EL NIÑO AND FLOODS

The relation between floods and El Niño or La Niña was examined in the modern and paleoflood records in select regions where seasonal precipitation shows a strong correlation with either the warm or cool phase of El Niño/Southern Oscillation (ENSO). Winter precipitation values from the northwestern and southwestern U.S. exhibit strong, opposite correlations with ENSO. In the Southwest, most floods occur during periods with a persistent negative Southern Oscillation Index (SOI), and the single largest flood on each river in this study was associated with a moderate to strong El Niño. In the northwestern U.S., although seasonal winter streamflow is strongly correlated with La Niña and a positive SOI, the largest individual floods in the historical record are equally split between La Niña and neutral SOI years and tend to cluster into decades dominated by moderate SOI values. In regions of India where annual monsoon precipitation correlates with years of positive SOI, large floods show a more consistent relation with decadal-scale periods of high ENSO variability than with interannual variations in the SOI.

The connections between paleofloods and Holocene ENSO variations is mixed. Paleoclimatic records from both sides of the tropical Pacific indicate the onset of a fully developed ENSO system by 5-7 ka BP. This change coincides with the earliest preserved paleoflood deposits at most study sites in North America and India, which may indicate the beginning of the modern hydroclimatic regime in these areas, including an increased climatic variability linked to ENSO. An increase in El Niño events ca. 3 ka BP does not coincide with peaks in the paleoflood records from any of the study areas. Over the last 1000 years, periods of frequent paleofloods in the southwestern U.S. correspond well with periods of more frequent El Niños. In the Northwest, paleoflood frequency on the Columbia River appears to be opposite that of the Southwest, while in other subregions the paleofloods are either randomly distributed or show similarities with the timing of the southwestern floods. In both the modern and paleoflood records, the effect of ENSO on annual precipitation is not a guide to the response of floods unless the specific atmospheric circulation patterns responsible for producing large floods are directly influenced by ENSO.