Floodplain deposits: A geochemical archive of flood history – A case study on the River Rhine, Germany
Based on the correlation between discharge and carbonate content of the suspended load of the River Rhine, Germany, a systematic geochemical, mineralogical and granulometric study was carried out to verify whether this geochemical signal is transferred to floodplain deposits and in what way these sediments and their chemostratigraphic characterization can be used as a tool for the reconstruction of the river flood history. The analysis of the time resolved changes in the composition of particulate matter during a flood event revealed that the increase of carbonate content (represented by CaO, Sr) with discharge was coupled to a simultaneous decrease in the relative amount of siliciclastics (K2O, Rb). The association of these two groups of diametrical parameters with specific grain size fractions (carbonates with 40–200 Î¼m; siliciclastics with >200 Î¼m) were found to be slightly shifted relative to each other and showed different gradients during the surging and fading flood wave. This, together with the covariance of elements pertaining to minerals with different density (e.g., carbonate and heavy minerals) suggests a chemical response to the changes in discharge, which is controlled primarily by hydraulic equivalence rather than grain size. There is also a time lag between the amount of suspended load and discharge, with a maximum in suspended load shortly after the peak discharge, when the flood has already started to abate. The flood plain sediments have similar composition to the suspended load, suggesting the direct transfer of the geochemical flood signal to the floodplain sediments. Three lithological units could be distinguished in a 240 cm long sediment core collected from the floodplain. Grain-size and geochemical composition indicate that only the top of the section (20–70 cm) represents sediments deposited in an abandoned channel of the river and may have preserved the geochemical flood signals as identified during actual flood events. The comparison of the chemostratigraphy of this part of the section (estimated with optical stimulated luminescence to be deposited between ca. 1650 and 1920), with a record of historically documented inundations, however, yielded only poor agreement. The possible reasons for this inconsistency are discussed and suggestions are made for continuing investigations. âº River discharge and composition of suspended material are correlated. âº Compositional gradients in floodplain sediments are correlative with flood intensities. âº Comparison of OSL ages and historically documented flood events.