The Ensemble Canonical Correlation Analysis (ECCA) forecast is one of the statistical seasonal tools that predict US surface temperature and precipitation. The ECCA uses Canonical Correlation Analysis (CCA), an empirical statistical method that finds patterns of predictors (variables used to make the prediction) and predictands (variables to be predicted) that maximize the correlation between them. The most recent available predictor data for different atmospheric/oceanic variables are projected onto the loading patterns to create forecasts. The ensemble refers to forecasts produced by using each predictor separately to create a forecast. The final forecast is an equally weighted average of the ensemble of forecasts. The model is trained from 1953 to the year before the present year to create the loading patterns.
Colors on the ECCA forecast maps denote the following :
Orange - above normal temperature ; Blue - below normal temperature
Green - above normal precipitation ; Brown - below normal precipitation
Predictor Selection :
The pool of possible predictors used in the forecasts are:
200mb global velocity potential global sea surface temperatures sea level pressure (north of 40N) US soil moisture
The predictors selected to be used in the ECCA are based on factors such as :
a) which climate signals/atmospheric variables play a large role in US temperature/precipitation for a certain season (ie. soil moisture is included in summer forecasts
b) status or strength of climate signals that impact US temperature and precipitation. For example, seasons with current or expected relatively strong ENSO years typically include sea level pressure as one of the ECCA predictors because of its ability to represent the ENSO signal.
The list of predictors used to make a certain lead forecast are listed above the forecast images on the ECCA forecast pages
Mo, K.C., 2003: Ensemble Canonical Correlation Prediction of Surface Temperature over the United States. J. Climate , 16, 1665-1683.
Barnston, A.G.: Linear Statistical Short-Term Climate Predictive Skill in the Northern Hemisphere. J. Climate , 7, 1513-1564.