Recent observational studies showed a multidecadal variation in the sea surface temperature (SST) in the central North Pacific Ocean and coherent variations in atmospheric circulation in the North Pacific and North American sector (Hu and Feng 2001a). A unique feature of this variation is the role of the variations in regulating the effect of El Nino/Southern Oscillation (ENSO) in the tropical Pacific on the summer (June-August) rainfall in the central United States. As a consequence, the ENSO teleconnection with the rainfall has alternated and completed nearly two cycles over the last 130 years; the ENSO effect was strong and significant from the early 1870s to middle 1910s and from 1948 to 1978, and was weak and languished in the two periods 1920-47 and 1979-present. In the former two epochs, ENSO dominated the interannual variation in summer rainfall in the central United States, serving as a reliable “predictor” of the summer rainfall anomaly. In the latter two epochs, the ENSO effect weakened.

The studies further showed that in the latter two epochs when the ENSO effect weakened, variations associated with the southerly flows from the Gulf of Mexico into the Great Plains dominated the interannual changes in summer rainfall in the central United States, indicating interactions of multidecadal scales between the circulations in the Pacific and North American sector and circulations in the Gulf of Mexico and the tropical Atlantic region (Hu and Feng 2001b).

The quasi-periodic change in strength of the ENSO teleconnection with and the associated alternation of the causes of rainfall variations in the central United States are mechanisms explaining the observed interannual variations in the summer rainfall in the central United States. These mechanisms must be contained in general circulation models (GCM) in order to correctly describe precipitation climate in the central United States and the North America. This model ability is especially important in light of the increase in application of GCM predictions in climate impact assessments and policy studies (IPCC 1998; National Research Council 1998). A climate impact assessment based on predictions from models unable to describe the multidecadal change in the ENSO teleconnection and, therefore, the interannual variations in summer rainfall in the central United States could yield misleading results. Similar consequences could result from model evaluation of greenhouse gas effect on climate change if these natural variations in the climate are not well resolved by the models.

In this study, we analyze the simulation results of the National Center for Atmospheric Research (NCAR) Community Climate Model (CCM3), and examine whether this atmospheric model is able to simulate the multidecadal variation in the ENSO teleconnection with the summer rainfall in the central United States, and if it can describe the alternation of the sources causing interannual variations in the summer rainfall.

Major results of this study are that the NCAR CCM3 can describe the quasi-periodic multidecadal variation in the ENSO teleconnection with the summer rainfall in the central United States, although some large differences exist between the model simulation and observation in details of these variations. Findings from explorations of the model dynamics, particularly the eddy effect using the E-P flux dynamic diagnostics, on maintaining the atmospheric processes facilitating the model multidecadal variations in the ENSO teleconnection will be detailed in my presentation at the meeting.