Representing slow, internally driven ocean evolution

Develop ocean climate emulators that represent the slow, internally driven evolution of the ocean interior and capture long-term trends in deep-ocean processes, including the accumulation of small trends over multidecadal and centennial rollouts.

Background

The autoregressive ocean emulators reproduce many upper-ocean responses to out-of-sample midHolocene boundary forcings but fail to reproduce the slow, internally driven evolution of the ocean interior. In particular, they do not accumulate the deep-ocean cooling and heat-transport changes present in the CESM2 reference response.

This limitation constrains the reliability of emulators for long-term climate perturbation experiments, especially in the deeper ocean and at high latitudes, where the response depends on internally mediated circulation and the accumulation of small forcing imbalances.

References

Capturing the long-term trends in deep-ocean processes remains an unsolved challenge, and we do not propose a solution in this work .

— Paleoclimate Boundary Conditions as an Out-of-Sample Test for the Forced Response of Ocean Climate Emulators  (2608.13494 - Subel et al., 13 Aug 2026) in Section 3, Evaluating Out-of-Sample Generalization; Section 4, Forced Response between piControl and midHolocene Climates

While groups have begun to evaluate against dynamical benchmarks for the atmosphere , the ability of emulators to capture nonlinear ocean feedbacks remains an open question essential for climate application.

— Paleoclimate Boundary Conditions as an Out-of-Sample Test for the Forced Response of Ocean Climate Emulators  (2608.13494 - Subel et al., 13 Aug 2026) in Section 5, Conclusion