing than is sometimes reported. The authors do not say that hazing is useless. They conclude that it can produce avoidance when applied consistently and intensively during early habituation, although effects may be short-lived and dependent on the animal, setting and continued availability of human food.
They object to treating hazing as a permanent answer after behavior has escalated.
Their proposed middle tier requires trained personnel to distinguish transient defensive behavior from patterned boldness. A coyote defending a den from a nearby dog may call for a temporary closure or restricted access. Repeated close approaches in heavily used areas, persistent loss of wariness or a continuing pattern after intervention presents a different problem.
The highest tier includes predatory stalking, predatory attacks on dogs and unprovoked aggression directed at people. The authors argue that such behavior constitutes an acute threshold breach requiring immediate removal.
That is the authors’ proposal. It is not a universally validated rule for every community, much less for wolves, bears or mountain lions. But its central distinction is sound: management should respond to verified behavior and context, not simply to the presence or identity of an animal.
The Toronto case—and its limits
The paper builds its argument around a 2024–25 conflict in Toronto.
The authors report hundreds of coyote incidents, including stalking or menacing behavior, attacks, two dog deaths and two attacks on people. Municipal authorities continued education and low-intensity hazing for months as incidents grew more frequent and severe. Two coyotes were eventually euthanized. Reported aggressive encounters then stopped.
That sequence deserves attention, but it does not prove by itself that removal caused the change. The authors explicitly describe the inference as correlational. The incident data also came from voluntary reporting systems, which can be affected by public awareness, reporting behavior and classification decisions.
The case therefore should not be turned into a slogan such as “hazing failed” or “removal solved the problem.”
Its stronger lesson is institutional. Toronto had no binding, predetermined rule for when verified escalation should move management beyond hazing. Decisions were made under growing public pressure after serious incidents had already accumulated.
The proposed remedy is not simply earlier killing. It is earlier clarity.
Set the rules before the crisis
A defensible threshold system would need more than a chart of alarming behaviors.
The authors identify several requirements:
- Behavioral markers must be clearly defined and grounded in evidence.
- Defensive behavior must be distinguished from predatory behavior.
- Incident reports must be standardized, centralized and released promptly.
- Serious reports must be investigated by trained personnel.
- The persistence of specified behavior after intervention must trigger the next management tier within a defined period.
- Agencies must establish decision authority and accountability in advance.
- The effectiveness of hazing or other contracted interventions should be evaluated independently.
Exact frequencies and time periods cannot simply be copied from Toronto. Each jurisdiction must calibrate them to local ecology, human density, resources and tolerance for risk.
This is crucial. A threshold system built on poor reporting or vague definitions would merely give administrative rigidity a scientific appearance.
The value of the proposal lies not in pretending that judgment can be eliminated, but in identifying the evidence that should discipline it.
Targeted removal is not population control
The most difficult part of the framework is removal.
Here, distinctions matter.
Removing a particular animal associated with verified predatory behavior is not the same as attempting to suppress a predator population across a region. The objectives, scale and ecological consequences differ.
Pitchstone’s archive contains ample reason to distrust indiscriminate predator control. Removing predators can disrupt social groups, change prey relationships and sometimes worsen the conflict it was intended to solve. Pitchstone's archive has also examined evidence that broad or poorly targeted wolf control can increase livestock losses under some circumstances.
The Frontiers paper does not advocate that kind of campaign. It argues for targeted removal after specified behavioral thresholds have been verified.
Even then, implementation is not simple. Individual coyotes may be difficult to identify. More than one animal may participate in an attack. Removing members of a family group may alter the behavior or movement of survivors. New animals may enter the territory if food attractants remain.
Removal cannot substitute for correcting the conditions that produced habituation.
Related Article:
Chris Gill
Nor should “nonlethal” automatically be treated as synonymous with humane. Repeatedly hazing an animal whose behavior no longer responds to deterrence may add stress without reducing danger. Relocation can transfer the conflict, expose an animal to territorial aggression or result in high mortality. Captivity can impose its own welfare costs.
There is no morally weightless option once serious conflict has developed.
The ethical task is to prevent avoidable escalation, act proportionately and minimize unnecessary harm—not to preserve the appearance that no hard choice was made.
Related article:
Chris Gill
People are part of the system
Predator policy often treats people as external to ecology.
They are not.
On working lands, coexistence has real costs. Fencing, range riders, guardian animals, carcass management, monitoring and lost livestock require money and labor. In cities, risk is also unevenly distributed. Residents living beside den sites or repeated conflict areas bear consequences that distant advocates may not.
Pitchstone has long argued that conservation must account for people, property and livelihoods as well as wildlife. That principle does not give anyone an unlimited claim against predators. It means policies that ignore incentives, costs and unequal exposure will eventually lose legitimacy.
Recent Forest Service research on wolves and livestock makes a related point. Interviews across six national forests found that conflict is embedded in economic pressures, divided authority, difficult terrain and sharply different public values. Managers often become intermediaries among livestock producers, wildlife agencies and other stakeholders. Perceived institutional responsiveness can matter alongside the technical effectiveness of conflict-reduction tools.
Wildlife management is therefore ecological, institutional and ethical.
It asks what people owe wild animals. It must also ask what institutions owe the people expected to live with them.
A threshold is a promise to use evidence
“Protect predators” is not a management plan.
Neither is “control predators.”
Both may express legitimate values. Neither tells an official, rancher or community what to do when circumstances change.
A credible policy needs room for tolerance, prevention, deterrence, restricted access, compensation and, in narrowly defined circumstances, removal. It also needs a transparent explanation of when management will move from one response to the next.
The Frontiers proposal is not the final word. Its thresholds need local calibration. Its Toronto evidence is observational. Its strongest recommendations—especially immediate removal after the highest-level breach—deserve continued scrutiny against other research and real-world outcomes.
But it asks the right institutional question:
What evidence will change the decision?
Coexistence remains a worthy goal only if it includes a responsible answer.
Related Article:
Chris Gill
External Sources:
Primary:
Dennis L. Murray and Emily Foucault and Keith Patton
Supporting institutional context:
Sylvia Kantor
This Pitchstone Waters institutional editorial was developed with AI assistance under human direction. It draws on Chris Gill’s published work, documented experience and the Pitchstone Waters archive, but it is not presented as a new article written personally by Chris Gill.
