Mike Dettinger provides the following abstract from his invited talk to the Geological Society of America Annual Meeting - Reno, Nevada, November 2000.
THE PACIFIC DECADAL OSCILLATION AND FLOOD FREQUENCIES IN THE UNITED STATES
Interdecadal variations of the climate of the North Pacific Ocean basin have been categorized by the first principle component of monthly sea surface temperature patterns there, an index commonly called the Pacific Decadal Oscillation (PDO). As during El Ninos, when PDO is in its positive phase, the central North Pacific is cool and the waters along the west coast of North America are warm; negative PDOs indicate a preponderance of the opposite temperature pattern. The PDO tracks climate variations that?in part?are distinct from, and complementary to, the tropical climate variations of the El Nino-Southern Oscillation (ENSO) process. Together the PDO and ENSO climate variations make significant contributions to year-to-year streamflow variability in North America. In order to identify possible PDO influences on the occurrence of floods, statistics describing, the magnitude of floods the magnitude of floods with 5- to 50-year return intervals have been analyzed in daily streamflow series from over 1100 gauging stations in undisturbed river basins across the conterminous United States.
In 34% of the flow records analyzed, the estimated magnitudes of 5- to 50-year floods, calculated from annual maximum daily flows in years during which a 5-yr moving averaged PDO index is positive, are significantly different (at 99% confidence level) from the corresponding flood magnitudes during years in which the averaged PDO index is negative. When the PDO is in its positive phase, for example, the 20-year floods in the northwestern rivers are +18% larger and the 20-year floods in the southwestern rivers are -34% smaller than when PDO is negative. November-March floods appear to reflect PDO status most in the northwest coastal rivers and through the interior southwest; April-October floods reflect the PDO status most in the interior northwest. In order to use such statistics for prediction of flood probabilities, and in recognition of the rapid response of atmospheric circulations and fluxes to sea surface temperature variations, the corresponding statistical analyses were repeated with a version of the PDO index that can be estimated from immediately antecedent sea surface temperature variations. Flood frequencies are almost as strongly conditioned by this concurrent estimate of the decadal PDO variations as by the retrospective PDO estimates, which encourages a new set of long-lead forecasts of flood-frequency variations in rivers throughout the country.