[Paper Review] The Influence of Time Series Distance Functions on Climate Networks

Leonardo N. Ferreira, Nicole Costa Resende+3 more|arXiv (Cornell University)|Feb 8, 2019

Complex Systems and Time Series AnalysisEconomics, Econometrics and Finance33 references3 citations

TL;DR

This paper evaluates 29 time series distance functions for constructing climate networks from global temperature data, demonstrating that while commonly used metrics like Pearson correlation produce similar networks, alternative distance functions reveal distinct long-distance connection patterns (teleconnections) critical for understanding climate dynamics and energy transport. The study advocates for broader use of diverse distance functions to capture varied underlying climate dynamics.

ABSTRACT

Network theory has established itself as an important tool for the analysis of complex systems such as the climate. In this context, climate networks are constructed using a spatiotemporal climate dataset and a time series distance function. It consists of representing each spatial area by a node and connecting nodes that present similar time series. One fundamental concern when building climate network is the definition of a metric that captures similarity between time series. The majority of papers in the literature use Pearson correlation with or without lag. Here we study the influence of 29 time series distance functions on climate network construction using global temperature data. We observed that the distance functions used in the literature generate similar networks in general while alternative ones generate distinct networks and exhibit different long-distance connection patterns (teleconnections). These patterns are highly important for the study of climate dynamics since they generally represent long-distance transportation of energy and can be used to forecast climatological events. Therefore, we consider that the measures here studied represent an alternative for the analysis of climate systems due to their capability of capturing different underlying dynamics, what may provide a better understanding of global climate.

Motivation & Objective

• To assess the impact of different time series distance functions on climate network topology.
• To identify whether commonly used metrics like Pearson correlation are optimal for capturing climate dynamics.
• To explore how alternative distance functions affect the detection of long-distance connections (teleconnections) in climate networks.
• To evaluate whether diverse distance functions can reveal distinct underlying climate system behaviors.
• To provide a foundation for selecting more informative distance functions in future climate network analyses.

Proposed method

• The study applies 29 time series distance functions to global temperature time series data across spatial grid points.
• Each distance function computes similarity between time series of adjacent or distant grid points to define network edges.
• Climate networks are constructed by connecting nodes (spatial locations) based on distance thresholds or fixed k-nearest neighbors.
• Network properties such as clustering, path length, and long-range connectivity are analyzed and compared across distance functions.
• The influence of each distance function on teleconnection patterns is assessed through visualization and statistical comparison of network structures.
• The analysis focuses on identifying differences in network topology, particularly in long-distance connections, across the 29 functions.

Experimental results

Research questions

• RQ1 How do different time series distance functions affect the structure of climate networks?
• RQ2 To what extent do standard metrics like Pearson correlation differ from alternative distance functions in capturing climate dynamics?
• RQ3 Do alternative distance functions reveal distinct long-distance connection patterns (teleconnections) not captured by conventional methods?
• RQ4 How do variations in distance functions influence the detection of energy transport pathways in the climate system?
• RQ5 Can alternative distance functions provide more informative representations of global climate connectivity?

Key findings

• Commonly used distance functions such as Pearson correlation produce highly similar climate network structures.
• Alternative distance functions generate distinct network topologies with unique patterns of long-distance connections.
• These alternative functions reveal different teleconnection structures, particularly in remote spatial connections.
• The variation in long-distance connectivity patterns suggests that different distance functions capture different underlying climate dynamics.
• The results indicate that distance function choice significantly influences the interpretation of climate network connectivity and dynamics.
• The study concludes that alternative distance functions offer valuable, complementary insights into climate system behavior beyond standard correlation-based approaches.