As warmer air has a higher capacity to carry moisture in form of water vapor, future climate raises the likelihood for strong rainfall events and particularly towards extremes. The 10-year return period rainfall episodes, such as the 5-day cumulative rainfall, is a good measure of these extremes. In many places around the world, the maximum expected amount of rainfall in a 10-year period is projected to increase, which can lead to flooding. As a result, power production can be largely affected. For example, the transportation lines for fuel can be interrupted by local flood, or distribution networks can be disturbed by excessive rainfall and flooding.
The boxplot shows recorded 5-Day Cumulative Rainfall for 1986-2005 and projected 5-Day Cumulative Rainfall 10-yr Return Level by 2050 under all RCPs of CIMP5 ensemble modeling. This indicator focuses on the maximum 5-day cumulative rainfall amount that can be expected within a 10-yr period.
Extremes are often behaving differently than the mean of a process. In many regions, the largest events are bigger than what would be expected when looking at more frequent events. Therefore, special treatment is necessary to capture the heavy tail of extreme rainfall. 10-yr return events are rare enough to warrant special analysis based on extreme value theory. Here, monthly maxima of the 5-day cumulative rainfall over any 20-year window were calculated and the values used to derive the extreme value parameters. Note, using just 20-year time series is somewhat short to best measure rare events, and thus the values might be somewhat noisy. For specific applications, a more targeted local study with the longest possible records should be performed and the local confidence interval should be consulted next the return levels.