Abstract
Abstract. Many recent studies have sought to characterize variations of the annual maximum flood discharge series over time and across space in Europe, including some that have elucidated different process controls on different statistical properties of these series. To further support these studies, we conduct a pan-European assessment of process controls on key properties of this series, including the mean annual flood (MAF) and coefficients of variation (CV) and skewness (CS) of flood discharges. These annual maximum flood discharge series consist of instantaneous peaks and daily means observed in 2370 catchments in Europe without strong human modifications covering the period 1960–2010. We explore how the estimated moments MAF, CV and CS vary due to catchment size, climate and other controls across Europe, where their averages are 0.17 m3 s−1 km−2, 0.52 and 1.28, respectively. The results indicate that MAF is largest along the Atlantic coast, in the high-rainfall areas of the Mediterranean coast and in mountainous regions, while it is smallest in the sheltered parts of the East European Plain. The CV is largest in southern and eastern Europe, while it is smallest in the regions subject to strong Atlantic influence. The pattern of the CS is similar, albeit more erratic, in line with the greater sampling variability of CS. In the Mediterranean, MAF, CV and CS decrease strongly with catchment area, suggesting that floods in small catchments are relatively very large, while in eastern Europe this dependence is much weaker, mainly due to more synchronized timing of snowmelt over large areas. The process controls on the flood moments in five predetermined hydroclimatic regions are identified through correlation and multiple linear regression analyses with a range of covariates, and the interpretation is aided by a seasonality analysis. Precipitation-related covariates are found to be the main controls of the spatial patterns of MAF in most of Europe except for regions in which snowmelt contributes to MAF, where air temperature is more important. The Aridity Index is, by far, the most important control on the spatial pattern of CV in all of Europe. Overall, the findings suggest that, at the continental scale, climate variables dominate over land surface characteristics, such as land use and soil type, in controlling the spatial patterns of flood moments. Finally, to provide a performance baseline for more local studies, we assess the estimation accuracy of regional multiple linear regression models for estimating flood moments in ungauged basins.
Highlights
Understanding the spatial distribution of statistical flood characteristics is important from both practical and scientific perspectives, assisting in estimating design floods in gauged and ungauged catchments and shedding light on the regional processes of flood generation from a probabilistic perspective (Rosbjerg et al, 2013).Much research has been conducted on identifying process variables and mechanisms controlling the magnitudesPublished by Copernicus Publications on behalf of the European Geosciences Union.D
The results indicate that mean annual flood (MAF) is largest along the Atlantic coast, in the high-rainfall areas of the Mediterranean coast and in mountainous regions, while it is smallest in the sheltered parts of the East European Plain
Over the entire dataset, the mean specific annual flood is 0.17 m3 s−1 km−2, while the mean specific annual flood normalized to a catchment area of 100 km2 (MAFα) is 0.21 m3 s−1 km−2
Summary
Understanding the spatial distribution of statistical flood characteristics is important from both practical and scientific perspectives, assisting in estimating design floods in gauged and ungauged catchments and shedding light on the regional processes of flood generation from a probabilistic perspective (Rosbjerg et al, 2013).Much research has been conducted on identifying process variables and mechanisms controlling the magnitudesD. Catchment area is usually the main control on the specific mean annual flood (MAF) as smaller basins tend to have larger specific MAFs than larger ones (Rosbjerg et al, 2013) because a large basin is less likely to be fully covered by a storm than a small basin. This tends to reduce the variance of extreme catchment-average precipitation and the MAF (Viglione et al, 2010). Based on an analysis of flood data in Slovakia, Austria and Italy, Salinas et al (2014) found both correlations of CV and skewness (CS) to decrease with catchment area, which they interpreted as the result of aggregation effects of the spatial heterogeneity of rainfall and the interaction between the spatial and temporal scales of rainfall and catchment size
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
