The Threshold of Tolerability

Since December 2023, Switzerland has been experimenting with a new approach to wolf management: proactive culling. Even before a pack reaches the thresholds that would trigger a reactive response, wolves can now be culled preventively to reduce future conflicts.

Three culling periods later, one crucial question remains unanswered: At what level of predation will we consider coexistence to be working?

In other words: what is the “tolerable threshold”? This question may seem theoretical. On the contrary, it is crucial. For without a measurable objective, how can we assess the effectiveness of a policy? And above all, how far should culling go when even packs causing very little damage become targets?

Anniviers-Réchy: When a Low-Conflict Pack Seems to Become a Target Despite Everything

The Anniviers-Réchy pack perfectly illustrates this paradox. Just like the packs in Les Toules, Nendaz-Isérables, and Les Hauts-Forts before it.

During the 2025–2026 proactive culling program, the canton of Valais requested a partial cull of the Anniviers-Réchy pack. This allowed for the removal of two-thirds of the year’s pups. Two wolf pups were ultimately killed in September 2025. It is also worth noting that this pack was monitored using a radio collar, which had been fitted to the breeding female F225 since early 2025. The authorities thus obtained a significant amount of additional data, which shows that this pack posed little threat to livestock.

In fact, when reviewing the pack’s movements since the arrival of the new breeding pair between 2022 and 2023, it becomes clear that Anniviers-Réchy does not exhibit the profile one would intuitively associate with a (highly) problematic pack.

On the contrary, the data at our disposal show very low predation on livestock in protected areas:

- one attack attributed to a subadult in 2025, where the protective fencing had non-compliant spacing between the wires at mid-height (40 cm instead of 20 cm), reducing its effectiveness, in addition to human monitoring occurring only once a week.

- two attacks in 2026, resulting in two sheep killed across two incidents. Here too, the protective measures appear to have had flaws, based on how they were operated and their location (carcasses were found at great distances from the night pen and the guard dogs, which are stationed there at night). To date, no DNA clearly linking one or both members of the breeding pair has been reported, and neither member has ever been identified on domestic animals since the pack’s establishment in 2022/2023. Added to this is the fact that the locations are in a peripheral area, where the presence of a buffer zone or the passage of dispersing wolves cannot be entirely ruled out.

This obviously does not mean that this pack will never prey on livestock. It may do so tomorrow. It could even cause more damage in 2027; it is currently impossible to say for certain.

But a possibility is not a prediction.

For preventive action to be based on a scientific risk assessment, it would first be necessary to identify the objective factors that would allow us to conclude that a pack currently posing little risk of conflict might become problematic in the future.

Otherwise, all packs could be considered future threats, since none can guarantee that it will never attack a domestic animal.

Not all predation incidents tell the same story

This is precisely where raw data reveals its limitations.

A pack responsible for four or five losses spread across several farms and multiple incidents over the course of a year does not necessarily exhibit the same behavior as a pack that regularly returns to the same herds over a period of several months and systematically circumvents all effective protective measures.

In both cases, a cantonal statistic might show five victims as of early August 2026. Biologically and in terms of management, however, the situations are radically different.

In the first case, we may be dealing with a pack whose diet consists almost entirely of wild prey and which causes only very occasional and minor damage to domestic animals.

In the second case, the repetition, spatial and temporal concentration of attacks, and the repeated circumvention of protective measures may indicate a problem requiring a targeted response. And this is entirely legitimate and understandable, as in the case of the Nanz and Ausgtbord packs, which have already totaled 25 attacks and 51 animals killed in protected areas since April 2026…

This is precisely why wolf management should not focus solely on the number of victims. It should analyze predation profiles: frequency of attacks, recurrence, concentration on certain farms, the actual effectiveness of protective measures, the persistence of the behavior over time, and the individuals involved.

Coexistence without any predation is not true coexistence

This distinction leads to a more uncomfortable question. Are we willing to accept that a wolf pack in a given territory may occasionally cause some losses?

If the answer is no, then we must state that clearly. Because coexistence between large predators and livestock farming does not mean the total elimination of conflict. It consists of keeping conflict at a tolerable level, minimizing damage as much as possible through protective measures, and intervening in a highly targeted manner—based on in-depth and objective analyses—when certain situations become truly problematic.

But this “tolerable” level has, in fact, never been clearly defined. How many losses would be compatible with successful coexistence? Five per pack per year? Ten? Twenty?

Should we instead consider the number of attacks? The proportion of herds affected? The frequency with which protective measures are circumvented? The economic cost? The recurrence of damage on the same farms?

The answer probably cannot be a single figure. But the total absence of acceptability criteria poses a major problem. For if no threshold for success is defined, any predation can become evidence that the current policy does not go far enough.

When damage decreases, culling can be presented as effective. When it increases, this can justify additional culling. And when a pack causes very little damage, it can still be culled on the grounds that it might cause more in the future.

In such a system, what situation would then allow us to conclude: this pack is compatible with coexistence; let’s let it live and observe how it evolves?

Living systems do not produce perfect curves

We must also accept a basic reality: predation rates will always fluctuate. Simply because nature has never been—and never will be—an accounting ledger.

Weather, the availability of wild prey, pack composition, dispersal, reproduction, the consequences of hunting or poaching, the location of herds, the duration of summer grazing, livestock behavior, the quality of fences, guard dogs, human presence, and a host of other factors are constantly changing.

An annual increase in damage therefore does not automatically indicate the failure of a policy, any more than a one-time decrease automatically indicates its success.

This is particularly important when evaluating proactive culling. In Valais, losses among herds considered protected have evolved as follows, according to the data we have compiled (sources: SCPF):

2022: 117 victims

2023: 104 victims

2024: 161 victims

2025: 126 victims

Proactive culling began in December 1st, 2023.

