The Gateway Threshold

01The evidence

• The hysteresis approach to the overturning circulation established that the system can have two stable states, and that the threshold depends on the freshwater forcing applied, which is the framework every later estimate uses.JGR Oceans, hysteresis approach
• Freshwater export experiments show that the response of the circulation depends on where and how freshwater is delivered, and that the largest export does not always coincide with the weakest circulation.Journal of Climate

02The tension

Estimates of the freshwater threshold differ by a factor of two across models, and the spread has resisted explanation. Models also differ in whether the Bering Strait is open, how wide it is, and how much Pacific water it admits, but those differences are treated as configuration details rather than as the cause of the spread.

03The theory

The overturning circulation is sustained by a density contrast, and the contrast is maintained by the balance between freshwater input at high latitudes and salt import from elsewhere. The Arctic gateways are the valves that set how much of the freshwater budget is exported versus retained.

If the Bering Strait admits more Pacific freshwater, the Atlantic's high-latitude surface is fresher before any hosing begins, so less additional freshwater is needed to cross the threshold. The threshold should then scale with gateway transport, not be a fixed number in Sverdrups of hosing.

This turns an unexplained inter-model spread into a prediction. Two models with the same physics but different gateway configurations should have thresholds that differ by an amount calculable from their salinity budgets. If the calculation works, the spread is explained and the threshold becomes a function of geography.

The idea also makes a palaeoclimate prediction. Gateways have opened and closed over the Cenozoic, so any abrupt circulation change in the record should coincide with a gateway reconfiguration rather than with a change in freshwater forcing alone.

04What this is not

05Predictions

• P1Models identical except for Bering Strait configuration should differ in collapse threshold by a predicted amount derived from their salinity budgets.
• P2The threshold should scale monotonically with gateway transport across a suite of models.
• P3Palaeoclimate records should place abrupt circulation changes at gateway reconfigurations more often than at freshwater anomalies alone.

06What would refute it

A threshold that is invariant to gateway configuration across a controlled model ensemble would leave geography irrelevant and the spread unexplained.

07What it would change

It would explain a long-standing model disagreement, make the threshold a geographic rather than universal quantity, and give palaeoclimate abrupt-change records a common mechanism.