Inferring interdependencies in climate networks constructed at inter-annual, intra-season and longer time scales

Deza, J., Barreiro, M. and Masoller, C. (2013) Inferring interdependencies in climate networks constructed at inter-annual, intra-season and longer time scales. The European Physical Journal Special Topics, 222 (2). pp. 511-523. ISSN 1951-6355 Available from: http://eprints.uwe.ac.uk/39887

Full text not available from this repository.

Abstract/Description

We study global climate networks constructed by means of ordinal time series analysis. Climate interdependencies among the nodes are quantified by the mutual information, computed from time series of monthly-averaged surface air temperature anomalies, and from their symbolic ordinal representation (OP). This analysis allows identifying topological changes in the network when varying the time-interval of the ordinal pattern. We consider intra-season time-intervals (e.g., the patterns are formed by anomalies in consecutive months) and inter- annual time-intervals (e.g., the patterns are formed by anomalies in consecutive years). We discuss how the network density and topology change with these time scales, and provide evidence of correlations be- tween geographically distant regions that occur at specific time scales. In particular, we find that an increase in the ordinal pattern spacing (i.e., an increase in the timescale of the ordinal analysis), results in climate networks with increased connectivity on the equatorial Pacific area. On the contrary, the number of significant links decreases when the ordinal analysis is done with a shorter timescale (by comparing consecutive months), and interpret this effect as due to more stochasticity in the time-series in the short timescale. As the equatorial Pacific is known to be dominated by El Nin ̃o-Southern Oscillation (ENSO) on scales longer than several months, our methodology allows constructing climate networks where the effect of ENSO goes from mild (monthly OP) to intense (yearly OP), independently of the length of the ordinal pattern and of the thresholding method employed.

Item Type: Article
Uncontrolled Keywords: climate networks, information theory, symbolic calculus
Faculty/Department: Faculty of Environment and Technology > Department of Architecture and the Built Environment
Depositing User: Dr I. Deza
Date Deposited: 19 Mar 2019 16:11
Last Modified: 19 Mar 2019 16:11
URI: http://eprints.uwe.ac.uk/id/eprint/39887

Statistics

Downloads
Activity Overview
0Downloads
0Hits
Origin of downloads

Additional statistics for this repository are available via IRStats2

Actions (login required)

View Item View Item