Signatures of Climatic Phenomena in Climate Networks: El Niño and La Niña events

Abstract:

We construct climate network based on surface air temperature data to identify distinct signatures of climatic phenomena such as El Niño and La Niña events which trigger many climatic disruptions around the globe with serious economic and ecological consequences. Climate networks are used to forecast various important climate phenomena, such as the monsoon, the North Atlantic Oscillation, El Niño events. Here we use correlation networks constructed out of the surface air temperature data (we use reanalysis data of the daily near surface air temperature (1000 hPa) from the ’NCEP/ NCAR reanalysis 1 project' (1979-2019) (https://psl.noaa.gov/)). We pick 725 grid points (7.5˚*12.5˚ resolution) for the entire globe and 105 grid points (2.5˚*2.5˚ resolution) for El Niño 3.4 region. Here we use correlation network constructed out of the surface air temperature of time series of different points in a global grid to analyse El Niño and La Niña phenomena.

We obtained monthly correlation of surface air temperature data during the arrival of the El Niño or La Niña events and for the preceding months to the event, both for the entire globe and for the El Niño basin region and constructed the heat map. This is performed by calculating the Pearson correlation co-efficient, for the all possible pairs of nodes, and calculating the average over the days of the given month. These values are used as the elements of the adjacency matrix of the grid nodes which is then visualised as a cluster heat map. We can also use the correlation matrix to calculate quantifers like the fraction of links with correlation values above a certain threshold and the entropy defined by –ln (no. of links with correlation above a given threshold / no. of total links).

The correlation matrix of the network shows a structure which has distinct characteristics for El Niño events, La Niña, and a time period of no events. We also identify the signature of the El Niño and La Niña oscillations in the heat map of the system and justify our cluster heat map with the quantifers. We hope these quantifers can be further used for the prediction of climatic events.