QUANTIFYING, MODELING AND PREDICTING SEDIMENT TRANSFER THROUGH A MOUNTAIN BASIN

Mountain basin sediment transfer has been conceptualized as a cascade in which sediment travels through cycles of storage and remobilization before exiting the basin. We developed a probabilistic sediment cascade model, SedCas, based on a sediment budget spanning 5 decades in the Illgraben, a debris-flow (df) prone basin in the Swiss Alps. We use this model to investigate the role of thresholds and storage dynamics in sediment transfer and to predict sediment discharge up to 2050 from downscaled climate simulations.

SedCas conceptualizes the fluvial system as a spatially lumped cascade of connected reservoirs representing hillslope and channel storages from which sediment may be remobilized by surface runoff. Sediment input is drawn from the observed probability distribution of slope failures and the model may be driven by observed or simulated climate. Although processes of sediment transfer and df generation are simplified, SedCas produces complex sediment discharge behavior which is driven by the availability of sediment and antecedent moisture (system memory) as well as climatic triggering.

The model reproduces the first order properties of observed dfs over the period 2000 – 2009 including their seasonality and probability distribution. The main control on the shape of the df distribution is a threshold discharge that defines runoff events with the potential to generate dfs. The stochastic element of hillslope sediment input is important to reproduce realistic sediment storage dynamics and occasional supply-limiting conditions. The correct reproduction of the relative occurrence of dfs for a range of rainfall magnitudes by the model despite not being calibrated to do so, gives us confidence in its use to predict future sediment discharge.

SedCas predicts substantial increases in both annual sediment discharge and the number of large dfs up to 2050 for all three climate models. Particularly large increases in sediment discharge in occur in autumn and spring driven by increases in precipitation falling as rain instead of snow and earlier snow melt. Future work will focus studying the sensitivity of the model to changes in climate versus changes in sediment input. This work is of significance in that it addresses the uncertainties involved in the reconstruction of climate change from sedimentary records.