A massive global study — marine, freshwater, and land ecosystems, a century of surveys — found that species turnover has slowed by a third since the 1970s. Not sped up. Slowed. Despite accelerating climate change, despite warming, despite everything that should make local ecosystems churn faster, the revolving door of species replacement is grinding down.
The naive interpretation: ecosystems are more stable than we thought. The warming isn't disrupting them. Perhaps resilience is higher than the models predicted.
The correct interpretation is darker. In healthy ecosystems, species replace each other constantly through internal biological interactions — competition, predation, parasitism, chance. This is the “multiple attractors” phase predicted by Guy Bunin in 2017: even under constant environmental conditions, the community composition keeps cycling because there are multiple stable configurations and the system drifts between them. What keeps the revolving door spinning is the regional species pool — the reservoir of potential colonizers available to take the place of a departing species. Each departure opens a niche. Each colonizer fills it. The turnover rate reflects the depth of the reservoir.
Habitat degradation and biodiversity loss shrink the reservoir. Fewer potential colonizers means fewer replacements. The door slows. Locally, the ecosystem looks stable — the same species persist year after year. But the stability is a symptom of the regional pool emptying, not a sign that the local community is robust. The engine isn't idling. It's running out of fuel.
The measurement problem is severe. If you monitor a single site and observe steady species composition, you might conclude that the ecosystem is healthy. Resilient, even. You would be measuring the absence of replacement and interpreting it as the absence of need for replacement. The stability is real — the species really aren't changing. But the mechanism producing the stability has reversed: it's no longer equilibrium (a well-defended configuration) but exhaustion (no colonizers available to challenge the incumbents).
This is the general form of a diagnostic error I keep encountering. A variable (turnover rate) can indicate health or disease depending on the mechanism behind it. Low heart rate means fitness in an athlete and cardiac failure in a patient. Low code churn means mature architecture in a well-maintained project and abandoned code in a dead one. Low species turnover means resilience in a healthy ecosystem and depletion in a degraded one. The number is the same. The meaning is opposite.
You can't distinguish the two without knowing the regional species pool — the background reservoir that determines whether stability is equilibrium or exhaustion. And the regional pool is exactly what most monitoring programs don't measure. They sample the site. They count the species. They compute the trend. They never look at the reservoir that determines what the trend means.
The paper's implication for conservation is stark: stable-looking ecosystems may be the most endangered ones. The damage is invisible at the local scale because it manifests as the absence of a process (replacement) rather than the presence of one (decline). You can't see what isn't arriving. You can only see what's already there, and what's already there looks fine.
Bunin's prediction was made for theoretical ecosystems with random interactions. That it describes real-world deceleration across marine, freshwater, and terrestrial systems suggests the multiple-attractors phase is not a theoretical curiosity but a widespread dynamical regime. The revolving door is the normal state. What we're observing — the slowdown — is the door being nailed shut from outside by the shrinking of the reservoir that kept it turning.