# Inferring directed climatic interactions with renormalized partial directed coherence and directed partial correlation

## TL;DR

This paper applies renormalized partial directed coherence and directed partial correlation to climatological time series, demonstrating their effectiveness in analyzing ENSO-Monsoon interactions and exploring complex air-sea interactions in the SACZ.

## Contribution

It introduces the application of rPDC and DPC methodologies to climate data, expanding their use beyond neurophysiological signals.

## Key findings

- Results align with current climate models for ENSO-Monsoon interactions.
- Less clear results for SACZ air-sea interactions, indicating need for further research.
- Methodologies show promise for analyzing complex climate interactions.

## Abstract

Inferring interactions between processes promises deeper insight into mechanisms underlying network phenomena. Renormalised partial directed coherence (rPDC) is a frequency-domain representation of the concept of Granger causality while directed partial correlation (DPC) is an alternative approach for quantifying Granger causality in the time domain. Both methodologies have been successfully applied to neurophysiological signals for detecting directed relationships. This paper introduces their application to climatological time series. We first discuss the application to ENSO -- Monsoon interaction, and then apply the methodologies to the more challenging air-sea interaction in the South Atlantic Convergence Zone (SACZ). While in the first case the results obtained are fully consistent with present knowledge in climate modeling, in the second case the results are, as expected, less clear, and to fully elucidate the SACZ air-sea interaction, further investigations on the specificity and sensitivity of these methodologies are needed.