Comparison of scenario-neutral approaches for estimation of climate change impacts on flood characteristics

Scenario-neutral assessments of climate change impact on floods analyse the sensitivity of a catchment to a range of changes in selected meteorological variables such as temperature and precipitation. The key challenges of the approach are the choice of the meteorological variables and statistics thereof and how to generate time series representing altered climatologies of the selected variables. Different methods have been proposed to achieve this, and it remains unclear if and to which extent they result in comparable flood change projections. Here, we compare projections of annual maximum floods (AMFs) derived from three different scenario-neutral methods for a prealpine study catchment. The methods chosen use different types of meteorological data, namely, observations, regional climate model output, and weather generator data. The different time series account for projected changes in the seasonality of temperature and precipitation, in the occurrence statistics of precipitation, and of daily precipitation extremes. Resulting change in mean AMF peak magnitudes and volumes differs in sign between the methods (range of -6% to +7% for flood peak magnitudes and -11% to +14% for flood volumes). Moreover, variability of projected peak magnitudes and flood volumes depends on method with one approach leading to a generally larger spread. The differences between the methods vary depending on whether peak magnitudes or flood volumes are considered and different relationships between peak magnitude and volume change result. These findings can be linked to differing flood regime changes among the three approaches. The study highlights that considering selected aspects of climate change only when performing scenario-neutral studies may lead to differing representations of flood generating processes by the approaches and thus different quantifications of flood change. As each method comes with its own strengths and weaknesses, it is recommended to combine several scenario-neutral approaches to obtain more robust results.