Why predicting California winter precipitation is challenging?

While influences of large-scale conditions, such as the El Nino/La Nina, on the year-to-year variability of California (CA) winter precipitation have been well recognized, our predictive skill for winter mean CA precipitation remains limited based on both statistical and dynamical approaches. For example, initialized in October, predicted winter (Nov-Apr) mean CA precipitation based on the coupled dynamical models participating in the North American Multi-Model Ensemble (NMME) Project only show a correlation of about 0.3 with the observations when validated over the last four decades. This analysis is aimed to provide insights into key processes underlying the challenge in predicting CA winter precipitation. Our analysis suggests that only about 25% of interannual variability of CA winter precipitation is contributed by tropical Pacific SST variability associated with El Nino/La Nina conditions; instead, interannual CA precipitation variability is largely associated with anomalous circulation independent from El Nino/La Nina, centered over the west coast US as a part of a short-Rossby wave-train spanning over the North Pacific. A linear regression model suggests that CA precipitation variability can be well predicted given perfect predictions of the winter El Nino/La Nina condition and the circulation anomalies over the west coast US. Predictions based on NMME models, however, suggest that while winter El Nino/La Nina condition can be skillfully predicted when initialized in Oct., these dynamical coupled models show almost no skill in predicting the El Nino-independent winter circulation anomalies associated with CA precipitation, therefore, leading to low predictive skill for CA winter precipitation. Low predictability of these circulation anomalies independent from the El Nino variability is further confirmed by a large number of AMIP-type AGCM simulations. Therefore, improved understanding of the formation mechanisms and large-scale precursors associated with the anomalous circulation independent from El Nino will be critical for a breakthrough in predicting CA precipitation.