The Siberian crane is the only one of the world's 15 crane species listed as critically endangered, and it looked like a clear win: 140 birds in 1981, an estimated 5,616 today. Yet 82% of the wintering flock now gathers on just two farmland sites beside Poyang Lake, among rice paddies and lotus ponds, where the grain it eats carries antibiotic residues and its droppings hold an average of 62 microplastic particles per gram. A perspective in The Innovation argues that such cases expose a flaw in how conservation keeps score, and proposes judging recovery by population health rather than headcount.
Numbers have ruled for a reason: they are cheap, comparable and easy to explain. The WWF Living Planet Report found vertebrate populations falling nearly two-thirds in under half a century, and recovery targets followed, defining success as more individuals. Headcounts, though, say little about functional integrity, genetic diversity or habitat connectivity. The IUCN's Green Status of Species, launched in 2021, was meant to widen the picture, but ambiguous baselines, heavy data demands and patchy taxonomic coverage have kept it to a small subset of species. Nearly 70% of emerging human infectious diseases originate in wild animals, giving the argument a public-health edge: watching a flock is also outbreak surveillance.
Population health comes from public health, where it means the health of a group measured by status indicators and their spread within it. Transplanted into conservation, it becomes concrete: physiology (body condition, stress hormones, blood chemistry), demography (birth rates, juvenile survival, sex ratio, age structure), genetics (heterozygosity, inbreeding, immune-gene diversity), disease (pathogens, parasites, gut microbiome), environment (food quality, pollutants, fragmentation) and behaviour (activity rhythms, migration, social structure). None substitutes for the others: birth rates follow physiology and disease burden, while a single stressor can move behaviour, hormones and genes at once.
The crane shows why the six must be read together. Poyang Lake's wetlands have degraded under pollution, climate-driven drought and hydrological change; earlier and longer low water collapsed the Vallisneria beds whose tubers the birds depend on, pushing them into farmland. In a December 2024 survey, water near one site held azithromycin at 882.6 ng/g and erythromycin at 1,182.5 ng/g, against a commonly cited 100 ng/g threshold; rice, lotus root and nutgrass carried residues of 13.4, 9.2 and 21.2 ng/g. Faecal virome work found no highly pathogenic viruses in the cranes, but detected Orthoavulavirus javaense and a SARS-related coronavirus in greater white-fronted geese sharing the fields. Elsewhere, H5N1 has killed more than 17% of adult Sandwich terns in dense colonies and nearly all chicks.
The framework is deliberately provisional. Six dimensions demand far more data than a headcount, so the authors would start with critically endangered species, populations behaving abnormally, or those swinging sharply year to year. Filling the gaps means large-scale sampling with RNA virome and population resequencing, and pairing field observation with satellite tracking. A bird that has recovered on paper, they argue, is not yet a bird that can survive the next outbreak.
