Assessment of Long-term Trends in Extreme Precipitation: Implications of In-filled Historical Data and Temporal Window-Based Analysis

Assessment of long-term trends in extreme precipitation data is critical for hydrologic design. Understanding of such trends is also essential from a climate change perspective and requires an extensive evaluation of trends in different temporal slices of historical precipitation data series. The current study focuses on evaluation of these trends via parametric and non-parametric statistical techniques and impact of in-filled missing precipitation data on the biases of long-term trends. A temporal window based approach is adopted for assessment of the trends in extreme values. The temporal windows are selected in such a way that they coincide with Atlantic Multi-decadal Oscillation (AMO) cycles. The frequency of occurrence of precipitation extremes over a pre-specified threshold is also analyzed. Long-term historical precipitation data available for 100 years at 53 NOAA rain gage stations in the state of Florida are analyzed using filled and un-filled missing precipitation data. Results indicate that in-filled missing precipitation data introduce biases in the statistical trend analysis and provide insights into the variability of precipitation extremes in AMO cool and warm phases along with the changes in the frequency of occurrence of extreme events over a threshold.