contributed by Tony Barnston², Ming Ji¹, Arun Kumar¹ and Ants Leetmaa²

A non-simple coupled ocean-atmosphere model has been developed for use for long-lead climate forecasts in the Coupled Model Branch of the Experimental Modeling Center (EMC) at NOAA's National Centers for Environmental Prediction (NCEP) (Ji et al. 1994a,b). The NCEP Medium Range Forecast (MRF) atmospheric model is used with a dynamic Pacific Basin ocean model originated at the Geophysical Fluid Dynamics Laboratory. The MRF has a reduced spatial resolution and is tuned for more realistic tropical circulation. The ocean thermal field, including SST and subsurface temperature, is initialized using an ocean data assimilation system (Ji et al. 1995). It has been shown that when observed SST fields are prescribed, the coupled model's atmospheric response is fairly reliable in the tropics but considerably less so in the extratropics, as is the case with most AGCMs. The extratropical response is most realistic during the warm or cold phase of ENSO as reflected in the SST. The nature of the model’s mid-latitude atmospheric response to the ENSO state has been treated in some detail in Livezey et al. (1997). However, while better prediction of the extratropical atmosphere is an ultimate goal, much attention has been given to the more basic prediction of ENSO itself--the tropical Pacific SST anomaly. Such an SST forecast is presented here.

In the September and December 1993 issues of this Bulletin, the expected forecast skill of the coupled model version used in 1993 (called CMP6) was shown. A horseshoe-shaped spatial pattern of maximum model skill was noted, with highest equatorial skill near the date line and higher skill just north or south of the equator than immediately along it to the east of 165^oW. The model generally outperformed persistence by a substantial margin in forecasting the Niño 3 and Niño 4 regions. A seasonal dependence in skill was noted, where forecasts were affected by a "spring barrier" as found in many other dynamical and statistical predictive models. As discussed in Barnston et al. (1994), the skill of CMP6's forecasts of tropical Pacific SST was competitive with that of other dynamical and statistical models in early 1994.

Starting with the forecasts presented in the September 1994 issue, the model was upgraded with a refinement of the flux climatology and the installation of a MOS (for "model output statistics") correction for the stress anomalies produced by the atmospheric model. Skill consequently improved and the high skill area extended farther eastward into the western part of the Niño 3 region (see Figures 2-1 and 2-2 of the September 1994 issue). The newer model version was named CMP9. In spring 1995, another improvement was implemented, resulting in the CMP10 version (Ji et al. 1996). While CMP9 contained a negative feedback procedure for coupling the anomalous net heat flux, CMP10 used anomaly coupling for the net heat flux forcing. While mean skill differences between CMP9 and CMP10 were not as marked as between CMP6 and CMP9, CMP10 behaved more realistically for high amplitude SST anomalies. CMP9, with its negative feedback mechanism, sometimes damped strong ENSO events too much and/or too soon.

The most recent improvement of the NCEP coupled model was completed during summer 1996, and is used in the current forecasts. The resulting model version (CMP12) has the highest hindcast skills in the model’s history over the central and eastern equatorial Pacific for 1981-95, exceeding 0.8 correlation skill in much of that area for 4-6 month lead hindcasts. Upgrades in CMP12 include improvements in the data assimilation system, and model improvements such as better mixing in the ocean model and more representative anomalous evaporation-precipitation (e-p) flux forcing in the coupling. CMP12 forecasts out to 6 months lead are now updated on a weekly basis and are available on Internet site http://nic.fb4.noaa.gov:8000 /research/climate/html.

The CMP12 coupled model forecasts for the SST anomaly field averaged over Apr-May-Jun, Jul-Aug-Sep and Oct-Nov-Dec 1998 are shown in Fig. 1, where the systematic model bias for hindcasts over the 1981-95 period has been subtracted. This forecast is actually the mean of an ensemble of 7 to 11 individual cases, each based on a different one- to two-week-apart initial ocean condition ranging from early February through early April 1998. The forecast for Apr-May-Jun shows a marked weakening of the extreme warmth that we have seen during the first three months of 1998, although a return toward normal between early summer and the end of 1998 proceeds only slowly, implying the possibility of neutral or even still slightly warm conditions for boreal winter 1998-99. Figure 2 shows the Niño 3 and Niño 3.4 forecasts in the form of time series for the three lead times used to form the 3-month forecast averages used in Fig. 1. In this format the same scenario is apparent: the month-to-month time resolution indicates a stalling of the dissipation during summer and some persistence of the residual warmth into last few months of 1998.

