Predictability of Zonal Means during Boreal Summer

Abstract

This study examines the predictability of seasonal means during boreal summer. The results are based on ensembles of June?July?August (JJA) simulations (started in mid-May) carried out with the NASA Seasonal-to-Interannual Prediction Project (NSIPP-1) atmospheric general circulation model (AGCM) forced with observed sea surface temperatures (SSTs) and sea ice for the years 1980?99. It is found that the predictability of the JJA extratropical height field is primarily in the zonal-mean component of the response to the SST anomalies. This contrasts with the cold season (January?February?March) when the predictability of seasonal means in the boreal extratropics is primarily in the wave component of the El Niño?Southern Oscillation (ENSO) response. Two patterns dominate the interannual variability of the ensemble mean JJA zonal-mean height field. One has maximum variance in the tropical/subtropical upper troposphere, while the other has substantial variance in the midlatitudes of both hemispheres. Both are symmetric with respect to the equator. A regression analysis shows that the tropical/subtropical pattern is associated with SST anomalies in the far eastern tropical Pacific and the Indian Ocean, while the midlatitude pattern is associated with SST anomalies in the tropical Pacific just east of the date line. The two leading zonal height patterns are reproduced in model runs forced with the two leading JJA SST patterns of variability. A comparison with observations shows a signature of the midlatitude pattern that is consistent with the occurrence of dry and wet summers over the United States. It is hypothesized that both patterns, while imposing only weak constraints on extratropical warm season continental-scale climates, may play a role in the predilection for drought or pluvial conditions.