What are the main regional and temporal controls of the hydro-sedimentary response of European and US catchments during floods ?

Floods are among the most common natural disasters worldwide. Caused by heavy rainfall events, they are responsible for significant damage to populations and infrastructure. Climate change is likely to increase the frequency and intensity of these extreme precipitation events. Furthermore, floods are often associated with significant sediment transport, which tends to increase damage to infrastructure and poses a threat to agriculture and river ecosystems. As a result, numerous studies have been conducted to estimate sediment production and transport, primarily using empirical models and, more recently, physically based models. However, whether empirical or physically based, it is generally difficult to estimate the parameter values of these models. Very few sediment transport measurements are available, and these generally consist of turbidity and/or suspended sediment concentration measurements at a single point, which do not provide information on the spatial variability of the processes at work.

The main objective of this study is to investigate the possible correlations between the characteristics of a catchment, such as its topography, soil texture, morphology or land-use, and its hydro-sedimentary response during flash flood. Calibration of hydro-sedimentary models could also benefit from knowledge of these correlations.

This work is based on open access databases containing discharge and suspended sediment concentration time series collected for more than 100 European and American catchments, together with complementary data on catchment characteristics. These databases are used to calculate catchment-scale indicators such as average slope, connectivity, percentage of sand or clay, precipitation intensity, etc. In addition, we characterized the hydro-sedimentary response using several signatures extracted from the time series of discharge and suspended sediment concentration, such as peak discharges, volume of sediment transported per year and per event, etc. Then, we used Spearman-rank correlations to measure the strength of the links between catchment characteristics and its hydro-sedimentary signatures. We calculated these correlations at different temporal and spatial resolutions to investigate whether the strength of these correlations is scale-dependent.

Preliminary results show a strong control of hydraulic connectivity and precipitation intensity on eroded volumes. Relationships already reported in the literature are also observed here, such as those between the Lloyd index and fine particle content in soils. Contrasting relationships are also found depending on the size of the catchments or their topography. This multi-scale approach could provide a more detailed understanding of sediment mobilisation and deposition mechanisms, and consequently suggest relevant indicators to characterise the sensitivity and origin of soil loss.

Further research includes using random forests to rank the catchment characteristics based on their importance in controlling the hydro-sedimentary signatures.