Chapter 12
Teleconnections Patterns
by Antonio Navarra
12.1 Objective Teleconnections
Walker and Bliss had devoted their entire life to a pure subjective search for
significant statistical relation among the myriads of correlation values that
their limited data set was producing, but a breakthrough was achieved in
1981 by two papers by Mike Wallace in collaboration with D. Gutzier and J.
Horel (Wallace and Gutzier, 1981; Horel and Wallace, 1981).
In modern terms, tele connections relations can be visualized by computing
the correlation matrix R. The elements ofR are the correlation coefficient rij
between the time series of the variable of interest (for instance, geopotential
or surface pressure ) at the grid point i with the temporal series at the grid
point j. The data usually consist of anomaly time series of fields, as for
instance monthly meean fields, since the correlation coefficients give more
stable results for anomalies. The grid points are assumed to be ordered
using only one index, rather than the more familiar two-indices ordering.
Considerable information is contained in the matrix R. The i-th column
represents the correlation information for the i-th grid-point and it is usually
called a one-point teleconnection map for the basis point i.
These patterns represents how each grid-point is connected with its neighbors. The patterns are characterized by a roughly elliptical area of positive correlations around the basis point. This area is the scale of the local
Acknowledgements: Many thanks go to Neil Ward and Mike Wa.llace for kindly a.llowing to reproduce the pictures from their papers. This work was partia.lly supported by the
EEC contract EVsV-CT92-0101
Teleconnections Patterns
by Antonio Navarra
12.1 Objective Teleconnections
Walker and Bliss had devoted their entire life to a pure subjective search for
significant statistical relation among the myriads of correlation values that
their limited data set was producing, but a breakthrough was achieved in
1981 by two papers by Mike Wallace in collaboration with D. Gutzier and J.
Horel (Wallace and Gutzier, 1981; Horel and Wallace, 1981).
In modern terms, tele connections relations can be visualized by computing
the correlation matrix R. The elements ofR are the correlation coefficient rij
between the time series of the variable of interest (for instance, geopotential
or surface pressure ) at the grid point i with the temporal series at the grid
point j. The data usually consist of anomaly time series of fields, as for
instance monthly meean fields, since the correlation coefficients give more
stable results for anomalies. The grid points are assumed to be ordered
using only one index, rather than the more familiar two-indices ordering.
Considerable information is contained in the matrix R. The i-th column
represents the correlation information for the i-th grid-point and it is usually
called a one-point teleconnection map for the basis point i.
These patterns represents how each grid-point is connected with its neighbors. The patterns are characterized by a roughly elliptical area of positive correlations around the basis point. This area is the scale of the local
Acknowledgements: Many thanks go to Neil Ward and Mike Wa.llace for kindly a.llowing to reproduce the pictures from their papers. This work was partia.lly supported by the
EEC contract EVsV-CT92-0101