Article ID Journal Published Year Pages File Type
5771345 Journal of Hydrology 2017 16 Pages PDF
Abstract

•We performed frequency analyses of urban extreme precipitation and runoff events.•Many assumptions on extreme precipitation-runoff relationships are less valid.•There is wide uncertainty ranges due to sampling variability.•The uncertainty has a large effect on reliability of design and management.•Extreme runoff process study is useful for design and management of infrastructure.

There is a need for assessment of uncertainties and hence effects on reliability of design and management of stormwater pipes due to the prevalence of urban floods trigged by modification of land cover and high precipitation intensities respectively due to increasing urbanization and changing climate.Observed annual maximum series (AMS) of extreme precipitation intensities of 17 durations (1-min to 1440-min) and runoff records of 27 years from a 21.255 ha (23% impervious, 35% built-up and 41% open areas) Risvollan catchment in Trondheim City were used. Using a balanced bootstrap resampling (BBRS) with frequency analysis, we quantified considerable uncertainty in precipitation and runoff quantiles due to the sampling variability of systematic observations (e.g., −43% to +49% relative differences from the quantile estimates for the original sample). These differences are higher than suggested increase in design rainfall and floods by many countries for climate change adjustment. The uncertainties in IDF curves and derived design storm hyetographs are found to have large effects on the reliability of sizing of stormwater pipes. The study also indicated low validity of the assumptions on extreme precipitation and runoff relationships in the return period-based method for the partially paved urban catchment: (i) maximum of only 46% of the AMS of extreme precipitation and runoff events occurred concurrently and (ii) T-year return period extreme precipitation events do not necessarily result in T-year flood events. These indicate that there are effects of snowmelt seasonality, and probably catchment moisture states and interactions between the flows in subsurface media and pipes. The results substantiate the need for better understanding of relationships between precipitation and runoff extremes and urban runoff generation process, and importance of uncertainty assessment and application of reliability-based methods for design and management of water infrastructure.

Related Topics
Physical Sciences and Engineering Earth and Planetary Sciences Earth-Surface Processes
Authors
, ,