ABSTRACT
This chapter reviews the possibilities and limitations for detecting and attributing changes in fluvial flood hazard and risk. The flood water level, the extent of inundation areas and the amount of flood damage are a result of complex physical and societal processes. The most often applied indicators in flood hazard trend studies are based on discharge observation. Among the discharge-based indicators, most studies use annual maximum streamflow. A severe problem in flood trend studies is the difficulty of distinguishing between changes as a consequence of natural climate variability and anthropogenically-induced climate change. There are change detection methods appropriate for normally distributed data, skewed data, for testing trends in different statistical moments or in autocovariance. Climate-related variables inherently display changes that are ascribed to natural variability of the climate system. Few studies analyse changes in vulnerability, flood damage and risk. Detection and attribution of past changes in flood hazard and risk are an open question.
