The climate over North Pacific and North America undergoes fluctuations on interannual-to-decadal time scales, with significant changes in precipitation, surface temperature and geopotential height. We will explore the hindcasting skill and physics of a climate system that hindcasts anomalies of heat flux convergence of the tropical and North Pacific oceanic mixed layer and converts them into the northern hemisphere climate anomalies using a coupled ocean mixed layer - atmosphere general circulation model.

The philosophy underlying the model states that North Pacific/North American climate anomalies on time scales of years to decades result from anomalous conditions in the Pacific equatorial upwelling regions, and the North Pacific Kuroshio-Oyashio extension (KOE), where anomalies in the ocean's thermocline are efficiently coupled to the surface. In these areas, changes of upwelling, thermocline depth, water mass properties and stratification, and advection impact a heat flux convergence of the ocean's surface mixed layer. In response to this forcing, the coupled atmosphere - ocean mixed layer system adjusts, and generates anomalies of sea surface temperature (SST) over the entire Pacific, including the central and eastern North Pacific where the ocean's mixed layer is effectively insulated from the thermocline.

Using the new method, forecast/hindcast of North Pacific/North American climate require prediction/hindcast of the heat flux convergences over the tropics and the KOE region. In the tropical Pacific, interannual SST anomalies associated with ENSO are routinely and skillfully predicted with a lead time of a few seasons. In the KOE, low-frequency anomalies of SST can be predicted/hindcasted from Rossby wave dynamics and observations of the surface wind stress during the preceding years.

The forecast system to be investigated here formulates these predictions in terms of ocean heat flux convergences and applies them as anomalous forcing in coupled ocean-mixed layer (or slab-ocean model/SOM) and atmospheric general circulation model (CCM3). The procedure of forecasting ocean heat flux convergence in the KOE and tropical Pacific region was explored utilizing various ocean GCMs (HOPE, POP and MIT). Ocean general circulation models were forced with the NCAR/NCEP reanalysis wind stress, while relaxing SST and sea surface salinity (SSS) to their climatological values. The air-sea heat flux is dominated in the KOE by the delayed effect of low-frequency ocean waves, i.e. by the predictable mid-latitude signal. Elsewhere in the mid-latitudes, the air-sea heat flux is much smaller and balances Ekman heat advection. In the tropics, the air-sea heat flux reflects primarily changes of the upwelling heat fluxes and advection, that are part of the predictable El Nino evolution. Therefore, the oceanic perturbations of the surface heat budget can be diagnosed from air-sea heat fluxes of these runs. Comparison with the observations showed that the oceanic models have good skill in hindcasting the mixed layer temperature and the thermocline depth anomalies in the KOE and NINO3.4 regions.

The hindcasted monthly oceanic heat flux convergence anomalies for the period of January 1960 to December 1999 in the KOE region and in the tropical Pacific were then applied as anomalous forcing in the coupled CCM3/SOM model. We performed a 40 year integration of the five member ensemble of CCM3/SOM forced with heat budget anomalies (the delta Q forcing) obtained from the ocean hindcast. This was achieved by augmenting in the tropics and KOE region the forcing in the oceanic mixed layer temperature equations with the time dependent oceanic heat convergences while elsewhere the undisturbed equation was retained. The members of the ensemble were obtained by shifting the initial conditions by one day each.

To identify the atmospheric response to the KOE delta Q forcing alone and to validate the statistical analysis, we repeated the five member CCM3/SOM ensemble integration, but applied the delta Q forcing in the tropics only. The differences between this integration and the corresponding CCM3/SOM integration with both tropical and KOE forcing indicates the effect of the KOE delta Q forcing on the mean atmospheric state and on the pdf of the intrinsic atmospheric modes.

Our preliminary findings can be summarized as follows: (1)The anomalous forcing over the tropical Pacific sets the global anomalous climate patterns over North Pacific/North America. These patterns include the SST and air-sea heat flux anomalies over North Pacific. Over the KOE region the SST and heat flux anomalies are strongly correlated with the heat flux anomalies over the tropical Pacific. (2)When the forcing over the KOE region is added, the relationship between the SST and air-sea heat flux anomalies changes. The air-to-sea heat flux anomalies over the KOE region are nor correlated with the tropical heat flux anomalies. However, there is a 0.4 correlation (as compared to 0.6 for the case with the tropical forcing only) between the SST anomalies over the KOE region and over the NINO3.4 region. (3) The additional forcing over the KOE region augments the teleconnection from the tropics. The changes are exhibited in the surface variables as well as in the free atmosphere. (4) Comparison with the observations reveals that the anomalous forcing over the KOE region sets the changes in the precipitation patterns over the west coast of the North US and Canada. The corresponding 2m temperature anomalies are most pronounced over Alaska and west coast of Canada