Response of a General Circulation Model to a Sea Temperature Perturbation

Abstract

Sea temperature anomalies which departed from the December climatic mean by approximately 2C off the coast of Newfoundland were inserted into the NCAR six-layer, 5° mesh, general circulation model of the atmosphere in order to test the model's response to small perturbations in sea surface temperature. The response of the model to the anomalies was analyzed with respect to pressure patterns, heat flux, and cyclone frequency, path and intensity. This response was compared with a statistical analysis of the response of the atmosphere to similar sea temperature anomalies based on approximately 80 years of observations as described by Ratcliffe and Murray. Analyses of the anomaly experiments are preceded by an analysis of the basic (control) statistics for both model and atmosphere. The most pronounced discrepancies between the two were noted in cyclone statistics. A calculation with double horizontal resolution greatly improved the model features. Detailed comparison was complicated by the fact that the model failed to achieve statistical stationarity. The extensive verification data of Ratcliffe and Murray proved valuable in distinguishing meaningful anomaly responses from those that could be attributed to the many limitations in the model, including a pronounced natural variability. Both warm and cold anomaly cases were tested. Best agreement with observed data was obtained for the case of the warm anomaly; this agreement was most evident during the middle portions of the integrations and then only in the North Atlantic sector. The response in the case with a cold anomaly was not as satisfactory although there were clear distinctions between the warm and cold anomaly cases.