Changes in intensity-frequency-duration relationship of heavy rainfalls at a station in Melbourne

Abstract

A primary concern in the context of climate change impacts is related to the increase in the intensity and frequency of heavy rainfall events. The risk of heavy rainfall events (i.e. floods and drought) has increased over the world, and the intensity and frequency of heavy rainfall events are very likely to increase in the future with the exception of the regions with very significant decreases in rainfall. More intense and frequent heavy rainfalls due to climate change and variability are a major concern for urban and rural flooding that causes damage to hydraulic structures such as dams, roads and stormwater drainage systems. Potential impacts of climate change on heavy rainfalls question the accuracy of current design rainfalls, which are used in estimating the hydraulic capacity of these structures. If there is a significant trend in heavy rainfall events, currently used design rainfalls obtained through Intensity-Frequency-Duration relationship may cause hydraulic failures and flood damages in case of underestimation. On the other hand, it would be a waste of money in the event of overestimation of design rainfalls. In this study, first heavy rainfall trends for daily and sub-daily storm durations were investigated. Rainfall data from the Melbourne Regional Office station over the period of 1880-2010 were used for trend analysis of heavy rainfalls. The non-parametric rank based Mann-Kendal and Spearman's Rho tests were selected for trend analysis in this study due to distribution free characteristic of these tests. The Mann-Kendal and Spearman's Rho tests were applied to detect trends in annual maximum rainfall intensities of the selected storm durations (i.e. 6, 12, 18 and 30 minute, and 1, 2, 3, 6, 12, 24, 48 and 72 hour). The trend tests showed statistically significant increasing trends for sub-hourly storm durations (6, 12, 18 and 30 minute). Also, the data sets of 1, 2 and 3 hour storm durations indicated increasing trends, but not significant even at 0.1 significance level. On the other hand, 6, 12, 24, 48, and 72 hour heavy rainfall data sets showed statistically insignificant decreasing trends. Then, the changing character of rainfall Intensity-Frequency-Duration (IFD) relationships were quantified through the analysis of four time slices (i.e. 1880-1909, 1910-1949, 1950-1979 and 1980-2010) using rainfall data from the Melbourne Regional Office station in Melbourne. The results of the frequency analysis of heavy rainfalls for 4 time slices showed that the IFD relationships vary over the period 1880-2010. Remarkable increases in sub-hourly design rainfall intensities were detected for the time slice 1980-2010. However, there is no evidence to state that increases in design rainfall intensities are statistically significant. This study aims to demonstrate a methodology for the investigation of heavy rainfall trends and IFD relationship using rainfall data from a single station in Melbourne. Therefore, it is not realistic to extrapolate the findings of this study without further analysis using data of the many stations.