Boreal winter predictions with the GEOS‐2 GCM: The role of boundary forcing and initial conditions

• Published in: Quarterly Journal of the Royal Meteorological Society
• Main Authors: Chang, Yehui, Schubert, Siegfried D., Suarez, Max J.
• Format: Article in Journal/Newspaper
• Language: English
• Published: Wiley 2000
• Subjects: Pacific; North Atlantic
• Online Access: http://dx.doi.org/10.1002/qj.49712656715; https://api.wiley.com/onlinelibrary/tdm/v1/articles/10.1002%2Fqj.49712656715; https://rmets.onlinelibrary.wiley.com/doi/pdf/10.1002/qj.49712656715

Abstract Ensembles of atmospheric General Circulation Model (GCM) seasonal forecasts and long‐term simulations are analysed to assess the controlling influences of boundary forcing and memory of the initial conditions. Both the forecasts and simulations are carried out with version 2 of the Goddard Earth Observing System (GEOS‐2) GCM forced with observed sea surface temperatures (SSTs). While much of the focus is on the seasonal time‐scale (January‐March; 1981–95) and the Pacific North American (PNA) region, we also present results for other regions, shorter time‐scales, and other known modes of variability in the northern hemisphere extratropics. Forecasts of indices of some of the key large‐scale modes of variability show that there is considerable variability in skill between different regions of the northern hemisphere. The eastern North Atlantic region has the poorest long‐lead forecast skill, showing no skill beyond about 10 days. Skilful seasonal forecasts are primarily confined to the wave‐like El Niño Southern Oscillation (ENSO) response emanating from the tropical Pacific. In the northern hemisphere, this is similar to the well‐known PNA pattern. Memory of the initial conditions is the major factor leading to skilful extratropical forecasts of lead time less than one month, while boundary forcing is the dominant factor at the seasonal time‐scale. Boundary forcing contributes to skilful forecasts at sub‐seasonal time‐scales only over the PNA region. The GEOS‐2 GCM produces average signal‐to‐noise ratios which are less than 1.0 everywhere in the extra‐tropics, except for the subtropical Pacific where they approach 1.5. An assessment of the sampling distribution of the forecasts suggests the model's ENSO response is very likely too weak. These results show some sensitivity to the uncertainties in the estimates of the SST forcing fields. In the North Pacific region, the sensitivity to SST forcing manifests itself primarily as changes in the variability of the PNA response, underscoring the need for an ...