Hydrological signatures of flood trends on German rivers: Flood frequencies, flood heights and specific stages
Highlights
► We systematically investigated stage and discharge time series of German rivers. ► Data were analysed for trends in flood stages, discharges, frequency and specific stages. ► No significant discharge trends were identified consistently for German rivers. ► Several river gauges with significant specific-gauge trends were identified. ► The identified trends in have contributed to past changes in flooding.
Introduction
Recent research has focused on assessing the impacts of diverse environmental and anthropogenic changes upon rivers. Both climate change (e.g., Petrow and Merz, 2009) and basin land use (change) are key factors controlling the hydrological behaviour of catchments and their discharge dynamics (Hörmann et al., 2005). Additionally, flood stages can be affected by changes in the stage-discharge relationship and therefore be influenced by local in-stream river modifications such as engineering measures (Pinter et al., 2006a, Pinter et al., 2006b). Pinter et al. (2006a) identified site-specific trends along the Rhine and Mississippi rivers and compared the contrasting river histories and management strategies to explain differences between the two river systems.
Following several severe floods in Germany during the past two decades (e.g., Rhine 1993 and 1995; Odra 1997; Elbe 2002 and 2006; Danube 1999, 2002 and 2006), mass media as well as scientists have debated the relative contributions of climate and/or anthropogenic processes to those floods. Mudelsee et al., 2003, Mudelsee et al., 2004 did not see any clear increase in flood occurrence rates for the Elbe and Odra rivers. Instead, based on long-term time series, they found decreasing winter flood occurrence, while summer floods showed no trends. Petrow and Merz (2009) analysed trends in flood peaks for German rivers between 1951 and 2002, from which they identified significant changes in flood behaviour in western, southern and central Germany (mostly increasing flood indicators). In northeast Germany, in contrast, they did not identify flood-related changes. As Radziejewski and Kundzewicz (2004) noted, however, failure to identify statistically significant trends in hydrological time series is different from proving that no trends are present. Trends may indeed be present but may not meet the threshold for significance depending on the magnitude of the trends, the interannual variability in the data, the length of the time series, as well as the choice of start- and end-dates of the time series.
Other studies have analysed trends in flood indicators in different parts of the world. For example, Douglas et al., 2000, Burn and Hag Elnur, 2002 and Adamowski and Bocci (2001) found few and only heterogeneous (some increasing and some decreasing) trends in maximum annual floods in North America (US and Canada). This previous result contrasts with the findings of some other researchers in the US (e.g., Groisman et al., 2001, Ya et al., 2004). In the UK, Robson et al. (1998) did not find any significant trends for annual time series, whereas in western Australia, Franks (2002) identified a marked increase in floods after 1945. In summary, flood-related trends might be detectable in some regions of the world, but they are often non-uniform, and trends at many stations are not statistically significant (Kundzewicz et al., 2005, Svensson et al., 2005).
All these studies investigated flood discharge (e.g., time series of annual maxima or peak-over-threshold flows), but few studies have rigorously analysed the related flood stages. Stage, or water-surface elevation, is traditionally viewed as a secondary factor, dependent largely upon discharge and therefore typically not the focus of independent time-series analysis. The public, however, perceives flood events most urgently in terms of flood stage and not (only) in terms of discharge. Different factors affect local rating curves and therefore stages over time, so that stage time series can provide discharge-independent information. For the Rhine River, Pinter et al. (2006b) found changes due to channel incision and to river engineering over the past ∼100 years; these changes on the Rhine were small compared to much larger alteration of stage-discharge relationships on navigable rivers in the US (Pinter et al., 2006a, Pinter et al., 2008, Pinter et al., 2010).
In this study, long time series of stage and discharge data from 78 river gauges in Germany are analysed for trends in flood frequency, peak discharge, peak stage and stage-discharge relationship. According to Blöschl and Montanari (2010), simultaneously analysing many catchments in a large region should help to reduce selection bias with respect to flood trends. These analyses involved analysis of flood stages in addition to discharges and investigated river engineering effects by quantifying changes in the rating curve using specific-gauge analysis. We looked for both local changes and broader regional patterns in order to try to explore the causal mechanisms driving the identified changes.
Section snippets
Data
For this study, daily river flows and stages from 78 German river gauges (Table 1 and Fig. 1) were obtained from the Global Runoff Data Centre (GRDC, Koblenz; discharges) and from the Bundesanstalt für Gewässerkunde (Koblenz; stages). For all gauges, time series of either stage or discharge or both variables have a minimum length of about 50 years. Before performing the analysis, all data sets were checked for completeness, outliers, and consistency in gauge datum and location. Changes in gauge
Historical changes in German river catchments
In order to relate hypothesized trends in flooding to environmental or anthropogenic changes, a general review of historic changes in climate, land use and river regulation is required.
Homogeneity of time series
According to the homogeneity tests, most of the time series of discharge (56 out of 78 gauges in total) and stage (47 out of 78) were found to be homogenous, while 17 discharge and 18 stage time series were inhomogeneous. For 5 discharge and 13 stage data sets, the two homogeneity tests suggested contrasting results.
