| description abstract | Three long-range forecasting methods have been evaluated for prediction and downscaling of seasonal and intraseasonal precipitation statistics in California. Full-statistical, hybrid-dynamical?statistical and full-dynamical approaches have been used to forecast El Niño?Southern Oscillation (ENSO)?related total precipitation, daily precipitation frequency, and average intensity anomalies during the January?March season. For El Niño winters, the hybrid approach emerges as the best performer, while La Niña forecasting skill is poor. The full-statistical forecasting method features reasonable forecasting skill for both La Niña and El Niño winters. The performance of the full-dynamical approach could not be evaluated as rigorously as that of the other two forecasting schemes. Although the full-dynamical forecasting approach is expected to outperform simpler forecasting schemes in the long run, evidence is presented to conclude that, at present, the full-dynamical forecasting approach is the least viable of the three, at least in California. The authors suggest that operational forecasting of any intraseasonal temperature, precipitation, or streamflow statistic derivable from the available records is possible now for ENSO-extreme years. | |
