Methods

The PD28 / Climatic Footprints project combined observational datasets, reanalysis products, and climate model simulations to investigate the drivers of variability in clouds and sea ice.
Our goal was to separate human-driven (anthropogenic) signals from naturally occurring internal variability.

🗂️ Data Used

Figure: Overview of the main datasets used in the project, including satellite-derived sea ice records, reanalysis-based cloud and climate variables, and CMIP6 climate model experiments.

Main Data Sources

• 📡 Satellite Observations: Passive microwave records for sea ice concentration.
• 📊 Reanalysis Products: ERA5 for cloud cover, surface air temperature, and circulation fields.
• 🌍 Climate Model Simulations: CMIP6 historical & pre-industrial control runs to separate forced vs. internal variability.

🔬 Analytical Methods

We applied a suite of advanced techniques to extract signals and attribute variability:

• Empirical Orthogonal Function (EOF) Analysis
  Extracted leading spatial patterns of cloud and sea ice variability, improving signal-to-noise ratio.

• Canonical Correlation Analysis (CCA)
  Linked global SST anomalies with polar cloud/sea ice fields to reveal coupled climate modes.

• Composite Analysis
  Characterized atmospheric circulation patterns during periods of extreme sea ice gain/loss.

• Convergent Cross Mapping (CCM)
  Tested causality between climate modes (e.g., AMO, IPO) and cloud/sea ice variability.

Together, these methods provided a robust separation of forced responses from natural climate variability, allowing for a clearer attribution of observed changes.