Abstract
A multidating approach was carried out on slackwater flood deposits, preserved in valley side rock cave and terrace, of the Gardon River in Languedoc, southeast France. Lead-210, caesium-137, and geochemical analysis of mining-contaminated slackwater flood sediments have been used to reconstruct the history of these flood deposits. These age controls were combined with the continuous record of Gardon flow since 1890, and the combined records were then used to assign ages to slackwater deposits. The stratigraphic records of terrace GE and cave GG were excellent examples to illustrate the effects of erosion/preservation in a context of a progressively self-censoring, vertically accreting sequence. The sedimentary flood record of the terrace GE located at 10 m above the channel bed is complete for years post-1958 but incomplete before. During the 78-year period 1880-1958, 25 floods of a sufficient magnitude (> 1450 m3/s) have covered the terrace. Since 1958, however, the frequency of inundation of the deposits has been lower:only 5 or 6 floods in 52 years have been large enough to exceed the necessary threshold discharge (> 1700 m3/s). The progressive increase of threshold discharge and the reduced frequency of inundation at the terrace could allow stabilisation of the vegetation cover and improved protection against erosion from subsequent large magnitude flood events. The sedimentary flood record seems complete for cave GG located at 15 m above the channel bed. Here, the low frequency of events would have enabled a high degree of stabilisation of the sedimentary flood record, rendering the deposits less susceptible to erosion. Radiocarbon dating are used in this study and compared to the other dating techniques.Eighty percent of radiocarbon dates on charcoals were considerably older than those obtained by the other techniques in the terrace. On the other hand, radiocarbon dating on seeds provided better results. This discrepancy between radiocarbon dates on charcoal and seeds is explained by the nature of the dated material (permanent wood vs. annual production and resistance to degradation process). Finally, we showed in this study that although the most common dating technique used in paleoflood hydrology is radiocarbon dating, usually on charcoal preserved within slackwater flood sediments, this method did not permitus to define a coherent age model. Only the combined use of lead-210, caesium-137, and geochemical analysis of mining-contaminated sediments with the instrumental flood record can be applied to discriminate and date the recent slackwater deposits of the terrace GE and cave GG.