J1.21 HISTORICAL SEASONAL FORECASTS WITH A SIMPLE GCM Hai Lin 1 , Jacques Derome 1 and Gilbert Brunet 2 1 McGill University, Montreal, Canada 2 Meteorological Service of Canada ABSTRACT A simple General Circulation Model (SGCM) driven by a time-independent forcing is used to perform a series of seasonal predictions. The predictions are made for 51 winter seasons (DJF) from 1948 to 1998. Ensembles of 20 forecasts are produced, with initial conditions of December 1st plus small perturbations. The model uses a forcing field that is calculated empirically from the National Centers for Environmental Prediction (NCEP) / National Center for Atmospheric Research (NCAR) reanalyses. The forcing used for a given winter is the sum of a climatological forcing plus an anomaly that is obtained from the preceding November NCEP data, and that is persisted through DJF. The forecast system does not use any data from the winter months (DJF) being predicted. The ensemble mean prediction for each of the 51 winters is verified against the NCEP reanalysis. The system is found to have statistically significant skill in forecasting the DJF mean 500 hPa height field in areas of the globe that are nearly the same as those of a full GCM, albeit at somewhat reduced levels, but a very much lower computational cost. The skill is observed not only in zero-lead forecasts (for DJF) but also in one- month lead forecasts (for JF). 1. INTRODUCTION Seasonal predictions are normally made either with statistical techniques (e.g., Palmer and Anderson, 1994) or with complex global dynamical models (e.g., Barnett et al., 1994; Derome et al., 2001). The purely statistical approach is hampered by the shortness of the observational record required to train the system. The complex global dynamical models do not suffer from that problem, but they are computationally much more expensive. The present study explores the usefulness of using a middle-ground approach, namely, a simple General Circulation Model (SGCM) driven by empirical forcing functions. We mimic operational forecasting conditions in that the mean DJF conditions are predicted without using any information of the state of the atmosphere or oceans for that period. * Corresponding author address : Hai Lin, Dept. of Atmospheric and Oceanic Sciences, McGill University, Montreal, QC H3A 2K6; e-mail: hai.lin@mcgill.ca 2. THE MODEL AND EXPERIMENTAL SETUP The simple GCM used in this study is the same as described by Hall (2000). It is based on a dry global spectral primitive equation model with linear damping and diffusion, and an empirically derived, predet- ermined, time-independent forcing. The model integrates prognostic equations for vorticity, divergence, temperature and log surface pressure at a horizontal resolution of T21 with 5 equally spaced sigma levels. For each time tendency equation of the model, the forcing is obtained as a residual, after evaluating all the dynamical terms with NCEP data on a daily basis, and time-averaging over a period of one month or one season, to obtain a time-independent forcing. Thus, contrary to a full GCM, in which the “forcing”, such as the diabatic heating, is calculated at every time step, the SGCM uses a predetermined time-independent forcing, hence its much lower computational cost. More details can be found in Hall (2000) and Hall and Derome (2000). The model was shown by Hall (2000) to have a good Northern Hemisphere climatology, not only in terms of the mean zonal wind and standing waves, but also in terms of the transient eddy statistics. All forecasts were made for the winter (DJF) season. We first computed the mean DJF forcing fields with NCEP data for each of the 51 winters (1948-1998) and the corresponding 51 mean-November forcing fields. When forecasting a given DJF we drove the model with a forcing constructed as follows. We first computed the anomaly in the November forcing of that particular year, defined as a deviation of that November forcing from the climatological forcing (the average over the 51 Novembers). We added this anomaly to the climatological DJF forcing (the average over the other 50 DJFs). The sum of these two forcings was then used throughout DJF. For each of the 51 winters, an ensemble of 20 forecasts was made. The initial cond- itions for the ensemble members were taken to be the December 1 st analysis plus small perturbations. 3. RESULTS The predictive skill is measured by the temporal correlation over the 51 winters between the predicted ensemble mean and observed seasonal averages. Fig.1 shows this skill for the 500 hPa geopotential height. The shaded areas have a significance level of 0.05 or better according to a Student-t test. Skilful DJF predictions are found over all the tropics and parts of the North Pacific, North America and eastern Asia.