Quantifying age and model uncertainties in palaeoclimate data and dynamical climate models with a joint inferential...

Area of Science:

• Paleoclimatology
• Climate Modeling
• Geochronology
• Time Series Analysis

Background:

• Paleoclimate studies utilize natural archives like deep sea cores.
• Conventional research involves multi-stage analyses: age modeling followed by climate model calibration.
• This sequential approach can propagate and amplify uncertainties.

Purpose of the Study:

• To demonstrate the potential for misleading conclusions in multi-stage paleoclimate analyses.
• To develop and present a joint inferential approach for integrated climate reconstruction, model calibration, and age estimation.
• To provide a more robust methodology for analyzing paleoclimate data.

Main Methods:

• A joint inferential framework was developed to simultaneously estimate age models and calibrate dynamical climate models.
• The approach was applied to two benthic sediment cores covering the past 780,000 years.
• A comparison was made between the joint analysis and a simulated multi-stage analysis using posterior age samples.

Main Results:

• Multi-stage analyses, when mimicking standard practice, yielded substantially different model selection conclusions depending on age samples.
• These conclusions contrasted significantly with those obtained from the joint inferential analysis.
• The study highlights the inadequacy of sequential analyses in the presence of uncertainty.

Conclusions:

• Multi-stage analyses are insufficient for robust paleoclimate research due to inherent uncertainties.
• A full joint inferential analysis is essential for drawing sound conclusions in climate reconstruction and model calibration.
• The proposed joint approach offers a more reliable method for understanding past climate dynamics.