The observed anomalous SST and subsurface equatorial temperature field for the week centered on March 4 (Fig. 3) shows strong positive sea temperature anomalies from the surface to about 150 m depth in the far eastern tropical Pacific Basin, and relatively weaker and shallower positive anomalies from about 115-180°W. Compared with the situation in late November 1997 (see the December issue of this Bulletin), the positive anomalies have somewhat weakened and have become markedly shallower westward of 115°W. Negative anomalies have been strengthening and expanding eastward from the western Pacific, now extending as far east as 115°W at only 50 m depth, underlying the weakening positive anomalies. This cold water may presage forthcoming conditions in the central and eastern tropical Pacific in fall 1998, based at least on an empirically documented 1 to 1.5 year lag between the anomaly in the subsurface western Pacific and the future SST farther east (Smith et al. 1995). The model forecasts shown here, however, imply that the anomalously cold water will not be replacing the positive anomalies at the surface within the coming 6 months.

Barnston, A.G., H.M. van den Dool, S.E. Zebiak, T.P. Barnett, M. Ji, D.R. Rodenhuis, M.A. Cane, A. Leetmaa, N.E. Graham, C.F. Ropelewski, V.E. Kousky, E. A. O'Lenic and R.E. Livezey, 1994: Long-lead seasonal forecasts--Where do we stand? Bull. Amer. Meteor. Soc., 75, 2097-2114. 

Ji, M., A. Kumar and A. Leetmaa, 1994a: A multi-season climate forecast system at the National Meteorological Center. Bull. Am. Meteor. Soc., 75, 569-577. 

Ji, M., A. Kumar and A. Leetmaa, 1994b: An experimental coupled forecast system at the National Meteorological Center: Some early results. Tellus, 46A, 398-418.  

Ji, M., A. Leetmaa and J. Derber, 1995: An ocean analysis system for seasonal to interannual climate studies. Mon. Wea. Rev., 123, 460-481. 

Ji, M., A. Leetmaa and V.E. Kousky, 1996: Coupled model forecasts of ENSO during the 1980s and 1990s at the National Meteorological Center. J. Climate, 9, 3105-3120.

Livezey, R.E., M. Masutani, A. Leetmaa, H. Rui, M. Ji and A. Kumar, 1997: Teleconnective response of the Pacific-North American region atmosphere to large central equatorial Pacific SST anomalies. J. Climate, 10, 1787-1820. 

Smith, T.M., A. G. Barnston, M. Ji and M. Chelliah, 1995: The impact of Pacific Ocean subsurface data on operational prediction of tropical Pacific SST at the NCEP. Wea. Forecasting, 10, 708-711.

Fig. 1. NCEP coupled model SST anomaly forecast fields for 1998, Apr-May-Jun, Jul-Aug-Sep, and Oct-Nov-Dec 1998. The CMP12 version of the model is used. Each forecast is an average of about 13 individual ensemble members, each based on a different mean of an ensemble of 7 to 11 individual cases, each based on a different one- to two-week-apart initial ocean condition ranging from early February through early April 1998 (see text).

Fig. 2. NCEP coupled model SST anomaly forecast time series for Niño 3 and Niño 3.4 for Jan-Feb-Mar, Feb-Mar-Apr and Mar-Apr-May 1998 (top panels); Apr-May-Jun, May-Jun-Jul and Jun-Jul-Aug 1998 (middle panels); and Jul-Aug-Sep, Aug-Sep-Oct, Sep-Oct-Nov and Oct-Nov-Dec 1998 (bottom panels). The broken line in each panel represents the SST anomaly forecast (^oC), and the solid line the observed SST anomaly. The predictions represent the mean of three ensemble mean forecasts, each for one of the 3 most recent months, respectively, and each produced by forecasts from two to three individual 1 to 2-week-apart initial conditions per month.

Fig. 3. Equatorial depth-longitude section of ocean temperature anomaly with respect to the 1983-92 mean for the week centered on March 4, 1998. Dashed contours denote negative anomalies.