When the map moves under the species: extinction models miss the climate shift

On 11 July 2026, a team at the University of Göttingen published a finding that puts a number on something conservation biologists have suspected for years: the standard tools used to predict climate-driven extinction systematically undercount species on the move.

The research, summarised by Science X, examined what happens when climatic conditions shift across landscapes rather than simply degrading in place. When a habitat's climate envelope slides poleward or uphill, the species adapted to it face a choice: track the conditions, adapt in place, or die out. Most predictive models, the team found, are calibrated for the second and third of those options and quietly treat the first as noise. The result is a structural undercount of risk, with consequences for protected-area design, species recovery budgets, and the global accounting frameworks that drive conservation funding.

What the models miss

The Göttingen team's central claim is a methodological one. Standard correlative models, the workhorses of extinction-risk projection, work by matching a species to its current climatic envelope and asking what happens when that envelope shrinks. The match is statistical, drawn from observed presence points and gridded climate data.

That approach handles two scenarios cleanly. A species stuck on a mountaintop as warming pushes its climate zone upward will, in many cases, run out of mountain. The model reads this and flags the risk. The same is true for species confined to low-lying islands or coastal flats as sea levels rise and habitat compresses. These are the cases where correlative models have earned their reputation.

What they handle less well is the case of a temperate-zone species whose entire climatic envelope slides several hundred kilometres north over a few decades. A boreal bird, a cold-water fish, an Arctic-edge mammal: the habitat they were adapted to is still there, somewhere, just not where they are. The correlative model can register habitat loss for the population at the trailing edge of the range and habitat gain at the leading edge. What it cannot easily capture is the cost of the journey itself: the mortality, the failed breeding seasons, the corridors blocked by agriculture, roads, and cities. A species can be expanding its range on paper while its actual population is collapsing.

The Göttingen finding is that this gap matters at scale. When the team re-ran projections accounting for movement costs, the headline extinction figures rose. The researchers did not, in the materials reviewed, publish a single global percentage; the significance is in the direction and the mechanism, not in a tidy headline number.

The counter-narrative

There is a counter-position worth taking seriously. Some ecologists argue that range-shift dynamics are already baked into conservation practice, that protected-area networks are designed with connectivity in mind, and that the existing literature has long flagged dispersal limitation as a known caveat. From this vantage point, the Göttingen paper is an important refinement, not a correction. The models were never claimed to be perfect; the field has always known that real animals face friction that statistical envelopes do not.

A second counter-narrative runs along different lines: that climate-driven extinction risk is a distraction from the more tractable, proximate drivers of biodiversity loss, habitat destruction, over-harvesting, pollution. On this reading, fine-tuning extinction projections for shifting envelopes is academic in the worst sense, while the bulldozers keep running. The Göttingen team would likely respond that their work is precisely about how proximate drivers interact with climatic ones: a species whose range is shifting into an already-fragmented landscape faces a compounded risk that the headline models cannot see.

The structural picture

The larger pattern here is the recurring lag between the tools institutions use to measure environmental risk and the dynamics those tools are supposed to capture. The same lag appears in flood modelling, in wildfire projections, in fisheries stock assessments. In each case, the dominant frameworks were built for a more stationary world than the one they are now being asked to describe.

For biodiversity, the consequences are concrete. National biodiversity strategies, the targets embedded in international agreements, and the prioritisation lists used by conservation funders all draw on projections of the kind the Göttingen team is interrogating. If those projections systematically undercount risk for species on the move, the strategies built on them will under-resource the very interventions, wildlife corridors, assisted migration policy, transboundary reserve design, that the moment demands. The lag is not a technical curiosity. It is a planning horizon.

What to watch next

Three developments would shift this story. First, whether the Göttingen framework is adopted by the major intergovernmental biodiversity assessments, where it would alter headline risk numbers cited in policy documents. Second, whether conservation funders begin to reallocate from static protected-area management toward connectivity and corridor investment, a redirection that would carry political weight in rural regions where land use would have to change. Third, how the next generation of species-specific recovery plans handles the species whose ranges are already in motion, whether the planners treat the corridor as a line on a map or as a budgeted, policed piece of territory.

There is also an honest unknown. The Science X summary reviewed here does not specify which taxonomic groups or geographic regions were most affected in the Göttingen reanalysis, and the public materials do not yet break out results by species. Without that detail, the practical question for any given conservation programme, should we redesign the corridor, the reserve, or the recovery plan?, remains harder to answer than the headline finding suggests.

This piece draws on a single peer-reviewed finding as summarised in a Science X news brief dated 11 July 2026. The full Göttingen paper would be the appropriate next stop for any reader sizing the implications for a specific species or region.

Wire provenance

This editorial synthesis draws on the following public wire/social posts:

• https://en.wikipedia.org/wiki/Climate_change_and_biodiversity
• https://en.wikipedia.org/wiki/Extinction_risk_from_climate_change
• https://en.wikipedia.org/wiki/Correlative_species_distribution_model