1. The recent news reports that the negative PDO indices in the last few years is a good indicator of below average rainfall over Southern California in the next five years is not entirely correct, and this can be proven through statistical analysis. The same analysis will show that the correlation between El Niño and Southern California rainfall would be a much better predictor of Southern California rainfall.

To check and see if PDO is a good predictor of Southern California rainfall, the seasonal rainfall at LA Civic Center and actual PDO values between the years of 1921-2000 were examined. The correlation was found to be 0.26, statistically significant at the 5% level, but accounts for only 6% of the variance of the record. A graph of the 10-year running mean of PDO was overlain with a graph of the LA Civic Center seasonal rainfall for the years 1921-2000 (Fig. 1). The graph shows that the correlation between PDO and seasonal rainfall would have worked up to 1980, but since then the correlation has been negative. In fact, since 1990 seasonal rainfall has been steadily increasing while PDO has been decreasing. The graph is proof that the correlation of 0.26 is small, marginally significant and accounts for a small percentage of the variance.

Using US climate division correlation maps from the CDC (Fig. 2), it was found that the correlation of seasonal precipitation over Southern California with NINO3 is 0.55, which meets the test of significance (<0.01). Despite being statistically significant, the variance of 27% accounts for less than 40% of the variance in the record, but yet is a much better predictor than PDO.

2. From the analysis above it can be concluded that in the next five years, it would not be prudent to say that rainfall over Southern California will be below average despite the predicted negative PDO indices for the next five years. It would be a safer assumption that the next five years Southern California will experience near normal rainfall, and that any reports of an El Niño during the winter months would be a better indicator of above average rainfall.

Fig. 1

Fig. 2