The homogeneity tests revealed higher percentages of inhomogeneity for those gauges with significant trends in discharge or stage: 9 homogenous versus 16 inhomogenous discharge time
Discussion
This study confirmed that – in agreement with previous findings (e.g., Petrow and Merz, 2009) – there is no uniform pattern in hydrological time series across all German rivers. We examined both discharge and stage time series and found, for example, predominantly decreasing trends for the Weser catchment, while annual maximum flows in the Rhine and the Main catchments increased significantly. However, for 42 (OLS; Mann–Kendall: 45) out of 78 stations no statistically significant trends were
Conclusion
This study presents a systematic investigation of flood-related trends in German rivers. Based on analyses of trends in flood frequency, flood height, and annual maximum discharges and based on specific gauge analysis, some findings of previous studies were confirmed and others supplemented. Most stations analysed on the German rivers did not show statistically significant trends in any of the metrics analysed here. Similar to previous studies, we confirm significant trends at about 30% of the
Acknowledgements
Funding for this research was provided by the Alexander von Humboldt Foundation. We also thank the Global Runoff Data Centre (GRDC) and the Bundesanstalt für Gewässerkunde (Koblenz) for providing data for this study.
References (53)
- et al.
Impact of landscape management on the hydrological behaviour of small agricultural catchments
Phys. Chem. Earth (B)
(1999) - et al.
Detection of hydrologic trends and variability
J. Hydrol.
(2002) Recent winter floods in Germany caused by changes in the atmospheric circulation across Europe
Phys. Chem. Earth
(1995)- et al.
Trends in floods and low flows in the United States: impact of spatial correlation
J. Hydrol.
(2000) - et al.
Climate development in the last century – global and regional
Int. J. Med. Microbiol.
(2008) - et al.
Statistical analysis of the flood situation and assessment of the impact of diking measures along the Elbe (Labe) river
J. Hydrol.
(2002) - et al.
The evaluation of land-use options in mesoscale catchments – prospects and limitations of eco-hydrological models
Ecol. Model.
(2005) - et al.
Landuse impacts on storm-runoff generation: scenarios of land-use change and simulation of hydrological response in a meso-scale catchment in SW-Germany
J. Hydrol.
(2002) - et al.
Trends in flood magnitude, frequency and seasonality in Germany in the period 1951–2002
J. Hydrol.
(2009) - et al.
Hydrodynamic and morphodynamic response to river engineering documented by fixed-discharge analysis, Lower Missouri River, USA
J. Hydrol.
(2005)
Trends in flood stages: contrasting results from the Mississippi and Rhine River systems
J. Hydrol.
Geostatistical regional trend detection in river flow data
Hydrol. Process.
A homogeneity test applied to precipitation data
J. Climatol.
Detection of climate change in Europe by analysing European atmospheric circulation patterns from 1881 to 1989
Theor. Appl. Climatol.
Climate change impacts – throwing the dice?
Hydrol. Process.
Changing runoff regimes of German rivers due to climate change
Erdkunde
Historical changes of the River Weser
Markieren die Winterhochwasser 1990 und 1993 das Ende der Stationarität in der Hochwasserhydrologie infolge von Klimaänderungen?
Wasser Boden
Umweltnutzung und Wirtschaft. Tabellen zu den Umweltökonomischen Gesamtrechnungen. Ausgabe 2010
Stromregelungen und Ausbau der Elbe
Wasserwirtsch. Wassertech.
Identification of a change in climate state using regional flood data
Hydrol. Earth Syst. Sci.
Heavy precipitation and high streamflow in the contiguous United States: trends in the twentieth century
Bull. Am. Meteorol. Soc.
Hydrolologic history of the Mississippi and lower Missouri rivers based upon a refined specific-gauge approach
Hydrol. Process.
Cited by (76)
A framework approach to address the trend and causes of flood stage change in a river reach downstream of a dam influenced by tributaries
2023, International Journal of Sediment ResearchTime-series analysis of Sentinel-1/2 data for flood detection using a discrete global grid system and seasonal decomposition
2023, International Journal of Applied Earth Observation and GeoinformationHydrologic and geomorphic effects on the reduction of channel discharge capacity in the Middle Yangtze River
2023, Science of the Total EnvironmentShallow lake water exchange process before and after water diversion projects as affected by wind field
2021, Journal of HydrologyCitation Excerpt :In this study, we investigated the water exchange process before and after the WDPs with the influence of hydrological characteristics and external forcing. First, in terms of the hydrological characteristics, we analysed the periodicity of long-series water-level data and the typical annual discharge of the main rivers that enter the lake, which laid the foundation for the analysis of the lake water exchange (Aronica and Candela, 2007; Bormann et al., 2011; Gualtieri et al., 2018). Second, in terms of the external wind field, the distribution of the uneven wind field in Dianchi Lake was further explored, which can lay a foundation for further studies.
Homogenization and polarization of the seasonal water discharge of global rivers in response to climatic and anthropogenic effects
2020, Science of the Total Environment