Medium term high frequency observation of discharges and suspended sediment in a Mediterranean mountainous catchment

In mountainous catchments, soil erosion and sediment transport are highly variable throughout time and their quantification remains a major challenge for the scientific community. Understanding the temporal patterns and the main controls of sediment yields in these environments requires a long term monitoring of rainfall, runoff and sediment flux. This paper analyses this type of data collected during 7 years (2007-2014), at the outlet of the Galabre River, a 20 km(2) watershed, in south eastern France, representative of meso-scale Mediterranean mountainous catchments. This study is based on a hybrid approach using continuous turbidity records and automated total suspended solid sampling to quantify the instantaneous suspended sediment concentrations (SSC), sediment fluxes, event loads and yields. The total suspended sediment yield was 4661 Mg km(-2) and was observed during flood events. The two crucial periods for suspended sediment transport at the outlet were June and November/December (63% of the total). The analysis of suspended sediment transport dynamics observed during 236 flood events highlighted their intermittency and did not show any clear relationship between rainfall, discharge and SSC. The most efficient floods were characterised by counter-clockwise hysteresis relationships between SSC and discharges. The floods with complex hysteresis were the more productive in the long term, during this measuring period exceeding a decade. Nevertheless, the current research outlines the need to obtain medium-term (five years) continuous time series to assess the range of variations of suspended sediment fluxes and to outline clearly the seasonality of suspended sediment yields. Results suggest the occurrence of a temporal dis-connectivity in meso-scale catchments over short time-scales between the meteorological forcing and the sediment yields estimated at the outlet. These findings have important methodological impacts for modelling and operational implications for watershed management.