Multiple serial correlations in global air temperature anomaly time series

Area of Science:

• Climate Science
• Time Series Analysis
• Statistical Physics

Background:

• Serial correlations in temperature data indicate temporal consistency of climate events.
• Global surface air temperature (SAT) time series contain inherent variability crucial for climate system understanding.

Purpose of the Study:

• To investigate serial correlations within global surface air temperature (SAT) anomaly time series.
• To identify and characterize short-term, long-term, and nonlinear serial correlations in climate data.

Main Methods:

• Preprocessing SAT time series to obtain anomaly data.
• Utilizing the first-order autoregressive model (AR(1)) for short-term correlations.
• Applying detrended fluctuation analysis (DFA) for long-term correlations.
• Employing the horizontal visibility graph (HVG) algorithm and a novel parameter (Δσ) for nonlinear correlations.
• Statistical significance testing using Monte Carlo simulations.

Main Results:

• Identified global patterns of short-term correlations linked to atmospheric phenomena like Rossby waves.
• Detected long-term correlations consistent with climate variability, such as the El Niño-Southern Oscillation (ENSO).
• Demonstrated the effectiveness of HVG topological parameters and Δσ in capturing and detecting temporal correlations, respectively.

Conclusions:

• Serial correlations are a universal feature of global SAT time series.
• The identified correlation types provide insights into climate dynamics and variability.
• These findings underscore the importance of considering serial correlations in climate science analyses.