Extended Memory of the Initial Conditions in Long-Range Forecasts of the January 1983 Atmospheric Circulation

Réjean Michaud Laboratoire de Météorologie Dynamique du CNRS, Paris, France

Search for other papers by Réjean Michaud in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

We compare four ensembles of long-range numerical forecasts made with the Laboratoire de Météorologie Dynamique general circulation model. These ensembles make use of mid-December initial conditions taken from distinct “non-El Niño” years and 1-day lagged initial conditions of mid-December 1982, a specific El Niño year. Both climatological sea surface temperatures (SSTs) and observed SSTs for 1982/83 winter are used to force these ensembles.

The comparison of both ensembles using mid-December 1982 initial conditions reveals a rapid impact of the SST anomalies on the forecast. On the other hand, at the end of the 45-day forecasts, namely, more than one month after their beginning, the ensembles making use of the same boundary conditions still display significant differences. The noteworthy feature is that the ensemble using both mid-December 1982 initial conditions and observed SST anomalies leads to an ensemble average January-mean predicted circulation that shows a much better agreement with the observations than does the ensemble using only the relevant boundary conditions. We suggest that this extended memory of the initial conditions is due to the extant influence of SST anomalies on initial fields, especially for the case studied here of mid-December 1982; indeed, those initial fields have been substantially modulated by the extraordinary SST anomalies already present in the autumn and the beginning of the 1982/83 winter.

Abstract

We compare four ensembles of long-range numerical forecasts made with the Laboratoire de Météorologie Dynamique general circulation model. These ensembles make use of mid-December initial conditions taken from distinct “non-El Niño” years and 1-day lagged initial conditions of mid-December 1982, a specific El Niño year. Both climatological sea surface temperatures (SSTs) and observed SSTs for 1982/83 winter are used to force these ensembles.

The comparison of both ensembles using mid-December 1982 initial conditions reveals a rapid impact of the SST anomalies on the forecast. On the other hand, at the end of the 45-day forecasts, namely, more than one month after their beginning, the ensembles making use of the same boundary conditions still display significant differences. The noteworthy feature is that the ensemble using both mid-December 1982 initial conditions and observed SST anomalies leads to an ensemble average January-mean predicted circulation that shows a much better agreement with the observations than does the ensemble using only the relevant boundary conditions. We suggest that this extended memory of the initial conditions is due to the extant influence of SST anomalies on initial fields, especially for the case studied here of mid-December 1982; indeed, those initial fields have been substantially modulated by the extraordinary SST anomalies already present in the autumn and the beginning of the 1982/83 winter.

Save