Analyses of extreme flooding in Austria over the period 1951–2006

Abstract

ABSTRACT:Analyses of extreme flooding in Austria is performed using daily discharge time series from 27 stations over the period . The main research questions revolve around: (1) temporal non-stationarities in the flood record, (2) upper tail and scaling properties of the flood peak records, and (3) relation between magnitude and frequency of flooding and the North Atlantic Oscillation (NAO). Two datasets are derived from the daily discharge time series: annual maximum daily discharge and peaks-over-threshold (POT) data. The validity of the stationarity assumption in the annual maximum discharge record is assessed by investigating the presence of abrupt and slowly varying changes using nonparametric tests. The time series are tested for abrupt changes both in the mean and variance of the flood peak distributions by means of the Pettitt test. The presence of monotonic trends is investigated by means of the Mann-Kendall and Spearman tests. Violations of the stationarity assumption are associated with abrupt rather than gradual changes. These step changes generally involve river regulation through construction of dams or other major engineering works. It is not possible to make conclusive statements about the presence of an anthropogenic climate change signal in the flood peak record. Similar conclusions are obtained when focussing on the frequency of POT floods. The Generalised Extreme Value distribution is used to study the upper tail and scaling properties of annual maximum daily discharge records. The location and scale parameters exhibit power-law behaviour as a function of drainage area. The shape parameters indicate that the flood peak distributions for Austria have a heavy tail. Non-stationary modelling of the annual maximum daily discharge and POT time series is used to explore the relation between flood magnitude and frequency and NAO. The results indicate that NAO is a significant covariate in explaining the magnitude and frequency of occurrence of flooding over a large part of Austria.

Inferences on weather extremes and weather-related disasters: a review of statistical methods

“…Of course, there is no problem as long as the processes underlying the data at hand are truly stationary, such as in the study of Wehner (2010) who estimates GEV distributions to pre-industrial control runs from 15 climate models, part of the CMIP3 dataset. The same holds for Villarini et al (2011) who apply GEV distributions for extreme flooding stations with stationary data over time only.…”

“…Loosely formulated, stationarity means that "the data" are stable over time: no trends, breaks, shocks, ramps or changes in variance over time. Methods for assessing stationarity are given by Diermanse et al (2010) and Villarini et al (2011) and references therein.…”

“…For a description in the context of Bayesian statistics please see Renard et al (2006). The frequency of exceedances may be analysed by a nonhomogeneous Poisson process (Caires et al, 2006) or by a Poisson regression model (Villarini et al, 2011).…”

Do climate-informed extreme value statistics improve the estimation of flood probabilities in Europe?

“…Villarini et al (2012) conducted a frequency analysis of annual maximum and peakover-threshold discharge in Austria with NAO as a covariate. López and Francés (2013) examined maximum…”

Comparative study of flood projections under the climate scenarios: links with sampling schemes, probability distribution models, and return level concepts

“…The POT arrival rates have in fact been 161 chronically accepted on faith to follow a homogeneous Poisson process under 162 stationarity (Shane and Lynn, 1964). However, such assumption has been reported to 163 be invalid due to two-type sources of nonstationarity which will be addressed herein: 164 (i) heterogeneity of Poisson process intensity (Cunnane, 1979;Villarini et al, 2012; 165 Silva et al, 2015), for which the Poisson distribution is retained no longer with 166 invariant Poisson process intensity, or rather, parameterized as functions of climatic 167 covariates; (ii) over-dispersion of observations. Theoretically, the Poisson distribution 168 holds identical variance and mean of population, whereas it is often the case that the 169 variance is rarely equal to, and even significantly higher than, the mean (Cunnane, 170 1979).…”