LaDochy et al. Findings: John Monteverdi's Results

1. "...West Coast precipitation anomalies are significantly related to Pacific indices and can be predicted in advance with a high level of confidence..." (Page 10)

"...significantly related..." implies statistical significance

"...Pacific indices..." implies all indices. "...can be predicted in advance with a high level of confidence..." implies statistically significant correlations, accounting for a high percentage of the variance.

• To reproduce results, one must examine the relationship of all Pacific indices to southern California rainfall. We looked at the relationship of three, PDO, NINO3 and SOI.
• To reproduce results, the correlation beteween all three of the indices and rainfall must be statistically signfiicant and account for a high percentage of the varaince.

2. "...the SOI explains nearly 40% of the variability of southern California's annual precipitation..."

This is highlighted in the conclusions, implying that it is the best measure of "foreshadowing" southern California's rainfall, if "...40% of the variability..." means 40% of the variance.

• To reproduce these results, of the three indices we are examining, the SOI should show the strongest correlation (in value, in statistical significance and in percentage of the variance explained).

3. "...with the probable PDO- phase continuing for at least this decade, it would be prudent to expect more cool anomalies with more frequent La Niña events. Overall means would suggest wetter than normal conditions along the north coast, with drier years to the south...."

Implies that PDO is in a negative phase and that it will continue negative. It also implies that the PDO phase is correlated with Southern California rainfall, that the correlation is statistically significant, and that it accounts for a high percentage of the variance.

• To reproduce these results both the climate division precipitation correlation with PDO and the correlation of actual PDO indices with southern California rainfall must show that (a) there is a correlation between the two; (b) that the correlation has a high value; (b) that the correlation is statistically significant; and, (c) that the correlation explains a great percentage of the variance.

1. "...West Coast precipitation anomalies are significantly related to Pacific indices and can be predicted in advance with a high level of confidence. The PDO and NP indices can explain a large percent of variability with a high level of confidence..." (Page 10)

• To reproduce results, one must examine the relationship of all Pacific indices to southern California rainfall. We looked at the relationship of three, PDO, NINO3 and SOI.
• To reproduce results, the correlation beteween all three of the indices and rainfall must be statistically signfiicant and account for a high percentage of the varaince.

2. "...the SOI explains nearly 40% of the variability of southern California's annual precipitation..."

This is highlighted in the conclusions, implying that it is the best measure of "foreshadowing" southern California's rainfall, if "...40% of the variability..." means 40% of the variance.

• To reproduce these results, of the three indices we are examining, the SOI should show the strongest correlation (in value, in statistical significance and in percentage of the variance explained).

SOI

Correlation, -0.55, Significant <0.01; Variance Explained 27%

NINO3

Correlation, 0.55, Significant <0.01; Variance Explained 27%

PDO

Correlation, 0.25, Significance, Borderline 0.05 (actually, does not quite meet the test of significance), Variance Explained 6%

Conclusions: SOI and NINO3 show the best correlations, and these are statistically signficant. However, they account for much less than 40% of the variance in the record. Further, although PDO is correlated with rainfall in southern California, the correlation is not statistically significant or barely statistically significant, and does not account for a significant portion of the variance.

The correlation of actual seasonal PDO values with LA Civic Center seasonal rainfall produces a sample correlation of 0.26 that is significant at the 5% level, but still only accounts for less than 10% of the variance in the record.

LaDochy's findings (1) and (2) above are not duplicated completely, and not at all for the relationship of PDO to southern California rainfall. Of the three indices, SOI and NINO3 are much better "predictors" of seasonal southern California rainfall than seasonal PDO values.

3. "...with the probable PDO- phase continuing for at least this decade, it would be prudent to expect more cool anomalies with more frequent La Niña events. Overall means would suggest wetter than normal conditions along the north coast, with drier years to the south...."

Implies that PDO is in a negative phase and that it will continue negative. It also implies that the PDO phase is (a) correlated with Southern California rainfall; (b) that the correlation is statistically significant; (c) that the correlation accounts for a high percentage of the variance.

• To reproduce these results both the climate division precipitation correlation with PDO and the correlation of actual PDO indices with southern California rainfall must show that (a) there is a correlation between the two; (b) that the correlation has a high value; (b) that the correlation is statistically significant; and, (c) that the correlation explains a great percentage of the variance.

Shows that seasonal rainfall at LA Civic Center has been increasing, along with the variability.

Shows that the 25 yr running means of both PDO and seasonal rainfall at LA Civic Center have recently been increasing.

Shows that the 10 yr running mean of PDO does show a decrease recently, but illustrates the danger in "running" with the small corrrelation of 0.26. Note that the correlation would have worked well as an estimate of the rainfall in any given 10 yr period for up to about 1980, but that the correspondence is now inverse.

Also, note that both on a 25 yr and 10 yr running mean basis, LA Civic Center rainfall is increasing, not decreasing, as the 10 yr running mean of PDO decreases.

The correlation of PDO on seasonal rainfall at LA Civic Center is small, is barely statistically significant, and only accounts for a small percentage of the variance.

LaDochy's findings (3) above are not duplicated, except for the decrease in PDO observed in the 10 yr running means. There is nothing in the results we tried to duplicate that suggests that PDO has any predictive value at all.