Forecast of Tropical Pacific SST using an intermediate ocean and statistical atmosphere model

contributed by In-Sik Kang and Jong-Seong Kug

Department of Atmospheric Sciences, Seoul National University, Seoul 151-742, Korea

An El-Niño prediction model was developed in order to improve ENSO prediction and to understand the physical basis for ENSO. The intermediate ocean and statistical atmosphere model (hereafter referred to as ISCM) is similar to Cane and Zebiak(1987). However, it has been modified in terms of the parameterization of subsurface temperature and the basic state. In addition, the atmosphere model replaced by an empirical statistical model which is based on the singular value decomposition (Kang and Kug, 1999).

Figure 1. compares the forecast skill of the ISCM and LDEO2 (Chen et al., 1998) for hindcast experiments during 1980-97. Each hindcast is run for 24 months. The skill is measured by the correlation and the root-mean-square(rms) error between monthly model forecast and observed SST anomalies in the tropical Pacific. The present model gives higher correlation and lower rms error for most lead times, particularly for lead times less than one year. The better performance of the present model is more distinctive when the correlation and rms are compared for the NINO4 SST. It is also noted that the correlations over tropical central and eastern Pacific exceed 0.7 for a 6 months lead and 0.5 for a 12 month lead.

The initialization strategy for the ISCM is to combine the observed SST and wind stress in initializing the ocean state. This initialization also differs somewhat from that used in the LDEO2 prediction system. The initialization is done by spinning up the model with an empirically derived wind stress for the period from January 1970 to the time of the forecast. The simple empirical wind stress combines the observed wind stress and the wind stress obtained from the statistical atmosphere using observed SSTA. This initialization technique forces the model initial state to be in close agreement with observations. Displayed in Fig. 2. is a comparison of these forecasts of NINO3 SST anomalies with observations at twelve month lag for the period from Jan. 1997 to Dec. 1998.

Fig. 3 and Fig. 4 show the twelve months forecast of ENSO indices including NINO3, NINO4 and seasonal SST forecast in tropical Pacific basin for the period of December, 1999 to November, 2000, respectively. Model forecast indicates that SST anomaly will be slightly negative in the next few seasons, but approach normal state in the eastern equatorial Pacific during the end of 2000.

Reference

Cane, M. A. and S. E. Zebiak, 1987: Prediction of El Niño Events using a physical model. Atmosphere and Ocean Variability, H.Cattel, Ed,Roy. Meteor.Soc., 153-182.

Chen, D., S. E. Zebiak, and A. Kaplan, 1998: The impact of sea level data assimilation on the Lamont model prediction of the 1997/1998 El Niño, Geophys. Res. Lett., 25, 2837-2840.

Kang, I.-S., and J.-S. Kug, 1999: An El-Niño prediction model with an intermediate ocean and statistical atmosphere system Submitted to Geophys. Res. Lett.