Discontinuous stochastic forcing in Greenland ice core data

Keno Riechers, Andreas Morr, Klaus Lehnertz, Pedro G. Lind, Niklas Boers, Dirk Witthaut, Leonardo Rydin Gorjão

TL;DR

By applying stochastic analysis to Greenland ice core data, researchers have discovered a fundamental difference in how temperature and atmospheric circulation "tip" during abrupt climate shifts. While atmospheric circulation behaves like a classic bistable system, temperature shifts appear to be driven by discontinuous, "jumpy" noise acting on a system with only one stable state—challenging long-held assumptions about climate bistability.

What Problem Does This Paper Address?

The Earth's climate history is punctuated by Dansgaard–Oeschger (DO) events—abrupt shifts where Greenland temperatures soared by up to 16°C in mere decades. For years, the prevailing theory has been bistability: the idea that the North Atlantic climate has two stable "wells" (cold and mild) and occasionally gets pushed from one to the other.

However, the exact physical mechanisms behind these transitions remain a subject of intense debate. Specifically, are these shifts caused by the underlying structure of the climate system (the "drift"), or are they triggered by the nature of the random fluctuations (the "noise") hitting the system? This paper seeks to move beyond simple visual correlations to determine the mathematical "DNA" of these transitions using two primary proxies: $\delta^{18}$O (temperature) and dust concentration (atmospheric circulation).

Methods

The authors treat the ice core records as Markovian stochastic processes. To dissect the dynamics, they employ the Kramers–Moyal (KM) expansion.

  • Drift and Diffusion: They estimate the first two KM coefficients ($D_1$ and $D_2$), which represent the deterministic "pull" of the system toward equilibrium and the intensity of the random noise, respectively.
  • Beyond Gaussian Noise: Most climate models assume "white noise" (Gaussian). To test this, the authors look at higher-order KM coefficients ($D_4$). If $D_4$ is significantly non-zero, it suggests the noise is discontinuous—meaning it contains "jumps" rather than just smooth, random wiggles.
  • Diagnostic Ratios: They use specific metrics called the $\Theta$-ratio and $Q$-ratio to distinguish between a standard Langevin process (continuous motion) and a jump-diffusion process (motion with sudden spikes).

Datasets

The study utilizes high-resolution data from the North Greenland Ice Core Project (NGRIP):

  • $\delta^{18}$O Record: A proxy for local surface temperature.
  • Dust Concentration: A proxy for hemispheric-scale atmospheric circulation (rescaled using a negative logarithm).
  • Timeframe: The analysis focuses on the period between 59,000 and 27,000 years ago, binned at a 5-year resolution.
  • Preprocessing: The data was detrended against global mean surface temperature reconstructions to remove slow, background climate changes.

Key Insights and Ideas

  • The Monostability Surprise: Surprisingly, the temperature proxy ($\delta^{18}$O) shows a monostable drift. This means that, mathematically, the system doesn't have two separate stable states. Instead, the "tipping" behavior is likely forced by the nature of the noise itself.
  • Dust is Truly Bistable: In contrast, the dust record—representing atmospheric circulation—exhibits clear bistability. It has two distinct stable equilibria that match the cold (stadial) and mild (interstadial) periods.
  • "Jumpy" Temperature Noise: The temperature record shows a high $D_4/D_2$ ratio ($\gtrsim 0.3$), providing strong evidence that the "noise" driving temperature is non-Gaussian and discontinuous. This suggests that external "shocks" to the system are more important for temperature shifts than previously thought.
  • Decoupled Dynamics: Even though temperature and dust are highly correlated, their underlying mathematical structures are qualitatively different. This implies that atmospheric circulation and temperature might be responding to different physical drivers during DO events.

Key Results

  • Drift Functions: The estimated drift for dust has two stable points (zero-crossings with negative slope), while the drift for $\delta^{18}$O has only one.
  • Noise Characteristics: The ratio of the fourth to the second KM coefficient is ten times larger for temperature than for dust, indicating a much higher presence of discontinuous "jump" components in the temperature forcing.
  • Scaling Behaviors: The $\Theta$-ratio for dust follows a $1/\tau$ scaling, which is a hallmark of jump-diffusion processes, even though its drift is bistable.

Limitations and Future Work

The authors acknowledge that their analysis is one-dimensional. In reality, the climate is a high-dimensional, coupled system. Reducing it to a single variable might oversimplify complex feedbacks.

Furthermore, while the KM framework is powerful, it cannot directly model $\alpha$-stable noise (noise with infinite variance), which some researchers believe is a better fit for climate "shocks."

Future Work: The findings highlight a need to investigate why the temperature noise is so discontinuous. Understanding whether modern anthropogenic warming is subject to similar "jumpy" stochastic forcing is critical for improving early warning signals for future climate tipping points.

Summary updated

Abstract

Paleoclimate proxy records from Greenland ice cores, archiving e.g. $$\delta ^{18}$$ O as a proxy for surface temperature, show that sudden climatic shifts called Dansgaard–Oeschger events (DO) occurred repeatedly during the last glacial interval. They comprised substantial warming of the Arctic region from cold to milder conditions. Concomitant abrupt changes in the dust concentrations of the same ice cores suggest that sudden reorganisations of the hemispheric-scale atmospheric circulation have accompanied the warming events. Genuine bistability of the North Atlantic climate system is commonly hypothesised to explain the existence of stadial (cold) and interstadial (milder) periods in Greenland. However, the physical mechanisms that drove abrupt transitions from the stadial to the interstadial state, and more gradual yet still abrupt reverse transitions, remain debated. Here, we conduct a one-dimensional data-driven analysis of the Greenland temperature and atmospheric circulation proxies under the purview of stochastic processes. We take the Kramers–Moyal equation to estimate each proxy’s drift and diffusion terms within a Markovian model framework. We then assess noise contributions beyond Gaussian white noise. The resulting stochastic differential equation (SDE) models feature a monostable drift for the Greenland temperature proxy and a bistable one for the atmospheric circulation proxy. Indicators of discontinuity in stochastic processes suggest to include higher-order terms of the Kramers–Moyal equation when modelling the Greenland temperature proxy’s evolution. This constitutes a qualitative difference in the characteristics of the two time series, which should be further investigated from the standpoint of climate dynamics.
BibTeX
@inproceedings{riechers2025discontinuous,
  author = {Keno Riechers and Andreas Morr and Klaus Lehnertz and Pedro G. Lind and Niklas Boers and Dirk Witthaut and Leonardo Rydin Gorjão},
  title = {{Discontinuous stochastic forcing in Greenland ice core data}},
  booktitle = {Climate},
  year = {2025},
  doi = {10.1007/s00382-025-07880-9},
}

Loading related content...

Figure