Ensemble Cannonical Correlation Prediction of Seasonal Precipitation Over the US

This paper presents preliminary results of an ensemble cannonical correlation (ECC) prediction scheme developed at the Climate and Radiation Branch, NASA/Goddard Space Flight Center for determining the potential predictability of regional precipitation, and for climate downscaling studies. The scheme is tested on seasonal hindcasts of anomalous precipitation over the continental United States using global sea surface temperature (SST) for 1951-2000. To maximize the forecast skill derived from SST, the world ocean is divided into nonoverlapping sectors. The cannonical SST modes for each sector are used as the predictor for the ensemble hindcasts. Results show that the ECC yields a substantial (10-25%) increase in prediction skills for all regions of the US and for all seasonal compared to traditional CCA prediction schemes. For the boreal winter, the tropical Pacific contributes the largest potential predictability to precipitation in the southwestern and southeastern regions, while the North Pacific and the North Atlantic are responsible for enhanced forecast skills in the Pacific Northwest, the northern Great Plains and Ohio Valley. Most importantly, the ECC increases skill for summertime precipitation prediction and substantially reduced the spring predictability barrier over all regions of the US continent. Besides SST, the ECC is designed with the flexibility to include any number of predictor fields, such as soil moisture, snow cover and regional regional data. Moreover, the ECC forecasts can be applied to other climate subsystems and, in conjunction with further diagnostic or model studies will enables a better understanding of the dynamic links between climate variations and precipitation, not only for the US, but also for other regions of the world.