Geographic Distribution of Trends in Tropical Pacific Rainfall Extremes

It is known that global warming has an effect on the prevalence of precipitation extremes, with both high-precipitation events and droughts becoming more common. These changes in rainfall distribution can have significant hydrologic consequences even if the average precipitation in an area does not change. Areas with limited groundwater and surface water capacity will face the greatest consequences, which describes most of the islands in the central Pacific Ocean.

Previous work has examined rain gauge data from the Pacific Rainfall Program (PACRAIN) and selected model results from the Climate Model Intercomparison Project (CMIP) with the goal of identifying trends in precipitation extremes in the central Pacific. Parameters looked at included rainfall intensity, the occurrence of high-precipitation events, and the length of dry periods. Trends in these parameters were suggestive of more extreme events (more intense rainfall separated by longer dry periods). However, most of these trends were not statistically significant and lacked the coherent spatial patterns that would be indicative of changes in large-scale physical mechanisms.

Work continues to refine these trend estimates. The Mann-Kendall test used to test for trend significance thus far is not as robust for highly periodic time series like rainfall, so other techniques are being explored. Once more reliable trend results have been calculated cluster analysis will be used to identify the spatial patterns in those results.