Data‐based modelling of runoff and chemical tracer concentrations in the Haute‐Mentue research catchment (Switzerland)

This paper presents a model for simulating discharge as well as chemical tracer concentration (silica and calcium) in stream flow for the Haute‐Mentue research basin (Switzerland). The model structure is based on a parameterization of the three components (acid soil, AS; direct precipitation, DP; deep groundwater, GW) of a hydrochemical mixing model. Each component is modelled through an identical structure consisting of a non‐linear gain, expressed by a three‐parameter logistic function, and a linear transfer function with two reservoirs (fast/slow) in parallel having a constant partition between them. The model is applied on an information‐rich 5‐week data set. Extensive Monte Carlo realizations (more than two billion models) have identified a representative sample of behavioural models able to satisfy quite stringent fit criteria on both discharges and tracers. A descriptive statistical analysis of the behavioural parameter sets reveals significant differences between the components. In particular, the AS contribution is activated for higher catchment storages and shows a steep, almost threshold‐like, increase. The partition coefficient (fast/total) for the three components is ordered as DP>AS>GW. The fast constants of the three components have a similar order of magnitude, but also show DP>AS> GW. The slow time constant of the GW component is almost an order of magnitude higher than that of DP and AS. The latter are of similar magnitude and generate a highly non‐linear interflow component. Copyright © 2005 John Wiley & Sons, Ltd.