In a Seabird Die-Off, Which Birds Died Matters More Than How Many

Since 2021 the H5N1 lineage circulating in northwestern Europe has killed at least 1.4 million wild birds and affected more than a hundred seabird species. Those totals are the figure everyone reports. They are also, on their own, close to uninformative about what happens next.

For a species that breeds slowly and lives a long time, a population's future depends less on how many individuals died than on which ones. A team working across the Netherlands, Belgium, Germany and Denmark had the data to ask that question of the Sandwich tern, which lost over 20,500 adults in 2022, more than 17 percent of its regional population.

Why it matters: Seabirds mature late and raise few chicks, so adult survival is what holds their numbers up. A die-off that falls evenly across ages is a setback. One concentrated in experienced breeders removes the part of the population that contributes most to growth.

What the ringing archive made possible

The analysis rests on decades of routine work. More than 50,000 terns had been ringed as chicks since 1995, which means their exact ages were known. Resightings across those years allowed the authors to estimate survival rates before the outbreak and reconstruct how many birds of each age should have been alive when it began. That reconstruction is the denominator, without which a pile of dead birds cannot be turned into a mortality rate.

The numerator came from 832 ringed birds of known age recovered dead during the outbreak, against a long-run average of about five ring recoveries a year over the preceding 27 years.

The age gradient is steep. In the southern Netherlands and Belgium, predicted mortality ran from 6.7 to 10.5 percent among four-year-old birds to 33.2 to 49.5 percent among 28-year-olds, with a similar pattern further north. Each additional year of age raised the odds of dying by roughly 8 to 9 percent depending on the model.

Sex showed nothing. Of 243 carcasses sexed, 123 were female and 120 male, which is as close to even as the sample allows.

As Courtens and colleagues put it in Nature Communications, HPAI did not act as a random mortality source, but selectively removed individuals that likely contribute substantially to population growth.

Ruling out the boring explanation

The obvious alternative is exposure rather than susceptibility. If older birds arrive at the colony earlier, they simply spend more time in the outbreak. The authors checked. First-time breeders did arrive later than established ones, but every age class was present well before the first cases appeared, with median arrival dates more than a month ahead of the outbreak. Timing does not explain the gradient.

Nest position was considered too. Young birds tend to nest at colony edges, but Sandwich terns pack in at up to seven nests per square metre, and transmission at that density is unlikely to respect the difference.

Which leaves an uncomfortable gap

That points toward older birds being more susceptible, presumably through declining immune function. The paper is notably unwilling to claim it. Experimental infections of domestic waterbirds have sometimes shown the opposite, with lower mortality in older individuals, and a study of common terns found no age-related immune decline at all. The authors state plainly that the mechanism remains poorly understood and warrants further investigation.

So the pattern is well established and the reason for it is not. That is a more useful place to stop than a speculative mechanism would have been.

The finding also has a use beyond this species. Colonial seabirds return to the same sites year after year and are marked in large numbers by long-running ringing schemes, which makes them unusually legible: few wild populations allow anyone to know the exact age of hundreds of dead individuals and how many of that age were alive beforehand. The authors suggest such populations could serve as sentinels for disease dynamics, particularly if ringing data were paired with blood sampling to separate who was exposed from who died.

What the study can't say yet

The survival method has a known blind spot: it cannot separate a bird that died from one that permanently moved elsewhere. The authors address this by running a second model that counts recoveries anywhere in Europe, and they note that both approaches probably underestimate true mortality, since many carcasses are never found or reported.

The link from age-biased death to slower recovery is an inference, not an observation. In fact the early numbers cut against it: mortality in the Dutch and Belgian colonies was put at 26.7 percent, while breeding pairs fell by 21.4 percent the following year. Non-breeding adults and immigrants appear to have partly filled the gap, which the authors read as short-term resilience that may come at the cost of the reserve available for the next outbreak.

This is also a single species in one region during one event. Whether age-biased mortality is a general feature of these outbreaks in seabirds needs the same kind of long-term marking data from elsewhere, which mostly does not exist.

Quick questions

Why does losing older birds matter more? In long-lived seabirds, experienced breeders raise more young. Removing them lowers the population's reproductive output beyond the headcount lost.

Could the older birds just have been unlucky? The age effect held across model variations and across thousands of simulations testing sensitivity to the survival estimates, staying positive and significant throughout.

What's the one-line takeaway? Bird flu killed Sandwich terns in proportion to their age, roughly 5 to 10 percent of young breeders against more than 40 percent of the oldest, and the reason older birds were more vulnerable is still unknown.