Combining point correlation maps with self-organising maps to compare observed and simulated atmospheric teleconnection patterns

We use a new method based on point correlation maps and self-organising maps (SOMs) to identify teleconnection patterns in 60 yr of National Centres for Environmental Prediction/National Centre for Atmospheric Research (NCEP/NCAR) sea level pressure (SLP) re-analysis data. The most prevalent pattern...

Full description

Bibliographic Details
Published in:Tellus A: Dynamic Meteorology and Oceanography
Main Authors: Freja K. Hunt, Joël J.-M. Hirschi, Bablu Sinha, Kevin Oliver, Neil Wells
Format: Article in Journal/Newspaper
Language:English
Published: Stockholm University Press 2013
Subjects:
Online Access:https://doi.org/10.3402/tellusa.v65i0.20822
https://doaj.org/article/63c339aa0c8f429dbc50cec74743d810
Description
Summary:We use a new method based on point correlation maps and self-organising maps (SOMs) to identify teleconnection patterns in 60 yr of National Centres for Environmental Prediction/National Centre for Atmospheric Research (NCEP/NCAR) sea level pressure (SLP) re-analysis data. The most prevalent patterns are the El Nino Southern Oscillation (ENSO), the North Atlantic Oscillation (NAO) and the Southern Annular Mode (SAM). Asymmetries are found between base points in opposite centres of action of the NAO and the Pacific North America pattern (PNA). The SOM-based method is a powerful tool that allows us to efficiently assess how realistically teleconnections are reproduced in any climate model. The degree of agreement between modelled and re-analysis-based teleconnections (or between different models) can be summarised in a single plot. Here, we illustrate this by assessing the skill of the medium complexity climate model FORTE (Fast Ocean Rapid Troposphere Experiment). FORTE reproduces some realistic teleconnections, such as the Arctic Oscillation (AO), the NAO, the PNA, the SAM, the African Monsoon and ENSO, along with several other teleconnections, which resemble to varying degrees the corresponding NCEP patterns. However, FORTE tends to underestimate the strength of the correlation patterns and the patterns tend to be slightly too zonal. The accuracy of frequency of occurrence is variable between patterns. The Indian Ocean is a region where FORTE performs poorly, as it does not reproduce the teleconnection patterns linked to the Indian Monsoon. In contrast, the North and equatorial Pacific and North Atlantic are reasonably well reproduced.