Deriving dynamical models from paleoclimatic records: application to glacial millennial-scale climate variability

Area of Science:

• Paleoclimatology
• Dynamical Systems Theory
• Statistical Inference

Background:

• Ice-core data provides invaluable records of past climate conditions.
• Dynamical models are crucial for understanding climate variability and predicting future changes.
• Integrating theoretical models with paleoclimatic data remains a challenge.

Purpose of the Study:

• To propose a systematic method for deriving simple nonlinear dynamical models from ice-core data.
• To provide a tool for integrating climate models and theories with paleoclimatic evidence.
• To apply this method to a North Greenland ice-core record for the glacial period.

Main Methods:

• Utilizing the unscented Kalman filter, a nonlinear extension of the standard Kalman filter.
• Adopting an abstract conceptual model of stochastically driven motion in a potential with two distinct states.
• Estimating model parameters, including potential shape and noise level, from ice-core data.

Main Results:

• A nonlinear dynamical model was successfully derived from the North Greenland ice-core record.
• The estimated potential for the glacial period (70-20 ky BP) was found to be asymmetric and nearly degenerate.
• The model identified a deep potential well for cold stadial states and a shallow well for warm interstadial states.

Conclusions:

• The proposed method effectively integrates paleoclimatic data with nonlinear dynamical models.
• The derived model provides insights into the distinct climate states during the last glacial period.
• This approach offers a robust framework for analyzing past climate dynamics using ice-core records.