Causality from long-lived radiative forcings to the climate trend

Area of Science:

• Climate Science
• Statistical Modeling
• Time Series Analysis

Background:

• Understanding the causal relationship between radiative forcings and climate trends is crucial for climate change research.
• Previous analyses often rely on complex models, necessitating model-free approaches.

Purpose of the Study:

• To provide model-free statistical evidence for Granger causality from long-lived radiative forcings (LLRFs) to climate trends (CT).
• To investigate the role of trend breaks in establishing causality.

Main Methods:

• A purely statistical, observations-based, model-free analysis was employed.
• The method relies on identifying locally ordered breaks in the trend functions of LLRFs and CT.
• Empirical analysis used global surface temperature series and aggregate LLRF data (CO2, N2O, CFCs).

Main Results:

• Empirical evidence supports the conditions for Granger causality: LLRF trend breaks precede CT breaks.
• Slope changes in both LLRF and CT trends were of the same sign.
• The findings hold even when the noise function in LLRF is negligible.

Conclusions:

• The study provides statistical evidence for Granger causality from LLRFs to CT.
• The presence of broken trends can influence causality conclusions in standard methods.
• Model-free analysis offers a robust approach to understanding climate forcings and trends.