For two decades, the Iberian lynx and the black kite shared the same patch of Spanish wetland and mostly ignored each other. Between 1993 and 2012, lynx destroyed 13 of 904 monitored kite nests, about 1 percent. Then, in 2013, a rabbit disease arrived, and drought took care of the rest. In the worst years that followed, lynx wiped out up to 46 percent of kite nestlings close to fledging, according to a study published on October 21 in Nature.

Twenty years of neutrality

The researchers, led by Fabrizio Sergio and Julio Blas at the Doñana Biological Station (CSIC) in Seville, watched a 10-square-kilometre plot of Doñana National Park for three decades. The plot supported 4 to 6 adult lynx and 105 to 223 pairs of black kites, which make up 82 to 90 percent of the local raptor community, the paper says. That makes the lynx the main nocturnal predator and the kite the main diurnal one.

Both lean on the same food. The European rabbit is a keystone prey for more than 30 Mediterranean predators, and lynx diets are dominated by it, 70 to 90 percent. When rabbits are scarce, lynx switch to waterbirds from the marshes, mainly ducks. For decades, that arrangement held. Lynx were a rabbit specialist, kites an opportunist, and neither troubled the other much.

A disease, then a drought

In 2013, a new variant of rabbit haemorrhagic disease, which originated at a French breeding farm and spread across the Iberian Peninsula, cut the local rabbit population by more than 70 percent. The paper treats this as an anthropogenic change in its own right. On its own it did not trigger attacks on kites.

Drought did. Across 1,624 kite nests monitored over 25 years, lynx predation increased exponentially as rabbits declined, but only in drought years, when the marshes dried out and duck numbers fell, the authors report. In wetter years it stayed negligible, even when rabbits were scarce, presumably because lynx could switch to ducks. Predation took off in drought years only once rabbit density fell below about one per hectare, which earlier work identified as the minimum for lynx survival and stability.

The authors call drought the switch that turned the interaction on and off. They define a drought year as one in which less than 20 percent of the maximum flooded area was inundated in February and March. Nine of the study years met that definition, including 2012, 2016, 2018, 2019, 2022 and 2023.

How lynx picked their nests

The data suggest the killing was not random. Lynx climbed trees up to 12 metres to reach well-feathered nestlings, and the nests they raided did not differ from other nests in height, brood size or parental age. What stood out was the ground beneath them. Nests that lynx predated had more than three times as many rabbit remains below, while the remains inside the nests were no different.

The authors' preferred reading is that lynx used rabbit remains on the ground as a signal that a competitor was feeding on their key prey, and removed it. They test and weigh an alternative, that lynx learned the remains meant a bigger reward, and note that broods under rabbit remains were not larger. They say further manipulative experiments are needed to confirm the competitor-removal idea.

What it cost the kites

Kite parents that lost broods to lynx had a lifetime reproductive success 27.7 percent lower than comparable birds that did not, and could not make up the loss in later years. They were four times more likely to move to another territory the following year, and abandoned territories were three times more likely to stay empty, probably because the remains of killed fledglings stayed visible for more than a month and deterred new recruits.

Kite numbers fell with lynx predation in the preceding year. In turn, little owls, a small insect-eating species that kites sometimes prey on, became more numerous in years after heavier predation. That ripple rests on only 19 years of data and should be read as a hint of cascading effects, not a proven chain.

Two stressors, one hidden weakness

The deeper message concerns the order of events. The disease weakened the system quietly, the authors argue, and the damage became visible only when drought arrived, by which time management might be ineffective. Drought alone had not caused this in decades. Disease alone had not either. Together, they created an interaction between two predators that behaved as if a new top predator had invaded the area.

That is a warning about forecasting. Models usually project how a species responds to warming or drying in isolation. This study shows that a stressor can change how species relate to each other, and that the change can stay hidden until another stressor lands. It sits alongside our earlier report on how most ecosystem services rise with each added species, another reminder that the roles species play depend on who else is in the web.

What it does not show, and what comes next

The study rests on a single 10-square-kilometre plot, and most of its evidence is correlational. The authors say they lack recent lynx diet data, and the competitor-removal explanation is not experimentally tested. The paper does not claim that either species is endangered as a whole. The finding is about how interactions can be rewired in one well-studied place.

It matters because Doñana is warming. Spring temperatures there have risen by more than 0.5°C per decade, and models project more frequent droughts across the region, the paper notes. The climate pathway behind that is laid out in UNEP's new emissions gap report, which puts current policies on course for 2.8°C by 2100.

The authors offer one practical idea. A manager who wanted to preserve abundant kite populations could provide supplementary food to lynx in drought years to dampen the spikes in predation. It is a suggestion, not a trial result. The larger point is that conservation plans built around single threats may miss the one that arrives second.

---