These figures do not allow us to conclude that culling increases damage. Such a conclusion would be scientifically unsound given the considerable number of variables involved. But for the same reasons, they do not allow us to assert that proactive culling reduces the damage either. Our analyses of the territories of the packs that were actually targeted show, moreover, comparable fluctuations so far.

Again, correlation does not imply causation. And that is precisely the problem: where is the scientific assessment that would allow us to measure the specific effect of the culling today? It is currently underway, but as noted in our previous article “link,” it is being conducted by the KORA Foundation, which is heavily subsidized by the Federal Office for the Environment (FOEN)—the agency that implemented this proactive culling program in 2023 under the leadership of federal councilor Albert Rösti.

Regulating a stable pack is not necessarily without consequences

Another aspect deserves closer attention: the social structure of the wolf. A pack that has been established for several years constitutes a system whose functioning we are gradually coming to understand. We can observe its territory, its breeding pair, better understand its movements, the management of shared areas with neighboring packs, and its history of predation. This is possible when data monitoring and analysis are conducted properly.

Intervening in this structure is therefore not necessarily neutral. Our monitoring in Valais has notably documented very rapid replacements of breeding males following their death, with an average time of between 3 and 8 weeks. This is because no pack has ever been completely culled in Valais during official culling campaigns, nor even disrupted to the point of abandoning the territory.

This means that removing a known individual does not guarantee the pack’s disappearance. It may simply lead to its replacement by a new individual whose behavior is unknown. In a highly conflict-prone pack, this disruption may potentially be part of the accepted risk of an intervention deemed necessary. But when a pack is stable and causes very little damage, the reasoning should be reversed:

What concrete benefit do we hope to gain by disrupting a system that, so far, has been functioning relatively well?

Regulation also involves uncertainties and risks. Therefore, not intervening is not the only decision that must be justified. Intervening must be justified as well. And the cantons still have considerable leeway in the event that a pack resumes its attacks, since it is possible to regulate the pack at any time during the allotted five months.

And yet, the accelerator is already pressed down

This issue becomes all the more urgent given that Switzerland is already considering expanding hunting allowances once again. Indeed, two parliamentary motions have led the Federal Council to put a new revision of the Hunting Act up for public consultation.

The first proposes allowing, under certain conditions, the culling of wolves belonging to a pack between February 1 and May 31, even though this period currently enjoys special protection precisely because it corresponds to the breeding season. The second aims to allow culling even within federal hunting reserves.

Perhaps the most troubling aspect is not even the content of these proposals. It is their timing.

The new provisions of the Hunting Act and the Hunting Ordinance did not fully take effect until February 1, 2025. During the parliamentary debates on these two motions, minority groups had specifically requested that the effects of the new provisions be studied first before proceeding further, emphasizing that their effectiveness had not yet been established.

Yet here we are, already considering the next step. We are speeding up before we have properly gauged how far we’ve come.

Before imposing further regulations, let’s define what we want to achieve

This is not to say that no wolves should ever be killed. Certain situations involving repeated predation can become extremely difficult for livestock farmers. Some packs may develop behaviors that make protection particularly complex, adding pressure to the pastoral community. A credible coexistence policy must be able to recognize these situations and respond to them.

But precisely for this reason, it must also be able to recognize situations that are working. Should a stable pack, living primarily on wild prey and causing only occasional damage, be considered a failure? If so, what level of predation would be acceptable?

And if the answer is “none,” then the debate is no longer really about population control. It is about the very possibility of sustainably conserving the wolf in livestock-rearing regions. Thus revealing, slowly but surely, the true objective that isn’t explicitly stated: eradication.

Before culling two-thirds of a litter in a pack because the pack might cause more damage the following year, before intervening during the breeding season, before expanding culls to game reserves, perhaps we should finally collectively answer a question we have so far carefully avoided:

What do we really expect from coexistence?

It will never be perfect. There will be good years and bad ones. Some packs will be unobtrusive, while others will be more confrontational. Protective measures will be effective in some situations and insufficient in others. There will be developments in methodologies and in the effectiveness of certain protective measures. There will be mistakes, adjustments, and an unavoidable element of uncertainty.

That is what it means to work with living beings. Mortality is an integral part of this and stems from various causes, many of which are never explicitly or publicly disclosed, whether by the pastoral community or the political sphere. Success, therefore, cannot be defined as the absence of all conflict or predation.

It should be our ability to keep this conflict at a tolerable level without turning every difficulty—or every hypothetical risk—into a justification for further population reduction.

And to do that, we would finally need to define what we are collectively willing to tolerate. For as long as the “threshold of what is tolerable” remains undefined, it can always be considered to have been exceeded.

A reflection at every level

Perhaps, then, it is time to open this discussion—collectively and calmly—even within our institutions. Not merely to determine how many wolves a territory can support or how much damage humans are willing to tolerate, but to ask ourselves what it truly means to accept the presence of another species alongside us.

Since the wolf’s return, we have essentially examined coexistence solely from our own perspective: our activities must be able to continue, our herds must be protected, our constraints must remain acceptable, and our interests must be preserved. These concerns are legitimate and must obviously be part of the equation. But coexistence may also require accepting that humans are not the sole measure of its success.

It raises a fundamental question: how much space are we truly willing to leave for the wild when that space requires us, too, to adapt? Because defining the “limit of what is bearable” is ultimately not just about determining what we are willing to lose. It is about deciding what we are truly willing to share.


Article : Team Wolf Mission
Photos : Ezio Giuliano Filmmaker & Wolf Mission