A nonstationary index flood model for precipitation extremes in transient regional climate model simulations

Abstract

[1] The generalized extreme value (GEV) distribution has often been used to describe the distribution of daily maximum precipitation in observed and climate model data. The model developed in this paper allows the GEV location parameter to vary over the region, while the dispersion coefficient (the ratio of the GEV scale and location parameters) and the GEV shape parameter are assumed to be constant over the region. This corresponds with the index flood assumption in hydrology. It is further assumed that all three GEV parameters vary with time, such that the relative change in a quantile of the distribution is constant over the region. This nonstationary model is fitted to the 1-day summer and 5-day winter precipitation maxima in the Rhine basin in a simulation of the Regional Atmospheric Climate Model (RACMO) for the period 1950-2099, and the results are compared with gridded observations. Except for an underestimation of the dispersion coefficient of the 5-day winter maxima by about 35%, the GEV parameters obtained from the observations are reasonably well reproduced by RACMO. A positive trend in the dispersion coefficient is found in the summer season, which implies that the relative increase of a quantile increases with increasing return period. In the winter season there is a positive trend in the location parameter and a negative trend in the shape parameter. For large quantiles the latter counterbalances the effect of the increase of the location parameter. It is shown that the standard errors of the parameter estimates are significantly reduced in the regional approach compared to those of the estimated parameters from individual grid box values, especially for the summer maxima.

Comparison of regional and at‐site approaches to modelling probabilities of heavy precipitation

“…The HW regional frequency analysis has already been incorporated in the evaluation of precipitation extremes in climate models over the UK (Fowler et al, , 2007 as well as for the construction of their future scenarios Fowler and Ekström, 2009). A similar regional approach to modelling precipitation extremes in climate model simulations, which incorporates also non-stationarity of the model parameters, was developed by Hanel et al (2009) and applied in the Rhine River basin. The ROI methodology may represent a useful alternative or even a step forward, particularly in areas in which the delineation of fixed regions is less clear.…”

Risk and uncertainty in hydrometeorological hazards

“…Cooley 2009;Bjarnadottir et al, 2011) or covarying with global mean temperature or an index of natural variability (e.g. Hanel et al, 2009;Hanel and Buisand 2011).…”

The evaluation of regional frequency analyses methods for nonstationary data

“…O'Brien and Burn (2014) applied index flood model only with timevarying growth curve (Eq. 1) and Hanel et al (2009) used the model with time-varying index flood and growth curve (Eq. 2).…”

“…This assumption tends to be violated due to climate change. Therefore, nonstationary regional frequency analysis (NSRFA), especially nonstationary index flood model (NSIFM), has been recently researched (Cunderlik and Burn, 2003;Cunderlik and Ouarda, 2006;Hanel et al, 2009;Leclerc and Ouarda, 2007;O'Brien and Burn, 2014). The several types of index flood models have been used for NSRFA depending on the time-varying term.…”