Trend Analysis for Precipitation Frequency Estimates in Puerto Rico

Pedro I. Matos-Llavona University of Puerto Rico at Mayagüez Hydrometeorological Design Study Center / NOAA

Abstract: In this project, a trend analysis was performed for Annual Maximum Series in Puerto Rico Region 1 (South of PR). The objective was to verify if there are significant differences between the precipitation frequency estimates (PFE) with the trend analysis incorporated to the statistical method and PFE without the trend analysis. To ensure the significance of the trend in mean of the magnitude, a t-test was performed for the 19 stations of region 1; and 8 of them showed a positive trend in mean with a 5% significance level. For these 8 stations, PFE was calculated by two different methods. The statistical method Maximum Likelihood estimates (MLE) method without trend analysis and the MLE method with the trend analysis incorporated (MLE(t)). MLE(t) always resulted in significantly lower return periods for extreme precipitation events than MLE. This study has proven the significance of trend analysis for regions with a cluster of stations with observed significant trend in mean. With a changing climate, it is practical to apply this analysis as a next step to climate adaptation. In addition to this research, the 24-hour Annual Exceedance Probability (AEP) was analyzed for an extreme precipitation event in Puerto Rico on October 6-7, 1985. This research will contribute to the accuracy of estimates for a better prediction and the protection of life and property.