13. Nonstationarity in Extremes and Engineering Design

Dealing with nonstationarity in hydrological extremes in the design of structures is a truly multidisciplinary undertaking; requiring expertise in hydrology, statistics, engineering and decision-making. This chapter gives a broad overview over relevant key aspects in these areas including definitions of key words like ‘extremes’ and ‘stationarity’. We briefly cover current knowledge of both climate variability and climate change and effects on hydrological extremes with particular emphasis on precipitation and floods. This is followed by a brief discussion on impacts of hydrological extremes, risk assessments and options for adaptation as well as hurdles. A large part of this chapter is dedicated to new statistical techniques (or extensions of existing techniques) to address nonstationarity in hydrological extremes through the use of time-varying parameters, moments, quantile estimates and the use of covariates. A changing climate may prove impetus to change some of the existing paradigms and explore new avenues. The need to reduce uncertainty, or alternatively derive more reliable uncertainty estimates, is exacerbated in a changing climate. One of the key strategies should be a move from deterministic to probabilistic approaches. Bayesian techniques are a promising framework in this context.