A system has been developed at COLA for making seasonal to interannual predictions of Tropical Pacific SST, using a coupled atmosphere-ocean general circulation model that incorporates an analysis of subsurface ocean measurements of temperature and salinity in the initial conditions (Schneider et al., 1997, 1998). The ocean component of the prediction model has a nearly global domain. The model is fully coupled (i.e. does not use anomaly coupling or flux correction). Instead, the approach of anomaly initial conditions is used to reduce problems associated with climate drift and the shock of inserting initial conditions. Initial conditions for the ocean are obtained from a near-global ocean analysis produced at COLA using an ocean data assimilation system (Huang and Kinter, 1997). The prediction system displays skill at leads up to 18 months for June 30 initial conditions.

The ODA uses variational optimal interpolation following Derber and Rosati (1989). The period of the analysis is 12 years from 1985-1997. The ocean model for the assimilation and for the coupled model is a nearly global version (0 ^oE to 360 ^oE, 60 ^oS to 65 ^oN) of the GFDL ocean model MOM 1 (Pacanowski et al., 1993). There are 20 levels in the vertical with 16 in the upper 400m. The zonal resolution is 1.5^o longitude and 0.5^o latitude between 10N and 10S. This tropical resolution and vertical structure are the same as used in the COLA anomaly coupled forecast system (Kirtman et al., 1997; Kirtman and Shukla, this issue of the Bulletin).

The atmospheric component of the coupled model is the COLA atmospheric GCM, which is a global spectral model with a state of the art suite of physical parameterizations. The AGCM resolution is triangular 30 with 18 unevenly spaced levels in the vertical. The anomaly coupled forecast system uses the same AGCM resolution and physics, except that, with the AGCM used here, the deep cumulus parameterization is the relaxed Arakawa-Schubert scheme of Moorthi and Suarez (1992) as implemented by DeWitt (1996), and the diagnostic cloud-radiative interaction scheme is modified following Kiehl et al. (1994, 1996) as described in DeWitt and Schneider (1997).

The coupled model climatology is obtained from the last six years of a 12 year coupled simulation after a three year spinup of the ocean with observed wind stress starting from the Levitus (1982) January climatology. The coupled model has a realistic annual cycle of SST at the equator, as well as vigorous interannual SST variability in the Tropical Pacific. However, the annual mean SST is too warm by up to 3 ^oC in the eastern equatorial Pacific, and the thermocline is only about half as deep as observed in the western Tropical Pacific.

Ocean initial conditions for the forecasts are obtained by adding the ODA anomalies to the climatology of the coupled model. The atmosphere is brought close to equilibrium with the initial SST anomaly through a one month spin-up. The system has been verified by performing one hindcast for each June 30 initial condition of the ODA.

The results of a hindcast verification for the NINO3 SST anomaly are shown in Fig. 1. The anomaly correlation drops to near 0.6 after six months and oscillates about this value for the rest of the period. The model anomaly correlation is better than the persistence anomaly correlation for all lead times. The RMS error levels off at 1 deg. C at three months lead time and then oscillates out of phase with the anomaly correlation. The model RMS error is smaller than the persistence RMS error for lead time longer than six months.

The system produces global SSTA predictions. The predicted SSTA in the low latitude Pacific from the latest forecast, made with June 30, 1997 initial conditions, is shown in Fig. 2. The anomalies are deviations from the coupled model climate without correction for the systematic error. The coupled system predicts a strong cold event beginning in the summer of 1998. Several features of the forecast should be noted that are probably erroneous, given the past behavior of warm and cold events. The cold anomaly extends nearly all the way across that Pacific rather than being small in the western Pacific, and the SSTA decays too rapidly away from the equator.

Acknowledgments: This work was supported under NOAA grant NA26-GP0149 and NA46-GP0217 and NSF grant ATM-93-21354.

References:

Derber, J. and A. Rosati, 1989: A global oceanic data assimilation system, J. Phys.Oceanogr., 19, 1333-1347.

DeWitt, D. G., 1996: The effect of the cumulus convection scheme on the climate of the COLA general circulation model. COLA Rep. 27, 58 pp.

DeWitt, D. G. and E. K. Schneider, 1997a: The Earth radiation budget as simulated by the COLA GCM. COLA Rep. 35, 39pp.

Huang, B. and J. L. Kinter III, 1997: A global ocean analysis for 1986-1992. COLA Rep. 38, 62 pp.

Kiehl, J. T., J. J. Hack, and B. P. Briegleb, 1994: The simulated Earth radiation budget of the National Center for Atmospheric Research community climate model CCM2 and comparisons with the Earth Radiation Budget Experiment (ERBE). J. Geophys. Res., 99, 20815-20827.

Kiehl, J. T., B. Boville, B. Briegleb, J. Hack, P. Rasch, and D. Williamson, 1996: Description of the NCAR Community Climate Model (CCM3). NCAR Technical Note NCAR/TN-420+STR.

Kirtman, B. P., J. Shukla, B. Huang, Z. Zhu, and E. K. Schneider, 1997: Multiseasonal predictions with a coupled tropical ocean global atmosphere system. Mon. Wea. Rev., 125, 789-808.

Levitus, S., 1982: Climatological Atlas of the World Ocean. NOAA Professional Paper 13. U. S. Dept. of Commerce, 173pp.

Moorthi, S. and M. J. Suarez, 1992: Relaxed Arakawa-Schubert: A parameterization of moist convection for general circulation models. Mon. Wea. Rev., 120, 978-1002.

Pacanowski, R. C., K. Dixon, and A. Rosati, 1993: The GFDL modular ocean model users guide, version 1.0. GFDL Ocean Group Tech. Rep. No. 2.

Schneider, E. K., Z. Zhu, D. G. DeWitt, B. Huang and B. P. Kirtman, 1997a: ENSO Hindcasts with a Coupled GCM. COLA Report 39, 40 pp.

Schneider, E. K., B. Huang, Z. Zhu, D. G. DeWitt, J. L. Kinter III, B. Kirtman, and J. Shukla, 1998: Ocean data assimilation, initialization, and predictions of ENSO with a coupled GCM. Mon. Wea. Rev., accepted.

Figure captions:

Fig. 1. Anomaly correlation (top) and RMS error (bottom, (C) of 12 NINO3 SSTA hindcasts from June 30 initial conditions 1985-1996. Persistence is shown by the dashed curve in each panel.

Fig. 2. Results of prediction from June 30, 1997 initial conditions. Top: evolution of NINO3 SSTA. Bottom: SSTA predicted for December, 1998.